Genel Energy Miran Bina Bawi Limited v The Kurdistan Regional Government of Iraq [2026] EWHC 1003 (Comm)

[2026] EWHC 1003 (Comm)Case No CL-2025-000198
IN THE HIGH COURT OF JUSTICE
KING'S BENCH DIVISION
BUSINESS AND PROPERTY COURTS OF ENGLAND AND WALES
COMMERCIAL COURT
Venue Royal Courts of Justice, Rolls Building, Fetter Lane, London, EC4A 1NLDate 1 May 2026
Before
THE HONOURABLE MRS JUSTICE DIASIN AN ARBITRATION CLAIM
Between
GENEL ENERGY MIRAN BINA BAWI LIMITEDClaimantTHE KURDISTAN REGIONAL GOVERNMENT OF IRAQDefendant
Mr Charles Graham KC and Ms Jade Fowler (instructed by Quinn Emanuel Urquhart & Sullivan LLP) for ClaimantMr Ricky Diwan KC (instructed by Wilmer Cutler Pickering Hale and Dorr LLP) for DefendantHearing Hearing date: 16 April 2026
Approved JudgmentThis judgment was handed down remotely at 9:30am on 1 May 2026 by circulation to the parties or their representatives by e-mail and by release to the National Archives..............................

INTRODUCTION

[1]Section 63(3) of the Arbitration Act 1996 (the “Act”) provides as follows:
“The tribunal may determine by award the recoverable costs of the arbitration on such basis as it thinks fit. If it does so, it shall specify – (a) the basis on which it has acted, and (b) the items of recoverable costs and the amount referable to each.” (a) the basis on which it has acted, and (b) the items of recoverable costs and the amount referable to each.”
[2]The matter before me relates to the costs of a two-week arbitration which took place in February 2024 arising out of the termination of two Production Sharing Contracts relating to the exploitation of oil and gas reserves in the Kurdistan Region of Iraq. On 2 December 2024, a distinguished tribunal issued a Partial Final Award holding that the claimant in the arbitration (“KRG”) had validly terminated the Contracts and dismissing the counterclaim brought by the respondent (“GEMBBL”) for damages for repudiation.[3]In subsequent costs proceedings, KRG as the successful party presented a claim which included the truly eye-watering sum of over US$35.5 million in respect of legal and expert fees over the roughly 2½ year period of the arbitration. It might have been expected that a claim for costs of this magnitude would be supported by lengthy schedules breaking down those costs by reference to particular workstreams, fee earners, rates and hours allocated to each workstream. Far from it. The only information that KRG condescended to provide in support of its claim consisted of:(a) A total aggregate figure for each category of fee earner (not even each individual fee earner) over the entire life of the proceedings, with the total hours spent by each such category and the range of hourly rates applied within each category;(b) The total aggregate figure charged in each month by all fee earners;(c) The total aggregate fees and expenses charged in each month by each expert.[4]There was no attempt to allocate costs to particular workstreams and no information to show how much time was spent on any particular item of work by any individual fee earner. This meagre information was supplemented to a limited extent in reply costs submissions by a very high-level summary of the work carried out in each month, although this consisted of little more than bald headline descriptions such as “Legal advice regarding issues in dispute”, “correspondence with GEMBBL regarding issues in dispute”, “procedural issues”.[5]In those costs proceedings, GEMBBL argued that this was insufficient to enable the tribunal to make any reasoned award on costs which complied with the requirements of sections 63(3)(a) and (b) set out above (to which I shall refer as the “Specificity Provisions”)and that it should therefore decline to do so and instead direct that the assessment of costs be referred to the court under s.63(4) of the Act. The Tribunal rejected this submission, holding that the provisions of s.63(3) and following subsections of the Act had been displaced by the agreement of the parties to arbitrate under LCIA Rules 2020 (the “Rules”)and, specifically, by Article 28(3) which gave it the power to decide the amount of Legal Costs (as defined in the Rules) “on such reasonable basis as it thinks appropriate.” In that regard, it commented that “… there is no basis from the Claimant’s costs submissions on which the Tribunal can determine whether given items of the work in question were suitable to be performed by a partner rather than more junior fee earners, nor whether the volume of hours spent on those given items of work was reasonable, in the light of the tasks performed….” Nonetheless, despite the paucity of material available, it concluded that it had sufficient information to allow it to reach a reasoned decision on the costs claimed and that in any event, it was appropriate for it to undertake the assessment itself given its familiarity with the proceedings.[6]In the event, the Tribunal determined that the costs claimed were not reasonable and applied a blanket reduction of 20% to the legal fees and 50% to the fees of one of the experts resulting in a total costs award for these heads of something over US$26 million. The Tribunal also made an award in respect of KRG’s disbursements and Arbitration Costs (as defined in the Rules).[7]Under the Rules, the parties agreed to exclude any appeal on a question of law. However, GEMBBL now seeks to challenge the Tribunal’s award of costs in relation to legal and expert fees for serious irregularity under s.68(2)(b) of the Act on the grounds that it exceeded its powers by issuing an award which failed to comply with the requirement in s.63(3) to “specify the items of recoverable costs and the amount referable to each.” No issue is raised in relation to the award for disbursements or Arbitration Costs.[8]KRG applied for summary dismissal of the s.68 challenge on the basis that it had no real prospect of success. The matter came before Robin Knowles J on the papers, who dismissed the application on the basis that it was unsuitable for summary determination and required a full hearing. In the reasons he gave for his decision, he held that it was arguable that “itemisation” for the purposes of s.63(3) required more than was set out in the costs award, but that the question of how much more was required and the consequences in the context of a s.68 application should be the subject of oral argument.[9]That hearing now comes before me. RELEVANT STATUTORY AND CONTRACTUAL PROVISIONS[10]Determination of GEMBBL’s s.68 challenge requires consideration of the following provisions of the Act: “General Principles[1]The provisions of this Part are founded on the following principles, and shall be construed accordingly –(a) the object of arbitration is to obtain the fair resolution of disputes by an impartial tribunal without unnecessary delay or expense;(b) the parties should be free to agree how their disputes are resolved, subject only to such safeguards as are necessary in the public interest; … 4.(1) The mandatory provisions of this Part are listed in Schedule 1 andf have effect notwithstanding any agreement to the contrary. (2) The other provision of this Part (the “non-mandatory provisions”) allow the parties to make their own arrangements by agreement but provide rules which apply in the absence of agreement. (3) The parties may make such arrangements by agreeing to the application of institutional rules or providing any other means by which a matter may be decided. … COSTS OF THE ARBITRATION Cost of the arbitration 59.(1) References in this Part to the costs of the arbitration are to – (a) the arbitrators’ fees and expenses; (b) the fees and expenses of any arbitral institution concerned, and(c) the legal or other costs of the parties. … Award of costs 61.(1) The tribunal may make an award allocating the costs of the arbitration as between the parties, subject to any agreement of the parties. (2) Unless the parties otherwise agree, the tribunal shall award costs on the general principle that costs should follow the event except where it appears to the tribunal that in the circumstances this is not appropriate in relation to the whole or part of the costs. … The recoverable costs of the arbitration 63.(1) The parties are free to agree what costs of the arbitration are recoverable. (2) If or to the extent there is no such agreement, the following provisions apply. (3) The tribunal may determine by award the recoverable costs of the arbitration on such basis as it thinks fit. If it does so, it shall specify – (a) the basis on which it has acted, and (b) the items of recoverable costs and the amount referable to each. … (5) Unless the tribunal or the court determines otherwise – (a) the recoverable costs of the arbitration shall be determined on the basis that there shall be allowed a reasonable amount in respect of all costs reasonably incurred, and (b) any doubt as to whether costs were reasonably incurred or were reasonable in amount shall be resolved in favour of the paying party.” (a) the object of arbitration is to obtain the fair resolution of disputes by an impartial tribunal without unnecessary delay or expense; (b) the parties should be free to agree how their disputes are resolved, subject only to such safeguards as are necessary in the public interest; … COSTS OF THE ARBITRATION Cost of the arbitration Award of costs The recoverable costs of the arbitration (a) the basis on which it has acted, and (b) the items of recoverable costs and the amount referable to each. (a) the recoverable costs of the arbitration shall be determined on the basis that there shall be allowed a reasonable amount in respect of all costs reasonably incurred, and (b) any doubt as to whether costs were reasonably incurred or were reasonable in amount shall be resolved in favour of the paying party.”[11]As already stated, the parties in this case agreed to arbitrate under the auspices of the LCIA Rules. Article 28 of those Rules provides in relation to costs as follows:
“28.1 The costs of the arbitration other than the legal or other expenses incurred by the parties themselves (the “Arbitration Costs”) shall be determined by the LCIA Court in accordance with the Schedule of Costs. The parties shall be jointly and severally liable to the LCIA and the Arbitral Tribunal for such Arbitration Costs. 28.2 The Arbitral Tribunal shall specify by an order or award the amount of the Arbitration Costs determined by the LCIA Court. The Arbitral Tribunal shall decide the proportions in which the parties shall bear such Arbitration Costs (in the absence of a final settlement of the parties’ dispute regarding liability for such costs). If the Arbitral Tribunal has decided that all or any part of the Arbitration Costs shall be borne by a party other than a party which has already covered such costs by way of a payment to the LCIA under Article 24, the latter party shall have the right to recover the appropriate amount of Arbitration Costs from the former party. 28.3 The Arbitral Tribunal shall also have the power to decide by an order or award that all or part of the legal or other expenses incurred by a party (the “Legal Costs”) be paid by another party. The Arbitral Tribunal shall decide the amount of such Legal Costs on such reasonable basis as it thinks appropriate. The Arbitral Tribunal shall not be required to apply the rates or procedures for assessing such costs practised by any state court or other legal authority.”
THE ISSUES[12]In the light of these provisions, the following issues arise for determination:(a) Whether the power exercised by the Tribunal in determining legal and expert costs derived from s.63 of the Act or from Article 28 or from a combination of the two and, in particular, whether the Tribunal was bound to comply with the Specificity Provisions. This requires an examination of the relationship between the relevant provisions in the Act and the Rules.(b) If the Tribunal was bound to comply with the Specificity Provisions in exercising its power to determine costs, whether those provisions circumscribed and limited the existence and/or exercise of that power or were merely incidental or ancillary obligations relating to the manner in which it fell to be exercised. In other words, would non-compliance constitute an excess of power or merely an erroneous exercise of power?(c) Whether the Tribunal’s award in fact complied with the Specificity Provisions;(d) If not, and if there was an excess of power, whether substantial injustice had thereby been caused or would be caused to GEMBBL.[13]Mr Ricky Diwan KC, who appeared for KRG, submitted that the answer to issue (b) was decisive in this case, since a challenge under section 68 could not get off the ground at all unless GEMBBL could satisfy the court that non-compliance with the Specificity Provisions was capable of amounting to an excess of power. He said that it could not and, as will appear, I agree with him. It follows that GEMBBL’s section 68 challenge must fail in any event. Nonetheless, the other issues were fully argued and raise matters of potential importance for arbitration practice. I will therefore address them in the order set out above. DISCUSSION (1) Relationship between section 63 and Article 28[14]In his skeleton argument, Mr Diwan referred to the following principles of statutory interpretation which were not controversial:(a) Statutory interpretation requires an objective assessment of the meaning which a reasonable legislature would be seeking to convey by the statutory language used:(b) A statute must be construed as a whole, since a provision in one part of an act may cast light on the meaning of provisions elsewhere;(c) There is a presumption against construing a statute in a way which would produce an impossible, unworkable, impracticable, inconvenient, anomalous or absurd result: see Bennion, Bailey & Norbury on Statutory Interpretation (8th ed.)(2020) para 13.1ff.[15]Applying these principles, he emphasised that sections 1 and 4 of the Act enshrine and give primacy to the concept of party autonomy in respect of those arbitrations to which the Act applies, subject only to the specified mandatory provisions. Furthermore, section 4(3) expressly contemplates that parties may make their own arrangements regarding matters to be decided by agreeing to the application of institutional rules. The correct approach[16]There was a fundamental difference of approach between the parties in their analysis of the relationship between section 63 and Article 28. Mr Charles Graham KC, who presented the appeal for GEMBBL with Ms Jade Fowler, treated it first and foremost as a matter of statutory interpretation. As I understood it, his argument ran as follows:(a) The intention underlying section 63(1) is to permit the parties to make an agreement which differs from the default rules set out in section 63(3)ff. If and to the extent they make such an agreement, the default rules are ousted pro tanto. On the other hand, if the parties simply make an agreement which is the same as the default rules, then the question of ouster becomes meaningless and the default rules continue to apply. It was slightly unclear to me whether in those circumstances, it was his case that the parties’ agreement fell to be ignored entirely or whether it was to be treated as merged or subsumed into the default rules. I return to this at paragraph 21 below.(b) It follows that, on a proper interpretation of section 63(1), the default rules will only be ousted by an agreement which makes it unnecessary to look further at section 63(3), i.e., an agreement that renders the remainder of that section otiose.(c) This can be achieved in two ways: (i) By agreeing that the relevant matter should be taken outside the purview of the tribunal altogether. This could be, for example, by agreeing on a specific amount of costs for a particular item, or by agreeing a mechanism for the assessment of costs which requires no further reference to section 63, for example Article 28.1 of the Rules which stipulates that Arbitration Costs are to be determined by the LCIA Court in a specified manner. (ii) Alternatively, by agreeing an arrangement which is incompatible or inconsistent with the default rules or which specifically excludes them. In this context, Mr Graham argued that there must be a specific agreement to do something which is incompatible or inconsistent with the default rules. He referred me to a footnote to paragraph 5-112 of Russell on Arbitration (24th ed.) (2015). This suggests that mere agreement to a set of institutional rules which confer a general power on the tribunal to decide or order something in circumstances where that general power could encompass an alternative approach to that in the Act, might not of itself be sufficiently specific to oust any non-mandatory provisions of the Act. He also drew attention to a passage in the DAC Report accompanying the draft Bill that became the Act where it is stated in relation to what became section 4 of the Act that the non-mandatory provisions of the Act can be changed or substituted by the parties “and exist as “fall-back” rules that will apply if the parties do not make any such change or substitution, or do not provide for the particular matter in question.” This, he submitted, supported the proposition it is only agreement which provides for something different to that contained in the default rules which will achieve their ouster. He further drew an analogy with sections 15, 16, 18 and 23 of the Act which all used similar language to section 63(1) in providing that the parties should be “free to agree” on specific matters and where such agreement would render the applicable default rules unnecessary. This was to be contrasted with the wording “unless otherwise agreed” used in other parts of the Act which he said operated as a general “get-out” without it being necessary to reach any specific alternative agreement.(d) In the present case, the second sentence of Article 28.3 addressing the assessment of costs was to all intents and purposes identical to the default rules in sections 63(3) and (5).(e) Section 63(3) has two limbs: a power to determine the recoverable costs of the arbitration on such basis as the tribunal thinks fit, and the Specificity Provisions in sections 63(3)(a) and (b). The Specificity Provisions are couched in mandatory language indicating a clear statutory intention that they should apply wherever a tribunal exercised its power to assess costs. As Mr Graham colourfully put it, the second limb containing the Specificity Provisions “had got its hooks … into the first sentence so firmly that the one cannot be looked at without the other.” Parliament must therefore be taken to have intended that a tribunal should not award costs without any reasoning or substantiation. In so far, therefore, as Article 28.3 simply replicated the first limb of section 63(3), that necessarily attracted the application of the Specificity Provisions.[17]Thus, on Mr Graham’s approach there is an inherent limitation as a matter of statutory interpretation on the nature and type of agreement falling within section 63(1) which is capable of ousting the default rules.[18]By contrast, Mr Diwan’s approach was to treat the matter primarily as one of construing the parties’ agreement. He submitted that there is no basis for ignoring the agreement made by the parties. To do so would be inimical to the concept of party autonomy enshrined in the Act. Party autonomy likewise means that - subject only to the application of any mandatory provision or considerations of public interest - the ability of the parties to reach their own bespoke agreement should not be constrained.[19]The starting point is therefore always to construe the parties’ agreement in order to assess whether and to what extent it has the effect of ousting the default rules. Mr Diwan accepted that, in principle, the agreement of the parties might leave a lacuna which would need to be filled by the default rules. Indeed, this is implicit in section 63(2) which applied the default rules “[i]f or to the extent there is no such agreement…” However, he submitted that there was no such lacuna in this case. (I return to this below.)[20]I agree that it is necessary to interpret section 63(1) in order to determine what needs to be covered by any party agreement in order to oust the default rules. Nonetheless, I cannot accept the premise of Mr Graham’s argument that there is any conceptual limit on the nature or type of agreement which can be effective for that purpose. There is nothing in the statutory language to suggest any such limitation. Indeed, I consider that it would be unduly prescriptive and contrary to the entire notion of party autonomy to provide on the one hand in sections 63(1) and (2) in apparently unqualified language that the parties are free to agree what costs of the arbitration should be recoverable and that such agreement will have the effect of ousting the default rules, but then on the other to impose an implicit qualification on the type of agreement which will have that effect. I agree with Mr Diwan that this is an unnecessarily complicated and strained approach which is not mandated by the language of the Act and which is contrary to the clear intention of the Act to provide a simple, straightforward framework for arbitration which is easy to apply and leaves maximum flexibility for the parties to agree their own procedures subject only to the mandatory provisions (of which section 63 is not one).[21]Nor do I accept (to the extent it was suggested) that an agreement made by the parties will be wholly ineffective to oust the default rules unless it is inconsistent with them. If the parties are free to agree that legal costs can be assessed by a tribunal on a basis which is inconsistent with or contradicts the default rules, I can see no reason in logic or principle why they cannot reach an agreement that the tribunal should assess legal costs on a basis which partially or entirely replicates what is already in the default rules. Particularly where parties sign up to institutional rules, this seems to me quite likely to be the case. Like Mr Diwan, I can see no principled basis on which the parties’ agreement can simply be ignored, even if it adds nothing to the default rules. On the contrary, the underlying ethos of the Act is to give primacy to such agreements.[22]In my judgement, therefore, the correct approach is entirely straightforward. First, it is necessary to determine what section 63(1) requires an agreement to cover in order for the default rules to be ousted. Then it is simply a question of construing the agreement to see whether it does cover those matters or, if not, to what extent it leaves a lacuna to be filled. Section 63(1)[23]By virtue of section 63(2), the default rules are ousted if and to the extent that no agreement is reached as referred to in section 63(1). Section 63(1) refers to an agreement as to “what costs of the arbitration are recoverable.”[24]Mr Diwan submitted that in the context of ouster, this can only mean an agreement as to those matters which would otherwise be covered by the default rules, viz:(i) the heads of cost defined as “costs of the arbitration” in section 59;(ii) the basis of assessment; and(iii) matters to be specified in the award. In the event, I did not understand this to be contentious and it makes practical sense since the rationale for ousting the default rules is because the parties have made express arrangements for matters which they would otherwise cover. Thus Mr Graham accepted that an agreement within section 63(1) could cover the recoverability of identified heads of cost, or items of costs, or a mechanism by which costs were to be determined.[25]Both sides further accepted that the parties could, of course, go much further in any agreement than simply the matters covered by the default rules, for example by agreeing specific amounts for specific items. However, for the purpose of seeing whether the default rules have been ousted, the only enquiry is whether the parties agreed to cover the matters which would otherwise be provided for. Article 28.3[26]There can be no real dispute that Article 28 contains an agreement that all the heads of cost identified in section 59 should be recoverable. Thus, the costs set out in sections 59(1)(a) and (b) (the fees and expenses of the arbitrators and of the LCIA itself) are both included in the definition of Arbitration Costs in the Rules and so covered by the provisions of Articles 28.1 and 28.2. Legal Costs as defined in the Rules are addressed by Article 28.3.[27]Article 28 also covers the basis on which both Arbitration and Legal Costs should be assessed. Arbitration Costs fall to be determined by the LCIA Court in accordance with its published Schedule of Costs. The amount of Legal Costs is to be decided by the tribunal “on such reasonable basis as it thinks appropriate.”[28]The battleground was as to whether Article 28 made provision for what, if anything, needed to be specified in the award in relation to recoverable costs. As to this, Article 28.2 provides that the tribunal should specify in an order or award “the amount of Arbitration Costs determined by the LCIA Court.” In relation to Legal Costs, Article 28.3 simply provides that the tribunal “shall decide the amount of such Legal Costs…”[29]Mr Diwan submitted that on its proper construction, Article 28.3 only required the Tribunal to state the amount of the Legal Costs awarded in its order or award. “Legal Costs” were explicitly defined in the Rules and there was no need or requirement to break them down further.[30]On that basis, he argued that Article 28.3 represented a complete package and implicitly excluded or was inconsistent with the Specificity Provisions. There was therefore no need to resort further to the default rules because there was no lacuna to be filled.[31]In response, Mr Graham returned to the footnote from Russell on Arbitration referred to in paragraph 16(c)(ii) above in support of his argument that signing up to a set of institutional rules is not necessarily a sufficient agreement to oust the default rules under section 63. I can certainly see force in the argument that mere agreement to institutional rules does not per se oust all non-mandatory provisions of the Act and, further, that a particular institutional rule may not in fact amount to an agreement on a relevant matter.[32]However, that is simply to beg the question as to what, on its true construction, Article 28.3 means, and whether it expressly or impliedly covers the matters provided for in the default rules. I accept Mr Diwan’s submissions that it is not possible to have two different regimes applying in parallel: either Article 28.3 covers the ground completely (in which case the default rules fall away), or it does not cover the ground at all (in which case they apply in toto), or it partially covers the ground (in which case they apply to fill the gap but no further).[33]Mr Graham argued that Article 28.3 does not contain anything which on its proper construction is effective to displace the Specificity Provisions. On the contrary, it is entirely consistent with the first limb of section 63(3) and so does not evince any objective intention to dispense with the essential and mandatory adjuncts contained in the Specificity Provisions in the second limb. Indeed, it does not make any provision at all for what needed to be stated in any award or order.[34]In response, Mr Diwan emphasised that although Article 28.3 is similar in its effect to the provisions of section 63, it is not identical. For example:(a) Costs are categorised and defined differently under the Act and under the Rules. This, he suggested, was an indication that the parties intended to apply a completely different and free-standing regime for the assessment of costs to that applicable under the default rules.(b) Under Article 28.3, the power conferred on the tribunal is to award costs “on such reasonable basis as it thinks appropriate”. Section 63(3), by contrast, provides that the tribunal shall determine costs “on such basis as it thinks fit”, but with the added proviso in section 63(5) that, unless it decides otherwise, it must award a reasonable amount in respect of costs reasonably incurred.(c) The power to determine costs in section 63(3) is permissive, whereas once the tribunal has determined the incidence of costs under the first sentence of Article 28.3, it is mandatory to determine the amount of those costs.[35]He further drew attention to the express provision in the final sentence of Article 28.3 which made clear that the tribunal is not constrained in assessing costs by any court practices or procedures.[36]In short, Mr Diwan submitted that Article 28.3 provides for an entirely free-standing regime which depends on a different categorisation of costs to that in the Act and which confers on a tribunal a discretion to assess costs on such reasonable basis as it thinks appropriate without any requirement to specify in its award the basis on which it has acted or to itemise the costs in any way beyond stating the amount awarded for Arbitration Costs (as determined by the LCIA Court) and the amount determined for Legal Costs, those being the only two categories of cost contemplated under the Rules. This regime is capable of standing on its own and there was therefore no lacuna which could only be filled by reference to the default rules. On the contrary, to apply those – non-mandatory – default rules by effectively bolting them on to Article 28.3 would actually be fundamentally inconsistent with the parties’ agreement that the Tribunal should have a discretion to assess costs on such reasonable basis as it thought appropriate because that would, by definition, constrain the exercise of such discretion.[37]I have considerable sympathy for GEMBBL. Even taking into account the amounts at stake, the quantum of legal and expert costs claimed by KRG was staggering for arbitration proceedings which lasted no more than 2½ years in total and culminated in a mere two-week hearing. It is profoundly unsatisfactory that those costs were not vouched in a manner which would have allowed GEMBBL to make properly informed submissions as to their reasonableness. Best practice would undoubtedly have been to compile a schedule in a format akin to that provided by GEMBBL in relation to its bifurcation costs. That said, I bear in mind that by agreeing to arbitrate in England under the Act, the parties accepted that there could be no appeal on a question of fact (such as the reasonableness of the costs awarded) and by agreeing specifically to arbitration under LCIA Rules, they also accepted that there could be no appeal for error of law. In other words, the parties deliberately and consciously elected for a form of dispute resolution which precluded any complaint that the Tribunal reached the wrong decision.[38]In those circumstances, and even accepting (as I do) that mere agreement to a set of institutional rules does not per se exclude all non-mandatory provisions of the Act, I would have held (had it been necessary to do so) that Article 28.3 provided a mechanism for the assessment of Legal Costs which was a complete package sufficient to exclude the operation of the default rules under section 63 in their entirety.[39]I regard this conclusion as being entirely consistent with section 4(3) of the Act and I reject Mr Graham’s suggestion that it would be uncommercial. On the contrary, I suspect that many parties would be surprised to learn that what they supposed to be a watertight package of institutional rules turned out to be no more than a leaky sieve which required ad hoc patching by bringing in odd bits and pieces from the Act. Indeed, requiring a particular level of specificity in institutional rules before they can be regarded as capable of ousting the default rules in the Act, would require a minute parsing of the rules in question. This would entail an unnecessarily cumbersome and complex exercise. There is also force in Mr Diwan’s submission that the LCIA Rules are international rules and that the logical consequence of Mr Graham’s argument is that their effect (and, indeed, that of any other institutional rules) might be different depending on the seat of the arbitration. I regard this as being the antithesis of what commercial parties anticipate or expect when they agree to a London seated arbitration under institutional rules. By signing up to a London seat, they also accept the principle of party autonomy. There are sufficient safeguards in the Act in the form of the mandatory provisions and to accede to Mr Graham’s arguments would be to confer on section 63 a quasi-mandatory status that Parliament did not see fit to accord it directly. Moreover, it is always open to parties to incorporate the default rules from section 63 into their agreement expressly if they feel that the institutional rules do not adequately cover the position. Alternatively, they could expressly exclude Article 28.3 in which case the default rules would apply in any event. If they do not adopt either course, then there is nothing in my judgment to warrant construing Article 28.3 as anything other than a complete free-standing agreement for the determination of Legal Costs which is effective to exclude the default rules in section 63. (2) Excess or erroneous exerciseof power[40]It follows from my conclusions above, that the Specificity Provisions did not apply in this case. If I am wrong about that and they were applicable (whether because the default rules applied in their entirety or because the Specificity Provisions applied alongside and in addition to Article 28.3), it is necessary to consider whether non-compliance amounted to an excess of power challengeable in principle under section 68(2)(b), or merely an erroneous exercise of power challengeable, if at all, only for error of law, which under the Rules is of course excluded. Mr Graham accepted that his challenge under section 68 can only succeed if he can demonstrate that non-compliance with the Specificity Provisions amounted to an excess of power.[41]The parties agreed that the starting point was the decision of the House of Lords in Lesotho Highlands Development Authority v Impregilo SpA, [2005] UKHL 43; [2006] 1 AC 221 and, in particular, the speech of Lord Steyn at [23]-[34]. As is well-known, that case concerned a challenge to a decision of an arbitration tribunal to make an award of damages in a foreign currency converted from the contractual currency of account at the rate specified in the contract notwithstanding a subsequent dramatic collapse in the value of that currency. Lord Steyn’s speech repays setting out at some length: “23 … I will now assume that the tribunal committed an error of law. In fact Lord Steyn would have held that the tribunal in that case did not make any error in their award. What he thereafter said was on the basis that he was wrong about that. What he said was expressly agreed with by a majority of the House which held that there had indeed been an error of law. Lord Phillips would have gone further and held that the error amounted to an excess of power, but he was a lone voice in that regard. That error of law could have taken more than one form. The judge and the Court of Appeal approached the matter on the basis that the tribunal erred in the interpretation of the underlying contract. Another possibility is that the tribunal misinterpreted its power, under section 48(4) to express the award in any currency. Let me approach the matter on the basis that there was a mistake by the tribunal in one of these forms. Whichever, is the case, the highest the case can be put is that the tribunal committed an error of law.[24]But the issue was whether the tribunal “exceeded its powers” within the meaning of section 68(2)(b). This required the courts below to address the question whether the tribunal purported to exercise a power which it did not have or whether it erroneously exercised a power that it did have. If it is merely a case of erroneous exercise of power vesting in the tribunal no excess of power under section 68(2)(b) is involved. Once the matter is approached correctly, it is clear that at the highest in the present case, on the currency point, there was no more than an erroneous exercise of the power available under section 48(4). The jurisdictional challenge must therefore fail. … 26… A major purpose of the new [Arbitration] Act was to reduce drastically the extent of intervention of courts in the arbitral process.[27]The legislative technique adopted to achieve this purpose was spelled out explicitly in the Report on the Arbitration Bill and in particular in discussion of clause 68, which became section 68 of the 1996 Act. The DAC observed about clause 68 that it “is really designed as a long stop, only available in extreme cases where the tribunal has gone so wrong in its conduct of the arbitration that justice calls out for it to be corrected”: p 58, para 280. On the other hand, the DAC recommended adoption of “the internationally accepted view that the court should be able to correct serious failure to comply with the ‘due process’ of arbitral proceedings: cf article 34 of the Model Law:” p 59, para 282. The ethos of the DAC report was that parties are entitled to a fair hearing leading to an impartial adjudication. But the idea that section 68 contemplated an adjudication which arrives at the “right” conclusion would have been wholly out of place in these recommendations. The DAC report was the matrix of the Parliamentary debates. …[29]It will be observed that the list of irregularities under section 68 may be divided into those which affect the arbitral procedure, and those which affect the award. But nowhere in section 68 is there any hint that a failure by the tribunal to arrive at the “correct decision” could afford a ground for challenge under section 68. On the other hand, section 68 has a meaningful role to play. An example of an excess of power under section 68(2)(b) may be where, in conflict with an agreement in writing of the parties under section 37, the tribunal appointed an expert to report to it. At the hearing of the appeal my noble and learned friend, Lord Phillips of Worth Matravers MR, also gave the example where an arbitration agreement expressly permitted only the award of simple interest and the arbitrators in disregard of the agreement awarded compound interest. There is a close affinity between section 68(2)(b) and section 68(2)(e). The latter provision deals with the position when an arbitral institution vested by the parties with powers in relation to the proceedings or an award exceeds its powers. The institution would exceed its power of appointment by appointing a tribunal of three persons where the arbitration agreement specified a sole arbitrator. …[31]By its very terms, section 68(2)(b) assumes that the tribunal acted within its substantive jurisdiction. It is aimed at the tribunal exceeding its powers under the arbitration agreement, terms of reference or the 1996 Act. Section 68(2)(b) does not permit a challenge on the ground that the tribunal arrived at a wrong conclusion as a matter of law or fact. It is not apt to cover a mere error of law. This view is reinforced if one takes into account that a mistake in interpreting the contract is the paradigm of a “question of law” which may in the circumstances specified in section 69 be appealed unless the parties have excluded that right by agreement. In cases where the right of appeal has by agreement, sanctioned by the Act, been excluded, it would be curious to allow a challenge under section 68(2)(b) to be based on a mistaken interpretation of the underlying contract. Moreover, it would be strange where there is no exclusion agreement, to allow parallel challenges under section 68(2)(b) and section 69.[32]In order to decide whether section 68(2)(b) is engaged it will be necessary to focus intensely on the particular power under an arbitration agreement, the terms of reference, or the 1996 Act which is involved, judged in all the circumstances of the case. In making this general observation is must always be borne in mind that the erroneous exercise of an available power cannot by itself amount to an excess of power.

(b) …

[34]I am glad to have arrived at this conclusion. It is consistent with the legislative purpose of the 1996 Act, which is intended to promote one-stop adjudication. If the contrary view of the Court of Appeal had prevailed, it would have opened up many opportunities for challenging awards on the basis that the tribunal exceeded its powers in ruling on the currency of the award. Such decisions are an everyday occurrence in the arbitral world. If the view of the Court of Appeal had been upheld, a very serious defect in the machinery of the 1996 Act would have been revealed. The fact that this case has been before courts at three levels and that enforcement of the award has been delayed for more than three years reinforces the importance of the point.”[42]In the light of these passages, both parties invited me to bring an intense focus to bear on section 63(3) although the lenses which each of them suggested I should apply to my eye were rather different.[43]Mr Diwan relied on the repeated emphasis in Lesotho that challenges for excess of power are not designed to provide a remedy if the complaint is simply that the tribunal reached the wrong answer. To the contrary, the Act was intended to limit the scope of possible challenges such that section 68 is a long-stop for exceptional cases where there has been a failure to comply with due process in the conduct of the arbitration. It is not concerned with errors of law or fact and therefore cannot be used as a vehicle for challenging errors in statutory construction, still less an assessment of reasonableness. Furthermore, the logic of Lord Steyn’s reasoning in Lesotho is that an excess of power only occurs if a tribunal purports to exercise a power which it does not have. Accordingly, section 68(2)(b) does not cover errors or mistakes made by a tribunal in the exercise of a power which it indisputably does have.[44]Mr Diwan submitted that, having regard to the high hurdle posed by section 68(2)(b), the relevant question to ask is whether the tribunal had no power at all to do what it did, or whether it simply went wrong in the exercise of its power. He referred to Essar Oilfields Services Ltd v Norscot Rig Management PVT Ltd, [2016] EWHC 2361 (Comm); [2017] Bus LR 227 as a good example of a case where Judge Waksman QC (as he then was) deprecated attempts to dress up an appeal on a question of law as an excess of power by arguing that the relevant power could only be exercised in a particular way. In that case, the tribunal had an undoubted power to award costs either under section 59(1)(c) of the Act or under the ICC Rules which also applied. The tribunal made an award in respect of the costs of third party funding and the paying party argued that it was not within its power to do so under either the Act or the ICC Rules.[45]Section 63(3) as such was not in issue but the following comments of

Judge Waksman are nonetheless relevant:

“41, As Lord Steyn noted, in order to see if what the arbitrator did fell within section 68(2)(b) as being in excess of his powers or whether it was no more than an erroneous exercise of a power that he did have, it is necessary to focus “intensely on the power concerned”
. In my judgment, the relevant power here is the undoubted power to award costs. If the arbitrator fell into error, it was an error as to the scope of such costs by reason of his allegedly erroneous interpretation of section 69(1)(c) and article 31(1).[42]I accept that, if one characterised the relevant power as being the power to order that one side pays the other side’s costs of obtaining litigation funding, or conversely, the power to order by way of costs such sums which do not include the costs of litigation funding, one could say as a matter of language that he was exercising a power that he did not have. But, if that was the correct approach, one could re-describe many, if not all, errors of law in that way. Indeed, an erroneous exercise of power itself could in theory almost always be re-described as an excess of power. However, according to the Lesotho Highlands Development Authority case, there is a real and vital distinction to be made between the two. In my judgment, to characterise the arbitrator’s error here in that way would be wholly unrealistic and artificial, and it goes against the grain of the strict and narrow confines in which section 68 is to operate.”[46]Basing himself on these propositions, Mr Diwan argued that in the present case, the Tribunal undoubtedly had the power to determine and award legal and expert costs under either section 63 or Article 28.3. However, the question of what evidence was required for that purpose and what needed to be specified about it in the award was nothing to do with the existence of the power itself but simply related to the manner of its exercise. Moreover, once it is accepted that the parties have autonomy to make alternative arrangements which may be inconsistent with the Specificity Provisions, it is difficult, if not impossible, to regard failure to comply with those (non-mandatory) provisions as so fundamental that non-compliance goes to the very existence of the power.[47]He also made the powerful point that if GEMBBL’s arguments were accepted, it would open up the prospect of post-award costs challenges under section 68(2)(b) in virtually every case since it would almost always be possible to argue that the tribunal had not sufficiently itemised the costs awards and had therefore exceeded its powers. He submitted that such an unpalatable outcome could only be avoided if the court laid down a prescriptive rule as to what the Specificity Provisions required by way of itemisation, which was capable of application in every case. This, he suggested, was a chimera given the infinite variability in arbitration procedures and the nature of the costs which might be incurred.[48]In response, Mr Graham reiterated that it was impossible to sever the Specificity Provisions from the power contained in the first limb of section 63(3). They were an inseparable part of that power which could not therefore be exercised save in compliance with them. Accordingly, they were properly to be regarded as delimiting the power itself. By contrast, the power in issue in Lesotho and Essar was entirely general and did not have any such adjuncts.[49]He accepted that the nature of the itemisation required would necessarily vary from case to case depending on the nature and complexity of the proceedings, the amount at stake, and the manner in which it was presented and that it was impossible to be prescriptive about it. He pointed to GEMBBL’s own bifurcation costs schedule as a paradigm of what a proper costs schedule should look like, and suggested that in 99 cases out of 100 there would not be a problem because the parties would either provide something similar or the tribunal would demand it.[50]Mr Graham further submitted that holding the Tribunal’s power to be circumscribed in this way would do no damage to party autonomy or to the reputation of London as a global centre for arbitration. He suggested that Mr Diwan’s submissions to the contrary were mere scaremongering which were not supported by any evidence. On the contrary, he argued, arbitrating parties have a legitimate interest in ensuring that they are only required to pay reasonable and proportionate costs and that they are provided with sufficient information to allow them to make meaningful submissions in that regard. If anything, therefore, they were more likely to be deterred from arbitrating in London if faced with a regime where a tribunal could make a wholly unspecific award of costs unsupported by any material which allowed transparent interrogation. It can be assumed that Parliament was of a similar view otherwise it would not have included the Specificity Provisions in section 63 to start with. An incidental beneficial effect of circumscribing a tribunal’s power in this way would be to discourage excessive and profligate expenditure on legal costs.[51]These were powerful points well made, but even taking them at their highest, I have little hesitation in preferring the submissions of Mr Diwan on this point. The fact remains that the Specificity Provisions are essentially adjectival and I could not really see an answer to Mr Diwan’s submission that if the parties were free to confer on a tribunal the power to act on a basis inconsistent with the Specificity Provisions, the latter could hardly be said to be fundamental to the existence of that power – a point only underlined by the fact that Parliament could have designated them as mandatory provisions but chose not to.[52]The concern that a contrary conclusion would open the floodgates to a deluge of costs challenges is neither illusory, nor a prospect that the court can view with equanimity given the clear objective of the Act to provide for one-stop adjudication and to limit the scope for challenges to arbitral decisions. For all these reasons, I am satisfied that even if the Tribunal was wrong in making an award in the manner and form in which it did, there was no excess of power but at most an erroneous exercise of that power. In an appropriate case, such an error might have been challengeable by way of an appeal under section 69 but it is GEMBBL’s misfortune in this case to have agreed to institutional rules which exclude any such possibility.[53]On this ground alone, therefore, the section 68 challenge must fail. (3) Compliance with the Specificity Provisions[54]The Tribunal clearly set out in paragraph 49 of its award the basis on which it acted in assessing legal and expert costs (even though, on KRG’s case, it was not obliged to do so). This explained that:
“The Tribunal must assess the Claimant’s costs on the ordinary basis in international arbitration – that is, whether the Claimant’s costs are reasonable and proportionate. The Tribunal accepts the general principle put forward by the Respondent that the Claimant must establish that its costs are reasonable, in the sense that it was necessary for the costs to be incurred, and that the expenditure referable to the workin question was reasonable in amount. This involved showing that the expenditure on that item was not disproportionate, given the work involved and its relative importance to the outcome of the dispute...”
[55]GEMBBL’s complaint is limited to its submission that in relation to the legal and expert costs the Tribunal did not sufficiently “specify the items of recoverable costs and the amount referable to each.” Mr Graham submitted that an award which simply awarded amounts for “Legal costs” and “Experts’ costs” was not good enough to comply with the Specificity Provisions. An “item of recoverable costs” for this purpose denoted(i) an item of work carried out by the lawyers or experts as appropriate;(ii) in respect of which costs had been incurred;(iii) which costs were both incurred reasonably and reasonable in amount. For such a determination to be carried out, the items in question needed to be sufficiently particularised so that a tribunal could make an assessment of each of these matters. As he accepted, it was impossible to be prescriptive about what would constitute an “item of work” in any particular case since this would vary depending on the circumstances. However, as a bare minimum each item specified should identify the particular fee earners involved, their seniority and the hours and rates charged by each of them for that particular item. Having carried out the necessary assessment, the tribunal should then specify in its award each item of work in respect of which it was awarding costs and set out the figure awarded. In most cases, it would be sufficient to refer to the receiving party’s costs schedule but where there was insufficient information to make a proper assessment, the tribunal should either request further particularisation or direct the parties to refer the matter to court under section 63(4).[56]Mr Diwan’s response was to point out that an “item of work done” is nowhere referred to in the Act and moreover is not the same as an “item of recoverable costs” The latter is the only concept referred to in the Specificity Provisions and this must refer back to the first limb of section 63(3), which in turn refers back to section 63(1), which in turn leads back to section 59 where “costs of the arbitration” are defined. Thus, he submitted that an “item of recoverable costs” meant no more or less than the headline categories set out in section 59. This was clear, simple and easy to apply in practice. By contrast, Mr Graham’s approach was impractical in the absence of any clear guidance as to what was required by way of itemisation, something which both parties agreed was impossible to provide a priori.[57]I have given some idea above of the nature of the costs claim advanced by the KRG and the extent to which it was (or was not) substantiated. Suffice it to say that I found it wholly unimpressive. If KRG was willing and able to spend over US$35 million on a two-week arbitration, it would hardly have been an excessive burden to require it to vouch its costs properly. As it is, I have considerable sympathy with GEMBBL’s complaint that the scant information furnished by KRG’s solicitors was completely inadequate to permit any proper examination of the reasonableness of the costs and so denied GEMBBL any opportunity to make informed submissions on the question.[58]The Tribunal was also clearly uneasy about the lack of particularisation but nonetheless felt that in view of its familiarity with the case it was in a position to do substantial justice. I emphasise that I am not here concerned with whether I would have taken the same view. The only question for me is whether in proceeding on that basis, the Tribunal complied with its obligation (on the hypothesis that the Specificity Provisions applied) to set out the items of recoverable costs.[59]With some hesitation, I have ultimately concluded that Mr Diwan is right about this as well as a matter of statutory interpretation and that “items of recoverable costs” in section 63(3)(b) refers back to the headline categories in section 59. I am fortified in this conclusion by the sheer impracticality of any other solution given that an “item of work done” is an impossibly elastic concept which will differ in every case. Moreover, there is scope for almost infinite disagreement as to appropriate “items of work” and a tribunal cannot know in advance what it needs to do in order to stay within its powers unless it takes pre-emptive precautions in agreeing with the parties what is required at an early stage, if only to ensure that appropriate records are compiled. Nothing of that nature appears to be contemplated by either the Act or the Rules.[60]However, the decisive point in my judgment is that I am considering this point on the basis that (contrary to my conclusions above), compliance with the Specificity Provisions was an essential pre-requisite to the proper exercise of the Tribunal’s power to determine costs. Given that premise, I consider it unsatisfactory that the exercise of a general power to determine costs should be circumscribed depending on the vagaries of the way in which the parties choose to present their respective cases and incur costs. This would mean, in effect, that a different power fell to be exercised in every single arbitration, the precise ambit of which could not be known until costs fell to be assessed – unless it had been presciently agreed in advance. I cannot accept that this was the intention of either Parliament or the parties. Rather, in my judgment, that intention was to confer on a tribunal a general power to assess costs which should be the same in every case to which the default rules applied.[61]On that basis, there was no failure by the Tribunal to comply with the Specificity Provisions and the challenge fails for this reason as well. (4) Substantial injustice[62]A challenge under section 68 requires the serious irregularity in question to be one which the court considers has caused or will cause substantial injustice to the applicant. As appears from RAV Bahamas v Therapy Beach Club, [2021] UKPC 8; [2021] AC 907 at [34], the test for substantial injustice is whether “had the irregularity not occurred, the outcome of the arbitration might well have been different… It is not necessary to show that the outcome would necessarily or even probably be different… In general, there will, however, be no substantial injustice if it can be shown that the outcome of the arbitration would have been the same regardless of the irregularity…”[63]Applying this test, if I had found that the Tribunal exceeded its powers by failing sufficiently to itemise the recoverable costs, I would have had no difficulty in concluding that the outcome of the costs award might well have been different and that a substantial injustice had been established in the sense set out above. Mr Diwan argued that there was nothing to suggest that the Tribunal would have reached any different conclusion even if full particularisation had been provided. However, that is pure speculation and I consider that it might, at the very least, have reached a different conclusion more favourable to GEMBBL. If KRG wished to rebut that suggestion, then it should have provided the court with the necessary material to substantiate its case. Indeed, it might be said that the very fact that we have no way of knowing what the Tribunal might have decided had it been provided with proper particularisation is a substantial injustice in itself. I also agree with Mr Graham that Essar (supra) provides support for the submission that there is substantial injustice in the very fact that the Tribunal awarded costs in a manner which it had no power to adopt.[64]I recognise, however, that my conclusion in GEMBBL’s favour on this point is likely to be cold comfort. (5) Appropriate relief[65]The only other matter debated briefly before me related to the nature of the order that I should make had I allowed the challenge. As to this, GEMBBL’s claimed relief asked for the question of costs to be referred to the court on the basis that the Tribunal could not realistically be expected to reach any different conclusion in the absence of further material. However, as I indicated during the course of argument, the obviously sensible course would be to remit the matter to the Tribunal with directions as to the further information that should be provided by KRG to substantiate its costs. It seemed to me that, in view of the Tribunal’s familiarity with the case and the nature of the arguments, it would make no sense at all to send the matter to the court which would have to approach it from scratch. Neither party dissented from this approach although, in the event, it does not arise. CONCLUSION[66]For all these reasons, GEMBBL’s application under section 68(2)(b) is dismissed.