[2022] SC (Bda) 30 Com (6 May 2022) In The Supreme Court of Bermuda

CIVIL JURISDICTION

COMMERCIAL COURT

2020: No. 369

BETWEEN:

AEOLUS RE LTD

(in respect of its KEYSTONE PF SEGREGATED ACCOUNT) Plaintiff -and-

(1) CS ILS SICAV-SIF

(in respect of CREDIT SUISSE (LUX) IRIS BALANCED FUND)

(2) CS IRIS ALHC FUND LIMITED

(3) CS IRIS C FUND LIMITED

(4) MANAGED INVESTMENTS PCC LIMITED

(on behalf of its IRIS BALANCED CELL)

(5) MANAGED IVESTMENTS PCC LIMITED

(on behalf of its IRIS ENHANCED CELL) (6) IRIS DYNAMIC SPC on behalf of its IRIS POST-EVENT FUND SP

(7) ALPHA Z ILS FUND LIMITED

Defendants Before: Hon. Chief Justice Hargun Appearances: Peter MacDonald Eggers QC, Mark Chudleigh and Laura Williamson of Kennedys Chudleigh Ltd, for the Plaintiff Timothy Howe QC, Peter Dunlop and Izabella Arnold of Walkers (Bermuda) Limited, for the Defendants 1

Dates of Hearing: Date of Judgment: 14-18 March 2022 6 May 2022

JUDGMENT

Contractual discretion in Industry Loss Warranty swaps; proper meaning and scope of determinations made “in good faith and a commercially reasonable manner”

HARGUN CJ

A. Introduction

1. In these proceedings the Plaintiff, Aeolus Re Ltd, claims damages for alleged breach of contract against each of the Defendants, and/or a declaration that it is entitled to be paid the sum of US$20 million deposited by the Defendants with Bank of New York Mellon (“Trust Accounts”) by way of collateral, in respect of certain Swap Confirmations that were concluded between the parties in July 2018.

2. The Defendants (funds that Credit Suisse Insurance Linked Strategies Ltd., a member of the Credit Suisse group of companies, manages) counterclaim seeking declarations that they are entitled to receive back their collateral from the Trust Accounts and damages for breach of contract against the Plaintiff for its wrongful refusal in and after early July 2020 (upon the automatic termination of the Swap Confirmations in accordance with their terms) to procure the release of the collateral to the Defendants despite having been given contractual notice to do so.

3. These proceedings concern seven Industry Loss Warranty (or "ILW") swaps entered into between the Plaintiff and each of the Defendants (together "the Swaps") on materially identical terms. Each of the Swaps was entered into upon the terms and conditions contained in a Swap Confirmation which supplemented, formed part of and was subject to the 2002 ISDA Master Agreement between the parties, and which also expressly incorporated by reference the 2006 ISDA Definitions. It is common ground that ILW 2

4. B. 5. Swaps such as these are not indemnity contracts; unlike traditional reinsurance they do not require the buyer to have suffered any loss, instead they are financial instruments which enable parties in effect to speculate on the occurrence and outcome of future natural catastrophe events including in order to hedge their underlying exposure(s). In summary, under the terms of the Swaps, the Plaintiff was the buyer from the Defendants in the event of the occurrence of a specified type of natural catastrophe in a designated geographical region during the defined Risk Period (1 June 2018 to 31 December 2018), for which the estimated losses incurred by the insurance and reinsurance industry as a whole arising therefrom, as reported by a particular nominated entity (known as a "Report Publisher"), exceeded a specified threshold amount within a defined period (and prior to the termination of the Swaps); in that event the Plaintiff would be entitled (subject to serving a Payment Notice) to receive payment from the Defendants of the sum of US$20 million per catastrophe event up to a maximum of two events (US$40 million in total). The Defendants were required to (and did) put up collateral for a potential future payment of such sum to the Plaintiff in the form of cash deposits in certain specified Trust Accounts totaling US$40 million less the Fixed Amount. The Fixed Amount was to be and was paid as the price by the Plaintiff under the Swaps. The relevant terms of the Swaps The Plaintiff entered into 7 Swaps, one with each of the Defendants, in early July 2018. The relevant terms of the Swaps (which were expressly governed by English law) are as follows:

1. The Swaps provided for payment by the Defendants to the Plaintiff of a set amount if there was a Trigger Event in a Covered Territory during the Risk Period, which was defined as 1 June 2018 to 31 December 2018 (and a valid Payment Notice was served prior to the Termination Date). 3

2. A Trigger Event was defined as a Covered Event where the Reference Amount was equal to or in excess of the Trigger Amount.

3. A Covered Event was defined as a Loss Occurrence in the Covered Territory with a Date of Loss occurring within the Risk Period in respect of which a Loss Report reports an Estimated Market Loss.

4. A Reference Amount was defined as the Estimated Market Loss for the applicable Covered Territory arising from such Covered Event. Estimated Market Loss was in turn defined as the estimate of incurred losses from the insurance industry and reinsurance industry as a whole, arising from a Loss Occurrence, as reported by the Report Publisher in a Loss Report.

5. A Loss Report was defined to mean any loss report originated and disseminated by the Report Publisher during the Reporting Period concerned (which for Japanese Windstorm was defined as the 2-year period ending on 31 December 2020) that identifies and assigns a number to a Loss Occurrence or gives preliminary estimates (or subsequently resurvey estimates) of losses arising from a Loss Occurrence.

6. Different Report Publishers were nominated for different sections of the Swaps. For the three sections of the Swaps which related to US events, Property Claims Services ("PCS") was chosen as the Report Publisher, while for the section concerning events in Australia, PERILS AG was selected. The Report Publisher for the section of the Swaps concerning Japanese windstorms was (at the Plaintiff’s election) nominated as NatCatSERVICE (“NCS”). NCS is part of the publication activities of the major reinsurance company, Munich Re.

7. Between 1 June 2018 and 31 December 2018, only one relevant Covered Event occurred, namely a Japanese Windstorm, Typhoon Jebi, pursuant to Section 4 of the Swap Confirmations. The Trigger Amount for a Windstorm in Japan was US$12.5 billion. In the event that the Estimated Market Loss was equal to or greater than that 4

Trigger Amount, the Plaintiff would have been entitled to payment of the sum of US$20 million from the Defendants.

8. It is common ground that both PCS and PERILS, which are commonly referred to as ‘official’ Report Publishers, provide loss estimates on a ‘paid for’ or subscription basis and explicitly contemplate the use of their information as an index in ILW swaps for a fee. According to the terms of their contractual engagement, their loss estimates are regularly updated in accordance with a specified schedule. By contrast NCS, which is commonly referred to as an ‘unofficial’ Report Publisher, is a publication service which is freely available. It is not intended or designed to be used as an index for ILW swaps.

9. As an ‘unofficial’ Report Publisher, NCS did not (and does not) commit to publish a certain minimum (or maximum) number of loss reports, nor to update its loss estimates over time, nor to observe any pre-defined time intervals between publishing any loss estimates for a given event and it is not contractually engaged or remunerated by the parties to an ILW swap to act as a Report Publisher. In contrast, an ‘official’ Report Publisher such as PCS is typically appointed or engaged by the parties to such transactions to act as a Report Publisher under a licensing agreement and for a fee, and it will typically make a contractual commitment to maintain the format, and to adhere to the specified schedule and frequency of publication, of its loss reports.

10. At the time of entering into the Swaps, in July 2018, for practical purposes there was only a choice of Report Publishers for Japanese events between NCS and Sigma (both ‘unofficial’ Report Publishers), as PERILS and PCS did not at that time cover Japan. Sigma is owned by the major reinsurance company, Swiss Re.

11. If at the Risk Period End Date (31 December 2018) a Covered Event had occurred but the size was either not yet known or was reported by the agreed report publisher at an amount which was not large enough to trigger the payment obligation, then a Development Period applied, namely a further time period in which to monitor whether the estimated insured loss developed sufficiently in order to trigger a payment under 5

the Swaps. By clause 9(f) of the Swap Confirmations, the Development Period End Date was 18 months after the Risk Period End Date (i.e. 30 June 2020) if the loss estimate for Typhoon Jebi was less than 80% of the Trigger Amount (i.e. less than US$10 billion) by that date. By clause 8.2(b) the Swaps were to terminate on the 3rd Business Day following the Development Period End Date - namely 6 July 2020.

12. The Reporting Disruption provision on which the Plaintiff bases its case is at Clause 7 of the Swap Confirmations and provides as follows:

7. Reporting Disruption: If the Report Publisher (“P”) ceases to provide any Loss Reports or materially changes the methodology, determination or reporting of a catastrophe or any loss estimates in any way that, in the determination of the Calculation Agent, makes such events or estimates unsuitable for the purposes intended herein, then the parties shall determine a successor Report Publisher in lieu thereof for the purposes of the Transaction by using the following methodology: 7.1 The parties shall use commercially reasonable efforts to choose a mutually acceptable industry recognised benchmark Report Publisher that uses the methodology most closely tracking the methodology used by P prior to the cessation or material change of reporting by P and such agency will be deemed to be the successor Report Publisher. 7.2 If a successor Report Publisher has not been determined under Paragraph 7.1 above as of the date of the cessation or material change and, at any time, a notice has been given or an announcement has been made by P specifying a replacement report publisher and the Calculation Agent determines that such replacement report publisher will use the same or substantially similar methodology as that which was used by P, determination or reporting methods of a catastrophe as P, such publisher will be deemed to be the successor Report Publisher. 6

7.3 If a successor Report Publisher has not been determined under Paragraph 7.1 or 7.2 above within 45 days following the cessation or material change, then the Calculation Agent shall ask five leading independent dealers in industry loss warranties or catastrophe-linked swaps or derivative products to state who the successor report publisher should be and, if three or more responses are received, and at least two or more leading independent dealers state the same report publisher, such publisher will be deemed to be the successor Report Publisher. 7.4 If a successor Report Publisher has not been determined under Paragraph 7.1, 7.2 or 7.3 above within 60 days following the cessation or material change, the Calculation Agent shall designate an industry recognized benchmark Report Publisher which uses a methodology which closely tracks the methodology used by P prior to the cessation of reporting by P and such agency will be deemed to be the successor Report Publisher …

13. Pursuant to clause 14.1 of the Swaps, the Defendants were obliged to exercise their duties as Calculation Agent in good faith and in a commercially reasonable manner. Clause 14.1 provides that: Additional Agreements: 14.1 Calculation Agent. For the avoidance of doubt, all Calculation Agent determinations, unless otherwise provided herein, will be made in accordance with the 2006 Definitions, in good faith and a commercially reasonable manner. The Calculation Agent shall provide a brief written explanation of each of its determinations promptly upon request of either party …”

14. There was a separate Trust Agreement between Aeolus, each of the Defendants and the Trustee relating to each Swap Confirmation. The sum of US$20,000,000 held by the Trustee was subject to the Trust Agreements. Each of the Trust Agreements contained the following provisions: 7

“…

3. Withdrawal of Assets from the Trust Account. (a) Without notice or the consent of the Grantor, the Beneficiary shall have the right, at any time and from time to time, to withdraw from the Trust Account, upon written notice to the Trustee in substantially the form of Exhibit C hereto (the “Withdrawal Notice”), such Assets as are specified in such Withdrawal Notice. The Withdrawal Notice may designate a third party (the “Designee”) to whom Assets specified therein shall be delivered. The Beneficiary need present no statement or document in addition to a Withdrawal Notice in order to withdraw any Assets. (b) Upon receipt of a Withdrawal Notice the Trustee shall promptly take steps to deliver or transfer or instruct the relevant depository to deliver or transfer such Assets to or for the account of the Beneficiary or such designee as may be specified in such Withdrawal Notice …

4. Application of Assets. (a) The Beneficiary hereby covenants to the Grantor that it shall use and apply any withdrawn Assets, without diminution because of the insolvency of the Beneficiary or the Grantor, for the following purposes only: (i) to pay the Beneficiary any amounts that are due and payable by the Grantor under a Swap Agreement …” C. The evidence adduced by the parties

15. The Plaintiff relied on the factual evidence of Mr Henry Kingham, who is a Partner and Portfolio Manager at Aeolus Capital Management Ltd, which is the designated insurance manager for various segregated accounts of the Plaintiff. The Defendants 8

relied on the factual evidence of Mr Niklaus Hilti and Mr Arthur Esteves-Ferreira, both of Credit Suisse Insurance Linked Strategies Ltd ("CSILS").

16. The Plaintiff also relied on the expert evidence of Mr Desmond Potter formerly of Guy Carpenter, who is an investment banker and participant in the insurance-linked securities market. The Defendants relied on the expert evidence of Mr Eduard Held, who has 26 years’ professional experience of working in the reinsurance industry, specialising in particular in the field of natural catastrophe reinsurance. D. The Plaintiff’s factual case

17. The Plaintiff’s factual case is set out in the two witness statements of Henry Kingham and is as follows.

18. The Swaps were drafted by the Defendants. In fact, throughout the relationship between the Plaintiff and the Defendants, the Defendants generally drafted the swap agreements, and these Swaps were no exception.

19. In the first draft of the Swaps, the Defendants suggested that the Report Publisher for windstorms in Japan should be both NCS and Sigma, and that the parties should use an average of these report publishers’ reported industry event estimates. The Plaintiff expressed a preference for using NCS alone because the majority of the swaps it had entered into prior to that time had used NCS and it wanted to maintain consistency across different deals. In the end, the parties agreed to have NCS as the sole Report Publisher for windstorms in Japan, although the Defendants made clear that their agreement in this regard should not be taken to suggest that they would agree to use just NCS for Japanese windstorms in any future ILWs between the parties.

20. At the time the Swaps were entered into, in mid-2018, Sigma and NCS were the only two potential report publishers covering windstorms in Japan and thus the parties’ only options in this regard. Whilst NCS was not a formal or “official” index, there was no 9

better alternative for a Japan-based swap at that time because Sigma was also “unofficial”. For the Japan section of the Swaps, therefore, the parties had a choice between two unofficial indices, neither of which formally endorsed being used for such purposes but both of which were in fact widely used (and had been since the beginning of placements of ILWs for losses outside the US), given the then lack of any official index outside the US.

21. Prior to the parties entering into the Swaps, NCS had followed a long established and consistently reliable reporting method in relation to losses caused by natural catastrophes. In their own words, NCS was “the world’s most comprehensive database of natural catastrophes, going all the way back to the eruption of Mount Vesuvius in AD 79”. NCS also claimed that “No other source offers more comprehensive, reliable and professional information on losses due to natural catastrophes than Munich Re’s NCS”. Contrary to what is said in the Amended Defence, therefore, and having regard to the evidence set out, NCS was indeed “reliably regular”.

22. NCS published detailed annual reports taking broadly the same format every year. Both Munich Re’s NCS and Swiss Re’s Sigma would publish annual reports in the first quarter of each year. Those reports would look back at the events of the previous year, assigning loss estimates to the various natural catastrophes that had struck as well as addressing other matters, such as the make-up and impact of the losses. Munich Re and Swiss Re were, and continue to be, market leaders, so the views expressed in their annual reports had credibility and were of interest to many in the industry, including Mr Kingham as both a broker at Willis Re and as a Partner and Portfolio Manager at the Plaintiff. Whilst the tangible use for these reports was for their estimated market loss figures, they also provided valuable information on industry events.

23. In addition to the regular annual reports, NCS provided supplementary reports on an ad hoc basis. These were often, though not always, published in mid- to late- December and sometimes again in early July. They provided a simple, “flash” report of the year 10

or half year, including estimates of economic loss, insured loss and fatalities, sometimes on just one page.

24. NCS also featured a freely available interactive online portal. This online portal included summaries of the greatest losses for the insurance industry, including up-to- date Estimated Market Losses for these events, clearly presented on maps. This online portal was widely used within the industry, including using the Estimated Market Losses reported therein to determine whether an event had met the trigger amount required under an ILW or swap for collection of payment to the buyer from the seller. Between reports, NCS updated its loss estimates online. The online portal could be readily accessed and did not require log-in details and there was no fee to use it.

25. When choosing NCS as a Report Publisher under the Swaps, the Plaintiff expected NCS to continue reporting in materially the same way that it had done in previous years. In particular, the Plaintiff expected a detailed annual report in Q1 of 2019 (and again in Q1 of 2020), updates as losses developed as required, and a readily (and freely) available online tool that it could use to check the latest Estimated Market Loss for an event. The Plaintiff also expected it to continue to report fairly and to be broadly consistent with industry views of particular losses as it had demonstrated in the past. There was no reason to expect that NCS’s reports would not continue, as, the Plaintiff alleges, happened in this case.

26. These expectations were based on the Plaintiff’s knowledge and experience of NCS and its reporting in relation to previous catastrophes, and in particular NCS’s reporting patterns and the way its estimates related to the estimates produced by other report publishers, over a number of years. Given the objective basis for the Plaintiff’s expectations of NCS, Mr Kingham assumes that the Defendants had similar expectations. Nevertheless, there would have been no reason to choose a Report Publisher if it was expected that they would cease producing their reports. 11

27. Having done things in a particular way for years, in the first quarter of 2019, contrary to all expectations and without warning, NCS failed to publish its usual annual report. Instead, it published an abbreviated report dated 8 January 2019. Munich Re also published a press release, also dated 8 January 2019, but nothing as detailed as its usual annual reports.

28. By April 2019, the usual annual report was well overdue. There was lots of speculation and discussion in the industry at this time as to what was happening at NCS and why the usual annual report had not been published. Mr Kingham learned after various discussions with brokers that NCS was moving to an online only service.

29. In January 2020, NCS again published just a very short report, rather than its previously detailed annual report. The abbreviated report was released on 8 January 2020, together with an infographic created by NCS depicting 2019 losses.

30. The fact that something was wrong at NCS in 2019 was widely discussed in the industry. There were lots of rumours and speculation that NCS may have been involved in litigation with PCS in relation to NCS using PCS data for US loss estimates. It was obvious that something was wrong with NCS because Swiss Re’s Sigma materially increased its estimated market loss for Jebi to reflect the market updates widely being shared by Japanese cedents in August 2019, but Munich Re’s NCS had gone completely silent. NCS’s Jebi loss estimate from January 2019 was not updated until November 2020.

31. Typhoon Jebi made landfall in Japan in early September 2018. According to a Cresta report, it was the most intense tropical cyclone to make landfall in Japan since 1993. The following media articles from insurance industry press sources relating to Typhoon Jebi, illustrate the development of the industry loss estimates over the relevant period and the significant “loss creep” arising from this event: (a) Typhoon Jebi industry insured loss estimated up to $4.5bn – 10 September 12

2018; (b) Typhoon Jebi to cost re/insurers up to $5.5bn: RMS – 17 September 2018; (c) Rising typhoon Jebi loss suggests broader ILS exposure – 19 October 2018; (d) Jebi officially the largest Japan typhoon insured loss ever – 22 November 2018; (e) ILS fund returns negative in October on Michael & Jebi/Trami loss creep – 30 November 2018; (f) Munich Re pegs typhoon Jebi industry loss at $6bn, Trami at $2bn – 6 December 2018; (g) Munich Re pegs 2018 catastrophes at $80bn, hikes wildfires & Jebi estimates – 8 January 2019; (h) Typhoon Jebi industry loss now over $10bn: Sirius – 22 February 2019; (i) Typhoon Jebi industry loss moving closer to $12bn – 28 March 2019; (j) ILW payouts on the cards after typhoon Jebi loss estimate rises – 11 April 2019; (k) SCOR’s results confirm Japanese typhoon loss creep trend from Jebi & Trami – 26 April 2019; (l) Market missed typhoon Jebi BI, industry loss now up to $13bn: Grandisson, Arch – 2 May 2019; (m) Jebi nears $13bn, why industry missed the loss creep: Swiss Re CFO – 6 May 2019; (n) Munich Re grows at higher prices, but EUR267m Jebi creep dents profits – 8 May 2019; (o) Typhoon Jebi industry loss may settle at $15bn to $16bn: Analysts – 28 May 2019; (p) Loss creep evident in combined $14.5bn Jebi, Trami, Japan rains claims from GIAJ – 6 June 2019; (q) Jebi creep impacts ongoing, as loss threatens fresh reinsurance layers – 5 July 2019; (r) SCOR into its retrocession for any further Jebi loss creep – 29 July 2019; (s) Further Jebi loss creep may impact RenRe’s retro portfolio: CEO O’Donnell – 30 July 2019; (t) Munich Re boosted by reserve releases, which offset €80m Jebi creep – 7 13

August 2019; (u) Hannover Re gets “long overdue price increases”, Jebi creep dents P&C – 8 August 2019; (v) Jebi above 1-in-40-year loss, development could continue: S&P – 21 August 2019; (w) Munich Re lifts Jebi industry loss to $13bn, some ILWs potentially in play – 17 December 2020.

32. The seriousness of the evident problem with NCS’s reporting of Jebi loss estimates is further illustrated by the insured losses confirmed as having been paid by the General Insurance Association of Japan. By October 2019, the GIAJ had recorded paid claims in relation to Typhoon Jebi of approximately US$10.7 billion, i.e. significantly higher than NCS’s estimate of US$9 billion, demonstrating concretely that the NCS estimate was woefully out-of-date and represented a clear underestimate of Jebi losses.

33. Mr Kingham says that there is no doubt in his mind that the Defendants must have been aware of the disruption and changes at NCS and/or its cessation of reporting, which, by at least December 2019, caused NCS’s Jebi estimate to be significantly out-of-step with the estimates produced by both PCS and Sigma and the quantified losses confirmed as having been paid by the GIAJ.

34. The Plaintiff further argues that NCS’s published methodology was that it would update its event loss estimates where new data was available. This is especially so in circumstances where (a) the losses arising from Typhoon Jebi were not just substantial but at record levels and certainly for the previous 25 years, (b) there was obvious “loss creep” meaning that NCS’s earlier estimate was substantially understated as increasing numbers of insurance claims were being submitted and paid, more than had been anticipated, for a considerable time after the Typhoon, and (c) NCS’s first estimate was plainly understated as it was prepared only four months after Typhoon Jebi, which was Japan’s most devastating catastrophe resulting in substantial and complex losses. 14

35. The Plaintiff contends that there was a wide-spread market expectation that the NCS estimate was unrepresentative of the true extent of the insured loss and had to be updated. Yet there was complete silence on the part of NCS in not updating its loss estimate until November 2020, where other loss reporters were providing updated estimates throughout 2019.

36. The Plaintiff argues that it would have been obvious to the Defendants as Calculation Agent that NCS had ceased reporting or had materially changed its methodology or reporting which rendered its January 2019 estimate of US$9 billion unsuitable for the purposes of the Swap Confirmations. Notwithstanding, the Defendants failed to take steps to replace NCS as the Report Publisher in breach of the Swap Confirmations.

37. The legal effect of the Defendants’ breaches of contract is, argues the Plaintiff, that (1) a pre-condition for the Termination of the Swap Confirmations and therefore the payment of the US$20,000,000 held by the Trustee to the Defendants in the case of an “Untriggered Section” is not satisfied, with the result that Aeolus is entitled to be paid the US$20,000,000, or (2) by reason of the Defendants' breaches of contract, Aeolus has suffered loss and damage in the amount of the US$20,000,000 and Aeolus is entitled to an award in the said amount (including any right of set off).

38. Despite the above, and the fact that the Defendants had contractual obligations (including an obligation to act in good faith) to remedy the situation by choosing a replacement report publisher, they took no steps to comply with their obligations under Section 7 of the Swaps to choose a successor Report Publisher. E. The Defendants’ factual case

39. Mr Niklaus Hilti was the decision maker for the Defendants in relation to this matter. At the relevant times he was the Managing Director and Head of CSILS. Mr Hilti has 21 years of professional experience in the reinsurance and insurance linked strategies market. 15

40. In Mr Hilti’s experience, NCS reported loss estimates for events both by way of short form reports published from time to time and also by way of periodical publications, such as the annual publication called "Topics GEO" which was typically published in the first quarter of each year and contained research and scientific articles that tended to focus on the natural catastrophes that had occurred in the preceding year. In addition, NCS’s loss data on prior year loss events might be made available as part of occasional ad hoc reports published on an irregular basis on Munich Re’s website. From about 2017/2018 onwards (and until after the automatic termination of the Swaps) NCS also made available a web-based online tool for accessing and interrogating NCS’s library database of published historical loss data.

41. Further, in Mr Hilti’s experience NCS gave no contractual or other commitment to report on any particular loss occurrence(s) with any stated frequency, intervals or regularity, and NCS did not generally tend to update its published loss estimates which was the exception rather than the rule. NCS would only report when it chose to and did not commit to provide any updates on its loss estimates in light of industry losses. It was therefore not known in advance, but was uncertain and unpredictable, whether or not, and if so when, NCS might produce an updated loss figure (if at all) for a particular natural catastrophe.

42. At the time of entering into the Swaps, the Plaintiff (acting through its brokers Aon UK and on Mr Henry Kingham’s instructions) insisted on the selection of NCS as the Report Publisher for Section 4 Japan Windstorm. The Defendants had suggested that Sigma, or else an average of the loss estimates of NCS and Sigma, should be used instead, and had put forward contractual wording which was aimed at taking account of the uncertainty surrounding these unofficial Report Publishers’ reporting practices and which catered for varying estimates being published by one, the other or both of them at different times. However, the Plaintiff rejected the Defendants’ suggestion and maintained its selection of NCS. In the face of this insistence by the Plaintiff, the Defendants agreed to accommodate its preference and accepted its choice of NCS as the Report Publisher for Japanese Windstorm. 16

43. Throughout the duration of the Swaps, the Defendants monitored and tracked the reporting of various Report Publishers, including NCS.

44. In early September 2018, Typhoon Jebi struck Japan, causing very substantial damage. On or about 5 December 2018, Munich Re (the owner of NCS) produced an estimate of insured losses caused by Typhoon Jebi of US$6 billion. The source of that estimate was Munich Re and the Japanese Meteorological Agency, not NCS. Accordingly, Mr Hilti did not consider at the time that this constituted a "Loss Report" by NCS for the purposes of the Swaps. It is evident that the Plaintiff also shared this view.

45. On 8 January 2019, NCS published a press release of the "Relevant natural loss events worldwide 2018" in which Typhoon Jebi was listed as a 2018 catastrophe causing US$9 billion in insured losses. That publication by NCS constituted a "Loss Report" for the purposes of the Swaps; the loss estimate was lower than the Trigger Amount of US$12.5 billion under the Swaps.

46. On 29 July 2019, the NCS publication "Relevant Natural loss events worldwide 2018" was downloaded at CSILS from the NCS online tool. The report was reviewed by Mr Esteves-Ferreira of CSILS. The publication again showed the loss estimate by NCS for Typhoon Jebi at US$9 billion.

47. In August 2019, during a standard assessment for reserving purposes of the value of assets managed by the Defendants, Milliman (an actuarial service provider) informed Mr Hilti that, in their view, the actual insured loss for Typhoon Jebi would ultimately be in the region of US$16 billion or more, and that therefore the Defendants should reserve on that basis. Mr Hilti did not agree with Milliman’s approach to reserving since he considered that the correct question to ask for reserving purposes in relation to the Swaps, given their terms (and as distinct from a traditional reinsurance contract), was not whether the actual insured industry loss was projected to reach a certain amount, but what was the probability (a) that the chosen Report Publisher, NCS, would update its published loss estimate prior to the end of the Reporting Period of the Swaps 17

and (b) that, if so, any such updated estimate would be higher than the Trigger Amount under the Swaps.

48. To that end, on 13 August 2019, on Mr Hilti’s instructions, the Head of Underwriting at CSILS asked Mr Aaron Koch of Milliman to provide a probability assessment of NCS and Sigma (because the Defendants also had another swap with an unrelated counterparty to monitor for the impact of Typhoon Jebi where the report publisher was Sigma) updating their loss estimates for Typhoon Jebi within the relevant period and to a level above the relevant trigger amount. At that time Sigma’s published loss estimate for Typhoon Jebi was US$9.77 billion (similar to NCS’s published loss estimate). Milliman’s analysis was principally based on an analysis of Sigma’s historic approach to updating. Mr Koch of Milliman said as much in an e-mail of 19 August 2019: "the information that we had only allowed for that depth of analysis with SwissRe [i.e. Sigma]". Milliman’s opinion in summary was that there was a two-thirds chance of there being an update of the loss estimate for Typhoon Jebi of greater than US$12.5 billion by either NCS or Sigma before the end of the Reporting Period of the Swaps.

49. Mr Hilti’s understanding at the time (based on his market experience) was, as set out in paragraph 41 above, that NCS did not generally update its estimates (and that the examples given by Milliman were the exception rather than the rule), and that if and when NCS might do so was uncertain and could not be predicted. Mr Hilti recorded his view that NCS was unlikely to update its published loss estimate for Typhoon Jebi in an internal e-mail of 14 August 2019, in which he stated that both NCS and Sigma were unlikely to update their estimates as they had no interest in an "actual number for the past year and changes post-publication are not really important to them".

50. It is evident that that view was widely held in the market: it was, for example, shared by individuals at Guy Carpenter (a major broker), others at CSILS as well as by Mr Held (the Defendants’ expert), and it had also been expressed by Ms Dalida Durkic of CSILS to the Plaintiff’s leading broker Aon during the negotiations of the Swaps in June 2018 without demur: "NatCat reports only once per event". 18

51. Having announced on 25 January 2019 that it would start to cover Japan geographically for the first time, PCS then produced an example or ‘trial run’ loss report using Typhoon Jebi in which PCS put the estimated total losses at US$10 billion. In August 2019, PCS revised this loss estimate to US$13 billion (which within a few days it increased to US$14.5 billion but then subsequently revised downwards in November 2019 to US$13.7 billion). Mr Hilti did not consider that the fact that PCS produced its first ‘trial run’ loss report for Japan catastrophes indicated that there was any reporting disruption at NCS.

52. In September 2019, Mr Hilti attended the annual Monte Carlo Rendezvous, an established gathering of insurers/reinsurers, brokers and service providers in Monte Carlo. During a meeting there, Mr Hilti was informed by Guy Carpenter that they believed, from their discussions with Munich Re, that there would be no update to NCS’s published loss estimate for Typhoon Jebi. Similarly, during meetings with brokers in London in September 2019, Mr Esteves-Ferreira was also informed by experienced individuals from Guy Carpenter that they understood (from information relayed by Aon from a source at NCS) that NCS would not be updating its published loss estimate for Typhoon Jebi.

53. That was consistent with Mr Hilti’s long experience of NCS’s practice. At no stage then, or subsequently (prior to termination of the Swaps), did Guy Carpenter, Aon or any other market participant suggest to Mr Hilti or anyone else to the Defendants that NCS had ceased publishing loss estimates or changed its methodology of loss estimates or determination or reporting in any way, or suggest that there might be any Reporting Disruption in respect of NCS. As Mr Hilti explains "we were (and are) in constant contact with brokers and others in the market and not one of them suggested at any time that there was anything "wrong" at NCS". Indeed, during the period until the termination of the Swaps brokers continued to contemplate the use of NCS as a report publisher for Japanese events for new ILWs which were proposed to CSILS. 19

54. In December 2019, Sigma updated its loss estimate for Typhoon Jebi to US$12.99 billion. This was noted by the Defendants.

55. Mr Hilti did not consider that this indicated anything untoward with regard to NCS’s loss estimate, which remained unchanged. In his experience, loss reports concerning the same catastrophe from different Report Publishers could, and often did, differ substantially from each other, and there was no reason to draw any adverse inferences or conclusions concerning NCS based on the fact that Sigma had issued an update. In Mr Hilti’s experience, for large catastrophes such as Typhoon Jebi it was commonplace to find wide variations between unofficial Report Publishers’ loss estimates for the same catastrophe, of as much as 50% (or sometimes more), and this was not a level of variation that would be considered remarkable or a cause for concern in the context of the insured loss amount as a whole.

56. More generally, Mr Hilti was aware that there had been ”loss creep” in the insurance/reinsurance market in relation to Typhoon Jebi. Mr Hilti explains that "loss creep" is not new or somehow unique to Typhoon Jebi. It is industry jargon for affected insurers and reinsurers initially estimating their losses for a loss occurrence at a lower level when compared to the actual claims coming in or, simply, new information becoming available, and the root cause is often a lack of available data. Loss creep is part of the risk that a swap party accepts when using a report publisher’s loss estimates rather than using cedants’ actual claims figures, as would be the case in a conventional indemnity contract. While that risk may to an extent be mitigated when using official report publishers such as PCS, which reports on a regular schedule, unlike unofficial report publishers such as NCS or Sigma, there is always an element of residual ‘basis risk’ because of the long time that it typically takes to establish the ultimate loss amount for a loss occurrence. Further, that basis risk will be heightened in relation to a market, such as the Japanese market, in which the insurers adopt a highly protective approach to their confidential (non-public) loss data, and there are a relatively limited number of data points available by reference to which to estimate losses: the Japanese insurers’ actual loss data was confidential and would not be available to NCS to use. 20

57. In December 2019, the GIAJ produced a report which estimated that the total industry paid loss for Jebi was 1,0678 billion yen, which equated at the time to an industry loss of around US$9.6 billion.

58. In the context of the year-end financial review for the CSILS-managed funds, on 24 December 2019, Mr Esteves-Ferreira mentioned in an email to Mr Hilti the possibility that the Plaintiff might try to allege a Reporting Disruption in respect of NCS based on the latest, higher Sigma loss estimate for Typhoon Jebi. Mr Hilti understood his e-mail to be a reflection on what the Plaintiff might seek to do commercially in the circumstances, but he did not understand it to be suggesting, or in any case himself consider, that there had been a Reporting Disruption in respect of NCS.

59. In January 2020, there was a further on-line publication on the Munich Re website referencing NCS’s estimates for insured losses during the preceding year 2019 (which therefore did not include Typhoon Jebi). That was entirely in line with what Mr Hilti understood to be NCS’s usual practice of annual reporting in the first quarter of each year only in respect of losses that had occurred in the year preceding that of the publication.

60. Notably, at no stage prior to the termination of the Swaps, did the Plaintiff suggest to the Defendants that there was a Reporting Disruption or, indeed, that there were matters which meant that the Defendants should consider whether there was such a Reporting Disruption, in respect of NCS. Had the Plaintiff genuinely believed that there had been a Reporting Disruption, Mr Hilti would have expected the Plaintiff to inform him/CSILS promptly that this was the case.

61. Furthermore, in mid-August 2020, following the termination of the Swaps when Milliman was asked by CSILS whether it agreed that the loss reserve for the Swaps could now be reduced to zero, Milliman did not suggest that the Defendants’ view that there had not been a Reporting Disruption was incorrect but considered that the Defendants could reduce the loss reserve in relation to the Swaps to zero. 21

62. Accordingly, in summary and prior to the termination of the Swaps in early July 2020, Mr Hilti was aware that: (a) NCS had produced a loss estimate for Typhoon Jebi in January 2019, as it had done for similar events in the past; (b) Consistently with its previous practice (including a number of events to which Milliman referred in their analysis) NCS had not updated its loss estimate for Typhoon Jebi, and (again consistent with previous practice) there was no certainty as to whether it would update it and, if so, when; (c) NCS had continued to produce loss estimates for other events, in accordance with its practice (for example, in 2020 for 2019 events); (d) The Plaintiff itself had not suggested that there was a Reporting Disruption in respect of NCS or that it had any concerns in that regard; and (e) It had not been suggested to the Defendants by any other market participants that there was a Reporting Disruption in respect of NCS or that they had any concerns in that regard.

63. In the event NCS did not update its loss estimate or publish a revised loss estimate for Typhoon Jebi during the lifetime of the Swaps up to and including the Termination Date (in early July 2020). G. Issues to be determined

64. Having regard to the pleaded case and the closing submissions of counsel, the following issues arise for determination by the Court: 22

(1) As a matter of construction of Clause 7, what is the scope of the determination to be made by the Calculation Agent? (2) As a matter of construction of Clause 14.1, what is the proper meaning of the expression “in good faith and commercially reasonable manner”? (3) As a matter of construction of Clause 7, are the steps set out in Clause 7.1 and 7.2 a precondition to the operation of the steps set out in Clause 7.3 and 7.4? (4) Was the Calculation Agent entitled to take the view that NCS had not ceased to provide any Loss Reports?1 (5) Was the Calculation Agent entitled to take the view that NCS had not materially changed its methodology? (6) Was the Calculation Agent entitled to take the view that NCS had not materially changed its methodology merely because of the abbreviation of the annual loss reports from 2019 onwards? H. Discussion and determinations Principles of construction

65. There is no dispute in relation to the relevant principles of construction to be applied in ascertaining the meaning of the provisions in the Swap Confirmations. These principles are set out in the speech of Lord Hoffmann in Investors Compensation Scheme Ltd v West Bromwich Building Society [1998] 1 WLR 896, which have been approved by the Court of Appeal for Bermuda in Aircare Ltd v Sellyeh [2015] CA (Bda) 1 The formulation of this issue in this way assumes that in the first instance it is for the Calculation Agent to determine this issue and not for the parties and this Court. 23

6 Civ (20 March 2015), para. 8 and by the Supreme Court of Bermuda in The Corporation of Hamilton v Bermuda Electric Light Company Ltd [2018] SC (Bda) 75 Civ (16 November 2018), para. 17-19 and in Annuity & Life Re Ltd v Kingboard Copper Foil Holdings Ltd [2021] SC (Bda) 52 Com (29 June 2021), para. 26.

66. As Mr MacDonald Eggers QC, on behalf of the Plaintiff, correctly submits, the aim of contractual construction is to divine the parties’ objective intention by reference to the language in which the parties have expressed the contractual terms in light of the factual background known or available to the parties and the commercial purpose of the contract and its provisions.

67. In Arnold v Britton [2015] UKSC 36; [2015] AC 1619, Lord Neuberger said (at para. 15-17): “15. When interpreting a written contract, the court is concerned to identify the intention of the parties by reference to “what a reasonable person having all the background knowledge which would have been available to the parties would have understood them to be using the language in the contract to mean”, to quote Lord Hoffmann in Chartbrook Ltd v Persimmon Homes Ltd [2009] AC 1101, para 14. And it does so by focussing on the meaning of the relevant words … in their documentary, factual and commercial context. That meaning has to be assessed in the light of (i) the natural and ordinary meaning of the clause, (ii) any other relevant provisions of the lease, (iii) the overall purpose of the clause and the lease, (iv) the facts and circumstances known or assumed by the parties at the time that the document was executed, and (v) commercial common sense, but (vi) disregarding subjective evidence of any party’s intentions.”

16. For present purposes, I think it is important to emphasise seven factors.

17. First, the reliance placed in some cases on commercial common sense and surrounding circumstances (eg in Chartbrook [2009] AC 1101, paras 16-26) 24

should not be invoked to undervalue the importance of the language of the provision which is to be construed. The exercise of interpreting a provision involves identifying what the parties meant through the eyes of a reasonable reader, and, save perhaps in a very unusual case, that meaning is most obviously to be gleaned from the language of the provision. Unlike commercial common sense and the surrounding circumstances, the parties have control over the language they use in a contract. And, again save perhaps in a very unusual case, the parties must have been specifically focussing on the issue covered by the provision when agreeing the wording of that provision …” (1) Role of the Calculation Agent under Clause 7

68. Mr MacDonald Eggers QC submits that on the plain language of Clause 7, the Report Publisher - NCS - should be replaced by a successor by the Defendants as the Calculation Agent in accordance with the procedure set out in Clauses 7.3 or 7.4 if the following conditions are satisfied: (1) NCS ceases to provide any Loss Reports; or (2) (a) NCS materially changes the methodology, determination or reporting of a catastrophe or any loss estimates and (b) the Defendants as the Calculation Agent determines that the said material change makes such events (the catastrophe) or (loss) estimates unsuitable for the purposes of the Swap Confirmations.

69. Mr MacDonald Eggers QC submits that the Calculation Agent has a limited role under Clause 7 in relation to the issues of whether the Report Publisher has ceased to provide any Loss Reports or has materially changed the methodology. It is submitted on behalf of the Plaintiff that the Calculation Agent’s role is limited to determining whether a material change in the Report Publisher’s methodology, determination or reporting of 25

a catastrophe or loss estimate would make “such events or estimates unsuitable for the purposes of the Swap Confirmations.”

70. It is Mr MacDonald Eggers QC’s submission that the Calculation Agent has no role in determining under Clause 7 the issues (i) whether the Report Publisher has ceased to provide any Loss Reports; or (ii) whether the Report Publisher has materially changed the methodology or reporting of the catastrophe, or any loss estimates in any way. These are issues, it is said on behalf of the Plaintiff, which must be determined by the parties and failing agreement between the parties, must be determined by the Court. According to this submission, the Calculation Agent’s role only comes into play once either the parties or the Court has made the determination that (i) the Report Publisher has ceased to provide any Loss Reports; or (ii) has materially changed its methodology. Once these determinations have been made, the Calculation Agent’s role is limited to determining whether the material change in methodology in question “make such events or estimates unsuitable for the purposes intended herein.”

71. The Plaintiff contends that in the event the parties or the Court determines that the Report Publisher has ceased to provide any Loss Reports, the Calculation Agent has no role to play in the opening paragraph of Clause 7 and must proceed to arrange for the selection of a successor Report Publisher in accordance with the mechanism specified in Clause 7.1 to 7.4.

72. The Court is unable to accept the construction of the opening paragraph of Clause 7 contended for by the Plaintiff.

73. First, as submitted by Mr Howe QC, the Plaintiff’s construction artificially compartmentalises and hives off “ceasing to provide any Loss Reports” and severs it from the assessment of suitability for the purposes intended by the contracts and leaves it entirely outside the discretion and determination of the Calculation Agent. 26

74. Second, on the contended construction, the “materiality” of the changes to the methodology is to be determined ultimately by the Court but that the closely related concept of “suitability” for the intended purpose is to be determined by the Calculation Agent. The separation of the termination of the closely related concepts for the intended purpose between two different decision-makers with potentially overlapping duplicative and wasteful processes of consideration and evaluation, as submitted by Mr Howe QC, is very unlikely to have been the sensible commercial intention of this provision.

75. Third, Clause 7 envisages that upon ceasing to provide any Loss Reports or material changes in the methodology, the parties will select a successor Report Publisher within 60 days utilising the procedure mandated by Clauses 7.1 to 7.4. However, if the issue whether the Report Publisher ceased to provide Loss Reports or has materially changed its methodology, has to be determined by the Court, it would be virtually impossible to comply with the 60-day timeframe to appoint a successor Report Publisher stipulated in Clause 7.4. Given that Clauses 7.1 to 7.4 are intended to operate during the currency of the agreement and within a timeframe of 60 days, it is highly unlikely that it was the intention of the parties that, before the machinery in Clauses 7.1 to 7.4 could be invoked, the parties obtain a ruling from this Court as to whether the Report Publisher has ceased providing Loss Reports or whether the methodology has materially changed.

76. The above considerations clearly indicate and support the natural construction of the opening paragraph of Clause 7 that the issues of whether (i) the Report Publisher has ceased to provide Loss Reports; (ii) the Report Publisher has materially changed its methodology; and (iii) the resulting estimates are unsuitable for the purposes intended, are all issues to be determined by the Calculation Agent. (2) Meaning of “in good faith and commercially reasonable manner”

77. Clause 14.1 of the Swaps provides that all determinations made by the Calculation Agent, unless otherwise provided, will be made in good faith and a commercially reasonable 27

manner. It also requires that the Calculation Agent shall provide a brief written explanation of each of its determinations promptly upon request of either party.

78. It follows that the determinations of the Calculation Agent in relation to whether (i) the Report Publisher has ceased to provide Loss Reports; (ii) the Report Publisher has materially changed its methodology; and (iii) the resulting estimates are unsuitable for the purposes intended, must be made in good faith and a commercially reasonable manner.

79. Mr MacDonald Eggers QC submits that the test for whether the Calculation Agent has acted in a “commercially reasonable manner” is an objective test for the Court to apply in the circumstances of this case. He submits that this requires the Calculation Agent to take into account the interests of both parties and also to adopt a reasonable and fair process in coming to its determination.

80. Mr Howe QC submits that a duty to exercise ‘good faith’ (which is usually to be contrasted with a duty to exercise reasonable care) connotes subjective honesty, genuineness and integrity, not an objective standard of any kind (whether reasonableness, care or objective fair dealing). A duty to act ‘in a commercially reasonable manner’ connotes an objective standard of rationality/reasonableness as to the process used (as distinct from the outcome of such process). The application of that standard leaves a bracket or range of processes - and a fortiori of outcome - within which the determining party may choose, even if the Court, carrying out the exercise itself, might have come to a different conclusion.

81. The meaning of the requirement that the Calculation Agent’s determinations must be made in good faith and a commercially reasonable manner is a matter of the proper construction of clause 14.1 of the Swaps having regard to the relevant factual matrix. Accordingly, the authorities relied upon by Counsel have to be considered by reference to the particular wording of the relevant clause and the factual matrix in those cases. 28

82. In support of his submission, Mr MacDonald Eggers QC relies upon the decision of Popplewell J in Barclays Bank Ltd v Unicredit Bank AG [2012] EWHC 3655; [2013] 2 Lloyd's Rep 1. In that case, Clause 12.1 of the guarantees granted the right of optional termination in four separate events, two of which required Barclays’ prior consent and provided that “such consent to be determined by [Barclays] in a commercially reasonable manner”. At [63] Popplewell, J said with respect to the phrase “commercially reasonable”: “The parties chose not to give Barclays an uncircumscribed discretion whether or not to consent but chose to confine it to behaviour which was commercially reasonable. The presumption is that by using express words, the parties were seeking to achieve some greater restriction than would have applied if no words had been used. Although public lawyers are familiar with the concept of reasonableness in its Wednesbury sense of not irrational, that is not the sense in which the word would commonly be used or understood by businessmen in a commercial agreement. That conclusion is reinforced by the use of the adverb “commercially”, which is apposite for a concept of objective reasonableness but rather less so for a concept of rationality.” (emphasis added).

83. Mr MacDonald Eggers QC also relies upon In Lehman Bros Special Financing Inc v National Power Corp [2018] EWHC 487 (Comm); [2018] 1 CLC 417, at para. 61, 70 and 81, in which Robin Knowles J construed an extended phrase (“commercially reasonable procedures in order to produce a commercially reasonable result”) in the ISDA 2002 Master Agreement as follows: “61. In Lehman Bros International (Europe) (in administration) v Lehman Bros Finance SA [2012] EWHC 1072 (Ch) Briggs J (as he then was) held, in relation to the 2002 ISDA Master Agreement: ‘[The provision] clearly imposes two objective standards. The first is that the procedures used should be commercially reasonable and the second is 29

that the result produced should also be commercially reasonable. Plainly, that leaves a bracket or range both of procedures and results within which the Determining Party may choose, even if the court, carrying out the exercise itself, might have come to a different conclusion. It nonetheless imposes an objective standard which, if for example the Determining Party refused to determine a Close-out Amount at all, could be applied by the court itself, for example on the application of the other party, as in Sudbrook Trading Estate v Eggleton [1983] 1 AC 444, where the breakdown of an agreed valuation procedure due to the refusal by one party to co-operate in the appointment of a valuer nonetheless left the court free to carry out the process itself on the application of the other party.’

70. Briggs J identified the wording ‘in order to produce a commercially reasonable result’ in the 2002 ISDA Market Agreement as adding what he termed a second objective standard, viz. that the result produced should be commercially reasonable. The wording is not simply describing the aim of commercially reasonable procedures …

81. Taking all of the above into account, in my view, the 2002 ISDA Master Agreement requires the Determining Party to use procedures that are, objectively, commercially reasonable in order to produce, objectively, a commercially reasonable result. If it does not do this the court or a tribunal will. This is also the view I understand Briggs J to have held.” (emphasis added).

84. Finally, in this context, Mr MacDonald Eggers QC refers the Court to the Court of Appeal decision in Gan Insurance Company Ltd v Tai Ping Insurance Company Ltd [2001] Lloyds Rep IR 667 and the judgment of Mance LJ (as he then was) at [67] and [76]: “67… I would therefore accept as a general qualification, that any withholding of approval by the reinsurers should take place in good faith after consideration of 30

and on the basis of the facts giving rise to the particular claim and not with reference to considerations of the particular reinsurance. …

70. In summary, the right to withhold approval was, here, Gan’s, and no one else’s. It is a right to be exercised in good faith after consideration of and on the basis of facts giving rise to the particular claim, and not with reference to considerations wholly extraneous to the subject matter of the particular reinsurance or arbitrarily. It is to be exercised by considering the claim as a whole. The Court cannot substitute its own view of the reasonableness of the reinsurer’s decision to withhold approval…”

85. In relation to the requirement of “good faith”, Mr Howe QC, for the Defendants, relies upon SNCB Holding v UBS AG [2012] EWHC 2044 (Comm) Cooke J at [72] and [110]: “A duty to exercise ‘good faith’ in doing something is one which is usually to be contrasted with a duty to exercise reasonable care. It connotes subjective honesty, genuineness and integrity, not an objective standard of any kind, whether reasonableness, care or objective fair dealing. It cannot be equated with "utmost good faith" and although its exercise in practice may involve different actions or restraint, the concept is not one which goes beyond the notion of truthfulness, honesty and sincerity" (emphasis added). "… A deliberate mistake would be made in bad faith, whereas an accidental mistake would not be, whether negligent or not. A reckless mistake, if made with shut-eye knowledge might also be made in bad faith, but the critical point is that bad faith in these provisions really does mean dishonesty or lack of integrity. It does not carry the meaning of ‘utmost good faith’. That is reinforced by the phrase ‘in good faith and in a commercially reasonable manner’ which draws a distinction between the two." 31

86. The main case relied upon by Mr Howe QC is the Court of Appeal decision in Barclays Bank Plc v Unicredit Bank AG [2014] EWCA Civ 302; [2014] 2 All ER (Comm) 115, [paragraphs 1, 7, 15 to 21] where the Court of Appeal considered the meaning of acting in a "commercially reasonable manner" in the context of a determination by a party to a financial instrument. As the Court of Appeal explained, the words "commercially reasonable manner" are directed at the process and manner in which the decision was made: they are not concerned with regulating the outcome of that process. Longmore LJ observed at [15] that it is "the manner of the determination which must be commercially reasonable; it does not follow that the outcome has to be commercially reasonable although, if it is not, that would no doubt cause one to look critically at the manner of the determination".

87. Longmore LJ held at [20] that as a matter of construction of the contract, the standard of rationality/reasonableness imposed in that case was the Wednesbury standard: that is, was it a determination which no reasonable party could come to: “[20] If, therefore, it is necessary or helpful to determine into which of the four categories, set out in para [14] above, the present clause falls, I would assign it to category (2) [“ (2) as equivalent to conferring a discretion to which the principles of AP Picture Houses v Wednesbury Corporation [1948] 1 KB 223, [1947] 2 All ER 680, 45 LGR 635applied”] as I think the judge did in para [67](4) of his judgment. This seems to me to accord, moreover, with previous decisions of this court or similar clauses such as Ludgate Insurance Co v Citibank NA [1998] Lloyd's Rep IR 221, paras 35-36 per Brooke LJ; Gan Insurance Co Ltd v Tai Ping Insurance Co Ltd (No 2) [2001] EWCA Civ 1047, [2001] 2 All ER (Comm) 299 paras 64, 67 and 73, [2002] Lloyd's Rep IR 667 per Mance LJ; Paragon Finance plc v Nash [2001] EWCA Civ 1466, [2002] 2 All ER 248, [2002] 1 WLR 685 para 41 per Dyson LJ and Socimer Bank Ltd v Standard Bank Ltd [2008] EWCA Civ 116, [2008] Bus LR 1304 paras 61-66, [2008] 1 Lloyd's Rep 558 per Rix LJ. This latter authority indicates that there is in any event little difference between categories (1) and (2).” (emphasis added). 32

88. That the Court of Appeal applied the Wednesbury test in Barclays Bank is confirmed by subsequent cases. Thus, in Crowther v Arbuthnot Latham & Co Ltd [2018] EWHC 504 (Comm) Walksman J noted at [28]: “[Longmore LJ in Unicredit] also said if it was necessary for him to decide it, that the nature of the reasonableness test here was by reference to Wednesbury unreasonableness or by reference to what today would probably be regarded as a form of a Braganza duty. In other words a test of rationality as opposed to the common law, reasonable man test."

89. In Mercuria Energy Trading PTE Ltd v Citibank NA [2015] EWHC 1481 (Comm); [2015] 1 C.L.C. 999 Phillips J confirm the same understanding at [116]: “Citi contends that the required opinion formed by Citi will fall within the clause unless it is unreasonable in the Wednesbury sense, that is, unless it is an irrational opinion that no reasonable party in the position of Citi could hold. Citi submits that similar clauses in banking and finance documents have been construed in this sense: … (ii) In Barclays Bank Plc v Unicredit Bank AG [2014] 1 CLC 342 Barclays, as the Guarantor in respect of credit default swaps, was required to determine certain matters “in a commercially reasonable manner”. The Court of Appeal, whilst emphasising that it was a matter of construction in the particular contract, held that the standard imposed was the Wednesbury standard.

90. Indeed, the approach of Popplewell J in Barclays Bank, relied upon by Mr McDonald Eggers QC, held at [67] that it is not for the Court to consider whether the decision is justified, but whether a reasonable commercial man in Barclays’ position might have reached such a decision: 33

“Accordingly in my judgment the correct approach to the application of the clause is as follows: (i) Barclays’ determination of whether to consent to early termination for a Regulatory Change under Clause 12.1(b) of the Guarantees must be commercially reasonable in an objective sense. It is not sufficient for Barclays to show merely that the decision was made in good faith and was not arbitrary, capricious or irrational. (ii) The question is not whether the decision is justified, but whether a reasonable commercial man in Barclays’ position might have reached such a decision. (iii) In determining what is commercially reasonable, Barclays is entitled to take into account its own commercial interests... (iv) Barclays’ commercial interests, in this context, comprise its interest in earning profits from its fee income under the Guarantees…”

91. The burden of the authorities reviewed above is that in cases where contractual discretion is given to one of the parties to make determinations which involve evaluation or assessment or the exercise of judgment, the court would apply the rationality test. Accordingly, the Court will not set aside the determination made by a party pursuant to such a contractual provision unless it can discern that no reasonable commercial person could have made the determination in the circumstances of the case. This test is supported by the Court of Appeal decision in Barclays Bank Plc v Unicredit Bank AG [2014] EWCA Civ 302; [2014] 2 AER (Comm) 115, [paragraphs 1, 7, 15 to 21]; the decision of Walksman J in Crowther v Arbuthnot Latham & Co Ltd [2018] EWHC 504 (Comm) at [28]; and the decision of Phillips J in Mercuria Energy Trading PTE Ltd v Citibank NA [2015] EWHC 1481 (Comm); [2015] 1 C.L.C. 999. 34

92. Clause 7 of the Swaps required the Calculation Agent to utilise its expertise in evaluating and forming a judgment in relation to the issues of whether there was a cessation of the provision of Loss Reports by the Report Publisher and/or a change of methodology and, if so, whether it rendered such events or estimates unsuitable for the purposes intended in the agreements. The Court considers that such a determination is to be reviewed by the Court applying the rationality test in accordance with the Court of Appeal’s decision in Barclays Bank. The Court notes that the wording of Clause 14.1 in the Swaps is on all fours with the wording in the Barclays Bank case.

93. The decision of Popplewell J in Barclays Bank, relied upon by Mr MacDonald Eggers QC, was of course appealed to the Court of Appeal and the judgment of the Court of Appeal makes it clear that the relevant test is the rationality test. That test is in accord with the holding of Popplewell J at first instance at [67(2)] that “The question is not whether the decision is justified, but whether a reasonable commercial man in Barclays’ position might have reached such a decision. ”

94. The relevant provision in Lehman Bros Special Financing Inc v National Power Corp [2018] EWHC 487 (Comm); [2018] 1 CLC 417, Robin Knowles J, (“commercially reasonable procedures in order to produce a commercially reasonable result”) in the ISDA 2002 Master Agreement was materially different and required the court to consider whether the determination was a “commercially reasonable result”. It is also to be noted that Robin Knowles J was following the decision of Briggs J in Lehman Bros International (Europe) (in administration) v Lehman Bros Finance SA [2012] EWHC 1072 (Ch) and Briggs J recognised at [82] that even if the concept of objective reasonableness were to be applied, a number of results may be commercially reasonable: “Plainly, that leaves a bracket or range both of procedures and results within which the Determining Party may choose, even if the court, carrying out the exercise itself, might have come to a different conclusion.”

95. In the circumstances, the Court concludes that Clause 14.1 imposes upon the Calculation Agent the standard of rationality (as opposed to a wholly objective test) in 35

the discharge of its functions under Clause 7 of the Swaps. It follows that a determination of the Calculation Agent would only be set aside if the Court is satisfied that a particular determination is one which no reasonable party could have made in the particular circumstances. This sets a high bar to impugn the determination of a contracting party and as Briggs J (as he then was) noted in Lomas v JFB Firth Rixson Inc [2010] EWHC 3372 (Ch) at [92] “It will nonetheless be a very rare case in which the apparently regular exercise of a purely contractual discretion can be successfully challenged…” (3) Operation of Clauses 7.1 to 7.4

96. The Defendants contend that the provisions contained in Clauses 7.1 to 7.4 set out a waterfall of alternative provisions for the selection of the successor Report Publisher. Mr Howe QC submits that each succeeding provision is conditional upon the previous provision having actually been invoked by the parties and having not resulted in the determination of a successor Report Publisher. Thus, the Defendants argue that, unless the provisions of Clauses 7.1 and 7.2 have been unsuccessfully invoked by the parties, it is not open to the Plaintiff to argue that a successor Report Publisher should have been appointed by the Calculation Agent under Clauses 7.3 and 7.4. The Defendants contend that if the parties have taken no action under Clauses 7.1 and 7.2, it is not open to the Calculation Agent to utilise the provisions of Clauses 7.3 and 7.4 to appoint a successor Report Publisher.

97. The opening paragraph of Clause 7 clearly imposes an obligation on the parties (“the parties shall”) to utilise the machinery provided in Clauses 7.1 to 7.4 (“using the following methodology”) to determine the successor Report Publisher.

98. Clause 7.1 imposes an obligation upon “the parties” to use “commercially reasonable efforts” to choose a “mutually acceptable” successor Report Publisher prior to the cessation of provision of Loss Reports or material changes to the methodology. 36

99. Clause 7.2 comes into effect “if a successor Report Publisher has not been determined under Paragraph 7.1” as of the date of cessation or material change in methodology.

100. Clause 7.3 comes into effect “if a successor Report Publisher has not been determined under Paragraph 7.1 or 7.2” within 45 days following the cessation or material change in methodology.

101. Clause 7.4 comes into effect “if a successor Report Publisher has not been determined under Paragraph 7.1, 7.2, or 7.3” within 60 days following the cessation or material change in methodology.

102. The scheme and purpose of these provisions, as Mr MacDonald Eggers QC correctly submits, is to ensure that if the Report Publisher ceases to provide Loss Reports or materially changes its methodology during the currency of the agreement, a successor Report Publisher is determined within a period of 60 days (or soon thereafter). The Court accepts Mr MacDonald Eggers QC’s submission that it is entirely irrelevant to this objective whether the parties made any effort to invoke Clauses 7.1 or 7.2. Furthermore, Clauses 7.2, 7.3, and 7.4 use the neutral language “if a successor Report Publisher has not been determined under” the previous paragraph. The language used does not require the parties to have made “commercially reasonable efforts” to choose a mutually acceptable successor Report Publisher. To impose such a requirement in relation to the machinery provided by Clauses 7.1 to 7.4 risks frustrating the clear objective of these provisions, namely, the appointment of a successor Report Publisher within a period of 60 days (or soon thereafter) after cessation or material change in methodology.

103. For the reasons set above, the Court concludes that it is not a condition to the invocation of Clauses 7.3 and 7.4 that the parties have invoked and discharged their obligations under Clauses 7.1 and 7.2. 37

(4) Was the Calculation Agent entitled to take the view that NCS had not ceased to provide any Loss Reports?

104. Having regard to the Court’s determination in relation to the role of the Calculation Agent at [68] to [76] above, the Court accepts that the question for the Court is not whether there was in fact a Reporting Disruption (as that term is defined in Clause 7 of the Swaps) but rather whether the Defendants, as Calculation Agent, acted in a commercially reasonable manner in determining that there was not.

105. In relation to this and the following issues, the Court has considered not only the written submissions provided on behalf of the parties at trial but also further documents setting out the transcript references and further correspondence dated 8 April 2022 and 12 April 2022.

106. In relation to this and the following issues, the Court heard factual evidence from Mr Kingham, on behalf of the Plaintiff, and from Mr Hilti and Mr Esteves-Ferreira on behalf of the Defendants. The Court also heard expert evidence from Mr Potter, on behalf of the Plaintiff, and Mr Held, on behalf of the Defendants.

107. The Court accepts Mr Howe QC’s submission that Mr Hilti, the primary factual witness for the Defendants, was honest and credible and gave direct evidence in relation to the issues before the Court. The Court also accepts that he was conscious that in addition to his other role he was also performing the separate obligations of the Calculation Agent.

108. In relation to the two experts, it is clear to the Court that Mr Held is by far the more experienced specifically in the field of loss reporting/report publishing and the reporting practices of the Report Publishers for ILW swaps. He was instrumental in the foundation of PERILS, a leading subscription-based loss reporter. Mr Potter accepted that his background is predominantly in commercial/investment banking and corporate finance. Mr Potter, unlike Mr Held, had no particular expertise in loss reporting. The 38

Court found Mr Held to be impartial, independent and balanced. To the extent that there is conflicting evidence, given by Mr Potter and Mr Held in relation to a particular issue, the Court prefers (unless indicated otherwise) the evidence given by Mr Held. Inherent risks/uncertainty in selecting an unofficial Loss Reporter

109. In considering this and the following issues, it is relevant to take into account that the parties accepted the inherent risk/uncertainty of using an unofficial Loss Reporter and of its loss reporting practices.

110. In his expert report, Mr Held states that, in his experience, there are a number of factors which would be considered relevant by the parties in considering the choice of the report publisher such as (a) the potential for basis risk, i.e. the risk that the index value correlates poorly with the actual loss experienced by the protection buyer (caused by one or more factors such as the methodology, reporting frequency, data sources, lack of data granularity, etc.), (b) the degree of independence of the report publisher (and any possible conflict of interests), (c) the charges and costs of the report publisher, and (d) an established track record of reporting in the region of interest. Mr Kingham accepted in cross-examination that when buying an ILW “you are inherently accepting the basis risk that goes with the report publisher” and that “there is an inherent uncertainty and lack of clarity of an unofficial loss reporter like NCS.” Mr Potter, the Plaintiff’s expert witness, also accepted that there is an inherent greater risk of uncertainty and lack of clarity in relation to an unofficial report publisher in an ILW swap as to whether there may or may not be a change of reporting or a change of determination of loss or methodology of loss estimates and of loss reporting.

111. In his expert report, Mr Held confirms that unofficial Report Publishers owe no obligations to parties to industry loss-based transactions. Indeed, they have never specifically intended that their loss information would be used as indices for industry loss-based transactions. They are not aware of any specific transaction, nor of any threshold, risk limit or any other terms of any particular transaction. They are not paid 39

for their information and their publications are not specifically designed for use as indices for industry loss-based transactions. Mr Kingham, in cross-examination, readily accepted these propositions.

112. Mr Held advises that NCS produces freely available publication series for Munich Re's clients and the public more broadly. Such publications include NCS Fact Sheets and Topics GEO. They are part of Munich Re's broad and highly regarded publication activities. NCS's information is used by market participants for risk management purposes and a range of research activities.

113. There are inherent risks attached to any loss reports produced by unofficial report publishers such as NCS. It is the evidence of Mr Held, which the Court accepts, that in relation to such reports, NCS does not commit to a certain minimum or maximum number of reports, nor to updating its estimates over time, nor to any pre-defined time intervals between the publication of any loss estimates for an event. That reflects the fact that NCS is simply a publisher providing information for market participants and is not specifically intended to be used as an index for loss-based transactions. Mr Kingham accepted that an unofficial report publisher: does not commit to any particular defined reporting schedule; does not commit to any particular defined reporting format; and does not commit to report on any particular catastrophe events with any given frequency or regularity.

114. As noted, unlike official report publishers, NCS does not commit to providing updates to its estimates and in Mr Held’s experience typically does not do so; an update is the exception rather than the rule. Generally, NCS provides just one estimate for a certain event. On occasion, it might mention an updated estimate for a loss occurrence in a subsequent publication but that could occur at any time and might be years after the event. It is not known in advance if NCS will update a loss number or if the first estimate given will be the only estimate ever issued by it. The Court accepts Mr Held’s evidence in this regard. 40

Calculation Agent’s belief that NCS typically reported once per event

115. As noted above at [41], it is the evidence of Mr Held that, generally, NCS provides just one estimate for a certain event.

116. Mr Hilti’s evidence is to the same effect. In cross-examination, Mr Hilti stated that his view was that “NCS typically would report once. So this would be typical when you enter such an industry loss warranty for a swap, you have the risk when you have a typhoon or an event later in the year that basically January reporting is – is not already having all the information available and it is an early point in time, that is part of that and, as I mentioned earlier, I think that is how we saw it, is that they typically report once and it was unknown to us whether if, and when they would report a second time on it. ”

117. Mr Held says that his experience is corroborated by the analysis he has conducted of the 382 natural catastrophe events included in the 2012 to 2021 issues of the NCS Topics GEO reports. Mr Held has sought to track whether any of the estimates for those events were updated by NCS in any subsequent NCS or Munich Re publication. Mr Held was only able to identify updates for 11 out of the 382 events. In other words, in 97% of the 382 events examined, NCS never produced an update.

118. To further illustrate this, Mr Held has analysed the NCS reports for the five largest Japanese natural catastrophes that occurred between 2011 and 2019. His analysis shows that the first loss estimate published by NCS in the first quarter of the year following the event in question typically remained the latest available loss estimate and was not updated. For Earthquake Tohoku (March 2011) NCS’s loss estimate of US$210 billion in Q1 2012 remained and was confirmed more than 5 years later. That remains the latest figure (which is 30% higher than Sigma’s estimate for the same event). For Typhoon Hagibis (which occurred in October 2019) the first loss estimate of US$10 billion published in 2020 continues to be the only figure published by NCS. (At the time of his first report, the NCS estimate was 25% higher than the latest Sigma number). For 41

Typhoon Faxai (September 2019) the loss estimate of US$7 billion from Q1 2020 continues to be the only figure published by NCS. For Earthquake Kumamoto (April 2016) NCS’s loss estimate of US$6 billion was published in January 2017. While a second estimate of US$6.5 billion was provided in June 2017, a different treatment of the fore- and after-shocks explains the difference.

119. In summary, Mr Held concludes that even for the four largest events occurring in Japan (other than Typhoon Jebi) there was no general practice of NCS publishing loss updates after releasing its first loss estimate (typically in its NCS Fact Sheet in Q1 of the year following the event). Did NCS cease to provide any Loss Reports?

120. Mr Howe QC emphasises that Clause 7 is referring to a situation where the Report Publisher ceases to provide any Loss Reports. He submits that in the circumstances of this case it is impossible to contend that this is the case. Mr Howe QC points to the fact that NCS in fact produced a Loss Report in relation to Typhoon Jebi in November 2020. The fact that the Loss Report in November 2020 was not within time for the Plaintiff’s purposes under the Swap, Mr Howe QC submits, is irrelevant for the purposes of this analysis.

121. Mr Howe QC also points out that, during this period, the market continued to use NCS as an unofficial Report Publisher on the basis that NCS was continuing to provide Loss Reports as it had done historically. Thus, correspondence produced to this Court shows that in January 2020 Guy Carpenter was proposing to designate NCS as the Report Publisher in relation to Japanese loss estimates in relation to another transaction. There is similar correspondence suggesting NCS as the Report Publisher for Japan on 10 February 2020 and 31 March 2020.

122. In the circumstances, the Court accepts Mr Howe QC’s submission that NCS plainly did not cease to provide Loss Reports. It provided a first loss estimate for Typhoon Jebi 42

and subsequently updated it in November 2020. It also continued to report loss estimates for other natural catastrophes in the meantime (e.g. in 2020 for 2019 events). The Plaintiff’s real complaint is that NCS did not update the estimate in time before the Development Period End Date on 30 June 2020. The Court accepts that was entirely consistent with NCS’s historic practice; even when it did update a loss estimate, that update might be years after the event, and there would be no certainty as to whether and if so when it would update. Accordingly, the Court concludes that NCS had not ceased to provide any Loss Reports in relation to Japanese windstorms. This finding supports the Court’s conclusion that the Calculation Agent was entitled to take the view that NCS had not ceased to provide any Loss Reports. (5) Was the Calculation Agent entitled to take the view that NCS had not materially changed its methodology?

123. The substance of this argument, on behalf of the Plaintiff, is based upon a publication by Munich Re in March 2018 headed “NatCatSERVICE Methodology” (the “Methodology Publication”).

124. The “Content” of the Methodology Publication is as follows:

1. Introduction

2. The Database 2.1 Database structure and content 2.2. Data sources 2.3. Peril classification 2.4. Geocoding 2.5. Loss data and loss estimation 2.6.Currency conversion 2.7. Inflation adjustment 2.8. Loss data normalization 2.9. Catastrophe classification (Cat Class) 43

2.10. Fatality data 2.11. Insurance density, insurance penetration

3. Products

4. Access and usage

125. The copy of the Methodology Publication produced to the Court, reproduces the document up to “Content” 2.11. The document produced to the Court does not expressly deal with the publication of this data and in particular the timing and frequency of any such publication.

126. The Plaintiff relies heavily upon “2.5. Loss data and loss estimation” and in particular the following passages: “Loss data and loss estimation Every nat cat loss event in the NatCatSERVICE database has an assigned direct economic loss expressed in nominal US$ values. Some events (where applicable) also include an insured loss … Our concept is to perform systematic loss estimations for all of the nat cat events ourselves and if our estimates are in line with other official estimates, we see this as confirmation of the robustness of our loss estimation procedures … The loss values in the NatCatSERVICE database are not static, i.e. they are updated every time there is new information available that leads to a re- evaluation of a loss estimate. This can affect both recent events and historic events.” Parties’ awareness of the Methodology Publication

127. In his second report, Mr Held confirms that, like Mr Potter, he was not aware of the Methodology Publication at the relevant time, and he does not believe that it was 44

generally known in the market. The fact is, Mr Held states, that in practice, as confirmed by the analysis in his first report, NCS did not in general update its loss estimates and if it did then it could not be predicted when it would do so, and (so far as he was aware) that was the general market view of NCS’s reporting practice (it was certainly his at the time). Mr Held considered that the market would have had regard to that practice in deciding whether there had been a reporting disruption (and he would have had regard to such practice as opposed to anything stated in a Methodology Publication, even if he had been aware of it at the time). In light of that practice as understood by the market, he did not consider that the fact that NCS did not provide any clarity as to if, and when, it might update its Typhoon Jebi estimate suggested a change in practice at NCS. On the contrary, providing such clarity would have been a change in practice by NCS.

128. Mr Held confirmed in cross-examination that he “was not really aware of [it] back in March 2018, 2019, I became only aware of this document in this discussion here and so if you would have asked me a year ago whether NCS has a published methodology my answer might have been I am not aware of.”

129. Mr Potter, the Plaintiff’s expert, gave the same confirmation in cross-examination that he was not aware of the Methodology Publication prior to this dispute in July 2020.

130. Likewise, Mr Kingham confirmed that he was not aware of the Methodology Publication when NCS was selected as the Report Publisher for the Swaps and only became aware of this publication some time in 2019. Relevance of NCS’s historical practice

131. As noted above, Mr Held considered that the market would have had regard to NCS’s practice in deciding whether there had been a reporting disruption and he would have had regard to such practice as opposed to anything stated in a Methodology Publication. 45

132. In cross-examination, Mr Kingham confirmed that, without a defined contractual benchmark or yardstick in relation to a particular ILW swap, by reference to which to measure the unofficial report publisher's conduct, there may be a period of uncertainty during which it may be unclear if an unofficial report publisher has or has not changed its practice. Mr Kingham accepted that is part of the inherent basis risk when the parties to an ILW swap choose a particular unofficial report publisher at the outset of the contract.

133. Mr Potter accepted in cross-examination that the question of whether there has been a reporting disruption depends, at least in part, on whether there has been a material change in the reporting of the particular report publisher in question which affects the suitability of its reporting for use for the intended purpose of the swaps. In his view, it is the essence of that determination to consider the past practice of the report publisher in terms of its reporting and compare it with its current reporting at the time at which it might be being suggested by somebody that a material change has occurred. Mr Potter stated in deciding whether there had been a reporting disruption in respect of NCS for Japanese windstorm in connection with section 4 of the Swaps with which this Court is concerned, the key is to analyse NCS's actual previous practice and compare it with its "current" practice, i.e. its practice at the time at which it may be suggested by somebody that a material change had occurred.

134. All the witnesses made it clear that their expectations of how NCS would perform as a Report Publisher were based upon how NCS had performed in the past.

135. Mr Kingham, in his first witness statement, confirms that when choosing NCS as a Report Publisher under the Swaps, the Plaintiff expected NCS “to continue reporting in materially the same way that it had done in the previous years.” He states that these expectations were based on his knowledge and experience of NCS and its reporting in relation to previous catastrophes, and in particular NCS’s reporting patterns and the way its estimates related to the estimates produced by other report publishers over a number of years, 46

136. Mr Potter’s evidence is to the same effect. In his first report he confirmed his belief that in considering whether an unofficial report publisher had ceased to provide loss reports or had changed its methodology, such a determination would include consideration of how the report publisher had acted prior to its selection and how their past behaviour compared with the report publisher’s current actions in response to a catastrophic event in relation to which the parties were relying on the report publisher to provide a loss estimate. Mr Potter states that in this case, that meant comparing NCS’s actions over the Development Period under the Swaps with its actions in the period immediately preceding the commencement of the Swaps.

137. Mr Hilti, in his first witness statement, states with reference to the Methodology Publication that in his experience there was no established methodology for periodic reporting, or contractual (or other) obligation on NCS to report on a particular loss occurrence. If it did, it was unpredictable whether or not it might produce an updated loss figure; it was not known in advance whether (and if so when) NCS would update its loss estimate for a specific natural catastrophe with any particular frequency or at any particular time. The Court accepts Mr Hilti’s evidence in this regard.

138. The Court has already held at [114], accepting the expert evidence of Mr Held, that, generally, NCS provides just one estimate for a certain event. On occasion, it might mention an updated estimate for a loss occurrence in a subsequent publication but that could occur at any time and might be years after the event. It is not known in advance if NCS will update a loss number or if the first estimate given will be the only estimate ever issued by it. The Court accepts that past practice of loss reporting by NCS is critical to determining whether there had been material change of methodology. The Court does not consider that NCS’s past practice, as accepted by the Court, supports the Plaintiff’s contention that there had been a material change in NCS’s methodology in relation to its loss reporting of Typhoon Jebi. Industry expectation 47

139. The Plaintiff argues that there was general widespread, uniform industry expectation that NCS was likely to revise its loss estimate for Typhoon Jebi upwards from the initial estimate of $9 billion in January 2019. The Court accepts Mr Howe QC’s submission that the evidence actually shows a diverse range of views based ultimately on the uncertainty, lack of information and inherent speculation about what might or might not happen in relation to the loss reporting of Typhoon Jebi by NCS. The Court accepts Mr Howe QC’s submission which is supported by the following evidence.

140. On 2 May 2019, Artemis, an industry publication, reported that the Bermudian insurance and reinsurance group Arch Capital believes the industry loss from Typhoon Jebi could now be as high as $13 billion. The article expressly acknowledged that Swiss Re’s Sigma had recently estimated an industry loss of $9.77 billion and “Munich Re’s NatCat which had stuck at $9 billion for Jebi.” In relation to future expectations from NCS the article stated “Neither of these are official reporting agencies for trigger purposes, but they have been widely used historically… They also tend not to continue updating losses for long after the events occurred.” (emphasis added).

141. On 16 August 2019, under the heading “Jebi shows ILW triggers are better when official & supported”, the Artemis article notes that Sigma’s estimate remains at $9.77 billion and that NCS’s is even lower at $9 billion. The article states that there was a view in the market that the industry loss estimates from these two sources remain far below the generally accepted market loss for Typhoon Jebi. However, the article goes on to explain why this may be the case: “To which it is important to note, these are not official ILW trigger data sources and never have they professed to be such a thing or been sold as one. Update: We’re told that in some cases Sigma and NatCat estimates are updated with development, so it’s possible they will be also in these cases. … In fact, Munich Re tends not to provide continuous updates to NatCat estimates and Swiss Re only does so for Sigma once a year. 48

… So, any parties to an ILW, or other industry loss based risk transfer arrangement, that is exposed to Typhoon Jebi and uses Sigma or NatCat linked trigger currently have no idea if or when estimates will be updated, leaving the contracts hanging and potentially still at risk. … This shows the importance of risk transfer triggers being official, supported, documented, defensible, updated on a reasonably regular schedule, as transparent as possible, and reported internally free of any potential for bias. … Buyer beware, if you really must use unsupported and unofficial sources of data within risk transfer triggers, there will be limited recourse when things don’t turn out quite the way you envisaged they would.” (emphasis added).

142. In September 2019, Guy Carpenter officials were expressing the view that “they are certain that NatCatSERVICE will not update their Jebi industry loss estimate.” Apparently, this was based upon information received from Aon.

143. On 17 December 2020, an article headed “Munich Re lifts Jebi industry loss to $13 bn, some ILW’s potentially in play” noted that the NCS industry loss estimate that was at US$9 billion for well over a year, below the US$10 billion point that is often selected as a trigger or attachment point for risk transfer contracts, but then rose 44% to US$13 billion, shows that there can be certain risks associated with the use of unofficial loss triggers. The article concludes, “That’s a moral hazard for those using an unofficial industry loss trigger though, as it’s not the responsibility of any unofficial source such as Munich Re, given its NatCat data is a research driven exercise to make catastrophe loss data available to clients and partners, not a service it ever intended to be used in this way.” (emphasis added). Variation in loss estimates between NCS and Sigma and PCS 49

144. As noted by Mr Held in his second report, Mr Potter suggests that the difference between Sigma and NCS prior to NCS’s update was such that the NCS loss estimate for Typhoon Jebi was then “out of date” and did not factor in the significant loss creep that occurred during 2019. Mr Held does not agree that there was something exceptional or unique about the loss development of Typhoon Jebi. Mr Held noted that the latest market loss estimate for Typhoon Jebi was US$12.052 billion (source: CLIX, from 31 December 2021), that is only 33% higher than NCS’s original estimate of US$9.0 billion from nearly 3 years prior. A loss development of around 33% cannot in Mr Held’s experience be considered unique, especially for events outside of the US. Further, Mr Held states that the loss development associated with Typhoon Jebi is in notable contrast to the much greater loss development observed over time for other large catastrophic events such as the tripling of industry loss estimates for the earthquakes in New Zealand or the more recent devastating floods in Germany from July 2021 for which the German Insurance Association GDV increased their initial estimate for the insured industry loss by more than 80%, from EUR 4.5 billion to EUR 8.2 billion.

145. In cross-examination, Mr Held confirmed that the fact that there was a big loss creep was not exceptional. He stated that there are many other events where there have big increases from initial estimates to final estimates. The floods in Germany, in his view, are a good example, being the latest big multi-billion dollar event. The German Insurance Association had to revise the loss of EUR 4.5 billion to EUR 8.2 billion within a period of less than six months. Accordingly, there was an 80% increase of the loss creep. Mr Held also referred to Storm Uri in 2021 where there was a similar 60% increase from US$10 to US$17 billion or US$18 billion. In the circumstances, it was Mr Held’s opinion that an increase of 30% or 32%, as is the case in relation to Typhoon Jebi, was no exception. Mr Hilti also confirmed the same view, noting that increase in the early days was normal and the loss increases in relation to Typhoon Jebi, in his view, were “normal”. 50

146. In this connection, the Court also notes that, on 31 October 2019, Aon/AG were clearly of the view that, despite the variance in loss estimates between NCS and other report publishers, this was not indicative of Reporting Disruption. In an email to the Plaintiff dated 30 October 2019, Aon/AG advised that “Mutual agreement will be required to refer to an alternative index. If PCS Japan, for example, CS ILS would be agreeing to a loss position. Aon/AG will, of course, approach CS ILS on the basis of seeking an alternative index but realistically cannot see Niklaus [Hilti] agreeing to this course, at this stage, given the above.” (emphasis added). This email reflects that, as at 30 October 2019, Aon/AG were not of the view that there had been a Reporting Disruption.

147. The Court accepts the evidence of Mr Held in relation to the lack of significance in the variation in loss estimates between NCS and Sigma, and that the loss creep in relation to Typhoon Jebi was not exceptional. Timing risk

148. The timing of the loss reporting by the Report Publisher is of course critical where market participants have used a particular index for ILW transactions.

149. Mr Held noted in cross-examination that, whilst there were similarities between NCS and Sigma, that did not apply to the updating practices of the two organisations. Mr Held noted that Sigma always updated their losses, so in the appendix to the Q1 publication they would show the top 40 insured losses which had been updated. Mr Held pointed out that that was not the case with NCS. NCS was not updating losses for publication on a regular basis. NCS was only updating single events on an exceptional basis.

150. It was put to Mr Held that the Methodology Publication clearly contemplated that NCS will revise loss estimates upon receipt of information which would justify such a 51

revision. Mr Held disagreed. Mr Held took the view that the wording of the Methodology Publication did not say when NCS would publish the revised loss estimate.

151. Mr Held remained of the view that NCS would publish revised loss estimates “if appropriate”: “I mean from the perspective of NCS to decide and I don't think they would do it every time because otherwise my analysis would not show the low number of 3% which are actually updated, because for every loss in that database there has been loss development and for every loss event there could have been reasons to say it is appropriate to do it. But if I say "appropriate" I really mean from their perspective.”

152. This is consistent with Mr Hilti’s evidence that the Methodology Publication is “silent on timing and exact thresholds and details, how they update or not update.”

153. The Court accepts the evidence of Mr Held and Mr Hilti in relation to the issue of timing of the publication of loss reports. The Court also accepts their evidence that the Methodology Publication is silent in relation to the issue of when the updated information in the database would be published to the world at large. The Court has already accepted that the historical evidence shows that NCS provides just one estimate for any given event. On occasion, it might mention an updated estimate for a loss occurrence in a subsequent publication but that could occur at any time and might be years after the event.

154. Having regard to the evidence of Mr Held and Mr Hilti, which the Court accepts, in relation to (i) the parties’ awareness of the Methodology Publication, (ii) the relevance of NCS’ historical practice relating to reporting of losses, (iii) the industry expectation of a revised loss publication from NCS, (iv) variation in loss estimates between NCS and Sigma, and (v) whether the loss creep in relation to Typhoon Jebi was exceptional, the Court concludes that there was no material change in the methodology of NCS. This conclusion accords with the fact that at no time during the relevant period did the Plaintiff ever suggest to the Calculation Agent that it was of the view that there had 52

been either a cessation of the provision of Loss Reports or a material change in the methodology. In the circumstances, the Court concludes that the Calculation Agent was entitled to take the view that NCS had not materially changed its methodology. (6) Was the Calculation Agent entitled to take the view that NCS had not materially changed the methodology merely because of the abbreviation of the annual loss reports from 2019 onwards?

155. In relation to this issue, the Plaintiff contends that, for the years 2019 and 2020, NCS abandoned its prior practice of publishing a substantial 60-page publication titled Topics GEO and introduced the new practice of publishing a fact sheet online. It is argued on behalf the Plaintiff that, had the previous practice continued, it was likely that there would be a review of the previous catastrophes including Typhoon Jebi and in review an opportunity could possibly have emerged to update the losses in relation to Typhoon Jebi.

156. In cross-examination, Mr Held stated that this was pure speculation, and he could not share the view that if there had been an extensive Topics GEO 2020 that there would have been a Typhoon Jebi update.

157. First, Mr Held does not agree with Mr Potter’s assertion that there was no Topics GEO report issued in 2020. Mr Held states that this is not accurate given that on 9 January 2020, there was a Topics GEO report published for events from 2019. It was labelled “Topics online” and it was in exactly the same format as the Topics online released in January 2019 for events from 2018.

158. Second, Mr Held refers to Mr Potter’s assertion that it was unusual that NCS’s annual fact sheet issued in January 2020 did not contain any loss updates for events which had occurred in 2018 (such as Typhoon Jebi) or earlier. Mr Held states that is not correct. In his experience, these Q1 reports typically only included estimated loss numbers from events which occurred in the immediately previous year and not for events which had 53

occurred in earlier years. In cross-examination, Mr Potter stated that he would not dissent from this observation.

159. Third, as noted earlier, Mr Held explained that of the approximately 382 natural catastrophe events examined by him during the period 2012 to 2021, NCS only updated estimates in relation to 11 of those events. In relation to those 11 events, only three were updated through a Topics GEO report and the other eight were not.

160. In the circumstances, the Court accepts the expert evidence of Mr Held that the suggestion that, had the previous format of Topics GEO continued, there was a likelihood that it would have contained an update in relation to the Typhoon Jebi losses, is speculation and can be ignored by the Court. Accordingly, the Court determines that the Calculation Agent was entitled to take the view that NCS had not materially changed the methodology merely because of the abbreviation of the annual loss reports from 2019 onwards. Relevant/irrelevant considerations

161. In his submissions, Mr MacDonald Eggers QC emphasised that, in relation to the requirement of “good faith”, he was not suggesting that the Calculation Agent had been dishonest in anyway but did suggest that the Calculation Agent either took irrelevant considerations into account or failed to take into account considerations which were relevant to the determination. For completeness, the Court does not accept this submission. It is clear from the evidence of Mr Hilti that relevant considerations were indeed taken into account. The Court accepts Mr Hilti’s evidence in this regard: Q. …Credit Suisse were also wrong not to even consider whether there was a reporting disruption by NCS by end of 2019? A. I think, according to our logic, we looked at everything we have to look, and we did not find it, that this is a reporting disruption. … 54

Q. I think your evidence is that you… consulted a lawyer… and you considered - enquired into, thought about whether or not there was a reporting disruption? A. Yes. … Q. And you also took into account the Milliman projected probability? A. That was still unchanged, yes. Q. And you took … into account what you were told by Guy Carpenter? A. That was the confirmation, yes. Q. So in those circumstances, no reasonable person could have come to the conclusion … that there was no reporting disruption; that’s right? A. That is what you say. I disagree. Implied term

162. Having regard to the findings made by the Court, it is unnecessary to consider in any detail whether an implied term that the termination of the Swaps would only validly occur if the Defendants properly exercised their duties as Calculation Agent. However, had it been necessary to decide that issue, the Court would have declined to imply such a term on the basis that (i) the implication of such a term is unnecessary since the contract works perfectly well without it, and (ii) the implication of such a term would be inconsistent with the express terms of the Swaps regulating their termination which form a complete contractual code. I. Conclusion

163. Having regard to the findings made by the Court, the Court concludes that, in the circumstances under consideration, NCS, as the Report Publisher, did not cease to provide any Loss Reports or materially change its methodology. In the circumstances, the Calculation Agent was entitled, in good faith and a commercially reasonable manner, to take the view that NCS, as the Report Publisher, had not ceased to provide any Loss Reports or materially changed its methodology. 55

164. Accordingly, the Plaintiff’s claim for damages and/or a declaration that the plaintiff is entitled to be paid the sum held in the Trust Accounts in the total sum of US$20 million is hereby dismissed. The Court grants the Defendants the relief sought in the Counterclaim, namely, a declaration that each of the Defendants is entitled to receive from the Plaintiff the collateral which they have deposited in the Trust Accounts.

165. The Court will hear the parties in relation to the issue of costs, if required. Dated this 6th day of May 2022 _______________________________

NARINDER K HARGUN

CHIEF JUSTICE

56