Soon Kok Tiang and others v DBS Bank Ltd and another matter

Chan Sek Keong CJ (delivering the judgment of the court):

[2] The material facts as pleaded by the parties are not in dispute. The Respondent launched the sale of the HN5 on 30 March 2007 in two tranches, one in Singapore dollars (“the SGD Tranche”) and one in US dollars (“the USD Tranche”). The Respondent’s offer of the HN5 was expressed to be made on the basis of the information contained in a 93-page pricing statement dated 29 March 2007 (“the Pricing Statement”) and a 168-page base prospectus dated 22 December 2005 as amended by a supplementary base prospectus dated 5 April 2006 (collectively referred to hereafter as “the Base Prospectus”). These documents set out in great detail, inter alia: (a) the terms and conditions applicable to the HN5; (b) the manner in which the Respondent and its related entity, Constellation Investments Ltd (“Constellation”), would use the funds of purchasers of the HN5 (“HN5 holders”); and (c) the risks involved in investing in the HN5.

[3] The sale of the HN5 was effected by the Respondent accepting applications made by prospective HN5 holders on prescribed application forms (“Application Forms”). Each Application Form contained an acknowledgement by the prospective HN5 holder that he:

(a) “agreed [to] the Terms and Conditions on the reverse of this form and the Terms and Conditions set out in … [the] Pricing Statement”; and

(b) “ha[d] assessed [the] suitability of the product against [his] risk attitude, financial means, and investment objectives”.

At the time the HN5 were marketed, it was not possible to assess the probability of default by any of the reference entities to which the HN5 were linked, except on the basis (since discredited) of the credit ratings given to them by the credit rating agencies relied on by the Respondent, viz, Standard & Poor’s, Moody’s and Fitch.

[4] The Pricing Statement described the HN5 as “structured notes” and warned prospective HN5 holders that “if a Credit Event [as defined at [8] below] … occur[red] before the Maturity Date [viz, 16 November 2012], [they might] lose their entire investment and [might] not receive any principal amount on the [HN5]” [emphasis in original omitted]. The circumstances in which and the reasons why this might happen were not clarified or explained. Prospective HN5 holders were also warned that the purchase of the HN5 involved certain risks, and that they should:

[5] The Pricing Statement described the HN5 as “5.5-year structured credit notes designed for investors seeking enhanced yield by providing exposure to a first-to-default basket of geographically diversified investment grade credits”. The credit basket (“the HN5 credit basket”) consisted of seven banks and one sovereign entity as follows:

These eight reference entities (collectively, “the Reference Entities”) were stated to have been selected with the following objectives:

[6] The HN5 were structured such that the Respondent would use the funds raised from the sale of the HN5 to purchase two structured notes issued by Constellation, a special purpose Cayman Islands trust company established by the Respondent in 2003. These two structured notes (collectively, “the Reference Notes”) were:

(a) Structured Retail Notes Series 75 (“SRN Series 75”) vis-à-vis the SGD Tranche of the HN5; and

(b) Structured Retail Notes Series 76 (“SRN Series 76”) vis-à-vis the USD Tranche.

The Pricing Statement explained Constellation’s role and the function of the Reference Notes as follows:

[7] It should be noted that although the Pricing Statement stated that Constellation would use the funds raised from the issue of the Reference Notes (“the Reference Notes funds”) to invest in “high quality bonds or structured securities with a rating of AA or better (as rated by Standard & Poor’s, Moody’s or Fitch)” to secure its obligations under the Reference Notes, it did not say that those funds would be used to purchase, specifically, what we will hereafter term “the Reference Obligations” (ie, the bonds or structured securities issued by the Reference Entities as set out in Appendix D of the Pricing Statement), all of which were rated lower than “AA” (or the equivalent). The Pricing Statement also did not explain what bearing Constellation’s investments would have on the risks undertaken by HN5 holders. To date, the Appellants have no knowledge of the identity of the actual high-quality bonds and/or structured securities purchased by Constellation to secure its obligations under the Reference Notes.

[8] The Pricing Statement stated that HN5 holders would receive:

(a) quarterly interest of 5% per annum and 6.5% per annum for, respectively, the SGD Tranche and the USD Tranche of the HN5 until 16 November 2012, which was the scheduled maturity date of the HN5 (“the Maturity Date”); and

(b) on the Maturity Date, 100% of the principal amount invested.

The Pricing Statement stated that the aforesaid payments would be made unless, inter alia, a “Credit Event” – ie, an event of default in respect of any of the Reference Entities – occurred prior to the Maturity Date. If a Credit Event occurred, the Reference Notes (which were stated to be linked to the performance of the Reference Entities) would terminate, and would also cause the HN5 to terminate. HN5 holders would then receive an amount equivalent to what the Respondent would receive as the “Credit Event Redemption Amount” (“CERA”) under the Reference Notes. This was explained in the Pricing Statement as follows:

The Pricing Statement stated that the Respondent, as the calculation agent under both the Reference Notes and the HN5 (“the Calculation Agent”), “ha[d] the sole discretion to determine whether a Credit Event ha[d] occurred in relation to any [r]eference [e]ntity”, and was “also responsible for determining, amongst other things, any … calculations … required under the Terms and Conditions of the [HN5]”.

[9] The Pricing Statement highlighted that the termination of the HN5 upon the occurrence of a Credit Event was on a “first-to-default” basis, ie:

[10] As for what would constitute a Credit Event, this was defined in Appendix D of the Pricing Statement as follows:

In this regard, “Failure to Pay” was defined as the failure of a reference entity to pay any amounts due under (inter alia) the Reference Obligation associated with it as specified in Appendix D of the Pricing Statement. In Lehman’s case, the relevant Reference Obligation consisted of US$1.25bn worth of 5.75% subordinated notes due on 1 March 2017 (ISIN Code: US524908UB47) (“the Lehman Note”).

[11] The HN5 were credit-linked notes and not asset-linked notes, and were a novel financial product in Singapore’s financial market at the time they were launched. From the foregoing survey of their nature and structure, the following features can be seen.

[12] First, the HN5 had some features of an ordinary bond, but they were not bonds as HN5 holders, apart from being exposed to the credit risk of the Reference Entities, were also exposed to the credit risk of Constellation and that of the issuer(s) of the high-quality bonds and/or structured securities purchased by Constellation to secure its obligations under the Reference Notes. As stated in the Pricing Statement:

In addition, HN5 holders were exposed to the credit risk of the Respondent, foreign exchange rate risks, liquidity risks and general market fluctuations. Vis-à-vis, specifically, the risk of default by one or more of the Reference Entities (ie, the risk of a Credit Event occurring), it would appear, at the time the HN5 were launched, that this risk was not likely to materialise, having regard to the high credit ratings given to the Reference Entities (we note in particular that at that time, Lehman had a credit rating of “A+” from Standard & Poor’s and Fitch, and a credit rating of “A1” from Moody’s). The Pricing Statement also stated:

Again, this would give the impression that it was unlikely that a Credit Event would occur vis-à-vis the Reference Notes, to which the performance of the HN5 was directly linked. The Pricing Statement nonetheless warned prospective HN5 holders that the credit ratings assigned by credit rating agencies might not fully reflect the true risks of investing in the HN5. In comparison, the Respondent assumed no risk in the event of a Credit Event occurring (and, consequently, the HN5 terminating before the Maturity Date) as the Pricing Statement expressly stated (at pp 2, 4 and 14) that the Respondent’s liability under the HN5 in that event was only to pay to HN5 holders what it would receive from Constellation as the CERA under the Reference Notes.

[13] Second, the components for calculating the CERA were such that if a Credit Event occurred, there was a high likelihood that HN5 holders would receive a zero payout under the HN5 because there was an equally high likelihood that the Respondent would receive a zero payout from Constellation under the Reference Notes. In fact, as it turned out, the computation of the CERA to be paid upon Lehman’s bankruptcy (the Credit Event in the present case) actually yielded a negative figure. In practical terms, this translated into a zero payout for HN5 holders since their losses were limited to the principal amount which they had invested in the HN5. Thus, stripped to their bare essentials, the HN5 were in substance a bet or wager (in legal jargon) that no Credit Event would occur during the 5.5-year tenure of these notes: if HN5 holders “won” this bet, they could potentially receive at the end of the 5.5-year period an aggregate return of 27.5% and 35.75% on, respectively, the SGD Tranche and the USD Tranche. In reality, the HN5 were a bet dressed up and marketed as an “investment”.

[14] Third, the explanation given at p 1 of the Pricing Statement for selecting eight reference entities (viz, the Reference Entities) to make up the HN5 credit basket – namely, to (inter alia) “reduc[e] the likelihood of a Credit Event” (see [5] above) – could give an unwary HN5 holder a false sense of security that the risk of a Credit Event occurring was low. It was only at p 9 of the Pricing Statement that the aforesaid explanation was qualified by the statement that “all else being equal, the greater the number of [r]eference [e]ntities in the [HN5] credit basket, the riskier the [HN5] are”. In this connection, we note that before both the Judge and this court, the Appellants did indeed labour under the misapprehension that having eight reference entities in the HN5 credit basket would provide diversification benefits. This was evidenced by their argument that they should not have lost their entire investment in the HN5 as Lehman was only one of eight reference entities.

[15] As mentioned at [2] above, the HN5 were launched on 30 March 2007. Altogether, the Respondent accepted 1,127 applications for the HN5, totalling S$91.915m for the SGD Tranche and US$7.93m for the USD Tranche. The HN5 were issued on 16 May 2007. Between 16 August 2007 and 18 August 2008, the Appellants received five payments of quarterly interest, amounting to S$128,303.27 for the SGD Tranche and US$22,741.07 for the USD Tranche.

[16] Unfortunately for HN5 holders, on 15 September 2008, Lehman became the first reference entity in the HN5 credit basket to default (“the Defaulted Reference Entity”) when it filed a voluntary petition under ch 11 of the US Bankruptcy Code (Title 11, USC). By a Credit Event notice dated 17 September 2008, the Respondent informed HN5 holders of the occurrence of this Credit Event and the termination of the HN5. Subsequently, by a letter dated 28 October 2008, the Respondent informed HN5 holders that the CERA under the Reference Notes was zero and, thus, the CERA under the HN5 would likewise be zero. The letter was accompanied by an appendix explaining how the CERA under the Reference Notes was calculated.

[17] The Respondent also issued a public statement on 27 October 2008 setting out a list of questions and answers in relation to the Credit Event triggered by Lehman’s bankruptcy (“the FAQs”) (see <http://www.dbs.com/dbsgroup/announcements/Pages/20081028HN5_Final_Valuation_Issuer_FAQ.pdf> (accessed 31 October 2011)). It is pertinent to note the answer to the question as to why the CERA under the Reference Notes (and, in turn, the CERA under the HN5) was zero:

[18] It should be mentioned that although the FAQs stated that the CERA was zero because of the “notional loss” on the Lehman Note, the Pricing Statement did not expressly highlight that the link between the CERA and the Defaulted Reference Entity’s Reference Obligation was “notional”. This fact becomes clear only upon a close analysis of the terms of the Pricing Statement. For the purposes of calculating the CERA to be paid upon the occurrence of a Credit Event, any loss sustained on the Defaulted Reference Entity’s Reference Obligation might only be a notional loss as the Pricing Statement did not require Constellation to use the Reference Notes funds to purchase that particular Reference Obligation or, for that matter, any of the Reference Obligations, and, thus, Constellation would not have suffered any loss if it had not invested the Reference Notes funds in the Defaulted Reference Entity’s Reference Obligation, whose value would have fallen upon the Defaulted Reference Entity’s default. In other words, the HN5 were structured such that it was not necessary for Constellation to have suffered any actual loss on the Defaulted Reference Entity’s Reference Obligation before a Credit Event could occur. All that was necessary was for one of the Reference Entities (by definition, the Defaulted Reference Entity) to (inter alia) default in paying any amount due under the Reference Obligation associated with it, which Reference Obligation might not even form part of Constellation’s actual investment portfolio. In this connection, we note that it is not known whether Constellation had in fact invested the Reference Notes funds in the Lehman Note (the Defaulted Reference Entity’s Reference Obligation in the present case).

[19] In the light of the terms and conditions of the contract governing the HN5 (“the HN5 contract”), the Appellants had little or no legal basis to argue that they were entitled to receive anything other than (as will be seen) a zero payment from the Respondent when the Credit Event triggered by Lehman’s bankruptcy occurred. They could not claim that they were not bound by the terms and conditions of the HN5 contract since (as mentioned at [3] above) they had expressly agreed to those terms and conditions in the Application Forms which they submitted to the Respondent. They also could not rely on the usual vitiating factors in contract law, such as fraud, mistake, misrepresentation, non est factum, etc. In particular, the Appellants could not claim that they did not understand the true risks of investing in the HN5 as they had acknowledged in their Application Forms that they “ha[d] assessed [the] suitability of the product against [their] risk attitude[s], financial means, and investment objectives”.

[20] The Appellants then discovered, probably to their surprise, that the Pricing Statement contained four different descriptions of the CERA (“CERA Descriptions”), which they interpreted to be inconsistent and irreconcilable with one another (and in the case of the fourth description, manifestly inconsistent with the third). This interpretation, if correct, would lead to different CERA payouts under the HN5, depending on which CERA Description was applied. Armed with this discovery, the Appellants commenced the OS against the Respondent for, in the main: (a) a declaration that the HN5 were void; and (b) a consequential order that the Respondent pay to each of them (as well as to each of the 192 other plaintiffs), inter alia, the principal amount invested in the HN5. The Appellants’ basis for seeking these reliefs was that:

[21] As mentioned at [1] above, the Judge dismissed the OS. The Appellants then brought the present appeal.

[22] Before we proceed to analyse the Judge’s reasons for dismissing the Appellants’ claim, we will first set out the four different CERA Descriptions in the Pricing Statement.

[23] The first CERA Description, which appears at p 2 of the Pricing Statement, reads as follows:

In this connection:

(a) “Charged Asset Adjustment Amount” is defined at p 14 of the Pricing Statement as “the shortfall of the market value of the Charged Assets over the Principal Amount of the Charged Assets” (the “Charged Assets” being the high-quality bonds and/or structured securities purchased by Constellation to secure its obligations under the Reference Notes (see [6]–[7] above)); and

(b) “Hedging Costs” are defined at p 15 of the Pricing Statement as “the losses, expenses and costs (if any) to Constellation in terminating, adjusting or re-establishing etc the underlying or related hedging arrangements”.

[24] The second CERA Description, which is at p 4 of the Pricing Statement, provides as follows:

[25] The third CERA Description (at pp 14–15 of the Pricing Statement) states:

For ease of reference, we will hereafter, where appropriate, use the same terms and abbreviations as those used in the above extract (ie, “Aggregate Principal Amount”/“APA”, “CAAA”/“Charged Assets Adjustment Amount”, etc) in discussing the CERA.

[26] The fourth CERA Description is set out at p 61 of the Pricing Statement in diagrammatic form. We reproduce below the Judge’s text-based transliteration of that description (see sub-para (d) of [14] of the HC Judgment):

[27] The Appellants formulated the four CERA Descriptions as follows:

[28] The issues raised by the parties and dealt with by the Judge in the court below can be summarised into two main issues. The first concerned the terms and conditions of the HN5 contract. The second was whether the four CERA Descriptions were different, inconsistent or irreconcilable, or otherwise unworkable such that they rendered the HN5 contract void for uncertainty.

[29] With respect to the first main issue, the Judge held that the HN5 contract was governed by “the terms and conditions set out on the reverse side of the Application Form and the entirety of the … Pricing Statement, including but not limited to the section of the Pricing Statement entitled ‘Terms and Conditions of the Notes’ and Appendix A” [emphasis added] (see [36] of the HC Judgment). The Respondent has not appealed against this finding, even though it is contrary to its argument in the court below that the terms and conditions of the HN5 contract consisted of only the terms and conditions on the back of the Application Form and certain portions of the Pricing Statement.

[30] With respect to the second main issue, the Judge rejected the Appellants’ contention that the CERA could not be ascertained due to the four inconsistent CERA Descriptions present in the Pricing Statement, and that the HN5 contract was thus void for uncertainty. His reasons for so ruling are set out in the HC Judgment as follows:

[31] In summary, the Judge accepted the Respondent’s arguments and held that the third CERA Description was the operative CERA Description because it was precise and was expressed in the form of a mathematical formula, whereas the first and second CERA Descriptions were verbal formulations and were therefore imprecise. The Judge also accepted the Respondent’s argument that contextually, the fourth CERA Description contained an obvious clerical mistake, in that a literal reading of the formula set out therein would give rise to an absurdity. The Judge further held that the clerical mistake could be corrected simply by construing the fourth CERA Description by removing “1 –” (which was the cause of the mistake). Applying the formulation set out in the third CERA Description, the Judge held that the Appellants were entitled to only a zero payout.

[32] The Appellants raised two issues (collectively, “the Issues”) before us for decision. The first was whether the Judge was right in ruling that the third CERA Description was the operative CERA Description (“the First Issue”). The second was, regardless of whether it was the third or the fourth CERA Description (these being the only two CERA Descriptions addressed by the Appellants on appeal) which was the operative CERA Description, whether the HN5 contract was void for uncertainty because both the third and fourth CERA Descriptions were unworkable in that they did not provide any agreed basis for calculating the CERA (“the Second Issue”).

[33] The Appellants’ arguments before us on the Issues were basically a reiteration of those arguments that had been rejected by the Judge.

[34] With respect to the fourth CERA Description, the Appellants argued that the formula set out in it was not absurd and would not lead to an absurd result because:

(a) under that formula, in the event of a Credit Event occurring, a HN5 holder would sustain significant losses of approximately 72% of the principal sum invested; and

(b) upon the occurrence of a “Constellation Event” (ie, early redemption of the Reference Notes for any reason other than the occurrence of a Credit Event or an “Issuer Call” (viz, the exercise by the Respondent of its right of early redemption of the HN5)), assuming the same factual conditions as those in the present appeal applied, the payout which a HN5 holder would receive would be the same as that which he would receive under the fourth CERA Description.

Hence, the Appellants contended, the fourth CERA Description did not contain any obvious clerical mistake. It followed that the Judge was not entitled to correct or rectify that CERA Description to make it consistent with the third CERA Description as the conditions for correction by construction as laid down in East v Pantiles (Plant Hire) Ltd [1982] 2 EGLR 111 were not satisfied.

[35] The Appellants further argued that regardless of whether this court were to uphold the Judge’s decision that the third CERA Description was the operative CERA Description or, alternatively, rule that the fourth CERA Description was the operative one, neither of these CERA Descriptions was workable as one of the elements common to both descriptions – ie, the Final Price – could not be calculated and, therefore, the CERA could not be ascertained. The reason, the Appellants explained, was that “Final Price” was defined under both the third and fourth CERA Descriptions as the price of the Defaulted Reference Entity’s Reference Obligation “expressed as a percentage”. However, as the phrase “expressed as a percentage” was nowhere defined in the Pricing Statement, there was no denominator that could be used to express the price of the Lehman Note (the Defaulted Reference Entity’s Reference Obligation in the present case) as a percentage.

[36] In support of their legal arguments on the issue of uncertainty, the Appellants referred to Foley v Classique Coaches, Limited [1934] 2 KB 1 and also May and Butcher, Limited v The King [1934] 2 KB 17. In the latter, the House of Lords held (per Viscount Dunedin at 21) in the context of a contract of sale:

[37] The Appellants’ counsel contended that the present case was one of conceptual uncertainty and not mere difficulty of application, relying on Megarry J’s analysis in Brown v Gould and Others [1972] 1 Ch 53 at 56–57 (citing the unreported case of In re Lloyd’s Trust Instruments (24 June 1970)) as follows:

[38] According to the Appellants, the present case was one of uncertainty of concept because:

(a) the third and fourth CERA Descriptions were inconsistent with each other;

(b) there was no way to determine conclusively which of these two CERA Descriptions was intended by the parties to be the operative CERA Description; and

(c) both of these CERA Descriptions were unworkable because they referred to the Final Price, which could not be determined as it had been expressed as a percentage without a denominator.

[39] The Respondent’s arguments on the Issues were likewise substantially a reiteration of their arguments in the court below. It was contended that the threshold to be satisfied before a court would declare a contract (particularly a fully-performed one such as the HN5 contract) void for uncertainty was very high and, as such, voiding a contract for uncertainty should be a measure of last resort. Reference was made to G Scammell and Nephew, Limited v H C and J G Ouston [1941] AC 251, where the House of Lords held that it was only where the words of a contract, “considered however broadly and untechnically and with due regard to all the just implications, fail[ed] to evince any definite meaning on which the court [could] safely act … that the court ha[d] no choice but to say that there [was] no contract” (at 268 per Lord Wright). The Respondent also referred to Sudbrook Trading Estate Ltd v Eggleton and Others [1983] 1 AC 444 in support of its contention that the threshold for voiding a contract was even higher in the case of executed (as opposed to executory) contracts. In that case, Lord Fraser of Tullybelton, quoting a passage from the English Court of Appeal’s judgment in the same case, said (at 484):

[40] On the basis of these authorities, the Respondent argued that the high threshold for voiding a contract for uncertainty had not been satisfied vis-à-vis the HN5 contract for the following reasons:

(a) The first and second CERA Descriptions were no more than general descriptions and were never intended to be comprehensive formulae.

(b) In any event, the first, second and third CERA Descriptions were accurate, consistent and reconcilable with each other.

(c) The Pricing Statement expressly designated the formula contained in the third CERA Description as the prevailing formula to resolve any uncertainty.

(d) There was ample basis for correcting the obvious clerical mistake in the fourth CERA Description through construction by removing “1 –” from the description such that it would become consistent with the first, second and third CERA Descriptions.

(e) There was no uncertainty about the fact that the CERA to be received by HN5 holders would be an amount equal to the CERA under the Reference Notes as calculated by the Respondent (as the Calculation Agent) and received by the Respondent (as the holder of the Reference Notes).

(f) In any event, the Pricing Statement stated that the calculation of the CERA by the Respondent was, in the absence of manifest error, binding on both the Respondent and HN5 holders.

[41] With respect to the First Issue, which stems from the alleged inconsistencies between and among the four CERA Descriptions in the Pricing Statement, we agree with the Judge that the first and second CERA Descriptions are verbal descriptions, but we do not agree that they are imprecise for that reason alone. We see no difficulty in expressing them in mathematical formulae as follows:

So expressed, it is clear that the first CERA Description is different from the second CERA Description, and both are different from the third CERA Description and also the fourth CERA Description.

[42] It should be noted that before both the Judge and this court, the Respondent did not expressly disagree that the first, second and third CERA Descriptions were expressed differently from one another. Rather, its argument, as summarised by the Judge in the HC Judgment at [41] and which it reiterated before us, was that these three CERA Descriptions clearly conveyed that the CERA was:

(a) directly related to the market value of the Lehman Note (the Defaulted Reference Entity’s Reference Obligation in this case); and

(b) inversely related to the shortfall in the market value of the CAAA and the HC.

On the other hand, the Respondent insisted that the fourth CERA Description was inconsistent with the third CERA Description as the former implied an inverse relationship between the CERA and the market value of the Lehman Note, which made the relationship absurd and, therefore, one that could not have been intended. Hence, the Respondent argued, the fourth CERA Description must contain a clerical mistake.

[43] In our view, the Respondent’s arguments as summarised in the preceding paragraph have not addressed the point that the first, second and third CERA Descriptions are formulated differently from one another, in that although they use (in substance) the same factors to calculate the CERA, they phrase these factors differently. Consequently, an issue does arise as to whether these three CERA Descriptions may produce different CERAs (or payouts) in the event of a Credit Event and, if so, which would be the operative CERA Description even if the fourth CERA Description were disregarded for this purpose. We will now analyse the four CERA Descriptions.

[44] In our view, given that the market value of the Lehman Note was only 0.005% of its nominal value, it is clear that the application of the first CERA Description would in effect result in a zero payout for HN5 holders as the computation of the CERA would yield a negative figure.

[45] In contrast, we are of the view that the literal application of the second CERA Description would not result in a zero payout for HN5 holders if Constellation had not invested the Reference Notes funds in the Lehman Note as, in that event, it would not have suffered any loss arising from Lehman’s bankruptcy. In that scenario, the CERA payout would be calculated as follows:

[46] On the above calculation, the second CERA Description would result in a CERA of S$1,392.52 for each HN5 under the SGD Tranche and a CERA of US$1,407.50 for each HN5 under the USD Tranche. Coincidentally, if Constellation had not actually invested any of the Reference Notes funds in the Lehman Note, the CERA under the second CERA Description would be the same as the CERA under the fourth CERA Description. It follows that the second CERA Description would contain a mistake in the same way that the fourth CERA Description is alleged by the Respondent to contain a clerical mistake.

[47] The difficulty with the literal application of the second CERA Description is that it assumes that the Pricing Statement required Constellation to invest all or substantially all of the Reference Notes funds in one or more of the Reference Obligations as it is only in this scenario that the CERA under the Reference Notes (and, in turn, the CERA under the HN5) would have a direct one-to-one correlation with the value of the Reference Obligations. This is, however, not the case as nowhere in the Pricing Statement was it stated that Constellation was required to invest the Reference Notes funds in any of the Reference Obligations. The Reference Obligations were used only for the purpose of triggering a Credit Event. In other words, the Reference Obligations were merely notional obligations and not real obligations. But, as mentioned at [18] above, this is clear only upon a close analysis of the terms of the Pricing Statement.

[48] The third CERA Description provides a precise mathematical formula to work out the CERA payable to HN5 holders upon the termination of the HN5 resulting from a Credit Event. It is not disputed that the formula set out in the third CERA Description will result in HN5 holders receiving a zero payout in the present case as the calculation, which both parties accept, is as follows:

[49] The application of the third CERA Description actually gives a negative result for the CERA, which, in practical terms, means a zero payout for HN5 holders. The absurdity of this theoretical result can only be explained on the basis that the third CERA Description is so constructed as to virtually guarantee that on the occurrence of a Credit Event, HN5 holders will get nothing back on their capital “investment”.

[50] It is the Respondent’s case that the fourth CERA Description contains a clerical mistake because the application of the formula set out therein would lead to an absurd result, in that the lower the value of the Defaulted Reference Entity’s Reference Obligation (as assessed by prospective investors in the relevant security), the higher the CERA to be paid to HN5 holders. Before both the Judge and this court, the Respondent did not explain how the mistake came about – an omission which we find surprising, given that the Respondent’s army of advisers would presumably have examined the Pricing Statement, especially the CERA Descriptions therein, with meticulous care. We earlier examined the second CERA Description, and found that it too could be said to contain a mistake as it would, if applied literally, give the same result as the fourth CERA Description. It seems incredible that in the context of such a humongous structured note programme as the HN5, the Respondent could have failed to detect two serious errors in the description of the CERA, which may be said to lie at the heart of HN5 holders’ “investment” in the event of a Credit Event occurring and which formed a key part of the basis on which the Respondent invited applications for the HN5.

[51] The Appellants’ case is based entirely on principles of contract law and not on facts. It rests, in essence, on two arguments, namely: (a) the four CERA Descriptions are inconsistent with one another, thereby rendering the CERA uncertain; and (b) in any event, regardless of whether it is the third or the fourth CERA Description which is the operative CERA Description, the CERA is unascertainable because the formulae in both the third and fourth CERA Descriptions are unworkable. Consequently, it is contended, the HN5 contract is void ab initio for uncertainty. We are unable to accept both arguments for the reasons given below.

[52] We have established that:

(a) the first CERA Description is in substance the same as the third CERA Description in so far as both will give, for practical purposes, a zero payout;

(b) the second CERA Description, if applied literally, will give the same payout as the fourth CERA Description in the circumstances of this case; and

(c) the first and third CERA Descriptions are therefore inconsistent with the second and fourth CERA Descriptions.

[53] In our view, the Pricing Statement can only stipulate one operative CERA Description for the purposes of the HN5 contract, and the question of which of the four (or two sets of) CERA Descriptions is the operative one is a question of fact that can and should be ascertained from the evidence which has been or can be made available to the court. In our view, it would not be wise or correct for us to decide the First Issue on the basis of preferring one set of legal arguments to another when the real dispute between the parties is really a question of fact as to what the operative CERA Description is. We know that the Pricing Statement states clearly that the CERA which HN5 holders will get upon the occurrence of a Credit Event is the CERA which the Respondent will get under the Reference Notes (see [8] above). We know that the Reference Notes exist, although their full terms have not been disclosed to the court. We know that the Respondent has alleged that the fourth CERA Description contains a clerical mistake. Hence, to ascertain whether the Judge was right in ruling that the third CERA Description was the operative CERA Description, all that we need to do is to look at the terms and conditions of the contract applicable to the Reference Notes to see whether the formula for calculating the CERA under the Reference Notes (which, as we have just reiterated, is also the CERA to be paid to HN5 holders) is the same as the formula set out in the third CERA Description. In short, the operative CERA Description can be conclusively established by examining the material facts.

[54] In this connection, neither the Appellants nor the Respondent appeared to be keen to have the Reference Notes and the relevant accompanying documents (collectively, “the Reference Notes documents”) disclosed to this court, even though we suggested, in the course of hearing this appeal, that such disclosure would be helpful to the court. We made this suggestion as we were effectively being asked to decide an issue of fact (albeit presented by the parties as an issue of law) without having before us all the material evidence pertaining to that issue of fact. There were other factors which prompted us to make this suggestion. Firstly, the HN5 are a novel and complex financial product. As such, the court should be slow to decide what the operative CERA Description is based purely on legal arguments. Secondly, the Pricing Statement (consisting of 93 pages of explanatory statements), which forms a key plank of the HN5 contract, would presumably have been carefully vetted by teams of lawyers and bankers. As such, we were not prepared to accept, without further scrutiny, the Respondent’s argument that the fourth CERA Description contained a clerical error. Thirdly, it seemed to us that if the fourth CERA Description did indeed contain a clerical mistake, then the second CERA Description would also contain a mistake as the second CERA Description, applied literally, would give the same result as the fourth CERA Description if Constellation had not actually invested the Reference Notes funds in the Lehman Note (see [46] above). That two mistakes could have been made in the Pricing Statement seemed to us (as mentioned at [50] above) incredible and called for further inquiry. Fourthly, if the formula for calculating the CERA as set out in the Reference Notes documents is the same as the fourth CERA Description, then the fourth CERA Description, and not the third CERA Description, would be the operative CERA Description. This is because the Pricing Statement states very clearly that the CERA to be paid to HN5 holders in the event of a Credit Event would be the same as the CERA received by the Respondent under the Reference Notes. In that scenario (ie, if the fourth CERA Description is the operative CERA Description), given the undeniable and irreconcilable inconsistency between the third and fourth CERA Descriptions, it would be the third CERA Description, and not the fourth, which contains a clerical error.

[55] In view of the above considerations, we directed the Respondent (notwithstanding both parties’ apparent reluctance) to produce for our consideration, inter alia, the Reference Notes documents. These documents, which were furnished on 11 July 2011, included:

(a) the pricing supplement dated 16 May 2007 for SRN Series 75;

(b) the pricing supplement dated 16 May 2007 for SRN Series 76;

(c) the swap confirmation dated 16 May 2007 for SRN Series 75; and

(d) the swap confirmation dated 16 May 2007 for SRN Series 76.

It should be noted that when producing the documents ordered by this court, the Respondent drew the attention of the Appellants’ solicitors to the application of the principle laid down in Riddick v Thames Board Mills Ltd [1977] QB 881 as well as the implied undertaking: (a) not to disclose the Reference Notes documents to any third party, and (b) not to use them, without the consent of this court, for any purpose other than in the present proceedings. In this regard, we confirm that the implied undertaking subsists and is not to be breached in any way.

[56] Having perused the Reference Notes documents, we observe that the formula for calculating the CERA under the Reference Notes (and, thus, the formula for calculating the CERA under the HN5) is consistently stated across several of the documents to be the formula under the third CERA Description – ie, the CERA Description with respect to the Reference Notes is the same as the third CERA Description. Accordingly, we conclude that the operative CERA Description for the purposes of the HN5 is indeed, as the Judge held, the third CERA Description (which description, in our view, the first CERA Description was intended to express, albeit in verbal and less precise terms). We also accept the Respondent’s argument that the fourth CERA Description contains a clerical mistake, although how this mistake (and also the mistake in the second CERA Description) came about remains a mystery.

[57] In the circumstances, the legal arguments advanced by the Appellants’ counsel on the issue of conceptual uncertainty in the terms of the HN5 contract are not applicable. The Appellants’ allegation of uncertainty in a critical term of the HN5 contract (viz, the term relating to the CERA) must, therefore, fail. The result is that owing to Lehman’s bankruptcy, the CERA payable to HN5 holders on their investment is indeed zero.

[58] We will now consider the Second Issue, which concerns the Appellants’ argument that both the third and fourth CERA Descriptions are not workable as the words “expressed as a percentage” in the definition of “Final Price” for both of these CERA Descriptions do not specify a denominator to give the said percentage (“the FP percentage”) a value.

[59] The Appellants’ submissions on this issue were as follows. First, there was no evidence that the parties intended, knew or understood that the denominator for the FP percentage would be the nominal value of the Lehman Note. Second, using the nominal value of the Lehman Note as the denominator for the FP percentage would only make sense if it were assumed that all of the Reference Notes funds were allocated to the Lehman Note; otherwise, there was no logical reason to assume that the loss suffered on those funds would be proportionate to the fall in value of the Lehman Note.

[60] In rebuttal, the Respondent contended that the denominator for the FP percentage was the outstanding principal balance of the Defaulted Reference Entity’s Reference Obligation (ie, the Lehman Note). This interpretation, the Respondent submitted, was borne out by the definitions of “Final Price” and other related terms in Appendix D of the Pricing Statement, which definitions the third CERA Description expressly referred the reader to (see the quotation at [25] above), and was also based squarely on the terms and definitions set out in the Pricing Statement. The material terms and definitions in this regard are as follows:

(a) “Final Price”, which is defined as “the price of the Reference Obligation of the Defaulted Reference Entity expressed as a percentage, determined in accordance with the Valuation Method” [emphasis added];

(b) “Valuation Method”, which is defined as “the Market Value determined by the Calculation Agent [ie, the Respondent] with respect to the Valuation Date [viz, a date falling within 80 business days from the “Credit Event Determination Date”, which is the date on which a Credit Event notice, delivered in accordance with the terms of the HN5 contract, is effective]” [emphasis added];

(c) “Market Value”, which is defined as, essentially, the arithmetic mean of “Full Quotation[s]” [emphasis added] obtained by the Respondent as the Calculation Agent;

(d) “Full Quotation”, which is defined as “in accordance with the Quotation Method [ie, the bidding method], each firm quotation obtained from a Dealer at the Valuation Time [viz, approximately 11.00am Singapore time], to the extent reasonably practicable, for an amount of the Reference Obligation with an Outstanding Principal Balance equal to the Quotation Amount” [emphasis added]; and

(e) “Quotation Amount”, which is defined as “the [APA] of the Notes [ie, the Reference Notes] outstanding”.

[61] In our view, in the context of the formulae under the third and fourth CERA Descriptions, the FP (Final Price) is the market value of the Lehman Note expressed as a percentage of the nominal value of the Lehman Note. Since the Lehman Note lost 99.995% of its nominal value, the FP percentage would be 0.005% (of the nominal value of the Lehman Note). The words “expressed as a percentage” in the definition of “Final Price” mean expressed as a percentage of the Aggregate Principal Amount (APA) of the Reference Notes, which the nominal value of the Defaulted Reference Entity’s Reference Obligation is equated with. This interpretation is consistent with the artificial and complex structure of the HN5, whose terms provide that the entire principal amount of the Reference Notes (ie, all of the Reference Notes funds) is notionally allocated to the Defaulted Reference Entity. In our view, the words “expressed as a percentage” cannot mean anything else. This is not a case of words having no meaning at all. It is, rather, a case of construing the relevant contractual documents to determine the intended denominator to be used for calculating the CERA under the HN5. Accordingly, we are of the view that contrary to the Appellants’ submission, there is no conceptual uncertainty in the words “expressed as a percentage” in the definition of “Final Price”. We would also add that considering the enormous number of words already used in the Pricing Statement to set out the terms of the HN5, the Respondent’s advisers could and should simply have added a few more words to give the phrase “expressed as a percentage” greater clarity. This would have made the definition of “Final Price” more intelligible, and would in turn have averted the obfuscation which gave rise to the Second Issue.

[63] In view of our decision in this appeal, we think it apposite and timely to remind the general public that under the law of contract, a person who signs a contract which is set out in a language he is not familiar with or whose terms he may not understand is nonetheless bound by the terms of that contract. Illiteracy, whether linguistic, financial or general, does not enable a contracting party to avoid a contract whose terms he has expressly agreed to be bound by. The principle of caveat emptor applies equally to literates and illiterates in such circumstances.