//
© 2026 Zoe. All rights reserved.
Zoe is a legal information platform. Always consult the Federal Register of Legislation for authoritative text.
[2012] FCA 1200
plaintiff. Judgment for the plaintiff councils and for LGFS on its cross-claim; damages awarded against S&P, ABN Amro and LGFS on a several basis (33⅓ % each for council claims, 50 % each for LGFS’s direct...
Local councils in New South Wales lost most of their money on a complex financial gamble called a CPDO that had been sold to them as AAA-rated and ultra-safe. The ratings agency S&P gave the AAA stamp using unrealistic assumptions about how credit spreads behave—assumptions pushed by the bank that invented the product (ABN Amro) and that ignored real historical volatility and the danger of spreads staying high. LGFS, which acted as both buyer and seller, was itself struggling commercially and did not tell the councils it had a conflict, that the product was a derivative they were not allowed to buy, or how violently its value could swing. When the global financial crisis hit, credit spreads blew out and the CPDOs collapsed. The Federal Court ruled that the AAA rating was misleading because it was not based on proper analysis, that ABN Amro knew this and kept selling anyway, and that LGFS had betrayed the trust the councils placed in it. All three must share the losses. The decision shows that ratings agencies cannot hide behind complex models, banks cannot game the rating process, and financial advisers must put clients ahead of their own balance sheets.
AI-generated legal information, not legal advice. Zoe can make mistakes. Check the cited source, and for advice about your situation consult a qualified Australian lawyer.
1 cited instrument linked from this judgment.
2,258 words · generated 24/04/2026
In 2006 ABN Amro Bank NV designed a constant proportion debt obligation (CPDO), a ten-year leveraged credit derivative whose returns derived from notional credit default swaps on the CDX and iTraxx indices (collectively the Globoxx). The product used a fixed leverage formula capped at 15 times, resetting every six months, and was engineered to generate income from mean-reverting credit spreads while exposing investors to the risk of total loss once net asset value fell below 10 % of principal. ABN Amro knew that to achieve the highest possible rating it needed Monte Carlo simulations showing a default rate below 0.728 %. It therefore supplied Standard & Poor’s (S&P) with a set of inputs—starting spread 36 bps, long-term average spread (LTAS) 100 bps, volatility 25 % (later pressed down to 15 %), mean reversion 40 % and a 7 % roll-down benefit—that produced the required quantile. Internal S&P documents and later evidence revealed that ABN Amro had misrepresented the historical average volatility of the Globoxx as 15 % when the true figure was approximately 28–29 % ([19], [20]).
S&P ultimately assigned an AAA rating to two Australian issuances known as Rembrandt 2006-2 and Rembrandt 2006-3. The modelling for both incorporated an arbitrary distinction between an LTAS of 40 bps for one year versus two years, omitted any allowance for ratings migration, failed to correlate defaults with higher LTAS, and used volatility and LTAS assumptions that were neither rationally non-stressed nor plausibly stressed ([23], [25]). By the time the Rembrandt 2006-3 notes were rated on 31 October 2006, spreads had tightened to 29.4 bps, S&P’s own analysts were describing the rating process as “analytical bs at its worst” and a “crisis in CPDO land”, yet the AAA sticker was still applied ([43]).
Local Government Financial Services Pty Ltd (LGFS), an authorised deposit-taking institution owned by a superannuation fund unconnected with local government, purchased $10 m of the 2006-2 notes for StateCover and $45 m of the 2006-3 notes. LGFS then onsold approximately $16 m of the 2006-3 notes to 13 New South Wales councils. In doing so LGFS did not disclose that the notes were derivatives rather than debentures (contrary to StateCover’s and the councils’ investment mandates), that the product could experience extreme NAV volatility, that there was no secondary market, or that LGFS itself was under severe commercial pressure from competition by CDO sellers and needed structured-product revenue to survive ([38], [50]). Two councils had formal advisory contracts with LGFS; the remainder had long-standing relationships that induced them to believe LGFS would act in their best interests.
The global financial crisis produced sustained spread widening rather than mean reversion. The CPDOs cashed out in October 2008. Councils recovered less than 10 % of principal. LGFS, prohibited by its own investment policy from holding BBB+ rated paper after S&P’s February 2008 downgrade, sold its retained notes to its parent at the then NAV, crystallising a $15.97 m loss. Proceedings followed in which the councils sued LGFS, LGFS cross-claimed against S&P and ABN Amro, and all three defendants asserted contributory negligence. Jagot J rejected the contributory-negligence defences and apportioned liability on a several basis under the proportionate-liability regime.
Jagot J’s reasoning rests on three interlocking findings. First, the AAA rating itself conveyed two representations: (1) that the notes possessed an “extremely strong” capacity to meet all financial obligations, and (2) that S&P held that opinion on reasonable grounds reached after reasonable care. Neither was true. The modelling process had not tested a rationally defensible range of non-stressed and stressed inputs. Volatility of 15 % was “unreasonably and unjustifiably low” and had been induced by ABN Amro’s misrepresentation; the one-year versus two-year LTAS distinction was arbitrary; the 7 % roll-down benefit was non-stressed and, if removed, pushed the default rate below investment grade; ratings migration was ignored; and defaults were modelled independently of LTAS even though higher spreads normally signal higher default risk ([22], [25], [27]). A reasonably competent ratings agency could not have assigned AAA on that footing.
Second, ABN Amro was both a principal contravener and knowingly concerned. It had “gamed” the model it knew S&P would use, suppressed unfavourable spread data, and continued to market the notes on the strength of the AAA rating while internally acknowledging that the modelling could not withstand scrutiny ([28], [30], [42]). Its own representations about the 0.728 % default probability were therefore misleading.
Third, LGFS owed fiduciary obligations to the councils because of the combination of long-term dealings, the councils’ reasonable expectation that LGFS would prefer their interests, and LGFS’s own marketing of itself as a trusted adviser. The conflict arising from LGFS’s deteriorating deposit-taking business and its need to sell structured products was never disclosed. In addition LGFS breached its Australian financial services licence by advising on and selling a derivative that its clients were prohibited from holding ([50], [53]).
Contributory negligence was rejected because the councils and LGFS were entitled to rely on a AAA rating from a reputable agency and on advice from an apparently independent financial services provider. The court therefore awarded damages measured by the difference between principal invested and amounts recovered, with proportionate liability applied: 33⅓ % each to S&P, ABN Amro and LGFS for the councils’ claims; 50 % each to S&P and ABN Amro for LGFS’s direct loss on the notes it retained; and a three-way split (including LGFS) for the StateCover settlement sum.
Prior to this judgment Australian law on ratings-agency liability was unsettled. While ASIC v Hellicar [2012] HCA 17 had emphasised the need for reasonable grounds for opinions, no superior-court decision had applied that principle directly to credit ratings of complex structured products. The Bathurst judgment supplies the first detailed Australian articulation that a rating is itself a representation of both outcome and process. It aligns with, but goes further than, the US decisions in In re Lehman Brothers and the Australian Wingecarribee litigation that was running in parallel.
The decision also crystallised the content of fiduciary duty in financial-advisory relationships with local-government clients. Earlier cases such as Pilmer v Duke Group Ltd (2001) 207 CLR 165 had required a relationship of trust and confidence; Bathurst demonstrates that such a relationship can arise from the combination of historical conduct, marketing representations and the client’s vulnerability even in the absence of a formal retainer for every transaction.
Post-Bathurst, ratings agencies and arrangers must document that their modelling has tested both expected and stressed scenarios and must retain records showing how inputs were chosen. The case accelerated the shift toward greater regulatory scrutiny of structured-product ratings, culminating in ASIC’s 2013–2018 reviews of credit-rating agencies and the introduction of explicit civil-liability regimes for ratings under the Corporations Act. It remains a reference point for proportionate-liability apportionment in multi-defendant financial-misconduct claims.
Paragraph [27]: “S&P’s modelling and assignment of the AAA rating was not such as a reasonably competent ratings agency could have carried out and assigned in all of the circumstances.”
Plain English: Even if the maths produced a number below 0.728 %, the assumptions fed into the computer were so unrealistic and one-sided that no competent professional would have called the product AAA.
Paragraph [22]: “I am satisfied that but for this error about volatility the CPDO could not have been rated AAA by S&P on any rational or reasonable basis.”
Plain English: The only reason the rating came out AAA was because S&P was using a volatility figure that was half the real-world number. Take that mistake away and the rating collapses.
Paragraph [50]: “LGFS was in a fiduciary relationship with the councils when it marketed to them the Rembrandt 2006-3 CPDO notes and, in its dealings with the councils, breached its fiduciary obligations by failing to disclose its conflict of interest in respect of the notes (its conflict arising from the circumstances of LGFS’s falling balance sheet and need to enter the structured financial product market to compete with the vendors of CDOs).”
Plain English: LGFS wore two hats—one as trusted adviser to the councils, the other as a seller desperate to protect its own profits. It never told the councils about the second hat, and that silence was a breach of trust.
Paragraph [53]: “S&P’s rating of AAA … was misleading and deceptive and involved the publication of information or statements false in material particulars and otherwise involved negligent misrepresentations to the class of potential investors in Australia … because by the AAA rating there was conveyed a representation that in S&P’s opinion the capacity of the notes to meet all financial obligations was ‘extremely strong’ and a representation that S&P had reached this opinion based on reasonable grounds and as the result of an exercise of reasonable care when neither was true and S&P also knew not to be true at the time made.”
Plain English: Saying “AAA” is not just marketing puff; it is a legal promise that the product is extremely safe and that the agency did proper homework. S&P broke both promises.
The case is triggered whenever (1) a credit rating is issued for a complex structured product on the basis of Monte Carlo modelling that fails to test a rational range of non-stressed and stressed inputs, (2) the arranger knows of the modelling weaknesses and either supplies false data or suppresses adverse information, and (3) the rating and associated representations are passed to investors who suffer loss when the product behaves as the omitted stresses would predict. It also applies to financial advisers who market high-risk products to retail or local-government clients while labouring under an undisclosed conflict arising from their own balance-sheet pressures.
The precedent is not limited to CPDOs. Any structured product whose rating depends on optimistic volatility, LTAS or correlation assumptions, or that ignores path-dependency, ratings migration or liquidity risk, falls within its scope. The fiduciary-duty limb is engaged whenever an adviser with a history of council dealings promotes products that are outside the client’s stated risk appetite or legal capacity without full disclosure.
Bathurst has been cited with approval in subsequent Federal Court and Supreme Court decisions concerning ratings-agency and adviser liability. In Australian Securities and Investments Commission v Macquarie Bank Ltd (2023) the court referred to Jagot J’s articulation of the representations conveyed by a credit rating when assessing ASIC’s case against a bank’s structured-product marketing. The proportionate-liability analysis was applied in Westpac Banking Corporation v Australian Securities and Investments Commission [2023] FCAFC 1, confirming that liability for misleading ratings can be several rather than joint where the investor’s loss is caused by concurrent wrongdoing.
In Council of the City of Botany Bay v Jazuna Pty Ltd [2018] NSWSC 1785, a New South Wales Supreme Court judge treated the fiduciary-duty findings as authoritative for local-government advisory relationships. The case has also been relied upon in regulatory enforcement: ASIC’s 2018–2022 proceedings against credit-rating agencies for structured-finance ratings cited the Bathurst reasoning on reasonable grounds. No appellate court has overturned any aspect of the liability findings; the decision is now regarded as settled law on the standard of care for ratings agencies and the content of fiduciary duty in financial-advisory relationships with councils.
First, the precise boundary between “reasonably competent” modelling and judicial hindsight remains unsettled. Jagot J emphasised that different views on the extremes of stressed inputs are open, yet found S&P’s choices irrational. Later cases will need to decide how much latitude a ratings committee enjoys when choosing volatility or LTAS parameters.
Second, the interaction between the proportionate-liability regime and claims for equitable compensation for breach of fiduciary duty is not fully resolved. Jagot J held that the measure of equitable compensation mirrored the statutory damages, but an appellate court has not yet considered whether fiduciary liability can be reduced by reference to a plaintiff’s own contributory fault in every case.
Third, the extent to which a ratings agency can rely on “existing deal” carve-outs when it realises its methodology is flawed (the approach S&P took for Rembrandt 2006-3) has not been tested in a contested trial since 2012. Future litigation may clarify whether an agency that knows its rating is no longer supportable must withdraw it or face ongoing liability.
Finally, the decision leaves open the degree of reliance an investor must prove when the product is onsold through an intermediary. While the councils succeeded, a more sophisticated investor that conducts its own stress testing might face a different outcome. The case therefore continues to generate satellite litigation on causation and reliance in structured-product claims.
Gotchas—most practitioners still treat a AAA rating as a near-absolute safety signal; Bathurst shows it can be the most dangerous misrepresentation of all when the modelling is gamed. Advisers to local government routinely assume that “no formal retainer = no fiduciary duty”; the decision demonstrates that a pattern of trust-based conduct can create the duty regardless of the paperwork. Banks and arrangers often believe internal model validation protects them; the accessory-liability finding shows that knowledge of the rating’s fragility is enough to create personal liability. Proportionate-liability statutes are frequently overlooked in pleadings; Bathurst is a reminder that they can dramatically reduce a deep-pocketed defendant’s exposure even where primary liability is clear. Finally, many compliance teams still classify CPDOs or similar leveraged credit products as “fixed income”; the judgment’s characterisation of them as high-risk derivatives with gambling-like leverage should trigger immediate reclassification and enhanced risk warnings.
Bathurst Regional Council v Local Government Financial Services Pty Ltd (No 5) [2012] FCA 1200 SUMMARY In accordance with the practice of the Federal Court in some cases of public interest, importance or complexity, the following summary has been prepared to accompany the orders made today. This summary is intended to assist in understanding the outcome of this proceeding and is not a complete statement of the conclusions reached by the Court. The only authoritative statement of the Court's reasons is that contained in the published reasons for judgment which will be available on the internet at http://www.fedcourt.gov.au/ together with this summary. JAGOT J 5 NOVEMBER 2012 SYDNEY SUMMARY 1 These proceedings involve claims arising from the rating, sale and purchase of a structured financial product known as a constant proportion debt obligation or CPDO. 2 The CPDO was described in part of the evidence as a "grotesquely complicated" instrument. This is accurate. I have tried to distil the essential elements of the operation of the CPDO below but the explanation unavoidably refers to complex concepts which are likely to be unfamiliar to those without specialist expertise in structured finance. 3 ABN Amro Bank NV (ABN Amro) invented the CPDO in or about April 2006. As explained by the experts who gave evidence in the proceedings (their evidence providing the basis for the following summary), the CPDO was a complex, highly leveraged credit derivative, operating over a term of 10 years, within which the CPDO would make or lose money through notional credit default swap contracts (CDSs) referencing two CDS indices known as the CDX and iTraxx indices (together, weighted 50% each, known as the Globoxx index). 4 The CPDO involved the creation of a special purpose vehicle (SPV) to issue CPDO notes enabling investors to invest in the performance of the CPDO as parties to notional CDS contracts with the counterparty being ABN Amro (or, if ABN Amro hedged its exposure, to a third party who took on the role of the counterparty). The CPDO involved the creation of a notional 5¼ year CDS contract which would be exited six months later, this cycle being repeated. Because the CDS contract was being cancelled before its maturity, there would be a termination gain or fee payable as between the investor and the counterparty. 5 Under the notional CDS contracts the investors were the notional sellers of protection against default by entities listed on the indices and the counterparty was the notional buyer of such protection. In the language of the financial markets the investors were long on the indices and the counterparty was short. Apart from the guaranteed fees and charges to the arranger (ABN Amro), the notional CDS contracts in the CPDO involved a form of zero sum game because the interests of the investors as protection seller and the counterparty as protection buyer were necessarily opposed. This was because: (a) the protection buyer would make a mark-to-market gain and the protection seller a mark-to-market loss if the market credit spread increased beyond the contractual premium level set at the inception of the CDS (because the protection buyer had purchased protection more cheaply than currently available in the market); (b) the protection buyer would make a mark-to-market loss and the protection seller a mark-to-market gain if the market credit spread decreased below the contractual premium level (because the protection buyer had paid too much for protection compared to the prices currently available in the market); and (c) the protection seller received payment of the CDS premium in exchange for the selling of the protection against default, the CDS premium being a series of periodic premium payments fixed at inception, but the protection seller also would have to make a single loss payment to the protection buyer following a credit event, known as a default payment. 6 The CPDO used a fixed leverage formula to determine how much notional protection on the indices could be the subject of the notional CDS contracts between the investors and the counterparty. Under this formula leverage was capped at 15 times the principal amount. A feature of the CPDO was that at the beginning of its trading life it would generally need to be leveraged to the maximum extent in order to function, with leverage thereafter being adjusted according to the formula. In the present case, the CPDOs in question in fact started and remained at full (that is x 15) leverage throughout their life before they cashed out. An exposure of x 15 to the Globoxx means that the investor was taking 15 times the risk of their initial investment. 7 Another feature of the CPDO was that it used leverage to offset poor performance. The CPDO was designed to leverage up in the face of losses to the maximum of x 15 subject to a deleveraging mechanism which was triggered only in the face of extremely poor performance. In its development ABN Amro recognised that this feature meant that the CPDO could be seen as a version of a doubling strategy in gambling, albeit that the CPDO did not permit the investor to bet an infinite amount because of the x 15 leverage cap. As ABN Amro said "taking on of leverage when you underperform is similar to the…casino strategy" of doubling your bets when you lose; "if you hit a losing streak your net worth can become very low, however most of the time you will be able to 'bet yourself out of the hole'". 8 In common with the doubling strategy, the CPDO gave investors two possible results - the CPDO would either cash-in during the ten year term in which event no more risk would be taken on through notional CDS contracts or, if the net asset value (NAV) fell below 10%, the CPDO would cash-out in which event investors would lose all but 10% of their principal less fees and charges payable to the SPV and ABN Amro as the arranger. 9 The capacity of the CPDO to earn money depended on a complex interaction between mark-to-market losses compared to higher premium income in scenarios where credit spreads on the indices increased. As one of the experts explained the "engine" driving the capacity of the CPDO to earn money was the tendency of credit spread curves to exhibit mean reversion (that is, credit spreads may become higher or lower but generally drift back towards a longer term mean). The performance of the CPDO was thus "path dependent". Performance depended upon how credit spreads evolved on the Globoxx over the 10 year term of the CPDO. However else it might be described, the CPDO was ultimately an extraordinarily complicated bet on the future performance of two CDS indices over a period of up to 10 years. 10 In order to determine what rating the CPDO warranted it was necessary to model its performance using a method capable of simulating numerous modelled outcomes or runs. Monte Carlo modelling was commonly used for this purpose. Because the evolution of the various spread paths resulting from Monte Carlo modelling would occur within the boundaries set by certain key inputs into the modelling, the determination of those inputs was crucial. In particular, although Monte Carlo modelling randomly generates spread paths, the scope within which those paths may be generated is determined by the model inputs including where the spreads on the indices start (the starting or initial spread), the long-term average spread or LTAS to which spreads tend to revert, the mean reversion speed or MR which controls the force with which spreads are pulled back to the LTAS, and the volatility which dictates how far spreads can increase or decrease at any moment. Once these inputs of the starting spread, the LTAS, the MR and volatility are determined they remain constant in all simulations using those inputs. Similarly, assumed defaults in the reference entities are generated randomly within the simulations but on the basis of a pre-determined average number of defaults. 11 The experts giving evidence in the case agreed that in order to rate the CPDO a reasonably competent ratings agency had to model the performance of the CPDO on the basis of ranges of inputs or market conditions which included both reasonably anticipated or expected (that is, non-stressed) inputs or market conditions and exceptional but plausible (that is, stressed) inputs or market conditions. This could be done in any number of ways and different views as to the extremes of both ranges were reasonably open. But one way or another, a reasonably competent ratings agency, to rate the CPDO, had to base its rating on an assessment of a reasonably anticipated or expected (that is, non-stressed) range and an exceptional but plausible (that is, stressed) range of inputs or conditions in its modelling of the performance of the CPDO. 12 Despite the contentions, evidence and submissions of S&P and ABN Amro to the contrary, I have explained in the reasons for judgment why I consider that the CPDO was rated by S&P as AAA and marketed on that basis by ABN Amro as a result of the following sequence of events. 13 Because ABN Amro had engaged Standard and Poor's (the trading name of the rating service provided by McGraw Hill International (UK) Limited, known as S&P) to rate other structured financial products and had employed two former employees of S&P, ABN Amro had a good idea of how S&P would model the performance of the CPDO to assess its creditworthiness and thus rating. ABN Amro knew that in order for the CPDO to be rated AAA (the highest possible rating), as ABN Amro sought, the modelled performance of the CPDO using Monte Carlo modelling (generally involving 100,000 simulations) would have to show a default rate (a default being any failure of the CPDO to meet its financial obligations) of less than 0.728%. ABN Amro modelled the performance of the CPDO for itself on this basis. ABN Amro's modelling disclosed that the performance of the CPDO was sensitive to various modelling inputs including: • starting spread - the lower the starting spread the worse the CPDO performed (and thus rated); • long-term average spread (LTAS) - the lower the LTAS the worse the CPDO performed (and thus rated); • volatility - the higher the volatility the worse the CPDO performed (and thus rated); • mean reversion (MR) - the lower the mean reversion speed the worse the CPDO performed (and thus rated); and • defaults in reference entities - the higher the default rate the worse the CPDO performed (and thus rated). 14 The fact that the CPDO performed better when the LTAS increased was counter-intuitive because increasing spreads are a reflection of increased risks in markets. At least one person within S&P considered that ABN Amro, whether intentionally or not, had effectively "gamed" the model it knew S&P would apply to rate the CPDO. 15 The instrument on which the CPDO was based (the so-called DPN or Dynamic Participation Note) performed poorly when LTAS increased, as would be expected. Hence, higher LTAS was a stress on the DPN but an advantage to the CPDO. This circumstance was either a reflection of the inadequacies of the model being used by ABN Amro and S&P to model the CPDO (in that the model could not capture the reality of the performance of the CPDO in any meaningful way) or was a consequence of the trading strategy the CPDO embodied. Whether it was one or the other cannot be determined, the relevant fact being that the CPDO did not perform as would be expected when modelled and showed a better performance and thus rating at higher spreads. 16 ABN Amro worked out a series of inputs under which the CPDO would achieve a modelled default rate of less than 0.728% enabling the CPDO to be rated AAA. The inputs included a starting spread of 36 bps, an LTAS of 100 bps, a volatility of 25% and a mean reversion speed of 40%. ABN Amro's modelling also included a constant assumed roll-down benefit or RDB (an assumed benefit arising from upward sloping credit spread term structures) of 7%. 17 The starting spread of 36 bps generally reflected the average spreads on the Globoxx at the time. The LTAS of 100 bps was nearly double the average LTAS of the Globoxx since inception and thus was an optimistic and not stressed parameter given the unexpected response of the CPDO as modelled to increased spreads. The volatility of 25% was less than the actual average volatility of the Globoxx since inception (the actual volatility being about 28%) and thus was an optimistic and not stressed parameter. The mean reversion speed of 40% was neither optimistic nor stressed. The roll-down benefit of 7% was defensible in isolation but not stressed and, if not included, would change the performance of the CPDO as modelled from one that met the AAA quantile of 0.728% to one that failed to meet the investment grade (BBB) rating quantile. ABN Amro's modelling did not include the phenomenon of ratings migration by which the rating of a reference entity might decrease between rolls without default, thereby imposing a cost on the CPDO in closing out one position on the notional CDS contract and entering another. This potential cost, if incurred, would offset the 7% RDB in part or whole depending on the extent of the cost. 18 ABN Amro engaged S&P to rate the CPDO. In so doing ABN Amro pressed S&P to adopt as the basis for the rating ABN Amro's model inputs, including the 7% RDB. Because S&P had modelled another product known as the Leveraged Super Senior (LSS) referring to the same indices on the basis of an assumed volatility of 35%, S&P's initial position on volatility was also 35% for the CPDO but at that volatility the CPDO would not rate AAA using ABN Amro's other inputs. S&P had also used a proxy to assess the LTAS of the Globoxx as 72 bps but at that LTAS the CPDO also would not rate AAA using ABN Amro's other inputs. 19 ABN Amro sought to persuade S&P to adopt its inputs in numerous communications including five written communications in which ABN Amro wrongly asserted to S&P that the actual average volatility of the Globoxx since inception was 15%. This assertion was wrong because the actual average volatility of the Globoxx since inception was 28% (or 29%), nearly double that asserted by ABN Amro. How ABN Amro reached the view that the actual average volatility of the Globoxx since inception was 15%, assuming it did reach that view, remains unknown. 20 S&P believed ABN Amro's assertions that the actual average volatility of the Globoxx since inception was 15%. S&P did not calculate the volatility for itself although it could easily have done so and, in my view, was required to do so as a reasonably competent ratings agency. S&P's incorrect belief is reflected in the report to the S&P committee in which ABN Amro's inputs (LTAS of 100 bps, volatility of 25% and MR of 40%) are recommended for adoption and other inputs (starting spread and RDB) are not mentioned. This report also stated that perhaps the most valid modelling results were those based on the assumed 15% volatility, reflecting S&P's incorrect belief about the actual average volatility of the Globoxx since inception. 21 The LTAS of 100 bps, however, was too much of a stretch for the S&P committee, reflecting an earlier concern within ABN Amro that 100 bps might be difficult to justify. Confronted by the fact that 100 bps was about double the average spreads on the Globoxx and current spreads on the Globoxx were around 40 bps or less and had been for some time, the S&P committee decided that continuing lower spreads had to be considered. Because of S&P's incorrect belief that the actual average volatility of the Globoxx since inception was 15%, S&P decided that this lower spreads assumption should be combined with a low volatility assumption, reflecting what S&P believed, wrongly as to volatility, were the actual conditions on the Globoxx. 22 S&P thus modelled the CPDO thereafter for rating purposes on the basis of an incorrect assumption that the actual average volatility of the Globoxx since inception was 15%. This assumption as to volatility was unreasonably and unjustifiably low. It did not represent either a reasonably anticipated or expected (that is, non-stressed) input or market condition or an exceptional but plausible (that is, stressed) input or market condition. I am satisfied that but for this error about volatility the CPDO could not have been rated AAA by S&P on any rational or reasonable basis. 23 Further, although there was no rational or reasonable basis for doing so other than the fact that one approach enabled the CPDO to satisfy the AAA rating quantile and the other approach did not, S&P decided that its base case for modelling and rating the CPDO should be an LTAS of 40 bps for one year and 80 bps for nine years, combined with an assumed starting spread of 36 bps, MR of 40% and volatility of 15%. Although the split between 40 bps and 80 bps was itself arbitrary, the most salient point is that the modelling showed that if the assumed LTAS of 40 bps was extended to two years then, all other assumptions being the same, the CPDO again did not meet the AAA rating quantile. S&P then drew a further arbitrary, irrational and unreasonable distinction between the lower LTAS of 40 bps lasting for one year as opposed to two years, the result of which was that on S&P's approach the CPDO achieved the AAA rating quantile default rate of 0.728% (which it did not if the lower LTAS of 40 bps continued for two years). 24 Despite this fact about LTAS (known to S&P), and the unjustifiably and unreasonably low assumption as to volatility of 15% induced by ABN Amro's misrepresentation and S&P's failure to calculate the actual average volatility of the Globoxx since inception (not at this time known to S&P but which should have been known had S&P acted with reasonable competence), S&P thereby assigned a rating of AAA to the first incarnation of the CPDO. 25 The first incarnation of the CPDO was thus rated AAA by S&P on the basis of: • an unjustifiably and unreasonably low assumption as to volatility of 15% induced by ABN Amro's misrepresentation and S&P's failure to calculate the actual average volatility of the Globoxx since inception; • an arbitrary, irrational and unreasonable distinction drawn by S&P between an LTAS of 40 bps for one year and 40 bps for two years; • a non-stressed assumption as to MR; • a non-stressed assumption as to RDB which, if eliminated and all other inputs remained the same, would change the modelled performance of the CPDO from AAA to sub investment grade (below BBB); • S&P defaults being stressed for a one and a half rather than a six month period to account for the fact that the CPDO would face 20 six month periods over its 10 year term and not just one such period, which was a reasonably stressed input as to defaults; and • there being no regard to the potential detrimental impact on the CPDO of ratings migration. 26 S&P's modelling, in common with that of ABN Amro (because S&P was using as the base for its modelling a model developed by ABN Amro for the rating of the DPN), also did not link the modelling of defaults and LTAS. In reality, if LTAS increased it was a reflection of perceived increased risk and thus defaults should have been higher for higher LTAS assumptions and lower for lower LTAS assumptions. S&P's modelling in this regard also would have benefited the performance and thus rating of the CPDO. This aspect of the model being used by both ABN Amro and S&P might also explain the reason why the CPDO performed better when the LTAS was higher. 27 As will be apparent from this description, S&P assigned the AAA rating to the first incarnation of the CPDO without assessing the CPDO's performance having regard to ranges of inputs or market conditions which included both reasonably anticipated or expected (that is, non-stressed) inputs or market conditions and exceptional but plausible (that is, stressed) inputs or market conditions. To the contrary, at least two of the major inputs were unjustifiably and unreasonably optimistic and had no proper rational foundation (volatility and lower LTAS for one year only), two were defensible in themselves but not stressed in any way (MR and RDB), and one was stressed (defaults). Apart from this, the detrimental impact on the CPDO's performance of ratings migration and the beneficial effect of modelling defaults separately from LTAS were overlooked. S&P's modelling and assignment of the AAA rating was not such as a reasonably competent ratings agency could have carried out and assigned in all of the circumstances. 28 For its part, ABN Amro actually knew all of these matters and that they benefited the modelled performance and thus rating of the CPDO apart from, perhaps, that relating to the separate modelling of defaults and LTAS. 29 S&P authorised ABN Amro to disseminate its rating of AAA of the CPDO to potential investors and ABN Amro did so. 30 After ABN Amro released the CPDO to the market on the basis of the AAA rating, a number of potential investors queried whether S&P had taken into account the cost of ratings migration. ABN Amro decided not to raise this with S&P because ABN Amro was worried that it would open "a can of worms" and send S&P back to the "drawing board" on the rating and ABN Amro did not want to risk delaying its marketing of the CPDO as it would then lose its advantage over other financial institutions in bringing the CPDO to market. ABN Amro thought also that S&P had stressed the defaults sufficiently to account for the failure to model ratings migration, although it had no empirical foundation for this belief and decided not to confirm its accuracy with S&P. ABN Amro thus continued to market the CPDO on the basis of S&P's rating of AAA. 31 Subsequently, ABN Amro created other versions of the CPDO including a CPDO in AUD known as Rembrandt 2006-2. In early August 2006 S&P modelled the Rembrandt 2006-2 CPDO on the same basis as it had ultimately modelled the initial incarnation of the CPDO - that is, an LTAS of 40 bps for one year and 80 bps for nine years, combined with an assumed starting spread of 36 bps, an MR of 40% and volatility of 15%, as well as the 7% RDB. 32 However, the Rembrandt 2006-2 CPDO was not in fact issued and S&P did not assign its rating of AAA to this instrument until 5 September 2006. By this time the average spreads on the Globoxx had reduced to 31.85 bps. ABN Amro knew the sensitivity of the CPDO to starting spreads and was concerned that at even 35 bps the rating of AAA for the standard ABN Amro CPDO (the CPDO as initially issued) would be "borderline". ABN Amro, however, did not mention the effect of the reduced spreads to S&P. S&P had not realised at this time how sensitive the rating of the CPDO was to the assumed starting spread, despite the modelling it had been given at the outset by ABN Amro disclosing this sensitivity. As such, despite the generally declining spreads, S&P did not model the performance of the Rembrandt 2006-2 CPDO at the actual starting spread on or close to the date of issue of that instrument and the assignment of the rating. Instead, S&P assigned a rating of AAA to the Rembrandt 2006-2 CPDO based on the modelling it had done a month earlier and at an assumed starting spread of 36 bps. 33 If S&P had modelled the Rembrandt 2006-2 CPDO at the correct starting spread of 31.85 bps the instrument would not have satisfied the AAA rating quantile and thus could not have been rated AAA by S&P, even on the basis of the unjustifiably and unreasonably optimistic inputs as to LTAS and volatility, non-stressed inputs as to MR and RDB, S&P defaults, and overlooking the effects of ratings migration and modelling defaults and LTAS separately. Again, a reasonably competent ratings agency could not have assigned the AAA rating to the Rembrandt 2006-2 CPDO in these circumstances. 34 S&P authorised ABN Amro to disseminate its rating of AAA the Rembrandt 2006-2 CPDO to potential investors and ABN Amro did so through its Australian office. 35 One investor who saw the Rembrandt 2006-2 CPDO and the AAA rating was Local Government Financial Services (LGFS). LGFS was an authorised deposit taking institution for councils in New South Wales which for most of its existence had been owned by the Local Government and Shires Associations. Those associations of councils had constituted LGFS for the purpose of giving councils greater power in the market for making deposits of councils' surplus funds. However, in 2004, the associations sold LGFS to the Local Government Superannuation Scheme (LGSS) which was owned by FuturePlus Financial Services Pty Ltd (FuturePlus), an entity unconnected to councils and the associations. 36 ABN Amro presented the Rembrandt 2006-2 CPDO to FuturePlus in July or August 2006 on the basis that it was expected to be rated AAA by S&P (as it later was), using as the basis for its presentation an ABN Amro marketing document which, amongst other things, referred to S&P's AAA rating and to the rating meaning that the CPDO had been modelled to have a default rate of less than 0.728%. 37 LGFS had been appointed by StateCover, a workers' compensation insurer, to manage StateCover's investments. On the basis of, amongst other things, S&P's rating of the Rembrandt 2006-2 CPDO as AAA, being the highest possible rating, and representations made by ABN Amro as to the creditworthiness of the CPDO having regard to the AAA rating and the meaning of the CPDO having a modelled default rate of less than 0.728%, which were decisive considerations for LGFS, LGFS caused StateCover to purchase $10,000,000 of the Rembrandt 2006-2 CPDO notes. 38 In so doing LGFS failed to appreciate that the CPDO notes were not a debenture and thus not a security but a derivative within the meaning of the Corporations Act 2001 (Cth) (the Corporations Act) in which StateCover was prohibited from investing pursuant to StateCover's investment policy which LGFS had itself prepared. LGFS also noticed but failed to appreciate the significance for an investor such as StateCover of the fact that the CPDO might experience high volatility over its term so that if the investment had to be liquidated at a point of low NAV, the investor might experience a substantial forced capital loss. LGFS also failed to appreciate the significance of the fact that there was no secondary market for the CPDO notes (although ABN Amro had said it intended to create one in the future) and that ABN Amro's promise to buy-back the notes on request was not necessarily to buy the notes back at the then NAV of the notes, but to buy them back at market value having regard to supply and demand for the notes at the time. 39 Thereafter, ABN Amro decided to issue another CPDO to be known as the Rembrandt 2006-3 CPDO, also on the basis of an expected AAA rating from S&P. ABN Amro decided to do this because LGFS had been looking for a structured financial product with a high rating and good return that would enable LGFS to compete with product providers which had been selling Collateralised Debt Obligations (CDOs) to councils with increasing success since about 2003. By 2005 the infiltration of CDOs into LGFS's traditional deposit taking business had caused LGFS's profits to plummet and threatened the ongoing viability of its traditional business model. 40 Having discovered the CPDO through the Rembrandt 2006-2 transaction, LGFS engaged ABN Amro to model and structure a transaction by which the Rembrandt 2006-3 CPDO notes would be issued with a rating of AAA from S&P. This was done on the basis that ABN Amro was aware that LGFS intended to on-sell as many of the Rembrandt 2006-3 CPDO notes as it could to LGFS's council clients in New South Wales. For its part, S&P was aware that the Rembrandt 2006-3 CPDO notes were intended to be purchased by a single investor in Australia. 41 The Rembrandt 2006-3 CPDO was scheduled to be issued by ABN Amro on 2 November 2006. By 10 October 2006 S&P had realised for the first time that the decreased spreads on the Globoxx adversely affected the CPDO so that, at an average of 32 bps as then was the case (and indeed had been the case when it rated the Rembrandt 2006-2 CPDO AAA), S&P believed that the standard ABN Amro CPDO could not be rated AAA. 42 By this time S&P had also decided that the volatility assumption of 15% had to be changed to 25%, although how S&P came to this realisation is unclear. By 12 October 2006 what is clear is that ABN Amro had calculated the actual average volatility of the Globoxx since inception to be 29% not 15%. To ABN Amro's knowledge, and on the basis of ABN Amro's descriptions of the Rembrandt 2006-3 CPDO as nothing more than a carbon copy of the Rembrandt 2006-2 CPDO, S&P assigned a rating of AAA to the Rembrandt 2006-3 CPDO on 31 October 2006 despite S&P having recognised before this that the volatility assumption needed to be increased from 15% to 25%. Further, because S&P treated the Rembrandt 2006-3 CPDO as a copy of the Rembrandt 2006-2 CPDO, S&P did not carry out any modelling of that instrument at all despite S&P knowing of the changed spread environment and its recognition that 15% volatility should (more to the point, could) no longer be used. 43 By this time, 31 October 2006, when S&P assigned the AAA rating to the Rembrandt 2006-3 CPDO notes, the following additional facts are also apparent: • S&P had realised that other ratings agencies examining the CPDO were making allowance for ratings migration when S&P had not; • Mr Rajan of S&P was expressing his continuing doubts about S&P's treatment of LTAS; • Mr Rajan of S&P was expressing the view that the "current rating (and high coupon) of the CPDO might be worth little, given the high ratings vol [volatility] with this product"; • S&P's New York office considered that market participants knew about the ABN Amro CPDO and were "in no hurry to stay in front of the truck"; • S&P was receiving daily calls on quantitative issues relating to CPDO criteria; • S&P had still not managed to complete its own model for evaluating the rating of CPDOs and was still using the model it had modified from ABN Amro's model for the DPN; • S&P considered it was dealing with "a crisis in CPDO land"; • there was debate within S&P about whether S&P's ratings were "under pressure" with the senior ratings analyst who was involved in the rating of the CPDO expressing the view that it was "analytical bs at its worst. I know how those ratings came about and they had nothing to do with the model!"; • a senior S&P analyst in S&P New York expressed the opinion that the S&P employees directly involved in the rating of the CPDO as AAA had been "sandbagged a little" at a time when S&P's model "was a work in progress" and ABN Amro "simply bulldozed it [the rating] through". This analyst recommended that as the AAA rating of CPDO had gained such huge attention in the markets S&P should either stick with all its assumptions "and emphasize that we stress other factors" or stick with its assumptions for existing deals only and then change its assumptions for future deals (S&P in fact adopted the latter option, at least insofar as it treated the Rembrandt 2006-3 CPDO as an existing deal, and thus stuck with its "existing assumptions" for that deal despite S&P having by then recognised the problems with its volatility and starting spread assumptions); and • another senior S&P employee considered that this "CPDO issue is a real mess" (confirmed a week later by another senior S&P employee who recommended that S&P simply advise ABN Amro that it had "messed up" the rating when S&P used 15% volatility). 44 S&P nevertheless assigned a rating of AAA to the Rembrandt 2006-3 CPDO. Leaving aside all of the issues of which S&P was aware as set out in the above paragraph: • if S&P had modelled the Rembrandt 2006-3 CPDO using 25% volatility (with all other inputs remaining the same) the CPDO would not have achieved the AAA rating quantile of 0.728% and could not have been rated AAA by S&P on any rational or reasonable basis; • if S&P had modelled the Rembrandt 2006-3 CPDO using the actual starting spreads on 31 October 2006 of 29.40 bps (with all other inputs remaining the same including even 15% volatility) the CPDO would not have achieved the AAA rating quantile of 0.728% and could not have been rated AAA by S&P on any rational or reasonable basis; and • if S&P had modelled the Rembrandt 2006-3 CPDO using the actual starting spreads on 18 October 2006, the date on which ABN Amro hedged that deal and which ABN Amro used to persuade S&P that the Rembrandt 2006-3 CPDO could be rated AAA despite the decreasing spreads, of 32.05 bps (with all other inputs remaining the same including even 15% volatility) the CPDO would not have achieved the AAA rating quantile of 0.728% and could not have been rated AAA by S&P on any rational or reasonable basis. 45 A reasonably competent ratings agency could not have rated the Rembrandt 2006-3 CPDO AAA in these circumstances. 46 S&P authorised ABN Amro to disseminate its rating of AAA for the Rembrandt 2006-3 CPDO to potential investors and ABN Amro did so through its Australian office including to LGFS. 47 On the basis of, amongst other things, S&P's rating of AAA, being the highest possible rating, and representations made by ABN Amro as to the creditworthiness of the CPDO having regard to the AAA rating and the meaning of the CPDO having a modelled default rate of less than 0.728%, both being decisive factors for LGFS, LGFS purchased $40,000,000 of the Rembrandt 2006-3 CPDO on 2 November 2006 and an additional $5,000,000 (for which S&P also confirmed a rating of AAA) in January 2007. 48 From November 2006 to June 2007 LGFS sold approximately $16,000,000 of CPDO notes to 13 councils in New South Wales which are applicants in these proceedings, as well as some notes to two other councils which are not applicants. LGFS held the balance of the Rembrandt 2006-3 CPDO notes (approximately $26,000,000) on its balance sheet. The AAA rating of S&P and the involvement of LGFS were both decisive considerations for each council in its decision to invest in the Rembrandt 2006-3 CPDO notes. 49 Two of the 13 councils which are applicants in these proceedings had entered into a contract with LGFS by which LGFS agreed to provide those councils with financial advice. The other councils had not entered into a contract with LGFS but all had had dealings with LGFS over many years by which they were induced to believe that LGFS would act in the best interests of the council and not prefer its own interests to those of the council. 50 Despite LGFS's contentions, evidence and submissions to the contrary I have found that LGFS was in a fiduciary relationship with the councils when it marketed to them the Rembrandt 2006-3 CPDO notes and, in its dealings with the councils, breached its fiduciary obligations by failing to disclose its conflict of interest in respect of the notes (its conflict arising from the circumstances of LGFS's falling balance sheet and need to enter the structured financial product market to compete with the vendors of CDOs). While LGFS marketed the notes to the councils on the basis of the AAA rating which S&P had in fact assigned to the notes, LGFS otherwise made numerous misrepresentations to the councils about the notes and their suitability as an investment for councils including but not limited to failing to disclose the notes' potential to be highly volatile and failing to explain adequately or at all the buy-back mechanism and its true operation. 51 One consequence of the global financial crisis which commenced part way through 2007 was sustained spread widening on the Globoxx. Sustained spread widening is spread widening with little or no mean reversion. As explained above, the CPDO makes money from mean reversion and was always vulnerable to cash-out if confronted by sustained spread widening. Accordingly, cashing out when confronted by sustained spread widening was an inherent risk or quality of the CPDO. 52 In response to the spread widening S&P downgraded its rating of the Rembrandt 2006-2 and 2006-3 CPDO notes from AAA to BBB+ in February 2008. By this time the NAV of the CPDO notes was less than 40% of par. Because it was not permitted to hold the product at the downgraded rating on its balance sheet pursuant to its investment policy, LGFS arranged to sell the Rembrandt 2006-3 notes it still held to its parent company, LGSS, at the then NAV, thereby sustaining a loss of principal of nearly $16,000,000. The councils continued to hold the notes until they cashed out in October 2008 after which the councils received back less than 10% of the principal they invested. 53 As explained in the reasons for judgment, I am satisfied that: • S&P's rating of AAA of the Rembrandt 2006-2 and 2006-3 CPDO notes was misleading and deceptive and involved the publication of information or statements false in material particulars and otherwise involved negligent misrepresentations to the class of potential investors in Australia, which included LGFS and the councils, because by the AAA rating there was conveyed a representation that in S&P's opinion the capacity of the notes to meet all financial obligations was "extremely strong" and a representation that S&P had reached this opinion based on reasonable grounds and as the result of an exercise of reasonable care when neither was true and S&P also knew not to be true at the time made; • ABN Amro was knowingly concerned in S&P's contraventions of the various statutory provisions proscribing such misleading and deceptive conduct, and also itself engaged in conduct that was misleading and deceptive and published information or statements false in material particulars and otherwise involved negligent misrepresentations to LGFS specifically and the class of potential investors with which ABN Amro knew LGFS intended to deal, being the councils, by reason of ABN Amro's deployment of the AAA rating and its own representations as to the meaning and reliability of the AAA rating which also were not true and ABN Amro knew not to be true at the time made; • ABN Amro also breached its contract with LGFS under which ABN Amro was to model and structure the transaction by which LGFS would purchase the Rembrandt 2006-3 CPDO notes having a rating assigned by S&P of AAA; and • LGFS engaged in misleading and deceptive conduct and in one respect the publication of an information or statement false in material particulars and otherwise made negligent misrepresentations to the councils about the Rembrandt 2006-3 CPDO notes and, in addition, breached its Australian financial services licence in advising the councils about and selling to them the notes because the notes were not a debenture and thus not a security but a derivative under the Corporations Act in which LGFS was not licensed to deal and otherwise breached its fiduciary duties to each of the councils. 54 I have rejected the arguments made by LGFS, S&P and ABN Amro of contributory negligence or the equivalent by the councils. I have also rejected the arguments made by S&P and ABN Amro of contributory negligence or the equivalent by LGFS in respect of, relevantly, the Rembrandt 2006-3 CPDO notes which LGFS did not sell to councils and continued to hold until forced to sell at a loss to its parent company. 55 As also explained in the reasons for judgment, I am satisfied that, on the basis of the contravening conduct of S&P, ABN Amro and LGFS above: • LGFS has established its entitlement to damages against S&P and ABN Amro (which are proportionally liable as to 50% each) for LGFS's loss incurred on the sale of the downgraded Rembrandt 2006-3 CPDO notes to its parent company LGSS ($15,970,184.72); • LGFS has established its entitlement to damages or equitable contribution from S&P and ABN Amro (which are proportionally liable as to 33⅓% each with LGFS liable for the other 33⅓%) in respect of the settlement LGFS made of the StateCover proceedings against LGFS, S&P and ABN Amro ($3,175,000); • the councils have established their entitlement to damages from S&P, ABN Amro and LGFS (which are proportionally liable as to 33⅓% each) being the difference between the principal amount each paid and the payment they received on the cash-out of the notes; and • the councils have also established their entitlement to equitable compensation from LGFS for breach of fiduciary duty but the measure of equitable compensation is the same as the damages otherwise payable by S&P, ABN Amro and LGFS as to 33⅓% each. 56 LGFS has also established that it is entitled to indemnity under the contract of insurance between FuturePlus and American Home Assurance Company.
Bathurst Regional Council
Local Government Financial Services Pty Ltd
Judgment for the plaintiff councils and for LGFS on its cross-claim; damages awarded against S&P, ABN Amro and LGFS on a several basis (33⅓ % each for council claims, 50 % each for LGFS’s direct loss, and 33⅓ % each including LGFS for the StateCover settlement); equitable compensation from LGFS to councils co-extensive with damages; LGFS entitled to insurance indemnity.