L’Inspection générale (SEGL/INS) a été saisie par le Comité exécutif du Groupe le 24 janvier 2008 pour conduire une mission d’investigation sur la fraude opérée par Jérôme KERVIEL (JK), trader au sein de l’activité de dérivés actions sur le marché des warrants « turbo » (Cf. Focus n°1).
L’objet de la mission, validé par le Comité spécial du Conseil d’administration de la Société Générale, a consisté à (i) décrire le mécanisme de la fraude afin de vérifier que les montants de positions et de perte calculés par SG CIB étaient exacts, (ii) rechercher les motivations et éventuelles complicités de JK, (iii) identifier les dysfonctionnements des contrôles et les responsabilités à l’origine de la détection tardive de la fraude, (iv) vérifier que de telles fraudes n’existent pas sur d’autres périmètres de SG CIB.
S’agissant des dysfonctionnements et des responsabilités, notre étude ayant exclusivement porté sur le périmètre de la fraude, nos conclusions ne s’appliquent qu’à la table DELTA ONE, au sein de laquelle JK opérait, et son environnement immédiat. L’organisation de SG, notamment celle de DELTA ONE, est présentée en Focus n°2 (Présentation de l’organisation SG).
Le présent rapport dresse les conclusions de la mission au 20/05/2008, les travaux sur la fraude ayant été achevés à cette date (1).
(1) Nos travaux de rapprochement n’ont pas pu être menés à leur terme s’agissant des positions titres en raison de travaux de réconciliation transversale en cours chez OPER. Nos travaux n’ont pas non plus couvert de manière exhaustive l’étude des suspens cash, leur traitement n’étant pas différencié selon le centre opératoire. Enfin, nous avons concentré nos travaux d’écoute des bandes téléphoniques et de revue de boites e-mails sur les interlocuteurs et périodes clés mais n’avons pas couvert de manière exhaustive la période 2005 – 2008 pour l’ensemble des protagonistes.
How Leeson Broke Barings part 3
The BoBS report notes "In each instance, the entries in the Contac system reflected a number
of spurious contract amounts at prices different to those transacted on the floor,
reconciling to the total lot size originally traded. This had the effect of giving
the impression from a review of the reported trades in account '92000&' that these had taken place at different times during the day. This was necessary to deceive Barings Securities Japan into believing the reported profitability in account '92000' was a result of authorised arbitrage activity. The effect of this manipulation was to inflate reported profits in account '92000&' at the expense of account '88888', which was also incurring substantial losses from the unauthorised trading positions taken by Leeson. In addition to crossing trades on SIMEX between account '88888&' and the switching accounts, Leeson also entered fictitious trades between these accounts which were never crossed on the floor of the Exchange. The effect of these [off-market trades, which were not permitted by SIMEX], was again to credit the 'switching' accounts with profits whilst charging account '88888' with losses."
The bottom line of all these cross-trades was that Barings was counterparty to many of its own trades. Leeson bought from one hand and sold to the other, and in so doing did not lay off any of the firm's market risk. Barings was thus not arbitraging between SIMEX and the Japanese exchanges but taking open (and very substantial) positions, which were buried in account '88888'. It was the profit and loss statement of this account which correctly represented the revenue earned (or not earned) by Leeson. Details of this account were never transmitted to the treasury or risk control offices in London, an omission which ultimately had catastrophic consequences for Barings shareholders and bondholders.
Figure 10.32, below, shows the number of cross-trades executed by Leeson. It is the difference between the solid line which represents all the Nikkei trades of account '92000' not crossed into account '88888' and the broken line which reflects the position Leeson reported to Barings management. The figure graphically illustrates the chasm between reported and actual positions. For example, Barings management thought the firm had a 'short' position of 30,112 contracts on SIMEX on 24 February; in fact it was long 21,928 contracts after ignoring the trades crossed with account '88888'.
Figure 10.3 Graph to show the Nikkei Position of Account '92000'. Reproduced by permission from the Report of the Board of Banking Supervision Inquiry into the Circumstances of the Collapse of Barings.
Footnotes:
1, 2) Report of the Banking Supervision Inquiry into the Circumstances of the Collapse of Barings, Ordered by the House of Commons, July 1995, Her Majesty's Stationery Office, London
How Leeson Broke Barings part 2
But Leeson's Osaka position, which was public knowledge since the OSE publishes weekly data, reflected only half of his sanctioned trades. If Leeson was long on the OSE, he had to be short twice the number of contracts on SIMEX. Why? Because Leeson's official trading strategy was to take advantage
of temporary price differences between the SIMEX and OSE Nikkei 225 contracts. This arbitrage, which Barings called 'switching', required Leeson to buy the cheaper contract and to sell simultaneously the more expensive one, reversing the trade when the price difference had narrowed or disappeared. This kind of arbitrage activity has little market risk because positions are always matched.
But Leeson was not short on SIMEX, infact he was long approximately the number of contracts he was supposed to be short. These were unauthorised trades which he hid in an account named Error
Account 88888. He also used this account to execute all his unauthorised trades in Japanese Government Bond and Euroyen futures and Nikkei 225 options: together these trades were so large that they ultimately broke Barings. Table 10.1 gives a snapshot of Leeson's unauthorised trades versus the trades that he reported.
For the rest of the chapter, contracts will be discussed or converted into SIMEX contract sizes.
Unreported positions (Fact)
The most striking point of Table 10.1 is the fact that Leeson sold 70,892 Nikkei 225 options worth about $7 billion without the knowledge of Barings London. His activity peaked in November and December
1994 when in those two months alone,
| Table 10.1 | Fantasy versus Fact: Leeson's Positions as at End February 1995. | ||
|---|---|---|---|
| Number of contracts1 nominal value in US$ amounts |
Actual position in terms of open interest of relevant contract2 | ||
| Reported3 | Actual4 | ||
| Futures | |||
| Nikkei 225 | 30112 $2809 million |
long 61039 $7000 million |
49% of March 1995 contract and 24% of June 1995 contract. |
| JGB | 15940 $8980 million |
short 28034 $19650 million |
85% of March 1995 contract and 88% of June 1995 contract. |
| Euroyen | 601 $26.5 million |
short 6845 $350 million |
5% of June 1995 contract, 1% of September 1995 contract and 1% of December 1995 contract. |
| Options | |||
| Nikkei 225 | Nil | 37925 calls $3580 million 32967 puts $3100 million |
|
| 1. Expressed in terms of SIMEX contract sizes which are half the size of those of the OSE and the TSE. For Euroyen, SIMEX and TIFFE contracts are of similar size. 2. Open interest figures for each contract month of each listed contract. For the Nikkei 225, JGB and Euroyen contracts, the contract months are March, June, September and December. 3. Leeson's reported futures positions were supposedly matched because they were part of Barings' switching activity, i.e. the number of contracts on either the Osaka Stock Exchange, the Singapore International Monetary Exchange or the Tokyo Stock Exchange. 4. The actual positions refer to those unauthorized trades held in error account '8888'. Source: The Report of the Board of Banking Supervision Inquiry into the Circumstances of the Collapse of Barings, Ordered by the House of Commons, Her Majesty's Stationery Office, 1995 |
|||
he sold 34, 400 options. In industry parlance, Leeson sold straddles. i.e. he sold put and
call options
with the same strikes and maturities. Leeson earned premium income from selling well over 37,000 straddles over a fourteen month period. Such trades are very profitable provided
the Nikkei 225 is
trading at the options' strike on expiry date since both the puts and calls would expire worthless. The seller then enjoys the full premium earned from selling the options. (see Fig 10.2 for a graphical presentation of the profit and loss profile of a straddle.) If the Nikkei is trading near the options' strike
on expiry, it could still be profitable because the earned premium more than offsets the small loss experienced on either the call (if the Tokyo market had risen) or the put (if the Nikkei had fallen.).
Figure 10.2 Payoff Profile of a Straddle.
How Leeson Broke Barings
| The activities of Nick Leeson on the Japanese and Singapore futures exchanges, which led to the downfall of his employer, Barings, are well-documented. The main points are recounted here to serve as a backdrop to the main topic of this chapter - the policies, procedures and systems necessary for the prudent management of derivative activities.
Barings collapsed because it could not meet the enormous trading obligations, which Leeson established in the name of the bank. When it went into receivership on February 27, 1995, Barings, via Leeson, had outstanding notional futures positions on Japanese equities and interest rates of US$27 billion: US$7 billion on the Nikkei 225 equity contract and US$20 billion on Japanese government bond (JGB) and Euroyen contracts. Leeson also sold 70, 892 Nikkei put and call options with a nominal value of $6.68 billion. The nominal size of these positions is breathtaking; their enormity is all the more astounding when compared with the banks reported capital of about $615 million. The size of the positions can also be underlined by the fact that in January and February 1995, Barings Tokyo and London transferred US$835 million to its Singapore office to enable the latter the meet its margin obligations on the Singapore International Monetary Exchange (SIMEX). Reported activities (Fantasy) The build-up of the Nikkei positions took off after the Kobe earthquake of January 17. This is reflected in Figure 10.1 - the chart shows that Lesson's positions went in the opposite direction to the Nikkei - as the Japanese stock market fell, Leeson's position increased. Before the Kobe earthquake, with the Nikkei trading in a range of 19,000 to 19,500, Leeson had long futures positions of approximately 3,000 contracts on the Osaka Stock Exchange. (The equivalent number of contracts on the Singapore International Monetary Exchange is 6000 because SIMEX contracts are half the size of the OSE.) A few days after the earthquake Leeson started an aggressive buying programme which culminated in a high of 19,094 contracts reached about a month later on February 17.
Figure 10.1 Baring's Long Positions against the Nikkei 225 Average. |
Lack of supervision
|
Theoretically Leeson had lots of supervisors; in reality none exercised any real control over him. Barings operated a 'matrix' management system, where managers who are based overseas report to local administrators and to a product head (usually based at head office or the regional headquarters). Leeson's Singapore supervisors were James Bax, regional manager South Asia and a director of BFS, and Simon Jones, regional operations manager South Asia, also a director of BFS and chief operating officer of Barings Securities Singapore. Jones and the heads of the support functions in Singapore also had reporting lines to the Group-wide support functions in London. Yet both Bax and Jones told the BoBS inquiry that they did not feel operationally responsible for Leeson. Bax felt Leeson reported directly to Baker or Walz on trading matters and to Settlements/Treasury in London for backoffice matters. Jones felt his role in BFS was limited only to administrative matters and concentrated on the securities side of Barings' activities in South Asia.
Leeson's reporting lines for product profitability are not clear cut since his supervisors have disputed who was directly responsible for him from January 1, 1994. His ultimate boss was Ron Baker, head of the financial products group. But who had day-to-day control over him? Mary Walz, global head of equity financial products, insists that she thought Fernando Gueler, head of equity derivatives proprietary trading in Tokyo was in charge of Leeson's intra-day activities since the latter's switching activities were booked in Tokyo. However, Gueler insists that in October 1994, Baker told him that Leeson would report to London and not Tokyo. He thus assumed that Walz would be in charge of Leeson. Walz herself still disputes this claim. Tapes of telephone conversations show that Leeson spoke frequently to both Gueler and Walz. (The bottom line however is that Gueler reported to Walz.) Two important incidents vividly illustrate the cavalier attitude Barings had towards supervising Leeson. The first involves two letters to BFS from SIMEX. In a letter dated 11 January, 1995; SIMEX senior vice-president for audit and compliance Yu Chuan Soo, complained about a margin shortfall of about US$116 million in account '88888' and that Barings had appeared to break SIMEX rule 822 by previously financing the margin requirements of this account, (which appeared in SIMEX's system as a customer account.) SIMEX also noted that the initial margin requirement of this account was in excess of US$342 million. BFS was asked to provide a written explanation of the margin difference on account '88888' and of its inability to account for the problem in the absence of Leeson. No warning lights went off in Singapore. No one investigated who this customer really was and why he was having difficulties in meeting margin payments or why he had such a huge position; or the credit risk Barings faced if this 'customer' defaulted on the margins that Barings had paid on its behalf. A copy of the letter was not sent to operational heads in London. Simon Jones did not press Leeson for an explanation; indeed he dealt with the matter by allowing Leeson to draft Barings' response to SIMEX. The second incident did come to the attention of London but again was dealt with unsatisfactorily, perhaps because Barings' personnel themselves are unsure about what really happened. At the beginning of February 1995, Coopers & Lybrand brought to the attention of London and Simon Jones the fact that US$83 million apparently due from Spear, Leeds & Kellogg, a US investment group, had not been received. No one is sure how this multi-million dollar receivable came about. One version of events is that BFS, through Leeson, had traded or broked an over-the-counter deal between Spear, Leeds & Kellogg, and BNP, Tokyo. The transaction involved 200 50,000 call options, resulting in a premium of 7.778 billion (US$83 million). The second version was that an 'operational error' had occurred; i.e. a payment had been made to a wrong third-party in December 1994. Both versions had very serious control implications for Barings. If Leeson had sold or broked an OTC option, then he had engaged in an unauthorised activity. Yet he was not admonished for doing so; nor is there any record of Barings' management taking any steps to ensure that it did not happen again. If the SLK receivable was an operational error, Barings had to tighten up its back-office procedures. |