No tears were shed in the marketplace when an embarrassed NatWest announced in mid-March that some of its interest rate options books had been mispriced over the last three years to the tune of £85 million.
A straw poll of options dealers and brokers suggests that NatWest’s mispricing had been an open secret. One US investment bank, having looked at NatWest’s prices and done its own sums, decided to call the bank’s bluff and bought, according to market sources, a large amount of out-of-the money options from NatWest. A US regional bank did likewise.
Mispricing of options is often a matter of opinion, and is easily forgiven by people who make money from it. But it was the arrogance of NatWest’s options and swaps traders which alienated brokers and dealers in the market.
Whatever the outcome of an inquiry by Cooper’s & Lybrand and Linklaters & Paines into the conduct of these derivatives traders and their controllers, there is some satisfaction among competitors that the NatWest traders got their come-uppance. Jean-François Nguyen, global head of derivatives since 1993, was not a popular figure outside the bank. And if former colleagues are to be believed, he exercised a reign of fear and cliquism within it.
Few excuses
NatWest Markets chief executive Martin Owen says of the slip-up: “We regard it as an unacceptable loss, taken on in an unacceptable way, in an unacceptable time-frame.” He also took a voluntary £200,000 cut in his £500,000 annual bonus.
But that exercise in self-criticism is not likely to exonerate NatWest and its managers in the eyes of the market. They have dealt another blow to derivatives’ reputation as a risk management tool. And they have given regulators another excuse to exercise a heavier hand. Says a former NatWest derivatives manager in Singapore: “A few more of these disasters will have a huge impact on the reputation of the business and the onslaught of the regulators.” It’s not just NatWest’s problem, it’s one for every firm that deals in derivatives.
Accordingly, NatWest’s reticence in explaining exactly what went wrong is not well judged. Owen said in an interview with Euromoney: “It’s not our intention to ever explain the whole series of incidents. The only thing we’re likely to make public is what we do in relation to people – because there’s a legitimate interest in it.”
Market traders and risk managers will be forced to draw their own conclusions about what went wrong at NatWest, unless, that is, the Bank of England does a Barings-type inquiry.
The legitimate question to ask of the NatWest case: Is there something inherently unstable in the pricing of long-dated out-of-the-money options? Or was it simply a question of flawed management?
In his interview, Owen used the terms “false accounting” and “false values”, but emphasized that the inquiry was still going on. “If you have a loss which has been covered over – the teaming approach,” he says, “you can set false values which are very valuable to a third party – very complicated.”
Baring Brothers chairman Peter Baring’s first reaction to his bank’s demise was to blame a conspiracy. NatWest’s was the branding of Kyriacos Papouis, the options trader who left the bank in December, as a “rogue trader”.
Comments Leslie Rahl principal of Capital Markets Risk Advisors in New York: “Being scammed by a rogue trader is somehow more socially acceptable.” Than what? In Owen’s case perhaps more acceptable than admitting the risk management culture of his bank had a serious flaw.
But first we must examine the question, are some kinds of option risk simply unmanageable?
Much has been written, since the NatWest incident, about the volatility smile. This is the tendency, empirically observed in the market, for options that move in or out of the money to go through more violent price fluctuations at the extremes than the theoretical Black-Scholes models predict. The probability of severe gains or severe losses seems to accelerate at the edges. Dealers know about this and they allow for it.
But pricing the smile scientifically is only partly possible. Filtering the results also requires common sense, says Stephen Myers, head of risk management at Bankgesellschaft Berlin in London. Myers is one of the few poachers-turned-gamekeeper, a former sharp-end derivatives dealer who minds the back office at Bankgesellschaft and knows most tricks in the dealer’s book.
“I also know how back offices check volatilities with the market,” says Myers, “they pay little attention to smiles. You need someone to ask with common sense ‘Is this a reasonable smile?'”
For financial controllers and risk managers to check the price of the smile independently, they have to go to the brokers in the market and ask for prices. If particular out-of-the-money options are barely traded then, if they get prices at all, they will have a wide bid-offer spread. A bid-offer spread of 2% or 3% can make a nonsense of an option’s profitability. If the option’s price is so uncertain, should a prudent bank have it on its books at all? Good question.
Digestible risk
The reply, even from responsible dealers, is that, in aggregate, options books make sufficient money. In many cases options are taken on first, because of customer demand, and the hedging costs and potential losses are dealt with later. On the swings-and-roundabouts principle the bank will still make money on its options business, even if it suffers some unexpected but digestible losses on the way. Some banks put aside reserves for “model risk”, the risk that their model is out of whack with the market.
But errors can be compounded, even if values are re-examined by independent financial controllers. If the bank is itself a big market-maker in the product, which NatWest was, or was trying to be, then trying to get independent prices without the input of NatWest’s own traders is near impossible.
The first line of defence for financial controllers is to generate their own interest-rate and volatility curves independently, from the market. Brokers are usually willing to send faxes to regular clients giving a list of dealing prices for selected instruments. On rarely-traded out-of-the-money options, brokers can get an indicative price from the market, but it is only as reliable as the dealer who makes the quote. To get a valuation on a large number of odd-date, odd-strike options is time-consuming for the broker and financial controller. The bank also risks giving its positions away to the market. A short-cut is for financial controllers to ask their dealers to enter their deals on a matrix, from which they will randomly check prices and volatilities. But the matrix must be updated to take account of changes in implied volatility.
One options trading expert argues that any good risk controller should be able to generate and calibrate from the market, a price matrix which is within a few basis points of actual market volatilities. The fact that NatWest’s hole was so big means that it wasn’t due to poor calibration of its trading or risk management models, he opines. More like no calibration at all, but simply allowing traders to put in their own volatilities, he says. “If you let traders do that, they can cheat.”
There have been many cases of this over the years. The only remedy is to have middle-office financial controllers and back-office risk managers who are clever enough, and are paid well enough, to stand up to superstar traders. Every investment bank chief should be aware of this, especially since the blow-up of Barings in March 1995.
Is there any possibility in the NatWest case that the models used could have been deficient? NatWest in its March 13 press release said emphatically no. But former NatWest derivatives staff have few good words to say of NatWest’s internally built Interest Rate Risk Management System (IRRMS). One remembers how IRRMS behaved during the September 1992 European exchange rate mechanism (ERM) crisis. “You couldn’t design a worse system than IRRMS for risk management. Positions that were worthless suddenly became valuable. It was like studying a piece of A4 paper through a drinking straw.”
IRRMS may have improved. But the technicians in 1992 said it would be a five-year job to upgrade it, the former NatWest man recalls.
As a back-up in London, NatWest installed Oberon, developed by Lombard Risk Systems. Oberon was “the first system in the market that allowed a credible smile effect for caps and floors,” says John Wisbey, chairman and CEO at Lombard. “But it’s up to a bank how it uses it.”
There’s the rub, say some former NatWest employees. It appears, although this is impossible to check without NatWest’s cooperation, that IRRMS was still in 1997 being used as London’s major risk management system. In New York a different system was used on the insistence of derivatives bosses there – Abacus, developed by Openlink Financial from software first used by Fuji Capital. Abacus has been introduced to other NatWest locations under the name GDS (Global Derivative System) and is also used by the treasury and FRA group in London. The two options books that went wrong in London, the Deutschmark and sterling books, were not handled by GDS, say former NatWest sources.
Most risk management systems have teething troubles. IRRMS was no exception. Here are some jaundiced comments from the early days: “In 1991 the in-house system was taking 16 hours to update the derivatives book,” says a former NatWest derivatives manager based in New York. “By 1993 it was taking 44 hours to do a P&L [profit & loss] report.” Across systems there are always reconciliation problems on the profit-and-loss. “With IRRMS and Oberon there was a P&L reconciliation problem once every two weeks,” the former manager says.
“IRRMS couldn’t value the forward volatility curves,” says a more recent ex-NatWester in New York, “although Oberon was reasonably state-of-the-art.” However, NatWest has steadily been upgrading its systems in London, says a software expert. Caps and floors were upgraded a year ago, before technicians moved on to include swaptions. It seems that IRRMS was handling options and Oberon swaptions, say sources close to NatWest.
Oberon has also been through periodic upgrades. Like most systems, say risk managers, it has adapted its mathematics as the science of calculating forward volatilities has progressed. Says a former NatWest derivatives manager :“Three years ago Oberon had a particular problem with volatility curves.” Other risk managers say it was not unique to Oberon, it was a problem of mathematics, not the system. “But maybe NatWest were using an outdated release [of Oberon],” the derivatives manager muses. There is certainly inertia among users to update every time a software house issues a new release. “There’s often no overriding reason to upgrade.”
Wisbey at Lombard emphasizes that NatWest’s inquiry immediately exonerated the “industry-standard models” used to calculate valuations of the products in question. While different software systems have various strengths and weaknesses and none has a totally bug-free history, that is not relevant in this case, he says.
If, as many in the market suspect, NatWest’s financial controllers were ignoring the volatility curve and assuming straight-line volatility, then Oberon’s capabilities weren’t being used to the full. It was an input problem, not a systems problem.
Discrepancies between IRRMS and Oberon should have helped alert risk controllers to any glaring anomaly in valuation, such as an £85 million hole. But if the two systems did throw up anomalies they were ignored until after December last year when, according to Owen “it was risk management and finance which eventually got it – continuing to go through the evaluation reports as part of the year-end financial accounting. They examined and examined but couldn’t get to the bottom of it”.
But how is it that NatWest’s risk managers allowed the use of straight-line volatility, or some other means of misrepresenting the mark-to-market value of these options? The trading risk team under managing director Bob Brooks was widely respected internally and externally.
Price assumptions
Risk managers typically do not monitor the financial inputs, they leave that to financial control. Apart from a few random checks they assume that financial control has verified dealer prices independently with brokers and available screen prices. Such price checks only began at NatWest Markets in 1995, according to a former employee. Those prices were compared with the output of Oberon. “The parameters and settings of Oberon were supposed to be part of the trading risk department, which puts in volatility,” says a financial control source. “When the report is kicked out of the system we assume it’s been tested by other people.” From this comment it seems that NatWest’s financial controllers weren’t naturally suspicious of the output from trading risk.
Outside observers are tempted to think that it was risk managers from Greenwich Capital, the US fixed-income securities broker/dealer which NatWest bought last June, who pinpointed the problem. Owen denies that, and so do other sources close to NatWest. But insiders say that Greenwich took over bond and bond options books as early as last July, and they were due to take over global fixed income from January. It would be surprising if Greenwich weren’t beginning to take a close look in January at the books they were supposed to be taking over. Owen says: “Greenwich have been used as part of the investigative process. The process of identifying the problem was handled by NatWest Markets people.”
The problem at NatWest appears to have been between the front office, the dealers, the trading risk reports that were produced, and financial control. But at the root, it may have been a cultural problem. A bank still run in part by traditional corporate bankers was trying to win market share by taking considerable trading risks.
Since Barings, a lot has been talked and written about the star culture and its dangers. It seems NatWest had been nurturing a star by the name of Jean-François Nguyen.
Nguyen joined NatWest in August 1992 from Credit Suisse Financial Products (CSFP) where he had been trading French franc swaps. He had tried to bring over a four-man team, but in the end only one other trader, Christophe Lanson, came with him. At CSFP Nguyen had been regarded as a good but not outstanding trader. So there was some surprise among his former bosses when Nguyen rose within a year at NatWest to become global head of derivatives.
The key was the September 1992 ERM crisis. Like many other houses NatWest was caught with bigger positions in its sterling interest rate options book than it had reckoned with. The consensus among senior derivatives managers in New York and London was that they shouldn’t hedge expensively but sit tight and ride out the storm. “But Jean-François had just been hired [as a senior trader for swaps and bonds] and immediately closed out the positions at a loss of £14 million,” says a derivatives colleague of that time. This made him look good and the originators of those positions bad.
Nguyen’s aggressive trading in the next few months alarmed some of his more experienced colleagues who knew that NatWest’s systems weren’t sophisticated enough to keep track. A senior derivatives manager recalls expressing “strong misgivings” to the NatWest management about the bank’s “ability to support a high volume of business in plain vanilla options let alone any exotic books”. The same person also voiced concerns about “the style of trading in the London derivatives group”.
But it seemed Nguyen could do little wrong. He had been hired into a bull market and was playing it for all it was worth. “The P&L was up five times what it was before he came,” admits a former colleague. His bonus in I993 was probably between £3 million and £5 million on derivative trading profits of perhaps £100 million, estimates a former NatWester. Nguyen was rewarded by being made global head of derivatives.
This child of the bull market was “arrogant, yet very parochial” says a former colleague. While he had been simply quiet and “not one of the boys” at CSFP, now, according to another former colleague, he was “very overbearing, very domineering, everybody was afraid of him, you had to be part of the clique. There was a lot of fear within the group. The traders felt vulnerable”.
Within six months, says this source, “the group lost its professional character. Once you feel above the mechanisms everything is open to abuse”.
It must be stressed that these observations come from people who have left NatWest and may be charged with some animosity. But at the least they reflect concerns at the time about the star treatment given to Nguyen.
“I did warn them,” says one senior ex-NatWester. “I spent some time with Martin [Owen], Phil Augur [then head of risk management, now vice-chairman of Schroders] and Phil Wise [subsequently head of rate risk management, now suspended]. Did Owen understand that [in the global derivatives group] you had to be part of the clique?”
Global ambitions
After the bull year of 1993, came the bear year for fixed income of 1994. Customer business tailed off and the only chance of maintaining last year’s profit level lay with proprietary trading. NatWest Markets, from chief executive Owen downward, had ambitions to be a global player. Nguyen appeared to be the answer in the area of derivatives. Did that make it difficult for the financial controllers to question what he was doing? In the words of one risk management expert: “Little squirts in the middle office earning £20,000 a year don’t stand in the way of a producer earning £200,000 or £2 million.”
Owen objects: “We’ve got some very highly paid people in the control functions, who earned more than Papouis [the options trader who left in December].”
However, it seems Nguyen’s traders weren’t questioned enough. Despite the downturn in customer derivatives business in 1994 and 1995 the profits recorded by the derivatives group appear to have been buoyant. “Nguyen has probably earned around £3 million to £5 million each year in bonuses on profits of about £100 million a year – although a lot of that came from the NatWest franchise,” says a former NatWest derivatives manager. Of the £8 million in bonuses for 1996, clawed back by NatWest in March after announcing the loss, half had been destined for Nguyen say market sources.
Competitors in the market say they were suprised by the level of bonuses awarded to these people. Perhaps another factor in the mismanagement of derivatives was Nguyen’s control over his traders’ bonus levels and the pressure to cut overheads to increase profits. Competitors were surprised that at NatWest Markets only three traders – Neil Dodgson, Papouis and Isabel Morresi – were running a plethora of complex options books.
Outside London, some NatWest derivative chiefs didn’t like the way Nguyen was trading and said so. One decided to pass an exotic swap book back to London to manage because “I didn’t agree with the way Nguyen was pricing”.
In early 1994, some say late 1993, the market began to notice that NatWest Markets was selling options cheaply. There can be a lot of reasons why options are aggressively priced. A bank may have a big offsetting position to lay off – although this is unusual if it is selling volatility. It may want to win market share. Or it may be mispricing because it has more faith in its own model than in the market.
But a sustained period of mispricing, from early 1994 until at least the middle of 1996, suggests something else. NatWest won’t detail the portfolio, but a market consensus believes NatWest’s sterling and Deutschmark options dealers were creating the illusion of profit by matching options they had sold, either at strike prices which had since gone out of the money, or which were initially out of the money, against options of other maturities and perhaps at other strike prices. They could take advantage of the straight-line volatility assumed by the financial controllers to show that the book was in net profit, ignoring the potential losses inherent in the mismatch.
The illusion of profit can also be created by matching options against swaptions. Swaptions have a lower volatility since they are a series of interest rate options. Buying swaptions and selling options can bring a net gain in premium and show a superficially matched position.
If a bank in this position is indifferent to the volatility mismatch it is also indifferent to pricing out of line with the market. In the rarefied world of out-of-the-money options there often is no liquid market.
Says one NatWest competitor: “We bought a lot of out-of-the-money options from them because we thought they were selling them too cheaply.”
Why didn’t someone in the market alert NatWest Markets? The answer is that firms today rarely forgive each other’s professional mistakes, they see them as an opportunity. Owen complains: “There’s evidence that certain financial firms would have made significant financial gains over a short period.” He notes that in earlier days “if we found something curious happening we would tell each other,” but now “we’ve become electronically connected not personally connected”.
Senior NatWest management with antennae out in the market might have got wind of something. But that is probably expecting too much. Did Barings or the Bank of England get wind of the Barings problem?
Chemical’s magic formula
An interesting case is that of Chemical Bank in 1988 and 1989. In early 1988 it was well known in the market that Chemical was selling interest rate caps too cheaply. In an article in Riskmagazine of April 1988 this writer related how other banks were able to buy caps from Chemical and on-sell them to clients at a profit. Chemical’s then managing director of global securities Steve Edelson hinted at a hidden source of supply and continued: “What we’re doing is very complicated – a building block on a building block – we keep finding new nuances. If you don’t understand the parts you can get into a lot of trouble. It’s like … a neutron bomb … a tremendous risk, but not a market risk.”
The neutron bomb took some time to blow up. However, there was some substance in Edelson’s reference to a source of supply. As one former Chemical trader explains: “On the Tuesday after the October 1987 stock market crash, US bonds blew up. Short-end volatility on the Wednesday was 51%.” While the street’s senior managers were ordering a close-out of positions Edelson and his team argued successfully that volatility was bound to come down.
“We sold about $5 billion of caps and sold straddles into Chicago,” says the trader. The upshot was that in 1988 “we were riding high on the fact that we were very short volatility, and volatility was coming down”. By the end of that year Chemical’s cap book showed reserves of $34 million. Senior management had brought in Arthur Andersen to check it out.
That 1988 triumph was the making and undoing of Edelson and his group. In 1989, believing their cap model was better than the street, they were happy to sell aggressively out-of-the-money caps on the US commercial paper rate, hedging them with Libor swaps. A shock loss of around $1 million alerted some traders to the basis risk, but that scarcely jolted the bandwagon. Edelson, a big position taker, was now convinced that volatility would rise. In fact it continued to fall. “Steve started taking big gamma and omega risk. Management believed in him,” says the trader. At one point the leverage was such that “for every 1% move in volatility, we gained or lost $13 million”.
The Chemical cap book had a small blow-up in March 1989 and a big one in September. That blow-up cost the bank the reserve of $34 million and a further $30 million, not to mention its reputation as a cap dealer. Edelson subsequently left the bank.
The lesson, for Chemical, NatWest and the rest of us, is that being out of line with the market is an alarm call. Chemical, having checked the first alarm, didn’t bother to keep checking.
Even though the cap and options markets have developed considerably since then the scope for mispricing and misvaluing the book continues.
Only last year Bank of Tokyo-Mitsubishi in New York discovered that a model used by one of its options traders was out of whack with the market. A revaluation showed that the book was $50 million adrift at mid-market prices, and that turned into a further loss of $33 million when senior management in Tokyo gave orders to unwind the book. “There was no violation of rules,” says Akira Watanabe, chief executive of Tokyo-Mitsubishi International in London. “Market volatilities were changing and they reviewed the option models they were using – to be conservative.” The new valuation was done according to a new “constant elasticity volatility model”, Watanabe says.
The close-out loss of $33 million is a conundrum for risk managers and regulators. The $33 million is in theory the difference between the mid-market price and the close-out price, that is, half the bid-offer spread. Either it was a big portfolio or there was a wide bid-offer on options at strikes the market didn’t want.
Prudent banks keep a close-out reserve, although for trading purposes firms tend to mark their portfolios to market at a mid-market price. Only if they have particular directional bias do they mark at the bid or the offer. The UK Securities & Futures Authority (SFA) asks for close-out valuations on trading portfolios and “we require bid-offer marking to market as a first step,” says Vicki Fitt, executive director at the SFA.
Mid-market valuation is suitable if you assume the business is a going concern, say many practitioners, but it gives no buffer for illiquidity or forced selling. However, a firm that valued everything at close-out would do no business at all. Traders at prudent firms, if they sold at the bid and have to value at mid-market, or even at the offer, find they sometimes are forced to book a loss on a deal from day one.
What is the difference between Bank of Tokyo-Mitsubishi’s little glitch and NatWest’s? The pragmatists would say “nothing”. The only reason NatWest’s glitch became public at that time, says Owen, is because a journalist, getting wind of it, threatened to write that it was a hole of £800 million to £1 billion. Tokyo-Mitsubishi had to report its hole in year-end accounts. Although the tale appeared on Reuters and in The Wall Street Journalno witch-hunt ensued.
Terms such as “false values” and “false accounting” have been used in the NatWest case. Practitioners argue that almost every case of option misvaluation – and there have been many – boils down to the same mistake: allowing traders, directly or indirectly, to determine their own valuation.
“It’s not wrongdoing but mismanagement,” says Watanabe at Tokyo-Mitsubishi of the NatWest case. “Similar things can happen at any institution.”
Till Guldimann, executive vice-president at Infinity Financial Technology, and former chairman of the market risk committee at JP Morgan, sees the problem as an absence of “clear language to describe the engagements of the traders. If the trader is motivated differently from the institution then you need checks and balances.” But, he warns, “there will always be accidents. The ability to independently check will reduce the occurrence of accidents.” JP Morgan has had its share, notably a failure to predict the potential behaviour of certain mortgage backed securities 1992 at a cost of $50 million.
In many cases it is difficult to ascertain whether the misvaluing was due to simple shortcomings of the risk model, or to humans exploiting the shortcomings of models or their risk managers.
Why did BZW disband its exotic structured products team at the end of last year? Why were the bond option traders at ScotiaMcLeod shut down at around the same time? Were these people unlucky, greedy or incompetent? The jury is still out on Joe Jett’s extraordinary false profits on government bonds at Kidder Peabody.
When traders’ compensation is so closely linked to their year-end profit-and-loss it is extremely difficult to separate the layers of motivation – greed, professional pride, egotism, peer pressure, loyalty, fear and the hunger for power.
Risk controllers cannot possibly monitor all these things. In the end they have to take into account that, beyond their best efforts, this is an area of operational risk – the risk of rogue traders, misused models, hoodwinked financial controllers – which is part of the cost of doing business.
There is a final dilemma. How publicly should the dirty linen be washed? If the incident is not aired in public, only one firm learns the lesson. If it is given a public airing there’s a chance that the entire industry will be a little wiser. Will NatWest do the decent thing and share its experience?
| How the options writing grew and grew | ||||||
| NatWest’s traded interest rate options contracts (notional principal in £bn) | ||||||
| 1 year or less | Between 1 and 2 yrs. | Between 2 and 5 yrs. | Over 5 years | Total | Fair values | |
| End 1995 | ||||||
| Options purchased | 87.6 | 25.6 | 22.6 | 4 | 139.8 | 0.596 |
| Options written | 27.7 | 20.1 | 26.1 | 8.7 | 82.6 | -0.671 |
| End 1996 | ||||||
| Options purchased | 85.5 | 26.1 | 21.7 | 12.6 | 145.9 | 1.052 |
| Options written | 49.2 | 28.2 | 33.4 | 10.5 | 121.3 | -0.972 |
|
Source: NatWest Group annual reports |
||||||
