Arthur Benson returned to German conglomerate Metallgesellschaft (MG)—as its US oil trading arm MG Refining & Marketing (MGRM) looked for trading expertise—from commodity trading house Louis Dreyfus Energy (LDE), where he had been a jet fuel trader. He rejoined MG after a big win in the jet fuel market, created mainly by the volatility around the First Gulf War.
From what I have researched, Benson had quite a mixed career as a trader, including during his previous stint at MG. It is claimed that he was ‘long and wrong’—i.e., he was betting on a rising market while it fell—at LDE, until President George Bush Snr’s decision to drive Iraqi strongman Saddam Hussein’s forces from Kuwait led to a spike in jet fuel prices and extreme market backwardation in most waterborne fuel products. It is rumoured he went from a losing long jet position to a $500mn profit!
Benson was a big believer in backwardation, where the price in the spot market and front end of the forward curve is higher than in later periods. So, the spike in jet prices on his jet fuel length created big profits; he earned very well on rolling those positions down the curve.
Rolling a long position from the front to a different dated part of a curve effectively allows you to cheapen your purchasing price when a market is backwardated. As a result, it is one of the principal drivers of returns when investing in commodity indices.
Contango—where contracts further along the curve are more expensive than the front end—is the reverse situation. Rolling an index position every month means you increase the price you paid for your long positions. Contango is the enemy of long positions.
Benson was hired to market long-term fixed-rate crude, gasoline and heating oil contracts to US consumers and to manage the risks associated with such an endeavour. Ideally, his mandate was locking in profitable margins and printing money.
Locking in prices
MGRM aggressively marketed their term fixed-rate oil products to consumers across the US. As a result, by the end of 1993, they had more than 160mn bl of physical short positions spread over ten years into the future. Moreover, these shorts had been sold with hefty margins, premiums of $3–6/bl, which was a big deal in a c.$20/bl outright price market.
The US oil consumer had recently lived through the price volatility of the Gulf War. The shortages and price spikes of the two Arab embargoes of the 1970s were also still prominent in its collective folk memory. There was, therefore, interest in locking in lower prices when the market fell in the early 1990s. MG saw an opportunity to satisfy its customers’ appetites and reap enormous profits.
MGRM agreed to allow customers to cash out early if the prompt futures price rose above the deferred contract price they had agreed to, on the difference between the prices at the time multiplied by the notional number of barrels. For this added feature, MGRM would share in this upside 50/50.
However, as we shall see, it was a peculiarly structured contract and difficult to price and hedge with the day's technology. It would also compound their issues when things turned sour.
Unsurprisingly, there was little historical data about market structure in 1993, given that the oil futures market had been around for less than ten years. For most of that decade, the market was backwardated or—over 12 months when contango and backwardation occurred—backwardation profits prevailed. So Benson built what passed for his ‘hedging’ strategy for managing risk around the idea that the oil and products markets would remain predominantly in backwardation.
In the early 1990s, futures markets' liquidity extended slightly further than the first three months. Similarly, access to the over-the-counter (OTC) oil swaps markets would allow only short-tenor hedges, partly because the savvier risk managers at banks liked to hedge their risks appropriately—working on the maxim that you must hedge it completely to price it correctly.
MGRM eventually held 55mn bl of futures positions in a market that traded about half that volume on an intraday basis and 110mn bl of OTC swaps with various dealers. And they were all in the first four months—despite being supposed to offset a rateable ten years of short positions.
This created an enormous spread position! So, while Benson had matched long and short barrel for barrel in terms of net volume exposures, he had ended up with a net delta and spread risks that ultimately would cost MG $1bn+. And that was back when $1bn bought a lot more things.
Stack and roll
MGRM planned to roll this front-loaded length every month into the market's backwardation, earning a little extra on the trade while surrendering a certain amount of the hedge as physical deliveries were made, while the outright deferred shorts volume effectively reduced. The term for such a strategy used at the time was ‘stack and roll’. It may be a method to manage risks of sorts, but it does not qualify as a hedge, nor does it remotely guarantee success.
On top of the implicit spread position, extra length had to be held to effectively hedge the delta of the unusual cash-out options they had granted to their customers. As a result, people were getting paid out when the spot or prompt market prices rallied above where their future purchase prices were struck, even if the real deferred market was lower than they had initially bought.
MGRM was very long in the front and short in the term structure. Furthermore, given the respective volatilities of prompt and future prices, it did not employ techniques that might have determined the right ‘notional’ to hold at the front. Crudely speaking, if there is a correlation between two points in time and those two points have different volatilities, then a different notional of futures is required at the front of the market to hedge the shorts at the back of the market.
The way a trader would look at spread risk management today, they would conclude that MGRM sat very long. Several academic papers published afterwards argued similar cases, and one suggested that MGRM had a net 81mn bl long position, on top of a massive spread position and an ill-thought-out option product.
Did Benson understand his real risk profile? Perhaps he knew his real risk and was bullish, believing length would constantly be rolled into backwardation. However, as we will see in the third and concluding part of our series, as 1993 drew to a close, the markets were to do what they always do. And it was not pretty.
This is the first in a three-part series on the Metallgesellschaft debacle. The first part, Oil trading’s biggest bust – MG: What started to go wrong?, and the final part Oil trading’s biggest bust – MG: The death spiral and aftermath are also available to read for free on the Petroleum Economist site.