Hedging with Futures Contracts 2


Oil Markets: Metallgesellschaft

Metallgesellschaft highlights a poor implementation of short-term futures contracts to hedge long-term oil commitments. The strategies may be analysed from a wider perspective as to the suitability of futures as a tool in reducing an underlying price risk across markets as a whole. Expert opinion remains mixed over the specific techniques employed by MGRM. It is fair to conclude, however, that while the actual approach may have incorporated greater risk than publicly acknowledged, the case does serve to illustrate the inherent benefits that futures may bring to the management of sophisticated risk management positions over a pre-defined time horizon.


In 1993 Metallgesellschaft AG (MG) was one of Germany’s largest conglomerates with interests in mining, metals and engineering. In December of that year Metallgesellschaft revealed that a US oil trading subsidiary called MG Refining and Marketing Inc. (MGRM), within its energy division, had declared losses of approximately USD1.5 billion on futures contracts. This lead to Metallgesellschaft’s near bankruptcy and the firing of its chief executive. There was subsequently a comprehensive reorganisation of the company resulting in the sale of a large number of its assets. The aim of this analysis is to investigate what led to this massive loss, how it could have been avoided, who was responsible, and general conclusions as to the suitability of futures contracts as a means of risk-reduction.

In 1989, as part of Metallgesellschaft’s plan to expand in to the North American market, MGRM acquired a 49 percent interest in Castle Energy. Castle Energy was a US oil exploration firm which became an oil refinery. They subsequently entered into long-term agreements with Castle Energy to buy all its production which provided MGRM with a guaranteed long-term supply of refined oil products. MGRM then entered into long-term fixed price contracts with customers that stipulated that MGRM receive fixed-rate prices every month for five or ten year periods. The price at which the refined oil products were sold was above the spot price at which they traded. However, in return for this premium the customer received protection against price rises and locked in historically low oil prices.

Fixed-Price Commitments

There were two main types of contract, most of which had a ten year maturity. The first required the buyer to buy a fixed amount of product per month at a fixed price. The second gave the buyer more of a say in when they would take delivery, as the customer could request 20 percent of its contracted volume for any one year with 45 days notice.

In signing these contracts, MGRM faced the risk of its customers defaulting if the spot price fell by a large amount. In order to limit this risk, MGRM only allowed buyers to take fixed-price contracts for a maximum of 20 percent of their total purchases. However, this meant that the customers could only be 20 percent protected against a surge in oil prices. In order to compensate for this, MGRM included a ‘cash-in’ option clause in the contracts that would let customers terminate the contracts if the future reached a predetermined exit level. Many contracts automatically terminated if the front month New York Mercantile Exchange (NYMEX) futures price reached this exit level. The first type of option stipulated that half the difference between the future price and the fixed price multiplied by the total volume remaining to be delivered, be paid by MG to its customers. The second contract contained a similar option clause, this time giving the customer the full difference multiplied by the total volume remaining to be delivered.

The contracts became so popular that by September 1993, MGRM were short of 102 million barrels of the first type of contract and around 50 million of the second. The agreements amounted to swap positions which left MGRM with a large price risk in crude, gasoline and heating oil and especially exposed to fluctuations in the crude price. Since MGRM received fixed prices, a rise in the price of crude would result in an extremely heavy loss.

Possible Hedging Strategies

As a result of the exposure, MGRM needed to hedge this price risk on oil. There were two ways of doing this. Both involved going long of futures or forwards with each strategy having positive and negative aspects. The futures/forwards acted as a hedge because the loss incurred on the fixed-rate contracts due to a rise in oil prices would be offset by a gain in the futures/forward position. On the other hand, any loss in the futures/forward position due to a fall in the oil price would be offset by gains on the fixed-rate side.

The first method involved buying a future for each forward contract with the same term structure. However as the contracts were between five and ten years long, futures with five and ten year maturities were needed. Unfortunately neither the NYMEX nor the International Petroleum Exchange (IPE) offered futures with such long maturities. This meant that the contracts would have to be hedged with over-the-counter (OTC) or tailored forwards. This was a very illiquid market and so it would have resulted in very high costs.

The second method involved a ‘stack and roll’ strategy, which meant hedging the contracts by going long of exchange traded futures that were readily available. This created a term structure problem as short-dated futures were being used to hedge long-dated short positions. Consequently a one-to-one hedge ratio could not be maintained, as the present value of the forward liability was less than the equivalent notional futures amount due to the discount factor. Another factor that needs to be taken into account when hedging with short-dated futures is that they must be rolled over each month into new futures. This can result in profits or losses depending on whether the future is trading below or above the spot price. However, spreading the positions in a wide variety of contract maturities can alleviate this exposure.

Strategy Chosen

MGRM chose the latter method but placed the entire hedge in front-end maturities. Despite obvious problems, Culp and Miller (1994) conclude that the approach was essentially the right one and did effectively lower the risk of oil price changes. However, in doing this, not only did they ignore the term structure discrepancies, but they knowingly took on the rollover risk. As the market was initially in backwardation, meaning that the spot price was higher than the futures price, this resulted in rollover profits that provided an additional source of income from the trading operation.

They purchased short-term futures on a one-to-one hedge ratio with their long-term forward liabilities even though a ten-year forward’s present value is worth far less than its notional amount. This meant that a large proportion of their hedge was not a hedge at all, but a speculative position. In a paper by René M. Stulz (1996) it is argued that out of a reported 154 million barrel futures position, only 86 million acted as a hedge with the rest (68 million) forming a speculative position. This view is supported by many experts, including Mello and Parsons (1995) who argue that instead of reducing its oil price risk, MGRM actually increased risk by using a grossly oversized hedge position.

As a result of the risks described, a substantial amount of capital was needed. A sustained series of drops would result in margin calls requiring the allocation of this capital. Prudent risk management should have considered allocating additional capital to cover the rollover risk. Even though a large amount of capital was eventually committed, this was many times more than what was considered necessary at the time. In March 1993 the price of crude oil started to fall as OPEC ministers could not agree on production quotas. This lead to losses on the futures positions and NYMEX predictably started making margin calls on MGRM’s positions. It has been reported that MGRM failed to persuade NYMEX that these calls were not needed because of the gains on their fixed-price contracts. Large amounts of capital were used up in increasing the margin amounts, which eventually resulted in a cash drain. The inability to cope with these short-term negative cash flows resulted in a funding crisis.

Market Reversal

As stated earlier, the hedging strategy relied on the NYMEX crude futures (West Texas Intermediate contracts, WTI) being in backwardation. However for much of 1993, there was a market reversal and futures prices were in contango. In other words, the spot price had now fallen below the futures price. Contango could initially have been due to factors that did not have to do with MGRM, such as the steady fall in the spot price resulting from disagreements within OPEC. Many believe, ironically, that as MGRM’s futures position increased, hedge funds and traders in general began to anticipate the need to roll over and started aggressively trading the basis between the futures and spot price. This created a distortion of the usual supply-demand balance, and lot of upward pressure on crude futures contributing to the contango effect.

Once the market, or more specifically the futures price curves, went into contango, all the rollover profits that MGRM had been making now turned into rollover losses. As the market continued to move downwards, these losses become much more severe. Unlike the losses due to the decrease in spot, these losses were not recoverable.

Metallgesellschaft’s problems were exacerbated by German accounting practises1). In Germany, the ‘Lower of Cost’ or ‘Market’ method was used, which meant that positions that hedged each other such as the swaps and futures positions could not be netted out. Any gains on the swap positions could only be first realised when contracts matured, which meant that these profits would first appear five or ten years down the line depending on the contract in question. On the other hand, the short-term negative cashflows that were being suffered from the futures positions had to be accounted for immediately. This had a devastating effect when Metallgesellschaft had to report earnings.


One aspect of this case that still has not been put to rest is the question of where the responsibility for the losses lies. In order to establish blame, one must first decide whether MGRM exclusively hedged the company’s risk or whether they also engaged in speculation. They tried to structure a novel program in order to assume rapid expansion in an evolving business, but the decision focused on a combination of OTC and futures with an element of speculation on the near and far-dated price connection. With a three-tier hierarchal structure descending from Deutsche Bank et al, through MGRM management to Arthur Benson, the Chief Trading Strategist, it follows that all were culpable. Benson, who initiated the tactics but as a trader must have remained fully aware of the risks, may certainly be held directly responsible for much of the losses. It also seems logical to conclude that Metallgesellschaft’s management were not fully aware of the risks that MGRM were exposed to. If they were, they are responsible for approving all-out gambling. If they were not, they must take the blame for not ensuring that sufficient risk management controls were in place. Metallgesellschaft’s management must also be held at least partially accountable as the company’s financial health was ultimately their responsibility.

What is more debateable is whether the group of institutional investors that controlled Metallgesellschaft AG must share any blame. When they were told of the losses, they immediately closed MGRM’s positions. Many traders argue that this was a rash over-reaction and that they should have been unwound in a more piece-meal fashion. Ironically, the futures positions were unwound just when the spot price reached one of its lowest points and later went on to recover. Deutsche Bank, one of the leading investors within the institutional group together with many industry analysts counter that waiting would have amounted to continued gambling, in the hope of an uncertain recovery. They further argue that MGRM positions had grown so great that they were themselves a significant cause of the rollover losses. Hence, waiting for backwardation to return would have been in vain and only resulted in continued rollover losses.


The losses were due to excessive levels of risk being taken on, some knowingly and some not. MGRM were right to hedge with futures contracts as this was the only financially realistic way of limiting the risk inherent in the fixed-price contracts. However, the way Arthur Benson and his team went about doing this was itself risky. For one, the size of the hedge was far too large as it did not take into account the discrepancies in the forward term structure. They intentionally put their entire hedge in front month futures as they were betting on backwardation continuing. As Culp and Miller quote from a Metallgesellschaft AG document, “At any given point in time, certain parts of the commodity market may be overvalued relative to that commodity’s own forward price curve, to other commodities, or to other markets…That, in turn, provides attractive opportunities from an arbitrage standpoint.”

There are many powerful arguments, not detailed here, that argue in favour of backwardation taking place during normal market conditions. But it was not MGRM’s job to bet on the validity of those arguments, and likely that the traders were at the very least aware of the rollover risk. With significant horizon-mismatches of the futures, the strategy evolved from simple risk-management into an overly complicated program. An alternative view on an implied time horizon over which risk is measured is likely to result in differing conclusions. Wahrenburg (1995) shows that hedgers with short time horizons tend to use a small hedge ratio of approximately one third. His analysis also indicates that the one-to-one hedge strategy undertaken by Metallgesellschaft was effective, but only when risk is measured over the total length of contract. The task was of eliminating as much risk exposure as possible was clearly not heeded.

Finally, MGRM drastically underestimated the funding capital that was needed. This is surely because they did not appreciate the risk that they knowingly took on could result in such heavy losses. By giving their customers the option to terminate, MGRM of course took the opposing short position. Described as an ‘embedded liquidity option’ by Ed Krapels (2001), this thereby forced MGRM to hedge by going long on short-dated futures contracts and utilising the rollover approach. They also failed to appreciate the effect that their futures position had on the market. It was more than big enough to be noticed, but not nearly enough to control an inevitable reaction as traders and hedge funds took advantage of this fact. As such, normal market conditions did not apply.

From a corporate perspective, this case emphasises that strict trading controls are necessary and that responsibility cannot be delegated with the assumption that traders will make prudent decisions in company’s best interests. It also seems evident that the management were not fully aware of the risks Arthur Benson was taking. The fact that the futures positions were unwound in such an abrupt manner seems indicative of human nature, where one is irrationally less tolerant of a loss after sustaining a large loss. In conclusion, had MGRM had a smaller hedge more reflective of their fixed-price exposures and had a realistic funding strategy, then this derivative disaster would never have happened.

Hedging crude oil futures contracts - discussion

In the absence of specific data from the period, the discussion of hedging to manage exposures to crude oil broadly refers to relationships between futures prices and maturities.

Backwardation or Contango

If the difference between the near dates and far-dates contracts on a regular basis during a period is positive (F0,n – F0,d > 0) the market is in backwardation. If the difference is negative (F0,n – F0,d < 0) the market is in contango. backwardation happens when the current futures price for the nearby contract is below spot price, but above the far dated contract (F0,d < F0,n < S0). At such times, a risk premium is incorporated in the futures price; leading to the current futures price underestimating the expected future spot price. One explanation for backwardation in the oil market is the convenience yield which is the additional return earned by holding a commodity in short supply. If the commodity is in short supply, such as oil, current consumption is higher relative to supply of goods. This could lead to a very high spot price. Under these market conditions investors will not be storing the commodity. As firms may need physical oil on hand to avoid inventory stock-outs, spot prices will rise above futures price to reward firms for lending their inventory to the current spot market.

Normal backwardation, contrasted with backwardation, refers to when hedgers have similar expectations which will lead them to make contracts with speculators instead of each other. The speculators will demand a return for the risk that the hedgers pass onto them and the hedgers will be willing to pay a premium to transfer the risk. Hence, the futures price will be lower than the expected spot price at expiration, to compensate speculators for their risk. Thus the future price will be expected to rise over the life of a contract.

With contango the current futures price is above expected future spot price as the current futures price overestimates the expected future spot price (F0,d > F0,n > S0). Normal contango refers to when hedgers are net long in futures contracts; the futures price will be higher than the expected spot price to compensate speculators for the risk of selling short and buying back later.

Cumulative loss of rolling over the hedge

The cost of rolling over a hedge relates to subtracting the price for the last trading day for the nearby contract from the cost of opening up the position in the distant contract, that over a period will amount to a cumulative sum. The implication is that it’s a lot more expensive to roll over a hedge when the market is in contango since the price of the futures contract rises over the period.

Relatively fewer margin calls when in backwardation

Initial margin is cash or collateral posted at clearing house to guarantee that traders fulfil their obligations, and is a function of the futures price each month. The margin still belongs to the trader but may be seized if the trader defaults on their obligation. Mark-to-market is the daily monitoring of futures price in relation to the initial margin. If the market moves against the trader they will be required to post more margin. If the futures price rises above the initial margin the trader can withdraw any excess cash. Maintenance margin is the lower limit at which a trader must post more cash or collateral otherwise the position is closed out.

When the market is in backwardation the number of margin calls will be fewer than under contango, due to the downward bias of the spot price leading to an increase in the futures price as time approaches maturity (positive return). Margin calls do happen in backwardation, due to the relative high/low price at which the position is opened. Thus, as the futures price rises as maturity comes closer, the possibility of a margin call decreases. During this market phase hedgers have a net short position and the speculators are net long of the futures contract. When the market is in contango there is an upward bias of the spot price. This leads to a decrease in the futures price as maturity comes closer (negative return profile) and one might expect relatively more margin calls then in backwardation. The initial futures price also plays its part, for example in a month where a position is opened at a very low price. Overall, however, one can expect more margin calls in contango than in backwardation.

The market condition during MGRM's hedging was clearly contango. MGRM clearly did not expect oil prices to fall to such an extent. Since oil markets were typically not in contango, because the convenience yield outweighs the cost of carry, MGRM probably did not expect the market to move into contango. Losses in cumulative rollover costs can be monumental if holding a belief that the market will stay in backwardation for the whole period.

Hedging crude oil futures contracts - Simplified example

A small oil company has signed a deal agreeing to sell 10,000 barrels of WTI oil for delivery three months from now. The spot price is USD 89.09 /barrel and the WTI futures price for delivery in three months is USD 89.51 /barrel. The sale price has been agreed by both parties to be the spot price in three months time, so the risk to the seller of oil is that oil prices may fall over the three month period until delivery is due (that they will receive fewer USD for their oil), and so they plan to use oil futures to reduce the risk. (The example ignores any issues of basis risk, or changes in the difference between spot and futures prices).

  • One WTI oil futures contract = 1,000 barrels of oil
  • S = Spot, F = Futures


S0 (spot price today) = USD 89.09 /barrel (USD 890,900)

S1 (spot price in three months) = USD 92.44 /barrel (USD 924,400)

Dollar gain on unhedged position = (USD 92.44 – USD 89.09) x 10,000 barrels = USD 33,500

The oil price rose over the period so remaining completely unhedged against oil price risk over this period would earn USD 33,500. The price could also have fallen for an equivalent loss.


With a long oil position, the firm wants to short the futures contracts (short hedge) to lock in a selling price. As they are able to accurately match their underlying spot oil position with a whole number of oil futures contracts, the profit/loss in the futures will offset any loss/profit in the underlying WTI oil spot price.

Today: Open (short) 10 futures contracts at F0 = USD89.51 (10,000 barrels / 1,000 barrels per contract = 10 contracts)

S0 (spot price today) = USD 89.09 /barrel

In three months: Close out (buy back) positions with loss in futures at F1 = USD 92.44/barrel (the oil futures price will have converged with the oil spot price over the three months to delivery date so that F1 = S1)

Loss in futures = Nf * (F0 - F1) * zf = 10 * (89.51 – 92.44) * 1,000 = (USD 29,300)

S1 (spot price in three months) = USD 92.44 /barrel Spot dollar receipt = S1 * (FVS0) = 92.44 * 10,000 = USD 924,400

Total receipt = (USD 924,400 – USD 33,500) = USD 895,100

This translates to an effective price of USD 89.51 /barrel (USD 895,100 receipt / 10,000 barrels), equal to the futures price F0. With a long WTI oil position of 10,000 barrels translating to USD 890,900 on the day of agreement, the total receipt demonstrates the hedging approach and outcome. Although the spot price rose and they are able to sell their barrels at a higher price, the short futures contracts position was opened at a lower price, and so the gain on the sale will be offset by the loss on the futures. This equates to the company locking in a price when the agreement is made at the corresponding futures price of USD 89.51 by hedging the sale. If the spot price had fallen instead of rising, a lower sale price would be offset by a gain on the futures.


Culp, C.L. and Miller, M.H. (1995) ‘Metallgesellschaft and the Economics of Synthetic Storage’, Journal of Applied Corporate Finance Vol. 7, No. 4

Kimbull, R.C. Failures in Risk Management

Kolb, R (1997) Understanding Futures Markets, 5th edition

Krapels, E (2001) Special Report – Re-examining the Metallgesellschaft affair and its implications for oil traders, Oil & Gas Journal March 26, 2001

Mello, A.S. and Parsons, J.E. (1995) Maturity Structure of a Hedge Matters: Lessons from the Metallgesellschaft Debacle, Journal of Applied Corporate Finance, 7.

Wahrenburg, M (1995) Hedging Oil Price Risk: Lessons from Metallgesellschaft

Finance, Derivatives

Metallgesellschaft AG: A Case Study- By John Digenan, Dan Felson, Robert Kelly and Ann Wiemert

QR Code
QR Code futures_contracts_2 (generated for current page)