Researchers have argued that the cause was a policy of hedging the same quantity and the funding of cash losses on futures. Risk management and finance staff can read from the course of events how hedge design and funding are related.

What happened

MG Refining and Marketing (MGRM), the US subsidiary of the German corporate group Metallgesellschaft (MG), ran a business selling refined oil products to customers at long-term fixed prices. In December 1993, MG took large losses in this oil business and was effectively bankrupt. The 1993 derivatives losses are reported at USD 1.25 billion (about 2 billion Deutsche marks at the exchange rate of the time) in a paper from the University of Cologne (hereafter "the paper"). More than USD 800 million of this was realized (the loss was locked in) in the fourth quarter. The source of this figure is MG's special audit report (1995)1.

The figure above is the reported 1993 derivatives loss. Separately, an officer of MG's main bank (Deutsche Bank) said that the balance sheet losses (expected losses on the books) from the oil business were expected to total about 4 billion Deutsche marks. The two are different kinds of figure. Without a large rescue by a consortium of major banks, MG would have had to file for bankruptcy1.

How the trades worked and how events unfolded

MGRM sold refined oil products at long-term fixed prices, so it needed to protect itself against price changes over the delivery period. But the market had too few instruments for long-term hedging2.

In 1991, MGRM began selling long-term over-the-counter contracts (contracts made directly with a counterparty, not through an exchange) to customers. Under these contracts, it delivered refined oil products every month at a price fixed in advance, for five or ten years. The price was set from the price of futures for the next 12 months at the time of the contract, plus a markup (margin) of 5 to 10 cents per gallon3.

MGRM's risk management policy prohibited speculative trading and holding one-sided positions. It therefore hedged the quantity it had to deliver with the same quantity, using short-dated futures and over-the-counter swaps, and repeated the same trades each time they matured, rolling them forward (a stacked hedge)2. The trader in charge could choose the futures used for the hedge from gasoline, heating oil, and crude oil2. The volume of contracts grew sharply in 1993. It expanded in the summer, when the market was in contango, with futures prices higher for later delivery (far-month) than for earlier delivery (near-month)4.

Time Event
1991 MGRM began selling long-term over-the-counter contracts3
Summer 1993 to September Contract volume expanded sharply while the market was in contango, and in September long-term delivery contracts reached more than three times their level at the start of the year and stayed there until the end of the year4
December 1993 Counterparties in over-the-counter trades refused to extend the maturity of swaps5
December 1993 The New York Mercantile Exchange (NYMEX) suspended the exemption for hedging trades and lowered the position limit. MGRM had to reduce its derivatives positions5
December 1993 to February 1994 According to MG, it reduced the delivery contracts and the hedges together, down to 109 million barrels by February 19946
Same period (another view) Another view holds that it liquidated the whole hedge first. At the time of the paper, the question was not settled6
Afterward MG abandoned the policy of hedging the same quantity it had to deliver6

The basics of futures are covered in Forwards and Futures.

Views on the cause

Researchers disagree on whether MGRM's hedge was appropriate. The ratio of the size of the hedge to the quantity to be delivered is the hedge ratio, and the policy of hedging the quantity to be delivered with the same quantity of futures is a "1-to-1 hedge", with a hedge ratio of 1.

There are two views. One view is that the 1-to-1 hedge was a textbook, sound strategy that MG's supervisory board and its banks simply did not understand. The other view is that the hedge was too large for MGRM's business and increased the variability of oil prices instead of reducing it. The latter claims that the hedge ratio that minimizes price variability (the minimum-variance hedge ratio) was below 0.57.

The difference between the two views can be explained by the period over which risk is measured. Measured over a short period, the appropriate hedge ratio falls. Measured over the whole life of the contracts, the 1-to-1 hedge is effective.

Even when measuring short-term risk, a 1-to-1 hedge had the potential to reduce oil price risk substantially. In the paper's estimate, this holds if the products to be delivered are taken to be gasoline and heating oil, and the hedge uses futures on crude oil, a different commodity (a cross hedge). But a 1-to-1 policy alone does not fix which product to use, and depending on the choice, short-term risk could have been larger than with no hedge at all8.

A third view holds that the real risk was funding, not price. When oil prices fall, the hedging futures lose money, and the loss is paid in cash. The gain on the delivery contracts with customers, on the other hand, is an unrealized gain that has not become cash. Because the cash loss can be offset only by this unrealized gain, the hedge can be sustained only while enough cash is available9.

The paper considers that, because MG had strong relationships with its main banks, it should have been able to borrow as long as it had unrealized gains to serve as collateral. The greater danger, in this view, was that the 1-to-1 hedge made it very likely that temporary short-term losses would wipe out MG's equity capital (the funds the company put in itself)9.

What changed afterward

MG abandoned the 1-to-1 hedging policy after the liquidation at the end of 19936. Changes to rules or institutions that followed the case are not covered on this page.

Notes

  1. Mark Wahrenburg, "Hedging Oil Price Risk: Lessons from Metallgesellschaft" (University of Cologne), page 2, "1 Introduction", and footnote 1. The expected loss of about 4 billion marks is the statement of a Deutsche Bank officer, and the source of the USD 1.25 billion is MG's special audit report (1995), page 58. ↩ ↩2

  2. The paper, page 3. It covers the risk management policy (no speculation or one-sided positions), the stacked hedge, the choice of futures (gasoline, heating oil, crude oil), and the lack of long-term hedging instruments. ↩ ↩2 ↩3

  3. The paper, page 3, "2 The Oil Business of MGRM". Sales began in 1991, for five or ten years (a few for two years), at the futures price plus 5 to 10 cents per gallon. ↩ ↩2

  4. The paper, pages 3 and 13. Volume more than tripled by September and stayed there until the liquidation began at the end of the year; the large expansion came in the contango summer. ↩ ↩2

  5. The paper, page 4. Counterparties refused to extend swaps, and NYMEX suspended the hedge exemption and lowered the position limit. ↩ ↩2

  6. The paper, page 4. MG's account is a simultaneous reduction to 109 million barrels by February 1994. Culp and Miller claim the whole hedge was liquidated first. MG abandoned the 1-to-1 hedge after the liquidation. ↩ ↩2 ↩3 ↩4

  7. The paper, page 2. The views of Culp, Hanke, and Miller, and of Edwards and Canter, Mello and Parsons, and Ross (a minimum-variance hedge ratio below 0.5). ↩

  8. The paper, page 1 (abstract), page 4, page 15 (the estimate assumes delivery contracts for gasoline and heating oil), and pages 17 to 18. ↩

  9. The paper, page 13, footnote 31. Mello and Parsons see the real risk as funding; the paper holds that the banks would have lent and sees the greater danger in losing equity capital. ↩ ↩2