Why energy futures markets merit support amid latest controversy

Edward N. Krapels
Energy Security Analysis Inc.
Washington, D.C.

• Non-commercials' net long position vs. nymex crude prices [32170 bytes]
• Correlations between net long (short) positions of noncommercials and Nymex fuels prices [27521 bytes]

Ten days later, the Wall Street Journal reported that-in the midst of a traditional discussion about oil inventories and the weather outlook-a new topic was put on the table by one of the industry participants: the influence of speculators on the New York Mercantile Exchange (Nymex) on oil prices.

In a nutshell, the participants were told that speculators had a lot to do with the current high prices of crude oil and heating oil.

This is a sensitive topic for all involved in the modern oil market. Nymex-which has the most to lose from stricter controls over paper oil trading-quickly issued statements defending its oil markets. The Commodities Futures Trading Commission (CFTC), which obviously does not want to be seen as exercising anything less than rigorous oversight over oil, also declared that the problem was with supply and demand for oil and not with trade in the commodities market.

A lot of people-especially in New York-have a great deal invested in the continued success of the energy paper markets. The oil and natural gas business has been revolutionized by the growth of paper trading. Electricity is about to follow suit.

We at ESAI have argued that the Nymex and associated paper oil markets have done nothing less than broken the hold of OPEC over oil and in the process created a highly liquid-and highly useful-paper commodity market that is a far better arena for price formation than the Vienna hotels in which ministers meet in secret.

Paper oil markets, in other words, merit safeguarding. And it is in this vein of support for Nymex and its energy markets that we have to examine the argument about the role of speculators.

Non-commercials' influence

ESAI first called attention to the apparently growing influence of what the CFTC calls "non-commercials" (which we believe to be mostly hedge funds) in a memo we sent to our clients in 1994.

Our argument can be summarized as follows:

Paper markets affect short-term prices. The flow of investor funds into commodities, into the fuels segment of commodities, into individual fuels, and from the long to the short side of particular markets has, by virtue of its size, a very significant-sometimes decisive-effect on world crude oil and petroleum product prices.

It would be indeed implausible to believe that trade of this magnitude would not influence the price of the commodity in question. Even though many crude oil producers, refiners, and petroleum product consumers do not hedge and even though paper markets go through phases when they appear to be thin and losing trading volume and open interest, the fact remains that Nymex-traded crude oil, heating oil, natural gas, and gasoline are benchmarks whose fluctuations affect the economics of the entire energy industry as well as those buying services from that industry. Thus, non-hedgers and those uninvolved in market activities are all just as affected by the force of paper energy markets as are the speculators and hedgers that use them every day.

Effects of 'virtual' trading

Futures trading is conducted by hedgers and speculators.

Data published for all commodities by the CFTC-the Commitment of Traders (COT) report-show the relative size of the positions of hedgers and speculators in each commodity contract.

ESAI examines these data carefully because they provide useful evidence of the differences between the commodity investment programs of speculators and those of the hedgers. These differences are reflected in the ratio between their respective long and short positions.

Given the sensitivity of discussions about speculative forces in the oil market, perhaps it is best to begin the discussion about the role of speculators with an anecdote from another market. On July 18, 1995, the Financial Times published a story about the potential for a retracement in the value of the yen. It noted that the fluctuations in the value of the yen "appear largely to be the result of heavy buying by a few U.S. hedge funds, which have had an exaggerated impact on thin summer markets."¹

This article about financial markets takes for granted what many in the oil markets still do not appreciate: professional speculators' demand for and supply of paper instruments can influence the underlying price.

That influence is felt in markets much larger than those for energy. Indeed, the topic of international capital mobility and its effect on equity prices, currency values, commodity prices, and even a nation's monetary and fiscal policy is routinely discussed in thoughtful periodicals and in newspaper stories. It is also the subject of a considerable amount of research in academia, government, and multilateral organizations.

One consequence of the growth of paper market trading is to render inadequate the analysis based exclusively on physical supply and demand. The Wall Street Journal regularly quotes market analysts who are surprised by the strength of oil prices. Some of these, just as regularly, profess to see no fundamental supply/demand reason for such high oil prices. It is tempting, then, to complain that the market is irrational, when in fact it is the analysis that is inadequate. The market, in other words, is challenging us to broaden our fundamental ideas about how commoditized energy prices are created.

We have to acknowledge that perfect short-term commodity price forecasts remain an elusive goal: no one has perfected one yet. The analytical problem that short-term price forecasting poses is one of the most challenging imaginable: the international oil markets are amazingly complex; timely information about oil supply, demand, and inventories, while better than before, is still far from complete; and, perhaps most importantly, the interaction between physical and paper markets is more intricate and obscure than ever. Anyone who confronts this system with simple-minded approaches-whether chartist or fundamentalist-will soon be humbled by the fact that a perfect price-prediction theory is like the search for a unified field theory in physics.

Nevertheless, the sheer size, scope, and volatility of paper oil and gas markets make it absolutely essential to address explicitly and creatively the relationship between physical market fundamentals and paper market fundamentals. But how to start? At ESAI, we have returned to basic economics, and have attempted to sketch out the underlying relationships that influence oil prices.

We have found it extremely useful to evaluate paper oil markets with the most basic framework of economic analysis: the supply/demand/price equilibrium diagram. In the wet-barrel market, this simple framework assumes that there is a relationship between a positively sloping supply curve (higher price means more supply) and a negatively sloping demand curve (higher price means lower demand). At some price, demand and supply are in balance.

The same framework is easily applied to paper markets. The terminology is a little different: the willingness to supply is called "being short," the willingness to buy is called "being long." Paper market participants add to paper supply (go short) or demand (go long). As in physical markets, structural changes in paper markets (i.e., a change in the number of barrels buyers are willing to buy at a given price or producers are willing to sell at a given price) can be represented as movements of the supply (short) and demand (long) curves inward and outward on the traditional supply/ demand chart.

Considered in this traditional way, oil prices in the paper markets move in reaction to changes in demand and supply for paper barrels. Therefore, analysts should be interested in any forces that cause sudden shifts in the quantity of paper barrels demanded or supplied-in other words, sudden changes in the number of longs and shorts.

Getting a grip on speculators

Our continuing evaluations of the ebb and flow of paper oil demand and supply has caused us to focus particularly strongly on the activities of professional speculators.

Speculators, such as commodity funds, move in and out of the oil markets for reasons that may have nothing to do with oil-for example, because a trading program has noted an historic propensity for oil to move one way when pork bellies move another. It is possible that a sudden decline in demand for paper barrels-occurring because one or more large players suddenly decides to abandon oil as a financial instrument-causes a decline in paper oil prices that quickly reduces the value of physical barrels.

How can the market's physical players anticipate such moves? Not easily, nor can financial market forecasters anticipate them. Not all players use the same trading theories or technical systems. While looking at charts is probably necessary-ideas of resistance and support levels do tend to gain consensus for a while-it is by no means a sure guide to the behavior of speculators and large commercial operators in the paper markets.

As mentioned, the best information on oil futures market activity comes from COT data. At first glance, COT futures data tell a simple story: There is always a short for every long. As the volume and open interest change, then the long and short positions taken by commercials (hedgers), non-commercials (large-scale speculators), and small traders change from day to day and week to week. The commercials, or hedgers, have the largest share of Nymex open interest. But we believe the impetus for change in the paper market's supply/demand balance at any given time is most likely to come from the speculators, or non-commercials. We also suspect that small traders are, by and large, speculators, and that they tend to follow rather than start trends.²

As with physical market data, the COT data are imperfect at best. COT data are simply clues in a complex detective story. Why did non-commercials increase their short positions? Why did commercials and/or small traders do the opposite?

We believe that a large and sudden increase in any given market position will tend to move prices up if the increase is in demand for long positions and down if the increase is in demand for short positions. For example, assume that on a given day an extremely large speculator decides to go short. His brokers will then attempt to purchase 5,000 short contracts. All other things being equal, news of such a large increase in the number of shorts demanded drives down the price. The order for 5,000 short contracts amounts to a search for 5,000 long contracts, and in the open outcry process on the floor of the Nymex, the bid price will fall until the necessary number of longs are attracted to take the offsetting positions for the 5,000 shorts.

This process can, of course, feed on itself. If the intra-day price decline forces prices below a support level, other bears will be attracted into the market, creating further downward price pressure. Fundamentalists in the physical market can only watch and wonder what led to the price decline since nothing of consequence was indicated in the physical supply/demand equation. In truth, the initial decision on the part of the speculator to go short in the oil market may have had more to do with changes in the Nikkei or in other commodities than with anything happening in oil. Once a bearish or bullish bandwagon gets going, technical buy and sell signals can and do reinforce the trend, lending credence to the old expression: "The trend is your friend."

Non-commercials and prices

The chart on p. 22 presents the striking visual correlation between the net long (or short) position of the non-commercials on the Nymex and the price of the prompt-month Nymex crude contract.³

Visually, the correlation is strong enough to support a conjecture that sudden changes in the preferences of speculators may be among the most powerful causes of changes in the price of oil.

Causality, however, is virtually impossible to prove. All we can do is conduct statistical evaluations of the relationships between sets of numbers. That will be done shortly. For the moment, however, we will confine our discussion to the qualitative level. Is it qualitatively reasonable to argue that the non-commercials-who command only a minority share of the trading in the Nymex contract-can have such a large effect on oil prices?

The answer appears to be: Yes. It is reasonable to assume that commercials' use of futures markets is largely for hedging purposes. Thus, it is reasonable to assume that the evolution of the gross amount of commercials' positions is rather steady and reactive to oil price changes. Consequently, as prices rise, it is reasonable to believe that the crude oil producer segment of the market's commercial cohort will sell (i.e., increase their short positions) increasing quantities into the futures market. Similarly, as prices fall, it is reasonable to believe that the end-users segment of the commercial cohort will buy (i.e., increase their long positions) increasing quantities into the futures market. The non-commercials' involvement in the oil market, in contrast, is likely to be unsteady and active.

We have subjected the COT data in their "raw" form (without daily interpolations) to various statistical analyses (see table this page). The absolute results of the regression and correlation analyses are less interesting than the trend. In these statistical tests, the apparent causal relationship between the net long or short positions of the non-commercials and the prompt month price of Nymex crude has grown closer since 1991.

What could explain such a result? It is probably the result of a change in the characteristics of the non-commercials category. In 1980, the amount of money under management by professionally run futures or hedge funds was estimated to be only about $300 million. According to the publication Managed Accounts Review, it was still only about $1.4 billion by 1986. By 1990, however, the estimate was $8 billion, and by 1995, it became $22 billion.

This increase in the amount of professionally administered investor money in commodities may help explain why the correlation between changes in speculators' net long or short positions and WTI prices has improved since 1990. It is possible that during 1986-96, the swings of money into and out of oil-and from the long to the short side of oil prices-have become more intense in relation to the steady business of the commercials.

We are continuing to develop these analyses of the relationship between non-commercials and crude oil prices. For now, however, we believe there is enough evidence to include these analyses in our current methodologies for forecasting short-term crude oil prices. The most robust price hypotheses that come out of our work in this area indicate that:

• Speculator trading patterns tend to reinforce price floors and ceilings until fundamental forces move them. Whenever crude oil prices are at the top of the presently established trading range-and non-commercials have a large net long position-we would expect prices to weaken because non-commercials will sooner or later abandon their net long position. As speculators abandon their long positions, prices trend downward back to the "equilibrium" trading range of about $19-20 for Nymex crude.

• Speculators' "turnover" of trading positions is one of the primary forces in the movement of oil prices between the top and bottom of the trading range.

• Speculators-especially professional hedge fund managers-are paid to make profits, not to sit on profits. Over the course of 10 years of COT monitoring of traders' positions, we can observe a typical pattern of bull-and-bear trading (going long/going short) of two to three cycles per year.

• Speculators affect backwardation and contango. Whether speculators are long or short also affects the structure of prices. Prolonged periods of backwardation are, naturally enough, facilitated by speculators' willingness to buy the prompt market. The bigger their net long position, the more likely the market is to be in backwardation. This reinforces the oil market's natural tendency to be in backwardation, because speculators are more often net long on Nymex crude than net short.

• Finally, since 1991, speculators have had a similar and observable effect on both gasoline and heating oil prices. While the effects are not as pronounced as they are in crude oil, they are nevertheless strong enough to support the argument that the positions of non-commercials are among the primary variables affecting product prices and thus refining margins.

That in no way diminishes the significance of the key physical determinants of refining margins: the level of refining capacity, the operating rates of that capacity, the competitive conditions within which refiners operate, and so forth.

To the contrary, ESAI's research into the physical market has indicated for years that the refining industry would be in for a prolonged period of pressure on its margins.

Conclusions

Those who argue that speculators are contributing to the currently high level of oil prices and to the degree of the present unusually large backwardation in oil prices are right.

As of Oct. 22, COT data show that non-commercials have a very large net long position in all oil contracts. While we cannot read fund managers' minds, they probably had these large net long positions because U.S. inventories of heating oil had been very low.

Thus, the two "fundamentals"-the long positions of the funds in the paper market and the low inventories in the physical market-are interrelated and will continue to collectively dominate heating oil prices until winter weather submits the decisive vote on the market.

References

1. "Options Market Signals Make Market Nervous," Financial Times, July 18, 1995.

2. In the crude oil market, CFTC defines non-commercials as nonhedgers holding more than 300 contracts.

3. It is important to note that the chart uses daily interpolations of the weekly COT data. A somewhat less striking correlation exists if weekly prices are used with weekly data. In the technical correlations analyses, we will use the weekly rather than daily interpolations.

The Author