On a cold day, natural gas from storage reservoirs may supply as much to markets as gas from producing wells. The ability to store gas underground not only ensures reliable delivery during periods of heavy demand, but also allows more level production and pipeline flows throughout the year. Thus, some believe that the cost of storage should be spread over all gas delivered during a year, not just gas delivered from storage sites to end-use customers during the winter. They believe storage improves the efficiency of the overall system and lowers cost of service to all customers throughout the entire year.
Gas reservoirs are increasingly important to the industry's inventory management.1 For heating years 1989-90 through 1993-94, average monthly injections and withdrawals per storage field rose significantly from an earlier five-year time period (heating years 1982-1983 to 1986-1987). Average injections during the heating season (November-March) increased 37 percent over the mid-1980s average, while working gas levels (the amount of gas available in storage to serve markets) increased by only 9 percent. Similarly, average withdrawals during the nonheating season (April-October) increased by 47 percent, while working gas levels increased only 10 percent. Withdrawals between
heating seasons 1986-1987 through 1991-1992 also increased systematically, even accounting for differences in the weather.
Data for 1991 through 1993 also show that the industry was able to operate with declining amounts of working gas. Interestingly, in addition to saving money by reducing their stocks of gas in storage, companies were also reducing their exposure to price risk between seasons.2
In today's active and transparent cash market, the rate at which storage reservoirs can be filled and emptied is crucial. Gas commodity market prices may now vary as much between days as they once varied between seasons. Newer salt-cavern storage sites that can be filled and emptied in less than a month enable companies to take advantage of daily and weekly price movements. (Many older, depleted oil and gas producing storage reservoirs take a year to fill and empty, and thus offer little flexibility and less strategic value.) High-deliverability salt-dome storage complements an inventory management style of almost constant injecting and withdrawing of gas. The attempt to keep stock levels low between seasons becomes almost irrelevant because the entire stock is turned over several times a year. Not surprisingly, new storage construction for 1994 through 1999 plans salt-cavern facilities that represent 68 percent of withdrawal capability but only 28 percent of working gas capacity.
Taking Storage to Market
Storage is a key adjunct to the development of natural gas trading or market centers. A market center offers various transaction services, at transparent prices, that support natural gas trading within liquid markets (em that is, markets evincing a similarity in the bid and offer prices for gas, transportation, and storage rights. This similarity of bid and offer prices arises when market centers are accessible to many candidate buyers and sellers.
Nearly all the physical services available at market centers involve some form of storage: short-term gas sales, parking of gas for short periods of time, loaning of gas, and balancing or adjusting amount purchased or sold with amount taken or delivered. A salt-storage site could be a market center if it were connected to a diverse set of supply sources and individual customer withdrawals were to some extent independent. Not surprisingly, many analysts see the development of an interconnected network of market centers as the next key step in creating a "seamless" North American grid.
Fungibility of gas between market centers is in part supported by storage operations. If a large price difference occurs between market centers because of a differential change in weather, storage and market center characteristics come into play in several ways. Gas may be withdrawn from storage and released to market where price is high. Simultaneously, this storage reservoir could be replenished from storage sites where gas is relatively inexpensive. The frequency of such transactions depends on the relative constancy, and transparency of the cost of transportation and storage space between market centers and on the flexibility of storage operations.
Available liquid markets and transparent prices at market centers enable storage gas to be evaluated on a daily basis. Transparent prices from liquid markets can be used as indicators of the current value of the gas stored nearby. This fact is notable because in today's gas marketplace the price of natural gas can change abruptly from day to day in response to sudden shifts in demand. During the winter, price tends to increase when daily temperatures decline sharply, and to decline just as quickly when daily temperatures return to normal. When uncertainty about available supplies increases, prices may soar. A company with flexible gas storage could readily release gas from storage to take advantage of these high prices. The greater the deliverability, the greater the expected return. The more flexibility in moving from withdrawing to injecting, the more times the storage user can take advantage of such opportunities.
Prices, Futures, and Premiums
Data from the Henry Hub market center helps quantify the changing value of stored gas over time and indicates how this changing value can be exploited. First, the Henry Hub natural gas futures price implicitly includes cost of storage and cost of money.3 A buyer of gas could purchase a futures contract for a future delivery month and avoid the cost of buying the gas now and storing it, because in some cases the futures price might exceed the cash price at least by cost of storage and cost of money. This situation would tend to occur when there is little uncertainty about the availability of current supplies.
Second, when uncertainty rises over available supplies (during the winter, for instance), spot prices may exceed futures prices. When this occurs, the term "premium" is used to indicate that owners of gas in storage near Henry Hub could obtain a return on each MMbtu (million British thermal units) almost equal to the difference. In fact, they could sell gas from storage at the current spot price and then replace the gas with a futures contract for next-month delivery, making a guaranteed return approximately equal to the premium.4 Alternatively, when the futures price exceeds the cash price by more than cost of storage and cost of money, a holder of rights to storage could purchase, store, and deliver gas under the futures contract and again make a risk-free return. Of course, any additional costs from such transactions would need to be counted as well. Owners and operators of highly flexible storage facilities are suited to such strategic behavior, as are companies with contract rights to such facilities that allow for flexible use.
Daily spot and futures prices and the difference between them illustrate how changes in the weather and other factors influence the natural gas market (see Figure 1 on p. 35). For example, the weather in early 1994 was unusually cold, particularly around January 15, and gas demand levels rose to historically high levels in many places. As a result, spot prices rose dramatically. Futures prices rose as well, but less dramatically. Most important, the premium rose to more than 80›/MMBtu on January 19 at the peak of the cold spell. Bad weather abated somewhat after January 20th, but the premium rose to $1.12/MMBtu on February 2 when the weather was forecast to turn cold again. By February 18, as the weather returned to normal and uncertainty about availability of supplies was reduced, the premium fell to zero.
Construction and constant monitoring of a data series such as the premium enables the owner of stored gas to track changes in the value of gas in inventory over time. Followed daily, this practice supports the same type of market discipline as the daily marking to market of futures positions. It also reveals market opportunities and market conditions (see Figure 2 below). For example, the magnitude of the premium upon a return to normal weather during late 1992 and early 1993 suggests that a positive premium was often associated with having stored gas to help balance the system and to serve markets should the weather change suddenly. The value of stored gas is also revealed by the premium during the March 1993 "Storm of the Century." (But note how quickly a premium can disappear as the market responds.) Finally, the value of the premium throughout much of 1994 is consistent with the growing oversupply that caused a large reduction in cash prices during latter part of year.
Series similar to the Henry Hub series could be constructed for different market centers using the purchased-gas cost plus cost of storage and money in place of the futures price. Alternatively, if transportation costs were relatively constant or predictable between market centers, cost of storage along with cost of gas at different market centers could be used to compute a premium for alternative market centers.
Statistical analysis of the premium is also possible. For example, in January through February 1994, the unit reduction in temperature below normal values for a single day was estimated to increase the premium by $0.0185 MMBtu nationwide. Similar estimates for market centers would enable the owner of gas stored near market centers to calculate, on average, either the premium to be received or the costs that could be avoided by having gas readily available for each unit decline in the daily temperature below normal.
Management could use this information for marketing and planning objectives. Such an equation could be used to evaluate service from a contract for flexible salt storage against a contract for relatively inflexible conventional storage reservoir. Thus, the equation could be used for a variety of scenarios to gauge the expected value of all premiums the company could gain if it acquired a contract for flexible salt storage. Weather and withdrawal records from previous years could be used to develop these scenarios. Equations could also be estimated using readily available data that includes additional variables such as the estimated level of storage on a day or week relative to expected deliveries on the next day or week.
These estimates could be used for planning purposes. Such tools are likely to grow in importance as distribution companies are forced to compete with independent producers that attempt to sell more gas to end-use customers and as other energy providers attempt to capture customers in the increasingly competitive and unregulated energy marketplace. t
John H. Herbert is currently senior economist at the Energy Information Administration, U.S. Department of Energy, and adjunct professor of statistics, Virginia Polytechnic Institute and State University. He has written extensively about natural gas storage, wellhead prices, and contracting practices in the gas, cash, futures, and related derivative markets over the last five years. Mr. Herbert has written more than 50 refereed articles in
energy, economic, and statistical journals; more than 20 proceedings, articles, and reports; and a book. He has over 15 years' experience as an energy consultant in the private and public sector and regularly conducts seminars for industry and professional audiences.
1. For a discussion and presentation of the details supporting the statistics reported in this section and elsewhere in the article see Energy Information Administration, "The Value of Underground Storage in Today's Natural Gas Industry" (DOE/EIA-0591), 1995. The views expressed are those of the author and do not necessarily represent those of the Energy Information Administration, U.S. Department of Energy.
2. During 1994, however, the share of capacity filled monthly exceeded that of 1993 throughout the nonheating season, and by September it even exceeded the 1991 share. The 1994 nonheating season was the first under Order 636, and thus the first during which a significant amount of interstate storage came under the management of parties other than the pipeline companies. For the most part, these parties are local distribution companies, and it is not surprising that during this first year they may have filled, in aggregate, a higher portion of available capacity than did pipeline companies. A decline in gas prices in late summer also may have encouraged more rapid filling of storage.
3. For a discussion of this see the classic article N. Kaldor, "Speculation and Economic Stability", VOl. 7, 1939, 1-27. For a modern treatment see J. Williams, "The Economic Function of Futures Markets", Cambridge University Press, Cambridge, 1986.
4. In this analysis we use the daily cash price at the Henry Hub for a delivery this month, the futures price for delivery in the next month, and the settlement price at the close of trading on each trading day. Both the cash price and the futures price quotes are for the same day. Unlike many other commodities, the cash contract market and the futures contract market are not simultaneously trading during a delivery month. The futures contract for a delivery month terminates trading on the very day the cash market for the same delivery month beings. The cash market is open near the end of the month preceding the delivery month and throughout the delivery month, whereas the futures contract never trades in the delivery month. Both the cash price and the futures contract are for the same trading day, except that delivery takes place in month t or t + 1 for the cash market and in month t+1 exclusively for the futures market.
Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.