4. Mid-Term Trends in Natural Gas Supply and Prices: Potential for Cyclic Price and Investment Behavior

Introduction

This chapter explores the issue of whether the domestic natural gas market can be expected to exhibit a new behavior pattern with respect to long-term wellhead natural gas supply, and the potential market implications of such behavior. In particular, did the recent rise and fall of natural gas prices signal a fundamental change in the long-term pattern of investment, drilling, wellhead supply, and prices in U.S. natural gas markets?

Commodity prices can exhibit four distinct patterns over time:

• A commodity’s price behavior can exhibit a long-term trend. For example, from 1993 through 1999 (Figure 19), monthly wellhead natural gas prices were relatively constant, averaging $2.00 per thousand cubic feet (in nominal dollars) with a standard deviation of $0.36 per thousand cubic feet. Although wellhead prices varied over the period, $2.00 per thousand cubic feet can be viewed as the long-term price trend.

• A commodity’s price can exhibit a seasonal pattern. Natural gas has demonstrated over the years a seasonal pricing pattern, with winter prices typically higher than summer prices due to greater winter consumption levels.

• A commodity’s price can reflect the influence of random events, such as oil embargos, pipeline ruptures, hurricanes, and abnormally cold or hot temperatures. Natural gas prices certainly have been influenced by such random events. For example, during a 2-week period in February 1996, Henry Hub spot prices went above $5.00 per thousand cubic feet because of unseasonably cold temperatures. The severe weather was short-lived, and spot prices quickly declined when temperatures rose.

• A commodity’s price can exhibit a periodic cycle of high and low prices over several years, reflecting a business investment cycle. To date, the natural gas supply industry has not conclusively demonstrated such cyclic price behavior. However, the run-up in wellhead prices during 2000 and their subsequent decline during 2001 raises a question as to whether the natural gas supply industry might be entering a new era of cyclic price fluctuations typical of a commodity business cycle.

Industry Restructuring and Market Price Behavior

Over the past 15 years there has been a significant transformation both in natural gas supply and natural gas consumption. Wellhead price deregulation has permitted natural gas prices to adjust freely to prevailing supply and demand conditions, and open-access transportation has allowed natural gas volumes to move freely from producers to consumers. As a result of industry restructuring, natural gas supply, demand, and prices are now subject to competitive market forces, which are largely responsible for recent efficiency gains in natural gas production.

Natural gas industry restructuring has heightened supply competition and caused natural gas suppliers to minimize per-unit fixed and operating costs by fully utilizing productive assets, in addition to using new technologies and management techniques to improve finding rates and decrease drilling and operating costs (see discussion on "Recent Efficiency Imrovements in the Natural Gas Production Industry"). In 1985, natural gas productive capacity was operating well below full utilization (Figure 20). Between 1985 and 1993, seasonal consumption variations were partly met by variable wellhead production. By 1993, capacity had come close to full effective utilization. (Full effective utilization appears to be in the range of 93 to 95 percent of total productive capacity. The seasonal variation in capacity utilization reflects seasonal variation in natural gas consumption.)

Since 1993, capacity utilization has continued to increase slightly. Probably the highest possible operationally feasible utilization level was achieved in late 2000. Near full capacity utilization, wellhead natural gas production is less responsive to short-term consumption variations. As a result, wellhead natural gas production has become less elastic with respect to price and  consumption in the short-term, as demonstrated by the decline in the monthly production variation since 1985 (Figure 21).

Competitive natural gas production markets have kept wellhead prices sufficiently low to encourage significant growth in natural gas consumption. Since 1983, when natural gas consumption bottomed out, it has increased by 35 percent, from 16.8 trillion cubic feet in 1983 to 22.7 trillion cubic feet in 2000.⁸⁵ Of the major natural gas consumption sectors, electricity generation has posted the largest growth, from 274 billion kilowatthours of electricity produced from natural gas in 1983 to 596 billion kilowatthours in 2000.⁸⁶ Natural gas was responsible for 16 percent of total electricity generation in 2000, compared with 12 percent in 1983. Commensurately, electricity production consumed 28 percent of total natural gas supply in 2000, compared with 17 percent in 1983.⁸⁷ Continuing growth in natural-gas-fired electricity generation could have profound implications for future natural gas markets.

The four largest energy sources for electricity generation are coal, nuclear, hydropower, and natural gas. Because natural-gas-fired generators have the highest fuel and operating costs, they typically are dispatched last to the transmission grid and provide mostly peaking load power and some intermediate load power.⁸⁸ Peak and intermediate electricity requirements are determined largely by variations in demand for summer cooling and winter heating. Because of the growth in natural-gas-fired electricity generation  capacity, natural gas consumption has become more sensitive to weather variations.⁸⁹

Natural Gas Industry Attributes That Could Facilitate Future Investment and Price Cycles

The natural gas production industry possesses the causal attributes necessary for business cycle behavior:

• Relatively inelastic supply and demand in the short term, which can cause large fluctuations in price during periods of relative scarcity or abundance of supply

• Large fluctuations in producer cash flows and investments, leading to large fluctuations in wellhead gas supplies, as a result of large price fluctuations

• Significant delays (approximately 6 to 18 months) between changes in price and changes in wellhead gas supply, which encourage overinvestment when prices are high and underinvestment when prices are low, relative to gas demand

• Rapid declines in production from new natural gas wells, which could rapidly turn a supply surplus into a deficit during a period of low producer investment.

Short-Term Inelasticity of Natural Gas Supply and Demand

From early 2000 through mid-2001, a scarcity of available natural gas supplies led to sustained high wellhead prices for natural gas for the first time since the early 1980s. This event was unique for a number of reasons. First, it occurred in a fully competitive market environment in which natural gas producers were operating their productive assets at full utilization. In contrast, the high wellhead natural gas prices of the early 1980s were caused by market distortions imposed by the regulation

of wellhead natural gas prices, when producers were operating at well below full capacity utilization. Second, the volume of weather-sensitive natural gas consumption has grown substantially, primarily due to the growth in natural-gas-fired electricity generation capacity. High wellhead utilization in conjunction with a colder winter caused high and sustained natural gas prices. In addition to other factors, such as a low level of Northwest hydroelectric capacity, the sustained high natural gas prices reflect the relative inflexibility of short-term natural gas production, because natural gas drilling—and ultimately natural gas production—could not immediately respond to the higher prices.

Like wellhead natural gas supplies, other sources of natural gas supply were also relatively inelastic. For example, while the volume of weather-sensitive natural gas consumption has grown, the capability of natural gas storage facilities to reduce high prices during periods of high winter demand appears to have diminished. In 1988, total working gas storage capacity equaled 4,324 billion cubic feet. By 2000, total working gas storage capacity had declined to 3,962 billion cubic feet.⁹⁰ The decline was offset to some extent by growth in capacity at salt cavern storage fields, which are more suited to supplying large volumes of natural gas over short periods of time. In 2000, salt cavern storage working gas capacity totaled 75 billion cubic feet.⁹¹ Moreover, the maximum withdrawal rates of some reservoir storage fields have been increased over the past decade through additional wells for withdrawing natural gas. Nevertheless, the volume of weather-sensitive natural gas demand has grown significantly while the capability of natural gas storage facilities has, at best, remained constant. As a result, natural gas supply from storage facilities would be expected to have become a less elastic natural gas supply source relative to the magnitude of potential swings in natural gas consumption.⁹²

Another natural gas industry attribute that contributes to the inelasticity of natural gas supplies is the fact that the U.S. natural gas market, although tightly integrated with the Canadian natural gas market, is relatively isolated from overseas natural gas supplies. In contrast, the well integrated international oil trade moves crude oil and petroleum products to or from the United States when there are relatively small price differentials between the U.S. and the overseas markets. These trade movements are feasible because the cost of bulk shipping is inexpensive and because there is considerable spare tanker capacity on the world market. Consequently, any relative scarcity or abundance of petroleum in the U.S. market can be moderated by shipments of petroleum to or from overseas markets. Given the limited U.S. infrastructure for imports of liquefied natural gas (LNG) and the limited extent of international LNG trade, the isolation of the U.S. and Canadian natural gas markets from overseas markets limits the degree to which spot LNG shipments can moderate price extremes.

In 1999, LNG imports accounted for 163 billion cubic feet or 0.7 percent of total U.S. and Canadian natural gas consumption.⁹³ Although 2000 LNG imports increased to 226 billion cubic feet, this still amounted to less than 1 percent of total U.S. and Canadian natural gas consumption. Two U.S. LNG terminals are expected to be reactivated soon, increasing sustainable U.S. import capacity to 718 billion cubic feet per year, or 3 percent of total U.S. and Canadian natural gas consumption, it is unclear whether the facilities will be delivering any LNG on a spot cargo basis.

If spot LNG cargos were feasible, then LNG could be a swing source of natural gas supplies by providing additional supplies during periods of high natural gas prices and by curtailing LNG shipments during periods of low natural gas prices. The feasibility of spot LNG deliveries depends on such issues as the availability of LNG tankers, whether current contract commitments at U.S. LNG terminals would preclude spot LNG deliveries, and the availability of spare overseas liquefaction capacity to load spot cargos. Currently, LNG is not a large enough portion of the total U.S. and Canadian market to alleviate extreme price fluctuations.

Natural gas consumption is also relatively inelastic to prices in the short term, because natural gas consumption equipment typically has lifetimes in excess of 15 years. The short-term inelasticity of natural gas demand increases the probability that natural gas supplies could be either comparatively abundant or scarce relative to prevailing natural gas consumption requirements. This natural gas market attribute of relative supply abundance or scarcity can lead to wide swings in natural gas prices, setting the stage for the possibility of surfeits or deficits of investment in new natural gas well productive capacity relative to potential natural gas consumption.

Price adjustments ultimately succeed in balancing supply and demand in the U.S. natural gas market, because neither demand nor supply is completely rigid or fixed in the short term. For example, natural gas producers can place skid-mounted compressors in the field to withdraw natural gas more rapidly from the reservoir. Or they can decide not to remove the natural gas liquids (ethane, propane, butane) from gross wellhead volumes, so that they can deliver larger natural gas volumes when the value of the natural gas exceeds the value of the entrained liquids. On the demand side, some industrial and electricity generation consumers can switch dual-fired equipment from natural gas to oil. Other industrial consumers can sell the higher value natural gas rather than use it as a feedstock to produce a lower value product (e.g., ammonium hydrate). The responsiveness of supply and demand ultimately depends on the cost or value of implementing a specific supply or consumption response, relative to the prevailing natural gas price, and the time delay associated with implementing that response. The extremely high and sustained natural gas prices of late 2000 and early 2001 indicate, however, that the costs could be higher and the delays longer than might have been thought to be the case.

Dependence of New Productive Capacity on Producer Cash Flows and Prices

Exploration and development activities for new natural gas wells are dependent on production cash flow for new investment capital. Cash flow equals revenues (price times production volumes) minus costs. Recent history has demonstrated that when prices are high, cash flows are also high, inducing investment and drilling (Figure 22).⁹⁴ Natural gas drilling activity, in turn, is directly related to the development of new productive capacity. A comparison of new natural gas discoveries and annual natural gas drilling rig levels shows that new natural gas discoveries are correlated with drilling rig rates (Figure 23).⁹⁵ Higher gas rig levels generally result in a higher level of natural gas discoveries. For example, during 2000, the high average gas rig count of 720 resulted in 19,138 billion cubic feet of new natural gas discoveries.

Given these relationships, high natural gas prices will result in high levels of new natural gas productive capacity, and low natural gas prices will result in low levels of new natural gas productive capacity. The industry potential for extreme price swings raises the question as to whether the level of new natural gas productive capacity will tend to overshoot or undershoot the level of supply necessary to match natural gas demand requirements. The inherent time delays between price level changes and wellhead supply changes suggest that this will be the case.

Delays Between Price Changes and Drilling Investments

New wellhead natural gas supplies will increase or decrease as wellhead prices increase or decrease, but the response is not immediate. Depending upon circumstances (e.g., onshore versus offshore production), the delay may range between 6 and 18 months. Because natural gas producers want some assurance that wellhead prices will be high enough for long enough to justify new investment, there is some tendency for producers to defer higher drilling investment levels until they are convinced that the higher prices do not represent a short-lived price spike.

In addition, there is a delay between the time when natural gas producers decide to increase their drilling budgets and when the new wells are completed. This delay encompasses contracting for the drilling, transporting and assembling drilling rigs on leases, drilling the wells, cementing the wells, fracturing the formations, and attaching the wells to gathering systems. If drilling rigs are in short supply, there will be an additional wait until a rig becomes available or a new rig is manufactured.

Wellhead prices will remain at high levels until sufficient new productive capacity is brought into operation, relative to demand, to cause wellhead prices to decline. Then, there will be another delay between the onset of the decline in prices and a reduction in drilling activity by producers. For example, average wellhead prices peaked in January 2001 at $8.06 per thousand cubic feet and fell to $3.39 per thousand cubic feet by July 2001, but the gas rig count did not peak until July 13, 2001, at 1,068 gas rigs, and had only declined to 876 rigs by October 26, 2001 (Figure 24). A key impediment to rapid reduction of drilling activity is that drilling contracts typically specify a minimum level of drilling activity by the service company over the life of a contract.

The recent natural gas price and drilling rig situation suggests that the time delay between natural gas price changes and wellhead natural gas supply changes could cause such an abundance of wellhead productive capacity that natural gas prices will become quite low in 2002. In 2001, the average annual gas rig count through October 26, 2001, was 969 rigs. If the rig count were to drop by 33 to 50 gas rigs per week for the remaining 9 weeks of 2001, then the year-end gas rig count would be 579 to 426, and the average annual rig count would be 924 to 910. At an annual average count of 910 to 924 rigs, 2001 natural gas discoveries could be in the range of 22 to 24 trillion cubic feet.⁹⁶ Yet, the average annual rate of new natural gas discoveries from 1995 through 1999 equaled 12.2 trillion cubic feet per year.⁹⁷ With new wellhead natural gas discoveries in 2000 replacing 99.6 percent of that year’s natural gas production,⁹⁸ and with the prospect for even higher reserve discoveries in 2001, the large additions to wellhead natural gas supply during 2000 and 2001 create the potential for a significant decline (“bust”) in wellhead natural gas prices. Whether or not a wellhead price bust will materialize should become apparent by the spring of 2002, at the conclusion of the winter heating season.

The delay between changes in price and changes in new wellhead supplies increases the propensity of natural gas producers to overinvest in new productive capacity during periods of high wellhead prices. During the “boom” phase of a business cycle, producers will continue to invest at high levels as long as wellhead prices remain high, even though there could be sufficient investment already underway to ensure adequate natural gas supplies. When wellhead prices begin to fall, the delay in cutting back drilling activity ensures that additional supplies will come into the market even after their economic justification has apparently disappeared. The resulting overabundance of wellhead natural gas supplies will then induce a precipitous drop in prices.

If natural gas prices fall below the long-run marginal price necessary to maintain adequate wellhead natural gas supplies, producers will be induced to underinvest in new wellhead natural gas capacity; and given the inherent time delays in the price and investment cycle, the underinvestment in new natural gas productive capacity will continue well after natural gas prices begin to rise, signaling the disappearance of surplus natural gas productive capacity. A period of extended underinvestment could rapidly turn an abundance of wellhead natural gas capacity into scarcity of supply.

More Rapid Declines in the Production Rate From New Natural Gas Wells

In recent years, production from new natural gas wells has been declining more rapidly than in the past. This phenomenon is best illustrated through the concept of a production half-life. The production half-life represents the amount of time that passes before a well (or a group of wells) produces natural gas at 50 percent of its (their) initial production level. This figure can be calculated for all wells that began producing in a given year (e.g., vintage 1990 wells), allowing a comparison across years (Figure 25). Natural gas well production rates were declining much faster at the end of the 1990s than they were at the beginning of the decade. Although there is some year-to-year variation in the trend, the 1990 and 1999 data illustrate the dramatic change in gas well half-lives⁹⁹ (Table 12).

Two trends are apparent from the data in Table 12. First, the average gas well half-life has dropped for all major production regions and for the lower 48 States. Second, the regional gas well production half-lives have converged to a value of between 23 and 25 months. Although individual well performance varies, the 1999 regional production half-lives show surprising uniformity.

A number of factors can influence a natural gas well’s production half-life, including the innate characteristics of the reservoir, the introduction of higher production rate technology (e.g., horizontal drilling and completions), low wellhead natural gas prices (which can cause producers to reduce production below a well’s full capability), lack of sufficient pipeline capacity, and State conservation regulations. Indeed, the fact that natural gas wells were not operating at full productive capacity before 1993 suggests that well production half-lives would be lower today than they were before 1993. Regardless of the cause, a more rapid decline in wellhead productive capacity requires producers to drill more wells per year in order to maintain a given level of natural gas production.

More rapid declines in natural gas well production have been actively pursued by producers through the application of technology, in order to improve their financial position. Increasing near-term cash flow is of paramount importance to the oil and gas business. Production rate technologies that enhance near-term cash flows increase the net present value of the discounted cash flow, improve rates of return, reduce the investment payback period, and lower producers’ exposure to future price risk. But better production rate technology is a double-edged sword. While it improves the near-term financial position of producers, it simultaneously increases the requirement for natural gas well investment in the next year and the year beyond.

If natural gas well drilling were to stop completely, productive capacity in the lower 48 States would decline by between 14 and 22 percent after 1 year and between 26 and 39 percent after 2 years.¹⁰⁰ Thus, next year’s stock of productive capacity depends on the extent of this year’s drilling. At the same time, natural gas well drilling activity depends on producer cash flow, which in turn depends on wellhead price levels. Consequently, low wellhead natural gas prices over any sustained period of time will lower producer cash flow, and could cause natural gas drilling to decline sufficiently to cause productive capacity to be less than the potential natural gas demand within a period as short as one year. In the bust phase of a business cycle, low prices, low cash flows and low investment levels will increase the probability that natural gas supplies will fall quickly and cause a deficit in wellhead productive capacity relative to natural gas demand.

Because neither productive capacity nor consumption is highly elastic with respect to price in the short term, a relative scarcity of wellhead productive capacity could be expected to cause very high natural gas prices, and a relative surplus could be expected to cause very low prices. When supply is relatively scarce, the short-term inelasticity of consumption and the delay associated with new natural gas supply investment will tend to cause natural gas prices to overshoot the long-run market-clearing wellhead price. This price behavior, in conjunction with the investment delays, would contribute to the tendency for natural gas productive capacity to overshoot demand, followed by an extended period of low prices and insufficient investment. Consequently, the natural gas industry embodies a set of dynamics that can cause periodic cycles in investment, drilling, supply, and prices.

Implications of Large, Unpredictable Price Fluctuations

Because the restructuring of the natural gas industry and its operation within a more competitive market structure are relatively recent, there is little experience from which to predict its future course of behavior. As discussed above, the market dynamics of the natural gas industry suggest that future cycles in natural gas prices and supply investments are possible. Of course, the market is also subject to unforeseen events. For example, oil price shocks, domestic economic cycles, and random weather events all tend to destabilize the balance of natural gas supply and demand. Consequently, a natural gas supply business cycle would probably not exhibit predictable, periodic fluctuations in price and supply investments. Rather, the future behavior of natural gas markets may be more analogous to a group of people kicking a ball suspended on a string. The suspended ball has a tendency to oscillate (around a long-run equilibrium), but the kicks change both the ball’s direction and the amplitude of its oscillations.

In the U.S. natural gas market, the extent of the price peaks and troughs during a market cycle will depend largely on weather conditions. Weather is a random force that can either reduce or exacerbate supply scarcity. When natural gas supplies are scarce, abnormally warm winters will moderate price increases; when supplies are abundant, warm winters will depress prices. Similarly, abnormally cold winter weather will exacerbate price increases when natural gas supply is already scarce and moderate price declines when supplies are abundant. Random weather effects, in conjunction with business cycle investment behavior, could result in thoroughly unpredictable prices and investment patterns. Such unpredictability could be further confounded by a whole range of events, such as oil price shocks and domestic economic cycles.

The unpredictability of future price behavior tends to obscure whether the natural gas industry is actually experiencing business cycle behavior and/or where it is situated with respect to any particular cycle. For example, wellhead productive capacity is not measured on any real-time basis. Wellhead natural gas prices are the only real-time measure of supply adequacy, but this measure is a relative one, which is also determined by prevailing consumption requirements. Although market analysts will associate causality to daily, weekly, and monthly wellhead price movements, the actual causes for price movements sometimes are not apparent until well after the fact, when more complete market data are available. Supply and demand uncertainty, in turn, encourages natural gas producers to delay reaction to market price changes, because of the risk of earning an inadequate rate of return on investment.

If large, unpredictable price excursions become common in the future, then the concept of a long-term price trend will be less meaningful to natural gas producers, consumers, and investors, because random short-term price excursions will confuse and obscure the longer term trends underlying the market at any point in time. Even if long-term price trends were clear, the risk associated with unpredictable price behavior could be financially devastating to natural gas producers, consumers, and investors in the short term.

The perceived financial viability of a natural gas supply investment depends, in part, on the price that the project sponsors expect to realize from the investment. Large, unpredictable price fluctuations expose potential projects to the possibility that when the project starts bringing its natural gas supplies to the marketplace, the price may be substantially lower than the price originally expected, imposing substantial risk on large capital-intensive projects that require long lead times, such as an LNG terminal or a natural gas pipeline from Alaska to the lower 48 States.

In contrast, short-lived supply projects that can be completed quickly during periods of high prices face less price risk. Projects that generally fit this description are conventional onshore drilling investments. Consequently, the potential for unpredictable future price behavior would result in an investment emphasis on near-term conventional drilling at the expense of investments in LNG terminals or an Alaskan pipeline, or in highly risky rank wildcatting, which is necessary to test new geophysical theories regarding natural gas formation and disposition. (Ironically, the future existence of extensive LNG infrastructure could potentially serve to moderate price fluctuations. LNG facilities could serve as “swing suppliers” of natural gas by providing incremental supplies during periods of high prices and curtailing shipments during periods of low prices.)

Unpredictable price fluctuations could also mask long-term trends in natural gas prices. For example, high prices during the natural gas shortages of the mid-1970s led many industry analysts to conclude that conventional natural gas supplies were insufficient to satisfy future natural gas consumption requirements. Many billions of dollars were invested in LNG terminals under the premise that inadequate domestic supplies would be reflected in high marginal production costs and high wellhead prices sufficient to justify LNG investments. Those expectations about long-term price behavior proved wrong, and the project sponsors lost substantial investment capital as a result.

Finally, unpredictable prices have deleterious consequences for natural gas consumers. For example, they obscure the value of appliances with higher energy efficiency ratings¹⁰¹ and can affect the financial viability of large industrial projects, such as electricity generation plants and fertilizer plants, where natural gas supply is the largest component of operating costs. Consequently, new coal-fired projects might become more financially attractive than new natural-gas-fired projects, simply because coal prices would be expected to be more predictable and much less likely to exhibit extreme fluctuations.

The deleterious effects of cyclical prices on suppliers and consumers can be mitigated through long-term, fixed-price contracts and price hedging; however, those financial instruments are limited in their duration and access. It is unlikely, for example, that natural gas supply contracts would be written for terms longer than 5 years without price re-openers. Moreover, they are generally not available to small consumers, especially residential natural gas consumers, although many local natural gas distributors do have constant payment programs for residential consumers, which may mitigate the financial impacts of unpredictable prices.

Notes