The Dynamics of Natural Gas Price Formation:
Implications for Gas Producers
J. Michael Bodell
The U.S. reference cash or spot gas price is set at the Henry Hub (HH) in Sabine, Louisiana. It depends on (1) weather-related, near-term supply-consumption fundamentals, (2) the marginal cost of supply, and to a minor degree on (3) the utility value of alternate fuels on a Btu-parity basis. Yet the dominant factor to the systematic variability in spot gas price is supply and consumption. The tension between these fundamentals is reflected in weekly underground natural gas storage inventories as reported by the Energy Information Administration (EIA).
Weather is the major short-term driver of storage inventories. Moderate weather increases inventories as supply advances over consumption, and extreme weather depletes inventories. This is a simplification because price over the long-term has an impact on future gas supply, whether it comes from domestic resources or imports, such as liquefied natural gas or Canadian pipeline exports. Furthermore, long-term change in use-efficiency lowers consumption per heating or cooling degree-day (weather) while a growing U.S. population - increasingly dependent on natural gas - boosts demand. In general, however, weather with its impact on storage inventory is the primary short-term driver to change in spot natural gas price.
Unlike weather driven-demand, domestic gas production is relatively flat or stable during the course of the year. In North America, during the spring, summer and fall, production typically exceeds demand thereby permitting gas to be stored for use during winter's heating season. But implicitly, the market understands that in order to maintain and/or to grow supply, a certain price level needs to exist over the long-term. Through timely price adjustments, the market discovers that one level will force producers to lower investment thereby threatening supply security while another price level induces 'irrational exuberance' in investment. Price, therefore, becomes the de facto long-term market arbiter to influence supply in the face of uncertain and variable demand.
The market gauges the strength of the sum of all gas supply in the context of consumption through change in storage inventories. More importantly, the weekly release of the EIA storage inventory data provides a regular measure of readily accessible gas supply. Storage volumes become critical in the short-term since they are the only reliable extant source of gas to balance a stressed market. In essence, the market understands whether prevailing storage levels are net long or net short relative to recent history (e.g., the most recent five years). Conceptually, the market expects a certain storage inventory or 'cushion' and it prices natural gas accordingly.
Spot price is highly correlated to the size of a particular storage cushion. We compare this to a specific moving benchmark inventory. This permits near real-time evaluation of the dynamic tension between supply and consumption - as gas inventories build or are depleted - and the attendant price response. Exhibit 1 is an economic price-vs-quantity diagram that shows this price variability in 2007 and 2008 as a function of change in net volumes. We call this price-quantity relationship a 'yield curve' because of important non-linear dynamics in the value of physical gas as correlated to comparative inventory position. In essence, when comparative inventories are net long, spot prices in general decline to a somewhat stable local minimum or plateau. Conversely when inventories are short, spot prices rise and when they are very short, rise very quickly (occasionally leading to high volatility as seen in larger daily price changes) in an attempt to ration remaining physical inventory. Viewed in this way, it is clear that over the longterm the market uses rational price signals to encourage the development of stable sustainable supplies and to discourage the maintenance of surplus supplies. And when storage inventories are in severe stress (i.e., a deficit position), the market uses price to ration consumption, sometimes destroying demand and frequently creating an economic incentive to switch to an alternate fuel.
Alternate fuel values, notably low sulfur No. 6 residual fuel oil and steaming coal, have several impacts on natural gas price. Firstly, alternative fuel values create price breakpoints related to fuel-switching economics for these fuels. For example, as oil price increases it can pull up the value of natural gas in parity to No. 6 fuel oil. The converse is true as seen in late 2008 as crude oil prices plummeted. Also, when the price increases for coal used in power generation, as it did in for the first half of 2008 in response to global supply concerns, it raises the effective floor value of natural gas.
Secondly, a change in oil price can cause in effect an up or down shift in the yield curve. In part, the effect may be seen as variability about the trend in the data (refer to Exhibit 1). For example, if crude oil price decreased by $1.00 per barrel, the value of No. 6 fuel oil typically declines. To maintain some semblance of Btu parity, the price of natural gas may respond likewise (~10 to 13 cents per mmBtu), although not always on a 1:1 basis. This relationship is complex and requires a detailed understanding of respective crude and natural gas inventories, supply, and demand seasonality. Moreover, natural gas price dynamics are separate from those of alternate fuels. In fact, an argument can be forwarded that in 2007 and 2008 crude oil values have had essentially no impact on the value of natural gas until December 2008 when No. 6 fuel oil declined below natural gas on a Btu-parity basis. In this case, because natural gas was net long, the declining value of crude essentially drove natural gas prices lower.
An evaluation of the historic price structure with respect to comparative inventory reveals a longer-term 'mid-cycle' price level or platform about which spot prices oscillate over a period of a year or longer. Note that in Exhibit 1 the mid-cycle price level at 0 (zero) on the x-axis and the storagerelated price variability about this point. Theoretically, a change to a long-term price platform is the mechanism the market employs to secure adequate supply. Therefore, we postulate that the midcycle price is the effective cost of marginal supply required to maintain long-term supplyconsumption balance and to ensure some semblance of supply security.
This price level specifically must be sufficient for at least 90 percent of producers earn at least their weighted average cost of capital of 6 to 8 percent on the exploration and development of new resources. The market will not normally reward over-investment or mal-investment (mal-investment refers to development of resources with a very high cost structure such that the play fails to ear