Daily Price Adjustments in the U.S. Market for Natural Gas
Bopp, Anthony E., Atlantic Economic Journal
ANTHONY E. BOPP [*]
A daily model of cash market for natural gas in the U.S. is presented and estimated over 1997. The model develops the notion that the expected rather than the actual amount of gas in storage (along with weather) and storage changes impact the current daily cash price. These notions are supported by the estimations presented. At the daily observation level, reported changes in storage levels are signals that actual consumption or production are either as expected and result in no price pressure or are not as expected and do result in price pressures. (JEL Q41)
Recent research has considered the informational efficiency in natural gas markets. Walls  presents evidence to support the notion that a single natural gas price prevails in the marketplace. That is, at any of the reporting points in the U.S., the price of natural gas equals the single price plus a transportation charge plus a random component. King and Cuc  then claimed that there is an east-west split in the U.S. natural gas market and Serletis  countered that there may not even be a composite east or west market. It has also been supported that the cash price of natural gas is cointegrated with the futures price. That is, essentially, each price conveys the same information about the present and expected underlying value of this commodity [Walls, 1995]. All of these studies deal with monthly gas prices, while gas is traded on a daily basis. This work extends the study of gas markets to the daily level by using results from the supply of storage literature and the notion that price expec tations in this market are based on relative, not absolute, levels of gas in storage.
Much gas is sold in daily cash markets, although more is sold through monthly, seasonal, and annual contracts. Those contracts are usually indexed to the daily average cash price of the contractual month [Energy Information Administration, 1995]. Understanding the workings of the U.S. natural gas market begins with an understanding of the daily cash market.
The second section presents background information about this market and the information supports the underlying structure of the model. The third section presents estimation results and an analysis of those results. Finally, some observations are made concerning where this work falls relative to the efficiency question debated in the above-mentioned literature.
Market Discussion and Model Development
A number of facts and assumptions regarding the workings of the U.S. natural gas market form the theoretical basis of the model presented. These factors are presented along with the development of the model.
First, Figure 1 shows U.S. natural gas consumption and production (including net imports) rates for 1996 and 1997 on a monthly basis (the shortest reporting frequency for these variables). Over this period, either gas production is relatively fixed or does not respond to contemporaneous factors or contemporaneous factors did not change in a way that would affect production decisions. Also, it is possible that production rates are predetermined by capital budgeting decisions made well in advance of production reporting periods. Whatever the reason, changes in production will not add much to an understanding of prices during this period. For purposes of explaining within month daily prices, it is assumed that production rates are fixed, so that:
[S.sub.t] = [S.sub.0] + [e.sub.t], (1)
where S is production plus net imports and e is an assumed random noise error term.
Figure 1 also shows considerable variation in consumption rates but of the kind that relates mainly to weather variation. The 1997 levels are very close to the 1996 levels primarily because 1997 experienced less extreme weather, which effectively negated any longer term changes related to relative prices and economic activity movements. …