Open Access and the Emergence of a Competitive Natural Gas Market
Vany, Arthur de, Walls, W. David, Contemporary Economic Policy
For more than 40 years, natural gas pipelines were regulated as natural monopolies. Federal regulations (i) authorized a single pipeline to link a city market with a producing area, (ii) limited entry, (iii) controlled transportation tariffs and gas prices, and (iv) required gas pipelines to own the gas they transported. But, in 1985, a new form of organization was authorized. The Federal Energy Regulatory Commission (FERC) partially deregulated natural gas pipelines and permitted them to become transporters of natural gas rather than merchants. Pipelines were allowed to unbundle natural gas from its transportation and to offer pure transportation services to their customers.(1) This change dissolved the barriers to markets that were inherent in the merchant carrier pipeline system and, for the first time in more than 50 years, authorized competition as a mechanism for distributing natural gas.
This paper examines the emergence of the competitive national natural gas market. This market was made possible by regulatory changes and by the creation of new decentralized market institutions that replaced the merchant carriage system for coordinating natural gas and transportation over the vast pipeline network of the United States. The primary focus here is on the institutional changes that underlie the emergence of natural gas spot markets and on the evolution of markets and prices from 1985 to the present. Removing the barrier that mandated merchant carriage between gas buyers and sellers brought forth new markets where none had existed. In five years, the number of markets reporting spot prices--in step with the number of pipelines offering transportation--grew from a handful to more than 50. Spot markets now exist in every production field and at many points where major pipelines interconnect. The trading is almost wholly decentralized. There is no central market for natural gas, and the bidweek institution is the primary means of coordinating gas and transportation.
Institutional and empirical evidence shows that as more pipelines elected to transport gas, new markets opened, the network of pipelines became more strongly connected, and prices converged throughout the network. Over the six year period of sample spot prices, one can observe the convergence of gas prices at geographically separated points in the network as competition began to enforce the law of one price.
II. THE EVOLUTION OF NATURAL GAS MARKET INSTITUTIONS
The three sectors in the natural gas industry are production, transmission, and distribution. Natural gas pipelines transport gas between the production and the distribution sectors. Of the three possible forms of transmission service that pipelines could have offered--merchant carriage, contract carriage, or common carriage--regulation mandated their structuring as integrated merchant carriers who sold gas bundled with transportation (see Daggett, 1955 for a discussion of the various forms of carriage). This structure restricted competition and balkanized gas markets.
A. Merchant Carriage
During the 1930s, the natural gas industry was vertically integrated to a significant degree. In some cases, producers owned pipelines. In other cases, distributors owned pipelines and production facilities in the gas fields. Under the Natural Gas Act, Congress re-organized the industry as a system of separate merchant carrier pipelines divorced from production and distribution. Vertical integration was discouraged, entry was controlled, and pipeline tariffs and gas prices were regulated.(2) Pipelines were required to tie the sale of gas to its transportation and could not offer pure transportation to their customers. Pipeline customers, usually local distribution companies and large end users, could purchase only the bundled package of services that included gas acquisition, storage, and transmission. Two qualities of bundled service were offered: interruptible deliveries and firm (uninterruptible) deliveries.
The process through which federal regulators certificated pipeline construction led to a dense but disconnected network of pipelines. Individual pipelines were constructed as new supplies were found and as the demand for natural gas increased. The Federal Power Commission (FPC) certificated construction of a new pipeline only after the pipeline had shown that its reserves could supply its downstream customers for a period of 15 to 20 years. To achieve this end, pipelines entered long-term contracts with producers under which the reserves in the gas wells were dedicated to the pipeline. Pipelines could not be abandoned, and contracts could not be renegotiated without FPC approval.
This regulatory process balkanized gas markets and created a disconnected network topology that prevented gas from flowing from each connected field to each connected city. The cities and producing fields connected by pipelines were isolated from the fields and cities connected by other pipelines. Pipelines operated independently of one another, each supplying its own cities with its dedicated gas supplies.(3) The disconnected network topology and trading limitations prevented markets and blocked competitive arbitrage between points of supply and use.
B. Deregulation and Open Access
Despite regulation that attempted to maintain high levels of reliability to users, federal price controls caused shortages of natural gas in the 1960s and 1970s (MacAvoy and Pindyck, 1975). In response to these shortages, Congress passed the Natural Gas Policy Act in 1978. The Act deregulated the field price of gas in steps and completely deregulated the price of some types of gas. In 1979, a major interruption of world oil supplies occurred. Reeling from gas curtailments, a run-up in oil prices, and uncertainty over the deregulation of field prices, many pipelines signed long-term contracts to buy large volumes of gas at high prices. When gas prices fell after wellhead prices were deregulated, these pipelines faced infeasible minimum purchase obligations of high-priced gas. Many of them renegotiated their contracts with producers. In exchange for partial release from their purchase obligations, these pipelines offered to transport gas for producers or their customers. This was the beginning of open access transportation in gas markets.
The FERC at first approved these transportation transactions individually. Then in October 1985, it issued Order 436 permitting interstate pipelines to transport gas for others under "blanket certificates." This regulatory order formally distinguished and separated the pipeline merchant and transportation functions. Earlier orders already had dismantled long-term contracts and left pipelines with few options to open access. After some initial skepticism, pipelines began applying to become "Open Access" pipelines. As figure 1 shows, the number of pipeline applications and approvals for open access carrier status grew rapidly from 1985 to 1990. Within three years of Order 436, nearly all the major pipelines had become open access pipelines.(4)
Open access pipelines continue to offer their traditional service of bundled gas and transportation. However, most of their throughput is gas transported on their customers' behalf. Between 1982 and 1987, transmission of pipeline-owned gas decreased 60 percent while transmission of customer-owned gas increased by 180 percent (U.S. Energy Information Administration, 1989). Transportation accounted for two-thirds of all interstate gas movements by 1987, and in 1991 customers owned over 85 percent of gas shipped in interstate commerce (FERC, 1991a). Figure 2 reveals the rapid shift from selling gas to transporting it for others by showing the amount of gas transported for brokers and local distribution companies from 1982 through 1987.
As open access spread through the pipeline network, spot markets opened at fields and interconnection points. The number of spot markets reporting prices to the Gas Daily, an industry periodical, grew from zero in 1985 to around 50 by 1990. Figure 3 shows the growth in spot markets. The number of markets doubled from 1987 to 1990. By 1990, it was possible to make delivery throughout most of the pipeline network, and a natural gas futures market opened on the New York Mercantile Exchange.(5)
C. Contract Carriage and the Decentralization of Transportation
Under merchant carriage, gas users contracted with pipelines for a firm (uninterruptible) supply of gas up to some maximum daily volume, the "callable" volume. The unused portion of the callable volume was sold as interruptible gas. This gas was subject to interruption, and the holder of a firm supply contract had priority over others. In making the transition from merchant to contract carriage, the FERC required the pipeline to permit customers who held firm purchase agreements to convert the callable volume to uninterruptible transmission rights. On making that election, the holder of transmission capacity became obligated to pay a reservation charge--which depended only on the volume of gas for which uninterruptible transportation was reserved--and a volumetric charge for each unit of gas shipped. Most of the transmission capacity in the pipeline grid is under contract to the companies who distribute gas in the city gate markets to wholesale and retail customers. These local distribution companies were the major gas buyers before open access. They inherited transportation capacity when their gas contracts were converted to transportation contracts.
Holders of firm transportation contracts may trade with one another or transfer their rights to brokers and other parties. However, the FERC has not permitted transportation to become a fully transferable property right (Smith et al., 1990; FERC, 1991b). Unused firm transmission capacity reverts to the pipeline, which sells it as interruptible transportation. Pipelines monitor throughput and post unused capacity for sale on electronic bulletin boards. These bulletin boards are accessible to all market participants who can buy interruptible transportation at the tariff posted by the pipeline. Because the tariffs may be discounted below the regulated tariff, the price of transportation adjusts continuously to clear the market, up to the maximum regulated tariff. In peak periods, the regulated tariff may become a binding upper limit on the market price of transportation.(6)
As a result of these changes, the transportation capacity has been reallocated from pipelines to their customers and the control of transportation has been decentralized. There are 21 major interstate pipelines on which 1,400 local distributors hold transportation contracts (Bradley, 1991). Instead of a single supplier's controlling transportation on a pipeline, there now are several suppliers of transportation on each pipeline. On average, each pipeline has nearly 70 suppliers of transportation rights. With active brokering of unused capacity, a competitive supply of interruptible transportation exists in many markets--even in those served by a single pipeline.
D. Institutions for Coordinating Gas and Transportation
If gas users hold or acquire transportation contracts, in each downstream market they can purchase from all the fields to which they are directly or indirectly connected. If the prices across fields are disparate, gas purchasers will demand transportation connections to gain access to fields with low prices. Gas producers in fields with low prices will demand transportation connections to gain access to customers in downstream markets with high prices. When new pipeline interconnections are made, gas can flow to reduce price disparities across the pipeline network. The growth of the number of markets reporting spot prices clearly supports the proposition that more pipelines opened their systems to access.(7) In addition, the emergence of the "pooling hubs" at pipeline intersections indicates that interconnecting was proceeding rapidly (De Vany and Walls, 1992b).
A transaction for gas specifies a volume to be transported on a pipeline from an injection point to a withdrawal point. Metering verifies the volume injected and withdrawn at these points. That is, the molecules are not traced from one location to another. The nature of a transportation contract then is to specify injection and withdrawal points and maximum volumes per month that can be transported over a pipeline. Using transportation rights and offsetting trades, gas traders can "clear" transactions on injection and withdrawal points when these points are against the direction of the gas flow on the pipeline. This "backhaul shipment" involves withdrawing gas that is headed to a user downstream of the trader's injection point. The trader then replaces the withdrawn gas at his injection point so that it reaches the downstream trader to whom it originally was destined. By holding contracts on different pipelines, traders can inject and withdraw gas at points on different pipelines even if those points do not interconnect. The injection and withdrawal of gas on each line, in combination with a clearing house for contracts, permits virtual trading between unconnected points.
Suppose that pipeline 1 connects field A and city B and that pipeline 2 connects field C and city D. To buy an amount of gas from field A for "delivery" to city D on pipeline 2, execute the following transaction: buy injection for that volume at A and withdrawal at D and sell withdrawal at B and injection at C. At the same time, buy gas at the injection points and sell it at the withdrawal points. For trading to be this sophisticated, traders must be large enough to clear these transactions internally. Otherwise, they require some kind of clearing house. Alternatively, the pipe-lines could be interconnected by adding links in the grid.
All three of the requirements to support gas trading have been satisfied: (i) Brokers capable of transacting throughout the network have come into existence. (ii) Mergers and interconnects have been made throughout the network.(8) (iii) Market institutions were created to coordinate gas and transportation trades. Brokers buy and sell gas throughout the pipeline network, even though they do not have uninterruptible transmission rights of their own. They aggregate producers' supplies and gas users' demands. By purchasing interruptible transmission from the pipeline, they can ship gas from the producers to the users. Essentially, brokers hold a portfolio of gas market transactions that they match in real time. Some brokers act as the purchasing agent for downstream local distribution companies. These brokers use the customer's transmission capacity to deliver the gas that they sell to the customer.
Pipelines coordinate their customers' transmission demands during "bidweek." During the bidweek--usually the third week of each month--pipeline customers nominate the volumes they plan to ship during the following month. These nominations specify the injection point, the withdrawal point, and the volume of gas to be shipped. Customers may nominate volumes only up to the amount of their firm transmission rights. Those pipeline customers who transfer their transmission capacity to third parties are responsible for nominating and paying for it.
During the bidweek, gas users who hold capacity rights purchase the volumes that they nominate for transmission during the following month. The spot contracts they enter are for volumes to be delivered to a specific injection point on the pipeline system. From this point, they exercise their transmission right and withdraw the gas from the pipeline at the downstream destination. The duration of these spot contracts usually is 30 days or less. The average transaction price of these spot contracts executed during the bidweek is the "bidweek price."
III. TRADING AND THE STRUCTURE OF GAS PRICES
With the institutions now in place, one can model the behavior of prices in the gas pipeline network. The central issue is how the change from merchant carriage to open access transportation alters prices. The findings show that, under the regulated system of merchant carriage, only the price spreads between fields and cities are determined. The levels of these prices are indeterminate. However, open access competition makes prices determinant. In the model, competition in the network is arbitrage and implies an inverse relationship between the connectivity of the network and the dispersion of prices.
A. Trading in a Simple Pipeline Network
Consider a simple pipeline system shown in figure 4. Point 1 is a field where gas is produced (a source). Points 2 and 3 are cities where gas is consumed (sinks). Under the regulated system of merchant carriage, gas could be transported from 1 to 2 and from 1 to 3, but not from 2 to 3: the pipeline was the sole buyer of gas at 1 and the sole seller at points 2 and 3. Entry was closed. If the pipeline rationally attempted to exploit this legally granted monopoly, it would do so by increasing the price spreads p2 - p1 and p3 - p1 and thus would drive the field price down and the end market prices up. Cost-based rate making would control only the spread between field and city prices, allowing the pipeline to charge a city price that exceeded the field price by the cost of transmission. City prices equal field prices plus tariff, and they are bounded between production cost and the maximum values that consumers are willing to pay. This method of regulation determines the spread between city prices and the field price but leaves their levels indeterminate. Whatever the pipeline pays for gas passes through as a cost to customers.
The transition to contract carriage changed this scenario. Under contract carriage, buyers and sellers meet in an auction for gas and establish a market price in the field. City prices equal this price plus the regulated transmission cost. Since the number of buyers and sellers on a pipeline typically is very large, the field price will tend to go to the competitive level. Even if there are only a few suppliers, the auction institution and suppliers' inability to withhold supply indefinitely will move price toward the competitive level.
B. Network Connectivity and Price Convergence
Prices converge as the triangle network in figure 4 is embedded in a larger network. Contract carriage does exactly this. It allows for the linking of other pipelines into the system so as to form a larger and more connected network. The production fields, pipeline hubs, and the city markets are elements of the set of vertices V of a network N, a network and the arcs of the pipeline network are the ordered pairs of the elements of V. Implementing open access on the network would permit any trader to inject gas at any source and to withdraw gas at any sink. This is precisely the "transitive closure" of the network, C(N). The transitive closure is the maximally connected network containing the vertices of N that preserves the direction of flow. The transitive closure of N indicates the market connections made possible by open access. No new pipelines are constructed. The trading of transportation rights and the interconnection of lines link markets. The process of interconnecting lines has progressed rapidly and created the emergence of pooling areas where new markets have come forth (FERC, 1991a).
Consider prices in a network N with vertex and arc sets (V, A). The vertices are supply and demand points, and the arcs connect vertices. The number of arcs, A, represents how many pipelines connect vertices and whether segments of those pipelines can be used to carry gas between vertices. Consequently, simply opening a segment of a pipeline to transportation between points on the line that previously had not been open to transportation may increase A. Given m supply points and n demand points, let P(V, A) be the m x n space of equilibrium vertex prices. Let the number of arcs connecting the vertices of N increase from A to A + 1 with no change in the network's vertices. When the number of connecting arcs increases, the number of effective constraints on prices increases and the equilibrium price set shrinks. Thus, P(V, A + 1) is the subset of P(V, A). The set of equilibrium prices satisfies the arbitrage constraints between all connected pairs of supply and demand points. This set must shrink as the number of arcs in the network grows because the number of competitive supply sources becomes larger and new arbitrage constraints are recruited over the network.(9)
IV. DATA AND EMPIRICAL ANALYSIS
Two separate sets of data reveal the evolution of natural gas market integration during the early, middle, and later periods of open access. One data set is regional and spans the time interval from merchant carriage through contract carriage. The second data set is more disaggregated and covers the period of contract carriage. Because the data sets are separate, the analysis here separately discusses the empirical results.
A. The Regional Data
The regional data are from the Energy Information Administration (EIA) and consist of monthly observations of the average spot price paid for natural gas in dollars per thousand cubic feet ($/Mcf) in five regions: Appalachia, Louisiana, Oklahoma, the Rockies and Texas. The EIA constructed these data by averaging the spot price that several industry periodicals reported in each region. Aggregating these disparate price series increases the noise in the data and biases the results away from finding significant relationships among prices. Aggregating allows constructing a longer time series of prices that are measured consistently. However, it also is true that individual reporting errors are less likely to influence significantly the empirical results because the data are a composite measure.
Consider the time series plot of the regional spot prices in figure 5. The figure shows a pattern of prices that clearly changed over time. The data follow three distinct patterns over three successive time periods: (i) from 1984-1985 prices move independently, (ii) from 1986-1987 all price movements are small, and (iii) from 1988-1989 each series moves in step with every other series. Examine the 19841985 period during which pipelines were disconnected sellers of bundled gas and transportation. In the summer of 1984, the price of gas in Appalachia fell while its price in the Rockies rose. During the following winter, the price rose in Appalachia while it fell in the Rockies. Even the prices in Louisiana and Texas, the two regions served by the most dense grid of pipelines, do not move together during 1984. By the end of 1986--about a year into open access transportation--the gas prices begin to move closely together, with each series declining at about the same rate. After 1987, spot prices in all regions move together.
(i) Correlations of Regional Spot Prices. The Pearson correlation coefficient between each pair of price series was calculated for three successive time periods: 1984-1985, 1986-1987, and 1988-1989. The series have a high degree of positive first order serial correlation. Therefore, the correlation analysis uses the first differences of the price series.(10) The analysis adjusts prices for seasonal demand--measured as heating degree-days--before calculating the correlation coefficients. Therefore, the correlation analysis uses Fisher's r to Z transformation tests whether the correlation between regions changed over the sample periods.(11) One can test the hypothesis that the correlation is equal between two independent samples of size |n.sub.1^ and |n.sub.2^ by computing the test statistic (|Z.sub.1^ - |Z.sub.2^) / |s.sub.12^ where |s.sub.12^ = (1/(|n.sub.1^ - 3) + 1/|(|n.sub.2^ - 3)).sup.1/2^. This test statistic follows a standard normal distribution.
Table 1 reports the correlations between pairs of regions for three sub-samples: 1984-1985, 1986-1987 and 1988-1989. The table also gives the value of the test statistic for the hypothesis that no change occurred in price correlations between the first and last period. The correlation between prices increased significantly for each region-pair.(12) These results support the hypothesis that these five regions functioned as distinct markets in 1984-1985 but evolved over time into one large market.
(ii) Convergence of Price Spreads. If the five separate markets did converge to a single market, then price spreads should have become less volatile. Tables 2a and 2b show the spreads in prices and associated descriptive statistics for each region-pair for 1984-1985 and 1988-1989, respectively. For each region-pair, the range of the spread (the maximum spread minus the minimum spread) decreased, and the standard deviation of the spread also decreased. In some cases, the magnitude of the average price spread increased--for example, the spread between Appalachia and the Rockies--but this new spread was more stable.
B. The Disaggregated Data
Gas Daily reports bidweek spot prices at over 50 pipeline interconnection vertices within the transmission network. The interconnection vertices are located either where smaller pipelines feed gas from the producing fields to the major trunk pipe-lines or where several trunk pipelines interconnect. In the analysis here, the sample of spot markets consisted of 25 vertices for which Gas Daily has continuously reported bidweek spot prices since February 1988. The vertices are located within six geographic areas: West Texas, East Texas, North Texas, South Texas, Oklahoma, and the Louisiana Onshore region. Thirteen of the major interstate pipelines are represented. These pipelines account for the majority of the gas flowing through interstate pipelines (EIA, 1989). Table 3 lists the vertices by the region in which they are located and the pipeline to which they are attached.
Bidweek prices are the weighted average price of gas purchased based on volumes and prices for spot deals struck during the bidweek. These prices are for gas injected into the pipeline at the vertex for which the price is listed. The prices include all gathering and transportation fees incurred to get the gas to the points for which prices are reported. All prices are based on dry packages of five million cubic feet stated in dollars per million Btu ($/MMBtu) for spot contracts of 30 days or less.
(i) Correlations of Bidweek Spot Prices. The correlations between the injection node-pairs are high, but table 4 reveals no-table patterns. In 1988, prices were more highly correlated between vertices in the same region than between vertices in more distant regions. Compare correlations of prices in West Texas with prices in other regions. The estimated correlation coefficient between bidweek prices on the ANR pipeline in North Texas and the El Paso pipeline in West Texas is 0.79. By comparison, within North Texas, even the lowest correlation is 0.97 (between the ANR and the Northern pipeline). Prices in Oklahoma are less correlated with prices in more distant regions such as South Texas than they are for nearby regions such as North Texas and Northern Louisiana. In general, 1988 prices were more highly correlated between closer points than between more distant points.
TABLE 1 Correlation of Price Differences Orthogonal to Seasonal Demand Appalachia Louisiana Oklahoma Rockies 1984-1985 Louisiana 0.012 Oklahoma 0.092 0.164 Rockies 0.063 0.062 0.351 Texas 0.299 0.548 0.400 0.166 1986-1987 Louisiana 0.506 Oklahoma 0.477 0.936 Rockies 0.787 0.671 0.597 Texas 0.545 0.920 0.916 0.658 1988-1989 Louisiana 0.860 Oklahoma 0.846 0.976 Rockies 0.701 0.800 0.781 Texas 0.793 0.962 0.962 0.801 Z-statistic for change in correlation between 1984-1985 and 1988-1989 Appalachia Louisiana Oklahoma Rockies Louisiana 4.101 Oklahoma 3.725 6.529 Rockies 2.580 3.318 2.181 Texas 2.468 4.342 4.956 2.989
In contrast with the results for 1988, the price correlations for 1990 are high between vertices in the same region and between vertices in different regions. The increase in price correlations from 1988 to 1990 is particularly noticeable between those vertices on pipelines in West Texas with vertices on pipelines in other regions. The correlation between ANR in North Texas and El Paso in West Texas now is 0.93, up from the 1988 correlation of 0.79. The correlation between vertices located in South Texas and Oklahoma also is much higher, about 0.98. The 1990 correlations differ from the 1988 correlations in two ways: for 1990, all correlations are higher, and the correlations between prices at near vertices nearly equal the correlations between prices at distant vertices.
TABLE 2a Inter-regional Price Spreads, 1984-1985 Regions Mean Spread Std. Dev. Min Max Appalachia-Rockies 0.446 0.220 -0.01 0.91 Appalachia-Louisiana 0.339 0.184 -0.12 0.65 Appalachia-Oklahoma 0.495 0.185 0.16 0.85 Appalachia-Texas 0.287 0.171 -0.23 0.51 Rockies-Louisiana -0.108 0.145 -0.44 0.11 Rockies-Oklahoma 0.049 0.136 -0.24 0.32 Rockies-Texas -0.160 0.169 -0.56 0.13 Louisiana-Oklahoma 0.156 0.113 0.04 0.51 Louisiana-Texas -0.052 0.105 -0.24 0.13 Oklahoma-Texas -0.208 0.155 -0.60 0.01 TABLE 2b Inter-regional Price Spreads, 1988-1989 Regions Mean Spread Std. Dev. Min Max Appalachia-Rockies 0.696 0.180 0.36 0.94 Appalachia-Louisiana 0.327 0.132 -0.06 0.55 Appalachia-Oklahoma 0.418 0.128 0.05 0.62 Appalachia-Texas 0.377 0.141 -0.02 0.59 Rockies-Louisiana -0.370 0.120 -0.67 -0.18 Rockies-Oklahoma -0.279 0.114 -0.54 -0.04 Rockies-Texas -0.320 0.112 -0.56 -0.12 Louisiana-Oklahoma 0.091 0.027 0.05 0.17 Louisiana-Texas 0.050 0.042 -0.02 0.17 Oklahoma-Texas -0.041 0.043 -0.12 0.07
One way to scale the increase in correlation between more distant vertices from the early to the later period is to compute the condition number for each correlation matrix. The condition number is the ratio of the largest characteristic value of the matrix to the smallest. It measures the degree of collinearity of the columns in the correlation matrix. A high condition number indicates high collinearity (Belsley et al., 1980). The condition numbers for 1988, 1989, and 1990 are 7.03 x |10.sup.16^, 1.43 x |10.sup.18^, and 2.28 x |10.sup.22^, respectively. The increasing condition numbers indicate that the correlations are becoming more equal between all pairs of vertices. The condition number correlations in the network indicate how strongly connected the network is and how many paths exist between vertices. Thus, the increasing condition numbers of the correlation matrices indicate that the network has become more connected.
TABLE 3 Interconnection Nodes listed by Region and Pipeline North Texas South Texas ANR Pipeline Company Natural Gas Pipeline of America Natural Gas Pipeline of America Tennessee Gas Pipeline Company Northern Natural Gas Company Trunkline Gas Company Panhandle Eastern Pipe Line Company East Texas West Texas Natural Gas Pipeline of America El Paso Natural Gas Company Tennessee Gas Pipeline Company Transwestern Pipeline Company Trunkline Gas Company Louisiana Oklahoma Texas Gas Transmission ANR Pipeline Company ANR Pipeline Company Natural Gas Pipeline of America Columbia Gas Transmission Northern Natural Gas Tennessee Gas Pipeline Company Oklahoma Natural Gas Trunkline Gas Company Panhandle Eastern Pipe Line Company United Gas Pipe Line Company Southern Natural Gas Company Natural Gas Pipeline of America
The complexity of trading gas and transmission in the large network of markets that emerged during open access could have overwhelmed traders' abilities to coordinate their trades. After more than 40 years of regulation and "bundled" gas sales, traders did not gain the experience necessary to permit "unbundled" gas and transmission markets to perform well. In addition, experimental research by Rassenti et al. (1988), McCabe et al. (1990), and Plott (1988) finds a potential for coordination failure when gas is purchased separately from transportation.(13)
A coordination failure could occur if gas and transmission are purchased sequentially and a buyer of gas is unable to arrange for its transportation. In this instance, the buyer either must pay a premium for the transportation or unload the commodity on short notice. Coordination failures of this kind would result in short term illiquidity in gas or transportation. Evidence of such failures would be episodes TABULAR DATA OMITTED of price volatility and unused transportation (the gas can always be left in the ground). Traders may unload commitments because they fail to coordinate gas with transportation. The prices of transactions made after the bidweek auctions closed would reveal this problem.
Have gas market institutions been designed to avoid coordination failures? The evidence suggests that the commodity and transportation markets have been coordinated successfully through three mechanisms. (i) The bidweek auction for the commodity is coordinated with advance nominations that shippers make for transportation. That is, shippers simultaneously are able to buy the commodity and arrange for transportation. (ii) Those who hold firm transportation contracts have guaranteed transportation up to the limit of the contract and, hence, are able buy the spot commodity with assurance of being able to ship any amount up to their limit. (iii) Brokers who buy the commodity to ship via interruptible transportation can make and unmake deals on the commodity throughout bidweek as they observe the amount of firm transportation nominated by the commodity holders. The brokers have real time information on the amount of capacity booked, and they use this information to make their commitments for transportation and natural gas.
Beyond this institutional evidence is the price evidence. Spot price volatility has narrowed geographically and temporally. Further evidence of successful coordination is that virtually all spot gas transactions made after bidweek closes are at the bidweek price. That would not be the case if gas were being sold by a buyer who was unable to ship it or by someone who was trying to cover transportation already acquired. There is no evidence that commodity or transportation "corners" create extreme spot prices.
TABULAR DATA OMITTED
There is evidence that lines in the grid become congested from time to time. The spot price spreads between different locations widen seasonally. The only persistently wide spread is between the Rockies and the remaining supply fields. This gas is far from markets. As a result, the field price drops to match the delivered prices from other fields to each market. It was not well-connected to the national grid during the sample period, but later a new pipeline was constructed to connect these fields to west coast markets. The spot price evidence correctly identified this market as needing additional transportation.
Although some links do become congested, there is no evidence of bottleneck monopolies in the price data. Prices are highly correlated over all the vertices, near and far, and there is no evidence that prices at vertices served by only one or two pipelines are less correlated than are prices at vertices served by three or more pipelines. Price spreads have narrowed over time through the network. This suggests that non-equilibrium price disparities existed early on as open transportation began to spread through the network. These early spreads were characteristic of the old merchant carriage system. Markets were separate and were not linked to the network. Merchant carriage, dedicated gas, closed entry, and a disconnected network prevented other suppliers from contesting markets. Regulation created market power, which rate regulation alone could not eliminate.
C. Markets versus Regulation
In the beginning of open access transportation, price spreads reflected the way pipeline tariffs were set under regulation. Regulated pipeline tariffs were based on historical cost and differed among pipe-lines serving the same market for reasons unrelated to the value of transportation or gas in that market. Two pipelines delivering gas to the same city might have different delivered prices simply because their regulated tariffs differed. Competition made them bring their tariffs into line so that both could deliver into the market at the going price. Early FERC transportation programs granted permission for pipelines to discount tariffs. This practice carried over to open access. Pipelines now aggressively discount tariffs to meet the competition.
Finding hard evidence of monopoly is unlikely because the capacity of pipelines now is in the hands of the approximately 1,400 shippers. No longer does a single agent hold all transportation capacity on any link of the network. The firm capacity on most links is in the hands of the local gas distributors, who buy and ship on their own behalf. Hundreds of brokers trade interruptible transportation. Cost of service rate regulation creates poor incentives for local distributors to lease or sell their transportation rights since the revenue they gain from selling transportation will reduce what they are permitted to earn from selling gas.
Could natural gas prices have fallen over the period as they have without open access? Declining oil prices surely helped, but at least some credit goes to open access. Spot markets came widely into existence only after transportation gave buyers access to the fields. The network interconnection points could not have become markets in the absence of transportation, which permitted traders to interconnect pipelines. Transportation allowed spot gas to become a genuine alternative to contract gas. Spot gas prices, which were market-driven, put competitive pressure on contract prices. The decline in spot prices led the decline in contract prices.(14)
As one would expect, prices now are more seasonal than before since regulated prices did not change over the season. The good side of seasonal prices is that gas curtailments are less likely. The seasonal pattern of prices also presents new opportunities. As pipelines and local distribution companies bring additional storage on line, seasonal price patterns will be smoothed. The gas futures market will further this process by spreading and shifting the risk of storing gas and by extending agents' ability to contract over the season. The maximum flow between almost all points has improved because open access has created many new paths through the network around former bottlenecks.
The evidence clearly indicates that markets have flourished in the brief period since FERC authorized open access. Markets have come into existence in every field and at most major pipeline interconnections. The volume of gas transported has increased dramatically. Markets succeeded where regulation failed in equalizing gas prices across the geographically dispersed production fields. However, regulation continues to block the full development of markets and competition. The transportation market could be more efficient if transportation rights were fully transferable and subdividable as to injection and withdrawal points (Smith et al., 1990). In its capacity release program of Order 636, the FERC has expanded the transferability of transportation. However, restrictions still remain, and pipelines cannot enter contracts that bundle gas and transportation. These restrictions limit the further evolution of the transportation and natural gas markets. Producers, brokers, utilities, and industrial users should be given the right to acquire firm transportation rights, either by contracting for new capacity or purchasing from a current holder of firm transportation. In light of the success of competition and markets in disciplining prices where regulation failed, policymakers should apply the lessons of open access to distributors and retail markets.
1. Michaels (1993) and Smith et al. (1988) discuss the role of energy price shocks and Federal Energy Regulatory Commission (FERC) rulings in promoting open access. FERC Orders 436 and 500 separated gas pipelines' merchant and transportation functions.
2. Mulherin (1986a, 1986b) shows that the Federal Power Commission created a disincentive for pipelines to integrate vertically into the production fields.
3. Both the Civil Aeronautics Board and the Interstate Commerce Commission chose similarly fragmented structures for the industries they regulated.
4. The District of Columbia Court of Appeals remanded FERC Order 436 for further rulemaking in 1987 but reaffirmed that the FERC had the power to impose open access status on interstate pipelines (Associated Gas Distributors et al. v. FERC, 824 F.2d 981 |D.C. Cir. 1987^). The intent of Order 436 was carried out in Order 500, issued in August 1987. Order 500 also contained a controversial cross-contract crediting scheme (De Canio, 1990).
5. Trading in the natural gas futures contract began on April 3, 1990. Thirty-day contracts for delivery to a pipeline hub in Louisiana are traded up to a year into the future. However, preliminary empirical evidence indicates that the spot and futures prices do converge reliably at contract expiration (Walls, 1993).
6. Because the pipeline's tariff and its price of gas are regulated and are not responsive to demands or capacity constraints, the market price of gas delivered at the maximum tariff may exceed the regulated system gas price of the pipeline in the peak winter heating period. The pipeline's system gas is adversely selected when it is less than the market price.
7. Open access came unevenly over the regions, beginning with applications dated as early as December 1985 and as late as June 1988. The FERC began granting approvals as early as February 1986 and continued as late as November 1988. By late 1988, all regions in the sample were connected via an open access pipeline. Even before applications for open access were approved, all the pipelines had interim transportation programs supporting limited spot market trading (De Vany and Walls, 1993).
8. Pipeline hub-and-spoke networks are forming through pipeline mergers (U.S. Energy Information Administration, 1986). Brokers are able to create hub-and-spoke subnetworks without merging pipelines by combining transportation rights on interconnected pipelines. The technology for interconnecting pipe-lines quickly developed after 1985, so that it now is possible to interconnect lines with different pressures and to change the flow between them. (See Oil and Gas Journal, August 6, 1990, pp. 41-48.)
9. The accessibility matrix becomes less sparse as the number of arcs increases, as the number of paths between vertices increases as a power of the number of arcs, and as the price matrix converges to singularity (De Vany and Walls, 1992a).
10. Stigler and Sherwin (1985) show that the correlation coefficient between two price series is the critical statistic in determining whether the two trading places are in the same geographic market. They argue that price spreads should remain stable even though price levels fluctuate. The correlation coefficient measures the ratio of the variance of the price spread over the variance of the average price.
11. Let |r.sub.xy^ be the sample correlation between x and y, and let |t.sub.xy^ be the population correlation between x and y. Then Z = 1/2 log ((1 + |r.sub.xy^)/(1-|r.sub.xy^)) is approximately normally distributed with expectation E|Z^ = 1/2 ((1 + |t.sub.xy^)/(1 - |t.sub.xy^)) and variance Var|Z^ = |(1/(n - 3)).sup.1/2^ where n is the number of observations used to compute the correlation coefficient (Hays, 1973).
12. Doane and Spulber (1992) also find that price correlations between regions increased after the move to open access transportation even when the analysis controls for the common influences of inflation and the prices of substitute fuels.
13. The experiments revealed the existence of coordination problems, but they did not find failure over time. In Plott's experiments, students eventually were able to achieve 90 percent efficiency while coordinating 19 gas and transportation markets. Further evidence shows that it also took some time (about 4 years) for gas markets to achieve a high degree of integration (De Vany and Walls, 1993).
14. The well-functioning spot market could provide the basis for reviving the long-term contract market. De Vany and Walls (1992a, 1992b) and Walls (1992) show that price shocks are absorbed across the pipeline network within a few days.
Belsley, D., E. Kuh, and R. Welsch, Regression Diagnostics, Identifying Influential Data and Sources of Collinearity, Wiley, New York, 1980.
Bradley, Jr., R. L., Reconsidering the Natural Gas Act, Southern Regulatory Policy Institute, Issue Paper No. 5, 1991.
Daggett, S., Principles of Inland Transportation, Fifth Edition, University of California Press, Berkeley, 1955.
De Canio, S. J., "Cross-Contract Crediting under FERC Order 500," Contemporary Policy Issues, April 1990, 159-175.
De Vany, A. S., and W. D. Walls, "When Barriers to Markets Fall: Pipeline Deregulation, Spot Markets, and the Topology of the Natural Gas Market," Working paper No. 123, Transportation Center, University of California, Berkeley, 1992a.
-----, "Open Access, Local Bypass, and the Structure of Natural Gas Production Area and City Gate Prices," Working paper, University of California, Irvine, 1992b.
-----, "Pipeline Access and Market Integration in the Natural Gas Industry: Evidence from Cointegration Tests," The Energy Journal, 14:4, 1993, 1-19.
Doane, M. J., and D. F. Spulber, "Open Access and the Evolution of the U.S. Spot Market for Gas," Invited paper, International Association for Energy Economics, 14th Annual North American Conference, 1992.
Federal Energy Regulatory Commission, "Importance of Market Centers," Office of Economic Policy, August 21, 1991a.
-----, "Notice of Proposed Rulemaking," Docket No. RM 91-11-000, 1991b.
Gas Daily, Pasha Publications, Arlington, Virginia, various issues.
Hayes, W. L., Statistics for the Social Sciences, Holt, Rinehart and Winston, New York, 1973.
MacAvoy, P., and R. Pindyck, The Economics of the Natural Gas Shortage (1960-1980), North-Holland, New York, 1975.
McCabe, K., S. S. Rassenti, and V. L. Smith, "Experimental Research on Deregulation, Natural Gas Pipeline and Electric Power Transmission," Working paper 90-21, Economic Science Laboratory, University of Arizona, 1990.
Michaels, R. J., "The New Age of Natural Gas: How Regulators Brought Competition," Regulation, 16:1, 1993, 68-79.
Mulherin, J. H., "Specialized Assets, Governmental Regulation, and Organizational Structure in the Natural Gas Industry," Journal of Institutional and Theoretical Economics, Vol. 142, 1986a, 528-541.
-----, "Complexity in Long Term Natural Gas Contracts: An Analysis of Natural Gas Contractual Provisions," Journal of Law, Economics & Organization, Vol. 2, 1986b, 105-117.
Oil and Gas Journal, "New U.S. Gas Lines Will Restructure North American Grid Flows," December 10, 1990, 33-39.
Plott, C., "Research on Pricing in a Gas Transportation Network," Federal Energy Regulatory Commission, Office of Economic Policy, Technical Report No. 88-2, July 1988.
Rassenti, S. J., S. S. Reynolds, and V. L. Smith, "Cotenancy and Competition in an Experimental Auction Market for Natural Gas Pipeline Networks," Working paper, University of Arizona, 1988 (forthcoming in Economic Theory).
Smith, R. T., A. S. De Vany, and R. J. Michaels, "An Open Access Rights System for Natural Gas Pipelines," in Interstate Natural Gas Rate Design Studies, Natural Gas Supply Association, Washington, D.C., 1988.
-----, "Defining a Right of Access to Interstate Natural Gas Pipelines," Contemporary Policy Issues, April 1990, 142-158.
Stigler, G. J., and R. A. Sherwin, "The Extent of the Market," Journal of Law and Economics, Vol. 28, 1985, 555-585.
Teece, D. J., "Structure and Organization of the Natural Gas Industry: Differences between the U.S. and the Federal Republic of Germany and Implications for the Carrier Status of Pipelines," The Energy Journal, 11:3, 1990, 1-36.
U.S. Energy Information Administration, "Pipeline Mergers and their Potential Impact on Natural Gas Markets," Natural Gas Monthly, DOE/EIA-0525. U.S. Government Printing Office, Washington, D.C., 1986.
-----, Growth in Unbundled Natural Gas Transportation Services: 1982-1987, DOE/EIA-0525, U.S. Government Printing Office, Washington D.C., 1989.
-----, Natural Gas Monthly, U.S. Government Printing Office, Washington, D.C., various issues.
Walls, W. D., Open Access Transportation, Network Competition, and Market Integration in the Natural Gas Industry, Ph.D. dissertation, University of California, 1992.
-----, "A Note on the Natural Gas Futures Market," Pacific and Asian Journal of Energy, December 1993, 221-228.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Open Access and the Emergence of a Competitive Natural Gas Market. Contributors: Vany, Arthur de - Author, Walls, W. David - Author. Journal title: Contemporary Economic Policy. Volume: 12. Issue: 2 Publication date: April 1994. Page number: 77+. © 2003 Western Economic Association International. COPYRIGHT 1994 Gale Group.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.