The Triumph of Markets in Natural Gas

Fortnightly Magazine - April 15 1995

During the last decade, the natural gas industry in the United States has been transformed from a heavily regulated business to one facing competitive markets. This transformation grew out of the failure of regulation; regulators, suppliers, pipelines, and customers all played a part. It continues today as the industry restructures and builds new institutions.A series of regulatory crises forced deregulation in stages: First, wellhead prices; second, gas contracts; and finally, pipeline transportation. As markets responded, the Federal Energy Regulatory Commission (FERC) and the state regulatory commissions were forced to change. The growing efficiency of markets demanded better access to services. "Bypass," that quaint industry term for access to markets, emerged as a goal for gas users and suppliers. Regulators answered with the now-famous series of "open access" orders permitting shippers to bypass pipelines as sellers of gas and letting pipelines elect to transport gas under contract. By about 1985, enough pipelines had opened up that gas buyers and sellers could deal directly with one another and make arrangements to transport gas with assurance. This development marked the beginning of the modern gas market.

By 1989, a mature form of competition had come to natural gas. Enough pipelines had opened their systems to form a pipeline network. Markets had evolved far enough to coordinate gas and transmission trading. The gas market had gained the broad participation of buyers and sellers, giving it the depth and liquidity characteristic of a competitive market. The gas pipeline industry is no longer a natural monopoly.

How did it all happen?

A way of thinking . . .

Gas flows from wells located in distant producing fields, through pipelines to users. The interstate pipelines end at state borders or at gateways to urban markets (the city gate), where gas is transferred to a distribution system for delivery to consumers. With certain minor exceptions, form follows function. Pipelines do not produce gas. Nor do producers own pipelines, beyond local connections to the trunk line. Distributors do not own pipelines or produce gas. But these functions could have been organized differently.

As it was, the gas industry splintered into segments to fit the jurisdictional boundaries that divided regulatory agencies. Regulatory policy barred shipper and producer joint-venture pipelines. Linear supply chains arose from production areas to city gates, largely as a consequence of federal certification of pipelines. That scheme required dedicated supplies and long-term contracts between pipelines, upstream suppliers, and downstream buyers. But certification of pipelines created a protected monopoly from the fields to the city gates. Pipelines operated independently of one another, each supplying cities with dedicated gas volumes. Regulators balkanized the industry into minute pieces and institutionalized monopoly.

that fell apart . . .

Interstate natural gas pipelines have long been considered natural monopolies, marked by economies of scale in size and output and making competition wasteful. With function following form, pipelines were built to serve a theory and were regulated as the theory prescribes: A monopoly is chartered to serve the market, prices are controlled, entry is closed, and cost-based rates are set. The state acts as central coordinator and planner; the regulated monopoly as its agent. Competition, it is argued, is not suitable for pipelines.

But this diagnosis is wrong, and the prescription faulty. It is wrong on several key points: It fails to consider that a pipeline forms but an element in a network. It assumes that the pipeline is organized noncompetitively. It misrepresents how coordination is achieved in a complex system. It lies at variance with how regulation works in practice. It badly misrepresents how markets work.

Pipelines are not natural monopolies, in spite of what the proponents of regulation have told us. They can be organized competitively. They can be stripped of monopoly power in a network of interconnected pipelines that offers many distinct paths between markets. Both these things can be accomplished by making transmission an asset that can be traded in a market open to producers, distributors, customers, brokers, and others.

when markets emerged . . .

Gas prices fell after they were deregulated. (This result has proved true for all the cases of deregulation we know of.) Thereafter, pipelines faced infeasible 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. Open gas markets began.

The FERC approved these transportation transactions individually until it issued Order 436, permitting interstate pipelines to transport gas and separating the pipeline merchant and transportation functions. 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.

Open-access transportation gave customers and suppliers many alternative trading partners. Because these partners could arrange their deals themselves, regulators no longer stood as a barrier to mutually beneficial exchanges. After gas traders had made their purchase arrangements, they could bring the deal to the pipeline who would transport the gas. Within a few years of the first open-access order, so many pipelines were open that buyers and sellers could come together through many paths in the network. These paths gave structure to markets. Where a path between markets is open, prices will reflect this openness. When there are many paths, prices are constrained by the wide number of alternate sources of supplies and buyers that can be reached from any point in the network.

and made prices competitive . . .

Competitive prices tend to move together. Consider how the two prices in Figure 1 behave over time. The price difference narrows over time; eventually they become almost correlated. These two properties (em initial narrowing, then eventual correlation (em are integral to competitive markets. Our research looks at prices in the spot market for evidence of these properties.

Throughout a network of markets, competitive prices should move together within a band related to transportation costs, so that price differences within bands do not become so large that a profit can be made by arbitrage. Moreover, it should be impossible to use information about a given price change in a spot market somewhere within the network to predict a price change later at another point in the network. Instead, every price will reflect the same information. Note in Figure 1 shows how the changes in these two prices coincide, and how narrow, over time, becomes the band within which the two prices move.

Our evidence shows that this convergence of prices took place all over as open access emerged and opened alternative network paths for transmitting and trading gas. New interconnections at pipeline hubs hastened the process, as did the creation of sophisticated gas trading techniques.

whether measured systemwide . . .

Our first look at markets considers the system of spot prices over the entire pipeline network. We wanted to test whether the "very strong form" of competitive pricing held. In this form, prices are so closely tied together that every price reflects the same information, making it impossible to exploit information about a prior price change anywhere within the system. (See sidebar "A Competitive Model," on p. 23.)

We tested this "strong" hypothesis (em that prices are competitive in every spot market simultaneously over various networks. To do so, we analyzed portions of pipelines that form parts of networks connecting supply basins. The supply basins and pipelines forming these networks are listed in Table 1.

During the period from July 1987 to June 1988, none of the six networks reveals "strong" competitive pricing. During the same period a year later, from July 1988 through June 1989, prices in the five-pipeline Network 6 in Oklahoma became "strongly" competitive. But after one more year, from July 1989 through June 1990, prices had become strongly competitive for all networks but Network 3. Prices in Networks 1, 2, 4, 5, and 6 all reflected the same information; no opportunities arose for profitable arbitrage in these market networks. By 1989 most pipelines had opened their systems to transportation. The number of paths in these networks expanded to bring about strongly competitive prices.

We also found that the competitive spot prices converged rapidly. To gain some understanding of the speed of price convergence, we simulated price shocks on Network 1 equal to one standard deviation, and measured how long it would take before the market damped the shock. We found that the damping period shrunk from seven days in 1988 to three days by 1990.

or between pairs of markets . . .

What about direct competition between any two specific markets? To examine competition between market pairs, we measured the cost of arbitrage.

Part of the cost of arbitrage is transportation. Thus, in a simplified example, if the price in market i increases by one unit, while the cost of transmission remains constant, then the price in market j should rise by the same amount as in market i to restore equilibrium. In other words, if the cost of arbitrage remains relatively constant (represented here by the cost of transmission) then the price difference between two spot markets should also remain relatively constant over time.

We found that market integration drifted upward over time in the production areas. By 1991, more than 65 percent of the production-area spot markets had become cointegrated. The level of cointegration did not depend upon the number of pipelines in a market. You can't measure market competitiveness by counting pipelines; you must look at prices.

as shown at hubs, city gates, and futures markets.

Next, we looked at market centers. Market centers are the points where pipeline systems intersect in a radial pattern of spokes through a hub. The significance of these hubs is that pipelines are close enough to be connected readily. This means that new links expand the number of arbitrage paths between markets. (In our forthcoming book, we show that the number of paths expands as a power of the links in the system; thus, opening a hub greatly expands the number of paths in the network.) By 1990, some of the hubs connected more than a hundred markets. We considered three networks: those centered on the Northern Town Border Station, on Maumee, OH, and on Broad Run, WV. Each of these pooling points offers 114 arbitrage paths. Our tests revealed competitive pricing in these networks.

We also examined prices at city gates. We asked how closely these prices are linked to prices in the production fields and hubs. We found that a few city gates are well integrated with production-area markets, but many are not. City-gate prices are in line with prices in the fields, but do not stay within the narrow limits one finds between the fields and market centers. It is not distance alone that makes the difference (em very distant market centers are highly integrated with production fields, and their prices track one another closely. Instead, it is access through the city gate that matters, and it turns out that the city gate is not yet open far enough. City-gate markets need have little fear their prices will be too strongly influenced by local conditions, with the exception, perhaps, of a few capacity-constrained city gates like Southern California. Price volatility is damped by links to the pool of gas throughout the United States and Canada.

Lastly, we found that most field and hub prices are highly cointegrated with the futures price. However, the futures contract offers a less satisfactory device for hedging price risk at some locations. The futures market is unarguably competitive. So, any price that is cointegrated with the futures price must also be competitive. Most markets we examined met this condition, leading one to conclude that their prices are competitive. t

This article is based on the forthcoming book The Emerging New Order in Natural Gas: Market versus Regulation (Quorum Books, Westport, CT). Arthur De Vany is a professor in the Department of Economics and the Institute for Mathematical Behavioral Sciences at the University of California, Irvine. David Walls is an assistant professor in the School of Economics and Finance, and a Fellow at the Centre for Urban Planning and Environmental Management at the University of Hong Kong.


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.