LexisNexis(TM) = Academic

Spring, 1995

16 Energy L. J. 299 =

LENGTH: 29913=20 words

ARTICLE: MARKET-BASED RATES FOR INTERSTATE GAS = PIPELINES:=20 THE RELEVANT MARKET AND THE REAL MARKET

Robert J. = Michaels *,=20 and Arthur S. De Vany **

* Professor of Economics, = California=20 State University, Fullerton, and Consultant, JurEcon, Inc., Los Angeles; = A.B.=20 University of Chicago, 1965; Ph.D. 1972, University of California, Los = Angeles.=20

** Professor of Economics, University of California, Irvine; = Ph.D. 1970,=20 University of California, Los Angeles. The authors have consulted and = prepared=20 expert testimony on topics covered in this article. The views expressed = here are=20 not necessarily those of their clients. The authors thank Robert L. = Bradley, Jr.=20 and Richard P. O'Neill for their helpful comments.
=20

SUMMARY:
... This norm is often = without=20 operational content and may often be incorrect as well. ... Open access = rules=20 may prevent them from withholding capacity from customers, in which case = a high=20 HHI will overstate market power. ... Electric utility mergers and power=20 marketing plans prior to the Energy Policy Act of 1992 (EPAct) make up = the=20 second class of FERC proceedings that use concentration statistics to = evaluate=20 competition. ... Open access means that both entitlement holders and = outsiders=20 can exchange capacity on a pipeline that links two = areas, and=20 that both can bid for unused space that reverts to the = pipeline=20 for mandatory offer as interruptible service. ... Under open access, all = points=20 in the network may be in the same market. ... The dynamics of market = reaction=20 also changed with open access. ... Whether observed at the producing = area, the=20 market center, the city gate, or the futures market, gas prices provide = abundant=20 evidence that a highly competitive market has arisen with open access. = ... We=20 have found no FERC proceedings in which shippers have alleged that an=20 open-access pipeline has willfully withheld capacity or = interruptible service in quantities that are significant by antitrust = standards.=20 ... Under open access, long-term requirements contracts become the = exception.=20 ...

TEXT:
[*299] =

I.=20 Introduction and Purpose

Under a cost-of-service regime, = regulators can=20 set a utility's rates without any explicit reference to markets. n1 = Employing=20 data on costs the utility expects to incur or has incurred, regulators = use=20 economically dubious formulas to allocate these costs to individual = services and=20 to set rates that recover the reported amounts. n2 Generally, regulators = need=20 not make conjectures about the products and terms of service the utility = would=20 offer if regulation vanished, or about the speed at which competitors = might=20 enter an unregulated market. More important, regulators need not base = any of=20 their decisions on estimates of the utility's prices in an unregulated = market.=20 An oft-asserted norm asks regulators to set prices and outputs at the = levels=20 that would arise in a competitive market. n3 This norm is often without=20 operational content and may often be incorrect as well. Perfect = competition=20 would usually be an inefficient as well as infeasible arrangement in = industries=20 where individual sellers enjoy substantial economies of scale. =

We will=20 only know competitive prices when competition arrives. A regulated price = set at=20 historical cost will only by accident also be the unregulated price in a = competitive market. Economics provide no expectations that a deregulated = competitive producer will have the same earnings as it would under even = the most=20 competent cost-based regulation. It is a fundamental error to confuse = the costs=20 on which buyers and sellers make market decisions with the costs on = which=20 regulators must base their decisions. Because historical costs bear no = necessary=20 relation to today's opportunities, they are irrelevant as guides for=20 economically rational decisions. Market actors look forward, because = future=20 costs are the only ones that are avoida- [*300] ble by = changing=20 today's decisions. Regulators look backward, calculating what must be = collected=20 today in order to recover past outlays. n4

The gas industry's = past=20 provides one outstanding example of the importance of disregarding = historical=20 costs. Hardly anyone today applauds either the efficiency or the = equitability of=20 the Federal Power Commission's (FPC) regulation of wellhead prices = between the=20 1950s and the 1970s. n5 Having been ordered to set maximum wellhead = prices, the=20 FPC used established cost-of-service techniques. n6 The FPC examined the = costs a=20 producer had incurred on a well or field (including the cost of = replacing=20 exhausted supplies) and then set a gas price that would recover those = costs. By=20 setting the price to recover recorded production and exploration costs, = the FPC=20 succeeded in creating an interstate shortage. The misallocations that = stemmed=20 from the shortages were surely more inefficient than whatever = imperfections=20 existed in a wellhead market that most observers agreed was competitive. = n7=20 Purchasers with access to controlled gas used it more than they would = have if=20 faced with the opportunity costs embodied in market prices. Producers, = whose=20 resources could have been valuably employed in exploration, allocated = them=20 elsewhere. Prospective users wasted resources in attempts to influence = politics=20 and regulations to obtain priority for allocations resulting from the = shortage.=20

In retrospect it is easy to discern the regulatory mistakes that = caused=20 the wellhead market to malfunction under price controls, and to = understand the=20 desirability of "market-based" field prices. By forcing prices to equal = a figure=20 that they called cost, regulators produced an outcome that was far from=20 competitive. Today, the FERC uses historical cost to set interstate=20 pipeline rates, but allows pipelines = with=20 excess capacity to discount below cost-based ceilings. If rate = regulation is=20 abandoned, the price of a given service will equal today's regulated = price only=20 by coincidence, regardless of whether the new market is competitive or=20 monopolized. Prices of services that are in short supply relative to = demand=20 (possibly because of regulatory policy) will rise to levels above booked = costs=20 and prices of those that are being overproduced under regulation will = fall below=20 those costs.

One can evaluate a market's performance only with = reference=20 to a well-specified alternative. Perfect regulation is not a reasonable=20 alternative to an imperfect market, and vice versa. A market may fail to = perform=20 competitively because an unregulated monopolist seeking private benefits = controls the price in it. The market may also fail because a regulatory=20 commission with the best of intentions sets economically inefficient = prices=20 [*301] based on historical cost. The relevant comparison is = between=20 imperfect markets and imperfect regulation. Comparing imperfections is = important=20 because markets and regulation have differing dynamics: An inefficiency = in a=20 market invites someone to profit by removing it, as when a new = competitor goes=20 up against an incumbent monopolist. If the wellhead price control = example is=20 general, an inefficiency in regulation is more likely to encourage = parties to=20 seek transfers of wealth from one another than to seek profits by = proposing an=20 efficient alternative.

II. Economic and Antitrust = Markets=20

A. Some Economics

In an economic market, = potential=20 buyers and sellers compare their valuations of a good in order to make = the=20 exchanges that best advance their individual interests. n8 Markets lower = the=20 cost of acquiring information because they are centralized points for = comparing=20 the offers of numerous possible trading partners. A market may be a = location=20 like a commodity exchange, or it may be a communications network such as = traders=20 use for spot gas. As people compare offers, "prices of the same goods = tend to=20 equality with due allowance for transportation costs." n9 The limiting = case of a=20 "perfectly competitive" market faces all participants with the same = price. For=20 all but the most standardized of commodities, however, perfect = competition is at=20 variance with the complexity of actual trading. One can better = characterize=20 participants in real-world markets as forming contracts rather than = trading=20 goods. The contracts they arrive at will vary with their individual = situations,=20 including their expectations and their available information. In a = contract=20 market, interpreting the convergence of price to one value as = "perfection" may=20 be quite misleading. There may simply be too great a diversity of = possible=20 services and contract terms to reach a uniform equilibrium.

By = an=20 economic standard, the market that matters for natural gas policy is a = broad=20 market for contracts. In Section IV below, we provide data which = indicates that=20 the gas market is becoming more competitive as institutions that = facilitate=20 transaction developments, and as open access pipelines = afford=20 more trading opportunities over an interconnected network. In this = market, gas=20 flows and contract terms adjust to eliminate opportunities to profit by=20 arbitraging price differences. The growing connectedness of the market = broadens=20 the range of potential trades a buyer or seller can execute. An agent = with more=20 options cannot be worse off than with only a subset of them, and that = agent may=20 find opportunities in the larger set which are superior to those in the = smaller=20 subset. In economic jargon, the efficiency of the market increases as = new trades=20 become feasible. Holding the same collective resources as before, buyers = and=20 sellers can make mutually advantageous trades that were infeasible when = the=20 market was narrower. A less efficient market contains more obstacles to=20 beneficial [*302] exchanges. Those obstacles might be = informational=20 (not knowing that a trading partner exists), or physical (no=20 pipeline links the buyer and seller). They might also = stem from=20 monopoly or regulation.

By withholding capacity, a=20 pipeline monopolist decreases market efficiency, since=20 withholding capacity drives a price so high that gainful exchanges go = unmade and=20 usable capacity goes unused. The economic objection to monopoly is = unmade=20 exchanges, not high profits. n10 A regulated price which does not equal = the=20 competitive price can likewise render efficient exchanges impossible. = For=20 example, a ceiling price that is set too low will discourage = economically=20 warranted investments in supply and may ration the good to buyers who do = not=20 value it highly. Regulation may give rise to other inefficiencies, for = instance,=20 when a seller who is insulated from competition loses the incentive to=20 effectively monitor its costs and thereby wastes scarce resources. The = economic=20 debate over deregulation is largely over a factual matter: Do the = misallocative=20 effects of unregulated monopoly (or monopoly subject to antitrust) = exceed the=20 misallocative effects of regulation? n11

B. Some = Antitrust=20

1. The Guidelines' Standards and the Clayton Act=20

Antitrust deals with the anticompetitive effects of the = acquisition or=20 exercise of market power. Its doctrines and analytical methods are the=20 foundation for the FERC's policies toward competition. In Market-Based = Rates=20 (MBR) proceedings, the FERC often employs the same tools that the = Department of=20 Justice (DOJ) and the Federal Trade Commission (FTC) use to examine = horizontal=20 mergers under the Clayton Act. n12 This law calls on the government to = halt=20 monopolization in its incipiency, specifically by prohibiting mergers = and=20 acquisitions which "may ... substantially ... lessen competition," "in = any line=20 of commerce ... in any section of the country." n13 To fulfill that = mandate, the=20 antitrust agencies have produced Merger Guidelines outlining their = preferred=20 economic model and the numerical thresholds that will bring a merger = into=20 question. n14 The process [*303] begins with the definition = of a=20 "relevant market" in which the merger may affect = competition,=20 and then examines the merger's effect on seller concentration and price = in that=20 market. n15 In practice the Merger Guidelines are a preliminary screen. = Since=20 economics offers few unambiguous standards for market delineation, the=20 authorities invariably go beyond the numerical standards of the Merger=20 Guidelines to examine institutional details of the market before = deciding the=20 legality of a suspect merger.

Their screening decision can go = wrong in=20 two ways: it may allow an anticompetitive merger to go forward, or it = may=20 prohibit a procompetitive merger. The Merger Guidelines appear to = encourage=20 erring on the side of caution. n16 Specifically, they ask if the merger = is=20 likely to produce a "small but significant and nontransitory" price = increase in=20 the relevant market. n17 The Merger Guidelines define = this as a=20 5% increase that persists for a year. n18 The economic standard for a = beneficial=20 merger is that the cost savings to the merging parties exceed the = allocative=20 inefficiency that results if post-merger market output falls. n19 By the = economic standard, the Merger Guidelines' price increase rule can lead = to=20 incorrect decisions: A merger might produce great cost savings but still = lead to=20 an increase in price. Despite the differences, the economic standard and = the=20 Merger Guidelines' standard share an important attribute: There is no = known=20 method that reliably predicts whether a proposed merger will be = beneficial under=20 either of them.

The "relevant market" of the = Merger=20 Guidelines is the smallest group of products (and producers) which if = controlled=20 or coordinated by a single entity could profitably impose a significant = and=20 nontransitory price increase. n20 The price increase standard errs on = the side=20 of pessimism. Sellers in the post-merger market (including non-merging = firms)=20 might form an overt or a tacit collusion that could act like such a = monopolist,=20 but this need not be the outcome. Beyond the threat of antitrust, the = colluding=20 parties must be able to deter members from price-shading that is = individually=20 profitable but decreases the group's profit. Without such enforcement, = the=20 collusion deteriorates into competition. More importantly, collusion is = only=20 [*304] one of many possible paths the market might take. A = market=20 that has experienced mergers might with equal plausibility be more = competitive=20 than it was prior to the mergers. The enlarged firms in the new market = may have=20 the resources and the risk-bearing ability to compete more vigorously = and=20 innovate more extensively than when they were small. Depending on the = freedom to=20 act as individual sellers and the expectations that those sellers have = of one=20 another's responses, almost any outcome is possible.

2. = Bounding=20 the Market - Product and Geography

Sellers in a relevant = market that includes good X have power over price only if = buyers have=20 few choices other than living with high prices. If buyers can easily = purchase=20 close substitutes for X at attractive prices, the relevant=20 market for a merger of two producers of X contains the two = producers=20 themselves, other producers of X, and producers of the substitutes. If = consumers=20 can substitute for good X, a monopoly in X alone is not worth having = because the=20 monopolist cannot exert power over the price. To successfully impose the = price=20 increase, producer(s) of X must combine or collude with producers of the = substitutes. Economic theory suggests the cross-elasticity of demand as = a=20 measure of interproduct substitutability by buyers. n21 Because = antitrust=20 enforcers generally lack the data needed to accurately compute = cross-elasticity,=20 qualitative judgments must usually supplement any calculations. Sources = of such=20 judgments can include documents (Whose prices does X look at when = deciding on=20 its own price?) or statistical observations (Do many = pipeline=20 users move between firm and interruptible service as their relative = prices=20 vary?). Statements by buyers about how they will react to price changes = often=20 bear no relation to their actual behavior when changes occur.=20

Substitution from the supply side also constrains a producer's = or=20 group's ability to raise prices. When buyers substitute against newly=20 monopolized good X, they increase the profitability of producing = substitutes for=20 it. If producers of substitutes respond quickly and substantially to the = increase in X's price, they are in the relevant market = along=20 with the producer of X. The responsiveness of producers of substitutes = lowers=20 the likelihood that the producer of X can impose the requisite price = increase in=20 the broader market. As with substitution by buyers, lack of data with = which to=20 predict producer responses necessitates the introduction of less = quantitative=20 judgments. To estimate supply substitutions, the economist must identify = existing producers, including those who do not currently produce the = substitute,=20 who will respond to the monopolization of X by increasing output = substantially.=20 Identifying such entrants may be highly conjectural, although history = might=20 provide examples of the sources of increased output following similar = market=20 changes. [*305]

If competition is not to be harmed = anywhere,=20 an analyst must also examine the market's geography. Monopoly power = exists=20 within an area if the producers located there can hypothetically act to = impose a=20 price increase of the requisite size and duration. n22 Their power = depends on=20 the availability of economic substitutes. Those substitutes might be = shipments=20 from outsiders who are induced to move goods into the area when price in = it=20 increases. The substitutes may also be produced locally by firms who can = switch=20 outputs, or by newly formed firms. The market's geographic scope then = depends on=20 elasticities of supply inside and outside of the region, and on the = costs of=20 moving the good over the boundary. n23 One frequently cited (and vague) = standard=20 requires that little of the relevant good moves into the area from = outside, and=20 little of it moves out of the area from inside. n24 There may be no = reasonable=20 product flow bounds on a geographic market, as might happen when the=20 availability of imports to substitute for domestic production creates a=20 worldwide market. Geographic markets are often important in analyzing = the=20 options of those who must use an "essential facility," such as a=20 pipeline that delivers to a certain city gate.=20

C. Market Concentration and Market Power

The = theoretical=20 monopolist who takes over a competitive market will lower output, raise = the=20 price, and possibly exclude sellers who wish to enter and compete in it. = n25=20 Since such clear monopolizations seldom occur, the antitrust agencies = typically=20 face harder questions. They must, for example, decide whether a = questioned=20 merger, short of a monopoly, will significantly increase the = price-setting power=20 of the merging firms, and analyze how the merged firm's competitors = (possibly=20 including new entrants) will respond. Economics provides numerous = theoretical=20 models of competitive responses in situations where sellers have some = market=20 power. n26 Not one of those models, however, is a clear conceptual or = empirical=20 winner for determining the effects of a merger on price in the market = where the=20 merging parties operate. The lack of a dispositive paradigm also = confounds the=20 comparison of similar markets with different supplier structures, e.g., = an=20 origin/destination pair linked by two pipelines with a = similar=20 pair connected by three of them. [*306]

To see the = possible=20 range of market outcomes under differing supplier structures, it is = helpful to=20 first characterize seller concentration. A monopolist has a market share = of=20 100%. n27 In a perfectly competitive market, each seller is a = price-taker, so=20 small that its own production decision cannot affect market price. The = dominant=20 firm with a competitive fringe has a market share determined in part by = the size=20 of that fringe and its aggregate response to the dominant firm's = decisions. n28=20 When several sellers are large enough to affect market price, economists = often=20 summarize their relative sizes in a Herfindahl-Hirschman index (HHI). = n29 The=20 HHI is the sum of squares of all sellers' market shares, expressed as = decimals.=20 A market whose two sellers control 80% and 20% shares has HHI =3D = .80[su'2'] +=20 .20[su'2'] =3D 0.68, and a market with N equally-sized firms has HHI =3D = 1/N. In a=20 monopoly market, HHI =3D 1, and as the number of equally-sized sellers = increases=20 without limit, HHI approaches zero. This limit rationalizes its use as a = measure=20 of competitiveness. Further, the HHI gives disproportionate weight to = sellers=20 with higher market shares, consistent with a possible intuition that the = total=20 market power of two large and equally-sized firms will more than double = if they=20 merge. Currently, the authorities usually approve mergers in markets = whose=20 post-merger HHI will remain less than 0.18, unless special circumstances = prompt=20 further investigation. n30

The HHI alone is an insufficient = basis for=20 inference about the effects of market concentration on price or profits. = An=20 economist also requires information about how sellers are expected to = respond to=20 one another's decisions. If one seller chooses to change output or = price, the=20 best responses of other sellers will often entail changes of their own. = The=20 Merger Guidelines seek to determine whether after a merger the aggregate = of=20 these responses yields a new aggregate output which is small enough to = be=20 saleable at five percent above the pre-merger price. Each seller's = decision=20 depends on its beliefs about how rivals will respond, as summarized in = the=20 economic concept of "conjectural variation." n31 To explicate = conjectural=20 variation, it is possible to derive a simplified theoretical = relationship that=20 links a market's price-cost margin, the elasticity of market demand, the = HHI,=20 and the conjectures sellers hold about one another. n32 Let all sellers = in the=20 market have the same constant unit costs, c, and let each hold the same=20 conjectures about the reactions of others, measured by a positive=20 [*307] number b (see below). Elasticity of demand in the = market is=20 denoted e, and there are no actual or potential suppliers outside of = this=20 market. n33 If each seller seeks to maximize profit, the percentage = margin of=20 market price (P) over cost is given by

(P - c)/P =3D Hb/e. =

If=20 other factors are unchanged in the above expression, the price-cost = margin in=20 the market is greater, the higher is the HHI and the lower the = elasticity of=20 demand. If the values of other terms are known, one can compare = pre-merger and=20 post-merger market margins (e.g., against a five percent standard) by=20 calculating them for the pre-merger and post-merger values of H. =

The=20 merger's effects depend critically on the conjectural term b. The = economically=20 meaningful limits for b lie between those of perfect competition and = perfect=20 collusion. It can be shown that if b =3D 0, sellers are acting as if = they were=20 price-taking perfect competitors, even though each individually is large = enough=20 to influence price. The industry's margin over cost is zero, and a = merger=20 between two sellers would not raise profits or prices. n34 No matter how = many or=20 how few sellers there are in the market, a merger between two of them = will be of=20 no economic consequence if they hold such beliefs about one another. At = the=20 other limit, if b =3D 1/H for all sellers, they are behaving like the = members of a=20 perfect collusion, sharing joint profits equal to those of a monopolist = who=20 controlled this market's entire supply. n35 A merger of two colluding = sellers=20 will not affect the market's monopoly-level price-cost margin. For any=20 conjecture between these extremes, a merger short of monopoly will = affect H, and=20 hence market price. In a theoretically important case, if b =3D 1, each = seller is=20 said to hold "Cournot-Nash" conjectures about the others. n36 Each = Cournot-Nash=20 seller chooses output in the expectation that others will not change = their=20 outputs in response. The other sellers may in fact respond, but the = Cournot-Nash=20 market generally settles at an equilibrium whose output and price lie = between=20 those of perfect competition and pure monopoly or collusion. In a = Cournot-Nash=20 market, total output approaches the perfectly competitive level as the = number of=20 sellers increases.

While the algebra provides some insights, = data=20 requirements make it difficult or impossible in practice to predict the = effects=20 of a merger. First, the algebra presupposes that the market has been = correctly=20 defined. Using different products or geography will affect the spectrum = of=20 substitutes, and hence the elasticity of demand on which the price-cost = margin=20 depends. [*308] Secondly, if sellers' costs differ or if = those costs=20 vary with output, inferences are more complicated. Third, conjectures = are an=20 empirical mystery. Economists have made only a handful of attempts to = measure=20 them in specific industries, and there is little agreement among the = estimated=20 values. If conjectures are not known, the seeming precision of an = algebraic=20 analysis of market power deteriorates into guesswork. Analytical = simplicity=20 makes Cournot-Nash conjectures a popular theoretical assumption, but = there are=20 no persuasive reasons to expect that real-world producers hold such = beliefs=20 about one another. n37 The modern economic literature has thrown up a = plethora=20 of models for oligopolies, while providing little guidance for empirical = predictions of industry behavior. n38 Moreover, competitive strategies = involve=20 choices and reactions on numerous dimensions, including product = characteristics,=20 discounting policy, distribution methods, inventories, capital = investment, and=20 the vertical scope of operations.

Inferences from the algebra = also=20 require an assumption that the number of sellers in the market is fixed. = n39=20 Entry that occurs in response to the higher price engendered by a merger = will=20 mitigate the merger's adverse effects. Empirically, it is generally = difficult to=20 estimate the likely volume of entry. Even if entry is estimable, = however, there=20 is no generally accepted economic theory of how incumbent sellers will = behave=20 when faced with it. n40 A producer in the market model that underlies = the Merger=20 Guidelines can further choose an output level at which some of its = capacity=20 remains idle, if doing so is most profitable. If a producer must offer = all of=20 its existing capacity to buyers, as is the case for open-access=20 pipelines, monopolistic outcomes that result from = restriction=20 of output become unlikely.

Perhaps the most problematic aspect = of the=20 HHI approach is its empirical irrelevance. Economists have failed to = find any=20 evidence of a critical HHI above which one can reasonably suspect that = overt or=20 tacit collusion will arise. n41 If there is no critical HHI that breaks = markets=20 into those with high and low likelihoods of collusion, rational merger = policy=20 (and pipeline MBR policy), cannot employ the HHI even = as a=20 preliminary screen. Using the HHI as a screen is equivalent to a=20 once-and-for-all random draw that chose 0.18 as a dividing line. Nor can = the HHI=20 be saved by appealing to a prior generation of econometric studies that = attempt=20 to relate profits and [*309] seller concentration. n42 Even = if such=20 statistical relationships are significant, there are equally plausible = models in=20 which efficiently functioning markets, rather than collusions, give rise = to a=20 positive relationship between concentration and profits. n43 =

D.=20 Regulation and Concentration

In unregulated industries, some = models of=20 seller behavior (e.g., Cournot-Nash) conclude that a market with a lower = HHI=20 performs less monopolistically than an equivalent one with a higher HHI. = In=20 regulated industries, few if any models produce such a conclusion. n44 = The=20 regulated firm may be a natural monopoly whose cost of serving a typical = customer falls as the firm grows. If competition is imposed on such a = market,=20 none of the competitors will enjoy costs as low as those of a single = large=20 server. Here, economic efficiency falls as market concentration falls.=20 Regulation may be necessary to ensure that a single seller serves all = who are=20 willing to pay its cost of serving them. n45 Whether a single server is=20 regulated competently, poorly, or not at all, however, the HHI in its = territory=20 for its product will equal 1.0 and be unrelated to the quality of the = server's=20 performance. A regulated utility with an obligation to serve cannot = restrict=20 output to earn supernormal profits. n46 If regulators require it to = serve some=20 customers at rates that do not recover cost, its 100% market share = indicates a=20 lack of market power. Competition to provide a regulated service can = only occur=20 with the approval of legislators and regulators. n47 They may constrain=20 [*310] competition, sometimes for the best of reasons, by = regulating=20 price, output, service offerings, quality, territory, capital = investment, and=20 the entry of competing servers. Regulation of these dimensions of = service can=20 also affect the HHI in the regulated firm's markets.

If = regulated=20 entities can compete for the right to offer service, measures of = concentration=20 are doubtful indicators of competition or its desirability. n48 Most = cities, for=20 example, can choose between corporate and municipal gas and electric=20 distribution utilities. n49 If only one entity can hold the franchise = and the=20 city limits define the geographic market, whoever serves at the moment = has a=20 100% share of it. That share tells nothing about the server's efficiency = or=20 about the strength of potential competition for the franchise. = Alternatively, a=20 city may be surrounded by a large corporate utility whose territory is = deemed to=20 define the market. If the city cannot serve outside of its limits, its = share of=20 sales in this market can never exceed a few percent, regardless of its=20 competitiveness. In markets for utility services, competitors may = operate under=20 substantially different constraints. An unregulated municipal utility = may be=20 able to price its services more aggressively than a regulated corporate = system.=20 The corporate utility may have "supplier of last resort" obligations = that=20 require it to serve loads in a municipal utility's territory that the = city finds=20 unprofitable.

Even if multiple servers are simultaneously = possible,=20 concentration statistics in utility markets may mislead. If entry is = suddenly=20 allowed into a franchise market that formerly had a single server, the = "share"=20 of entrants will be small at the outset for reasons unrelated to the = behavior of=20 the franchised seller. Alternatively, if the incumbent seller is acting=20 anticompetitively toward the entrants (e.g., because it denies them = access to a=20 facility that is essential if they are to compete) an HHI will not help = in=20 identifying that exclusion or inferring its likelihood. The speed with = which=20 seller shares change may also depend on regulation. If entry is allowed = but=20 prices are regulated, a new seller may be unable to discount rates below = current=20 costs in hopes of expanding quickly. If every one of fifty franchised = taxi=20 companies in a city must charge the same fare, adding an additional = provider=20 decreases measured concentration but has few other effects on = competition. n50=20 If existing regulated sellers gain the right to MBRs, inferences from = their=20 "shares" may be tenuous. Open access rules may prevent = [*311] them=20 from withholding capacity from customers, in which case a high HHI will=20 overstate market power. Alternatively, regulators may prohibit the entry = of new=20 competitors while allowing incumbents to withhold capacity. Here, the = HHI will=20 probably understate market power. The plausibility of any prediction = about=20 deregulation depends heavily on the details of the scheme. =

III.=20 Relevant Markets and Market Power at The FERC =

The FERC=20 has examined antitrust markets in numerous electricity, gas and oil=20 pipeline proceedings. In nearly all of these diverse = dockets,=20 it has employed a single basic model. The method borrows heavily from = the supply=20 structure analyzed in the Department of Justice's (DOJ) delineation of = the=20 relevant market in Otter Tail Power Co. v. United = States. n51=20 This method has guided the Commission in dockets extending from = electrical=20 transmission foreclosures in the late 1970s to interstate=20 pipeline MBRs in the mid-1990s. We begin with an = overview, not=20 intended as a detailed history, of Otter Tail and subsequent = applications of its=20 market theory at the FERC. We then look more closely at the market = definitions=20 and inferences the FERC has made in various studies of = pipeline=20 MBRs.

A. Otter Tail

Otter Tail brought antitrust = markets=20 to regulated industries. That company, a vertically integrated corporate = utility=20 serving Minnesota and the eastern Dakotas, refused to transmit ("wheel") = power=20 from federal hydroelectric projects to newly-formed municipal = distribution=20 utilities in its territory. n52 Unlike any other utility, Otter Tail had = also=20 filed a statement with the FPC that it did not hold itself out to serve = new=20 municipal utilities at wholesale. The government charged Otter Tail with = leveraging a transmission monopoly into a monopoly of retail service. = Its denial=20 of transmission on facilities that small towns could not economically = duplicate=20 discouraged the formation of competing municipal systems. Although the = Federal=20 Power Act did not allow federal orders to wheel, the courts determined = that=20 wheeling could be ordered as antitrust relief.

The affected = cities had=20 either received or were applying for allocations of inexpensive federal = power to=20 which they had a preference by law. If Otter Tail obtained the federal = power=20 because municipals did not exist, it was obligated to fold the price of = that=20 power into its regulated retail rates. The government's relevant = market was for retail franchises, and its market share = calculations=20 were unconcerned with institutional details of regulated rates, = Preference=20 Power, and obligations to serve. The government bounded its = relevant=20 market by the territory in which Otter Tail was obliged to = serve those=20 who would not serve themselves. Otter Tail's territory was in fact = largely=20 served by others. Municipals, cooperatives, and [*312] other = corporate utilities sold over 70% of all power retailed in the = territory. By the=20 antitrust standards of its time, Otter Tail's 30% market share might = have been=20 insufficient to support allegations of monopoly power. Perhaps for this = reason,=20 the government claimed (and Otter Tail agreed) that the relevant = market was for franchises in individual cities. Since 45 of the = 510=20 towns in its service area had municipal utilities, Otter Tail had, on = the=20 government's calculation, a 91% market share, in an area where it sold = under 30%=20 of the power.

The government produced no analogous quantitative = measure=20 of Otter Tail's market power over transmission, the market it had = allegedly=20 succeeded in monopolizing. The government, however, did not assert that=20 transmission was a relevant market, perhaps because = Otter Tail=20 owned only 6.2% of the 87,000 miles of transmission in its area. Since = Otter=20 Tail had such a low "share" of the market (and nearly 50 other = transmission=20 owners in its territory), the government instead claimed that the = company had=20 "strategic dominance," a previously unknown term in both economics and=20 antitrust.

Both the franchise and transmission market shares = mislead,=20 but for different reasons. Otter Tail's high proportion of city = franchises can=20 only be correlated with monopoly power at retail if its rates are = unregulated=20 and it has no obligation to serve. If it can charge only regulated rates = (and=20 regulation is effective) it cannot earn the supernormal returns of a = monopolist.=20 n53 With a service obligation, the company's control of numerous = franchises=20 might indicate little more than their unprofitability to others. Otter = Tail's=20 low proportion of transmission ownership, however, might still endow it = with=20 some monopoly power. If it refuses to wheel for a town without = alternative=20 connections that seeks to municipalize, denial of access to that single = line may=20 harm competition. The harm will occur regardless of how many or how few = total=20 miles of transmission the company owns in its territory. To act=20 monopolistically, the company need own no more than this one line.=20

B. Markets at the FERC

At the FERC, today's = market=20 analyses are easily identifiable as Otter Tail's descendants. In over = twenty=20 years since Otter Tail, there have been no conceptual or empirical = advances that=20 might better rationalize the Commission's use of that case's methods.=20 Commodities, geographic areas, and measures of seller concentration have = changed=20 over these years, but the structural methods of Otter Tail remain with = us.=20 [*313]

1. Electricity

The FERC has = applied=20 Otter Tail's methods in three broad classes of electricity proceedings. = n54 The=20 first includes dockets from the 1970s and 1980s dealing with topics that = are now=20 defunct or moribund. In some of them, the FERC examined wheeling as a = potential=20 remedy for anticompetitive arrangements between jurisdictional corporate = utilities and nonjurisdictional municipal systems. n55 Although the = Federal=20 Power Act (FPA) n56 at the time foreclosed the FERC from ordering = utilities to=20 wheel in wholesale rate cases, the Commission considered, but never = issued,=20 wheeling orders on antitrust grounds using as precedent Otter Tail and = certain=20 rulings of the Nuclear Regulatory Commission (NRC). n57 FERC staff and=20 intervenors sometimes calculated retail "market shares" and compared = them to the=20 shares that the antitrust courts had declared suggestive of monopoly = power. n58=20 In other dockets, municipal utilities charged that price squeezes = resulting from=20 disparities between state and federal rates inhibited competition = between them=20 and their wholesale suppliers for industrial loads. n59 Using the same=20 territorial standards that governed Otter Tail, the FERC ruled that a = common=20 border between a jurisdictional utility and its wholesale customer = sufficed as=20 evidence of competition. n60 The courts later rejected the Commission's = holding=20 that a price squeeze intervenor needed to show no active competition for = an=20 identifiable load. n61

Electric utility mergers and power = marketing=20 plans prior to the Energy Policy Act of 1992 (EPAct) n62 make up the = second=20 class of FERC proceedings that use concentration statistics to evaluate=20 competition. Such statistics have at times rationalized the FERC's = conditioning=20 of approval on [*314] offers of open access to transmission = by the=20 applicant utilities. n63 In these dockets, various bulk power and = transmission=20 markets have replaced retail service as the focus of analysis. n64 These = markets=20 are more in keeping with geographical reality than were Otter Tail's, = extending=20 beyond the applicant's service area to include systems that will be = accessible=20 under a more liberal transmission policy. n65 In approvals of MBRs for = other=20 bulk power coordination services, the FERC has also examined the = concentration=20 of their potential providers. n66

As concentration in generation = markets=20 falls, the FERC policy in a third class of cases has moved to a = near-presumption=20 that the market for power sales from new generation capacity is = competitive. n67=20 The FERC has also employed concentration statistics to evaluate MBR = applications=20 by non-utility generators. n68 In contrast, a recent appellate ruling on = transmission access for a utility seeking MBRs for bulk power sales did = not rely=20 on concentration statistics. Instead, the court remanded the stranded = investment=20 provisions of an open-access transmission tariff to the FERC on grounds = that the=20 Commission had not examined their potential anticompetitive impact as a = tying=20 arrangement. n69 The court invoked none of the FERC's market = concentration=20 analysis, possibly indicating a turn from the logic of Otter Tail.=20

2. Oil Pipelines

The EPAct = required=20 that the FERC establish a "simplified and generally applicable = ratemaking=20 methodology for oil pipelines." n70 In 1993, FERC staff = responded with a proposal that rates in competitive markets be=20 [*315] put under price cap regulation, and that increases in = rates=20 in noncompetitive markets be limited by increases that the=20 pipeline put into effect in competitive markets. n71 = The=20 staff's proposal largely formalized the standards embodied in the = Commission's=20 1990 three-year experimental ratemaking for Buckeye Pipe Line Company. = n72=20 There, the FERC found the Merger Guidelines useful as a first screen for = evaluating markets for transportation between individual origins and=20 destinations. As a supplement to the pipeline HHI, the = staff=20 recommended examination of the ease of entry, including entry of other = modes of=20 transportation. Geographically, the staff approved as destination = markets the=20 Department of Commerce's Bureau of Economic Analysis Economic Areas = (BEAs),=20 which delineate economic areas surrounding cities. Costs of substitution = bounded=20 BEAs, since data showed that trucks carrying oil products are good = substitutes=20 for pipelines within a BEA, but not necessarily between = BEAs.=20 n73 Neither the staff nor the Buckeye decisions consider the choices = that might=20 actively constrain an oil pipeline from withholding = capacity in=20 order to raise price. Instead, the decisions, (but not the staff) = concentrate on=20 the relationship between the HHI and the likelihood of collusion that = might=20 exist in unregulated industries. n74

3. Gas Inventories = and=20 Storage

Prior to its recent work on interstate transportation = tariffs,=20 the FERC examined MBRs in two other gas industry contexts. In the first = episode,=20 no longer operative, Commission Order 500 of 1987 allowed=20 pipeline resale ("sales") tariffs to incorporate gas = inventory=20 charges (GICs) that compensated the pipeline for = standing ready=20 to serve. n75 If a pipeline showed that its gas supply = came=20 from a competitive market and that it offered users comparable = transportation,=20 it could file for a GIC that varied with the market price of gas. n76 To = determine competition in the market for gas delivera- [*316] = ble to=20 the pipeline ("divertible" gas), the FERC chose to = start from=20 an HHI, after which it would consider transportation quality. n77 =

In the=20 second context, the FERC has begun to entertain MBR applications by gas = storage=20 operators. The necessary showing of competition includes calculation of = an HHI=20 (e.g., for storage facilities in the applicant's area that are connected = to the=20 same pipeline). n78 Here, too, the HHI is intended as a = screen,=20 to be augmented by findings on potential market entry and alternatives = to=20 storage in the relevant market. n79 In a prior = experimental=20 program for market-based storage rates, the market so constrained the = applicant=20 that it never negotiated a rate that was as high as the cost-of-service = cap=20 imposed by the FERC. n80

C. MBRs for Interstate=20 Pipelines

The FERC has produced three studies = analyzing=20 competition among pipelines. All of them utilize = measures of=20 supply concentration as screens for the market ability of=20 pipelines to somehow coordinate their actions in = reducing the=20 availability of capacity. n81

1. The Gallick Study=20

Published in 1993, economist Edward Gallick's study of potential = pipeline competition derives from his work for the FTC = in the=20 late 1980s. n82 Using data from the early and mid-1980s, he used HHIs to = analyze=20 the concentration of pipelines serving Standard = Metropolitan=20 Statistical Areas (MSAs). He found that few areas were served by enough=20 independently owned lines to pass an HHI screen for competitiveness. = Gallick=20 then examined the costs and delays of extending nearby=20 pipelines to reach areas that might otherwise be harmed = by=20 overly high supplier concentration. n83 He concluded that after two = years,=20 competitive entry of new pipelines to these markets = would=20 significantly lower the concentration of pipelines = serving=20 them. Sixty percent of all MSAs, and 90% of the larger ones, could have = HHIs=20 below 0.25 within two years. n84 Since 90% of all gas would be consumed = in=20 competitive areas, he concluded that the FERC should consider MBRs as an = alternative to cost-of-service regulation. n85 [*317] =

Like=20 the FERC's later analysts, Gallick rationalizes the HHI as "a crude = index of the=20 likelihood of successful coordination or collusion among the colluding = group of=20 gas suppliers." n86 He does not discuss the effects of regulation on the = ability=20 to withhold supplies or on the observed concentration of=20 pipelines at city-gates. n87 Prior to the optional = expedited=20 certificate provisions of Order 436, however, regulation should have = been=20 particularly important in determining concentration. n88 Virtually all=20 then-extant pipelines had been constructed under = certificate=20 procedures that required them to have long-term contracts with producers = and=20 local distribution companies (LDCs), and allowed rate-basing only upon a = showing=20 of market need. Although open access had begun and transportation had = replaced=20 sales as the majority of pipeline throughput at the = time of=20 publication, its author notes the change only in passing. n89 He does = not=20 analyze the conditions, if any, under which to expect that a=20 pipeline's market power is independent of the = transactional=20 structure that governs its industry.

2. The 1993 Task = Force=20

a. The Task Force Report

In May 1992, the FERC=20 established a Pipeline Competition Task Force, chaired = by=20 then-Commissioner Branko Terzic. His May 1993 report for the group = discussed how=20 competition might evolve as pipelines filed and = implemented=20 their settlements under Order 636. n90 Addressing transportation between = market=20 centers, the report identified two "key issues" for judging market = power: "how=20 much transporter concentration is acceptable?" and "how much do=20 pipeline paths from other producing areas (or to other = market=20 areas) count as competitors?" n91 The Task Force used the HHI to measure = concentration, noting that the index "does not measure competition = directly,"=20 but that "high concentration can suggest market power." n92 After = examining two=20 examples of competition along parallel pipeline paths, = the Task=20 Force concluded that "under the right condi- [*318] tions, = three to=20 five competitors may contest key routes." n93 Although it did not detail = what=20 those conditions might be, the Task Force understood that "for a=20 pipeline to exercise market power, it must be able = profitably=20 to withhold or restrict its customers' access to transportation = capacity,"=20 including "any relevant capacity on other pipelines." = n94=20 Instead of examining the record under open access, the report offered = only=20 hypothetical examples of capacity withholding. n95

b. = The=20 Discussion Paper

The FERC staff's Discussion Paper for the Task = Force=20 summarizes the antitrust reasoning that apparently underlies the Task = Force's=20 recommendations. n96 Its bibliography contains numerous economic = references on=20 the relationship between supplier concentration and the competitive = performance=20 of a market. n97 Two appendices to the Discussion Paper contain articles = on the=20 theoretical links between market structure and performance. The first = contains a=20 mathematical exposition of how concentration-performance relationships = in=20 oligopolies depend on assumptions about seller conjectures. n98 Its = author does=20 not examine the possible effects of rate regulation or open access on = the=20 theoretical relationships he derives. n99 More relevant to=20 pipelines, the paper contains no analysis of markets = where a=20 regulated capacity provider in effect competes with customers who can = release=20 their entitlements to others. The second appendix particularizes the = economic=20 model of the first appendix to network industries which are regulated = and have=20 large sunk costs. n100 Although its authors acknowledge = [*319] that=20 regulation can restrict competitive entry, they do not examine the = predictive=20 value of an HHI in an industry with restricted entry and obligations to = serve.=20 n101

The Task Force's work is consistent with the theory = exposited in=20 the Staff's Appendices, but the relevance of that theory to=20 pipelines is not made clear. As we have noted above, = the=20 conjectures that sellers hold about one another's behavior are critical=20 determinants of any conclusions about the relationship between market = structure=20 and performance. Much of the theoretical work in the FERC staff's = appendices=20 assumes that sellers hold Cournot-Nash conjectures, which typically lead = to an=20 inverse relation between a market's competitive performance and its HHI. = n102=20 The staff, however, acknowledges that few attempts have been made to = estimate=20 conjectures numerically. Neither of the two attempts cited finds = substantial=20 evidence of Cournot-Nash conjectures in real-world markets. n103 The = staff=20 rationalizes its reliance on such theories as follows: =

=20
Many studies are forced to make simplifying assumptions by the = lamppost=20 principle. The lamppost principle is simple: work where the light = (theory or=20 data) is good. But we are often trapped by the lamppost principle. = Because H=20 [the Herfindahl index] appears in models when behavior is Cournot, the H = is=20 given more credibility as a statistic. This statistic is only a valid=20 descriptive statistic for performance, market behavior, or welfare = effects when=20 behavior is Cournot. In markets more or less competitive than Cournot, H = does=20 not describe performance. To the extent the market diverges from a = Cournot=20 equilibrium, H is an arbitrary measure of performance. n104
=20

3. Staff's 1995 MBR Study

The 1993 Task = Force=20 stressed that the FERC's future MBR policies would depend on the content = of=20 Order 636 settlements, on how secondary markets and market centers = developed,=20 and on how pipeline operations evolved. n105 In = response to=20 evolving competition since Order 636, in 1995 the FERC issued a Notice = of=20 Proposed Rulemaking on MBRs. n106 In connection with that docket, the = FERC staff=20 produced a paper that summarized the Commission's earlier market = analyses and=20 broadly recommended continuation of existing methods. n107 The staff = proposed no=20 major departures from established procedures for evaluating competition, = and=20 eluci- [*320] dated no new intellectual rationales for those = procedures. The borrowings from antitrust pre-merger review would = remain, and=20 the staff's proposed method "would be the same for all types of = services." n108=20

Although data on market performance (including electronic = bulletin board=20 records containing prices) were becoming available, the staff performed = no=20 quantitative analysis of these or related data on transactions. n109 At = first=20 glance, the observed evolution of the industry under open access should = have=20 caused the staff to reexamine its presumptions. Seller concentration in = many=20 parts of the industry was virtually unchanged from before open access, = but=20 markets with unchanging supply structures were producing increasingly = more=20 competitive outcomes. The 1993 Task Force apparently had good reason to=20 recommend close examination of the effects of the new regulations. In = the next=20 section, we summarize some of the new knowledge.

IV. = Competition=20 Since Open Access

A rational choice of pipeline = ratemaking policies requires that we first examine the market for the = gas they=20 transport. By virtually all evidence that we present below, the gas = market has=20 become unified, national, and increasingly competitive since the = institution of=20 open access. To understand the consequences of gas market developments = for=20 pipeline MBRs, we begin by examining markets that = encompass=20 networks. Network analysis and our ensuing examination of network data = provide=20 strong logic and evidence for a major revision of policy. They show that = pipeline markets defined over an origin, a destination, = and the=20 direct links between them are inappropriate for today's industry, and = are likely=20 to lead to erroneous conclusions about MBRs. n110

The FERC will = be=20 better able to tailor procompetitive policies if its future analyses are = informed by a network model. Competition is at base a situation of = abundant=20 alternatives. Those alternatives are better summarized by examining open = access=20 to a network rather than to individual pipelines or = paths. The=20 past decade's expansion of interconnections and trading institutions has = so=20 increased competition that the markets the FERC believes are relevant = are the=20 ones that its policy has already rendered irrelevant. Origin-destination = analysis describes opportunities in a balkanized, weakly connected=20 pipeline network that no longer exists.

Open = access to=20 the network provides opportunities to arbitrage price differences and = move gas=20 in paths that were unavailable in the past. This expanding set of = arbitrage and=20 trading activities is fundamental to the determination of competitive = prices.=20 The price of gas at any point in an interconnected network is determined = by the=20 supplies and demands for gas not just at that point, but at all points = on the=20 network. The network encom- [*321] passes a competitive gas = market=20 if prices over it converge quickly to eliminate profitable arbitrage=20 opportunities. We infer competition by examining prices. The evidence = from=20 prices is directly available, and conclusions from observed prices are = better=20 grounded than conclusions drawn from market shares. The case in favor of = the=20 price evidence is clear: In many areas, the market shares on which the = FERC=20 might rely have hardly changed at all since open access began, while = market=20 behavior has become strikingly more competitive, both for gas and its=20 transportation.

A. Using Prices to Measure Competition=20

In a competitive market with costless exchanges, only one price = will=20 prevail. If there are two distinct prices, sellers in the low-price area = can=20 profit by moving their goods into the high-price area, and buyers in the = high-price area gain by purchasing from low-price sellers. The market = may=20 contain arbitrageurs whose specialty is to profit by eliminating such = price=20 differences. In a perfectly competitive market equilibrium, there are no = profitable arbitrage opportunities. In a monopolized market, the = monopolist's=20 high price is sustainable only if supplies from elsewhere cannot enter = the=20 market. For a good that is costly to transport, such as gas, the market = is=20 competitive if prices at different locations are within arbitrage = limits, i.e.,=20 they do not offer anyone the opportunity to profit by shifting gas from = one area=20 to another. If many traders can reach a customer over many paths from = many=20 sources of product, price cannot long exceed the competitive level. An = area=20 where price is within arbitrage limits of prices at many other locations = is=20 economically linked with all of those locations, whether they are = directly=20 connected or not. They are in the same integrated economic market and = the array=20 of prices is competitive, no matter how many or how few=20 pipelines connect them.

Figure 1 presents an = example of=20 competitive behavior for two gas markets that are separated from one = another and=20 located on pipelines that do not interconnect directly. = The=20 prices are Gas Daily's monthly spot contract index values for East Texas = gas=20 entering Tennessee Gas Pipeline's interstate system, = and for=20 Oklahoma gas entering ANR Pipeline. n111 The evolution = of these=20 prices shows a broad pattern typical of prices since open access. Their=20 difference diminishes with time, and their fluctuations become more = closely=20 matched. The evolution of these prices is consistent with an increase in = competition since open access, as prices come to lie within a band whose = limits=20 are set by transportation and arbitrage costs. If price differences = exceed the=20 width of that band, arbitrage profits are possible. By late 1991, the = prices=20 move together so closely that the two areas appear to be in the same = market.=20 n112 The commonality of movement occurs [*322] despite the = fact that=20 the two areas are not directly connected. The prices have converged = because=20 producers in the areas can deliver gas to interconnected downstream = locations=20 among which arbitrage is possible. [br n:fig,l(.10,.10)][mg=20 f:'[data.egy.16-2]eps20101.eps',w30.,d24.]

Information contained = in the=20 prices in different areas also determines whether or not they are in the = same=20 market. Competitive markets are characterized by easily accessible = information=20 about prices and other market conditions that allows traders to quickly = discover=20 and exploit profitable opportunities. If a competitive market = encompasses two=20 locations, it will be impossible to use information about a price change = at one=20 location to make a profitable arbitrage trade at the other. If, for = example, the=20 two prices rose and fell in identical patterns, but those in one area=20 consistently lagged behind those in the other, profitable arbitrage = could occur.=20 One could profit by buying gas (or selling it short) in the lagging = area, using=20 price in the leading area as a guide. The manner in which the two series = on the=20 graph move together appears to indicate that information about one=20 [*323] price will not be of help in predicting the other = price. n113=20 Prices in the example are so informative that they do not allow = profitable=20 arbitrage between the present and the future. When prices contain the = same=20 information, they are in the same market. A market encompassing both = producing=20 areas (and possibly more) will be the relevant market = for=20 evaluating the competitive consequences of MBRs, since this is the area = within=20 which information and trading constrain prices to competitive levels.=20

Prices at two locations (or at two dates) are said to be = informationally=20 efficient if both usually contain the same relevant economic = information. n114=20 They contain such information if arbitrage leads to prices that are=20 arbitrage-proof. If, for example, price movements in one area = consistently=20 lagged behind price movements in the other, these prices would not = contain all=20 relevant information. If so, arbitrage profits remain possible despite = the=20 underlying correlation. The prices in Figure 1 demonstrate informational = efficiency. In an informationally efficient market, price changes = reflect only=20 the arrival of new information or the appearance of new constraints. = Examples of=20 such factors include changes in transmission prices and changes in the = cost of=20 buying or selling gas. Again, arbitrage may act through markets other = than these=20 two, e.g., through exchanges at a common intersection. =

B.=20 Competitive Pricing in a Network

Next, consider a more complex = network=20 of open-access pipelines that encompasses a single = market. Gas=20 delivered to a point on the network must arrive at a price that is no = higher=20 than that of gas from any other possible source. Under open access, gas = supplies=20 can come from interconnected production areas, from storage, and from = consuming=20 areas if purchasers there can reassign their contractual takes. Open = access=20 means that both entitlement holders and outsiders can exchange capacity = on a=20 pipeline that links two areas, and that both can bid = for unused=20 space that reverts to the pipeline for mandatory offer = as=20 interruptible service. The economic link between the locations is not = the=20 pipeline, but rather the supply paths within the line = that are=20 held by the individual parties. The allocation of these paths is = reconfigured in=20 real time in the misleadingly-named secondary markets for capacity and = reverted=20 interruptible service. Economically, it is immaterial whether the = distribution=20 of entitlements gives their holders rights to use a single pipe or = several=20 parallel pipes. The HHI for pipelines between two = points will=20 change with their number, but that number is of no relevance to = competitive=20 outcomes under open access.

The appearance of a capacity = bottleneck=20 along a single pipeline link in the network will bring = forth a=20 cascade of reallocations. The reallocations will occur whether the = bottleneck=20 stems from the pipeline's engineering limits or from an = attempt=20 to withhold usable capacity in hopes of raising its price. First, the = high price=20 for released capacity and interruptible service [*324] will=20 encourage those shippers who do not value their space so highly to trade = it to=20 shippers who do. The latter need not be current entitlement holders. = Second,=20 space on pipelines that are alternative (not = necessarily=20 direct) paths between the two points can be reallocated to those who = sold their=20 rights in the first bottleneck. Reallocations on those alternative=20 pipelines will give rise to still more reallocations = and their=20 associated price changes on a third tier of alternatives. Reallocations = of=20 transportation capacity may make it advantageous for some gas buyers to = change=20 their supply areas. On all affected pipelines, there = may be=20 reallocations of firm and interruptible capacity. Even if one believes = that a=20 high HHI for pipelines between two points usefully = measures the=20 ability to withhold capacity, an HHI over only the direct path will = overestimate=20 the market power held by the pipelines comprising that = path.=20 n115

By the FERC's origin-destination standard, a line that = connects two=20 formerly unconnected points on a network is a monopoly with an HHI of = 1.0=20 between the points. This monopoly, however, is of little significance. A = textbook monopolist gains by restricting the availability of services in = its=20 market. A new link in a network will often increase (and never decrease) = the=20 network's total potential throughput. The new link will also increase = the number=20 of paths between a given buying area and possible suppliers. If the = owners of=20 pre-existing segments of the system had monopoly power, the construction = of a=20 new link gives system users alternatives they did not have before, even = if they=20 are not directly connected with that link. By increasing the = alternatives for=20 buyers, the new link decreases any monopoly power of existing line = owners,=20 including those who are not directly interconnected with the new link. = If the=20 new link both adds capacity and adds to the alternatives of buyers, = their=20 welfare cannot possibly worsen.

Under open access, all points in = the=20 network may be in the same market. Any point upstream or downstream of a = consuming area can supply gas to it. For example, a downstream LDC might = release=20 transmission capacity to a broker who uses it to sell gas to an upstream = market.=20 In effect, the downstream LDC is supplying the upstream market with = delivered=20 gas on which the LDC has a right of first refusal. Arbitrage in the = market need=20 not require physical deliveries. Offsets and futures trades can move = prices=20 toward equality between points that are only indirectly or not at all = connected.=20

If pipelines function as clearinghouses to = offset=20 trades for one another they in effect allow their customers to arbitrage = over=20 disconnected markets. Gas does not have to flow between the=20 pipeline systems in order to mitigate regional price=20 differences. A customer can use the competitive futures market as a = source of=20 supply by taking delivery at the Henry Hub (the contract's standard = delivery=20 point) and using connections there and elsewhere to effect delivery. = Many other=20 sophisticated clearing and trading [*325] strategies are in = use and=20 under development. n116 Such strategies can move prices to within their=20 arbitrage limits even when connections to the pipeline = network=20 are few. Circumventing monopoly price offers a pure arbitrage profit to=20 ingenuity.

Open access facilitates the entry of traders whose = activities=20 push price to its arbitrage limits between points on the network. If = prices at=20 two points indicate potential arbitrage profits, a newcomer can = construct a path=20 to ship between them by acquiring interruptible transmission or = short-term=20 released capacity. The combinatorics n117 of the network may provide a = large=20 number of possible paths between the points, and arbitrage will be more=20 profitable if the gas flows along less expensive paths that are = otherwise=20 underutilized. Short-term transportation transactions make it possible = to enter=20 and exit a market quickly and without making irreversible commitments. = With open=20 access, a market entrant need not incur the sunk costs and delays of=20 constructing a pipeline that would otherwise be = barriers to=20 competitive entry. Because "hit and run" entry is so easy under open = access, gas=20 markets have become "contestable." A contestable market produces=20 nearly-competitive outcomes regardless of the current number of sellers. = n118=20 Their ability to charge supracompetitive prices is limited by the = constant=20 threat that doing so will encourage rapid entry.

C. = Unification=20 of the Gas Market

If a market encompasses a unified network, = policies=20 that nominally apply to only a subset of the network will in fact affect = the=20 entire system. This "spill-over" problem is well-known in other economic = contexts. The imposition of a price ceiling that causes shortages in one = market=20 will affect prices in uncontrolled markets. If prices of substitutes for = the=20 controlled good rise to politically unacceptable levels, the government = must=20 also control the prices of those substitutes. Conversely, deregulating a = single=20 market in a regulated system may produce unacceptable dislocations in = those=20 markets which remain regulated. In this and subsequent sections, we = provide=20 evidence that both gas and pipeline services are traded = in a=20 unified, nationwide market. On spillover reasoning, selective grants and = denials=20 of MBRs to origin-destination pairs in such a market will affect gas = flows and=20 efficiency everywhere. n119 [*326]

In a fragmented = market,=20 producers in different areas face differing prices. The differences will = exceed=20 arbitrage limits and prices will be poorly correlated because markets = are=20 incompletely linked. Buyers who purchase at high prices will do so for = the long=20 term because they do not have easy access to lower cost supplies. Those = supplies=20 might come from other producers or from other buyers engaged in = arbitrage. Such=20 arbitrage is a manifestation of competition. In the eyes of buyers, = shipments by=20 arbitrageurs to their area are perfect substitutes for shipments by = producers.=20 If arbitrage flows to that area can originate from anywhere in the = network,=20 then, after allowance for transportation costs, gas from anywhere is a=20 substitute for gas from anywhere else. If substitutes define a market, = the=20 entire network is now a single market.

Economists have recently = begun to=20 use correlations of prices among commodities and areas to estimate the=20 boundaries of markets. n120 Correlation analysis has occasionally = appeared in=20 antitrust litigation, but litigation remains dominated by the methods of = market=20 definition described above. n121 Price correlations provide evidence of=20 competition that is independent of market structure, and hence less = prone to the=20 inferential difficulties associated with concentration measures. n122 = Instead of=20 requiring complex assumptions about the behavior of market participants, = price=20 correlations indicate how buyers have conducted business in the face of = actual=20 opportunities. The strength of correlations is evidence of how well = arbitrage is=20 facilitating efficient trades. Further, the geographic scope of = competition is=20 endogenous in a price correlation analysis, as opposed to = [*327] =20 assuming that an origin-destination pair forms a market. We next provide = summaries of four such investigations of gas price behavior. =

1.=20 Cointegrated Production Area Markets

If two areas are in the = same=20 competitive market, their prices will inhabit a band whose width = reflects the=20 cost of arbitrage. Those costs include transportation, risk exposure, = and=20 information about profitable opportunities. If competition exists, it = will=20 quickly bring disparate prices back within their arbitrage limits. If, = for=20 example, bad weather increases price in area i while price at area j and = transmission cost are unchanged, transactions in a competitive market = will=20 restore an equilibrium at which the two prices again differ by no more = than the=20 arbitrage limits. If the cost of arbitrage varies little over time, two = areas=20 are in the same market if the difference between their prices is = relatively=20 constant. The statistical technique known as cointegration provides a = criterion=20 under which to determine the relative constancy of such a difference. = n123 If=20 the prices are not cointegrated, there are no well-defined bounds on the = difference between them.

If prices in two areas are = cointegrated, the=20 areas are in the same economic market. Although the difference between = the=20 prices varies with some randomness, there is a high probability that it = will=20 remain within arbitrage bounds. De Vany and Walls have shown that at the = beginning of open access daily average prices in only a handful of pairs = of=20 production areas were cointegrated. n124 Subsequently, additional=20 pipelines began open access operations and new network = links=20 were constructed. n125 By 1991, over 65% of production area pairs had = become=20 cointegrated, i.e., behavior of prices at the locations indicated that = they were=20 in the same market. As open access progressed, market forces came to = operate=20 over greater distances. At the end of 1991, distant pairs of markets = were=20 cointegrated to the same degree (in percentage terms) as more proximate = pairs.=20 This finding is consistent with a progressive unification of gas markets = that=20 tracks [*328] the evolution of open access, increased=20 interconnections, and the growth of market centers. n126 =

2.=20 Network Arbitrage

Generalizing the above findings that apply to = pairs of=20 markets, the statistical technique of vector autoregression (VAR) = permits=20 analysis of simultaneous arbitrage possibilities at multiple points in a = network. VAR analysis builds on the concept of informational efficiency, = i.e.,=20 that if prices contain the same information they will only change with = the=20 arrival of new information. If all price changes are driven by new = information,=20 past price changes or price changes elsewhere in the network are of no = value for=20 predicting future price changes in an area. n127 VAR implements these = ideas by=20 modeling the dependence of price change in a producing or consuming area = on past=20 changes in that area price and in all other prices on the network. VAR = produces=20 regression equations from whose coefficients one can infer the presence = or=20 absence of profitable arbitrage opportunities between all pairs of = markets in=20 the net. The criterion of no arbitrage opportunities between any pairs = of=20 markets is a strong one, given the large number of possible ways of = choosing=20 pairs of markets from larger sets. n128

De Vany and Walls = performed VAR=20 tests for possible arbitrage on six sub-networks of supply basins and=20 pipelines. n129 Examining data for July 1987 to June = 1988, when=20 only a small number of pipelines had begun open access=20 operation, they found profitable arbitrage opportunities on all six = networks,=20 inconsistent with a unified market. By June of 1989, prices on one = network (five=20 pipelines leaving Oklahoma) had become arbitrage-proof, = but not=20 on others. By 1990, with open access on all major = pipelines,=20 five of the six networks exhibited strongly competitive, arbitrage-proof = pricing.

The dynamics of market reaction also changed with open = access.=20 As markets become more interconnected, in effect reserve stocks in all = markets=20 become available to dampen price shocks in any one of them. Both the = speed and=20 the accuracy of price convergence have improved with open access. = Examining the=20 same sub-networks, De Vany and Walls found that in 1988 a standardized = price=20 shock at any one location was dampened to [*329] nearly zero = within=20 seven days. n130 By 1990, such a shock would be dampened in three days, = and=20 nowhere on the network would prices move by more than a few cents in = response.=20 This small and widespread response of area prices is itself evidence for = unification of the market. Such a response pattern can occur only if = there are=20 numerous changes in gas flows among regions not directly connected with = the area=20 where the shock occurred.

3. Market Hubs

Under = open=20 access, the industry has seen the growth of market centers, or hubs, at = which=20 several pipelines can interconnect. n131 New hub = junctions can=20 expand the scope of markets by expanding the number of possible = arbitrage paths=20 between producing and consuming areas. n132 By 1988, all but two pairs = (out of=20 45) of the twenty largest pipeline nodes were connected = by two=20 or fewer links. By 1990, some of these hubs connected more than a = hundred=20 markets with one another. Examining three such hub-centered networks = near major=20 markets, De Vany and Walls found that prices between market pairs = reachable=20 through the hubs were cointegrated, and that the likelihood of = cointegration did=20 not vary with the number of paths through a hub. n133

4. = The=20 Spot and Futures Market

Futures markets offer both physical = deliveries=20 and contracts which can be held to hedge risk. Empirical studies = generally agree=20 that futures markets are competitive and informationally efficient. n134 = If the=20 gas futures market is competitive, a nodal market on the network will be = competitive if its price is cointegrated with that of gas futures. De = Vany and=20 Walls examined whether spot prices at eight different hubs and three = city gates=20 were cointegrated with futures prices. All of the city gate prices and = all but=20 one hub price were cointegrated with futures prices. n135 This is a = small=20 sample, but it is highly suggestive that the market has achieved a = degree of=20 integration that appears to cross both geography and time. n136=20 [*330]

D. MBRs in a Unified Market =

Whether=20 observed at the producing area, the market center, the city gate, or the = futures=20 market, gas prices provide abundant evidence that a highly competitive = market=20 has arisen with open access. Prices are arbitraged to the extent that it = is=20 generally impossible to consistently profit from predictable = transactions.=20 Delivered gas prices are the sum of producer prices, = pipeline=20 charges, and miscellaneous market-related costs. Producer prices are now = unregulated, and most miscellaneous services sell at market prices. = There=20 remains the possibility that monopolistic pipeline = rates are=20 still causing dislocations or will do so when MBRs for transportation = arrive.=20 For example, if the pipelines in Figure 1 both act=20 monopolistically, withholding capacity and marking up price by identical = amounts, there will still be no profitable arbitrage on the network, but = the=20 pipelines will earn supernormal returns.

Most = city=20 gates are served by more than one pipeline. For those = that are=20 not, there are usually numerous paths from numerous fields that = terminate at=20 points near the city gate. For monopolistic pipeline = pricing to=20 be general under MBRs, pipelines would need to form and = police=20 collusions. Such a collusion would be easy to detect, since customers = have=20 access to public and nonpublic statistical data on capacity, as well as = their=20 own experience to go by. They can observe throughputs and can observe = the=20 activities of the pipeline and others through = electronic=20 bulletin boards (EBBs) on which released capacity and interruptible = reversions=20 are traded. n137 Treble antitrust damages and class actions will = encourage=20 pipeline shippers and entrepreneurial plaintiff's = lawyers to=20 watch closely.

A collusion among pipelines = would also=20 be more difficult to organize and coordinate than a collusion in a = non-network=20 industry. If there are a large number of possible paths between fields = and=20 consuming areas, the collusion must contain a large number of=20 pipelines. If the market for flowing gas is national, = only with=20 extensive membership can a collusion put meaningful restrictions on the=20 alternative paths that customers might use to escape the exaction. Even=20 pipelines not directly interconnected must join such a=20 collusion. The larger the colluding group, the more difficult it will be = to=20 reach agreement on sharing the gains. The sharing task is further = complicated by=20 the multiplicity of paths available to shippers, since prices on = individual=20 pipeline links must be set so that customers cannot = easily=20 arbitrage among those paths. Random gas discoveries and demand shifts in = consuming areas would require that the agreement be flexible enough to=20 accommodate such changes.

We have found no FERC proceedings in = which=20 shippers have alleged that an open-access pipeline has=20 willfully withheld capacity or interruptible [*331] service = in=20 quantities that are significant by antitrust standards. n138 There is = instead=20 considerable evidence against withholding of service. = Pipelines=20 regularly announce expansion plans when their capacity limits are = reached. n139=20 With the help of expedited certification, pipelines = under open=20 access have persistently built new capacity, with $ 6.1 billion in new=20 facilities proposed, under way, or completed during 1994. n140 Further = lowering=20 barriers to entry, pipeline construction costs have = fallen by=20 37% in the past decade. n141 Market centers have grown substantially, = but a=20 pipeline wishing to exercise monopoly power should be = reluctant=20 to join with others in forming a market center. Such centers facilitate=20 switching by shippers, both among pipelines and among=20 alternatives to firm transportation, including storage.=20 Pipelines themselves are investing in new storage = facilities,=20 although under no regulatory compulsion to do so, and are sometimes = requesting=20 authority to charge MBRs for it. n142 Oil pipelines are = being=20 converted to gas pipelines, and new = pipelines=20 are being built to competitively maximize, rather than minimize,=20 interconnections with others. n143 Instead of behaving monopolistically, = some=20 pipelines are attempting to facilitate capacity = release,=20 presumably to win shippers from others. n144

The evidence points = to a=20 conclusion that pipelines are already operating under a = regime=20 in which MBRs play a dominant role. n145 Discounted rates for short-term = and=20 released pipeline capacity are market rates, since they = do not=20 bump against the cost-of-service ceiling. On the other side of = cost-of-service,=20 a party that holds rights to capacity whose uncontrolled market value = would=20 exceed the nominal ceiling can realize their value by engaging in "gray = market"=20 transactions. In the gray market, a capacity holder transports gas on = its own=20 account for a third party which has actually arranged the gas purchase, = and=20 charges a single price for the bundle of gas plus transportation. Since = the=20 price of gas is unregulated, the market premium for the capacity can in = effect=20 be paid in the price at which the gas [*332] is resold to = the actual=20 user upon delivery. The gray market is not a perfect substitute for the = open=20 market since transactions in the former are more costly. Although it is = hard to=20 estimate the volume of gray market transactions, there is little = evidence of=20 capacity shortages and non-price rationing that would occur if a = cost-of-service=20 cap were a binding constraint on price. n146 If prices of=20 pipeline services are flexible in both directions and = capacity=20 withholding is infeasible, the entire controversy over MBRs may be moot, = since=20 we already have them.

The behavior of both gas prices and=20 pipeline rates provides abundant evidence of = competition and=20 little evidence of monopoly power. There is no known evidence that the = degree of=20 competition, in either gas or pipelines, rises with the = number=20 of pipelines serving an area. With wellhead prices=20 decontrolled, it is clear that gas is priced competitively. With the = development=20 of an open-access pipeline network, there is growing = reason to=20 believe that pipeline services are being priced = competitively.=20 Rather than restricting the availability of capacity or allowing = bottlenecks to=20 grow, pipelines are expanding capacity and = interconnecting more=20 intricately with one another. If pipelines already = operate in=20 competitive markets, controlling their rates at historical cost of = service (or=20 any other amount) is more likely to inhibit the forces of competition = than to=20 temper the forces of monopoly.

V. Market Definition in = the New=20 Gas and PipelineMarkets

With open access, a unified national gas = market=20 has emerged, with buyers and sellers linked by a complex yet accessible = net of=20 interconnected pipelines. When the FERC considers the=20 competitive impacts of its policies, it must keep its eyes on this broad = market=20 rather than the narrow ones it examines in MBR proceedings. Because the = FERC's=20 tools embody incorrect presumptions, they will probably point to = incorrect=20 policies. After looking at the logic of the FERC's presumptions, we = propose a=20 reversal in the commission's methods of analysis. Our proposal leads us = to=20 reexamine the logic of the staff's relevant product markets in light of = recent=20 history, and finally to reexamine the staff's arguments about=20 pipeline undersizing.

A. Presumptions = and=20 Policy

In overseeing competition, the FERC places its initial = reliance=20 on principles taken from the Department of Justice's (DOJ) Guidelines = for=20 mergers. The Clayton Act, however, gives the DOJ clearer guidance than = the NGA=20 gives the FERC. Starting from a generally plausible presumption that = markets are=20 competitive (possibly less than perfectly so), the Clayton Act asks the = DOJ to=20 stamp out a monopoly before it can cause harm. Merger policy begins from = where=20 it wants the market to stay - in competi- [*333] tion. The = DOJ=20 constructs its relevant markets and employs its chosen = decision=20 criteria with the objective of approving only those mergers that will = maintain=20 an existing state of competition.

When the FERC looks at = competition, it=20 must start elsewhere. The FERC's very existence as a regulator of price = and=20 entry is predicated on a belief that unregulated markets would be = uncompetitive.=20 The FERC expects that its regulation will mitigate market power by = disallowing=20 the prices a monopolist would charge and the outputs it would produce. = In=20 pipelines, the underlying conditions have changed. Open = access=20 has made possible a degree of competition, but probably one that falls = short of=20 perfection. The FERC must decide whether allowing competition to govern = the=20 changed market will produce better outcomes than its existing policy = does.=20 Regulation may in fact produce market outcomes superior to unregulated = monopoly,=20 but this observation tells nothing about whether regulation is = preferable to=20 competition.

Whatever the value of the DOJ's methods in merger = cases,=20 the FERC is using them to analyze the wrong problem. The DOJ's = relevant=20 market is the smallest producer group that might be able to act = in=20 concert to impose a small and sustainable price increase. That increase = is over=20 the current price, which is assumed to be competitive, or at least not=20 anticompetitive. n147 If regulated rates bear no necessary relation to=20 competitive rates, predictions from the HHI about how MBRs might rise = above=20 regulated levels are of little economic value. They are only predictions = about=20 how market rates might differ from rates that recover historical costs, = a=20 prediction unlikely to improve the quality of an MBR decision. The = FERC's=20 maximum acceptable price increase under MBRs is one "at or below the = applicant's=20 approved maximum cost-based rate plus 15%." n148 An applicant is thus = more=20 likely to get MBRs if it has already expended large amounts. An = applicant whose=20 investments were more productive per dollar than those of its = competitors might=20 violate the 15% threshold, while its competitors would not. The staff's=20 rationale for starting from existing rates is another application of the = lamppost principle:

The regulated price has been = used as=20 the prevailing price - a proxy for the competitive price. This is = necessary=20 because almost all prices for transportation are regulated and a = competitive=20 price level would be, at best, a guess. However, the use of prevailing = prices=20 presents analytic problems. For example, three = pipelines that=20 follow parallel courses may have radically different rates because of = different=20 historical costs, despite the fact that in a competitive market they = would offer=20 almost identical services at almost identical prices. Which of the = alternative=20 pipelines' prices should be used as the "prevailing" = price?=20 This question would have to be addressed in deciding whether the prices = of=20 alternatives are appropriate references. n149
=
[*334] =20 In effect, competition will prevail if a randomly selected number (the=20 cost-of-service rate) that is unrelated to the likely post-MBR price = falls into=20 the right range.

On the FERC's criterion, MBRs might not produce = even=20 the rudimentary outcome of a single market price. If a market contains = several=20 pipelines with differing historical costs, the staff = proposes=20 an investigation of whose costs should be assumed to be the competitive = price.=20 n150 Let there be three pipelines with historical costs = of $ 1,=20 $ 2, and $ 3. If the FERC chooses $ 3 (and that also becomes the = prevailing=20 market price), two of the pipelines make unacceptably = high=20 profits. If the Commission chooses $ 1 or $ 2, it faces charges of = confiscation=20 and the problem of keeping the costly pipeline(s) = whole. If it=20 chooses a different benchmark for each, they cannot possibly compete to = a single=20 price without violating the 15% standard.

Competing under MBRs, = some or=20 all of these pipelines might earn high returns (or = realize=20 losses) relative to historical costs. MBRs are of value not because they = link=20 prices to historical costs, but because they facilitate rational = planning for=20 the future. Market rates send economically correct signals about where=20 investment in new facilities will be valuable and where it will not, a = signal=20 that cost-based rates are unlikely to send as effectively. Upon = institution of=20 MBRs, the rates that prevail may or may not equal long-run competitive=20 equilibrium rates. Only the future pattern of investments and = disinvestments can=20 determine those rates. Whatever the short-run returns to=20 pipelines under MBRs, open access, capacity release, = and=20 interruptible reversion will foreclose withholding of capacity or = service. n151=20 It is an absence of withholding, rather than a temporarily high or low = price,=20 that distinguishes competition from monopoly.

Beyond inducing = efficient=20 short-run allocations of capacity and long-run allocations of = investment, MBRs=20 can also resolve the inefficiencies associated with rates that are = unnecessarily=20 discriminatory. To discriminate profitably a pipeline = must=20 group its customers in accord with their elasticities of demand, in = order to=20 charge lower prices to those with a lower willingness to pay. The=20 pipeline must also keep these customers separated or = otherwise=20 prohibit resales among them. With open access and MBRs, such a ban on = capacity=20 resale cannot stand because the pipeline's customers = (and=20 interested third parties) become its competitors. If there are = disparities=20 between price and cost among customers, capacity release facilitates = arbitrage=20 that eliminates those differences. Further, in a network market the = potential=20 victims of discrimination have alternative paths for delivery beyond = capacity=20 released by shippers on the discriminating pipeline.=20 [*335]

B. Examining Shares or Examining = Barriers?=20

The HHI standard of the DOJ's Guidelines determines when a=20 relevant market can likely endure a horizontal merger = without=20 noteworthy effects on post-merger price. If a merger meets the HHI = standard (and=20 the underlying economic theory is correct), the DOJ can avoid a more = speculative=20 examination of competitive entry. If the DOJ wishes to find whether the = merger=20 will adversely affect any relevant market, it is = appropriate to=20 examine concentration first and barriers to entry later, if at all. n152 = To=20 evaluate the competitive effects of pipeline MBRs, the = FERC=20 should reverse that sequence. n153 Capacity release and interruptible = reversion=20 create the opposite of an entry barrier. A pipeline = with idle=20 capacity faces regulatory sanctions if it does not bring that capacity = to=20 market. If shippers can economically substitute among services of = different=20 terms and firmness (which depends on their prices), they constrain the=20 pipeline's monopoly power in ways that do not occur in=20 unregulated industries.

A pipeline's efforts to = exact a=20 high price for renewal of an expiring firm capacity contract will be=20 self-defeating. A high quoted price raises the opportunity costs of = other=20 capacity holders and encourages them to release. If a = pipeline=20 attempts to withhold larger amounts of firm capacity, its interruptible = service=20 becomes more reliable and a better substitute for firm service. Unless, = contrary=20 to experience, shippers view the two services as poor substitutes = regardless of=20 their price and reliability, the attempted monopolization will be = futile. A=20 customer thus coerced into interruptible service can cushion the pain by = using=20 some smaller amount of released firm capacity, making contingency and = exchange=20 arrangements behind its city gate, and arranging for storage in both = producing=20 and consuming areas. n154 The customer need not even substitute gas for = gas. It=20 can invest in financial instruments and insurance policies to deal with = either=20 price or deliverability risks. n155

Barriers to entry also = provide a=20 helpful perspective for examining possible analogies between=20 pipeline MBR policy and deregulation in other = industries. The=20 staff's 1995 report contains an appendix on railroads, = telecommunications, and=20 airlines. n156 Broadly, the staff finds their perform- = [*336] ances=20 improved under deregulation, but also finds that deregulation in all = three=20 industries has brought with it some encounters with monopoly power. At = some=20 junctures, regulators and courts have accepted market definitions that = resemble=20 those being proposed for pipelines. Since these = industries are=20 in some ways similar to pipelines (e.g., network = structure and=20 high fixed to variable cost ratios), one might reason that those = relatively=20 successful deregulations bode well for MBR determinations that rely on = the=20 staff's proposed markets. n157 None of the deregulated industries, = however, is=20 characterized by open access, secondary markets, and mandatory offers of = unused=20 capacity.

One final potential "entrant" into deregulated markets = is=20 unlikely and unneeded. Specifically, large buyers with monopsony power = may=20 exercise countervailing power against monopolistic = pipelines.=20 n158 Economics provides no general theory of bilateral monopoly, and no = reason=20 to expect that the outcome of such bargaining will be desirable. The = typical=20 "power buyer" from a pipeline is an LDC that may have = no good=20 substitutes for firm transportation. n159 If the LDC is a highly = inelastic=20 demander of firm transportation rights to cover its worst-day situation, = it will=20 have few weapons with which to fight the victimization. It is also = unclear why=20 an LDC should care. n160 It recovers the pipeline's = charges in=20 rates to captive customers and cannot earn more than a regulated return. = By=20 holding large amounts of firm capacity that it chooses not to release = (or to=20 release only with onerous recallability provisions), the LDC can exert = localized=20 monopoly power against end-users behind its city gate. Economic = misallocations=20 and monopolizations by LDCs, however, are largely state regulatory = issues. n161=20

C. Undersizing

If a pipeline = cannot=20 credibly withhold capacity, its monopolistic potential shrinks and = becomes=20 unrelated to statistics of concentration. Unless users of long-term firm = capacity have a general inability to substitute against it, release = markets=20 allow them to trade their holdings in accordance with their individual=20 valuations, and open-access network markets allow them to do so on other = pipelines. A case that substitution by customers is = difficult,=20 might be a case against MBRs. Instead of providing a theoretical or = empirical=20 study of that subject, the staff approaches the withholding = [*337] =20 problem by proposing a hitherto unstated objection to MBRs. = Specifically, the=20 staff claims that if MBRs are in effect, builders of new=20 pipelines will undersize them. n162

In the = staff's=20 illustrative example of undersizing, ABC is a hypothetical interstate=20 pipeline seeking MBRs for firm transportation.=20

ABC Pipeline might also allege = [before=20 the FERC] that released capacity on its own system and on other=20 pipelines would provide good alternatives for [a = certain LDC in=20 this example]. However, in one very important respect released capacity, = especially on ABC Pipeline itself, will have little, if = any,=20 impact on the assessment of ABC Pipeline's underlying = market=20 power in the primary long-run FT market. An analogy might help. Suppose = there=20 were only one manufacturer of automobiles, but robust used-car and = leasing=20 markets. Would the manufacturer have monopoly power? Yes. Even with a = perfectly=20 competitive secondary market for automobiles, the manufacturer could = "contrive"=20 a scarcity by making fewer new automobiles and charging a higher price = than=20 necessary to cover costs.

Similarly, if a = pipeline has=20 market power, it would exploit it by "contriving a scarcity." Although a = pipeline with a well-functioning capacity release = program might=20 not withhold existing capacity, it could choose not to expand. Customers = can=20 only release capacity they don't need; they can't build. As demand = grows, a=20 pipeline with market power could simply enjoy higher = prices and=20 refuse to build even if its customers were willing to pay the = incremental cost=20 of expansion. It would build only when the market clearing price for FT = went=20 above the monopoly price. n163

If undersizing becomes a = problem,=20 there is a relatively unintrusive solution consistent with MBRs: If any = entity=20 (including speculators) wishes to bear the cost of a = pipeline's=20 capacity expansion and the pipeline refuses, allow the = FERC to=20 order construction. The customer bears the risk of the investment made = at its=20 request, and can enjoy the value of a successful project. The financier = can act=20 as an entrepreneur in marketing the space, in competition with all other = holders=20 of rights. n164 In the staff's hypothetical example [*338] = of an=20 unregulated automobile manufacturer, no government agency has the power = to order=20 construction.

At the time of construction, nobody, including the = FERC,=20 knows how the demand for an unconstructed pipeline's = services=20 will vary over a pipeline's long lifespan. n165 Even if = the=20 statistical range of future demand is known, there is no easy formula = from which=20 to estimate the economically efficient size of the line. From society's=20 viewpoint, a single large line built today is not necessarily a better = use of=20 resources than smaller, seemingly duplicative, investments spread over = the=20 future. The large line will foreclose future alternative uses for = resources that=20 will be of no value to pipeline users for years. = Likewise, it=20 is wasteful to build an overly large line whose full capacity is seldom=20 utilized. Additions to existing facilities such as looping or = compression can=20 substitute for new construction. Open access and new market institutions = also=20 provide substitutes for new construction, because they offer choices of=20 transportation paths and prices before them that were impossible under = the old=20 regime. In a network market, capacity additions by = pipelines=20 not directly connected with an origin-destination pair can undermine = monopoly=20 power held by pipelines that link those points. =

MBRs=20 provide signals of scarcity that cost-based rates cannot, both to = allocate=20 today's capacity and to indicate where new facilities are most valuable. = If=20 pipeline users expect that capacity on some link will = be scarce=20 a year from now, capacity releases extending beyond a year will carry = premium=20 prices that anticipate that scarcity. This price signal gives an early = warning=20 to investors that construction of competing facilities or expansion of = existing=20 ones may be warranted. Cost-of-service rates reveal nothing about where = users=20 expect new facilities to be valuable, or how valuable those facilities = might be.=20 The competitive consequences of MBRs extend both geographically and into = the=20 future. Suppressing price signals with a cost-of-service ceiling on any = type of=20 capacity transaction increases the risk that new capacity will be built = in the=20 wrong places, or in the wrong amounts. [*339]

The=20 relationship between competitive entry and undersizing is itself = unclear.=20 Gallick's finding that spurs from nearby pipelines to = consuming=20 areas can be constructed economically and quickly led to his = recommendation that=20 the FERC give thought to MBRs. n166 The FERC staff offers no rationale = for its=20 recommendation that the easy entry of competitors within a year is a = necessary=20 precondition for MBRs in markets that do not meet its HHI standard. n167 = If=20 entry takes more than a year, the staff appears to be saying that it is = worse to=20 have two years of monopoly rates followed by competition than to have = rates set=20 by historical cost for the indefinite future.

An alternative = theoretical=20 perspective gives reason to question the FERC's interest in undersizing = and=20 capacity withholding. n168 A monopolist's best strategy for maximizing = long-term=20 wealth may entail oversizing its facility beyond the economically = efficient=20 level at the time of construction. Such an investment can protect some = of the=20 monopolist's profit against potential builders of competing capacity. = Simply=20 threatening a price war against newcomers is insufficient, since the = incumbent's=20 best response to actual entry will be to accommodate the entrant, and = the=20 entrant knows this. To credibly threaten a price war that will ruin the = entrant,=20 the incumbent should build an oversized plant with low operating costs. = With an=20 unrecoverable investment in excess capacity, the incumbent can produce = large=20 quantities if entry occurs and sell each unit at slightly over operating = cost.=20 If forced to sell at such a low price, the potential entrant cannot = recover the=20 cost of a new plant, and will choose not to build it. Undersized = facilities=20 promise a profit to competitive entry rather than a loss. This theory is = probably of little relevance to pipelines, since under = open=20 access an attempt at preemption by oversizing will probably be = self-defeating.=20 All of the capacity in an oversized facility must reach the market, = where in the=20 early years it can only sell at prices that do not cover the builder's = cost.=20

D. Monopoly Power, Substitutes, and Prices

If = the FERC=20 reorients itself to market analyses that start from barriers to entry, = as we=20 recommend, it needs a method for identifying these barriers and = determining=20 their likely effects in an MBR regime. Competition is competition among=20 alternatives, for both buyers and sellers. Access to alternatives is = determined=20 by the range of economic substitutions, and cost- [*340] = of-service=20 rates harm competition to the extent that they block either buyers or = sellers=20 from exercising their powers to make substitutions. Some of the best = guidance=20 for future policies toward competition is already on the record in the = form of=20 observed changes in markets since the inception of open access. Taken in = conjunction with the evidence on market unification, they indicate = competition=20 of substantial scale and diversity. If this picture of the industry is = an=20 accurate one, when the FERC institutes MBRs it will be acknowledging = market=20 reality rather than changing it.

1. Which Products = Matter?=20

If substitution defines the relevant market in = which=20 MBRs might affect competition, that market is the one in which=20 pipeline customers arrange for supplies that are = optimal from=20 their individual viewpoints. Supply is a package of gas purchases,=20 transportation, reliability, and financial arrangements. The optimal mix = of=20 supply-related contracts will vary with the prices of the package's = components.=20 With MBRs for pipelines and competitive conditions = elsewhere,=20 the price of any service will equal the marginal buyer's willingness to = pay.=20 n169 Buyers whose situations do not allow easy substitution will have a = higher=20 willingness to pay, but in a competitive market they will pay the same = price as=20 those buyers who are on the brink of switching.

In its studies = of MBRs,=20 the FERC has been most concerned with firm and interruptible = transportation.=20 This concern does not equate to economic relevance. "Firm" and = "interruptible"=20 are regulatory pigeonholes into which pipelines = allocate costs.=20 With MBRs, the market-clearing prices for those two forms of = transportation=20 might be quite distant from costs allocated according to existing rules. = At the=20 new prices, consumers might also have quite different views about = substituting=20 between firm and interruptible service, and about substituting either = for the=20 other services that comprise a supply package. Under MBRs, firm and=20 interruptible transportation, as the FERC knows them, will survive only = in=20 competition with other rate and service offerings, some of which will be = marketed by parties releasing capacity. Economic choices for=20 pipeline users will depend on market conditions, which = are=20 themselves determined alongside the prices of the various services that = make up=20 a supply package. Some prices may be contractually fixed or relatively=20 inflexible, while others may allow short-term flexibility.

With=20 heterogeneous customers and the heterogeneous service offerings that = will=20 probably arise under MBRs, the analysis of competition requires = additional care,=20 particularly when comparing prices that are under longer-term contracts = from=20 those that are not. In its example, the staff states that interruptible = and firm=20 service are only substitutes for a hypothetical LDC if interruptible=20 transportation (IT) is of sufficiently high qual- [*341] = ity, as=20 measured by the probability of interruption. The = pipeline=20 seeking MBRs would "need to present evidence that IT was provided at a = price=20 that rendered the price of delivered gas using IT at or below the price = of=20 delivered gas using FT (firm transportation)." n170 A higher price for=20 infrequent IT deliveries than for everyday FT deliveries is quite = consistent=20 with competition. n171 The LDC's least-cost strategy might be to avoid = the high=20 cost of holding firm capacity for its worst-day requirements, at the = occasional=20 risk of an exorbitant price for interruptible delivery. n172 The LDC's = decision=20 against full reliance on firm capacity indicates that IT is a substitute = for it=20 at market-determined prices. n173 By the staff's quality standard, a = market may=20 be competitive only when customers are buying supplies of a quality that = would=20 be too expensive for them at MBRs. If state regulators insist that an = LDC hold=20 an inefficiently large amount of firm interstate capacity, they will = create=20 economic misallocations with or without interstate MBRs. Such a state = policy=20 will lead to high prices for firm interstate service under MBRs, but = such prices=20 are scarcely a reason for the FERC to deny MBRs to the = pipeline=20 in question.

2. Patterns of Substitution =

Long-term firm=20 service deserves primacy in market analysis only if an open-access=20 pipeline can meaningfully monopolize it. Virtually all = of the=20 industry's experience since open access indicates that economic = substitutes for=20 long-term firm service are abundant. Since open access, all relevant = facets of=20 the gas industry have moved from long-term arrangements to short-term = ones.=20 Because gas has become an ordinary commodity, purchasers construct = increasing=20 fractions of their supplies under short-term arrangements (longer-term = contracts=20 usually reference price changes to the spot market). The market's = ability to=20 alter prices, production, and shipping patterns as the economic = environment=20 changes allows traders to deal with contingencies that would otherwise = require=20 long-term contracts. Today's gas price efficiently incorporates = information=20 about today's market conditions and traders' expectations of the future. = n174=20 Long-term fixed-price contracts are superior only if individual traders = are=20 better predictors than the [*342] market. As predictors of = future=20 prices, spot and futures markets have substantially outperformed = long-term=20 fixed-price contracts. n175

The FERC imposed open access on an = industry=20 dominated from birth by long-term full-requirements contracts. Those = contracts=20 may have been economically efficient for an industry without open = access. n176=20 Long-term contracts save the costs of repetitive contracting and can = deter=20 opportunism by parties who have made large and highly specialized = investments.=20 n177 Pre-open access, a producer could best protect its exploration and=20 production investments with a pipeline contract whose=20 provisions would be in effect for the life of its wells. A pre-open = access=20 pipeline would contract with the producer because it = needed a=20 dependable source of gas to resell at predictable prices under its = long-term=20 contract with an LDC. A long-term requirements contract between the=20 pipeline and the LDC would likewise facilitate payoff = of the=20 location-specific investments that each of the parties had made in order = to=20 serve and be served. Under open access, long-term requirements contracts = become=20 the exception. The producer now faces a nation of alternative buyers = reachable=20 by pipelines that cannot refuse to provide service. The = LDC has=20 a similar richness of competitive gas suppliers. If both parties have = the=20 abundant alternatives of competition, each will probably prefer = short-term=20 transactions, hedged according to their individual preferences. The = long-term=20 fixed-price LDC contracts that remain are overpriced insurance policies = that=20 persist because LDCs can transfer their risks to captive end-users. n178 =

An LDC or end-user in the unified market will probably be buying = spot=20 gas under short-term agreements. Someone who purchases gas on these = terms from=20 different production areas will place little value on long-term firm=20 transportation over a given pipeline segment. = Inflexible=20 long-term transportation and flexible short-term gas purchases are = unlikely to=20 be a useful combination. n179 Long-term firm capacity reservation will = be=20 economic in a market dominated by types of long-term gas contracts that = are=20 [*343] in eclipse. The staff's hypothetical of an LDC with a = highly=20 inelastic demand for long-term firm capacity is fast becoming = counterfactual.=20 Recently, and more so over the near future, long-term capacity contracts = between=20 pipelines and LDCs dating from before open access will = expire.=20 n180 The available evidence is that those contracts will be replaced by=20 shorter-term arrangements, for volumes that do not necessarily give the = LDC=20 "worst day" firm rights. n181 If pipelines prefer the = income=20 guarantees of long-term firm capacity contracts, their inability to = renew such=20 contracts on the old terms is another indication of the growth in = competition=20 that they face under open access.

Transportation service has = responded=20 to institutional changes. In the early years of open access, = interruptible=20 transportation became the choice of end-use transporters who could not = obtain=20 firm interstate rights from their LDCs. Likewise LDCs used interruptible = service=20 to reach short-term opportunities outside of their usual supply areas. = As=20 capacity release markets have become organized, their growth has = diminished the=20 role of interruptible transportation. Shippers have become increasingly = able to=20 obtain shorter-term (and sometimes long-term) firm capacity, and the = fraction of=20 interstate throughput moving by interruptible transportation has begun = to fall.=20 n182 One pipeline applying for MBRs has told the FERC = that the=20 growth of capacity release on its system is evidence that it lacks = market power=20 in interruptible transportation. n183

Capacity releases provide = evidence=20 of competitive substitution and evidence of the forces at work to = circumvent=20 restrictions on market activity. Where users value released capacity at = more=20 than the legal ceiling rate, gray market transactions have increased. = There is=20 growing awareness that the caps inhibit desirable competition, = accompanied by a=20 growth in policy discussions about their elimination. n184 On the other = side,=20 pipelines with excess capacity to offer (and capacity = release=20 markets on these pipelines) frequently discount prices = below=20 the maximum that the FERC allows them to recover. n185 While not itself = evidence=20 against monopoly power, discount- [*344] ing suggests a test = of the=20 FERC's structural theory. n186 After correcting for differences in = market=20 conditions, if the FERC's theory is true, discounting should be less = frequent or=20 less deep in areas with high HHIs, other things equal. One recent = statistical=20 study finds no differences in the depth of discounts offered at city = gates=20 served by a few pipelines (i.e., high HHI) and those = served by=20 many. Additionally, that study found that seasonal differences in = discounting=20 were insignificant, indicating that pipelines could not = even=20 exact higher prices for deliveries in peak heating periods. n187 =

Growing=20 markets for unbundled services further constrain the pricing of = transportation.=20 Unbundling allows pipeline users greater freedom to = choose the=20 size, interruptibility, and timing of their transportation. Capacity = release has=20 opened up new types of markets and ways of linking them. n188 In gas = itself, the=20 bidweek spot market is being joined by a growing swing market for those = who wish=20 to change their gas purchase or delivery plans. Increased availability = of=20 storage allows the substitution of off-peak pipeline = service=20 for on-peak, short-term firm capacity for long-term, and interruptible = service=20 for firm.

Market centers have grown both in number and in the = range of=20 services offered. n189 These services include gas and transportation = exchanges=20 (sometimes including market making), imbalance trading, scheduling, = accounting,=20 and short-term storage injections ("parking") and withdrawals = ("advances").=20 There is no obvious limit to the locales that might serve as market = centers.=20 n190 The growth of access to pipeline electronic = bulletin=20 boards mandated under Order 636 (and growth in their standardization=20 [*345] and detail) makes it possible for a wider set of = users to=20 evaluate and acquire alternative transportation paths. Although they are = not=20 parts of gas supply technology, the growing markets for gas futures, = options,=20 and derivative contracts further facilitate transactional flexibility. = They give=20 traders and speculators new instruments for coping with all types of = risk,=20 including risks that transportation will be unexpectedly overpriced or=20 unavailable.

VI. Conclusions

Cost-of-service = regulation=20 does not produce competitive prices. The only place to find those prices = is in=20 competitive markets. If MBRs for interstate pipelines = replace=20 regulation, the FERC must gain a reasonable expectation of whether the = markets=20 that develop will be competitive or monopolistic. The tools the FERC has = chosen=20 to use, however, are those of an antitrust enterprise whose subject = markets and=20 policy goals are not the FERC's. Application of the antitrust tools to = regulated=20 industries, and particularly to pipelines, will be=20 inappropriate at best and disastrous at worst. Both the antitrust = paradigm and=20 the NGA tell the FERC to determine competition by looking at the parties = it=20 regulates. In most markets, this makes sense, since corporate owners of=20 facilities make decisions about their output and price. In=20 pipelines, it does not. Under open access, their owners = cannot=20 decide to put less than all of their capacity on the market, and they = must trade=20 it in competition with customers and others who participate in the = release and=20 interruptible markets. The FERC must look behind the corporate veil of=20 pipeline ownership.

In a regulated natural = monopoly,=20 efficiently served markets will have only a small number of=20 pipelines to choose from. If pipelines = in an=20 area do not exhibit a high HHI, regulators have done their job poorly, = because=20 they have certificated an inefficiently large number of inefficiently = small=20 pipelines. The lines themselves are technological = natural=20 monopolies, but this fact is of little value for market analysis. = Competition is=20 feasible if space in them can be allocated by a market, as in capacity = release,=20 and if owners cannot withhold capacity, as is guaranteed by = interruptible=20 reversion. It is important to distinguish the structure of a market from = the=20 institutions under which people trade in it. In some cases, such as = bidweek for=20 spot gas (and most other unregulated markets), structure and = institutions match=20 in one important respect. There are many independent buyers, sellers, = and=20 speculators, each with its own corporate identity. The available = theories of=20 competition in this market may be infirm, but there is no question of = who is=20 executing the trades and might exercise market power. In interstate=20 pipelines, structure and institutions do not match. = Under open=20 access, the pipeline whose capacity is being traded = gives up=20 important rights over that capacity, and in doing so ceases to be = identifiable=20 as the sole seller of it. The FERC mistakenly treats the=20 pipeline as the sole seller and calculates market = shares on=20 this basis for sales of capacity that it cannot refuse to market.=20 [*346]

There are costs of regulating and costs of = not=20 regulating. The FPC imposed cost-based wellhead price controls on a = naturally=20 competitive producing sector and generated shortages that caused far = more loss=20 to the economy than they saved for gas consumers. Wellhead prices were = the final=20 prices to be controlled in an industry where regulation had frozen all = of the=20 other sources of market flexibility. Interstate = pipeline rates=20 seem to present the obverse. Wellhead decontrol and open access have = brought=20 highly competitive markets to all other important sectors of the = industry. n191=20 The wealth of options available to the participants in those markets has = allowed=20 them to mitigate some adverse effects of cost-based = pipeline=20 rate regulation. They can contract around monopoly by choosing gray = market=20 transactions, alternative paths in the network, unbundled substitutes = for firm=20 transportation, and financial hedges. Given the observed patterns of = trade and=20 substitution, we must seriously entertain the possibility that the = interstate=20 pipelines already operate predominantly under an MBR = system,=20 and that the FERC can best adapt to the reality by officially adopting = an MBR=20 policy. If we do have MBRs, however, the competitive forces that act to=20 determine them bear no discernible relationship to the = origin-destination=20 concentration statistics on which the FERC bases its analysis of = competition.=20

FOOTNOTES:
n1. = Regulators=20 must, however, acknowledge that the regulated firm's costs are = determined in=20 markets where it must bid against other purchasers of inputs into = production.=20

n2. For a discussion of the arbitrariness of regulatory = allocations of=20 cost, see William J. Baumol et al., How Arbitrary is "Arbitrary"? - or, = Toward=20 the Deserved Demise of Full Cost Allocation, Pub. Utils. Fort., Sept. 3, = 1987,=20 at 16, 16-19.

n3. See, e.g., William J. Baumol & J. Gregory = Sidak,=20 Transmission Pricing and Stranded Costs in the Electric Power Industry = 21-23=20 (1995). See also Northern Natural Gas Co. v. FPC, 399 F.2d 953, 959 = (D.C. Cir.=20 1968).

n4. If regulators accept forecasts that the future will = differ=20 substantially from the past, they may allow automatic adjustments or = costs=20 projected on the basis of a "future test year."

n5. For the = history of=20 wellhead regulation and its consequences, see Richard J. Pierce, Jr.,=20 Reconstituting the Natural Gas Industry from Wellhead to Burnertip, 9 = Energy=20 L.J. 1 (1988); Arlon R. Tussing & Connie C. Barlow, The Natural Gas=20 Industry: Evolution, Structure, and Economics 59-114 (1984).

n6. = Phillips Petroleum Co. v. Wisconsin, 347 U.S. 672 (1954).

n7. = Even if=20 there were compelling reasons in equity to transfer wealth from = producers to=20 consumers, less disruptive methods were certainly available to effect = the=20 transfer.

n8. Armen A. Alchian & William R. Allen, = University=20 Economics 56 (3rd ed. 1972).

n9. Alfred Marshall, Principles of=20 Economics 324 (9th variorum ed. 1920).

n10. The capacity will = not be=20 wasted if the owner can discriminate among customers, offering otherwise = unused=20 capacity at low prices while extracting high prices from those willing = to pay.=20

n11. Economists' interest in these so-called deadweight losses = (losses=20 of exchange value gained by no one) may not be commensurate with their = empirical=20 importance. A franchised seller, for example, may waste substantial = wealth=20 (possibly up to expected monopoly profit) competing against other = applicants=20 (who are also wasting wealth) to obtain the franchise. For theory and = empirics=20 of this phenomenon see Richard A. Posner, The Social Costs of Monopoly = and=20 Regulation, J. Pol. Econ., Aug. 1975, at 807, 827; Thomas W. Hazlett = &=20 Robert J. Michaels, The Cost of Rent Seeking: Evidence from Cellular = Telephone=20 License Lotteries, 59 S. Econ. J. 425, 425-35 (1993).

n12. = Clayton Act,=20 15 U.S.C. 12(a)(1988).

n13. 15 U.S.C. 18 (1988). The = Hart-Scott-Rodino=20 Antitrust Improvements Act of 1976 allowed the government to challenge = most=20 mergers before the fact. 15 U.S.C. 18(a) (1988).

n14. The = Guidelines=20 have evolved since their 1982 unveiling, but their analytical methods = and=20 numerical criteria are little-changed. U.S. Department of Justice Merger = Guidelines, 4 Trade Reg. Rep. (CCH) 13,101 (1984) [hereinafter Merger = Guidelines=20 1984]; Merger Guidelines, 49 Fed. Reg. 26,823, reprinted in 4 Trade Reg. = Rep.=20 (CCH) 13,102 (1984); U.S. Department of Justice Horizontal Merger = Guidelines, 57=20 Fed. Reg. 41,552, reprinted in 4 Trade Reg. Rep. (CCH) 13,104 (1992)=20 [hereinafter Merger Guidelines 1992].

n15. For a comprehensive = survey of=20 the topic, see Gregory J. Werden, The History of Antitrust Market = Definition, 76=20 Marq. L. Rev. 123 (1993).

n16. For a description of how = numerical=20 criteria affect the tradeoff between the two errors, see John R. Morris = &=20 Gale R. Mosteller, Defining Markets for Merger Analysis, 36 Antitrust = Bull. 599,=20 622-32 (1991).

n17. Merger Guidelines 1992, supra note 14, 1.0.=20

n18. Merger Guidelines 1984, supra note 14, 2.11. The standard = is a=20 policy choice that does not stem from any theoretical importance of the = numbers=20 chosen. Both the deadweight loss and the potential price increase depend = on=20 elasticities of supply and demand that relevant market=20 definition does not account for explicitly.

n19. Oliver E. = Williamson,=20 Economies as an Antitrust Defense, 58 Am. Econ. Rev., Mar. 1968, at 18, = 18-36.=20

n20. Id. In a merger case, only a likelihood, as opposed to a = certainty,=20 of the price increase need be shown. Brown Shoe v. United States, 370 = U.S. 294,=20 323 (1962); United States v. General Dynamics, 415 U.S. 486, 505 (1973). =

n21. The cross-elasticity of demand for A with respect to the = price of B=20 is the percentage change in purchases of A when the price of B changes = by one=20 percent, all else equal. High values indicate easy substitution between = the=20 goods. See generally Richard A. Posner, Antitrust Law: An Economic = Perspective=20 125-30 (1976), for details and critiques of its use.

n22. The = geographic=20 market should consist of an area in which the defendants operate and = which the=20 plaintiff can reasonably turn to for supplies. Tampa Elec. Co. v. = Nashville Coal=20 Co., 365 U.S. 320, 327 (1961).

n23. Elasticity of supply is = defined as=20 the percentage response of output in a market for a one percent increase = in=20 market price. Note that this definition loses value if producers have = power to=20 set the price.

n24. Kenneth G. Elzinga & Thomas F. Hogarty, = The=20 Problem of Geographic Market Delineation in Antimerger Suits, Antitrust = Bull.,=20 Spring 1973, at 45, 45-81.

n25. If it is feasible, price = discrimination=20 will raise profits above those achievable under a single-price policy, = and will=20 also yield a smaller shortfall from the market's former competitive = output. We=20 do not consider price discrimination further in this section. =

n26. See,=20 e.g., Philip Areeda & Louis Kaplow, Antitrust Analysis (4th ed. = 1992);=20 William M. Landes & Richard A. Posner, Market Power in Antitrust = Cases, 94=20 Harv. L. Rev. 937 (1981); Comment, Landes and Posner on Market Power: = Four=20 Responses, 95 Harv. L. Rev. 1787, 1848 (1982).

n27. We are = assuming that=20 the monopolist either faces no potential entrants or that it chooses = price in=20 disregard of their likely reactions.

n28. Landes & Posner, = supra=20 note 26, devote much of their theoretical analysis to markets of this = type.=20

n29. See, e.g., Merger Guidelines 1992, supra note 14, 1.5 = (1992); ANR=20 Pipeline Co., 56 F.E.R.C. 61,293 (1991). In this gas = inventory=20 charge (see section B infra) proceeding and others, the FERC chose a ten = percent=20 threshold price increase, for reasons that are unclear.

n30. = Merger=20 Guidelines 1992, supra note 14, 1.51 (1992). Mergers that produce a = post-merger=20 HHI exceeding 0.18 are subject to case-specific investigation. = Algebraically, an=20 HHI exceeding that value arises in a market with approximately five=20 equally-sized sellers.

n31. Dennis W. Carlton & Jeffrey M. = Perloff,=20 Modern Industrial Organization 229-80, 382-430 (2nd. ed. 1994). =

n32.=20 Robert D. Willig, Merger Analysis, Industrial Organization Theory, and = Merger=20 Guidelines, in Brookings Papers on Economic Activity, Microeconomics = 1991, at=20 281, 287.

n33. Elasticity of demand measures the price = sensitivity of=20 buyers. It is the percentage change in the aggregate quantity they will = purchase=20 when faced with a one percent change in price. Higher absolute values = indicate=20 greater responsiveness.

n34. Costs here include the "normal" = profit=20 foregone by investors who have chosen to place their capital in this = industry=20 rather than their best expected alternative.

n35. This can be = seen by=20 noting that in a monopolized market H =3D 1 and conjectural variation is = irrelevant because the monopolist has no rivals.

n36. If sellers = hold=20 expectations about one another that are at variance with actual = behavior, the=20 theoretical difficulties become virtually intractable. See, Dennis W. = Carlton=20 and Jeffrey M. Perloff, Modern Industrial Organization 248-69 (Harper = Collins,=20 2d ed. 1994).

n37. For examples of the theoretical usefulness of = Cournot-Nash, see Joseph Farrell & Carl Shapiro, Horizontal Mergers: = An=20 Equilibrium Analysis, 90 Am. Econ. Rev. 107, 126 (1990). If demand is = linear and=20 marginal costs are constant, a Cournot-Nash market containing N firms = produces=20 (N-1)/N percent of the output of a similar perfectly competitive market. =

n38. For two views, compare Franklin Fisher, Games Economists = Play: A=20 Noncooperative View, Rand J. Econ. 113, 124 (1989) with Carl Shapiro, = The Theory=20 of Business Strategy, Rand J. Econ. 125, 137 (1989).

n39. The = assumption=20 may not be reasonable. If all sellers have identical costs, as = postulated in the=20 equation, it may well be possible to easily open a firm with similar = costs that=20 will enter the market and affect the price-cost margin.

n40. See = Carlton=20 & Perloff, supra note 36, at 382, 415.

n41. Noel D. Uri = &=20 Malcolm Coate, The Department of Justice Merger Guidelines: The Search = for=20 Empirical Support, 7 Int'l Rev. L. & Econ. 113, 120 (1987). =

n42. The=20 mass of empirical studies on concentration-profits and = concentration-performance=20 relationships (in unregulated industries) also provides little rationale = for the=20 FERC to rely on concentration measures, even as a preliminary screen. A=20 comprehensive survey (cited in the Staff's 1993 bibliography) concludes: =

[Concentration] profit studies had been the mainstay of=20 industrial organization research for more than two decades, but they are = now=20 largely discredited. The numerous studies of the relationship between = price and=20 concentration within particular industries tended to find the = relationship to be=20 positive and statistically significant but typically not particularly = important=20 quantitatively.

Gregory J. Werden, Antitrust Div., U.S. = Dep't of=20 Justice, A Review of the Empirical and Experimental Evidence on the = relationship=20 between Market Structure and Performance, Economic Analysis Group = Discussion=20 Paper EAG 91-3 (1991). This document does not necessarily reflect the = views of=20 the Department of Justice.

n43. Harold Demsetz, Industry = Structure,=20 Market Rivalry, and Public Policy, 16 J.L. & Econ., April 1973, at = 1, 10;=20 Sam Peltzman, The Gains and Losses from Industrial Concentration, 20 = J.L. &=20 Econ., Oct. 1977, at 229, 264.

n44. For more general discussions = of=20 antitrust in regulated industries, see Keith S. Watson and Thomas W. = Brunner,=20 Monopolization by Regulated "Monopolies': The Search for Substantive = Standards,=20 22 Antitrust Bull. 559 (1977); Andrew N. Kleit & Robert J. Michaels, = Antitrust, Regulation, and Rent-Seeking: The Past and Future of Otter = Tail, 39=20 Antitrust Bull. 689, 725 (1994).

n45. This statement presupposes = that=20 the server cannot price discriminate so that each customer becomes a = profitable=20 addition to its clientele, in which case all those who would be served = under the=20 regulation will also be served in an unregulated market.

n46. = Service=20 obligations do not necessarily preclude such other inefficiencies as = failing to=20 produce at the lowest possible cost.

n47. The regulated firm may = compete=20 for consumer dollars with unregulated sellers who produce substitute = goods that=20 are not in the regulatory jurisdiction. When regulators control their = domain,=20 they also indirectly control its competitive posture toward unregulated = sellers.=20

n48. Some economists endorse franchise competition as a method = by which=20 a jurisdiction's residents can capture the cost savings of natural = monopoly for=20 themselves by taking competitive bids. See Harold Demsetz, Why Regulate=20 Utilities?, 11 J.L. & Econ., April 1968, at 55, 66. Others believe = that the=20 contracting process itself may be too costly relative to the expected = benefits,=20 and that it will invite opportunistic behavior by both the government = and the=20 incumbent server. See, e.g., Oliver Williamson, Franchise Bidding for = Natural=20 Monopolies - In General and with Respect to CATV, 7 Bell J. Econ., = Spring 1976,=20 at 73, 104.

n49. In reality, electricity franchises seldom turn = over,=20 and some franchise changes reflect no more than the effects of = tax-exempt=20 finance and the preferential availability of inexpensive = federally-produced=20 power to municipal utilities. See Robert J. Michaels, Deregulating = Electricity:=20 What Stands in the Way, 15 Reg., Winter 1992, at 38, 47.

n50. = The=20 quality of service might change, e.g., if after entry the typical = customer has a=20 shorter wait for a taxi. This outcome, however, would also occur without = changing the HHI if each existing company expanded its fleet. =

n51. 410=20 U.S. 366 (1973).

n52. Sources for all of the text's factual = statements=20 on Otter Tail are provided in Kleit & Michaels, supra note 44. =

n53.=20 Here we disregard the remarks of Section A, supra, regarding the = unlikely=20 equality of regulated and competitive prices.

n54. A more = detailed=20 survey of Otter Tail's descendants appears in Kleit & Michaels, = supra note=20 44, at 714, 724.

n55. These included Florida Power & Light = Co., 9=20 F.E.R.C. 61,366 (1979); Kentucky Utils. Co., 23 F.E.R.C. 61,317 (1983); = and=20 Pacific Power and Light Co., 26 F.E.R.C. 63,048 (1984).

n56. = Federal=20 Power Act, 16 U.S.C. 791a (1988) [hereinafter FPA].

n57. The NRC = can=20 impose orders to wheel as licensing conditions for a nuclear powerplant = if it=20 finds inconsistency with the antitrust laws. 42 U.S.C. 2135(c)(5) = (1988). The=20 NRC decisions included language to the effect that the regulated status = of=20 applicants and intervenors did not vitiate the inferences of monopoly = power from=20 market shares, for reasons that were not made explicit. The courts later = stated=20 that the NRC's antitrust language was not necessarily to be construed as = equivalent to that of antitrust law. Alabama Power Co. v. Nuclear = Regulatory=20 Comm'n, 692 F.2d 1362 (11th Cir. 1982), cert. denied, 464 U.S. 816 = (1983).=20

n58. In ruling that, the Public Utility Regulatory Policies Act = of 1978,=20 Pub. L. No. 95-617, 92 Stat. 3117 (codified at 16 U.S.C. 2601-2645 = (1988)=20 [hereinafter PURPA], prohibited wheeling orders that promised to upset = existing=20 competitive relationships, the FERC also used such market definitions.=20 Southeastern Power Admin. v. Kentucky Utils. Co., 25 F.E.R.C. 61,204 = (1983).=20

n59. FPC v. Conway Corp., 426 U.S. 271 (1976). For a summary of = price=20 squeeze issues and references, see Paul L. Joskow, Mixing Regulatory and = Antitrust Policies in the Electric Power Industry: The Price Squeeze and = Retail=20 Electric Competition, Antitrust and Regulation: Essays in Memory of John = J.=20 McGowan (Franklin M. Fisher ed., 1985).

n60. Connecticut Light = and Power=20 Co., 8 F.E.R.C. 61,187 (1979).

n61. Town of Concord, Mass. v. = Boston=20 Edison Co., 915 F.2d 17 (1st Cir. 1990).

n62. Pub. L. No. = 102-486, 106=20 Stat. 2776, 2905 (codified at 42 U.S.C.A. 13,201-13,556 (West Supp. = 1995)).=20

n63. Merger opinions containing market share analyses include = among=20 others, Opinion No. 318, Utah Power & Light Co., 45 F.E.R.C. 61,095 = (1988);=20 Southern California Edison Co., 47 F.E.R.C. 61,196 (1989); Opinion No. = 364,=20 Northeast Utils. Serv. Co. (re Public Serv. Co. of New Hampshire), 56 = F.E.R.C.=20 61,296 (1991); Opinion No. 385, Entergy Servs., Inc., 65 F.E.R.C. 61,332 = (1993);=20 Cincinnati Gas and Elec. Co., 64 F.E.R.C. 61,237 (1993); Midwest Power = Systems,=20 Inc., 71 F.E.R.C. 61,386 (1995). Power marketing opinions include, = Opinion No.=20 349, Public Serv. Co. of Indiana, 52 F.E.R.C. 61,260 (1990); Entergy = Servs.,=20 Inc., 58 F.E.R.C. 61,234 (1992).

n64. E.g., In Utah Power & = Light,=20 the FERC determined that firm bulk power, nonfirm bulk power, firm = transmission,=20 and nonfirm transmission were relevant markets. 45 = F.E.R.C.=20 61,095, at 61,284. In Entergy Servs., the markets were for short-term = firm bulk=20 power, long-term firm bulk power, and transmission. 58 F.E.R.C. 61,234.=20

n65. 58 F.E.R.C. 61,234, at 61,729.

n66. E.g., Pacific = Gas and=20 Electric Co., 44 F.E.R.C. 61,010 (1988).

n67. Kansas City Power = &=20 Light Co., 67 F.E.R.C. 61,183 (1994). In this and related power = marketing=20 dockets, the FERC has accepted the applicant's proposal subject to = filing of an=20 open-access transmission plan.

n68. E.g., Enron Power Enter. = Corp., 52=20 F.E.R.C. 61,193 (1990). The FERC sometimes examines the ratio of total = bids to=20 total acceptances in a utility procurement as a measure of competition, = a=20 process for which there is no rigorous economic justification. Id. at = 61,714-15.=20 The FERC has expressed concern over too small a ratio in one rejection = of MBRs.=20 See TECO Power Servs. Corp. and Tampa Elec. Co., 52 F.E.R.C. 61,191 = (1990).=20

n69. Cajun Elec. Power Coop. v. FERC, 28 F.3d 173 (D.C. Cir. = 1994). For=20 arguments that the provisions did not implement an anticompetitive tie, = see=20 Alfred E. Kahn, Can Regulation and Competition Coexist? Solutions to the = Stranded Cost Problem, 7 Elec. J., Oct. 1994, at 23, 35.

n70. = EPAct,=20 supra note 62, 1801(a).

n71. Fed. Energy Reg. Comm'n, Staff = Proposal for=20 Revisions to Oil Pipeline Regulation Pursuant to the = Energy=20 Policy Act of 1992 (1993) [hereinafter Staff Proposal].

n72. = Opinion No.=20 360, Buckeye Pipe Line Co. Ltd., 53 F.E.R.C. 61,473 (1990). The staff = views the=20 experiment as successful. Buckeye's prices in all markets that the FERC = deemed=20 competitive only rose by modest percentages, all by less than thresholds = that=20 could trigger the FERC's instituting a suspension or investigation. See = 1995=20 Staff Paper, Fed. Energy Reg. Comm'n, Market-Based Rates for Natural Gas = Companies, A Staff Paper, at 18 (1995) [hereinafter 1995 Staff Paper].=20

n73. Buckeye Pipe Line Co. Ltd., 50 F.E.R.C. 63,011 (1990). The = staff=20 followed the recommendations of a DOJ report on oil = pipelines=20 and recommended a critical HHI screen of 0.25 rather than the Merger = Guidelines'=20 0.18. That report broadly concluded that competition was sufficiently = strong in=20 all origin markets (and crude oil destination markets) in the contiguous = United=20 States and that cost-of-service regulation was not warranted. Staff = Proposal,=20 supra note 71, at 33. The DOJ's analysis appears in U.S. Dep't of = Justice, Oil=20 Pipeline Deregulation (May 1986).

n74. E.g., 53 = F.E.R.C. 61,473 (1990); Staff Proposal, supra note 71, at 36, citing = Buckeye, 53=20 F.E.R.C. 61,473, at 62,665.

n75. Regulation of Natural Gas=20 Pipelines After Partial Wellhead Decontrol, Order No. = 500,=20 [1986-1990] F.E.R.C. Stats. & Regs. 30,761, 52 Fed. Reg. 30,334 = (1987).=20

n76. Transwestern Pipeline Co., 43 F.E.R.C. = 61,240=20 (1988); Transcontinental Gas Pipe Line Corp., 55 F.E.R.C. 61,446 (1991); = El Paso=20 Natural Gas Co., 49 F.E.R.C. 61,262 (1989). The courts ruled that the = FERC could=20 only approve a market-based GIC contingent on a finding of competition. = See=20 Tejas Power Corp. v. FERC, 908 F.2d 998 (D.C. Cir. 1990).

n77. = E.g., El=20 Paso Natural Gas Co., 49 F.E.R.C. 61,262 (1989).

n78. E.g., = Richfield=20 Gas Storage System, 59 F.E.R.C. 61,316 (1992); Petal Gas Storage Co., 64 = F.E.R.C. 61,190 (1993).

n79. See Petal Gas Storage Co., 64 = F.E.R.C.=20 61,190, at 62,573.

n80. Koch Gateway Pipeline = Co., 66=20 F.E.R.C. 61,385, at 62,301 (1994).

n81. For a summary of other = early=20 work on MBRs see Dan Alger and Michael Toman, Market-Based Regulation of = Natural=20 Gas Pipelines, 2 J. Reg. Econ. 263, 280 (1990). =

n82.=20 Edward C. Gallick, Competition in the Natural Gas = Pipeline=20 Industry: An Economic Policy Analysis (1993).

n83. Gallick used=20 construction cost data to estimate a critical distance of 140 miles = below which=20 new construction would be economic. Gallick, supra note 82, at 57. =

n84.=20 Gallick, supra note 82, at 89. His choice of 0.25 rather than the DOJ's = 0.18=20 rests on another FTC study. Gallick, supra note 82, at 90.

N85. = Most of=20 the problem markets were in Florida, due to that state's odd location = and shape.=20 Gallick, supra note 82, at 79.

n86. Gallick, supra note 82, at = 29.=20

n87. Gallick, supra note 82, at 38. Below we argue that even if = output=20 restrictions are feasible, collusion should be peculiarly easy to detect = in this=20 industry.

n88. Order No. 436, Regulation of Natural Gas=20 Pipelines After Partial Wellhead Decontrol, [1982-1985] = F.E.R.C. Stats. & Regs. 30,665, 50 Fed. Reg. 42,408 (1985), vacated = and=20 remanded, Associated Gas Distribs. v. FERC, 824 F.2d 981 (D.C. Cir. = 1987), cert.=20 denied, 485 U.S. 1006 (1988).

n89. Gallick, supra note 82, at = 92.=20

n90. Branko Terzic, Fed. Energy Reg. Comm'n, = Pipeline=20 Competition Task Force on Competition in Natural Gas Transportation (May = 24,=20 1993) [hereinafter cited as 1993 Task Force Report]. See generally Order = No.=20 636, Pipeline Service Obligations and Revisions to = Regulations=20 Governing Self-Implementing Transportation and Regulation of Natural Gas = Pipelines After Partial Wellhead Decontrol under Part = 284 of=20 the Commission's Regulations, III F.E.R.C. Stats. & Regs. 30,939, 57 = Fed.=20 Reg. 13,267 (1992).

n91. 1993 Task Force Report, supra note 90, = at 10.=20

n92. 1993 Task Force Report, supra note 90, at 12. The report = does not=20 further discuss the occasions on which such a suggestion may be = legitimate.=20

n93. 1993 Task Force Report, supra note 90, at 8. The paths = studied=20 originated in Louisiana and extended to Pennsylvania and Illinois. =

n94.=20 1993 Task Force Report, supra note 90, at 21 (emphasis in original). The = Task=20 Force states that "market power appears as prices that are above market = clearing=20 levels." 1993 Task Force Report, supra note 90. It does not explain how = to=20 identify such prices, and does not mention our point supra that = regulated rates=20 which recover historical costs will seldom equal competitive equilibrium = prices.=20

n95. Under open access there has been extensive discounting of=20 transportation services. If a pipeline can easily = withhold=20 capacity, withholding should sometimes be a more profitable strategy = than=20 discounting. We discuss this matter in Section V infra.

n96. = Staff=20 Report, Fed. Energy Reg. Comm'n, On Developing a Framework for Assessing = Competition in Natural Gas Transportation, A Discussion Paper For = Commissioner=20 Terzic's Task Force, Nov. 20, 1992 [hereinafter cited as 1992 Discussion = Paper].=20

n97. 1992 Discussion Paper, supra note 96. We know of no = bibliographic=20 entry that examines the relationship in industries that are under = regulation=20 such as affects pipelines.

n98. Richard P. = O'Neill,=20 Common Framework for the Discussion of Structure and Performance, Nov. = 1991,=20 reprinted in, 1992 Discussion Paper, supra note 96, app. B.

n99. = The=20 paper only analyzes markets in which new competitors do not enter even = if=20 incumbent sellers are making substantial economic profits. This may be = an=20 attempt to model regulation of entry apart from regulation of rates.=20

n100. Richard P. O'Neill et al., A Further Discussion of Market = Power=20 and Competition Measures of Firms Within Connected Multi-Market = Industries, July=20 1989, reprinted in, 1992 Discussion Paper, supra note 96, app. C. The = role of=20 regulation is unclear in the paper, since the entities modeled in it are = apparently free to choose their own rates and policies regarding = capacity=20 availability.

n101. 1992 Discussion Paper, supra note 96, at 4.=20

n102. 1992 Discussion Paper, supra note 96, app. B, at 20-39.=20

n103. 1992 Discussion Paper, supra note 96, app. B, at 54. The = cited=20 studies are Frank Gollop and Mark Roberts, Firm Interdependence in = Oligopolistic=20 Markets, 10 J. Econometrics 313-331 (1979); Gyoichi Iwata, Measurement = of=20 Conjectural Variation in Oligopoly, 42 Econometrica 947-966 (1974).=20

n104. 1993 Task Force Report, supra note 90, at 7. The Task = Force's=20 Report refers to the choice of a concentration measure that summarizes = the=20 "potential for market power abuse" as "[a] fairly small methodological = issue."=20 1993 Task Force Report, supra note 90, at 8.

n105. 1993 Task = Force=20 Report, supra note 90, at 19-28.

n106. Pricing Policy for New = and=20 Existing Facilities Constructed by Interstate Natural Gas=20 Pipelines: Notice of Public Conference and Opportunity = to File=20 Written Comments, 59 Fed. Reg. 39,553 (1994); Request for Comments on=20 Alternative Pricing Methods, 70 F.E.R.C. 61,139 (1995).

n107. = 1995 Staff=20 Paper, supra note 72.

n108. 1995 Staff Paper, supra note 72, at = 23.=20

n109. As noted above, however, the staff commented favorably on = the=20 quantitative performance of the MBRs that the FERC had offered Buckeye = Oil=20 Pipeline. 1995 Staff Paper, supra note 72, at 18. =

n110.=20 In its 1995 paper, the FERC staff uses its customary methods to screen a = hypothetical origin-destination market for MBRs. 1995 Staff Paper, supra = note=20 72, at 38-59.

n111. The prices are based on samples of daily=20 observations of into-the-pipeline prices that are = regularly=20 checked for statistical consistency. See Gas Daily, Monthly Contract = Index=20 (Fri., April 15, 1994).

n112. Empirically, price fluctuations = occur even=20 when prices in two areas are within arbitrage limits of each other, = i.e., a=20 small but nontrivial degree of randomness continues late into the period = of=20 price observations in Figure 1. After 1991, the coefficient of = correlation=20 between the prices equals 0.95 (A coefficient equal to zero indicates = pure=20 randomness of their movements relative to each other, and a coefficient = equal to=20 one indicates perfect correlation of their movements). Since there is no = numerical criterion for determining the magnitude of correlation that = puts the=20 two areas in the same market, we will shortly introduce the clearer = standards of=20 cointegration. These issues are discussed in De Vany and Walls, Network=20 Connectivity and Price Convergence: Gas Pipeline = Deregulation,=20 3 Research in Transportation Economics 1-36 (1994).

n113. See = the=20 discussion on informational efficiency for networks infra part IV D.=20

n114. Merton Miller, Financial Innovations And Market Volatility = (1991).=20

n115. On links of some major pipelines there = may be as=20 many as 60 or 70 LDCs who hold and offer firm capacity. See generally=20 Transcontinental Gas Pipe Line Corp., 48 F.E.R.C. 61,399 (1989). =

n116.=20 Catherine Abbott, The Expanding Domain of the Nonjurisdictional Gas = Industry, in=20 Prager, New Horizons in Natural Gas Deregulation (Jerome Ellig & = Joseph Kalt=20 eds., forthcoming 1995).

n117. Combinatorics is the applied = mathematics=20 of counting and selection. Examples of combinatorial problems include: = (1)=20 finding how many different ways three distinct objects can be selected = from a=20 set of eight, and (2) as in the text, finding how many distinct paths = through a=20 network can link two points in it.

n118. William Baumol et al.,=20 Contestable Markets and the Theory of Industry Structure (1982). =

n119.=20 "Asserting that "no pipeline is an island,' the = [Natural Gas=20 Supply Association] paper said rates and tariff conditions that may make = sense=20 for a pipeline when considered in isolation might = actually=20 inhibit or prevent economically attractive gas supplies from flowing to = end-use=20 markets physically located on other pipelines." NGSA = Urges FERC=20 to Consider Development of National Gas Grid When Reviewing=20 Pipelines' Order No. 636 Compliance Filings, Foster = Nat. Gas=20 Rep., Nov. 12, 1992, at 2.

n120. See Philip Areeda & Donald = F.=20 Turner, 2 Antitrust Law, 351-57 (1978); See also George J. Stigler & = Robert=20 A. Sherwin, The Extent of the Market, 28 J.L. & Econs. 555 (1985); = Noel D.=20 Uri & Edward J. Rifkin, Geographic Markets, Causality, and Railroad=20 Deregulation, 67 Rev. Econ. & Stat. 422 (1985); Margaret E. Slade,=20 Exogeneity Tests of Market Boundaries Applied to Petroleum Products, 34 = J.=20 Indust. Econs. 291 (1986); Pablo Spiller & Cliff J. Huang, On the = Extent of=20 the Market: Wholesale Gasoline in the Northeastern United States, 35 J. = Indust.=20 Econ. 131 (1986); David T. Scheffman and Pablo T. Spiller, Geographic = Market=20 Definition under the U.S. Department of Justice Merger Guidelines, 30 = J.L. &=20 Econ. 123 (1987); and Phillip A. Cartwright et al., Price Correlation = and=20 Granger Causality Tests for Market Definition, 4 Rev. Indust. Org. 79 = (1989).=20

One other, more limited, price correlation study of the gas = market=20 reaches conclusions similar to those of the text but does not discuss = their=20 implications for pipeline regulation. See Michael J. = Doane=20 & Daniel F. Spulber, Open Access and the Evolution of the U.S. Spot = Market=20 for Natural Gas, 37 J.L. & Econs. 477 (1994).

n121. The two = major=20 antitrust cases to use price correlations are Marathon Oil Co. v. Mobil = Corp.,=20 530 F. Supp. 315 (N.D. Ohio), aff'd, 669 F.2d 378 (6th Cir. 1981), and = United=20 States v. Archer-Daniels-Midland Co., 695 F. Supp. 1000 (S.D. Iowa = 1987), rev'd,=20 866 F.2d 242 (8th Cir. 1989). In the former, the courts rejected the = defendant's=20 price correlations, and in the latter the appellate court did so. As we = note in=20 the text, the competitive issues addressed in such antitrust cases are = quite=20 different from those that the FERC entertains. The inappropriateness of = price=20 correlations for antitrust does not imply their inappropriateness for = the FERC.=20 A summary of the difficulties in applying price correlations to = antitrust=20 appears in Gregory J. Werden & Luke M. Froeb, Correlation, = Causality, and=20 All that Jazz: The Inherent Shortcomings of Price Tests for Antitrust = Market=20 Delineation, 8 Rev. Indus. Organization 329 (1993).

n122. Price=20 correlations might also be evidence of collusion, but no one to our = knowledge=20 has proposed collusion as a plausible structure of today's gas market.=20

n123. Mathematically, let the prices in markets at time t be = given by=20 p[in'i,t'] and p[in'j,t']. They are cointegrated if we can find = constants a and=20 b such that

p[in'j,t'] - a - bp[in'i,t'] =3D <t>,where = <t> is=20 stationary, as determined by statistical tests. Intuitively, = stationarity means=20 that the difference does not become arbitrarily large as time passes to = the=20 limit. A more technical discussion and additional applications appear in = G.E.P.=20 Box & G. Jenkins, Time Series Analysis (1970); and Arthur S. De Vany = &=20 W. David Walls, Pipeline Access and Market Integration = in the=20 Natural Gas Industry: Evidence from Cointegration, 14 Energy J., Winter = 1993, at=20 1, 1-19.

n124. De Vany & Walls, supra note 123. While daily = prices=20 seem most useful for the study of spot markets, other prices might be = employed.=20 Using regional monthly average spot prices for 1991-92, prior to=20 pipeline operations under Order 636, the 1993 FERC Task = force=20 found a small number of relationships between prices and gas flows that = were=20 inconsistent with competitive markets. See 1993 Task Force Report, supra = note=20 90, at 33.

n125. The growth of markets since open access is = evidenced by=20 the increase in area prices reported in Gas Daily, which have now grown = to over=20 fifty. The growth of pipeline transportation over those = years=20 is documented in U.S. Energy Information Administration, Growth in = Unbundled=20 Natural Gas Transportation Services: 1982-1987 (1989).

n126. De = Vany=20 & Walls, supra note 123, calculate the number of paths through the=20 pipeline network for different years. They show that = through=20 some hubs it is possible today to reach over 140 markets on a path two=20 pipeline links long.

n127. Formally, two = markets can be=20 arbitraged if the price change at market j at time t - 1 can be used to = predict=20 how price in market i will change at time t.

n128. E.g., For 10 = markets,=20 45 different pairs can be chosen, and for 20 markets, 190 pairs. Using a = 10=20 market network and three time periods gives a VAR model with 320 = coefficients to=20 be estimated. If all 320 coefficients are zero (indicating no predictive = ability=20 between any pairs), a highly unlikely event a priori, we have very = strong=20 evidence of competitive pricing.

n129. Such subsets were a = necessity=20 because of the computational difficulties associated with larger sets. = Each one=20 of their six sub-networks contained 252 pairs of spatial and temporal = arbitrage=20 possibilities. See De Vany and Walls, supra note 123.

n130. = Arthur S. De=20 Vany & W. David Walls, The Emerging New Order in Natural Gas: = Markets versus=20 Regulation 88-100 (1995). The standardized amount is one standard = deviation of=20 the observed distribution of price at that point.

n131. See 1993 = Task=20 Force Report, supra note 90, at 36-47.

n132. The number of paths = between=20 any two points in a network expands as an exponent of the number of = links in the=20 system. A new hub produces more paths in the network than lead to it, = and the=20 larger the number of paths in existence, the more new paths will be = produced by=20 that hub. See De Vany & Walls, supra note 130, at 68-72. =

n133. See=20 De Vany & Walls, supra note 123. The hubs were Northern Town Border = Station,=20 Washington; Maumee, Ohio, and Broad Run, West Virginia.

n134. = See, e.g.,=20 Michael Hartzmark, Luck Versus Forecast Ability: Determinants of Trader=20 Performance in Futures Markets, 64 J. Bus. 49, 72 (1991); and Charles = Cox,=20 Futures Trading and Market Information, 84 J. Pol. Econ. 1215 (1976).=20

n135. De Vany & Walls, supra note 130, at 127-40. =

n136. E.=20 Brinkman & R. Rabinovitch, Regional Limitations on the Hedging = Effectiveness=20 of Natural Gas Futures, Energy J. (forthcoming 1995). In this more = recent work,=20 the authors have shown that the Henry Hub futures contract is a less = effective=20 hedge for gas moving into Rocky Mountain and West Coast areas than it is = for the=20 Gulf Coast, Appalachia and eastern Canada. The Kansas City Board of = Trade is=20 currently planning a new futures market whose deliveries will be made at = Waha,=20 located in West Texas, a region identified by De Vany and Walls as = failing the=20 cointegration test. Linda Micco, Is the Midwest Futures Market Being = Stalled?, 3=20 Gas Daily's NG 22 (Apr. 1995).

n137. Since other colluding=20 pipelines can look in as well, however, EBBs might ease = the=20 pipelines' task of monitoring the agreement. =

n138. A=20 pipeline faced with charges of withholding might claim=20 operational necessity as a rationale for its behavior. The FERC seems an = ideal=20 forum in which to resolve the matter.

n139. By installing some = 98 miles=20 of 30-inch diameter pipe to parallel and loop its Havasu crossover line, = ... [El=20 Paso Pipeline] said it could "quickly react to changing = market=20 conditions by moving gas from any supply source to any market on its = system."=20 ... El Paso maintained that it no longer controls the gas-flow patterns = on its=20 system and that they rather are "dictated by its shippers."
=
PGT=20 May Expand Mainline Again; El Paso Can Loop Havasu Line, Inside FERC's = Gas=20 Market Rep., April 21, 1995, at 17.

n140. Capping an Extended = Period of=20 Vibrant Construction Activity, Inside FERC, May 15, 1995, at 12. =

n141.=20 Exhibit accompanying remarks of Dr. Kenneth Lay, Enron Corp., before = Pacific=20 Coast Gas Association Annual Business Meeting, Houston, Sept. 13, 1995.=20

n142. "[Eighteen] new [pipeline] storage = projects were=20 placed into service during 1994, representing 130 Bcf of additional = working-gas=20 capacity." Id.

n143. E.g., Crossroads to Provide=20 Pipeline Link from Midwest to East Coast, Inside FERC's = Gas=20 Market Rep., June 2, 1995, at 3.

n144. E.g., CIG wants to Speed = Capacity=20 Awards to Keep up with Hedging Timetables, Inside FERC's Gas Market = Rep., June=20 30, 1995, at 18.

n145. For similar views, see Comments of the = Interstate=20 Natural Gas Association of America, The Path to Market-Based Pricing for = Gas=20 Services, Docket No. RM95-6-000, at 3 (April 21, 1995).

n146. = See Philip=20 Marston, The Rumble of Bundles: A Review of Experience Under the = Capacity=20 Release Experiment (Repro, Hadson Corp., Washington, 1994); and When = Once Just=20 Isn't Enough (Graphs from presentation at DOE/NARUC Annual Natural Gas=20 Conference, Orlando, 1995).

n147. More precisely, the pre-merger = price=20 rises if there is no collusion. Each producer in the DOJ's pre-merger = market may=20 have some choice about its price, and there is no necessary assumption = that=20 competition has driven price in the pre-merger market down to marginal = cost.=20

n148. 1995 Staff Paper, supra note 72, at 27 (emphasis in = original,=20 footnote omitted).

n149. 1995 Staff Paper, supra note 72, at 27. = The=20 staff's assertion about identical services and prices may run into = difficulty if=20 shippers and pipelines negotiate heterogeneous = contracts from=20 an array of competitive alternatives. It is unclear how one might extend = a=20 superficially appealing criterion about rates to situations that are = governed by=20 complex contracts.

n150. 1995 Staff Paper, supra note 72, at 27. =

n151. ANR Outlines Market-Based Pricing Proposals for Released = Capacity,=20 Foster Nat. Gas Rep., Dec. 2, 1993, 4.

n152. For more on the = distinction=20 between antitrust markets and economic markets, see David T. Scheffman = &=20 Pablo T. Spiller, Geographic Market Definition Under the U.S. Department = of=20 Justice Merger Guidelines, 30 J.L. & Econ., Apr. 1987, at 124, = 124-28.=20

n153. Some economists have recommended a similar screening of = mergers in=20 unregulated industries, first examining barriers to entry and only later = examining concentration if barriers are high. See Steven C. Salop, = Symposium on=20 Mergers and Antitrust, 1 J. Econ. Perspectives 3, 7 (1987); Lawrence J. = White,=20 Antitrust and Merger Policy: A Review and Critique, 1 J. Econ. = Perspectives 13,=20 17 (1987).

n154. There is still a role for the FERC in dealing = with=20 opportunistic behavior, e.g., if the pipeline curtails=20 interruptible service without good cause in hopes of making its = overpriced firm=20 capacity contracts look more attractive.

n155. "Says [Greg = Lander,=20 Chairman and President of the National Registry of Capacity Rights], = "You deal=20 with in money what you don't know in operational certainty.'... But for = those in=20 the utility business, says Lander, "It can come as a shock that someone = might=20 settle the physical problem in money.' " Bruce W. Radford, Simplify and=20 Exaggerate, 133 Pub. Utils. Fort., Aug. 1995, at 6, 5-6.

n156. = 1995=20 Staff Paper, supra note 72, app., Analysis of Other Industries. =

n157.=20 The link between the market analyses and the favorable experiences is = unclear.=20 The fact that interested parties (including regulatory staffs) used = these market=20 definitions in no way implies that they were dispositive in producing = the good=20 outcomes. The staff also does not look at the performance of markets = affected by=20 decisions that rejected such market definitions.

n158. 1995 = Staff Paper,=20 supra note 72, at 37; see also Mary Lou Steptoe, The Power-Buyer Defense = in=20 Merger Cases, 61 Antitrust L.J. 493 (1993).

n159. The staff = asserts that=20 this is so, without providing evidence for the view. 1995 Staff Paper, = supra=20 note 72, at 37.

n160. If it operates under performance-based = ratemaking=20 or is at risk of disallowances for imprudent policy, the LDC might care. =

n161. They can become matters for the FERC if a large user can = feasibly=20 bypass the LDC and tap an interstate line directly.

n162. Now a = "very=20 important" issue, undersizing is not mentioned in either the Gallick = study,=20 supra note 82, or the 1993 Task Force Report, supra note 90. One other = study of=20 undersizing seems inapplicable to the largely unintegrated gas industry. = In the=20 late 1970s, the DOJ claimed that the vertically integrated owners of = regulated=20 oil pipelines undersized them in order to recover = profits in=20 downstream markets that they could not extract from = pipeline=20 rates. For a summary of the issues, see Michael E. Canes & Donald A. = Norman,=20 Pipelines and Public Policy, Oil = Pipelines=20 & Pub. Pol'y, 1979, at 141, 141-63.

n163. 1995 Staff Paper, = supra=20 note 72, at 45 (emphasis in original, footnote omitted). The omitted = footnote=20 cites as authority Judge Hand's decision in United States v. Aluminum = Co. of=20 America, 148 F.2d 416, 424 (2nd Cir. 1945). Reliance on the logic of = this case=20 need not lead to the staff's conclusions. See Darius Gaskins, Alcoa = Revisited:=20 The Welfare Implications of a Secondhand Market, 7 J. Econ. Theory, = 1974, at=20 254, 254-71. The question of monopolistic undersizing depends in complex = ways on=20 whether the monopolist sells or rents the good. In = pipelines,=20 the analogous distinction is whether capacity holders can resell their = holdings=20 (not currently allowed) or are constrained to a release market. See = Jeremy=20 Bulow, Durable-Goods Monopolists, 90 J. Pol. Econ., 1982, at 314, = 314-332; see=20 also Comments of Hadson Gas Systems, Inc., FERC Docket No. RM 95-6-000, = at 9.=20

n164. Here it may be important that regulation change to allow = full=20 resale of a holder's capacity rights, so that parties who do not = directly=20 finance construction are treated symmetrically with those who do. There = has been=20 one request for a FERC rulemaking on this issue. Associated Gas = Distributors and=20 United Distribution Companies Request Rulemaking to Allow Direct = Marketing of=20 Firm Pipeline Capacity Rights, Foster Nat. Gas Rep., = Dec. 30,=20 1993, at 26.

n165. We have found only one FERC decision since = open=20 access that mentions pipeline undersizing. In a rate = case, the=20 Commission argued against an applicant's request that it reconsider a=20 reservation charge based on 100% of a proposed = pipeline's=20 design capacity. The applicant claimed that the 100% factor would = increase its=20 risk of insufficient throughput, and in response it might choose to = build an=20 undersized line. The FERC responded:

TransColorado = further argues=20 that high throughput percentages may force pipeline = applicants=20 to undersize proposed pipeline projects to reduce their = risk.=20 As we stated in the preliminary determination, "The central feature of = the=20 optional certificate procedures is the requirement that risk taking be = entirely=20 voluntary." This means that rates charged for new service be based on = optimal=20 use of the new facility as designed. If the sponsor of an optional = certificate=20 project "undersizes" its proposed pipeline project to = reduce=20 its risk as posited by TransColorado, the sponsor is simply making the = kind of=20 market-based decision that forms the foundation of the optional = certificate=20 procedures. An optional certificate applicant sizes its project based on = the=20 market conditions it perceives to exist. If it believes that a certain=20 pipeline capacity cannot be optimally used, it is = appropriate=20 that it decide to lower the scale of its project to reduce its risk. =
=20
TransColorado Gas Transmission Co., 67 F.E.R.C. 61,301 (1994). =

n166.=20 Gallick, supra note 82, at 89-92. Regulatory delays for each of the four = major=20 pipelines or expansions reaching California since 1991 = have=20 exceeded construction times. Pacific Gas Transmission's expansion = required 610=20 days for construction and 966 days for FERC approval. Newly-built Mojave = Pipeline required 182 days for construction and 1,745 = days for=20 approval. These are unpublished figures supplied by Robert L. Bradley, = Jr., of=20 the Enron Corporation.

n167. 1995 Staff Report, supra note 72, = at 26.=20

n168. A. Michael Spence, Investment Strategy and Growth in a New = Market,=20 10 Bell J. Econ., Spr. 1979, at 1, 1-19; and Richard G. Gilbert & = David M.G.=20 Newbery, Preemptive Patenting and the Persistence of Monopoly, 72 Am. = Econ.=20 Rev., June 1982, at 514, 514-26. For an opposing view about the = likelihood of=20 such preemption, see Easterbrook, Predatory Strategies and = Counterstrategies, 48=20 U. Chi. L. Rev. 263 (1981).

n169. In economics, a consumer's = demand=20 price for an extra unit of some good is the maximum price it would be = willing to=20 pay rather than go without that unit. Those with a high willingness to = pay=20 (marginal valuation) will outbid those with lower willingness. The = lowest=20 winning bidder, like all other buyers in a competitive market, pays the = market=20 equilibrium price.

n170. 1995 Staff Paper, supra note 72, at 42. =

n171. By this pricing standard, even the spot gas market may not = be=20 competitive. If there has been an unexpected rise in demand after = bidweek, the=20 price of swing market gas [increments or decrements to bidweek = contracts] will=20 be higher than the bidweek price, an increase that is necessary to = elicit=20 supplies that would have been shut in at the bidweek price. A user can = protect=20 itself against high swing prices by contracting for its worst-case = requirements=20 during bidweek, but this strategy need not be best for it.

n172. = The=20 staff does not examine price/quality differences of the opposite sign, = although=20 the analysis might be useful. Specifically, when can IT sell at too = large a=20 discount relative to FT, i.e., one which overstates the difference in = quality=20 between IT and FT?

n173. On the other side of the actual market, = prospective capacity releasors are complaining that competition from IT = has cut=20 the prices they receive for their capacity to unacceptably low levels. = See=20 Pipeline IT Deals Are Undercutting Released Capacity, = SoCal Gas=20 Says, Inside FERC, Dec. 19, 1994, at 15.

n174. Here, efficiency = is used=20 in the sense of Section IV A supra.

n175. Robert J. Michaels, = When=20 Captive Customers Bear the Risk, Pub. Utils. Fort., Nov. 15, 1993, at = 25.=20

n176. In contrast to the choices that market participants have = made=20 since open access, some economists continue to argue for the superiority = of=20 long-term contracting. See Thomas P. Lyon & Steven C. Hackett, = Bottlenecks=20 and Governance Structures: Open Access and Long-Term Contracting in = Natural Gas,=20 9 J.L. Econ. & Organization, 1993, at 380, 380-398; and David J. = Teece,=20 Structure and Organization of the Natural Gas Industry, 11 Energy J., = Mar. 1990,=20 at 1.

n177. Oliver Williamson, The Economic Institutions Of = Capitalism=20 103-30 (1985); Benjamin R. Klein et al., Vertical Integration, = Appropriable=20 Rents, and the Competitive Contracting Process, 21 J.L. & Econ. 297 = (1978).=20

n178. See Michaels, supra note 175, which also discusses lack of = analogy=20 between an investor holding a mix of securities and a gas purchaser = holding a=20 mix of gas contracts of different terms.

n179. In its analysis = of Order=20 636, the FERC staff argues that the Order will facilitate the formation = of new=20 long-term gas contracts, an outcome that has not materialized. As the = markets=20 have evolved, traders do not find the contracts desirable, and = economists do not=20 find them efficient. One could better praise Order 636 because it has so = equalized access to pipeline transportation that it = facilitates=20 the making of short-term gas contracts that are more responsive to = market=20 conditions. Office of Economic Policy, Fed. Energy Reg. Comm'n, Costs = and=20 Benefits of the Final Restructuring Rule, at 7-9 (1992).

n180.=20 Pipelines, Wall St. Focusing on Cost Exposure from = Capacity=20 Turnback, Inside FERC, Mar. 27, 1995, at 1.

n181. Transwestern = to Cover=20 70 % of Costs of Relinquished SoCal Gas Capacity, Inside FERC's Gas = Market Rep.,=20 May 5, 1995, at 1; El Paso Eyes Capacity-Turnback Fee; PG&E to Walk = Away=20 from 1.14 bcf/day, Inside FERC's Gas Market Rep., June 30, 1995, at 10.=20

n182. INGAA Study Finds "Sustained and Rapid" Growth in Number = and Size=20 of Capacity Release Transactions During First Three Quarters of 1994, = Foster=20 Nat. Gas Rep., Feb. 23, 1995, at 5; INGAA's Annual Transportation Survey = Shows=20 Continued Decline in Interruptible Volumes, Dramatic Growth in Capacity = Release,=20 Foster Nat. Gas Rep., Aug. 24, 1995, at 20.

n183. Order 636 = Experience=20 Raises Secondary-Market Issues on Transco, Inside FERC, Nov. 7, 1994, at = 7.=20

n184. Commissioner Hoecker Says FERC Must Ultimately Consider = Additional=20 Approaches to Capacity Releasing, Foster Nat. Gas Rep., Jan. 13, 1994; = Secondary=20 Market is the Place to Start Negotiated Rates, FERC Told, Inside FERC, = May 8,=20 1993, at 11.

n185. In March 1994, normally a high-load month for = gas=20 flowing to California, released firm capacity on El Paso=20 Pipeline was being heavily discounted: Four percent was = flowing=20 at 20% or less of the pipeline's maximum rate; 53% at = 30% of=20 the maximum rate; 22% at 55% of maximum, and 19% at 100% of the maximum. = The=20 entities paying 100% were all California "core aggregators," prohibited = by state=20 regulation from using discounted interstate capacity. Competition for = Throughput=20 Means Lower Rates to California Markets, Inside FERC's Gas Market Rep., = April=20 22, 1994, at 1.

n186. Discounting has been substantial since the = beginning of open access. Between 1984 and 1993, the average=20 pipeline markup per Mcf to the city gate (mainline=20 transportation cost) fell from $ 1.67 to $ 1.19, in constant 1993 = dollars. The=20 More Things Change..., Nat. Gas Week, Feb. 13, 1995, at 1.

n187. = Paul W.=20 MacAvoy et al., Federal Energy Regulatory Commission Order No. 636 as = the=20 Penultimate Regulatory Reform of the Gas Industry, John M. Olin = Foundation=20 Working Paper 38, at 35 (1995).

n188. Capacity release could = help K N=20 Gas Marketing's bottom line by increasing revenues from sales to new = markets=20 that otherwise would not have been made ... In the summer months when = released=20 capacity was selling at bargain-basement prices, [K N] was able to pick = up that=20 capacity at such a discount that it was able to sell to markets it never = would=20 have cracked in the pre-Order 636 environment.

Marketers = Find New=20 Opportunities through Capacity Release Deals, Inside FERC's Gas Market = Rep.,=20 Dec. 2, 1994, at 1.

Capacity release also was a factor in = KCS=20 Energy's decision to expand its market presence on the west coast. The = Director=20 of Transportation stated the "Capacity release has opened up an = opportunity to=20 take advantage of basin swaps, ... We have a much larger capability of = doing=20 pipe-to-pipe transfers now than we did prior to Order 636."
=
Id.=20 at 2.

n189. Ray Klempin, Here Come the Hubs!, Gas Daily's NG 1 = (Sum.=20 1993), 20-23.

n190. As examples, one proposal uses compression = changes=20 in a large interstate line to accommodate the gas flows of a market = center.=20 Another proposal intends to use the pipe network under New York City to = do the=20 same. Yet another proposed market center will not have a single=20 pipeline operator. CNG Transmission Files for Rates; = Requests=20 on Noram Exceed Capacity, Inside FERC's Gas Market Rep., May 20, 1994, = at 17;=20 Gas Industry Competition Goes Underground, Energy Daily, July 14, 1995, = at 2.=20

n191. LDC monopoly power over customers behind city gates is a = matter=20 for state regulators. The impact of interstate deregulation is being = felt in LDC=20 territories, and pressure is increasing for LDCs to institute unbundled=20 transportation programs. See Suedeen G. Kelly, Intrastate Natural Gas=20 Regulation: Finding Order in the Chaos, 9 Yale J. on Reg. 355 (1992).=20

	Home =20 S= ources =20 How=20 Do I? Site=20 Map What's = New Help =20