Geographic market delineation for electric utility mergers.
This article addresses methodologies to delineate geographic markets for use in evaluating the likely effects of electric utility mergers on market power based on ownership of electric generators. The article describes the methodology used by the U.S. Department of Justice (DOJ) and the Federal Trade Commission (FTC), contrasts this with the methodology used by the Federal Energy Regulatory Commission (FERC), demonstrates that the differences between these methodologies are important, and recommends use of the methodology in the DOJ/FTC Merger Guidelines(1) rather than that in FERC's Merger Policy Statement.(2)
Section II of this article explains geographic price discrimination, which plays an important role in this article. Section III discusses the purpose of defining geographic markets for use in evaluating market power and explains the methodology for identifying geographic markets that is used by DOJ and the FTC. The DOJ/FTC methodology, which is based on sound economic principles, features the hypothetical monopolist test.
Section IV explains the different methodology for delineating geographic markets that is set out by FERC in appendix A of its Merger Policy Statement and subsequent orders in merger cases, and further elaborated in an April 1998 notice.(3) FERC's appendix A methodology begins with individual (or similarly situated) wholesale buyers called destination markets, and for each one delineates a geographic market using the delivered price test.
Section V explains that the appendix A methodology rests on implicit assumptions that sellers in electric power markets can practice geographic price discrimination and that mergers can affect the ability or incentive to discriminate. In most cases, however, price discrimination is not likely to be significant in these markets. Section VI provides illustrations that demonstrate that, absent price discrimination, the appendix A methodology for defining geographic markets is likely to produce misleading diagnoses regarding the competitive effects of mergers and appropriate remedies. The appendix A methodology for defining geographic markets leads to substantial violations of the competitive analysis screen standards for some potential mergers that would not create or enhance market power, and produces no violation for some other potential mergers that would in fact create or enhance market power. In other words, the appendix A methodology produces false positives in some circumstances and false negatives in others.
Sections III through VI focus on assessment of market power arising from ownership of generating capacity(4) on the assumption that there are no transmission constraints.(5) Section VII addresses implications of transmission constraints for geographic market delineation. Section VIII brings together points made in earlier sections to focus on circumstances under which it would be appropriate to analyze individual destination utilities as separate relevant markets for purposes of merger evaluation. Section IX states conclusions and recommends that FERC should revise its approach to geographic market delineation to make it consistent with the hypothetical monopolist test in the Merger Guidelines.
A structural analysis of market power involves several steps beyond delineation of markets, including identification of the competitors in relevant markets and measurement of market shares. This article focuses principally on geographic market delineation. Elsewhere I have explained that the appendix A methodology for measuring market shares has a number of significant shortcomings. These shortcomings may make those shares unreliable as a basis for inferences about market power even if an individual destination utility is a relevant antitrust market for purposes of analyzing a merger.(6)
II. Geographic price discrimination
One must be familiar with the concept of geographic price discrimination in order to understand how geographic markets should be delineated for use in analyzing market power, and in order to understand the differences between the Merger Guidelines and appendix A methodologies for defining geographic markets. This section begins with an explanation of geographic price discrimination and then addresses geographic price discrimination in markets for electric power.
A. Explanation of price discrimination
A producer that practices geographic price discrimination charges different prices for the same product sold at the same time to customers in different areas. To test for discrimination, the relevant prices to compare exclude transportation costs from the producer to the customer: the relevant prices are ex factory, not delivered, prices. Furthermore, if transportation capacity from the producer to the customer is limited and fully utilized, the relevant transportation costs include the opportunity cost (scarcity value) of the transportation capacity used and not just the out-of-pocket transportation charge.
The following four examples illustrate these points:
1. Power & Light charges delivered prices of $25 per megawatt-hour (MWh) for spot energy sold to customers in East Centralia and $30/MWh for spot energy sold to customers in West Centralia. Transmission service from Power & Light to East Centralia and West Centralia is readily available at $3/MWh and $8/MWh, respectively.(7) There is no price discrimination here because the ex generator price for sales to both sets of customers is $22/MWh.
2. Electric & Gas charges an ex generator price of $40/MWh to local customers in North Centralia and a delivered price of $75/MWh to customers in South Centralia. South Centralia is located on the other side of a congested transmission interface, and the general market price in South Centralia is $751MWh. The out-of-pocket cost of transmission from Electric & Gas to South Centralia is $10/MWh. In this case there is no price discrimination. The difference between the prices Electric & Gas charges to customers in North Centralia and South Centralia is accounted for by the opportunity cost of transmission service to South Centralia.
3. Energy Corp. sells spot energy to adjacent Public Service at a delivered price of $24/MWh and at the same time to all other customers at an ex generator price of $22/MWh. Energy Corp. charges $5/MWh for unbundled transmission service between any two points on its system. Unless there is some other explanation based on facts not provided here, there is price discrimination. Energy Corp. is discounting the ex generator price of energy to Public Service by $3/MWh below the ex generator price at which Energy Corp. is making sales to other customers.
4. Energy Corp. sells spot energy to all buyers at an ex generator price of $22/MWh. Energy Corp.'s ceiling transmission rate is $5/MWh. Although no part of its transmission system is congested, Energy Corp. discounts transmission service--to $2/MWh--only for deliveries to its interconnection with Public Service. Based on a comparison of ex generator prices, there does not appear to be price discrimination for energy. However, the way in which transmission service prices are discounted may have the same effect. In fact, from a competitive perspective, this example is the same as the preceding one.
Sellers, even monopolists, can practice price discrimination only if they can prevent buyers who are able to purchase at the lower price from reselling to buyers who are asked to pay the higher price--that is, only if they can prevent arbitrage. Sellers of certain types of services--for example, haircuts--can price discriminate because resale is not possible. By contrast, sellers in a commodity market with active arbitrageurs that are in the business of buying low and reselling high cannot price discriminate.
As an illustration of how arbitrage defeats price discrimination, suppose that a utility, Electric Power, quoted a delivered price of $27/MWh for energy sales to Central Energy while selling at an ex generator price of $21 fMWh to other buyers. Suppose that the price of transmission service from Electric Power to Central Energy is $4/MWh. In that case, Central Energy would buy no energy from Electric Power at a delivered price of $27/MWh. Power marketers--acting as arbitrageurs--would be able to purchase energy from Electric Power at an ex generator price of $21/MWh, and they would be able to purchase transmission service to Central Energy at a price of $4/MWh. Consequently, power marketers would be willing to deliver energy from Electric Power's generators to Central Energy at a delivered price of $25/MWh plus a thin margin for their own services. Competition among power marketers would enable Central Energy to purchase the energy in question at a delivered price slightly above $25/MWh. Electric Power would not succeed in price discriminating, because all of its sales of energy would take place at the lower ex generator price of $21/MWh. It follows that Electric Power would have no incentive even to try to discriminate against Central Energy.
C. Price discrimination for electric power
Several conditions in markets for electric energy and capacity make it difficult for sellers to practice geographic price discrimination. Energy and capacity are commodities that are actively arbitraged by power marketers and by utilities themselves. Power marketers accounted for an estimated 1.2 billion MWh of energy sales in 1997, representing an estimated 50% of total wholesale sales.(8) Also, in a number of regions energy is traded in power exchanges in which all buyers face the same prices for a given service. In addition, there are a number of hubs that serve as liquid markets for energy traded at prices specified on an "into hub" basis.
Based on these facts, it is reasonable to conclude that analyses of market power in markets for electric power should not be based principally on an assumption that utilities can practice geographic price discrimination, including price discrimination that can target one wholesale buyer at a time. Statements by FERC suggest that it does not expect substantial geographic price discrimination. For example, FERC stated, "we would expect prices to vary little from customer to customer in the same region during similar demand conditions (if there are no transmission constraints)."(9)
Nonetheless, it is not unreasonable for an analysis of market power to include an evaluation of the facts of a case to determine whether sellers are likely to have the ability to practice geographic price discrimination, and for the analysis to take into account the implications of discrimination in analyzing market power when such discrimination is likely. In the case of electric power, two strategies might in principle be used in certain circumstances to discriminate among customers in different geographic areas. These strategies are illustrated by the third and fourth examples in section AR--the examples involving Energy Corp. as a seller.
The first of these strategies is for the seller to make sales on a delivered basis, at least to customers to which sales are made at prices (computed ex generator) below the highest price charged. If sales are made only on a delivered basis, and there are only two customer locations, arbitrage would not prevent a discriminatory price increase (computed ex generator) to Customer 1 relative to Customer 2. The discriminatory price increase could be as large as the excess of (i) the cost of transmission along the indirect path from the producer to Customer 1 and onward to Customer 2 over (ii) the cost of transmission along the lowest cost path from the producer to Customer 2.(10)
The second of these strategies is for the seller to make sales at a uniform ex generator energy price but to discount prices for transmission service to some delivery points but not to others. These two strategies might have similar results in terms of delivered prices and sales.
If geographic price discrimination is likely in a particular set of circumstances, such discrimination may result in delineation of geographic markets for merger analysis that are smaller than the ones that would be appropriate absent such discrimination. This point is addressed further in section V.
Geographic price discrimination is not the only potential reason that geographic markets may be narrower than the ones that would generally be appropriate. Transmission constraints may also result in narrower geographic markets. This point is addressed in section VII.
III. Delineating relevant antitrust markets
The delineation of markets for purposes of antitrust analysis of mergers and other issues is the subject of a large economics and legal literature going back for decades.(11) Based on this literature and experience with thousands of mergers in hundreds of industries, over a period of many years the federal antitrust agencies developed the methodology for delineating markets for merger analysis that is set out in the Merger Guidelines. FERC should use that framework, but the methodology laid out by FERC in appendix A and elsewhere does not use it.
A. Initial statement of principles
Traditional antitrust analyses of market power rely heavily on inferences based on market shares and market concentration, along with other structural characteristics of markets, such as entry barriers. In order to compute market shares and evaluate barriers to entry for a market, one must first identify a relevant antitrust market for the analysis. In many cases, the product and geographic area identified as a relevant market largely determines the results of an analysis of market power. If a relevant market is delineated too narrowly, market shares, concentration and entry barriers all may be exaggerated, and in some cases the effect of the merger on competition may be substantially overstated. In other cases, use of an overly narrow market may cause the market share of one or both merging companies to be understated--or even appear to be zero. In that case, the effect of the merger on competition may be substantially understated or overlooked entirely. As a result, the methodology used to delineate markets is a critical determinant of the reliability of a structural analysis of market power.
An antitrust market consists of a product and a geographic area. In identifying markets for merger analysis, one delineates the geographic area or market in question first on the assumption that sellers do not practice geographic price discrimination. Then, if there is evidence that sellers are reasonably likely to engage in such discrimination, one delineates additional geographic markets that take account of this discrimination. The remainder of section III of this article deals with situations in which sellers cannot price discriminate. Section V addresses situations in which sellers can price discriminate.
As a general matter, a geographic market used in antitrust analysis refers to an area in which a product is sold. However, under the assumptions in the present section of this article (no price discrimination, transmission constraints, or transmission market power) one can assume that all electric energy is sold ex generator, and a geographic market for electric energy refers to an area in which electric energy is produced as well as sold.
In principle a relevant geographic market is delineated beginning with each generator of each merging company. Many of these geographic markets would normally be the same. Indeed, if a company's generators are subject to the same transmission conditions (for example, prices and constraints to each destination), then a single geographic market will be delineated for all of that company's generators. For simplicity, this is the way a utility's generators are treated in this article unless otherwise stated.
In defining a relevant antitrust market for purposes of analyzing an electric utility merger, the objective is to identify a product (such as electric energy during a representative summer peak hour) and a geographic area in which generators are located that satisfy the following criteria:(12)
1. The geographic area includes generators of at least one of the merging companies. (The merger will have no effect on shares and concentration unless the market includes generators of both merging companies.)
2. If one company owned all electric generating units located in the geographic area, that company would raise the price for the product to at least 5% above the price that would prevail if the merger did not occur.(13) That is, the company could increase prices by at least 5% without losing so many sales that the price increase in question would be less profitable than some price increase between 0% and 5%.
3. The geographic area is the smallest area that satisfies the preceding criterion.
The easiest way to achieve an understanding of the process of delineating a relevant antitrust market is to work through an example, such as the following one based on a hypothetical set of facts.
B. Illustration: defining a relevant market for a bread merger
Suppose that one wanted to analyze the effect on market power of a proposed merger of two white bread bakery companies in Tennessee. Suppose too that equipment used to produce other items, such as whole wheat bread, cannot be used to produce white bread. Finally, suppose that an opponent of the merger has offered an analysis of competitive effects based on an alleged relevant market limited to white bread produced in Tennessee. How would one decide whether this proposed antitrust market was sufficiently large?
One would decide this issue by asking whether the price of white bread that would maximize the profits of a company that owned all Tennessee white bread bakeries would be at least 5% above the price that would prevail absent the merger. If the answer to this question is yes, then white bread in Tennessee is a relevant antitrust market for purposes of this merger.(14)
In that case, if white bread bakers in Tennessee raised their prices, these bakers would, of course, lose some sales. Some customers would switch from white bread to whole wheat bread and from bread made in Tennessee to bread made in Kentucky. However, by assumption, competition from whole wheat bread and bread made in Kentucky would not be sufficient to prevent a price increase of at least 5% from being most profitable for a company that owned all white bread bakeries in Tennessee.
In this fact situation, why is it useful to identify white bread produced in Tennessee as a relevant market for purposes of analyzing the competitive effects of this merger? The answer is that whole wheat bread and bread baked in Kentucky are not sufficiently important competitors to warrant their inclusion in the relevant antitrust market. Consider the following property of this relevant market: elimination of the competition that would exist, absent the proposed merger, among white bread bakeries in Tennessee would result in a price increase of at least 5% In that case, a merger of two Tennessee white bread bakers with significant market shares may result is higher prices for consumers.
Now suppose that the profit-maximizing price increase for an owner of all white bread bakeries in Tennessee would be less than 5% because a large share of its customers would respond to a price increase of 5% by switching to whole wheat bread. In that case, whole wheat bread should be added to the product market. If the profit-maximizing price increase for an owner of all white and whole wheat bread bakeries in Tennessee still would be less than 5% because many customers would switch to bread from bakeries in Kentucky, then Kentucky should also be added to the proposed market. If we suppose that the profit-maximizing price increase for an owner of all white and whole wheat bread bakeries in Tennessee and Kentucky would be at least 5%, then white and whole wheat bread produced in these two states would be a relevant antitrust market for analysis of the proposed merger.
C. Formal statement of the hypothetical monopolist test
The preceding bakery case illustrates what is known as the hypothetical monopolist test. This test is the central feature of the methodology used by the federal antitrust agencies and widely accepted by courts to delineate relevant antitrust markets for merger cases in all industries.(15) DOJ has recommended that FERC use the hypothetical monopolist test as the basis for defining geographic markets used to analyze electric utility mergers.(16)
For purposes of defining geographic markets absent price discrimination, the Merger Guidelines ([sections] 1.21) describe the hypothetical monopolist test as follows:
[T]he Agency will delineate the geographic market to be a region such that a hypothetical monopolist that was the only present or future producer of the relevant product at locations in that region would profitably impose at least a "small but significant and nontransitory" increase in price, holding constant the terms of sale for all products produced elsewhere.(17)
D. Antitrust markets vs. other types of markets
Relevant antitrust markets are delineated for a specific purpose--to aid in the analysis of market power. Furthermore, market shares and concentration indexes computed in antitrust markets are compared to various thresholds in reaching at least preliminary conclusions regarding market power. In order for market power analyses based on market shares and concentration to be reliable, and for the thresholds to be applied consistently, relevant antitrust markets must be identified using the appropriate criteria.
To emphasize the importance of methodology, it is useful to point out that relevant antitrust markets frequently are not the same as markets that are identified for other purposes. For example, a relevant antitrust market is not the same as an economic market. The latter term is frequently used to refer to a set of products and a geographic area for which prices tend toward uniformity, allowing for transportation costs. Because antitrust markets and economic markets are delineated differently and for distinct purposes, it should not be surprising that antitrust markets can be larger or smaller than economic markets.(18)
E. Antitrust markets for different purposes
This article addresses how relevant geographic markets are delineated for purposes of analyzing whether mergers would create or enhance market power. While the hypothetical monopolist framework should also be used to delineate antitrust markets for use in nonmerger cases involving market power issues, one detail of the test may need to be handled differently in some nonmerger cases. In merger cases, the baseline price for the 5% test used to delineate markets is the price that would prevail absent the merger (and absent collusion).(19) The price that would prevail may be above the competitive level because of unilateral market power that would exist without the proposed merger in the pertinent future period.
However, in a nonmerger context, the purpose of an antitrust analysis may be to determine whether there is significant market power under existing circumstances. If that is the issue, the relevant baseline price for the test is the competitive price. It follows from this discussion that it may be appropriate to use different antitrust markets for purposes of evaluating market power in different contexts.
IV. The appendix A method for delineating geographic markets
In its Merger Policy Statement, FERC appropriately adopted the DO J/FTC Merger Guidelines as the analytical framework for use in evaluating the effects of electric utility mergers on market power. However, in appendix A to that statement and subsequently, FERC set out a detailed algorithm for identifying geographic markets for which market shares and Herfindahl-Hirschman indexes (HHIs), which measure market concentration, must be computed in merger cases. Contrary to assertions in the Merger Policy Statement, the methodology set out in appendix A is not consistent with the Merger Guidelines.
The appendix A methodology involves two steps. First, one identifies utilities that are considered to be potential victims of market power; these utilities are called destination markets. In most cases, FERC requires that each of the following utilities be treated as a separate destination market in analyzing a proposed merger of utilities Alpha and Beta:
Alpha's transmission dependent utilities as a group. Typically, Alpha's transmission dependent utilities would be municipal and rural electric cooperative systems that are interconnected only with Alpha. Beta's transmission dependent utilities as a group. Each other utility that is directly interconnected with Alpha and/or Beta. Each other utility that has been a significant wholesale customer of Alpha and/or Beta during the past 2 years.
I will refer to this step in the appendix A methodology as involving identification of individual destination markets. In limited circumstances, FERC has accepted a destination market that is delineated more broadly to include a set of utilities that are, at least during pertinent time periods, subject to the same transmission conditions, for example, utilities in the Pennsylvania-Maryland-New Jersey interconnection power pool.
Second, for each destination market, one delineates a geographic market consisting of the set of generating units from which energy could be generated and transmitted to the destination market in question at a variable out-of-pocket cost that does not exceed by more than 5% the price that would prevail at that destination absent the proposed merger. This part of the methodology is referred to as the "delivered price" test.(20)
Appendix A's focus on individual destination utilities is an appropriate basis for analyzing the geographic scope of competition if and only if a hypothetical monopolist could and would increase prices to each of these individual utilities separately without increasing prices to buyers located elsewhere. In short, the analysis of individual destination utilities implicitly assumes not only geographic price discrimination but discrimination that permits targeting of individual utilities for independent price increases.
The appendix A method for delineating geographic markets also implicitly assumes that the price paid by a given destination utility is determined by competition among only the generating units that could supply electric power to that particular utility at prices no more than 5% above the prevailing level. Absent geographic price discrimination among individual utilities, this assumption would frequently be incorrect.
When geographic price discrimination cannot be practiced, an exercise of generation market power that raises the price for one buyer would typically raise prices for buyers in surrounding areas as well. In that case, the profitability of raising prices would be reduced by losses of sales not only to the targeted buyer but also to buyers in surrounding areas. As a result, market power would depend not only on the competitive alternatives available to the targeted buyer--alternatives that appendix A attempts to identify using the delivered price test--but also on alternatives available to surrounding buyers. From this it follows that generating capacity that would not satisfy the delivered price test with regard to a particular buyer would, nevertheless, reduce the profitability of increasing prices to that buyer--and hence would constrain market power.
The focus of appendix A on individual destination markets has been criticized by DOJ in comments to FERC. For example, DOJ recently made the following statements to FERC:
The delivered price test would be the Guidelines' analysis only if price discrimination were possible and there was no possibility of arbitrage, so that prices in adjoining areas were free to move entirely independently.(21) While that [geographic price discrimination] certainly is possible in the electric power industry, it does not appear that it is likely to be common.(22)
DOJ has advised FERC to delineate geographic markets using a methodology that incorporates the hypothetical monopolist test. I turn next to delineation of geographic markets based on the hypothetical monopolist test when sellers can practice geographic price discrimination.
V. The Merger Guidelines approach to price discrimination
The Merger Guidelines deal explicitly with how geographic markets should be delineated both absent geographic price discrimination and when there is such discrimination. In both cases, one carries out a hypothetical monopolist test. However, the description of that test provided in section III relates to cases without price discrimination.
According to the Merger Guidelines, in using the hypothetical monopolist test to delineate geographic markets, one should first assume that a hypothetical monopolist would not practice geographic price discrimination. One should then ask whether a hypothetical monopolist would impose a discriminatory price increase on buyers in one or more areas. The Merger Guidelines state:
[I]f a hypothetical monopolist can identify and price differently to buyers in certain areas ("targeted buyers") who would not defeat the targeted price increase by substituting to more distant sellers in response to a "small but significant and nontransitory" price increase for the relevant product, and if other buyers likely would not purchase the relevant product and resell to targeted buyers, then a hypothetical monopolist would profitably impose a discriminatory price increase. ... The Agency will consider additional geographic markets consisting of particular locations of buyers for which a hypothetical monopolist would profitably and separately impose at least a "small but significant and nontransitory" increase in price.(23)
Gregory J. Werden clarifies this procedure:
The Guidelines' approach to market delineation initially assumes that price discrimination is not possible and delineates markets on the basis of that assumption. Consequently, markets are delineated in geographic space on the basis of points of production, rather than points of consumption. If price discrimination is possible, the Guidelines delineate additional, smaller markets, on the basis of arbitrage possibilities. In the case of geographic discrimination, markets are delineated on the basis of the locations of consumers. If the cost of arbitrage among consumers were sufficiently high, each consumer could be in a distinct relevant market.(24)
When relevant markets for use in merger analysis are delineated on the basis of the locations of consumers, the hypothetical monopolist test identifies the smallest set of market participants that would raise prices in the market by at least 5% above the level that would prevail absent the merger, without raising prices to other customers, assuming these participants formed a single company.
If a hypothetical monopolist practicing price discrimination would raise prices to an individual wholesale customer by at least 5% above the level that would prevail absent the merger, without raising prices to other customers, that targeted customer would be what the Merger Guidelines identify as an additional, smaller geographic market that would be delineated to reflect opportunities for price discrimination. Even in that case, however, the appendix A methodology for identifying, and for measuring market shares of suppliers that compete in that market based on the delivered price test departs in important ways from the approach of the Merger Guidelines based on the hypothetical monopolist test.(25) Two illustrative examples of these differences between the appendix A and Merger Guidelines approaches follow.
First, appendix A requires that applicants use ceiling transmission prices when they implement the delivered price test, although applicants are permitted to submit an additional analysis using alternative transmission prices. When a question arises regarding which prices should be used in analyzing market power, the approach of the Merger Guidelines is that one should use the prices that are likely to prevail in the future periods under consideration. Absent information to the contrary, the most reasonable assumption regarding transmission prices would generally appear to be that prices in the near future would be similar to prices in the recent past during the same times of the year (for example, summer peak hours). However, it would also make sense to investigate how competition would be affected if companies that might realistically exercise increased market power as a result of the merger would reduce or eliminate discounting of transmission prices.
When discounting of transmission prices below regulated ceilings is significant, implementation of the delivered price test using ceiling rather than discounted prices may cause a merger to have a substantially larger or smaller effect on concentration. Use of ceiling rates is not inherently either conservative or the opposite. Given an assumed price for power in the destination market, use of ceiling transmission prices will tend to reduce the amounts of power that can be delivered economically to the destination market from all utilities other than the local utility. In some cases, potential deliveries from the merging companies will be disproportionately reduced, and the merger will therefore have less effect on concentration, when ceiling rates are used. In other cases, the opposite will be true. It is difficult to predict the direction of the effect in any particular case without doing the analysis, because the impacts of changes in assumptions about transmission prices depend on such things as the configuration of transmission constraints and the variable costs of generation in different locations. For example, transmission prices will not affect a particular supplier's potential deliveries to a destination if those deliveries are limited by transmission capacity rather than costs, or if the supplier's generators have such low costs that they are economic no matter which transmission prices are used. In other cases, a small change in transmission prices may lead to inclusion or exclusion of a significant seller for a destination market.
Second, the appendix A delivered price test assumes that a seller's willingness to sell power in a particular destination market hinges only on the seller's out-of-pocket variable costs for generation and for transmission service to the market. Appendix A ignores the fact that a seller's willingness to sell in a particular destination market will depend on what the seller could obtain for its power elsewhere, that is, its opportunity costs. Suppose a seller has 1000 MW of hydroelectric capacity with a variable cost of $0/MWh, 1000 MW of contracts to purchase energy from nonutility generators on a take or pay basis with a variable cost of $0/MWh, and 1000 MW of nuclear capacity with a variable cost of $5/MWh. Suppose that during summer peak hours the market price for energy where the seller is located is $30/MWh. Suppose also that the cost of delivering energy from the seller's area to destination X is $5/MWh. According to appendix A, this seller would be willing to sell 3000 MWh per hour of energy to destination X at any delivered price greater than $10/MWh. In fact, the seller would not sell a single MWh of energy to destination X at any price below $35/MWh. The seller's willingness to sell to any particular destination is based on the seller's opportunity costs, which may greatly exceed its out-of-pocket variable costs for generation and transmission.
Appendix A's neglect of opportunity costs may cause a merger to have a much greater or a much smaller effect on concentration for a destination market than would be the case if the analysis were done correctly. The direction of the effect is not easy to predict, because it will depend in a complex way on the relationship between market prices and variable costs of generation in different areas, transmission costs, and transmission constraints throughout a region of the country.
VI. Differences between the appendix A and DOJ/FTC methods are important
DOJ has stated that appendix A's focus on individual destination utilities and use of the delivered price test can lead to substantially different results than the hypothetical monopolist test if price discrimination is not practiced. Referring to an analysis based on the hypothetical monopolist test, DOJ stated:
The result of this [hypothetical monopolist] analysis, in theory, can be a narrower or broader geographic area than that produced by the delivered price test, but the latter is more likely, and the difference in the areas produced by the two tests can be dramatic.(26)
The following two examples illustrate the point made by DOJ in the preceding quotation. In the first case, the appendix A method would indicate that a merger would substantially reduce competition when in fact the merger would have no adverse effect on market power--a false positive. In the second case, the appendix A method would indicate that a merger would have no effect on competition when in fact the merger would be likely to lead to a substantial increase in market power--a false negative.
A. Illustration: Appendix A may overstate the importance of competition between merging companies by understating the geographic scope of competition
Let us begin by assuming the following facts regarding the entities depicted in figure 1:
Island Light, the utility that provides distribution service on Island, is directly interconnected to four other utilities--Alpha, Beta, Gamma and Delta. The latter utilities are identical to each other (not to Island Light) and are the only members of the Central Pool. The price of transmission service between each of the Central Pool members and Island Light is $3/MWh. There are no charges for transmission service among the Central Pool members, and there are no transmission constraints among the Central Pool members or between the Central Pool members and Island Light. There is retail competition, and Central Pool members compete with each other and with Island Light to serve Island's retail customers. The prevailing prices of energy (measured at wholesale) are $20/MWh in Island Light's area and $17/MWh in the Central Pool. Central Pool members are also interconnected with numerous members of two large pools, West Pool and East Pool. There are no transmission constraints between the Central Pool, on the one hand, and either the West Pool or the East Pool, on the other. The transmission tariff between members of the Central Pool and the West Pool is $1.5/MWh, while the tariff between members of the Central Pool and the East Pool is $5/MWh. The prevailing price of energy in the West Pool is $17/MWh, and at prevailing prices there are no transfers between the West Pool and the Central Pool. However, if the price in the Central Pool rose even slightly above $18.5/MWh, there would be large transfers of energy from the West Pool to the Central Pool. West Pool members have no generating capacity with a dispatch price below $17/MWh but have a large amount with a dispatch price of $17/MWh. The prevailing price of energy in the East Pool is $22/MWh, and at prevailing prices there are substantial transfers (2000 MWh/hr) from Central Pool members to the East Pool. East Pool members have no generating capacity with a dispatch price below $14/MWh but have a large amount with a dispatch price of $22/MWh. As a result, the amount of energy supplied to the East Pool by Central Pool members does not have a significant effect on the East Pool price. If the price of energy in the Central Pool rose above $17/MWh, East Pool members would stop buying energy from the Central Pool, because the delivered price of energy from the Central Pool would rise above $22/MWh. At a price 5% above the prevailing price on Island--that is, at $21/MWh--Island Light and the four Central Pool members each would supply 20% of the energy available to customers on Island. Even a monopolist cannot practice geographic price discrimination or change transmission prices to specific destinations in this example.
Using Appendix A's methodology and terminology, if one were evaluating a proposed merger of Alpha and Beta, one would delineate Island as a destination market, since it is interconnected with Alpha and Beta. Also:
Using the delivered price test, one would delineate the geographic market for this destination market as the region formed by combining the Island Light and Central Pool areas. At a price of $21/ MWh on Island--5% above the prevailing price of $20/MWh--it would not be economical to deliver any energy from the West Pool or the East Pool to Island given the assumptions that have been made about variable costs for generators in the West Pool and the East Pool and given transmission prices. Therefore, members of these two pools are not in the appendix A geographic market for the Island destination. Island Light and each Central Pool member would have a 20% market share, and the premerger HHI would be 2000. The HHI is the sum of the squared market shares of the sellers in the market. A merger of Alpha and Beta would increase the HHI by 800 to a postmerger level of 2800. The merger would therefore violate FERC's competitive analysis screen. In fact, Alpha could not acquire more than 6.25% of Beta's economic generating capacity without violating FERC's screen, because an acquisition of a larger share would increase the HHI by more than 50.
Now let us consider how the merger would be assessed if one used the hypothetical monopolist test for geographic market delineation. In order to implement the hypothetical monopolist test to delineate the relevant geographic market or markets in which to analyze a merger of Alpha and Beta, one begins with Alpha's area (or Beta's area) and the prevailing price at that location absent the merger, $17/MWh. Skipping over some smaller candidate geographic markets that one would consider, one would soon ask whether a hypothetical monopolist for energy in the Central Pool and Island region would profitably raise prices in the Central Pool area by at least 5% above the prevailing level, that is, to at least $17.85/MWh.
Suppose this hypothetical monopolist raised its ex generator price in the Central Pool area above $17/MWh. What would happen? Given the facts assumed above, as soon as the Central Pool price increased above $17/MWh, the delivered price of energy from the Central Pool to the East Pool would rise above $22/MWh. As a result, the monopolist would lose all energy sales to East Pool members, whose demand curve is perfectly elastic (horizontal) at a delivered price of $22/MWh. As a result, the hypothetical monopolist would lose the excess of revenues earned on these sales over their variable costs.
Readers interested in technical details can consult figure 2, which shows:
The hypothetical monopolist's marginal or incremental cost curve for energy (MC in Central Pool and Island).(27) The demand curve for energy in the Central Pool and Island region (Demand in Central Pool and Island). The demand curve for energy exports to the East Pool, which is perfectly elastic at an ex generator price in the Central Pool of $17/MWh. (This ex generator price is equivalent to a delivered price in the East Pool of $22/MWh, given the price of $5/MWh for transmission service.) (Demand for Exports to East Pool.) The supply curve for energy imports from the West Pool, which is perfectly elastic at a delivered price in the Central Pool of $18.50/MWh (Supply of Imports from West Pool).
Absent the merger, assuming competitive behavior, energy output in the Central Pool and Island region would be determined where the MC curve crosses the $17/MWh line (point Y), or at 5000 MWh. Consumption in the Central Pool and Island region would be 3000 MWh. Exports to the West Pool would be 2000 MWh. The ex generator price in the Central Pool would be $17/MWh.
At the output of 5000 MWh, the hypothetical monopolist's variable costs of generation would be measured by the area under the MC curve, while revenues from sales of energy (ex generator) would be measured by the area under the $17/MWh line. Therefore, if the monopolist operated in the same manner in which the individual utilities would operate absent the merger, the excess of the monopolist's revenues over its variable costs ("profits" for the purpose at hand) would be the area XYZ.
If the monopolist raised the price in the Central Pool slightly above $17/MWh, its sales would decline to 3000 MWh of energy sold to meet loads in the Central Pool and Island region. The monopolist's profits would decline from area XYZ to area XRSZ.
Once the price in the Central Pool was slightly above $17/MWh, the hypothetical monopolist would increase its profits (given inelastic demand in the Central Pool and Island region) by increasing the Central Pool price to $18.50/MWh. However, the monopolist would not attempt to raise the Central Pool price above $18.50/MWh, because if it did so it would lose all of its sales in the Central Pool and Island region to competitors in the West Pool.
Thus, the question for the hypothetical monopolist is which of the following two price and output choices gives it greater profits:
Produce 5000 MWh and sell it at an ex generator price of $17/MWh, obtaining a profit equal to area XYZ. Produce a bit less than 3000 MWh and sell it at an ex generator price of $18.50/MWh, obtaining a profit equal to the area XTUV.
If area XYZ is larger than area XTUV, the monopolist would find the first of these alternatives more profitable than the second. Suppose that the first alternative is the more profitable.(28) In that case, a hypothetical monopolist of the Central Pool and Island region would not raise prices above the competitive level, and the candidate relevant geographic market would have to be expanded to include the West Pool. The incentives of a hypothetical owner of all Central Pool and Island generation to raise prices in this case is eliminated by two features of competition that are ignored by the appendix A methodology: (1) an increase in prices in the Central Pool would result in a loss in Central Pool exports to the East Pool, and (2) the extent to which prices in the Central Pool could be increased would be limited by competition from the West Pool.
Put simply, in this case the appendix A methodology understates the geographic scope of competition and as a result ignores competition that would prevent an exercise of market power by a company that owned all generation that would pass the appendix A delivered price test for the Island Light destination. In the case just discussed, a merger of Alpha and Beta would not create or enhance market power over energy during the pertinent time period. Thus, the appendix A methodology produces a false positive for this period.
B. Illustration: Appendix A may overlook important competition between the merging companies by understating the geographic scope of competition
While appendix A's focus on individual utility destination markets often leads to understatement of the size of the geographic markets relevant for merger analysis, under some circumstances reliance on the appendix A methodology would lead one to overlook a significant loss in competition that would result from a merger.
Suppose that four utilities--P&L, E&G, Edison, and Corp.--are interconnected in the manner shown in figure 3. Suppose also that:
Transmission capacities among these utilities are sufficient that they are never fully utilized. The charge for transmission service on each link is $5/MWh. These four utilities are not interconnected with any other utilities. P&L and E&G each have a portfolio of generating plants with variable costs ranging from $10/MWh to $30/MWh, while Edison and Corp. each have a portfolio of generating plants with variable costs ranging from $17/MWh to $30/MWh. The pattern of generating capacities, generating costs, transmission prices, and loads for the utilities, combined with competitive behavior, results in market clearing prices $15/MWh at P&L and E&G and $20/MWh at Edison and Corp. Both P&L and E&G sell energy to both Edison and Corp.
In this situation, continuing to assume competitive behavior on the part of other utilities, an attempt by any one of the four utilities to exercise market power by reducing output would result in an increase in output by each of the other three utilities. For example, if Edison reduced its output, the prices at Edison and Corp. might rise to $21/MWh while the prices at P&L and E&G would rise to $16. In response to these price increases, P&L, E&G, and Corp. would all increase output.
Let us assume that the supply responses of each of the utilities to higher prices is such that a hypothetical monopolist would have to own the generating plants of all four utilities in order profitably to raise prices at any one of the utilities by more than 5% above the initial prevailing levels. Also, assume that it is not possible to practice geographic price discrimination. In that case, based on the Merger Guidelines methodology, the relevant geographic market in which to analyze a merger of any two of these utilities would be the combined area of all four utilities.
If we assume that there is retail customer choice, and that the output levels of the four utilities would be equal at prices above the initially prevailing levels, before any merger the HHI in the relevant market would be 2500. A merger of any two of the utilities, such as Edison and Corp., would raise the HHI by 1250 to 3750. In order to avoid increasing the HHI by more than 50, the merging companies would have to divest (or otherwise remove from their market shares) 48% of their combined capacity at generators with variable costs sufficiently low to be in the market.
Now consider how a merger of Edison and Corp., the two utilities with higher costs and prices, would be analyzed under appendix A's individual destination market and delivered price test methodology. The appendix A methodology would identify each of the four utilities as a destination market for purposes of the analysis. However, there is no destination market among these four for which energy from both Edison and Corp. would pass the delivered price test, and hence there is no destination market in which both would have a market share. Using the appendix A methodology, one would ask which suppliers could deliver energy to each of the destinations at prices equal to 105% of the initial prices in those areas. Edison could not deliver energy to any destination (other than Edison) at those prices. Since Edison's cheapest generator has a variable cost of $17/MWh, it would be unable to deliver energy to P&L or E&G at a delivered cost below $22/ MWh, and it would be unable to deliver energy to Corp. at a delivered cost below $27/MWh; Corp. is in a symmetric position. In short, the merger would not affect HHIs and would pass the appendix A screen with flying colors.
Why does the appendix A screen miss the competitive problem raised by the merger of Edison and Corp.? The answer is that appendix A is designed to answer the wrong question. Appendix A focuses on which parties can supply energy to a particular buyer. Since Edison and Corp. cannot economically supply energy to any area other than their own, appendix A assumes they play no role in constraining prices outside their own areas. This is not the case. Even though there is no buyer that can be supplied by both Edison and Corp., Edison's generation constrains pricing at Corp. As we saw above, if Corp. attempts to raise prices by reducing its output, all three other utilities will increase their outputs. Corp. will lose sales in its own area to increased imports from P&L and E&G. Sales from P&L and E&G to Corp. will increase not only because P&L's and E&G's outputs increase, but also because Edison's output increases. The increase in Edison's output causes P&L and E&G to divert energy that they were initially selling to Edison over to Corp. Thus, prices at all four destinations are determined by competition among generators located at all four.
C. Implication of the preceding illustrations
The preceding examples indicate that by focusing on individual destination markets rather than geographic markets delineated using the hypothetical monopolist test, the appendix A methodology is likely to overstate competitive problems for some mergers and understate them for others. The overstatement and understatement may be so great that for some mergers the appendix A methodology may find what appear to be serious problems when no problems exist at all, while for other mergers the appendix A methodology may entirely overlook a serious competitive problem.
VII. Geographic markets when there are transmission constraints
Thus far, in discussing geographic market delineation I have assumed that there are no binding transmission constraints--that is, transmission capacity would never be fully utilized, even if market power were being exercised. When that assumption is correct, it is appropriate to ignore limits on transmission capacity in analyzing market power. In fact, however, transmission constraints often play an important role in delineating geographic markets because, under some relevant conditions, transmission capacity would be fully utilized.
In discussing the effects of transmission constraints on geographic markets, I will address two types of situations: (1) situations in which transmission capacity would be fully utilized both without a proposed merger and if a hypothetical monopolist in a relevant market maximized its profits, and (2) situations in which transmission capacity would become constrained if a hypothetical monopolist in a relevant market maximized its profits. For purposes of the discussion of transmission constraints in section VII, I will assume that utilities cannot practice geographic price discrimination.
A. Situations in which transmission capacity would be fully utilized regardless of the merger
Suppose there are two regions, North and South. Suppose that within each region there are no transmission charges or transmission constraints and that the charge for transmission service between North and South is $1.50/MWh. Suppose further that, absent any mergers, there would be significant competition among utilities in the North and among utilities in the South, the price of energy in the North would be $20/MWh, transmission capacity from North to South would be fully utilized, and the price of energy in the South would be $24/MWh (see figure 4).
Now suppose that two utilities in the South propose to merge. The South would be the relevant geographic market for this merger, because a hypothetical monopolist that was the only seller in the South would raise energy prices by at least 5% above the level that would prevail (given competition among utilities in the South) absent the merger. Because transmission capacity from the North to the South would be fully utilized even without a price increase in the South, buyers in the South could not increase their purchases from the North, and prices in the North would not increase, in response to an increase in prices in the South.
Now suppose instead that two utilities in the North propose to merge. One might think that the South should be included in the relevant geographic market for this merger, because transmission capacity from the South to the North is available. However, the South would not be in the relevant geographic market. The relevant geographic market would be the North, because a hypothetical monopolist that was the only seller in the North would raise prices by more than 5% above the level that would prevail (given competition among utilities in the North) absent the merger. Suppose that a hypothetical monopolist that owned all generators in the North raised prices by 5% to 10%, to a level of $21 to $22 per MWh. With prices in the South at $24/MWh, this price increase in the North would not cause buyers in the North to switch any purchases to generators in the South; indeed, there would not even be a reduction in exports from the North to the South. It follows that a 5% to 10% price increase in the North would not cause a price increase in the South.(29)
Finally, suppose that a utility in the North proposes to merge with a utility in the South. As long as maximization of profits by a hypothetical monopolist of both North and South would not cause a reduction in use of transmission service from North to South, the North and the South would be separate geographic markets during the pertinent time period for the reasons discussed above. If maximization of profits by a hypothetical monopolist of both North and South would cause transmission capacity between the North and the South to become less than fully utilized, however, it is likely that prices in the North and South would move together and that the North and South combined would be an additional relevant geographic market.(30)
In the cases discussed here, if the North (the exporting area) is a separate geographic market, generators located in the South would not be competitors in the North and should not be given shares in the North market. The situation is more complex if the South (the importing area) is a separate geographic market. Even though imports into the South cannot increase in response to an anticompetitive output reduction in the South, the market value of those imports in the South would increase. The entities to which that increase in market value would accrue would benefit, other things equal, from an exercise of market power in the South.
A reasonable case can therefore be made for attributing shares in the South market to the entities that would capture the increased market value of imports into the South.(31) When transmission to the South is constrained, an increase in prices in the South increases the scarcity value of transmission service from the North to the South, and therefore increases "rents" accruing to the parties that obtain that scarcity value. Market power in the South may be influenced by how these rents are distributed. There are numerous possibilities. If transmission prices adjust to reflect increasing scarcity, as they would under congestion pricing, then the recipients of the higher transmission prices benefit. If transmission prices do not adjust, then there will be nonprice rationing of scarce transmission service that is underpriced (that is, priced below the level at which demand would equal supply). The entities that obtain the underpriced transmission service will capture the scarcity value of the service. Various ways that nonprice rationing might in principle be used to allocate underpriced transmission service include priority for certain uses, such as native load; first-come, first-served; or allocation to the transmission system owner or its affiliates.
In any event, entities that receive rents based on the scarcity value of transmission from the North to the South and as a result have an incentive to raise prices in the South could reasonably be given a market share in the South. It follows that market shares relating to imports by the South should not necessarily be attributed to generators in the North.(32) In a particular case, it may be appropriate to allocate market shares based on imports to some combination of distribution utilities in the South, generators in the North, transmission companies, power marketers, etc. This point is one reason that control over transmission rights across constrained interfaces plays a role in merger analysis.
B. Situations in which transmission capacity would become fully utilized if a hypothetical monopolist maximized profits
The preceding discussion requires modification if transmission capacity would not be fully utilized absent a merger but would become fully utilized if a hypothetical monopolist in a relevant market maximized its profits. For illustrative purposes, assume the following facts:
There are two regions, West and East, and two utilities in the East propose to merge. There are no internal transmission charges or constraints in the East or in the West. Absent the merger, there would be significant competition among utilities in the West and among utilities in the East, and there would be 500 MW of available (unused) transmission capacity from West to East. As one uses the hypothetical monopolist test to evaluate successively larger candidate geographic markets centered around one of the merging utilities, one begins to add generators located in the West. However, before one arrives at a geographic market that satisfies the hypothetical monopolist test, buyers in the East would have increased purchases from the West by 500 MW, so that transmission capacity from the West to the East would be fully utilized.
While one could debate semantic issues, in this case it seems reasonable to identify the East as a relevant geographic market.(33) In that case, one would include the megawatts of transmission capability from West to East in measuring the size of the East market. The difficult issue is to figure out which parties should be assigned market shares based on these megawatts of transmission capability.
Under the assumptions made here, prices in the West may or may not increase as a result of an exercise of market power in the East. To the extent prices in the West would increase, it would be reasonable to assign market shares in the East (based on megawatts of transmission capability from West to East) to generators in the West in proportion to the benefits these generators would obtain from these higher prices. To the extent that rents based on transmission capacity from the West to the East would increase as a result of an exercise of market power in the East, it would make sense to assign market shares in the East (based on megawatts of transmission capability from West to East) to whichever entities would capture those rents.
C. Conclusions regarding transmission constraints
Transmission constraints have two important implications for analyses of the geographic scope of competition. First, transmission constraints can reduce the size of geographic markets that are appropriate for analysis of market power in the electric power industry. This is true not only when one is analyzing market power on the importing side of a constrained interface but also when one is analyzing market power on the exporting side.
Second, when transmission is constrained, appropriate assignment of market shares based on import capability into a market can be difficult. The problem is to determine which entities would benefit from increases in the scarcity value of transmission capacity and from any price increases that occur in exporting areas. Of the methods used in appendix A analyses to date, proration of transfer capability based on economic (or available economic) capacity would be appropriate under some conditions but not others; assignment to the cheapest generating capacity outside the constraint that is unlikely to be appropriate.
VIII. Individual utilities as antitrust markets
Appendix A and subsequent Commission orders specify that to evaluate the effects of a merger on generation market power one should delineate as a separate destination market each wholesale customer that is a potential victim of newly created or enhanced market power. However, a single wholesale customer will be a true antitrust market for purposes of merger analysis only under limited circumstances. Consequently, before delineating an individual wholesale customer as an antitrust market for a merger, one should determine that these circumstances exist. The purpose of section VIII is to discuss these circumstances.
For discussion purposes, I divide the appendix A individual destination markets for a merger of Alpha and Beta into three groups: (1) the destination market for Alpha's transmission dependent utilities (TDUs), and the separate destination market for Beta's TDUs; (2) other destination markets that are directly interconnected to Alpha and/or Beta; and (3) other destination markets that are not directly interconnected to Alpha or Beta. I discuss these three groups in sequence.
A. Under what circumstances would Alpha's TDUs be a relevant market?
In carrying out an antitrust analysis of a merger of Alpha and Beta, one assumes that absent their merger Alpha and Beta would each set its prices in a manner that would maximize its profits. Alpha might exercise a certain amount of market power over its TDUs absent the merger. However, in order for Alpha's TDUs to be a relevant market for purposes of analyzing a merger involving Alpha, one must identify a hypothetical company that would have the ability and incentive to raise prices for Alpha's TDUs by at least 5% above the levels that would prevail absent the merger--while at the same time, by assumption, prices would not be increased for customers in surrounding areas.
Sections II through VII of the present article discuss two factors that might in principle limit the size of relevant geographic markets--price discrimination and transmission constraints. Section VIII.A discusses the limited circumstances under which each of these two factors separately would lead to a relevant market limited to Alpha's TDUs.
1. PRICE DISCRIMINATION In order for one to justify identification of Alpha's TDUs as a geographic market based on price discrimination, utilities would have to be in a position to target Alpha's TDUs for a significant price increase relative to both the price that would prevail absent the merger and prices in surrounding areas. Generally, the conditions for such price discrimination appear unlikely to hold (see sections II and IV).
2. TRANSMISSION CONSTRAINTS The next issue is to identify circumstances under which, absent price discrimination, a transmission constraint would justify delineation of Alpha's TDUs as a relevant market for purposes of analyzing a merger involving Alpha. A discussion of the role of transmission constraints is provided in section VII. However, that discussion focuses on mergers between companies that are located on the same side of a transmission constraint. If Alpha's TDUs were identified as a relevant market because of a transmission constraint, generally the constraint would surround Alpha's area, and Alpha's merger partner (Beta) would be on the other side of the transmission constraint from Alpha.
In order to determine whether it is appropriate to delineate Alpha's TDUs as a relevant market for a merger based on limited import capability into Alpha's area, the first step is to set out a competitive theory that would be consistent with delineation of such a market. The next step is to determine whether the theory is consistent with the facts of the particular area. For purposes of this discussion, I will consider the following case:
Alpha owns all generating capacity in its area and makes sales to its TDUs. While Alpha may exercise some unilateral market power, absent the merger Alpha's profit-maximizing prices would be constrained by competition from generators located outside its area. Absent the merger, there would be net imports of electric power into Alpha's area during the time period (e.g., representative hour) in question;(34) however, the import capability into Alpha's area (from pertinent directions) would not be fully utilized.
Under the assumptions made above, one might consider the following competitive theory: A merger between Alpha and Beta would give the merged company an incentive, which Alpha alone would not have, to reduce output from Alpha's generators. This output reduction would lead to a price increase, and this in turn would cause Alpha's TDUs to switch some of their purchases of electric power from Alpha to imports from generators outside Alpha's area. Moreover, the output reduction for Alpha's generators would be large enough so that the import capability of Alpha's area would become fully utilized. Once that import capability would be fully utilized, the merged company would continue to reduce Alpha's output until there was a substantial price increase for Alpha's TDUs. The latter price increase would be large enough so that the merged company would increase its profits by adopting this strategy, even though the merged company's sales of output from Alpha's generators would be reduced by enough so that the import capability of the Alpha area would be fully used. So far, or course, this story is only a theory.
Before delineating Alpha's TDUs as an antitrust market based on this theory, one should confirm that the theory makes sense given the facts of the case at hand. Specifically, one should identify a hypothetical company that would have the ability and incentive to raise prices in the manner described by the theory while Alpha alone would not. This hypothetical company (or hypothetical monopolist) would include Alpha, but it would have to be larger than Alpha because Alpha would not independently pursue this strategy.
Suppose that one identifies a hypothetical monopolist with the ability and incentive indicated by the theory. Presumably this hypothetical monopolist's incentive would differ from that of a stand-alone Alpha because--given an anticompetitive price increase in Alpha's area--the hypothetical monopolist would obtain the price increase on a substantial share of imports into Alpha's area, while Alpha alone would not obtain the price increase on these imports. That is, the hypothetical monopolist might somehow capture a significant share of any increase in the scarcity value of transmission into Alpha's area that would result from an anticompetitive reduction in the output of Alpha's generators (see section VII).
The preceding competitive theory may not make sense in a particular case for a number of reasons. For example, suppose that in response to a price increase targeted at Alpha's TDUs, those TDUs would be able to switch all their purchases from Alpha to other suppliers before the import constraint around Alpha's area would become fully utilized (from pertinent directions). In that case, the import constraint around Alpha's area would be of no consequence insofar as Alpha's TDUs are concerned and could not justify delineation of Alpha's TDUs as a relevant market. The strategy described by the competitive theory would not be profitable; Alpha would lose all sales to its TDUs.
Also, before accepting the theory, one should determine that, once the import constraint became fully utilized, the hypothetical monopolist would raise prices by enough so that it would earn higher profits on its lower volume of remaining sales. Suppose that absent the merger Alpha would have sold energy to its TDUs at a price of $30/MWh. Suppose that in order to make the strategy described by the theory profitable, a hypothetical monopolist would have to raise the price to $75/MWh. Finally, suppose that for "political" reasons--for example, a threat of regulation--the hypothetical monopolist could not increase prices above $60/MWh. In that case, the hypothetical monopolist would not engage in the strategy described by the theory, and it would not be proper to treat Alpha's TDUs as a separate destination market.
3. CONCLUSIONS ON ALPHA'S TRANSMISSION DEPENDENT UTILITIES AS AN ANTITRUST MARKET Price discrimination and transmission constraints are unlikely to provide a justification for delineating Alpha's TDUs as a relevant antitrust market for purposes of analyzing the effects of a merger between Alpha and Beta except under limited circumstances. Therefore, before one analyzes the competitive effects of a merger on prices in a target area limited to Alpha's TDUs, one should establish through an investigation of the market in question that it is reasonably likely that Alpha's TDUs would be targeted for a significant price increase, relative to the price that would prevail absent the merger, by a hypothetical monopolist. A necessary (not sufficient) condition for customers in Alpha's area to be a relevant antitrust market based on a transmission constraint is that the megawatts of purchases that the customers would be free to switch to suppliers outside the area exceed the import capacity available to these customers (from pertinent directions). In addition, there should be evidence that a hypothetical monopolist could and would raise prices high enough to offset foregone profits on sales it would lose both inside and outside Alpha's area.
This article focuses on geographic market delineation, but it is important to keep in mind that market delineation, market shares and HHIs are merely tools. If properly used, these tools can give insights into whether a merger is likely to create or enhance market power. However, these tools can easily be misused.
An illustration of the misuse of HHIs based on the topic of this section may be useful. Suppose that the hypothetical monopolist test indicates that it is appropriate to define Alpha's TDUs as a relevant market based on a transmission constraint, taking into consideration the facts in the particular case. Suppose that in the market defined as Alpha's TDUs, Alpha has a share of 45% while imports account for the remaining 55%. Suppose that it is appropriate to attribute 2% of the total market to Beta. In that case, a merger of Alpha and Beta would give the merged firm a share of 47% and would increase the HHI by 180, while the postmerger HHI would exceed 2000. Comparing these HHI figures to the thresholds in the Merger Guidelines, one might be tempted to jump to the conclusion that the merger would be anticompetitive, assuming entry barriers.
However, market shares and HHIs are aids in thinking carefully about competitive effects; they are not a substitute for analysis. Even if a hypothetical monopolist would exercise significantly more market power than would Alpha alone, it is not obvious that a market share increase from 45% to 47% for the owner of Alpha's generating plants would cause the merged firm to have an incentive, which Alpha alone would not have, to reduce generation at Alpha's plants to the point at which import capability into the Alpha area would become fully utilized.
Against this background, it is difficult to think of a rationale for requiring that the TDUs of each merging company be treated as a relevant market for every time period for every merger, or for requiring that applicants take steps to prevent the HHI in any such market from increasing by more than 50 or 100.
B. Under what circumstances would other individual utilities that are directly interconnected to the merging utilities be relevant markets?
Aside from differences in detail, there is one major distinction between analyzing Alpha's TDUs as a potential target area for a price increase and analyzing another utility, Gamma, that is directly interconnected to Alpha and/or Beta as a potential target area. The implication of this distinction is that Gamma is less likely to be a relevant market than are Alpha's TDUs.
To understand this distinction, suppose that the only thing that could make it possible for anyone to target a single wholesale customer for a price increase is that import capability into that customer's area would be fully utilized under at least some relevant conditions.
Suppose that Alpha owns a substantial share of generating capacity located inside any transmission constraint surrounding its area. In that case, given entry barriers, when transmission capacity into its area is fully utilized Alpha may have some ability to raise prices to its TDUs above prevailing levels.
By contrast, if neither Alpha nor Beta owns a significant amount of generating capacity located in Gamma's area, then neither Alpha nor Beta (nor a merged Alpha/Beta) would have the ability to raise prices in the Gamma area when transmission capacity into the Gamma area is fully utilized. As a result, in analyzing a merger of Alpha and Beta, one cannot justify analyzing Gamma as a separate antitrust market based solely on the existence of a transmission constraint--even a fully utilized transmission constraint--surrounding Gamma. The justification for delineating Gamma as a separate market would have to be based on something else--for example, price discrimination.
There is a further limitation on the usefulness of treating Gamma alone as a relevant market if one is focusing on an exercise of market power engaged in solely by the merged company. If Gamma has interconnections with utilities (say, Delta and Epsilon) that are not directly interconnected to Alpha or Beta, the merged company will often be unable to prevent arbitrage between Delta and Epsilon, on the one hand, and Gamma, on the other. In that case, a merged Alpha/Beta could not target a price increase to Gamma; any action by Alpha/Beta that would increase prices to Gamma would increase prices in a larger region including at least Delta and Epsilon as well. From this one can conclude that analyses based on individual utility destination markets will often have to assume not only some form of geographic discrimination but also exercise of market power by more than one utility.
C. Under what circumstances would other individual utilities that are not directly interconnected to the merging utilities be relevant markets?
FERC requires that each important trading partner of either merging company be treated as a separate destination market for purposes of merger analysis. However, I have found it difficult to imagine a scenario in which a merger could result in a price increase for a wholesale customer that is not directly interconnected with either merging party without also resulting in a price increase for intervening utilities. The existence of important trading partners that are not directly interconnected with the merging companies would be likely to support a larger regional market rather than additional individual utility markets.
Notwithstanding criticisms of its reliance on a methodology that assumes widespread geographic price discrimination, FERC has proceeded in the direction of codifying the appendix A methodology.(35) FERC's position appears to be that even if its methodology for delineating geographic markets and for other steps in analyzing market structure differs from that used by the Merger Guidelines, appendix A is a reasonable and conservative screening device that permits FERC to identify mergers that do not require further analysis, and is superior in this regard to alternatives that are available or that FERC could develop in the near term.(36)
There are two difficulties with FERC's apparent position, even if I limit my focus to the portion of appendix A dealing with geographic market delineation. First, the methodology based on individual destination markets and the delivered price test is not necessarily conservative; rather, it is unreliable. The way in which geographic markets are delineated by FERC's method may lead to false negatives as well as false positives. Moreover, where FERC's method identifies a real problem, its method may lead to an inaccurate diagnosis regarding the remedy required to mitigate the concern.
Even if potential geographic price discrimination warranted the delineation of a narrower geographic market in a particular merger case, the narrower market would not necessarily be a single utility. And even if one utility was a relevant antitrust market, it would not follow that each of what appendix A designates as a destination market would be a separate relevant antitrust market.
Also, even if a particular utility is a relevant antitrust market, the delivered price test methodology set out in appendix A for identifying the competitors in that market and for measuring market shares often is not appropriate.(37) Finally, even if there were a narrow relevant market based on price discrimination, in applying the Merger Guidelines methodology one would typically delineate an additional, broader geographic market that does not assume price discrimination.
Second, even if the appendix A methodology were unambiguously conservative, this fact would not warrant a conclusion that the methodology was in the public interest. Applicants generally act as though their mergers must pass FERC's appendix A competitive analysis screen, or concerns raised by violations of that screen must be mitigated, in order to win approval by FERC. The possibility of a trial-type hearing at FERC to test the merits of an additional, alternative analysis appears to be seen more as a threat than an opportunity. Because the screen is the final analysis for most electric utility mergers, FERC should not use a screen that is more conservative than the antitrust standards that apply to virtually all other industries.
There are at least three reasons that FERC should revise its methodology for defining geographic markets so that it is based on the hypothetical monopolist test and consistent with the Merger Guidelines. The first and most important reason is to assure that FERC will reach correct conclusions in assessing market power and remedies. The second is that merging companies must now satisfy both FERC, which applies one methodology to delineate geographic markets, and the federal antitrust agencies, which apply another. Use of a single methodology--that in the Merger Guidelines--would be helpful. Third, FERC's methodology may not withstand a legal challenge, while the Merger Guidelines methodology has been widely accepted by courts.
(1) Department of Justice and Federal Trade Commission Horizontal Merger Guidelines (1992, rev'd 1997), reprinted in 4 Trade Reg. Rep. (CCH) [paragraph] 13,104 [hereinafter Merger Guidelines], particularly [subsections] 1.0-1.22.
(2) Inquiry Concerning the Commission's Merger Policy Under the Federal Power Act: Policy Statement, Order No. 592, FERC Stats. & Regs. [paragraph] 31,044 (1996), order on reconsideration, Order No. 592-A, 78 FERC [paragraph] 61,321 (1997) [hereinafter Merger Policy Statement].
(3) Notice of Proposed Rulemaking, Docket No. RM98-4-000, 63 Fed. Reg. 20,340 (April 24, 1998).
(4) For simplicity, throughout this article I assume that there are no long-term capacity or energy contracts. In real cases, one should take these contracts into account in analyzing market power.
(5) Throughout the present article, which addresses generation market power, I assume that electric power companies do not exercise transmission market power. Transmission market power refers to the ability of one or more companies profitably to raise prices for electric power by affecting adversely the availability of transmission service required by competing sellers to reach customers. In recent orders in merger cases, FERC has stated that Order Nos. 888 and 889 make it appropriate to analyze mergers on the assumption that utilities do not exercise transmission market power. It follows that transmission market power is not FERC's rationale for focusing on individual destination utilities in analyzing market power,
(6) M.W. Frankena, Analyzing Market Power Using Appendix A of FERC's Merger Policy Statement: Rationale, Reliability, and Results, 1 CCH POWER & TELECOM L. 29 (Jan./Feb. 1998).
(7) All transmission prices in this illustration and other illustrations include ancillary services and losses.
(8) POWER MARKETS WEEK, Feb. 16, 1998, at 1, 10-12.
(9) Notice of Proposed Rulemaking, supra note 3, at 31.
(10) See Merger Guidelines [sections] 1.21 n. 12.
(11) See, for example, G. J. Werden, The History of Antitrust Market Delineation, 76 MARQUETTE L. REV. 123 (1992).
(12) Other criteria restrict the shape of the market to prevent gerrymandering. See G. J. Werden, Market Delineation and the Justice Department's Merger Guidelines, 3 DUKE L. J. 514, 528 (1983).
(13) DOJ has stated that, for purposes of delineating markets for merger analysis, "The threshold typically used for what is a `significant' price increase is five percent above levels that likely would prevail absent the merger." DO J, Application of the Horizontal Merger Guidelines to Mergers of Electric Utilities, appendix to Comments of the U.S. Department of Justice, In the Matter of Merger Policy under the Federal Power Act, FERC Dkt. No. RM96-6, May 7, 1996, at A2. G. J. Werden, Market Delineation under the Merger Guidelines: A Tenth Anniversary Retrospective, 38 ANTITRUST BULL. 517, 529 n.31 (1993), explains that the 5% threshold is applied to the product of the merging firm around which the market is delineated, and discusses circumstances in which a threshold higher or lower than 5% may be appropriate.
(14) This conclusion assumes that no geographic area smaller than Tennessee satisfies the criteria for a relevant market.
(15) For court cases that have followed the Merger Guidelines approach to market delineation, see Werden, supra note 13, at 519 n.8.
(16) DOJ, supra note 13, and Comments of the U.S. Department of Justice, Inquiry Concerning the Commission's Policy on the Use of Computer Models in Merger Analysis, FERC Dkt. No. PL98-6, June 11, 1998.
(17) The Merger Guidelines ([sections] 1.11 n.10) explain that "The terms of sale of all other products are held constant in order to focus market definition on the behavior of consumers. Movements in the terms of sale for other products, as may result from the behavior of producers of those products, are accounted for in the analysis of competitive effects and entry."
(18) D.T. Scheffman & P. T. Spiller, Geographic Market Definition Under the U.S. Department of Justice Merger Guidelines, 30 J.L. & ECON. 123 (1987); J. R. Morris & G. R. Mosteller, Defining Markets for Merger Analysis, 36 ANTITRUST BULL. 599 (1991). For a study of the geographic scope of economic markets for electric energy, see E. Bailey, Electricity Markets in the Western United States, 11 ELECTRICITY J. 51 (July 1998).
(19) The Merger Guidelines ([sections] 1.11) state that in defining relevant antitrust markets "the Agency will use prevailing prices of the products of the merging firms and possible substitutes for such products, unless premerger circumstances are strongly suggestive of coordinated interaction, in which case the Agency will use a price more reflective of the competitive price." Coordinated interaction refers to tacit or other collusion.
(20) Appendix A uses the term "geographic market" differently than the Merger Guidelines do. Appendix A implicitly assumes that individual destination utilities are what the Merger Guidelines would call relevant geographic markets if sellers practiced price discrimination among individual destination utilities. However, appendix A refers to these areas as "destination markets" rather than "geographic markets." The step that appendix A describes as delineation of a geographic market around a destination utility is treated by the Merger Guidelines not as part of the process of market delineation but rather as the separate step of identifying the sellers that are competitors in the relevant market.
(21) DOJ, supra note 16, at 3 n.7.
(22) DOJ, supra note 13, at A7 n.8.
(23) Merger Guidelines [sections] 1.22, footnote omitted, emphasis added.
(24) Werden, supra note 13, at 533, footnote omitted.
(25) Frankena, supra note 6.
(26) DOJ, supra note 16, at 3 n.8.
(27) For simplicity, figure 2 ignores transmission costs between Island and the Central Pool.
(28) Given the other assumptions in this example, which alternative is more profitable depends on the shape assumed for the MC curve.
(29) It is possible to construct an example in which a hypothetical monopolist in the North would raise prices in the North so high that transmission from the North to the South would no longer be constrained. In that case, there may also be a relevant geographic market consisting of the North and South combined.
(30) Even if transmission capacity from North to South were fully utilized during one period (for example, a representative summer peak hour), transmission capacity between North and South might not be fully utilized during another period (for example, a representative winter peak hour). As a result, North and South combined might be a relevant geographic market during winter peak hours even if the combined area were not a relevant geographic market during summer peak hours.
(31) When imports cannot be increased in response to an exercise of market power, a case can also be made for doing an additional calculation in which imports are not given a market share.
(32) Even when it is appropriate to attribute market shares relating to imports to generators in the North, it is not likely to be appropriate to attribute these shares solely to the cheapest generators in the North. All generators in the North that can generate energy at a variable cost below the market price in the North have the same opportunity cost for selling to the South.
(33) A reasonable argument could be made for delineating the East and West combined as an additional geographic market in order to avoid missing competitive problems that would appear if one used a lower threshold for the price increase imposed by a hypothetical monopolist. This argument would appear to be consistent with the discussion in Werden, supra note 13, at 529 n.31.
(34) Other things equal, if there were net exports from Alpha's area, the chances that Alpha's TDUs would be a relevant market would be lower, because the anticompetitive strategy discussed below would sacrifice profits on export sales.
(35) After this article was written and circulated, a number of FERC staff members indicated that they agreed with various points made here, and it appears that FERC has moved its codification to a back burner.
(36) The data required to carry out the hypothetical monopolist test are not substantially different from those already required by the appendix A methodology. Both methodologies rely on data for generation capacities and costs, transmission capacities and costs, and loads over a wide region. The two methodologies simply analyze the data in different manners.
(37) Frankena, supra note 6.
MARK W. FRANKENA, Principal, Economists Incorporated, Washington, D.C.
AUTHOR'S NOTE: This article is based on an article prepared for the Edison Electric Institute and used with their permission.
|Printer friendly Cite/link Email Feedback|
|Author:||Frankena, Mark W.|
|Date:||Jun 22, 2001|
|Previous Article:||Critical thinking about "critical loss" in antitrust.|
|Next Article:||The role of the essential facilities doctrine.|