MARKET DOMINANCE AND ANTITRUST POLICY doc

And fourthly, what relation is there between the monopoly illustratedand firms in an antitrust case in a position of market dominance?. If theincumbent firm continues to restrict output

MARKET DOMINANCE AND ANTITRUST POLICY, SECOND EDITION

For Mark and Harry

Market Dominance and Antitrust Policy, Second Edition

© M.A Utton 2003

All rights reserved No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without the prior permission of the publisher.

Published by Edward Elgar Publishing Limited Glensanda House

Montpellier Parade Cheltenham Glos GL50 1UA UK

Edward Elgar Publishing, Inc.

136 West Street Suite 202 Northampton Massachusetts 01060 USA

A catalogue record for this book

is available from the British Library

Library of Congress Cataloging in Publication Data

Utton, M A (Michael A.), 1939–

Market dominance and antitrust policy / Michael A Utton.–2nd ed.

p cm.

includes index.

1 Industrial concentration 2 Market share 3 Antitrust law I Title.

HD2757 U88 2003 338.8–dc21

2002029832

ISBN 1 84064 728 0 (cased)

Printed and bound in Great Britain by Biddles Ltd, www.biddles.co.uk

List of figures and tables vi

Preface to the first edition vii

Preface to the second edition ix

2 Market dominance in practice: current perceptions and trends 27

3 The antitrust response: an outline of antitrust policy in Europe

v

Figures and tables

Figures

1.2 Price and output under competition and monopoly with

9.1 Incentives to integrate under different production conditions 211

Tables

2.1 Average five-firm concentration for a sample of 121 manufactured products in the UK, 1958–77 (unadjusted and

2.2 Long-run shares of market leaders in six UK industries 39

8.1 Calculation of the Herfindahl (H) index in a market with six

vi

Preface to the first edition

Even to the casual observer it might appear that problems of market dominanceand antitrust policy are almost daily in the news Amongst the more prominentexamples are the following: officials of the European Commission pay unan-nounced visits to the headquarters of some of the most famous companies inthe world and seize documents that appear to show that they have been colluding

on prices; a takeover bid for a highly respected and long-established UK fectionery firm is made by a foreign company and is allowed to proceedunhampered despite widespread protest; in the aviation industry a company whosename is a household word is accused of using predatory tactics to ruin a muchsmaller competitor; in the UK the most prestigious and successful brewingcompanies are horrified to learn that the Monopolies and Mergers Commissionhave recommended that they should be forced to divest themselves of a largeproportion of their retail outlets or pubs; in the US the most successful computersoftware company is accused of anticompetitive behaviour Many other examplescould be cited and in subsequent chapters we will look in detail at cases fromthe EU, the US and the UK involving collusion, mergers, the market conduct

con-of dominant firms and the market power that may or may not derive both fromvertical integration and from vertical restraints

The issues and institutions involved, like industry itself, are complex Thebook is therefore structured in a way which we hope will allow the reader tomake sense of the complexity There are four sections Part I contains ananalysis of market dominance and its possible extent, with a preliminary review

of the institutions used to deal with it The core of the book is then contained

in Parts II and III, distinguishing horizontal from vertical issues In Part II,particular attention is paid to the market conduct of dominant firms, which hasreceived so much recent theoretical attention It also contains a discussion ofcollusion, where the antitrust response has perhaps been the most uniform, andhorizontal mergers, where despite very intensive study many issues still remainunresolved In Part III, the emphasis switches to vertical issues: that is, thoseinvolving the relationships between firms and their input suppliers or their dis-tributors Both theory and policy in this area have undergone significant changes

in recent years Finally, in Part IV we raise a number of controversial questionsabout the effectiveness of antitrust policy, including a discussion of the appro-priate sanctions both against those who infringe the law and against those whoattempt to mould its application to their own purposes We also take up sensitivequestions involving conflicts between antitrust and trade policies, the interna-tional ‘reach’ of antitrust and foreign takeovers of domestic firms

vii

In courses in industrial and business economics in British universities,antitrust policy tends to be relegated to a brief final chapter or passing reference

to a few well-known cases Even though much of the preceding analysis mayhave led up to some apparently important policy conclusions, the next step –how these are or are not translated into actual policy – is often left unanswered

or merely given a fleeting reference On the other hand, students of tion law or competition policy may acquire a detailed knowledge of manycases without appreciating the economic analysis that may or, in some notoriousinstances, may not underpin them By bringing together in each chapter of Parts

competi-II and competi-III a discussion of the economic analysis and then the treatment of theissues in European, British and US antitrust policy, we have attempted toovercome this limitation

The intention, therefore, is that the book should provide a useful niment to courses in industrial and business economics, competition law andinstitutions, and in some instances microeconomics where there is an emphasis

accompa-on market power issues The level of ecaccompa-onomics assumed is no more than thatusually taught to first and second year undergraduates and what little algebra

is used has been largely relegated to the appendices

Many of the topics have been discussed over several years with businesseconomics students at Reading, and I have benefited greatly from their scepticismand readiness to challenge the conventional wisdom I would especially like tothank Lauraine Newcombe who coped superbly with the daunting task of deci-phering my handwriting and preparing the final draft for the publisher

M.A.U

Preface to the second edition

I have used the opportunity of a second edition to make a number of substantialchanges Some of these were necessary because the antitrust laws themselveshave changed, and some were required to discuss important recent cases

In the first category was the British Competition Act passed in 1998 and whichcame into effect in March 2000 The changes embodied in the Act wereprobably the most substantial made in the fifty years of British competition policy

In the process of aligning British policy closely with that of the EuropeanUnion, much of the previous machinery was swept away and existing institutionswere fundamentally altered The detailed discussion of the previous policy,especially that involving the Restrictive Practices Act, has therefore beendiscarded to make way for coverage of the new policy Similarly, fundamentalchanges have been made to certain aspects of European policy, especially thatdealing with vertical restraints The second edition, therefore, focuses on thechanges that came into effect in 2001

In the second category are a number of cases which not only are highlysignificant for antitrust policy but have been widely reported and extensivelydiscussed Probably the most prominent was the case brought by the USDepartment of Justice against Microsoft, allegedly for trying to monopolize themarket for operating systems The proposed acquisition of McDonnell Douglas

by Boeing was widely reported for different reasons The merger of twoquintessentially US companies was challenged by the European Unioncompetition authority Until a compromise was agreed, the case threatened torupture US–EU commercial relations Discussion of these and other cases areincluded in the new edition I have also renumbered references to Articles 85and 86 as Articles 81 and 82 throughout, in accordance with the current Treaty

of Amsterdam

Antitrust policy continues to evolve and inevitably, at the time of writing(February 2002), the precise details of further changes in Britain (promised inthe Enterprise Bill) and in the EU (concerning the future structure of cardistribution) are not finalized By the time the book is published they will be.The outcome of these and subsequent changes will have to await a further edition!

M.A.U

ix

PART I

ANALYTICAL AND INSTITUTIONAL BACKGROUND

1 The economic analysis of market dominance

I Monopoly and market dominance

The immediate task of this opening chapter is to discuss some of the basicconcepts that will be used throughout the book We begin by setting out insimple terms the economic case against market dominance using the well-known tools of static theory to highlight the inefficiencies that can arise frommonopolized compared with competitive markets The discussion also allows

us to compare the ‘monopoly’ of economic theory with the looser notion of

‘market dominance’ which often finds its way into antitrust cases

Although the static analysis highlights the core of the problem, real markets

do not remain frozen but are constantly undergoing change, even if the majorparticipants try to resist In Section II, therefore, we extend the preliminaryanalysis into the difficult area of dynamics and introduce a question which willrecur at many subsequent points in our discussion: namely, if in many cir-cumstances positions of dominance generate a faster rate of growth or can beeroded by pressures in the market, is ‘benign neglect’ a more efficient solution

to the problem than direct intervention by antitrust action which may be costly,cumbersome or even wrong? In other words, we need to consider the question

of how selective antitrust policy should be In Section III we try to allay thesuspicions of those who feel that antitrust action is unimportant because thelosses it tries to correct and repair are trivial There is now quite a lot of evidencethat the costs, broadly interpreted, of market dominance can be considerableand, although not as important, say, as hyperinflation or the depletion of theozone layer, are nevertheless great enough to merit detailed enquiry

A monopolist in economic theory is the sole producer of a good or service forwhich there are no close substitutes Some impediment also exists whichprevents other firms from entering the market and competing with the incumbent.Under these circumstances the firm can choose that output and thence pricewhich maximizes profit The barrier to entry will also ensure that, whatever thesize of the profit, there will be no competition to affect the firm’s performance

It will be useful for our subsequent discussion to represent these familiarresults in a simple diagram Thus, in Figure 1.1, the monopolist’s demand and

marginal revenue are shown as AR and MR, respectively, and long-run average cost, denoted by LRAC, is assumed to be roughly L-shaped and constant beyond output Q E This representation of the average cost curve has substantial empirical

support (see Johnston, 1960; Wiles, 1961) If average costs are constant beyond

Q E , marginal costs, denoted LRMC, will equal average costs over this range For

profit maximization the firm will equate marginal cost with marginal revenue,

3

resulting in an output OQ M which can be sold for a price of P M The monopoly

profit is represented by the rectangular area ABCP M

We can now use this simple analysis to get a preliminary answer to somebasic questions of concern in antitrust First, in what sense, if any, is themonopoly represented in Figure 1.1 inefficient? We can also introduce therelated question of whether efficiency is the legitimate concern of antitrustpolicy Secondly, what is the character and role of the impediment to entry inthe market shown? Thirdly, what is the significance of the profits shown in thefigure? And fourthly, what relation is there between the monopoly illustratedand firms in an antitrust case in a position of market dominance?

In order to address the first question we need to focus on the relationshipbetween marginal cost and price at the equilibrium output It is clear from the

figure that at output OQ M there is a difference CB between demand price and

marginal cost Since this difference is fundamental to the economic case againstmonopoly we shall spend some time examining its full implications Any point

on the demand curve represents the valuation at the margin by consumers of theunit of output specified Thus, assuming the output is infinitely divisible,

Figure 1.1 Price and output under simple monopoly

consumers’ marginal valuation of the Q M unit of output is CQ M The marginal

resource cost of this unit, however, is only BQ M(in the absence of production

externalities) The difference, shown as CB in Figure 1.1, amounts to the

mark-up of price over marginal cost Over the entire output range Q M Q Cconsumerswould be prepared to pay a price (as indicated by the demand curve) greater thanthe resource cost of production, indicating that such output would generate aconsumer benefit In principle consumers would be prepared to pay the

monopolist a sum equal to the area ABCP Min exchange for an increase in output

to Q C,sold at a price OA The consumers’ net gain would then be the equivalent

of the triangle BCD The fact that the monopolist would restrict output below

Q Cimplies that resources are being misallocated: too few resources are devoted

to monopoly production in order that price can be maintained above marginalcost The monopolist’s pricing behaviour therefore leads to allocative ineffi-ciency At the heart of economists’ case against monopoly is this price–costdivergence and the resource misallocation that results

The concept of efficiency, however, is one that is often heard in discussions

of the relative merits of competition and monopoly and, unless the severaldifferent meanings are kept clear, confusion is likely to occur Thus althoughthe monopoly may be allocatively inefficient it produces at a technicallyefficient scale of operations By this we mean that the firm has built a plant of

a size large enough to take advantage of all available economies of scale, where

we include not only physical plant but also optimal organizational and marketingpractice By combining all factors in a way which embodies best availablepractice the firm is able to achieve minimum unit costs for producing and dis-

tributing its output Q M As shown in Figure 1.1, costs can be minimized for a

scale of operations Q E or greater For scales smaller than Q Eunit costs would

be higher, as shown by the LRAC curve The monopolist is thus technically

efficient but allocatively inefficient

Those familiar with the perfectly competitive model will know that in run equilibrium the industry will also be efficient in this sense All firms will

long-be producing on the minimum point of their long-run average cost curves which

in this model are assumed to be U-shaped, owing to the influence of onomies of scale beyond the optimum output On the assumptions currentlymade, therefore, the main difference between monopoly and competition isthat, in the former, prices are greater than marginal cost and this leads toallocative inefficiency Both forms of market organization, however, aretechnically efficient in the long run.1

disec-It is useful at this stage to introduce a further type of efficiency to which wewill refer in subsequent chapters A number of writers, stretching back as far

as Adam Smith, have noted that, whereas the representation of costs in Figure1.1 refers to the firm using the best available technology and organizationalmethods so as to obtain minimum costs for any level of output specified, when

firms are in a protected market, incentives to achieve minimum cost may beblunted In particular the amount of effort put into achieving the firms’objectives may be reduced at all levels of the organization so that it may operatewith a considerable amount of slack The term used for this kind of internal dis-organization is ‘X-inefficiency’ In terms of Figure 1.1, this would mean that,

instead of operating at a point B on the LRAC and LRMC curve, the firm would

be operating with a cost level above B for an output Q M However, if that was

the case its chosen output would be to the left of Q M, since in effect it would

be attempting to obtain its objectives from a higher perceived unit and marginalcost curve than that shown We have hesitated to use the term ‘profit maxi-mization’ in this context because if the firm allows its costs to rise in the wayindicated its objectives have clearly shifted and it may be pursuing ‘employeeand management satisfaction’, rather than profit maximization.2 For ourpurposes, however, it is sufficient to note that the stronger the protection for amonopoly the weaker are the incentives for X-efficiency If competitive forcescan be strengthened in a market then X-inefficiency will tend to disappear

It is appropriate at this point to introduce the question of whether efficiencyshould be the major goal of antitrust policy or whether broader or indeednarrower objectives would lead to more satisfactory results In Chapter 3 we willlook in detail at what are the stated objectives of antitrust policy in the EU, theUSA and the UK For the moment we consider the issue in more general terms.One approach is to argue that ultimately the consumer interest should beparamount in antitrust questions and that this can be best achieved by the pursuit

of greater efficiency through antitrust policy Bork certainly takes this view:

‘The whole task of antitrust can be summed up as the effort to improveallocative efficiency without impairing productive efficiency so greatly as toproduce either no gain or a net loss in consumer welfare’ (Bork, 1978, p 91).Thus, on this view, the reduction of monopoly by removing impediments toentry is likely to have the effect of improving allocative and X-efficiency If theincumbent firm continues to restrict output and pays insufficient attention tothe level of its costs it will lose market share and its profits will be eroded.Consumers will gain through lower prices The argument can be sustained eventhough the increased competition may ultimately lead to the demise of one ormore of the competitors In fact if the process of competition is effective weexpect over time some firms to disappear either through bankruptcy or merger,and their place to be taken by new organizations The speed with which theprocess works is one of the main concerns of Chapter 2

Although the ultimate objective to antitrust policy may be greater efficiencyand consumer welfare, by placing a greater emphasis on the competitive means

to that end some writers may unwittingly allow antitrust authorities to giveundue emphasis to the protection of competitors rather than to competition.Thus if the goal of antitrust is seen primarily in terms of maintaining competition

this may be interpreted by lawyers, judges or antitrust administrators asrequiring the continued presence of existing firms even though it might meandamaging the consumer interest because they are inefficient and are attempting

to use the antitrust machinery to preserve their position The pressures for thisoutcome should not be underestimated, especially where an existing marketstructure is dominated by one firm and where the complaint or action comesfrom a smaller rival Unless the antitrust body keeps efficiency and the consumerinterest firmly in mind it may serve the interests of weak competitors

We turn now to the second question mentioned above, the character and role

of impediments to entry in the monopolized market Some impediments orbarriers are quite straightforward and their effects clear-cut The governmentgrant of a monopoly to a single firm or individual, for example, means that noother firm can legally enter that market in the short or long run The Stuartkings found this a useful means of raising revenue and the governments of somedeveloping countries have granted exclusive import licences to individualsowed a favour

Other barriers are more complex and ambiguous The government grant ofexclusive rights to a new product or process in the form of a patent has thespecific purpose of encouraging and rewarding invention and innovation eventhough, for a time, a monopoly performance can be expected The optimumlevel of innovation and change can be thought of as dynamic efficiency There

is clearly a trade-off involved between static and dynamic efficiency Staticefficiency is impaired by the use of the legal monopoly (the patent) but thegrant of the patent may be the means by which new products are introduced orthe costs of existing ones reduced, so that over time there are positive benefits

to consumers Furthermore, even with a patent system the impediment to newentry is likely in many instances to be less complete than in the case of agovernment-granted monopoly Invention without patent infringement mayfrequently be possible, especially by firms already working along similaravenues of research

Control of crucial raw materials is also often cited as an importantimpediment to entry The most extreme case would be where known high-grademineral deposits are very limited and owned by a single firm Competitors mustthen either make do with inferior (more costly) materials or depend on themonopolist for their supplies In either case their costs will be higher than those

of the monopoly A similar effect may be felt through control of prime sitesfor distributing a product If the existing firm controls the best locations, otherfirms will have to be content with slightly inferior ones, which would againput them at a cost disadvantage Although such extreme cases do undoubtedlyexist and have been maintained sometimes for long periods, others have beenmuch less secure simply because the very scarcity and high price of the resource

provides strong incentives for finding new deposits or other methods ofproduction and distribution.

If we move a little further down the hierarchy of entry barriers the ambiguity,not to say controversy, increases In the markets for consumer products whereincumbent firms may have very large market shares they may also have built

up over a long period strong consumer loyalty and brand preference throughtheir advertising and marketing policies Such firms thus have a very valuableintangible asset which entrants may be unable to create for themselves in theshort term, even though they may have access to the same advertising andmarketing media as the incumbent Even in the long term the degree ofuncertainty surrounding the creation of a comparable intangible asset is likely

to be far greater than that associated, for example, with operating a productionline In this sense, therefore, the entrant is at a cost disadvantage compared withthe incumbent Whether this kind of cost disadvantage should rank with agovernment-granted monopoly in the analysis of entry barriers has, however,recently been the subject of considerable dispute A number of writers haveargued that the new entrant faces similar expenditures in creating consumerbrand loyalty as the incumbent firm (or firms) have incurred in the past Aslong as the entrant has access to the same facilities as the incumbent, then nolong-term impediment exists To sustain their argument they can rely on a well-known definition of entry barriers by Stigler: ‘a cost of production (at some orevery rate of output) which must be borne by a firm which seeks to enter anindustry but is not borne by firms already in the industry’ (Stigler, 1968, p 67)

If the entrant can create a similar intangible asset by replicating the past ditures of the incumbent, on this view, no entry barrier exists The difficultyarises, however, precisely because although similar costs may be incurred by

expen-an entrexpen-ant the market environment that it faces with expen-an established dominexpen-antfirm is not the same as when the latter was building its market share

Similar considerations surround the treatment of economies of scale If anentrant to a market where such economies are substantial can effectivelyreproduce the production facilities used by an incumbent firm, there is noimpediment to entry, according to the Stigler view Yet, if entry on a large andtherefore technically efficient scale were to occur in such a market, the increase

in output might be so great as to ensure that the market price would fall belowaverage costs Entry would not be viable and therefore is unlikely to take place,even though in principle a potential entrant could achieve the same cost level

as the incumbent

Clearly there are many sources of impediment to entry to particular marketsand often different types will occur simultaneously It is also evident that theirprecise impact on the performance of existing and entrant firms will vary con-siderably In Parts II and III of the book we will be very much concerned withthese effects For the moment we simply underline the significance of entry

barriers to the performance of the monopoly shown in Figure 1.1 The presence

of some impediment to entry ensures that the firm retains its monopoly notonly in the short term but also in the long term If the impediment were removedthe firm would still be able to charge a monopoly price in the short term butthe profits that resulted would attract new resources to the market and overtime the increased supplies that result would cause the price to fall to thecompetitive level

We thus come to the third question posed above: what is the significance

of profits shown in the simple monopoly case illustrated in Figure 1.1? As long

as the firm is protected by barriers to entry its policy of output restriction and

simple monopoly pricing will generate the excess profit shown as ABCP M By

‘excess’ we mean an amount greater than that required to retain the resources

in their present use On the usual convention that amount, the ‘normal’ return

or opportunity costs of capital, is included in the long-run average cost curve

In a market economy the key role of returns greater than ‘normal’ is to attractadditional resources The presence of entry barriers frustrates this mechanism

To the extent that additional resources are kept out of a market because ofsome barrier (of the kind mentioned above) which allows the incumbent firm

or firms to earn excess returns, the resource misallocation identified inFigure 1.1 will persist

This conclusion would appear to have important implications for antitrustpolicy Excessive profits identified amongst monopoly firms might give a goodindication of a poor allocative performance, and thus be of direct policy concern.The British Monopolies and Merger Commission (MMC) often used suchindicators to assist its investigations, but they have to be treated very cautiouslyand may be misunderstood or completely misleading.3In practice it may bevery difficult to distinguish short- from long-term influences and those thathave a benign effect (due to innovation) from those that are malign (due toentry deterrence) It may also be very difficult to obtain data that reflect evenapproximately the concepts used in the economic analysis of the problem.Furthermore, if monopoly firms are prone to X-inefficiency as we suggestabove, recorded profits may appear modest simply because internal slackaccounts for the rest On the other hand, a monopoly which has earned onlymodest returns may be constrained by the threat of entry Simply because afirm has a very large market share does not mean that it can sustain a monopolyprice and earn a monopoly return Only if its share is buttressed by barriers toentry will this be possible

So far we have deliberately used the notion of a monopolist as the sole seller

of a product for which there are no close substitutes, taken from the theory ofmonopoly We have thus been able to introduce a number of concepts thatwill be particularly useful in the succeeding chapters Since, however, a majortask of the book is to show how economic analysis can be used to analyse

antitrust problems we need now to address the question of the relationshipbetween the monopoly of economic theory and firms in antitrust cases inpositions of market dominance.

By ‘market dominance’ we mean the ability of a firm or group of firms sistently to hold price above long-run average costs without thereby losing somany sales that the price level is unsustainable We have deliberately used theterm ‘market dominance’ rather than the more frequent ‘market power’ in order

per-to emphasize the central role played in many markets by one or a few largefirms The terms can, however, be used interchangeably In the light of theforegoing discussion we can note the following points about this definition.First, a market may be dominated by more than one firm and they may eitheract together (through secret collusion, for example) or tacitly arrive at a pricesolution close to the monopoly level by acting on what they believe is theirbest strategy, given their anticipations of what the policy of the others will be.Thus, whereas our previous discussion proceeded on the basis of a single firmmonopoly, in practice we have to recognize that a frequent case will beoligopoly and the mutual interdependence that that implies

Secondly, the price–cost difference has to be persistent Temporarily highprices in relation to costs may be caused by a variety of factors, including short-run fluctuations in demand or input prices High short-term profits may resultbut they will not be due to market dominance An unforeseen increase indemand will generate windfall gains even in a highly competitive market Inantitrust cases, however, ‘persistence’ has to be properly interpreted Clearlywhether increases in demand cause prices to rise for one, two or five yearsdepends very much on the type of industry involved In some cases additionalsupplies from new entrants may be available within weeks or months of ademand increase They may come either from imports which were previouslyuneconomic because of transport costs or from new firms entering the market

In other cases technological complexities may mean that additional output fromentrants will only be available after several years Nevertheless the price increasecaused by the shift in demand will not be sustainable against new entry, and theexisting firms cannot be said to dominate the market in the sense defined.Thirdly, the definition refers to prices successfully held at a level greaterthan average costs Although not involving directly the concepts of perfectcompetition and barriers to entry, it is clear from our previous discussion thatthey have an important indirect bearing on the definition In a perfectlycompetitive market, prices will eventually fall to the level of marginal cost.Even though there is no suggestion that the real world is populated by perfectlycompetitive markets, nevertheless, in those where a reasonable approximation

is found, prices will in the long run be aligned with costs The corollary is that,where serious and prolonged divergences occur between price and cost, thecause can be traced back to some impediment to entry We may note in passing

that the definition can clearly encompass different degrees of dominanceimplying large or small divergences between price and cost Clearly, given thelimited resources that will always be available for antitrust enforcement, it isdesirable that serious cases should take precedence over minor cases, not leastbecause these may be expected to create the largest welfare losses forconsumers.

We have left until last perhaps the most difficult question of all arising fromthe definition of market dominance This concerns the apparently innocuousreference to ‘the market’ If a firm or firms are dominating ‘the market’ and ifthis can have important policy implications it is evident that we need to have

a clear idea of what constitutes the relevant market in a particular case In Figure1.1 we were able to avoid the practical problems by simply drawing adownward-sloping demand curve facing the monopolist, on the assumptionthat the product in question had no close substitutes The greater the gap in the

chain of substitutes, the smaller will be the price elasticity of demand, ceteris

paribus Hence a firm or group of firms controlling all of the supply of a product

for which no close substitute exists will have greater scope for raising priceabove cost

When we move, however, from theory to antitrust policy some assessment

of what actually constitutes the affected market has to be made As we shallsee in later chapters, large quantities of ink have been spilt and many hairs havebeen split in trying to catch the elusive concept Clearly the firms themselveswill argue in defence of their position that the correct market is very wide andthat consequently their dominance is relatively small Complainants and antitrustauthorities are likely to think otherwise and draw the boundaries more narrowly.One approach which has received much attention but which in practice may

be very difficult to employ involves the use of cross-elasticity of demand andelasticity of supply The cross-elasticity of demand between products X and Ymay be defined as the percentage change in the quantity of X demandedresulting from a small percentage change in the price of product Y In effect this

is merely a more formal way of identifying products which are close substitutesfor each other Where the cross-elasticity is high, that is where a small change

in the price of Y results in a relatively large change in the quantity of Xdemanded, the products can be regarded as part of the same market Consumerswill evidently switch from one product to another very readily

Although the concept may help to clarify thinking about the scope of markets,there are practical reasons why it is of limited policy use Estimates of cross-elasticities are not readily available or easily made An antitrust case does notprovide the ideal environment for making such estimates Furthermore it is notclear where the cut-off points should be, even if estimates were available In thecontext of an antitrust case, what constitutes a ‘relatively large’ cross-elasticity:

2, 5, 25 or higher? Additional adjustments would also have to be made to take

account of the fact that ‘close substitutability’ depends not only on physicalcharacteristics but on geographic location Identical commodities producedseveral hundred miles apart and costly to transport do not constitute part of thesame market.

Similar considerations apply to the elasticity of supply The concept is defined

as the percentage change in the quantity of X supplied in response to a smallpercentage change in the price of X Thus, if a small increase in the price of Xcauses a disproportionate increase in the supply of X, supply elasticity will behigh In other words existing producers face few production difficulties inincreasing their output but also other firms, using flexible production methods,can readily switch to producing X following a price increase A further source

of additional supplies may be imports If a small price change makes thedifference between profitable and unprofitable imports, clearly existingdomestic firms may have little room to manoeuvre on price; that is they willhave limited market dominance In principle therefore high supply elasticitywill have an important influence on the way the true ‘market’ is perceived.4The principle is of increasing importance in European Union (EU) cases wherethe continued removal of trade and other barriers between member states hashad the effect of ‘widening’ the market in the way indicated This is not to say,

of course, that statistical estimation of supply elasticities is actually carried out,because the same problems arise as in the case of cross-elasticities of demand.The underlying logic, however, does have a bearing on the way a particularmarket is viewed and consequently how much discretion the leading firm orfirms have

The joint emphasis in the elasticity approach to both demand and supply stitutabilities is therefore correct and avoids the mistake of taking too narrow

sub-a view of the msub-arket by simply sub-accepting sub-at fsub-ace vsub-alue the sub-alleged msub-arket shsub-are

of existing firms A seemingly different approach has been suggested by Areedaand Turner, who define a market as ‘a firm or group of firms which, if unified

by agreement or merger, would have market power in dealing with any group

of buyers’ (Areeda and Turner, 1978, p 347) The definition appears to godirectly to the heart of the problem by drawing the boundary of the marketaround those firms which, if acting together successfully, could raise prices It

is close to the Guidelines now used by the US antitrust authorities in mergercases which we discuss fully in Chapter 8 Some assessment has to be made bothabout the substitutability in demand of the outputs from the firms to be includedand about the possible switches in production that may take place in firmsostensibly supplying a different market In fact commercial history is strewnwith the remains of many restrictive agreements that collapsed precisely becausethe participants failed to take account of the increases in supply that wouldoccur if the price was raised Participants clearly have a very sharp incentive

to try to ensure that all suppliers are included in the agreement to make it

effective Yet they often fail As has often been pointed out, the managements

of firms have a far greater knowledge of their own industry than either lawyers

or antitrust authorities It is difficult therefore to envisage either doing any better

in their attempts to define a market by drawing a ring round all actual andpotential suppliers

There is thus a considerable gulf between the concepts defined in economicanalysis and their close approximation in an actual case This does not detractfrom the importance of the analysis which, as we shall see, is a powerful toolfor distinguishing the relevant from the misleading or false

II Market dominance: extensions to the preliminary analysis

The caution is reinforced when we move away from the simple, static analysis

to consider more complex cases At this stage we merely introduce three issueswhich will be of more detailed concern later on: significant scale economies,price discrimination and incentives for innovation

The first and most obvious point arising out of the case illustrated in Figure1.1 is that we implicitly assumed that costs were unchanged whether the marketwas supplied by one firm or by a larger group of firms If there are significanteconomies of large-scale organization, as seems evident from the experience ofsome industries, then monopoly may be the most efficient form of organization

In particular, if the cost-reducing economies are important enough, price may

be lower and output higher with monopoly than with a multi-firm competitivestructure These results can also be usefully illustrated in Figure 1.2 In the

figure, market demand is shown by AR and, if production is organized by a monopoly, MR is marginal revenue The monopolist’s long-run average and marginal revenue curves are shown by LRAC and LRMC, respectively We

have thus retained the assumption that under monopoly unit costs decline to a

minimum (at output Q N) and then remain constant over the relevant range The

profit-maximizing monopoly price is then P M and output Q M All of this is thesame as in Figure 1.1 However we now show the competitive supply curve as

S C In other words, if production is organized by a large number of small-scale

firms, their aggregated costs are embodied in the horizontal line S C.5In this

case the market-clearing price is P C and corresponding output Q C Price wouldtherefore be higher and output lower than under monopoly

What can we say about the comparative efficiency of the two contrastingcases? As far as technical efficiency is concerned there is no question that themonopoly is much more technically efficient than the competitive industry It

uses far fewer resources to produce output Q C or its preferred output Q M Theresulting lower price means that consumers who, as it were, enter the market

at prices between P C and P Mhave their demands satisfied by the monopolist,whereas they would have remained uncatered for by the competitive industry.6

As far as allocative efficiency is concerned the answer to the question is more

complex The fact that a greater output is sold at a lower price under monopolyimplies that consumers are better off than under competition, and yet it is clearfrom Figure 1.2 that the monopoly output is sold at a price greatly in excess of

marginal cost At output Q M marginal cost is FQ M , whereas price is EQ M, giving

a mark-up over cost of EF Thus despite the lower price the monopolist is

generating considerable allocative inefficiency Consumers are better off butthey could be made even better off if a means were found of making the

monopolist charge a price P R, equal to marginal cost Consumers would thenhave the full benefit of the scale economies, and the allocative inefficiencywould be eliminated We state the possibility without suggesting that there is

a means readily to hand which would produce these desirable results.Indeed, in a very real sense, the possible conflict between technical andallocative efficiency lies at the heart of many antitrust questions In addition tothe case shown in Figure 1.2, it frequently arises in slightly less extreme form

in large horizontal mergers where there may be a strong probability that themerger would increase market dominance but at the same time reduce coststhrough economies of reorganization.7Similarly, a fragmented market structure

may be allocatively efficient but provide an inhospitable environment forinnovation A concentrated structure may innovate more rapidly, even though

in static terms it is allocatively inefficient In each case the antitrust authoritymay have to decide where the balance of advantage to consumers really lies.Where a case involves two types of efficiency pulling in opposite directions,

an antitrust objective which simply seeks to ‘improve efficiency’ is clearlyambiguous The ambiguity is made worse if the third kind of efficiencyintroduced in Section I above is also considered Suppose, for example, themarket represented in Figure 1.2 is initially organized competitively but then

a consortium of UK, Australian and US entrepreneurs hold out the promise ofconsiderable resource savings by being allowed to create a monopoly in theindustry and so produce the results already described The promise may thus

be of an increased output and a lower price (with also an admittedly large profitfor the monopoly’s shareholders) Once created, however, the monopoly faces

no actual competition and, assuming it has created barriers to new entry(possibly a strong assumption), no potential competition either The eventualperformance of the firm may therefore not live up to expectations Lethargyand inertia may mean that achievable economies remain unrealized Actualcosts stay above those shown In other words it is possible (although difficult

ex ante to predict with confidence) that X-inefficiency may cause prices to rise

after the monopoly is created In Figure 1.2, if costs are reduced to only LRMClthen the implied profit-maximizing price will be above P Cand output below

Q C Consumers are worse off than before (they are paying a higher price) but,despite the X-inefficiency, costs for the diminished output have been reducedfollowing the creation of the monopoly As we shall see in Part II, the trade-offs involved in such cases can be exceedingly complex and have been thesubject of much controversy

So far we have assumed that the monopolist or monopolistic group charged

a single price for the product In such a case the contrast between competitiveand monopoly pricing is thrown into sharpest relief Simple pricing, however,

is probably less common than discriminatory pricing Anyone who has recentlyenquired, for example, about the price of an airline ticket between, say, Londonand Paris, Cologne and Madrid, or New York and Chicago is likely to havebeen bemused by the variety of different offers that were made, depending onsuch information as time of travel, date of booking, whether the trip involved

a weekend, whether they were flying on to another destination, how frequentlythey flew with the particular airline, and so on The information contained inthe answers to these questions in effect is used to slot customers into a particularcategory, depending on their (approximate) demand elasticity Instead of anairline offering a uniform price for a ticket between two destinations, differentcategories of customer are offered different prices where those differences bearlittle relationship to differences in the marginal cost of providing the service

Even in the case of different classes of ticket (‘first’ and ‘tourist’, for example)there is no suggestion that the difference in quality of service provided isaccurately reflected in different levels of cost Many other examples of pricediscrimination could be quoted and the subject is dealt with at greater length inChapter 5 For the present we shall rely on the definition offered by Stigler:price discrimination is the sale of two or more similar goods at prices which are

in different ratios to marginal cost (Stigler, 1968, p 209) The discriminationmay occur between different quantities to the same group of consumers (themore purchased per time period, the lower the price) or between different groups

of consumers (businessmen and tourists) or between different markets insulatedfrom each other (the new equipment and the replacement market; the home andthe foreign market) A firm with a degree of market dominance may be able todiscriminate on price In the absence of such dominance the attempt will failbecause of competition

We argued above that many antitrust problems involve a possible conflictbetween technical and allocative efficiency In particular, market dominancewas likely to produce divergences between price and marginal cost and hencecause allocative inefficiency Price discrimination may make such conclusionsmuch more ambiguous and difficult to interpret The point can be illustratedwith reference to a modified Figure 1.1, shown as Figure 1.3 The monopolist

has up to now been assumed to charge a uniform, profit-maximizing price P M.However, suppose the firm has been able to gather sufficient information aboutthe precise demand requirements of its customers to charge different prices fordifferent quantities purchased Consumers only qualify for the lower pricesonce they have purchased the initial quantity, per period Thus consumers are

offered the initial quantity at P1and further quantities at P2, P M , P3and so on.The motive for such a complex price structure is clearly to increase profits Forthe structure shown, profits amount to the irregular step-shaped area in thefigure They are thus considerably greater than the profits generated by simple

pricing and given by ABCP M More to the point in the present context is theeffect on allocative efficiency Output is considerably increased and a large

amount is sold below the simple monopoly price P M The discriminatingmonopolist will find it worthwhile to increase output up to the point where thelast block of output is sold at marginal cost In the limit, therefore, output will

be at the allocatively efficient level

For the purposes of illustration at this stage we have deliberately chosen what

is probably the easiest case, price discrimination by quantity It brings out,however, the important point that, once we move away from simple pricing,conclusions on the allocative inefficiency resulting from market dominancemay be much less clear-cut It also brings into greater prominence the question

of the income distribution effects of market dominance It is clear from Figure1.3 that improved allocative efficiency has also created a much greater profit

for the firm and one which is only possible because of its position of marketdominance If part of such profits could be used in a number of ways to preservedominance and stifle the emergence of competition it may be thought that theprice of the improved allocative efficiency is too high and that price discrimi-nation itself is a proper subject for antitrust scrutiny These issues are taken up

in more detail in Chapters 5 and 6

A third qualification to the simple analysis presented in Section I aboveconcerns the relationship between market structure and innovation Ever sinceSchumpeter (1965) argued in his classic work on capitalism, socialism anddemocracy that the really important aspect of competition was innovation(broadly interpreted) that struck at the cost levels or demands of existing firmsand, further, that adequate incentives for the inherently risky process ofinnovation were only available to firms with some degree of market dominance,there has been enormous controversy, both over the plausibility of his analysisand over the correct role of antitrust policy in such questions

The efficiency concepts that we have introduced are also useful in nating this issue One conclusion was that a firm or group with marketdominance would create resource misallocation through operating at ineffi-

illumi-Figure 1.3 Price discrimination and allocative efficiency

ciently low levels Under competitive conditions this misallocation woulddisappear That analysis, however, was static in that it referred to the allocation

of a given bundle of resources at a particular time The question that reallyconcerned Schumpeter, however, was the growth of resources through timeand he saw innovation, in products, processes and methods of organization, asthe means to achieve economic growth If it is true that markets dominated byone or a few firms will generate a faster rate of growth than those with afragmented and highly competitive structure, questions of static, allocativeinefficiency may become relatively unimportant Initially the allocative inef-ficiency of an economy where market dominance prevails may cause it to have

a lower national product than if there was universal perfect competition.However, if Schumpeter was correct, it would only be a matter of time beforethe concentrated economy was enjoying a higher national product ‘A system– any system, economic or other – that at every given point of time fully utilisesits possibilities to the best advantage may yet in the long run be inferior to asystem that does so at no given point in time, because the latter’s failure to do

so may be a condition for the level or speed of long-run performance’(Schumpeter, 1965, p 83)

For Schumpeter the incentive to undertake the greater risks and challengesthat innovation involved was the prospect of future monopoly profits, but he alsosaw the means of financing such activities as deriving from the greater thancompetitive returns that only firms with some prior market dominance can earn.Several writers developed his ideas in a number of ways Galbraith (1963), forexample, focused more particularly on oligopoly rather than monopoly andargued that the muted form that competition took in oligopoly allowed firms togenerate sufficient profits to finance the enormous research and developmentcosts that significant innovations now required in many industries Furthermore,once innovations were made, the same conventions of oligopolistic competitionwould ensure that the innovating firm gained most of the resulting profits, ratherthan have them dissipated rapidly amongst a large number of rivals.8Comparedwith a fragmented industry, therefore, oligopoly was better equipped and suitedfor innovation Similarly it could also be argued that the incentive to innovatewas greater under oligopoly than monopoly, precisely because no single firm

in an oligopoly could allow a rival to gain a significant advantage, whereas amonopolist was in sole command of the market and had no immediatecompetitors Just as a monopolist may allow itself to become X-inefficient, soalso it may be loath to take on the special problems associated with innovationand change

However, this may be an unduly restrictive view of monopoly It may apply

to a few cases where barriers to entry are especially high and where technology

is relatively static In many other cases firms currently in sole command of amarket and earning high profits may nevertheless feel impelled to undertake

extensive research and development simply because it is impossible to predictwhere a path-breaking innovation may next emerge It may come from anapparently unrelated technology and involve firms not hitherto regarded ascompetitors Furthermore, if the current monopolist is successful in, forexample, initiating a cost-reducing production method, it will retain all of theadditional profit rather than having to share it with oligopolistic rivals Despitepatent protection, the nature of oligopolistic rivalry may be such that aninnovation can often be imitated (Scherer and Ross, 1990) On this view,therefore, the incentive to innovate may be greater for a monopolist than for oli-gopolists Ultimately the question of which market structure is most conducive

to technical change or ‘dynamic efficiency’ is an empirical one which may raise

a number of difficult antitrust problems For example, will a merger betweenleading firms in what may already be a concentrated oligopolistic marketstructure intensify or weaken incentives for innovation? How far shouldindependent firms be allowed to collaborate on expensive research anddevelopment before this affects competition between them? Are joint ventures

on experimental products and processes between competing firms desirable onpolicy grounds or are they likely to be used as a means of collusion?

Although the factors mentioned in this section are by no means exhaustive,they should be sufficient to warn us that the simple analysis of marketdominance, despite highlighting some very important features, must be treatedwith considerable caution when trying to apply it to policy questions Somemarkets may only be able to achieve technical efficiency if they contain fewfirms Very often firms having some control over their price will be able tosegment the market and use price discriminations to increase their profits.Market structures which generate a faster rate of growth through innovationmay have to be heavily concentrated and create static inefficiencies All of thesecircumstances raise complex issues for antitrust policy because they involve atrade-off between different types of efficiency, usually, although not exclusively,between technical and dynamic efficiency on the one hand, and allocative andX-efficiency, on the other One way of looking at antitrust policy, therefore, is

as an attempt to reconcile these different types of efficiency

III The costs of market dominance

We have so far argued that market dominance raises important issues for publicpolicy, but the sceptical reader may quite rightly enquire how great is theirpractical importance The problems may appear significant from a theoreticalpoint of view, but if their practical importance is negligible the scarce resourcesdevoted to the application of economic policies could be more usefullyemployed elsewhere

In this section, therefore, we discuss the probable size of the losses that mayresult from market dominance It is convenient to start with a slightly modified

Figure 1.1 In Figure 1.4 we assume that costs remain constant over the entire

output range The line AF can be interpreted initially as the supply curve of a

competitive industry (with input prices invariable with industry output) Given

that market demand is AR, then under competitive conditions the price is OA and a quantity OQ Cis sold Consumer satisfaction is then maximized in thismarket since marginal valuation, as reflected by the demand curve, is justequated to marginal resource cost Another way of looking at this is to note that

at this price and output configuration consumer surplus, ADG, is maximized Consumers place a value on the amount of the product purchased of ADG

greater than the resource costs involved in production with the demand and costconditions shown Any higher or lower price would reduce consumer surplus

If the industry shown were to be taken over by a monopoly with no change in

cost conditions occurring, price would be increased to OP M A close examination

of the effects of this change on consumer surplus is of direct relevance to thequestion of how significant the effects of market dominance may be Under the

monopoly the area P M CBA becomes the monopoly profit and amounts to a direct

transfer from consumers Although this transfer may raise questions aboutincome distribution and the relative merits of consumers as opposed to share-

Figure 1.4 Welfare losses from market dominance

holders’ interests and the uses to which it may be put, it is usually not treated

as a matter of antitrust concern The triangle GCP Mremains as a somewhat

depleted consumer surplus This leaves the triangle CDB On present

assumptions, when the competitive industry is taken over by the monopolist,

the value represented by the triangle CDB disappears altogether It is not

transferred from one group to another; in effect the monopoly destroys thatamount of value

On the basis of this analysis, a number of researchers have therefore tried toquantify the extent of the destruction, or welfare loss By adding up all of thewelfare losses in each of the markets where dominance occurs, an estimate forthe loss to a sector or a whole economy can be made Merely to state thisrequirement should give some indication of the estimation difficulties involved.Consequently the figures referred to below should be taken as rough approxi-mations rather than precise estimates Different procedures have been used torender operational the theoretical result that output restriction leads to a welfareloss If the full monopoly price is charged, for example, then some elementary

geometry makes it clear that the triangle CDB is equal to half the area P M CBA;

that is, the welfare loss is equal to half the size of the monopoly profit Estimates

of monopoly profits can then be used to calculate the likely welfare loss On theother hand, if a less extreme view of market dominance is taken, the welfare losswill be considerably less For example in Figure 1.4 the expectation that the

full monopoly price P Mwould attract new entry and thus erode the market share

of the incumbent firm (or firms) may constrain price below that level to, let us

say, P L In this case the welfare loss, represented by HDK, would be much less.

Indeed unless the incumbent firm is protected by very strong and substantial

entry barriers the price is unlikely to be as high as P M If the existing market is

in the hands of a few firms with large shares then most oligopoly theory predictsthat, in this case also, price may settle below the monopoly level

Clearly, depending on which view is taken, the empirical estimates of actualwelfare losses imposed by market dominance are likely to differ At one levelthis distinction lies at the heart of the different estimates made, for example, byHarberger (1954) on the one hand, and Cowling and Mueller (1978) on theother The former favoured the more cautious approach of accepting thatexisting monopolists would feel some constraints on their pricing behaviour Forthis assumption (as well as others which we discuss below) he was taken totask by Cowling and Mueller, who argued that the assumption inevitably meantthat estimated welfare losses would be small Harberger was, however, in verygood company because, as Scherer (1987) has pointed out, Marshall in his

Principles of Economics, first published in 1890, and characteristically in a

footnote, showed that welfare losses from this source would be small, althoughthis did not deter him from being highly critical of monopoly

In their paper, Cowling and Mueller gave estimates using both their own andHarberger’s methodology In the present context it is convenient to contrasttwo sets of their estimates rather than referring to the original set presented byHarberger.9Using samples of the largest enterprises in the USA and UK,Cowling and Mueller estimated that the welfare losses produced if such firmswere constrained to price below the monopoly level (as Harberger maintained)amounted to less than one-half of 1 per cent of the product generated in thecorporate sector These were very much in line with Harberger’s originalestimates and, even allowing for wide margins of error, are hardly worthbothering about However, if it is assumed that the largest firms can charge thefull monopoly price (as Cowling and Mueller suggest), the estimates rise toapproximately 4 per cent of corporate sector product This is not insignificant,but if these were the total losses, potentially removable by an active antitrustpolicy, many observers might feel that it should be given a rather low priority,especially since the costs of administering, monitoring and enforcing such apolicy have not been included in the calculations.

Cowling and Mueller’s main contribution, however, was to focus on themuch wider implications of the effects of market dominance, and to give sometentative estimates of their extent The figures just mentioned represent thewelfare losses due to allocative inefficiency caused by the power to raise priceabove marginal cost Their additional estimates suggest that losses from a relatedbut distinct source may be more substantial It was made clear in Section Iabove that firms’ ability to maintain prices above costs rested on the presence

of barriers that impeded competition From a firm’s point of view, tures that help to create and maintain such barriers will thus enhance their prof-itability However, expenditures of this kind amount to investment in marketdominance and therefore should be regarded as creating a welfare loss For thefirms concerned, the expenditure will simply be recorded as part of their costsand affect their pricing and output decisions In practice it may be difficult for

expendi-an outside observer to distinguish legitimate production expendi-and distribution costsfrom those designed to maintain or increase market dominance In their study,Cowling and Mueller used the controversial assumption that all advertisingexpenditures had this ultimate purpose and should therefore be treated as part

of the social costs of monopoly Adding these losses to those directly due toallocative inefficiency raises the losses of gross corporate product to approxi-mately 12 per cent in the USA and 6 per cent in the UK (the differences can belargely attributed to the much greater reliance on advertising in the USA).10

The activity referred to in the previous paragraph is usually termed seeking’ A rent is the amount earned by a factor of production greater thanthat required to keep it in its present use Profits due to market dominance fallclearly into this category If the working of an economy allows the emergence

‘rent-of such rents, many individuals will find it pr‘rent-ofitable to expend resources for

this purpose Some advertising may have this objective but so also may ditures on lobbying, public relations, political contributions and even bribery.

expen-To the extent that the Cowling and Mueller estimates effectively draw attention

to the potential waste involved in these activities they have served their purpose.However, the argument can be taken one stage further Firms will compete inorder to obtain such rents but, as in most competitions, the majority will lose

Unlike the caucus race in Alice in Wonderland, everyone will not win prizes.

The resources used by the losers in their unsuccessful attempt to gain rentsshould also be treated as one of the welfare losses indirectly resulting from thepresence of market dominance in an economy For fairly obvious reasons it isextremely difficult to measure empirically the size of such losses and there isclearly scope for a great deal of disagreement over where to draw the linebetween expenditures which are a normal part of the competitive process (which

of course can involve the costs of bankruptcy) and those which are rent-seekingand wasteful The tentative estimates given by Cowling and Mueller suggest thatsuch expenditures may add a further 1–2 per cent to the social costs of marketpower The estimates suggest therefore that the total effects of market powerare likely to be considerable

The estimates referred to so far have focused explicitly on the effect of marketdominance on allocative efficiency and rent seeking The potential loss due toX-inefficiency is also considerable Referring again to Figure 1.4, suppose that

the incumbent firm was confident enough of its position while charging P Mthat

it became complacent and internally slack so that actual costs for its output rose

above AF Precisely because the X-inefficiency extends over the entire output

range rather than simply to the output contraction that occurs with marketdominance, the losses it generates are likely to be much greater than those due

to allocative inefficiency The central question is then how large are ciency losses? Unfortunately here we have far less to go on than the estimates

X-ineffi-so far discussed The ‘evidence’, such as it is, tends to be piecemeal, qualitativeand indirect Thus, for example, observers point to the rapid increases in internalefficiency that occurred in some UK industries in the 1960s when many cartelswere abandoned in the wake of early decisions under the Restrictive Practices

Act (see, for example, Swann et al., 1974), which led to intensified competition.

In a study of the US electricity power industry it was found that costs were nificantly lower in cities where the municipality-owned plant competed with aprivate installation than where there was a local municipal monopoly It hasalso been observed that large firms are able to reduce, sometimes drastically,their central office staff and expenditures when profitability has been reducedthrough competition (Scherer, 1987) A particularly interesting case involvedthe productivity of management consultants After the main recommendations

sig-of the consultants were implemented in a sample sig-of firms for which resultscould be quantified it was found that productivity increased by more than 50

per cent on average, and this represented an average return of more than 200per cent over the costs of the service (Johnston, 1963) Although there was not

a direct link in this case with the degree of competition, the results do suggestthat firms frequently run with considerable slack This conclusion also appears

to apply to British Airways, which, in the period prior to privatization in 1987when domestic competition was also increasing, was able to reduce itsworkforce and subsequently to increase its profitability (Kay and Bishop, 1988;McGowan and Trengove, 1986) Thus although the evidence is fragmentarymany observers are convinced that X-inefficiency is both quantitativelysignificant and directly related to the intensity of competition It is thereforelikely to figure prominently in any discussion of existing market dominance orits creation through merger.11

Ideally we would like to be able to say how much welfare gain there can be

by ensuring that the ‘correct’ degree of competition is maintained We wouldthen know the importance of antitrust policy for innovative efficiency Unfor-tunately the evidence on this issue is also imprecise and far from clear-cut Forexample, it is not clear from the empirical evidence which type of marketstructure is most conducive to rapid innovation Industries vary considerably inthe opportunities for innovation that present themselves For long periods someindustries may experience practically no alteration to their production methodsand product range but then go through a period of dramatic change Someimportant innovations have taken years to develop to a commercially viablestage and have involved vast expenditures, while others, equally significant,have been developed by a single person at modest cost and been quickly diffusedthroughout the industry.12After a detailed review of the empirical work, Schererand Ross (1990) arrived at the following, balanced conclusion: ‘What is neededfor rapid technical progress is a subtle blend of competition and monopoly,with more emphasis in general on the former than the latter, and with the role

of monopolistic elements diminishing when rich technological opportunitiesexist’ (p 660) We are not in a position to say by how much market dominancemay retard innovation and thus affect economic welfare, although theindications are that actions which enhance dominance or facilitate collusionwill tend to worsen this aspect of efficiency

IV Conclusion

The essence of market dominance is the ability to control the market in such away that prices can persistently be raised above costs, leading to excessiveprofits Where this occurs the market will fail to allocate resources efficiently.The task of antitrust policy can be viewed as an attempt to remove the source

of that failure or to prevent its emergence Since barriers to entry play a keyrole in impeding the flow of resources to their most highly valued uses weshould expect antitrust authorities to be closely concerned with their analysis

and control They are likely to be especially sensitive to those actions bydominant firms which seem deliberately designed to reinforce impediments tonew competition.

We saw in Section II, however, that the seemingly straightforward concept

of market dominance may be very difficult to identify and interpret in practice,for a variety of reasons Prominent amongst these was, first, the likelihood inmany cases that large relative size is required for technical efficiency Supplyfrom a few firms rather than many could therefore be made available at lowerprices, even though they were consistently above costs Secondly, firms in aposition of dominance would rarely charge a single price, preferring a morecomplex system of discriminatory prices which increased profitability whilesimultaneously making some supplies available at prices close to marginalcosts Thirdly, if the static inefficiencies caused by market dominance pale intoinsignificance beside the improvements that come from a faster rate of growthand if such growth is a function of a degree of market dominance, the antitrustauthorities have the formidable task of deciding how much dominance to allow.Frequently antitrust issues may boil down to a trade-off between technicaland X-efficiency, on the one hand, and allocative and innovative efficiency,

on the other We could determine analytically that X-inefficiency was liable togenerate considerably greater welfare losses than allocative inefficiency

Attempting to estimate empirically the size of the losses from marketdominance is hazardous and the figures should be treated with great caution.However, the general conclusion from a number of estimates is that likelywelfare losses from X-inefficiency and rent seeking are considerable andcertainly larger than those arising from allocative inefficiency Direct evidence

on the size of possible losses from an inferior innovative performance due to the

‘wrong’ level of market dominance is generally lacking Indirect and piecemealevidence, however, suggests that potentially the losses are considerable

In short, the welfare losses from market dominance that antitrust policy tries

to correct are likely to be substantial Economies that are successful in removingthe worst effects of market dominance or preventing its emergence are likely

to register substantial gains This does not mean, of course, that all currentantitrust action is correct and effective In some cases it may have made mattersworse or been unable to eliminate the abuses of dominance

Notes

1 We are ignoring for the moment the case where demand is relatively small in relation to available scale economies For example, in Figure 1.1, if demand shifted sharply and

permanently to the left so that it cut the LRAC curve to the left of E, the firm could still make

positive profits even though it was operating at a technically inefficient scale In this case, the firm could be classed as a natural monopoly; see Stigler (1968, ch 6).

2 Although the concept of X-inefficiency is now widely used and seems to be backed by empirical evidence (see Section IV below) there is still considerable controversy about its analytical usefulness (c.f Leibenstein, 1978, with Stigler, 1976).

3 For an early airing of the controversy, see Rowley (1969) and Sutherland (1971) More recent and opposing views appear in Fisher and McGowan (1983) and Kay (1987).

4 This line of reasoning has led two distinguished US authorities to argue that where appropriate all sales of ‘distant’ producers including foreign firms should be included in the calculation

of market shares in antitrust cases See Landes and Posner (1981) A full discussion is reserved for Chapter 4 below.

5 For simplicity we are assuming that the competitive industry can supply under conditions of constant costs If input prices rose (fell) as the industry expanded the supply curve would have a positive (negative) slope.

6 In other words, a changeover in the organization of production from a competitive structure

to monopoly creates an increase in consumer surplus of P C AEP M.

7 See Chapter 8 below.

8 This part of Galbraith’s analysis does not really square with the more apocalyptic peterian vision of the innovator sweeping away all competition in a gale of creative destruction.

Schum-9 The interested reader can find Harberger’s estimates in his 1954 paper and may also wish to consult subsequent efforts and commentary, for example in Bergson (1973), Kamerschen (1966), Schwartzman (1960) and Littlechild (1981) For a critical summary, see Scherer and Ross (1990, ch 17).

10 The figures quoted are based on continued use of the assumption that full monopoly prices can be charged Using the more conservative, Harberger, assumption produces much lower estimates; see Cowling and Mueller, tables 2 and 4.

11 Further examples are given in Frantz (1988).

12 For contrasting examples, see Jewkes et al (1969).

2 Market dominance in practice:

current perceptions and trends

I Introduction

The application of any policy to a set of economic problems is clearly notcostless Organizing and carrying out the policy uses scarce resources but also,since policy tools are by no means perfect, mistakes will be made which withhindsight show that a better economic performance would have resulted had

no policy intervention occurred Antitrust policy is no different in this respectfrom other forms of economic policy Those in charge of antitrust policytherefore have to be confident that the benefits of enforcement will outweighthe costs, not only those incurred by the firms directly involved but also thecosts of applying the policy itself

The point may appear obvious, but behind it lies a very important issue whichhas been at the centre of antitrust discussions for at least the past decade Itconcerns the speed with which unregulated markets are self-correcting, even inthe face of market dominance If those inefficiencies identified in the previouschapter are fairly quickly eroded by the forces of competition it may be better

to withhold any policy intervention; the market may do the job more efficiently.Even if the record of a dominant firm appears to show that it has consistentlyand over a long period abused its position, it does not necessarily follow that

in the absence of antitrust action it will be able to continue An innovation from

a totally unexpected source may very rapidly undermine its market share

In the analysis of antitrust questions, therefore, great emphasis needs to beplaced on the sources of market dominance Are they likely to be short-livedand therefore require no direct policy intervention or are they so firmlyentrenched that they are likely to be weakened only in the long run? In the lattercase, knowledge of the foundation on which market dominance depends shouldact as a guide to the most effective antitrust remedy This is what Easterbrookseems to have in mind when he says that ‘the central purpose of antitrust is tospeed up the arrival of the long-run’ (Easterbrook, 1984, p 2) Section II ofthis chapter contains a more detailed discussion of the sources of marketdominance and their susceptibility to antitrust remedies

Section III then reviews some evidence on the pattern of market dominance

in the UK and the USA, but focuses particularly on an important shift ofemphasis that has occurred in the last 20 years or so Until well into the 1970sthe empirical analysis of market structures as a screening device for possibleantitrust action was largely based on static measures, such as the combined

27

market share of the largest four or five sellers Often the measurement appliedonly to sales of domestically produced output and took no account of interna-tional trade More recently there has been much greater concern for thedynamics of market structure and also foreign competition Thus for the correctanalysis of market dominance the important points may not be that one firmhas a current market share of, say, 60 per cent, but whether or not the rate ofchange of that share has been rapid and which firm or firms are providing themost immediate competition The evidence on such issues is far less compre-hensive than for the simple static measures but there is now broad agreementthat it is far more important.

II The foundations of market dominance

In their authoritative review article, Hay and Vickers (1987) distinguish betweenthe acquisition, exercise and maintenance of market dominance They recognizethat in many ways the three are interrelated, but nevertheless find it convenient

to deal with them separately For our purposes it is useful to focus in this chapter

on the acquisition of market dominance while reserving for later sections ofthe book a more detailed treatment of its exercise and maintenance Althoughthey are not intended as an exhaustive list, Hay and Vickers mention five factorsthat can lead to the acquisition of dominance: government grant, ‘skill, foresightand industry’, explicit and implicit collusion, merger and predatory behaviour

We discuss each of these in turn

Probably the strongest foundation of market dominance comes from thegovernment Since sovereign power resides with the government, firms thatare able to harness that power for their own ends are likely to be in animpregnable position In important cases in the UK, firms had such powerforced upon them Thus the nationalization of industries such as gas, electricity,telecommunications, coal-mining, railways and civil aviation created unifiedmonopolistic enterprises which were protected by statute from direct entry totheir industries and from the reach of antitrust policy.1At the same time,however, they were constrained not only in their pricing, product and investmentpolicies but also in their growth: they could not diversify or (for the most part)integrate vertically In addition, of course, the power and transport industriesmet intense competition in some of their markets from non-nationalizedenterprises (for example, the oil and road haulage industries) Assessment oftheir subsequent performance has produced varied results with, if anything, anemphasis on the probable X-inefficiency that resulted from their dominancerather than excessive profits.2

More directly relevant to the present discussion is the recent policy ofprivatizing or returning to the private sector most of these enterprises and, inparticular, the conditions under which the policy was carried out Thus in thecase of British Telecom, which was privatized in 1984, initially a licence to

provide fixed link services was granted to only one other firm (Mercury munications) with the explicit understanding that no other entrant would beallowed for at least five years The position of British Gas was, if anything,even stronger Similarly the privatization of the electricity industry was in theend only able to proceed after the two generation companies had beeneffectively granted an exemption from any competition for 80 per cent of theirsales for eight years This last example illustrates a related point concerningthe grant of positions of dominance by governments In the USA the so-called

Com-‘utility industries’ that we are discussing, although for the most part retained

in private hands, have been regulated by special agencies for many years The

UK has been busily setting up rather similar agencies to regulate its newlyprivatized industries What was increasingly recognized about the US system,however, was that the regulators who were supposed to ensure that positions

of dominance were not abused were frequently ‘captured’ by the firms theywere supposed to regulate As a result they tended to operate in favour of thefirms’ interests rather than those of the consumer The continued evolution ofthe regulatory process in the UK plus technical changes have probably meantthat the tendency has been minimized Thus, for example, deregulation in theenergy sector has led not only to a complete restructuring of the once unifiedBritish Gas (it has split into two distinct companies) but to a replacement ofthe original and separate gas (OFGAS) and electricity (OFFER) regulators by

a unified regulator OFGEM (Office of Gas and Electricity Markets) Intelecommunications, persistent market growth and technical change havecontinued to erode BT’s market share, although it remains by far the largestsupplier in most markets

Nationalization and privatization probably exemplify the difficultiesassociated with state-granted dominance in its most acute form It would bewrong, however, to conclude that they were the only kinds Wherever the stateuses its authority to exclude potential competitors it is likely to enhance marketdominance It may take the form of exclusive licences, tariffs, quotas or evenoutright bans For example, although the European Union (EU) is often seen(by Europeans) as a means of intensifying competition and eroding existingpositions of dominance, those outside may have a different perception Quiteapart from the common external tariff which obviously inhibits imports intothe EU, it has also been claimed that pressure by member firms has causedthe anti-dumping procedures to be used in a highly restrictive and anti-competitive fashion.3

The cases of state-granted privileges mentioned so far have generally beentreated with some scepticism by economists There may be cases where theycan be justified on a broader basis than simply profiting the industry concerned,but they are likely to be comparatively rare However, a much more positivecase can be made for the next category of privilege to be considered, namely