The consumer guide to competition - International Development

The consumerguide to competition:A practical handbook

March 2003

Com

petit

ion

and

com

petit

ion

polic

y

Con

clus

ions

Con

sum

ers

and

com

petit

ion

Con

sum

ers

and

regu

latio

n

The consumerguide to competition:A practical handbook

1

Phil Evans

Acknowledgements

The Consumer Guide to CompetitionHandbook was written by Phil Evans atConsumers Association, UK and edited byLezak Shallat from Consumers InternationalRegional Office in Santiago, Chile. This workwas undertaken within the ConsumerMovement and Competition Policy Post Dohaprogramme, which is funded by the UKDepartment For International Development(DFID). However, in addition to gratefullyacknowledging the support of DFID, this workbuilt on existing activities funded by the FordFoundation, the ministry of foreign affairs ofthe Dutch Government, the European Union,the International Development and ResearchCentre (IDRC), and Oxfam.

Designed and produced by Steve Paveley

© Consumers InternationalMarch 2003

ISBN 1 902391 42 X

2

Acknowledgements

3

Contents

Introduction 6Assessing assumptions 6

Section 1: Competition and competition policy 9

Part I: What is competition? 10

Part II: What competition does 12Direct gains 12

Lower prices 12More choice 14New entrants 14Better service 14

Indirect gains 15Efficiency and productivity gains 15What about jobs? 17

Part III: Impact of Competition Law and Policy 19

Section 2: Consumers and competition 23

Part I: Markets and consumer behaviour 24Classical model of consumer behaviour 24

Utility maximisation 25Stable preferences 26Optimal information 26

Undermining the model of rational consumers 28When the rational model does not apply 29

Part II: Defining the market 31Geographic market definition 31

How it is currently done 31The end of geography? 32Van Thunen Circles 32Internet market definition 32Matrix of key questions 33

Contents

The consumer guide to competition: A practical handbook

Product market definition 33How it is currently done 33Substitutability 33

What’s needed: Time in substitutability assessments 34Working time into assessments 35

Demand-side complementarities 36Supply side substitution 37

Part III: Calculating market shares 38How it is currently done 38Doing the sums 38Step down differences between market shares 40Temporal aspects to concentration 41What’s needed: Impact of external trade 41Problems with indices 42

Part IV: Assessing vertically integrated markets 43How it is currently done 43Two-stage process 44Tying 45

Part V: Mitigating factors in assessments of market power 46How it is currently done 46Countervailing power 47Technological innovation 48

Part VI: Barriers to entry, establishment and exit 49How it is currently done 49Barriers to entry 49Structural barriers and market entry 52Barriers to establishment 52

Part VII: Strategic behaviour 52How it is currently done 52Bertrand or cournot? 53Predatory pricing 54Fighting brands, fighting ships and reputation 55The chain store paradox and signalling 55

Part VIII: Aiming for workable competition 56

Section 3: Consumers and regulation 57

Part I: Regulation: What it is, why it has grown 58What exactly is regulation? 58Why regulation has grown 59

Part II: Regulation and market failure 61Why regulation should happen 61

4

Contents

Part III: Why regulation happens 63Why regulation does happen 63Theory of regulation: Stigler-Peltzman 63Role of interest groups 64Watching out for the bootleggers 66

Part IV: Government failure, regulators and regulation 67Government failure and the role of regulators 67Identifying government failure 68Comparing market with non-market failure 70

Part V: Getting regulation to work for consumers 72When will regulation be successful? 72Attempts at reforming regulation 74

Section 4: Conclusions 77

Targets for regulation 80The process for regulating in normal markets 80Who can best deal with the spillover effects? 81How will improved efficiency and redistribution affect rent seeking? 81

Appendices 83

Appendix I: Behaviour, structure, strategy/conduct, performance, policy 84

Appendix II: Hints on decision-making and the analytical approach 85

Appendix III: Possible errors in market assessments 86

Appendix IV : Factoring in consumer behaviour 87

Appendix V: Structure-conduct-performance paradigm 90

Endnotes 91

What is Consumers International? 95

5

“Chaos, illumined by flashes of lightning”– Oscar Wilde on Robert Browning’s style 1

The analysis of markets is not an exactscience but a combination of evidence-based research and analytical interpretation.This handbook does not aim to provide anexact roadmap of ‘how-to’ proportions butto describe the parameters of thinkinggenerally contained in market structureanalyses and competition investigations.More importantly, this handbook aims toindicate the range of issues that competitionanalysis can cover and point out potentialproblem areas for those of us involved inrepresenting the consumer interest incompetition cases.

In addition to providing a guide to competitionpolicy investigations, this handbook alsosuggests improvements to the existingmethodologies in the area of consumerbehaviour, plus new approaches in geographicmarket definition and substitutability.

Incorporation of consumer behaviour morefirmly into the assessment of competition cases focuses on developing a matrix ofindicators to assess the degree to which theimplicit model of consumer behaviour willhold true. The key problem here is the scarcityof hard and fast indicators of consumerbehaviour other than ones derived from actual purchasing behaviour. As a result, theindicators we propose are more aboutidentifying barriers to consumer rationalitythan indicating the degree of that irrationalityitself. The advantage of this approach is that itallows the pinpointing of specific barriers torationality in individual markets and gives

6

structure to assessments both of competitionproblems and the likely success of regulatorysolutions.

The additions we recommend in the definitionof relevant geographic markets are also relatedto the ways consumers behave within normallydefined geographic markets. Establishedmethodologies generally use proxy indicatorsand measures of substitutability in the marketas guides to defining the market. We are notproposing to scrap those indicators but tocomplement them with more consumer-focusedones. In particular, we recommend additionalmeasurements of consumer mobility andpurchasing behaviour within markets toascertain the degree to which consumers define their own geographic markets.

We also propose several ways forward toincorporate the development of e-commerceinto geographic market definition. We argue for the use of data about Internet usage byconsumers that is tied to an understanding ofthe distribution of consumers within themarket. This will help pinpoint the potential for e-commerce to have an effect on off-lineretailers and suppliers.

In the field of substitutability analysis, wepropose a series of indicators to gauge thelikelihood that consumers will invest time inseeking substitutes. The measures are similar inscope to those used in defining the role of theconsumer in competition investigations.However, they are more capable of having realdata tied to them and are more explicit aboutthe judgmental weightings built into the overallscale. Again, the aim is not to supplantestablished methodology but to supplementestablished methods with consumer-focusedindicators.

Assessing assumptions

No one comes to the topic of competitionwithout a host of prejudices and historicalbaggage, and assessment of markets rests on anumber of these views. Aside from theestablished rules of economics and lessons fromthe study of markets, the followingassumptions make a useful starting point:


Introduction

Introduction

7

Markets exist only because of consumers

Consumption is the sole end and purpose ofproduction, and the interest of the producerought to be attended to only so far as it may benecessary for promoting that of the consumer. –Adam Smith 2

We sometimes forget that the market is only inexistence because of the consumer. Thedefinitions of markets tend to focus on thefirms that operate in them. If we think ofmarkets in this way we miss the centrality ofthe consumer to understanding how the marketfunctions. We also end up with regulatorysolutions that favour businesses overconsumers and aid collusion and abuse ofcompetition.

Competition operates in the real world

The social object of skilled investment should beto defeat the dark forces of time and ignorancewhich envelop our future – John Maynard Keynes 3

You can’t step twice into the same river –Heraclitus

We must subscribe to the post-Keynesian viewof the world:

• Economic and political institutions are notnegligible

• The economy is a process in historical (real) time

• In a world where uncertainty and surprisesare unavoidable, expectations have anunavoidable and significant effect oneconomic outcomes.4

Markets are complex and operate in real time.Any assessment of a market must take accountof this and not rely on simple snapshots orstatic views of competition.

There is more collusion in ordinarymarkets than most people presume

Adam Smith argued that '(p)eople of the sametrade seldom meet together, even for merrimentand diversion, but the conversation ends in a

conspiracy against the public, or in somecontrivance to raise prices.’ 5

Many consumers and regulators like to thinkthat the natural state of markets is competition.In fact, the natural state of markets is morelikely to be collusive. The desire of companiesand their managers to have a 'quiet life'encourages collusion and underminescompetition. This is not to say that allindustries are riddled with anti-competitivebehaviour – some are, and some are not.However, it does indicate that there are manymore abuses of competition and consumersthan most people would expect.

Government is not blamelessThere used to be essentially two views on thesource of monopoly power.6 These views weredivided into:

• self-sufficiency: firms gain monopoly powerunder their own steam:

• interventionist: monopoly power is accruedas a result of government intervention or failure.

In the real world, it is unlikely that either viewwill always be correct. In most cases, theaccrual of monopoly power will arise from aninteraction of the two sources.

Established approaches only tellpart of the storyCompetition regulation tends to analyse firmsin markets and has less experience in dealingwith consumers in markets. Inability to factorreal consumer behaviour into analysis will limitresponses to cases, reduce the ability to proposereal solutions and hamper the recognition of‘emergent properties’ in markets that a holisticview of a market will provide.

Competition is not an end in itself

Cecil Graham: What is a cynic?Lord Darlington: A man who knows the price ofeverything and the value of nothing.– Oscar Wilde, Lady Windermere’s Fan (1892)

While the study of markets and promotion ofcompetition are important exercises, it must beremembered that the object of the market is not


8

competition. We must raise our eyes from theworkings of the market to really see what sortof world we want to live in.

Markets are good at providingallocative efficiency, notdistributional equityCompetition analysis and investigations aregood for improving allocative efficiency but notfor dealing with distributional equity. It is quitedangerous to confuse the two targets.

Allocative efficiency, or Paretoefficiency/optimality, occurs when it is notpossible for anyone to be made better offwithout making someone else worse off. Paretoefficiency can be seen in conjunction with aNash equilibrium. A Nash equilibrium occurswhen a firm aims for profit maximisation at thesame time as competing firms, and means thatno firm can gain by changing its strategy fromthe one it has independently chosen.

Distributional equity is a broader social concerndriven by a desire to strive towards whatTobin7 has called 'specific egalitarianism'. Thisis the concern to ensure that distribution of aresource or service is done in a manner that isless unequal than the simple ability-to-paysolution would provide for. A distributionalequity solution may involve measures outsideand beyond the assessment of allocativeefficiency. As we discuss below, the decision-making process should assess and aim for theallocative efficiency position before consideringwhat distributional impacts may be.

Competition and competition policy

9

Competition and competitionpolicy

Raymond Carver, one of the finest Americanshort-story writers of the 20th century,produced a collection entitled “What Do WeTalk About When We Talk About Love?’ Thesame question can be asked aboutcompetition. Like love, we know competitionwhen we see it. But when we try to pin downexactly what it is, we find ourselves trippingover words and definitions. Like love,competition tends to be viewed in differentways by different people. This sectionoutlines what competition is, what it is notand what it might become. It also discussessome important distinctions betweencompetition, competition law, competitionpolicy and regulation.

At its most basic, competition is the process ofrivalry between firms and other suppliers forthe money and loyalty of customers over aperiod of time. The nature of this rivalrydepends to some degree on the structuresapparent in the marketplace and the historyand culture both of consumers and producersin that market. This rivalry tends to focus onone of two routes, or a combination thereof:

• Price-based competition: rivals compete tocut their costs and prices to catch theattention of customers.

• Service-based competition: rivals go beyondprice offers and offer differentiated serviceoffers. These may sometimes includeinnovations in product and service markets.

The package that the consumer sees in acompetitive market is a often combination ofthese two approaches.

10

The classic approach to competition stems fromAdam Smith’s notion that competition occurredwhen rivals acted in isolation from one another.Collusion would not generate competition.Following individual self-interest, whencollected together in a market, would generatethe best economic outcome. Smith’s focus onthe ‘Invisible Hand’ that guides individualactors in an economy to a wider end to whichthey had not strived has taken on almostmythical proportions in the modernunderstanding of markets.

Smith’s approach to competition was abehavioural one. Competition worked on thebasis of the behaviour of individual players inthat market and would be stymied by themonopolistic behaviour of those same players.The market was almost an exterior result ofindividual behaviours which none of theplayers aimed for but arrived at, courtesy of theInvisible Hand.

During the 19th century, a more structural viewof competition emerged. This approachgenerated the widely held view that a marketcould be defined as competitive when therewas a sufficiently large pool of sellers of ahomogenous (identical) product so that nosellers had a enough of a market share toenable them to influence product price byaltering the quantity they placed on the market.

The more structural approach of the 1930-40sled to the development of the Structure,Conduct, Performance approach to industrialorganisation. This approach marked a hugestep forward in the understanding of marketsand is still remarkably influential today.

The behavioural approach tends to be strongestamong the business community. It focuses onthe behaviour of firms and tends to argue thatif there is a competition problem, amending thebehaviour of firms within that market willdeliver results. The behaviouralist error focusesalmost exclusively on the appearance of rivalrypresented by businesses.

In contrast, the structural approach looksalmost exclusively at the market structure andassumes a deterministic relationship between


Part I: What iscompetition

structure, conduct and performance. It tends toargue that getting the structures right in themarket will drive behavioural changes.

The behavioural and structural approaches canlead to opposite errors in the analysis ofcompetition in a marketplace. Taking tooabsolutist a line in looking at markets in eithertradition can lead to the following errors:

• The behaviouralist error assumes that firms,left to their own devices, will tend towardscompetition. This is typified in the ChicagoSchool /Robert Bork approach to competitioninvestigation which rests on the implicitmoral superiority of capitalist enterprisesover government regulation. It assumes thatmarkets are best left to themselves and that ifthere is a breach of competition rules,behavioural remedies are best – simplydirecting the firms to their previous narrowpath of competitive virtue. This type of errorpredisposes the proponent to leave marketsalone for too long.

• The structuralist error assumes that allbehaviour in markets is driven by thestructure in that market. This tends to leadproponents to a reductionist, deterministicapproach that assumes that all you have todo is get the structures right (e.g.: wherethere are few firms, they must be colludingand must be broken up) and the desiredbehaviour will follow. This type of errorleads to a more intrusive regulatory approachto competition problems.

The two errors tie rather neatly into FMScherer’s acute observation, as described in hisclassic textbook8, about the difference betweenrivalry and competition. He defines rivalry asthe process by which business peopleconsciously jockey for position against rivalfirms. This, he argues, is often defined by thosebusiness people as competition. However, youcan have rivalry without vigorous competitionand competition without rivalry (e.g. oncommodity trading systems).

This distinction between rivalry andcompetition is an important and useful one forconsumer organisations. We are often faced

11

with markets where firms maintain that theyare competing vigorously for consumers’money, whereas we sense that the market is notworking as well as it could in the interest ofconsumers.

Part I: What is competition


The process of competition and itsmanifestation in rivalry between firmsdelivers a number of things in thosemarkets where it is strong. The effects ofcompetition can be generally grouped intotwo broad areas: those directly relevant forconsumers and those indirectly relevant for consumers.

Direct gains

It is always more straightforward to prove thedisadvantages for consumers where there is nocompetition than it is to prove how good it isfor consumers when competition exists. This isfor a number of reasons, most notably thatwhen measuring the impact of the blocking of amerger, or the liberalisation of a market it isalmost impossible to separate out each of thepossible reasons for a change in pricing orchoice. There are always ‘other factors’ thatcome into play. However, there are somegeneral things that can be said to flow directlyfrom competition.

Lower pricesThe greater the competition within a market,the more likely vigorous price competition willoccur. In contrast, monopolists’ only incentiveis to keep prices high. This poses afundamental problem for consumers in dealingwith pricing, a problem that rests on whetheryou can tell from price alone whether a marketis competitive or not.

Economic theory tends to argue that in aperfectly competitive market, prices will bedriven as near as possible to marginal cost. Inmost national markets, this will likely lead to

12

Part II: What competitiondoes

harmonised prices. At the same time, a marketrun by a cartel, or in which there is price-fixingor collusion, will also tend toward a singleprice. The conundrum for consumerorganisations is whether this single price isevidence of near-perfect competition or near-absolute lack of competition. The answer to thisconundrum requires steering a path betweenthe behavioural/structural errors outlinedabove.

Looking for collusionWhile there are no cast-iron pointers to guidethe consumer activist, the following signs maypoint to collusion:

• What opportunities does the industry haveto meet and discuss prices and output? If there is a strong industry lobby group,regular conferences and regular closedmeetings, the industry has the opportunity toengage in anti-competitive behaviour moreeasily than an industry without suchopportunities. It is a lot easier to create a cartelat the fringes of a regular industry meeting orconference than to organise a stand-alonegroup that has to be established specifically for the task.

• Does the industry tend to speak with thesame voice? In a highly competitive market, it is rare that allplayers will have the same view of an issue.While there will be many common issues onwhich they may agree (e.g. levels of taxation orregulatory burdens), beware of the industry inwhich there appears to be one voice only. Thispoints to an industry where the culture is oneof co-operation, not competition.

• How much price and cost signalling is there? Collusion is easier if you know your rivals’prices and costs. Look at the market and seehow easy it is for firms to see each other’sprices and understand each other’s costs. Forexample, in petroleum retailing, most firms aretotally vertically integrated and the productthey deal with is pretty much identical. Eachfirm thus knows that its own costs are unlikelyto be significantly different to its rivals. Thenature of retailing in this oligopoly marketmeans that each firm knows its rivals’ pricesand has a fair understanding of their costs.

Part II: What competition does

13

show that prospective competition (as proxiedby the number of years remaining toliberalisation) and effective competition (asproxied by the share of new entrants or by thenumber of competitors) both bring aboutproductivity and quality improvements andreduce the prices of all the telecommunicationsservices considered in the analysis. No clearevidence could be found concerning the effectson performance of the ownership structure ofthe industry (as proxied by both the publicshare in the PTO and years remaining toprivatisation). Severin Borenstein, Nancy Rose.Competition and Price Dispersion in the U.S.Airline Industry. July 1991. NBER WorkingPaper No. W3785

The pattern of price dispersion that we find doesnot seem to be explained solely by costdifferences. Dispersion is higher on morecompetitive routes, possibly reflecting a patternof discrimination against customers who are lesswilling to switch to alternative flights or airlines.We argue that the data support an explanationbased on theories of price discrimination inmonopolistically competitive industries.

Eric Kodjo Ralph, Jens Ludwig. Competitionand Telephone Penetration: An InternationalStatistical Comparison. By using the naturalexperiment in world telephony markets wherenations have chosen vastly different regulatoryregimes, this paper shows how competitionspurs telecommunications penetration. Further,we show that moving from two to three ormore firms is more important than movingfrom one to two, and that actual entry mattersmore than the threat of entry. This is ofeconomic as well as policy interest since game-theoretic models yield ambiguous predictionsabout oligopoly and monopoly when entry isthreatened.

Aaron S Edlin. Do Guaranteed-Low-PricePolicies Guarantee High Prices, and CanAntitrust Rise to the Challenge? Harvard LawReview, Vol. 111, No. 2, December 1997.

This article argues that there is an analogybetween a seller offering (and agreeing) tomatch a price for a buyer and other buyer-selleragreements that violate the Sherman Act. Thisarticle also considers a wholly new avenue for

Firms can estimate margins and the ability ofits rivals to engage in different levels of pricecompetition. If each player in a tight oligopolymarket knows price, costs and long runmargins, competition on price is extremelyunlikely.

What experience has taught usThe positive price effects of competition areenormously important. For consumerorganisations in some nations, competition maybe seen as a luxury or as a tool for the middleclasses. In other nations, consumerorganisations may ask: “What does competitionpolicy mean to a consumer living on less than adollar a day?” The simple answer lies in beingable to stretch that dollar further and havinggreater choice in where that dollar is spent.Turn the question around, and it is even easierto understand: “What good is a monopoly to aconsumer living on less than a dollar a day?”

Prices matter: they matter most for those withthe least to spend. An interest in driving pricesdown through competition is an interest sharedby all consumers. If competition can drivedown prices, then we must look to the policymix that helps deliver competition. The mixwill include reducing barriers to entry inmarkets (through trade liberalisation, removingrestrictive regulations and getting governmentsout of some markets) as well as rules to stopcompanies from simply controlling a market fortheir own benefit. Competition is not created bycompetition policy but competition policy canprotect competition once it occurs.

EVIDENCE FILE No. �

Oliver Boylaud, Giuseppe Nicoletti. OECDEconomics Department. Regulation, MarketStructure and Performance inTelecommunications Economics Department.April 21, 2000. OECD Economics DepartmentWorking Paper No. 237

Controlling for technology developments anddifferences in economic structure, panel data(long-distance (domestic and international) andmobile cellular telephony services in 23 OECDcountries over the 1991-1997 period) estimates


attacking price matching, asking whether theprice discrimination involved in matchingviolates the unfair-competition or price-discrimination laws. In so doing, this articleexamines whether price matchers should beable to protect themselves from such an attackwith a "meeting competition" defence. Breakingwith conventional wisdom, this articleconcludes that the defence should be rejected in cases in which meeting competition maysignificantly injure competition among sellers.

More choiceThe other clear and obvious benefit ofcompetition is that it can generate more choice.Instead of the one monopoly supplier,consumers can choose from different options.For example, in the telephony market, thechoice for many years was a land-line, if youcould get one at all. This land-line tended to becontrolled by the national, government-ownedmonopoly. The arrival of competitors in land-line operations in many countries changed thisstate of affairs. However, the biggest shift in themarket has been driven by the arrival of mobiletelephony. This has not only provided morechoice but also helped open access to themarket for many more consumers than couldpreviously access a fixed land-line.

New entrantsA competitive market can trigger market entry.(Of course, market entry can be triggered byflabby monopolists.) The argument about entryreflects the idea that entry will only tend tooccur where a potential new player seesattractive margins being earned by existingplayers. Thus a lack of competition triggersentry. However, a firm entering a market canonly survive if it can compete with rivals. If themarket is sewn up through collusion orgovernment regulation, then the new entrantwill simply not survive.

What experience has taught usChoice is sometimes seen as some terribleproblem for the middle classes, but whereconsumers can exercise choice, we should beempowered to do so. Competition can deliverchoice at its most basic level. When a countrylike India effectively has only one make of carfor many years, consumers are denied choice.

14

But when other firms are allowed to compete,consumers have a choice. What’s more,innovation is driven into the sector, aspreviously dominant players have to fight forconsumer money. However, choice does notdeliver everything. In some markets, choice isdifficult either because the market is complexor because the market is made complex byexisting players to confuse consumers intosticking with the firms they recognise. Thereare real barriers to choice that must berecognised in drawing up regulations.


Mary W Sullivan. US Department of Justice.The Effect of the Big Eight Accounting FirmMergers on the Market for Audit Services.March 17, 2000. Working Paper No. EAG 00-2.The research assesses how the two Big Eightmergers of 1989 affected the market for auditservices. A data set of 1,978 firms over a 12-year period is used to test four theories of howthe mergers could have affected competitionand consumer welfare. The study finds that themergers reduced the marginal costs of auditinglarge clients. There is no evidence that themergers were anticompetitive or that theyreduced costs for all types of audit buyers.

Better serviceIt may seem a strange to argue that competitiondelivers better service. It helps to identify whatwe mean by service. For example, in theaviation market, service is generally defined toinclude frequency and punctuality as well asthe cup of tea or coffee. Back in the 1970s, priorto liberalisation, US airlines competed almostentirely on ‘service’ rather than prices (whichwere fixed). This led to such insane ‘service’offerings as in-flight cocktail piano bars andplayboy bunnies. Bloated fares and lack ofcompetition encouraged the airlines to wastemoney on gratuitous service ‘innovations’.

What experience has taught usMonopoly providers offer little choice and getlazy, which drives down service quality (e.g.Aeroflot in the 1970-80s) or encourages them toinvest huge amounts of money on excessivefrills (e.g. US airlines in the early 1970s). Either


15

way, the consumer loses. Service quality tendsto improve when firms are forced to competewith other firms. Why would a monopolyprovider bother to provide a better servicewhen it knows it will not be punished if it does not?

Indirect gains

While the direct gains of competition arerelatively easy to identify, the indirect gains aremore difficult. However, the relationshipbetween the direct and indirect gains isimportant to investigate. Indirect gains throughenhanced efficiency, productivity and evenprofitability can result in even greater gains forconsumers. In a competitive market, therelationship between direct and indirect gainscan form a virtuous circle. In monopolisedmarkets, the opposite may be true: lack ofdirect gains for consumers can form a viciouscircle resulting in the destruction of efficiency,productivity and profitability that reduces thepossibility of future gains for consumers.

Efficiency and productivity gainsEfficiency and productivity gains for firms arenot usually viewed as being a benefit deliveredto consumers from competition. But whencertain conditions are met, such gains can forma virtuous circle of competition, efficiency andproductivity. When consumers choose betweenproducts on an individual basis, they signaltheir preferences. When these individualpreferences are aggregated together, firmsreceive a clear market signal of what consumerswant (and do not want) and what to produce.When a firm loses out, it is spurred on to findways to recapture consumer preferences eitherby cutting costs (to undercut their competitors),or by innovating and producing a betterproduct to recapture these preferences, or both.This constant need to capture consumerpreferences forces firms into a never-endingsearch for the productivity gains and efficiencyenhancements that will allow them to go one-up on their rivals.

This competition-productivity-efficiency drivecan bring about a situation where consumersbenefit from better service and lower priceswhile firms in the market see profit

enhancements. An example of this comes fromthe UK supermarket sector, where retail priceshave been declining steadily over recent yearsas profits have been rising. Consumers havebeen reaping a fair share of the benefits ofcompetition between the main players.

What experience has taught usThe indirect gains from competition may appearto be of secondary interest to consumers, butthey are of fundamental importance over thelong-term. More productive and efficient use ofresources helps an economy to develop morequickly and leads to a higher standards ofliving. (It does not, however, deal with issues ofincome distribution). More dynamic firms tendto innovate more and be more responsive totheir consumers. This triggers new products,new markets and new approaches to consumer-industry relations. Efficiency and productivitygains from competition are actually the mostimportant long-term benefit a consumer canreceive from competition.


Richard Disney, Jonathan Haskell, Ylva Heden.Restructuring and Productivity Growth in UKManufacturing. May 2000 CEPR DiscussionPaper No. 2463

We find that (a) 'external restructuring'accounts for 50% of labour productivity growthand 90% of TFP growth over the period; (b)much of the external restructuring effect comesfrom multi-establishment firms closing downpoorly-performing plants and opening high-performing new ones, and (c) externalcompetition is an important determinant ofinternal restructuring.

Francesco Trillas Jane, The Structure ofCorporate Ownership in Privatised Utilities.September 2002. CEPR Discussion Paper No. 3563

In the benchmark case where the governmentmaximizes privatisation proceeds, it is shownthat the optimal level of concentration increaseswith a tougher regulatory climate for investors.A more lenient regulatory regime increases thevalue of the commitment not to interfere

implicit in a more dispersed ownershipstructure. Deregulation (through increasingmonitoring costs) also pushes corporatestructure in the direction of more ownershipconcentration. When political objectives areadded to the analysis, it is shown that lobbyingwith managers induces levels of shareholderdispersion that are higher than in thebenchmark case. Collusion with largeshareholders, however, may yield higherconcentration levels than in the benchmark. Inthe case of managerial lobbying, the leniency ofthe regulatory climate does not have anyimpact on the equilibrium stake of the blockholder, and has a negative impact on thedifference between the political and thebenchmark outcomes.

Mariko Sakakibara, Michael E Porter.Competing at Home to Win Abroad: Evidencefrom Japanese Industry.

We find robust evidence that domestic rivalryhas a positive and significant relationship withtrade performance measured by world exportshare, particularly when R&D intensity revealsopportunities for dynamic improvement andinnovation. Conversely, trade protectionreduces export performance. These findingssupport the view that local competition, notmonopoly, collusion, or a sheltered homemarket, pressures dynamic improvement thatleads to international competitiveness.

Frank R Lichtenberg. Industrial De-Diversification and Its Consequences forProductivity. January 1990. Jerome LevyEconomics Institute Working Paper No. 35

Using plant-level Census Bureau data, we showthat productivity is inversely related to thedegree of diversification: holding constant thenumber of the parent firm's plants, the greaterthe number of industries in which the parentoperates, the lower the productivity of itsplants. Hence de-diversification is one of themeans by which recent takeovers havecontributed to U.S. productivity growth. Wealso find that the effectiveness of regulationsgoverning disclosure by companies of financialinformation for their industry segments waslow when they were introduced in the 1970sand has been declining ever since.

16

Michael Gort, Nakil Sung. Competition andProductivity Growth: The Case of the USTelephone Industry. Economic Inquiry.

Abstract: Both the estimation of total factor productivitygrowth and the analysis of cost shifts show amarkedly faster change in efficiency in theeffectively competitive market than for the localmonopolies. The results support, byimplication, a policy of permitting entry andcompetition in local telephone markets.

Jalana D Akhavein, Allan N Berger, David BHumphrey. The Effects of Megamergers onEfficiency and Prices: Evidence from a BankProfit Function. January 1997. FEDS PaperNumber 97-9.

We find that merged banks experience astatistically significant 16 percentage pointaverage increase in profit efficiency rank relativeto other large banks. Most of the improvement isfrom increasing revenues, including a shift inoutputs from securities to loans, a higher-valuedproduct. Improvements were greatest for thebanks with the lowest efficiencies prior tomerging, who therefore had the greatest capacityfor improvement. By comparison, the effects onprofits from merger-related changes in priceswere found to be very small.

Sumit K Majumdar. The Hidden Hand and theLicense Raj: An Evaluation of the RelationshipBetween Age and the Growth of Firms in India.Imperial College of Science, Technology andMedicine, Management School Forthcoming inJournal of Business Venturing

The evidence suggests that entrepreneurialbehaviour is an important feature ofcontemporary Indian industry. Recentanecdotes about Indian firms, particularly inthe information technology sector, suggest thatthere has been a resurgence of industrialactivity in the country. These beliefs are borneout by the analysis. The "hidden hand" is aliveand well in India! Additionally, the relationshipbetween size and the growth of Indian firms isnegative. This suggests that a process ofindustrial fragmentation may be taking place inIndian industry, with small firms growingfaster than larger firms and reducing the


importance of large firms in Indian industry.This has important implications for the futurecompetitiveness of Indian industry.Allen N Berger, David B Humphrey. Bank ScaleEconomies, Mergers, Concentration, andEfficiency: The U.S. Experience

Scale and scope economies in banking are notfound to be important, except for the smallestbanks. X-efficiency, or managerial ability tocontrol costs, is of much greater magnitude – at least 20% of banking costs. Mergers have nosignificant predictable effect on efficiency –some mergers raise efficiency but others lowerit. Market concentration results in slightly lessfavourable prices for customers, but has littleeffect on profitability.

Allen N Berger, Timothy H Hannan. TheEfficiency Cost of Market Power in the BankingIndustry: A Test of the ‘Quiet Life’ and RelatedHypotheses.

We find the estimated efficiency cost ofconcentration to be several times larger thanthe social losses from mispricing astraditionally measured by the welfare triangle.

What about jobs?Employment and unemployment are majorconcerns in all countries. It is often argued inthe popular press that greater competitionleads to higher unemployment and worseemployment prospects for workers. This beliefis supported by the evidence of immediatepost-privatisation restructurings that oftencentre on job losses. Stories of job loss generallyreceive more coverage than stories of jobcreation. When a factory shuts down and athousand workers are laid off, it is not hard toidentify the immediate losers from the process.In contrast, having ten firms employ 100 extrapeople does not grab tomorrow’s headlines.The problem with the relationship betweencompetition and employment is that measuringdirect effects is complex. For example, when anew firm sets up shop and enters a market, theimmediate employment impact is positive. Inthe long run, however, that firm might displaceanother or force the rival firm to engage inefficiency measures that require it to cutemployment levels.

17

A basic argument about the relationshipbetween competition and employment comesfrom its obverse relationship. In monopolysituations, a share of monopoly power is oftenexercised by the employees. This can bebeneficial in terms of wages and workingconditions. The correlation between monopolypower (public or private) and restrictivepractices by workers is fairly close. It makesintuitive sense that when a firm has aprivileged position, its employees are able toextract greater benefits from that firm at alower cost to that firm, because the firm canpass on these additional costs on to itscustomers. Almost every nationalised postal,railway, utility and national airline companyhas faced this problem. We often find,particularly in developing countries, that thesefirms are akin to arms of the state designed toemploy large numbers of people as analternative to the welfare state.

The problem with workers extracting a share ofthe benefits from the monopoly firm is that ittends to be a short-term benefit which laterendangers the entire enterprise whencompetition is introduced. Sharing monopolyrents only works as long as there are monopolyrents to share. When firms lose money, themonopolist is left either to extract morerevenue from customers or from the taxpayerthrough the state. The subsidy afforded by thestate is thus shared between the monopolistand its employees.

When competition is introduced, the incumbentmonopolist tends to be lumbered with abloated cost base (as its monopoly positioninvited inefficiency and encouraged its workersto extract ever-higher wages at the expense ofconsumers). New entrants are not faced withthis legacy problem and can thus come in withsignificantly lower costs. The impact on theincumbents will vary by industry.

The immediate likely impact on employment isthus negative from competition. Determiningmedium to long-term impact, however, is moredifficult. Competition drives efficiency intoresource allocation. If investors find that capitalinvested in a monopolist comes under pressure,they will likely seek to invest that capital wherethe return is better. Loss of capital from a


monopolist almost certainly means that anotherfirm or sector will benefit from increasedinvestment and will be better able to employ people.

What experience has taught usConsumer organisations have a real dilemmawith employment issues in competitionmatters, for diverse reasons. As concernedcitizenry with current or potential alliances toother groups that may benefit from maintainingthe status quo, it is easy for consumerorganisations to get drawn into protectionistarguments against competition. The high-profile nature of many deregulation initiativesand liberalisation efforts makes every job loss apolitical issue. It is always difficult to sitopposite people who will lose their jobs andargue that the move will lead to long-termgains and better employment prospects forothers. (As Keynes argued, in the long run weare all dead.) But the truth is that morecompetition, particularly among dominantincumbents, is good for the consumer, good forthe economy and good for employment.


Pietro F Peretto. Market Power, Growth andUnemployment. March 1998 Duke University,Economics Working Paper No. 98-16.

Labour market reforms that reduce the cost oflabour have effects in the product market thatreinforce the modern view that a morecompetitive labour market leads to lowerunemployment. This implies that such reformsare even more attractive than previouslythought. In agreement with the idea thatproduct market competition matters, moreover,I show that lower barriers to entry in theproduct market lead to lower unemployment.

Marianne Bertrand, Francis Kramarz. DoesEntry Regulation Hinder Job Creation? Evidencefrom the French Retail Industry. November 2001.CEPR Discussion Paper No. 3039

We show that stronger deterrence of entry bythe boards, and the increase in large retailchains' concentration it induced, slowed downemployment growth in France.

18

Michael C Burda. Product Market Regulationand Labour Market Outcomes: How CanDeregulation Create Jobs? January 2000. CESifo Working Paper Series No. 230.

A non-negligible component of the recentDutch employment miracle could be attributedto product market deregulation, in particularliberalization of shop-closing laws effected inthe mid-1990s. I sketch a model, based onBurda and Weil (1999), which can rationalizepotential public interest aspects of suchregulations as well as identify theiremployment and output costs.


The case that competition is generally good forthe economic well-being of a country is astrong one. Similarly the argument thatcompetition damages workers’ rights is weak.Of greater interest is the relationship betweencompetition and competition policy. This mayseem an academic point, but it is also importantfor practical policymakers. If you want toencourage competition, what balance of toolsand policies are needed to achieve this end?

All advocates of competition and competitionlaw and policy recognise that these are only one element of the toolkit to foster a morecompetitive economy. Within the area ofregulation that we loosely refer to ascompetition policy, a range of policies existsaimed at diverse areas of activity. We can dividethese policies along the following continuum:

19

Competition law and policy take differentforms in different countries. Most countriesseek a phased approach to their development.Cartel rules and rules on price-fixing areusually the first to be brought in, followed byincreasingly tough rules on mergers andcollusive behaviour. (Most developed countries,for example, had fairly strong laws on cartels,and price fixing before they had firm rules onmergers.) Sectoral regulation only comes intoits own during deregulation efforts. Obviously,sectoral regulation rules were not thought ofbefore countries started to privatise their public utilities.

As the table on page 20 illustrates, the patternof adoption of different elements of acompetition regime can be done on a graduatedapproach. While competition advocates willdiffer about which element goes where, itprovides a useful typology for thinking abouthow countries new to competition enforcementshould deal with competition law and policy.

Part III: Impact of competition law and policy

Part III: Impact ofcompetition lawand policy

Area of activity Type of policy

Government-supplied services Monitoring, target setting, competitive neutrality with private sector

Public utilities Basic sectoral regulationIntroduction of competitorsCompetitive tendering

Privatised public utilities Sectoral regulation Competition regulation

Private firms acting individually Rules on anti-competitive behaviour, price fixing,abuse of dominance, predatory pricing

Private firms knowingly acting together Rules on market division, market sharing, price fixing, collusion on keeping others out, cartel formation

Private firms effectively acting together Oligopoly rules for firms that effectively do not compete, market review mechanisms to inject competition

Private firms merging Rules on mergers

If a country wishes to privatise a publiclyowned company, it must recognise that it ismoving from what is termed a ‘high trust’ formof regulation to a ‘low trust’ form of regulation.High-trust regulation assumes that thosecarrying out the activity have a wider remitand broader accountability to elected officials.Low-trust regulation assumes that the firm willbehave in a narrowly self-interested way (asprivate firms are supposed to). This move fromhigh-trust to low-trust demands sectoral andcompetition regulation. When privatising, wemust assume that the newly minted privatefirm will abuse whatever power it has.

What does experience tell us?Experience shows that the best way to dealwith this problem is to ensure as muchcompetition as possible before privatisation ofthe domestic monopoly. Once competition hasstarted to bite, additional liberalisation can beintroduced, provided it is followed by sectoralregulations to ensure that competition thrives.If privatisations are not handled properly, theconsumer interest is undermined and serioussocial problems, including unrest, can occur.

Competition policy is flexible; no one-size-fits-all approach is possible. Countries can offeradvice and recommend reforms but theythemselves will be re-assessing on a permanentbasis how they conduct competition reforms.Competition policy is about getting the toolkitright to protect consumers and foster

20

competition, not about requiring countries to follow a rigid path to development.


George Symeonidis. Price Competition,Innovation and Profitability: Theory and UKEvidence. University of Essex – Department ofEconomics; Centre for Economic PolicyResearch (CEPR) May 2001. CEPR DiscussionPaper No. 2816

The econometric results suggest that theintroduction of restrictive practices legislationin the UK had no significant effect on thenumber of innovations commercialised inpreviously cartelised R&D-intensivemanufacturing industries, while it caused asignificant rise in concentration in theseindustries. In the short run profitabilitydecreased, but in the long run it was restoredthrough the rise in concentration.

Brian R Cheffins. Investor Sentiment andAntitrust Law as Determinants of CorporateOwnership Structure: The Great Merger Waveof 1897 to 1903. Faculty of Law, University ofCambridge.

One theme the paper develops is that mergersmatter with respect to the evolution of systemsof ownership and control. A second topic thepaper deals with is the process by which a


Different stages of development for a national competition regime

Start Enhancement Advancement Maturity

Competition advocacy and Merger control Regulation Second generationpublic education international agreements

Control of horizontal Vertical restraints International Pro-active competitionrestraints co-operation advocacy

agreements

Checking abuse of Development ofdominance effects doctrine

Exceptions and exemptions, including on public interest grounds

Technical assistance

Source: Gesner Oliviera after Shyam Khemani and Mark Dutz. 1996. The Instruments of Competition Policy and their Relevance for Economic Development. PSD Occasional Paper No. 26. World Bank. Quoted in P Mehta. 2003. Friends of Competition. CUTS, Jaipur.

country's investors become sufficientlycomfortable owning publicly traded shares topermit a transition from concentrated todispersed share ownership. A third theme thepaper emphasizes is antitrust law's significance.The experience in the U.S. and Germanysuggests that the legal status of anti-competitive alliances is a potentially importantdeterminant of corporate ownership structures.

J Gregory Sidak, Michael K Block, F C Nold.The Deterrent Effect of Antitrust Enforcement.Journal of Political Economy, Vol. 89, No. 3, 1980.

We show that a cartel's optimal price is likely tobe neither the competitive price nor the pricethat the cartel would see in the absence ofantitrust enforcement but rather anintermediate price that depends on the levels ofantitrust enforcement efforts and penalties. Ourempirical results reveal that increasing antitrustenforcement in the presence of a credible threatof large damage awards has the deterrent effectof reducing mark-ups in the bread industry.

J Gregory Sidak. Rethinking AntitrustDamages. Stanford Law Review, Vol. 33, 1981.

Part I analyzes the consumer's economic injuryfrom exploitative behaviour and shows that,prevailing contrary opinion notwithstanding,the Clayton Act does not unambiguouslyestablish a consumer right to be free from suchinjury. Because the prevailing interpretationmay cause allocative inefficiency, Part Iproposes a countervailing producer's right anda corresponding damage rule. Part II analyzesthe kind of injury that competitors suffer fromexpansionary behaviour. It criticizes thecompetitor's right suggested by the currentdamage rule and proposes an alternative rightand damage rule that would improve socialwelfare by enhancing productive efficiency. PartIII proposes implementing the economic rightssuggested in Parts I and II through a judicialtest for calculating antitrust damages thatwould restrict the availability of such damages.

Bruce H Kobayashi. Antitrust, Agency andAmnesty: An Economic Analysis of theCriminal Enforcement of the Antitrust LawsAgainst Corporations. George Mason Law &Economics Research Paper No. 02-04.

21

Because criminal fines are not accuratemeasures of loss, and because of the vicariousnature of corporate liability, there is a greatdanger that higher-than-optimal penalties willinduce corporations to incur excessive costs inan attempt to avoid these high fines. Thepotential over deterrence costs resulting fromhigher-than-optimal fines is exaggerated by theAntitrust Division's expanded use of theCorporate Leniency Policy. Ironically, the costsof over deterrence will result in higher prices toconsumers, a decrease in welfare, and,ultimately, in the exact effects that the criminalantitrust laws are intended to prevent.

Margaret C Levenstein, Valerie Y Suslow. What Determines Cartel Success? University ofMichigan Business School Working Paper No. 02-001.

Our examination of cartel duration concludesthat cartels are neither short-lived nor long-lived; they are both. Similarly, our analysis ofthe effect of cartels on prices and profitabilityfinds that there is enormous variance in cartelsuccess at raising price to the joint-profitmaximizing level. In our examination of cartelbreakdowns we find, as suggested by recenttheoretical literature, that cheating is a commoncause. Occurring even more frequently,however, are entry, external shocks, andbargaining problems, suggesting that theseissues should be given deeper consideration infuture work. Stigler's hypothesis that largecustomers contribute to cartel breakdowns isborne out in a few case studies. But thereappear to be more cases in our sample in whichlarge customers help to stabilize the cartel.Only the oldest of suppositions, that highlyconcentrated industries are more prone tocartelisation, seems to hold true across studies.Our inability to find more commonality amongthese studies and among cartels does notsimply reflect our ignorance of carteloperations or secrecy on the part of cartels (orthe different methodological approachescovered in this survey). Rather, it reflects theinnumerable possibilities for organizing asuccessful cartel, and the interdependence ofthose factors determining cartel success.

Part III: Impact of competition law and policy

22


23

Consumers and competition

Consumersand competition

Increasingly, competition regulators mustface up to the difficulties of conductingmarket analysis in areas where consumerbehaviour is an important factor. This needis most immediately felt in the investigatoryphase, when regulators map the market.While the best regulators are attempting tounderstand how consumers behave in themarket, this has tended to proceed on anad-hoc basis largely based on the interest ofstaff and the willingness of panel membersto ensure its operation. However, theimportance of consumer behaviour must notbe underestimated in identifying possibleregulatory solutions to competitionproblems. It is here that an understanding of drivers for consumer behaviour isparticularly important.

This section identifies ways to assessconsumer behaviour in competitioninvestigations. It starts from thepresumption that optimal consumerbehaviour in a market is characterised byclassical economic theory. However, acombination of structural and behaviouralfactors hamper the ability of consumers toattain the model of rationality established inclassic economic theory. We thereforepresent a matrix of factors and a scale ofimportance to estimate the likelihood thatconsumers will conform to this rationalconsumer model. We hope these categoriesof effects will provide a guide for indicatorsof consumer behaviour to supplement themore established methods of marketinvestigation.

24

Why should we care?The behaviour of consumers in markets is oftenviewed as a ‘given’ in the analysis ofcompetition problems. Classically trainedeconomists have tended to view consumers asrational beings who will pursue the maximumbenefit to themselves in a selfish way. Themessages transmitted to the market from thisindividual behaviour will help direct resourcesin an efficient manner and help drive marketsto more efficient operation.

If competition investigations and analyses areconducted on the basis that consumers act in adetermined manner, such investigations willmake certain assumptions at both the front andback end of their deliberations. Thus, in theinvestigation phase, regulators may assume acertain pattern of behaviour corresponding towhat one would expect from a rational model ofconsumer behaviour. Similarly, at the final stageof investigation, regulators may assume thatconsumers will behave in a classically ‘rational’manner when remedies are introduced.However, two key questions must be asked:

• To what degree does consumer behaviourmatch the classical model?

• If it does not, what effect should this have on the manner in which markets are viewedand regulated?

Classical model of consumer behaviour

The classical model of consumer behaviourassumes that consumers do essentially threethings in making decisions: “All humanbehaviour can be viewed as involvingparticipants who [i] maximise their utility [ii]from a stable set of preferences and [iii]accumulate an optimum amount of informationand other inputs in a variety of markets.” 9

To a large extent, the degree to which theserules apply will indicate the degree to whichthe assumption of consumer rationality applies.For analytical purposes, we thus need to askthe degree to which each rule applies.

Three key caveats need to be placed on theclassical approach:


Part I: Markets andconsumerbehaviour

25

Part I: Markets and consumer behaviour

1 Bounded rationality: Human cognitiveabilities are not infinite. (We all have limitedcomputational skills and flawed memories.)

2 Bounded willpower: People often takeactions in the short term that they know to bein conflict with their own long-term interests.

3 Bounded self-interest: People generally care,or act as if they care, about others, evenstrangers, in some circumstances.10

Given these bounds on behaviour, we need toidentify the degree to which consumers act inthe classically rational manner. To do this, weneed to identify all the problems that get in theway of consumers making rational decisions.We can thus identify factors over which wehave little control and those that regulators and the operation of the market can affect.

Utility maximisationThe idea of the utility-maximising individual is key to the classical view of consumerbehaviour. However, it is also one of theweakest links in the chain of argument thatseeks to place consumers within the rationalchoice model. A number of important caveatsmust be placed on the idea that consumerspursue a utility-maximising approach. Theseinclude cultural and peer group issues andmore basic structural processing issues on thepart of individuals.

The following factors can limit the operation ofa utility-maximising consumer.

Cultural aspects to decision-makingConsumers do not behave in a vacuum in anygiven market; their behaviour is circumscribedby their own cultural norms and thoseprevalent in the market. A clear understandingis needed of a market culture and the degree towhich the consumer can operate comfortablywithin it.

Tests:

• What is the culture of the market in question?• In what sub-culture does the market sit?• How will changes to that market affect the

cultural interactions of participants?

The endowment effect: Any product that is already part of theindividual consumer’s existing endowment

will be more highly regarded than a productthat is not. Individuals tend to rate what theyalready own higher than products that they donot own.

Tests:

• Proxy measures may include markets withheavy advertising budgets and campaignsaimed at boosting ‘new’ features to existingproducts. This may spill over into misleadingand deceitful advertising.

• Pressure advertising/marketing may be a problem

Sunk costs and the momentum theory The momentum theory argues that individualswill complete a task once work has begun,irrespective of the continuing validity of theoriginal decision. A sunk cost is an already-borne cost that is not easily recoverable.Individual sunk costs do affect decision-making.

Tests:

Is this a market where:• sunk costs are common, and legitimately so?• final decisions are arrived at over time?• consumers are required or encouraged to pay

for goods over time?• the final quality of work is only ascertainable

long after payment?• consumers have little opportunity to revisit

original decisions?

Psychic costs of regretPresent decisions may be affected whenindividuals feel unable to trust themselves tomake correct decisions in the future.

Tests:

• Is there a mismatch between ‘objective’measures of consumer need and ‘subjective’assessments?

• Is there a need for compulsion in markets?(e.g. health insurance)

• Is the market one that attracts lower incomeconsumers through its ability to discountlong-term fallibility in decision-making? (e.g.Christmas Clubs)

• Is this business proposition risky enough towarrant higher charges, opaque charginginformation or high credit charges?

• Will comparative information limit thepotential for abuse?

26

• What sources of information are available toenable decisions to be made?

• How accurate is information about products/services that incorporate losses (investmentproduct) and gains (M&S voucher)?

Stable preferencesThe idea that consumers make decisions from afoundation of stable preferences can beundermined by many factors, including thoserelated to the risk and perceived risk of adecision and the cost of getting a decisionwrong. The key difficulty, in analytical terms, ofidentifying factors in the operation ofpreferences is the close link between thosepreferences and the idea of utilitymaximisation. Many factors in consumerbehaviour that undermine the idea of theutility-maximising individual also underminethe idea that consumers operate from a set ofstable preferences.

The operation of a stable set of preferences canalso be undermined by the nature ofconsumption behaviour. For example, if theconsumers purchase a service or product thatrequires them to use an intermediary foradvice, the preferences of the sellers will be asimportant as those of the consumers. Similarly,the transparency of the transaction will beimportant, as will the degree to which theproduct and/or service are bundled in apackage whose individual parts are difficult toidentify and accurately price.

Optimal informationMuch work has been carried out on theproblems of information transmission inmarkets. The role of information economics hasgrown over recent decades in response to theproblems faced by regulators in liberalisingmarkets and by companies operating incomplex markets. The 'common understanding'reached is that information acts as a grease toeffective markets and the more information thatcan be made available, the more effective thisgrease will be. Unfortunately, the ability ofconsumers to compute the information theyreceive cannot match the desire of regulatorsand companies to furnish that information. Asa result, information can actually limit theeffective operation of a market, or make itslower to react to market mechanisms.


Self-control and pre-commitmentConsumers often recognise that their existingconsumption patterns are incapable of meetingcertain future needs (e.g. Christmas spending,retirement). This prompts saving and tying theconsumer into patterns of committedexpenditure.

Tests:

• Is this a market where regular payments forfuture goods are commonplace?

• How does the decision-effect-feedback loopwork in this market?

• How clear to the consumers are the costs of gradualism?

• How independent is the informationavailable to the consumer about the decision-effect-feedback loop?

Losses and gains treated differentlyBecause losses and gains are treated differently,we need to have an understanding of how themarket in question treats losses and gains inmarketing and product/service provision.

Options for dealing with combinations of losses and gains:

• Segregate gains: Individuals prefer to treatmultiple gains as a series of individual gains.(e.g. two gifts wrapped separately arepreferable to two gifts in a single wrapping).

• Integrate losses: Individuals like to place alltheir losses in one basket.

• Let big gains cancel small losses: If theoverall balance of gains and losses is positive,losses should be pooled with the gains tocancel them out.

• Segregate ‘silver linings’: When large lossesout-weigh small gains, gains may beseparated out as a ‘silver lining’ to the cloudof the large loss.

• The picture becomes less clear when dealingwith smaller gains and losses. Here,integration may be the preferred option.

Tests:

• What opportunity exists in the market forconsumers to bundle and re-bundle lossesand gains?

• How clear is the quality/price information inbundled products/services? Can objectivecalculations be made of relativecosts/benefits?

27


The following are constraints on the optimalacquisition of information by consumers:

Framing and informationThe quality of information available to theconsumer is key. However, the importance of thedecision and the time the consumer will/willnot want to allot to the decision cannot beunderestimated. Markets with low levels ofclarity in information and time investment mayforce consumers to make irrational or incorrectdecisions, or no decisions at all. Solutions insuch markets are unlikely to centre on moreinformation. Solutions may range from lessinformation to clearer information,benchmarking and basic product design.

Problems are presented to consumer in two ways:

• Transparent: Choice behaviour does notviolate basic tenets of rationality.

• Opaque: People may well violate basicprinciples.

Tests for transparency of information:

• How transparent/opaque is the informationpresented to the consumer?

• In what environment does the consumermake the decision?

• How much time does the consumer have tomake the decision?

• What other sources of information areavailable to the consumer to make adecision?

• What effect will a merger/behaviour have onthe transparency/opacity of information?

The structure of a problem The way in which a problem is presented to theconsumer may affect the choices made.Prospect Theory tells us that the same problempresented in different ways may influence thedecisions taken.

Tests:

• How many ways is the information in themarket presented?

• How uniform is the presentation of marketinformation?

• How much time will the consumer need toinvest before a decision is made and howmuch gain will the consumer get from that decision?

• How will the consumer make a cost-benefitanalysis in the market?

How consumers learnMarkets with limited learning opportunitiescontain a number of incentives to abuse marketpower and consumers (e.g. investmentmanagers making decisions whose effects willnot be uncovered for decades). While somemarkets have limited opportunity for learning,others try to limit learning for reasons ofcontrol (e.g. many insurance and investmentproducts, travel agents). Consumers needsufficient, clear information to learn. Solutionsin markets where this is not the case mayinclude information requirements and clearupdates on long-term decisions.

Necessary feedback on decisions is oftenlacking because:

• Outcomes are delayed and not easilyattributable to a specific action.

• Variability in the environment degrades thereliability of the feedback, especially whereoutcomes of low probability are involved.

• There is no information about what theoutcome would have been if another decisionhad been taken.

• Most important decisions are unique andtherefore provide little opportunity forlearning.

• Outcomes received with certainty are over-weighted compared to outcomes that areuncertain.

• Gains are treated differently to losses. Lossesgenerate a risk-seeking response while gainsproduce a risk-adverse response.

Tests:

Delay and variability• How long does it take before a consumer

understands the effect of a consumptiondecision? The obvious application here isfinancial services (e.g. credit cards vs.pensions). A rule of thumb: the longer thetime between decision and effect andfeedback, the weaker the potential foreffective competition.

• What is the transmission mechanism forinterpreting the decision-effect-feedbackloop? Who controls it?

• How easy is it to understand thetransmission mechanism and how accessibleis the information contained in it?

Uniqueness and alternatives• How often is the consumer in the market?• How many other decisions will the consumer

have made in the market?• Do any accurate proxy measures exist in the

market for the consumer to rely upon?

Time and importance• How important is the decision-effect-

feedback to the consumer?• What other decisions-effect-feedback loops

will the consumer be dealing with during asimilar time frame?

• Will peer group pressure have any influenceon the seeking/acceptance of information?

• How are consumers likely to value the time needed to interpret the transmissionresults?

Over-weighting certainty• How certain is the relationship between

decision-effect-feedback? How much is theconsumption decision a bet (e.g. aninvestment) and how much a certainty (e.g. a tin of beans);

• What is the consumer understanding of thebalance of probabilities in the market?

• How clear is the information in the marketabout probabilities?

Gains and losses• Is the product/service a bundle?• How clear are the gains and losses in the

market?• Do consumers have the opportunity to

re-bundle gains and losses according topreference?

Search costs Any difference in price between goods is seenin relation to the total price of the goods. Thusconsumers will spend time searching for alower-priced television, but not for a tin ofbeans. This has greatest implications for substitutability assessments and for thepossibility of local monopolies.

28


Tests:

• How important is this consumer decisionrelative to the potential saving?

• What is the likely peer group view of thegains from shopping around?

• What is the consumer understanding of themarket within which they are shopping?

• What are the physical bounds of the market?• How long will a consumer normally shop

around for in this market?• Are impediments placed in the way of a

consumer to limit the ability to find searchinformation?

• Do stores stock products that makecomparison shopping easier?

• How much of the market is migrating to the Internet, both in consumer and retailer terms?

Undermining the model ofrational consumers

Competition analysis is replete with indicesand mathematical calculations. These measuresall relate to the structural characteristics of themarket while tending to skate over theelements that relate more directly to consumerbehaviour. Attempts to bring more cohesion tounderstanding consumer behaviour in a marketare prone to the same difficulties seen in tryingto uncover strategic behaviour by companies.Pointers and proxy measures are thus useful inindicating certain patterns for assessing thedegree to which consumers will behave in aclassically rational manner.

The factors outlined in the matrix discussedbelow attempt to identify barriers to rationality.Given the scale and identification of problems,they can only be used as a rule-of-thumb guideto assessing markets. The advantage of usingsuch rules-of-thumb rests primarily on theability to identify those individual barriers andcombinations of barriers, which will hamperthe effective operation of consumers in amarket. As such, they may help to framesolutions or contribute to the assessment ofproposed policies.

Matrix of factors hindering consumerrationality

Scale: 1=low, 2=medium, 3=high

How likely is it that consumers will maximise utility?

Length of time for the decision process to be made

Cost of searchThe opportunity cost of the decision

Cost to revisit a decisionCost to revise the decision

Sub-total for utility maximisation

How stable are consumer preferences likely to be?

Cost of getting decision wrongNumber of people involved in decision

Risk involved in the decisionBundling of other products/services to choice

Sub-total for stable consumer preferences

How good will the information be?Length of time post-decision to see effect

Volume of information requiredBackground information needs

Diversity of information presentationDecision feedback time

Likelihood not in the market againLikelihood not in related markets again

Likelihood proxy measures of performance not available

Sub-total for optimal information

Overall totalLikelihood that rational consumer model

will not apply

When the rational modeldoes not apply

The matrix for assessing the degree to which therational consumer model applies can only act asa guide in assessment of markets and possibleregulatory or market behaviour. The matrixpresents both structural and behavioural aspectsof markets. The object of regulators/marketparticipants should be to move the marketcloser to the ideal type for consumer behaviourto become more rational. This would producelower scores for each market.

29

The matrix should also provide a useful tool forassessing the likely effects of remedies incompetition cases. The question here will be thedegree to which a merger or a remedy in acomplex monopoly case or regulatory decisionincreases or decreases the likelihood thatconsumers will act in a rational manner.

While use of this matrix provides rule-of-thumb assessments for factoring-in consumerbehaviour, this approach must be employed inconjunction with other tools for competitioninvestigations.

The role of the matrix will differ betweenmarkets where direct sector-specific regulationexists and markets where no such regulationexists. In the former, it helps point to specificremedies or structural/behavioural indicators.In the latter, it helps identify the need toaddress broader policy goals (e.g. informationprovision or development of proxymeasurements) to allow informationtransmission to work more effectively.

Worked examplesThe matrix of questions to be asked in a givenmarket is applied to a series of specificexamples below (see page 30). The marketschosen range from complex, to verticallyintegrated to relatively simple. The workedexamples test the relative application of the matrix.

Totals represent the sum of each score; theestimation of the applicability of the rationalconsumer model reflects the degree to whichthe total score relates to the 'worst case' market(where all questions receive a 3 rating). In orderto maintain consistency in the application ofthe scale, some questions are presented the'wrong way round'. This keeps the scale andrankings relatively unsullied. (Please note,however, that a degree of bias in the estimationof scale is unavoidable, particularly when it issubjective.)


30


Worked examples of rational consumer matrix

Ideal Pension Cars Holiday Tin of beansmarket

How likely is it that consumers will maximise utility?Length of time for the decision process to be made 1 3 2 2 1Cost of search 1 3 3 2 1The opportunity cost of the decision 1 2 2 2 1Cost to revisit a decision 1 3 3 1 2Cost to revise the decision 1 3 3 1 2Sub-total for utility maximisation 5 14 13 8 7

How stable are consumer preferences likely to be?Cost of getting decision wrong 1 3 3 2 1Number of people involved in decision 1 2 2 2 1Risk of the decision 1 3 2 3 1Bundling of other products/services to choice 1 2 1 3 1Sub-total for stable consumer preference 4 10 8 10 4

How good will the information be?Length of time post-decision to see effect 1 3 1 2 1Volume of information required 1 3 2 2 1Background information needs 1 3 2 2 1Diversity of information presentation 1 3 3 3 2Decision feedback time 1 3 1 3 1Likelihood not in the market again 1 3 2 2 1Likelihood not in related markets again 1 2 1 2 1Likelihood proxy measures of performance not available 1 2 1 3 1Sub-total for optimal information 8 22 13 19 9

Overall total 17 46 34 37 20Likelihood that model will not apply 90.20 66.67 72.55 39.22

Scale: 1=low 2=medium 3=high

Bounds: most likelihood that model will apply: 17most likelihood that model will not apply: 51

Defining the relevant market is acornerstone of competition investigations.Definitions have emerged from a learningprocess based more on organic evolutionthan on the development of hard and fastrules. While this provides endlessopportunities for interpretation, it doesafford some degree of flexibility in thesystem. We propose a set of additionalfactors to be taken into account in thedefinition of relevant markets. In thegeographic market analysis stage, weadvance a number of proxy measuresavailable to regulators to assess the degreeto which consumers define the geographicbounds of their own markets. We alsopropose a route to understanding thepotential impact of electronic commerce onthe operation of specific markets: thatregulators utilise market segmentation dataand consumer research work to identify thelikelihood of consumer groups transferringtheir purchasing behaviour to onlinemarketplaces.

Why bother?Once the degree to which a consumer is likelyto behave in a classically rational manner hasbeen established, it is necessary to analyse theextent of the market within which thatconsumer will operate. Central to anycompetition investigation is the definition ofthe relevant market. One shorthand rule usedin monopoly cases in the US is that the relevantmarket is the market that is capable of beingmonopolised. Another, less circular, shorthandmethod for assessing a market involvesanswering the following question (referred toas the ‘Hypothetical Monopolist Test’):

31

Part II: Defining the market


• Could a potential monopolist raise its pricein that market for a significant period of timeand hold it there?

Essentially, two elements are relevant to marketassessment:• Geographic market definition• Product market definition

Geographic market definition

Competition problems are commonlyassessed in a static manner and mostmeasures reflect this bias. We recommendsome changes to the spatial analysis ofrelevant geographic markets to make themrelate more closely to the way in whichconsumers behave.

Once the likely pattern of consumer behaviourhas been mapped, it is necessary to place thatconsumer within a bounded market of somedescription. Most consumption activity takesplace in a physical environment of some sort,and this involves some form of bounds to amarket. To get a sense of how that market isconstrained, it is necessary to place the activityof that market correctly within those bounds.

How it is currently doneThe following key facts help to define thegeographic bounds of a market and reflect thequestions usually asked by competitionregulators in Europe.11

• Where is the product marketed?• What is the geographic market into which

the product is sold?• Is this defined by regulation?• Is the geographic area contiguous with a

similar one?• What are the physical barriers to entering

that geographic market?• Where is the market under which

homogenous market effects can be found?• Where are the conditions under which firms

operate fairly uniform?

To these questions, we can add:• Is the market a service market or product

market? This is important because servicesare less easily tradable on a cross-border

32

analogous to the Central Business District. The size of the market within which theconsumer is willing or able to travel varies forevery product. One would assume (not alwayscorrectly) that the more expensive the product,the larger the concentric rings. This appears tobe borne out by two developments in retailing:1) the situation of retail premises in manycountries; 2) the price-promises made byretailers indicating the assumptions they makeabout consumer shopping behaviour. TheDixons Group, for example, offers two differentprice-matching guarantees. Dixons stores(which tend to be on high streets and in out-of-town retail parks) offer a price-matchingguarantee for products brought within 10 milesof a store. However, the PCWorld part of thechain offers a price-matching commitmentwithin 30 miles of a store.

Several methodological enhancements can aid consumer organisations in geographicmarket definitions for competitioninvestigations:

Von Thunen CirclesMany consumer organisation seek to mapconsumer behaviour through surveys whichidentify the distances that consumers arewilling to travel to purchase certain goods.Whether simple interview questionnaires ormore complex diary exercises, these surveyscan provide a useful indication of thegeographic market definition applicable incompetition cases. This can then be comparedto price-matching guarantees and similar workcarried out by firms operating in the market.For example, in its recent work on new carpricing, the UK Consumers’ Associationshowed consumer willingness to travel up to 50 miles to purchase a car.

Internet market definitionThe impact of the Internet as a force forchallenging existing models of competition ismore talked about than actually analysed.However, geographic market definitionprovides a useful rule-of-thumb that may allowcompetition regulators to assess the potentialimpact of the Internet on a market.

As a first step, an investigation should ask two questions:


basis and are consumed on the spot. Services often require cultural specificity (e.g. advertising agencies operate withinculturally highly specified bounds whereconcentration in a market is unlikely to spurmajor entry by foreign firms).

• What pattern of spatial distribution doesconsumer purchasing follow?

• How does the consumer view the geographic market?

The end of geography?Much has been made of the power of theInternet and e-commerce to alter markets. Interms of competition regulation, one of the keyimpacts of the Internet will be on geographicand product market definition. Almost bydefinition, the Internet broadens the geographyof the market within which the consumer willoperate. Internet widens the potential realm ofconsumer shopping from that which isphysically at hand to that which is possible bycourier. The Internet allows the consumer toavoid the bounds of the physical location ofstores by providing almost unlimited access tosites. However, the constraints placed on thischoice include customs duties and legal limitsas well as old-fashioned delivery problems. Inthis sense, the Internet lies somewhere betweena real entrant and a potential entrant envisagedby contestable market theory.

Consumers define their own geographicbounds to the marketplace through theireveryday consumption behaviour. A usefulanalogy comes from the work of von Thunen ineconomic geography. He developed adeceptively simple theory of land allocationbased on an isolated town surrounded byfarmland. The allocation of that land woulddepend on the value of the crop grown and therelative cost of getting that crop to the town.Von Thunen ended up with a now instantlyrecognisable concentric rings model of landallocation. This model was later adapted fortown planning purposes and lead to theidentification of the Central Business Districtand suburbia. In many ways, consumersoperate a series of von Thunen circles for theirconsumption decisions.

The foundation for most consumptiondecisions is the home, which can be viewed as

• To what degree does the existing retailmarket correspond to the Pareto 80/20 rule(e.g. what portion of the consumers in themarket account for what portion of profits)?As a general rule, the greater the share ofprofit accounted for by the smaller number of consumers, the greater the potential forseemingly small-scale entry to have asignificant effect.

• If the market does display a loading towardone small segment of the population, to whatextent is this population likely to migrate tothe Internet (or other technologicaldevelopments)?

The latter factor helps the regulator identify thepotential impact of the Internet on a retailmarket. As a rule-of-thumb, it falls somewherebetween an analysis of the contestability andthe problem of anticipatory dominance.

Matrix of key questionsTo update the standard approach to geographicmarket definition, we propose incorporation ofthe following questions:

Geographic market definition• How big are von Thunen circles of

consumption?• What areas do price-matching guarantees

cover?• How does the Pareto Rule apply to this

market?• Are Pareto consumers switching/likely

to switch to the Internet?

Product market definition

With analysis of consumer behaviour a firststep and identification of geographic boundinga second step, the definition of product-basedcompetition follows naturally as a third step inthe analysis of the effective operations of a market.

How it is currently doneThere are two main elements to the product-market definition:

• Demand side• Supply side

33

On the demand side, two key tests to analysethe correct product market are needed:

• Substitutability • Complementarity

SubstitutabilitySubstitutes are near products to the product in question. They are important because a price rise in one product will theoretically lead to consumers switching to substituteproducts.

There are a number of physical measures ofsubstitutability:

• Price elasticity of demand: Given a scarcityof data, this textbook calculation is difficultto operate in practice. The arrival of scannerdata in supermarkets makes calculation ofprice elasticities of demand easier. The priceelasticity of demand is:

% change in demand

% change in price

The higher the figure, the greater the degree of price elasticity and substitutability. While a low ratio will indicate either low substi-tutability or market power, it is not clear simply from this measure which one you aredealing with.

• Cross-elasticities of demand: The degree towhich a price change in one product willlead to a demand change in another

• Chain of substitution: A broad range ofproducts that would not normally be seen assubstitutes (e.g. Rolls Royce vs. Skoda) butwhere a price change in one has a knock-oneffect on its nearest substitute, leading to achange in prices down the substitution chain.The justification for supply chains is weakand is often used to excuse abusivebehaviour. However, chains can operatewithin sub-markets (e.g. qualitynewspapers/mid-market/tabloid). Thedegree to which a chain exists is largelydetermined by the degree to which theconsumer recognises the chain.


• Innovation: Is this a market in which:– new substitutes are being created

(e.g. Playstation II for Dreamcast)? – new substitutes are likely to be created

(e.g. WAP for Internet)?

If the answer is yes, the ability of a firm toleverage market power is restrained and mayebb by the time regulatory processes havetrundled into action.

Other important characteristics to be taken into account include:

• Physical characteristics of products, ascharacterised by: physically substitutable(e.g. Mac vs. Windows software) includingpath dependency (a situation where a choiceof technological path determines the sort ofcomplementary products that will flourishe.g. Mac, Windows, VHS-Betamax, Qwertykeyboards)

• Perceived-to-be substitutes, which view therelevant comparative performance of theproduct (difficult to assess); service substi-tutability, which is even more difficult tomeasure than product substitutability (e.g. in airlines, financial services, burgers, etc.)

• Price spreads: What is the range of pricesseen for this product range? (e.g. are the top-10 Playstation games (c£35) substitutes foreach other? What effects do discounted olderPlatinum games (c£20) have on the relevantmarket)?

• Intended use: What will the consumer usethe product for? (e.g. in India, many top-loading washing machines were purchasedto make Lhassi.)

• Income elasticity of demand: Similar to priceelasticity of demand. Will demand for theproduct change relative to consumerincomes/perceived consumer incomes(consumer confidence regarding vacations,hotel rooms, car purchases, etc.) Thecalculation here is:

% change in demand

% change in income

34

• Structural issues: How easy is it forconsumers to actually substitute betweendifferent products? This question oftenfocuses attention on distributionmechanisms. Key aspects include thephysical manner of sale (e.g. size of shop,nature of sales process (catalogue/retail),presence of tight selective and exclusivedistribution, SED) and the way the product is distributed (e.g. SED, nichemarket).

• Company intention: Are firms marketingproducts as substitutes or is a firm trying tomarket a product as a substitute for aproduct it currently is not a substitute for?(e.g. the attempt in the 1990s by Volvo toreposition itself as a competitor for BMW.)

What’s needed: Time in substitutability assessments

While the need to take account of spatialfactors has been addressed with somesuccess, treatment of time in competitioninvestigations has been less successful. Inthis section, we propose indicators to helpregulators take better account of consumervaluation of time in investigations.Collection of some basic market data,combined with some relatively straightfor-ward consumer research, will illuminate therole of time in the issue of substitutability.The aim of the indicators is to highlight the likelihood that consumers will substitute products and to identify limitsthat factors in the market will place on thatlikelihood.

A particularly important element of any substi-tutability assessment is the issue of time. Overwhat sort of time period should behaviour beassessed? This question is enormouslyimportant in markets based on high technologyproducts and services. When assessing howlong the hypothetical monopolist is able toraise prices, the question has to be asked: “How long is ‘how long’?”

The answer must comprise the followingassessments:


• What is the current pattern of consumerbehaviour in the market? This must includeassessment of its cultural characteristics (e.g.Japanese consumer electronics market)

• How quickly has this market responded inthe past to changes in service/productprovision? What is the division of consumersin this market? How many are earlyadopters? gradual adopters? late adopters?

• How important are early adopters? – Will they drive down later prices for later

adopters?– Is their number sufficient to ensure

successful launch of products?– Are consumers serial early adopters,

wherein the adoption of one new productwill quickly be followed by adoption ofanother (e.g. as in the effect of potentiallaunch of Playstation II on Dreamcast sales).

– How much product loyalty is there in themarket? (e.g. impact of Sony Playstation onNintendo and Sega).

• The essential question to be addressed is:Does the consumer behaviour in questioninvolve changes over time that seem out-of-kilter with existing patterns of behaviour inthe market? 12

Time

• What is the consumer valuation of time inthis market?– Is the market in question important to the

consumer? (e.g. car purchase, housepurchase)

– How quickly can the consumer grasp thebasic information needed to make adecision? (e.g. car performance vs. pension performance)

– How quickly can the consumer undertakethe process to carry that decision through?(e.g. car purchase vs. pension purchase)

– Is this time period artificially inflated as abarrier to entry? (e.g. financial services)

• How long will it take consumers to assess thesubstitutes on offer?– What is the process by which consumers

become aware of substitutes (e.g.advertising/leaflets)

– What is the existing pattern of informationin the market? (e.g. visual, verbal, leaflet,poster, performance based)

– What opportunities are afforded toconsumers to allow them to gather

35

information on the performance ofproducts/services?

• How long does this process take relative tothe likely valuation of time by the consumer?

Consumers value time relative to the expectedgain from the decision.

Working time into assessmentsThe matrix of questions aimed at identifyingthe degree to which consumers will behave inthe classically rational model of behaviourincludes a number of measures to take timeinto account in competition investigations.However, there appears to be a need for a moretargeted means of assessing the potential role oftime in the process of substitution. Such ameasure combines market data with asubjective ranking (which can be based on areal figure) and weighted according to thedegree to which a decision is likely to bedirected or constrained by competingdistribution channels.

The key measures are:

• Number of competitors in the market. Thisindicator can be easily drawn from a simplehead count of players in the market, limitedby the geographic market definition carriedout previously.

• Number of competitors that the consumer isaware of. This measure can be ascertainedthrough opinion polls conducted to ascertainthe awareness of consumers to players in amarket. Another method is to conduct theVon Thunen circle analysis to indicate thenumber of firms a consumer has physicalaccess to. This would indicate the number offirms from which a consumer can actuallychoose.

• Time required to understand theinformation provided. This figure can beascertained from a simple reading of theinformation necessary to understand theproduct or service concerned. In a pensiondecision, for example, one can simply takethe documents provided as part of the salesprocess and log the time needed to read andunderstand. Similarly, in the purchase of acar, one can measure the time needed to read


documentation, carry out a test drive andread a magazine review. While the accuracyand uniformity of this information will differby market and is subject to some degree ofsubjectivity, it provides an indication of theamount of work a consumer will have to putinto a market prior to purchase.

• Importance of potential savings. The scalemeasure used to indicate the importance of asaving can be based on calculation. Forexample, one could calculate the size of theprice spread within a market as a percentageof the average price of the product (e.g. a 4oz burger may differ in price by 2 pence on atotal spend of £1, whereas the price spreadon a pension may run into hundreds ofpounds on a spend of thousands). Thismeasure would indicate the degree to which'shopping around' would be worthwhile.However, such a measure would probablyhave to be supplemented by a calculation ofthe share of the average price of theproduct/service as a percentage of averagedisposable income. Thus, the relative pricespread for a car may be low in percentageterms, whereas the actual value of the pricedifference as a share of disposable incomemay be high.

• Degree of vertical integration. Theweighting factor tacked on as the lastquestion to be asked is based on theassumption that vertical integration tends toincrease the time taken to substituteproducts. For example, a consumer wishingto shop around for a hatchback car, cannotvisit one site to access all models. He or shehas to visit a number of geographicallydispersed sites to view vehicles produced bymany manufacturers, which increases thetime needed to make decisions. If one islooking for a certain shade of lipstick, thecontrol that cosmetics houses have over theirdistribution requires the consumer to visitdifferent stands within an 'official' retailerrather than view all product categories in oneplace. We have thus chosen to weight theexistence of the vertical integration as amultiplier of the time required to make adecision.

36

The relationship between the factorsThe overall figure is arrived at by multiplyingthe number of competitors the consumer isaware of by the amount of time required tounderstand the necessary information providedby each company. This is then multiplied by theimportance of the decision to the consumer andthen multiplied by the weighting factor of theexistence of vertical integration.

The original figure, which indicates the numberof actual competitors in the market, is a usefulstarting point from which to assess the knownnumber for consumers. A calculation based onthe awareness of competitors as a share of totalfirms in the market is also useful. For example,in a market with many hundreds of potentialsuppliers, the number of those firms thatconsumers are aware of is an important piece ofinformation in itself. In markets where largenumbers of firms operate but few are able toreach consumers, one tends to find proxiesoperating for consumers. This is the case in thefinancial services industry, for example. In suchmarkets consumers tend to base their decisionson proxy measures or proxy firms. Consumerswill tend to rely on branding as an indication ofperformance or on intermediaries who canguide them through the choices available.

Demand-side complementarities

This assessment involves looking atproducts/services related to the product inquestion (e.g. bread/butter, CD players/CDs,game consoles/games) where an increase inprice in one adversely affects demand for the


Consumer evaluation of time

How many competitors are there Numberin the market?

How many competitors is the Numberconsumer aware of in the market?

How much time does it take to Hoursunderstand information provided?

How important will a saving be? Scale of 1-3

Is sales process vertically 2=yes, 1=nointegrated?

37


Worked example: Consumer evaluation of time

Consumer evaluation of time Pension Cars Vacation Tin of beans

How many competitors are there in the market? Number 100 40 8 6

How many competitors is the consumer aware of in the market? Number 50 20 4 2

How much time does it take to understand information provided ? Hours 1 1 3 0.01

How important will a saving be? Scale 1-3 3 3 2 1

Is sales process vertically integrated? 2=yes, 1=no 2 2 2 1

Scale 33.33 13.33 12.00 0.02

complementary product. (e.g. the increase inthe price of petrol may have an effect on thedemand for cars that consume large amounts of petrol). The key questions in this estimation include:

• Are the products physically complementary?(e.g. cars and petrol, CD players and CDs)

• Are the products/services perceived byconsumers to be complementary? (e.g.TV/phone/Internet bundles)

• Are the products/services bundled ascomplementary products/services? (e.g.TV/phone/Internet bundles)

Supply side substitution

The issue of supply side substitution relates towhether a small increase in price would inducea new firm (operator) to enter the market, thusdeterring the incumbent from raising its price.For example, if a firm produces work boots andnotices that a firm is hiking prices in walkingboots, it may enter the market relatively easilyas it already produces a similar product. Thekey elements of the supply side substitutabilityassessment relate to:

• Is the product/service made with similartechnology that other producers/servicesuppliers have access to?

• Is the product/service a complementaryproduct to one already made? For example,is the product made alongside others (e.g. jetfuel) for which costs are shared in theproduction?

The issue of supply side substitutability raisesthree important issues:

• How easy is it for a firm to switchproduction into the different/associatedproduct?

• How quickly can that firm switchproduction?

• How willing will a firm be to enter thatmarket?

The answers to these questions raise importantissues we will address later. In particular, thesupply side substitutability raises issuesrelating to:

• Contestability• Barriers to entry (both real and perceived,

domestic and foreign)• Sunk costs• Reaction time of incumbents• Strategic behaviour of incumbents

(e.g. predatory pricing)

38


Part III: Calculating marketshares

Competition regulators have developed anumber of tools to help them to weigh therelative size and strength of competitors inany given market. This section discusses therelative merits and demerits of each schemeand supplements the more accepted toolswith less often used indicators of marketdynamism. It looks at means ofincorporating the possible impact of externaltrade on market share calculations, defines aseries of potential scenarios (verticallyintegrated, large number of small playersand small number of large players) andassesses the potential impact of each onstandards in worked examples.

How it is currently done

Once you have a sense of how the consumerbehaves, what the market is and what sort ofmechanisms operates within it, you must gaugehow market power is distributed within thatmarket. The most straightforward way to dothis is with market share calculations. There are essentially two options for looking atmarket shares:

• Volume data• Value data

Volume data is often the fall-back position forconsumer associations simply because it ismore easily available than value data.However, it is definitely a second-best solutionand is best employed in conjunction with valuedata, for the following two reasons:

• Volume data can mask the nature of themarket. A firm may have 30% of the volume

of business but only 10 % of the value of themarket. This information is vital in assessinga merger.

• Volume data can give false readings ofmarket share:– they are more easily manipulated to give

the appearance of sales where none exist(e.g. pre-registered cars);

– they can hide stocks at retailers and thus thereal pattern of purchasing by consumers.

– Volume data presents particular problemsin vertically integrated firms.

Volume data is a useful addition to value databut it should be used in isolation only where:• Value data is not present.• Products/services in question are basically

identical (e.g. passenger numbers on airlines).Here value/volume data should look similar.

Doing the sums

There are essentially two methods ofcalculating market share:

• Concentration ratio• Indices

The concentration ratio is the most straightfor-ward. It simply involves adding together themarket shares of (CR) the top 4, 6, 8, 10, firms.Usually firm market shares are summed to thetop 4, 5 or 10.

Assuming a market with seven firms, one cansee from the table below that the sum of themarket shares of the top four firms leads to aCR4 figure of 80%. This means that the top fourfirms have 80 % of the market in question.

Market share % CR4Firm A 35 35Firm B 25 60Firm C 10 70Firm D 10 80Firm E 10Firm F 5Firm G 5Total 100

Baldwin (14) has developed a useful estimationof a series of market definitions based on the

39


CR4 measure. They are:• Highly concentrated oligopoly (75-100)• Moderately concentrated oligopoly (50-74.9)• Slightly concentrated (or low-grade)

oligopoly (25-49.9)• Atomism (0-24.9)

The main drawback with the CR4 share is thatit fails to:• Inject any dynamism into the calculation

(unless time series data are used);• Ignores the remainder of the market.

A partial solution to this problem lies in the useof indices. Two closely related indices are incommon use:

• Herfindahl Index• Herfindahl-Hirschman Index

Both indices involve calculations of marketshare based on the squaring of a market sharefor each firm and a summing of that squaredmarket share to cover the entire market. Wherethey differ is essentially in where the decimalpoint goes. The Herfindahl Index uses figuresthat sum to 1 and the Herfindahl-HirschmanIndex uses percentages that sum to 100. Bothindices have the advantage over aconcentration ratio in that they sum the totalmarket, not just a subset. Importantly, the HHIalso gives greater weight to the market sharesof those firms with larger market shares, asituation that more correctly reflects the greatermarket power afforded to such firms.

The Herfindahl Index has one major advantageover the more commonly used Herfindahl-Hirschman Index – a derivative of the HI cangive a snapshot of the number of effectivecompetitors in a market at any given time.

Column 1 Col 2 Col 3 Col 4Herfindahl Index Market share HI HI A+B HI E+FFirm A 0.35 0.1225 0.36 0.1225Firm B 0.25 0.0625 0.0625Firm C 0.1 0.01 0.01 0.01Firm D 0.1 0.01 0.01 0.01Firm E 0.1 0.01 0.01 0.0225Firm F 0.05 0.0025 0.0025Firm G 0.05 0.0025 0.0025 0.0025Total 1 0.22 0.395 0.23No. of effective competitors 4.5 2.5 4.3

This is useful in merger assessments as itindicates the potential competitive loss from anagreement. As always, the indices are bestshown in tabular form. In all tables, the firstcolumn indicates the current market shares ofeach firm in the market (A-G). Column 2 showsthe initial Index measure and the sum of thatindex. Column 3 shows the potential effect of ahypothetical merger of firms A and B. Column4 shows the potential effect of a hypotheticalmerger between firms E and F.

While the Herfindahl-Hirschman Index showsthe same result in terms of the total effect oncompetition, the Herfindahl Index gives auseful measure of effective competition. Themeasure of effective competition is arrived atby inverting the final measure of concentration(the Total line); that is, by running thecalculation as one over the sum:

No. of effective competitors = 1/0.22 = 4.5

In the example below (based on ourhypothetical market), the first mergereffectively reduces the number of firms in themarket by two (despite actually only removingone) and the second merger only involves theeffective loss of 0.2 of a competitor. Anyregulator or consumer activist would thus beworried about merger 1 but not merger 2.

The Herfindahl-Hirschman Index is the moreregularly used of the two indices (see page 40).The HHI is arrived at by summing the squares ofmarket shares (expressed in percentage terms).The HHI can reach a maximum of 10,000 (100x100) for a monopolist. The advantage of the HHIis that it gives a usable index figure and allowsfor an assessment of the increase in concentrationcaused by a particular merger proposal.

40


Herfindahl Hirschman Index Market share HHI HHI A+B HHI E+FFirm A 35 1225 3600 1225Firm B 25 625 625Firm C 10 100 100 100Firm D 10 100 100 100Firm E 10 100 100 225Firm F 5 25 25Firm G 5 25 25 25Total 100 2200 3950 2300

If one uses our hypothetical market again, onecan see that the market starts with an HHI of2200. The increase in concentration caused by themerger of firms A and B is 1,750 (a very largeincrease in concentration), while that for firms Eand F is relatively small, at 100 points.

The 1992 Federal Trade Commission HorizontalMerger Guidelines (amended in 1997 with newsection on efficiency defences) established auseful set of rules-of-thumb for assessingchanges in concentration in merger cases usingthe HHI (13). To quote Section 1.51 GeneralStandards in full:

In evaluating horizontal mergers, the Agencywill consider both the post-merger marketconcentration and the increase in concentrationresulting from the merger. Marketconcentration is a useful indicator of the likelypotential competitive effect of a merger. Thegeneral standards for horizontal mergers are as follows:

a) Post-Merger HHI below 1000. The Agencyregards markets in this region to beunconcentrated. Mergers resulting inunconcentrated markets are unlikely to haveadverse competitive effects and ordinarilyrequire no further analysis.

b) Post-Merger HHI between 1000 and 1800.The Agency regards markets in this region tobe moderately concentrated. Mergersproducing an increase in the HHI of less than100 points in moderately concentrated marketspost-merger are unlikely to have adversecompetitive consequences and ordinarilyrequire no further analysis. Mergers producingan increase in the HHI of more than 100 pointsin moderately concentrated markets post-merger potentially raise significant competitiveconcerns depending on the factors set forth inSections 2-5 of the Guidelines.

c) Post-Merger HHI above 1800. The Agencyregards markets in this region to be highlyconcentrated. Mergers producing an increase inthe HHI of less than 50 points, even in highlyconcentrated markets post-merger, are unlikelyto have adverse competitive consequences andordinarily require no further analysis. Mergersproducing an increase in the HHI of more than50 points in highly concentrated markets post-merger potentially raise significant competitiveconcerns, depending on the factors set forth inSections 2-5 of the Guidelines. Where the post-merger HHI exceeds 1800, it will be presumedthat mergers producing an increase in the HHIof more than 100 points are likely to create orenhance market power or facilitate its exercise.The presumption may be overcome by ashowing that factors set forth in Sections 2-5 ofthe Guidelines make it unlikely that the mergerwill create or enhance market power orfacilitate its exercise, in light of marketconcentration and market shares.

The guidelines developed by the FTC are auseful starting point for assessments of mergersand of possible accruals of market power.

Step down differencesbetween market shares

Of equal importance in any market sharecalculation is the assessment of the differencesbetween market share of incumbents. Thus if amarket has one or two very large competitorsand a large number of very small competitors(as does our hypothetical market, above), themanner of operation of that market will besignificantly affected by restraints oncompetition. The relative size of the secondstring competitor is important. The differencesin market share can be easily modelled. One can:

41


• sum the Herfindahl or Herfindahl-Hirschman Index for the second ordercompetitors;

• calculate the marginal-concentration ratio ofthe second group of four firms;

• calculate the size of the second group of fourfirms to the first group of four firms.

Health warning I: It should be noted that thesedata are only really useful on either a:

• significant time series basis: 10 years of dataare always a useful starting point

• comparative basis, either across industries,markets or countries.

Health warning II: These data on relativeconcentration need to be treated with cautionwhen used in comparative mode and on a timeseries basis. Evidence of an increase inconcentration/widening of gap is evidence ofnothing in isolation from an understanding ofthe dynamics of the specific market and ofmarkets being used in comparison.

Temporal aspects toconcentration

One of the key drawbacks to an index-basedapproach to measuring market power is thefact that such measures tend to be ‘snapshot’measures and lack a dynamism that wouldhelp to explain a market. This can be solved byusing time series data. Such data would includea number of years’ worth of market share data(with all firms involved at each stage of themarket’s evolution.) The key advantage to sucha calculation is that it allows for a more fluidanalysis of the market and a betterunderstanding of the degree of turnover in a market, and thus the likelihood of entry and exit.

What’s needed: Impact ofexternal trade

There is a real problem in dealing with cross-border trade in goods in measures of marketpower and market share. When looking attrade data, we must ask:

• Is the trade registered as trade between partsof a firm or independent operators?

• Is the import purchase market oligopolisedor monopolised? (with gains from tradecaptured by intermediaries)

• What is the pattern of market structure in theexporting country? Is the exported good:– exported by an export cartel? (e.g. Voluntary

Export Restraint on Japanese vehicles)– exported by a very large number of very

small producers? (e.g. leather goods)– produced under contract by importing

firms?• What are the regulations governing the

traded goods?– Do they have a high ad valorem tariff?– Are there quantitative restrictions on the

import of the good?– Do other regulations effect the cost of

the goods?

All these factors must be taken into accountbefore injecting an assessment of trade into amarket share calculation. In doing so, a numberof scenarios exist to factor in the potentialmarket share of trade. The following optionsare available:

• Vertically integrated trade: This should bereflected in the existing market share dataand will have no effect on the share allocatedto firms already operating there. Forexample, if the UK imports 30% of itspetroleum through existing verticallyintegrated operations, the effect on marketshare is zero because of the vertical nature ofthe firms in question. – Competitive effect – very small– Competitive threat (for contestability

assessments) – very small/zero

• Independent trade from a large number ofsmall exporters: The effect on existing marketshare is likely to be small as exporters into themarket are unlikely to accrue a good deal ofmarket power. For example, one can look atthe import share of independently producedtoys and shoes from developing countries as ashare of total imports in those sectors.– Competitive effect – very small– Competitive threat (for contestability

assessments) – very small/potentiallymedium

• Independent trade from a small number oflarge exporters: Depending on the number ofexporters, this can be dealt with as a normalmarket share calculation. It is, however,necessary to understand how exporters feedtheir products into the market: is it a standalone strategy? does it involve intermediaries(in which case the competitive effect isabsorbed by others)? does it involvefranchising/licensing (again transferring/absorbing the market effect by others)? Onecan look at the purchase of petrol on the spotmarket by supermarket retailers as a marketwhere a small number of importers purchaseexported oil. Here the competitive calculationrests on sales of petrol by supermarkets, notby exporters into the UK– Competitive effect – medium – subject to

mediation/capture– Competitive threat – medium – can be

leveraged into significant scale undercertain circumstances.

Let us return to our example of the theoreticalmarket (seven players, highly concentrated). Ifwe introduce the three possible trade impacts(vertically controlled, small exporters, large

42

exporters), we can see the effect that analteration in methodology can bring about.

In example one (above), we have redistributed5% each from Firm A and B to importcompetition and re-run the scenarios. The initialreallocation of market share has significantlyreduced the pre-merger HHI. However, becausethe number of exporters into the market arelarge and their individual market shares aresmall, the potential constraining effect of theseplayers on the post merger A+B is very smalland the likelihood of the exporters affecting thismarket is low.

If, however, one redistributes the market sharetaken by the multitude of smaller exporters andallocates it to a single exporter (below), thedegree of concentration increase significantly.But one is also potentially limiting the impactof any mergers in second order firms.

Problems with indices

Concentration measures pose a number ofproblems but these should not be overstated.


Large number of small independent traders

Herfindahl Hirschman Index Market share HHI HHI A+B HHI E+FFirm A 30 900 2500 900Firm B 20 400 400Firm C 10 100 100 100Firm D 10 100 100 100Firm E 10 100 100 225Firm F 5 25 25Firm G 5 25 25 25Import agents 10 1 1 1Total 100 1651 2851 1751

Single independent exporter into market

Independent trade/one supplierHerfindahl Hirschman Index Market share HHI HHI A+B HHI E+FFirm A 30 900 2500 900Firm B 20 400 400Firm C 10 100 100 100Firm D 10 100 100 100Firm E 10 100 100 225Firm F 5 25 25Firm G 5 25 25 25Import agents 10 100 100 100Total 100 1750 2950 1850

43


The drawbacks to the use of indices have moreto do with their possible abuse rather than theirintrinsic utility. As with all indicators, the indicesand concentration measures should never bestretched beyond their breaking point in termsof their explanatory powers. A concentrationmeasure on its own proves little and shouldalways be used in conjunction with more workon the operation and nature of the market.

• The measure is largely static.• The measure assumes that a post-merger firm

will have the same market share as the sumof its parts does prior to merger.

• The measure of concentration presumes alink between greater concentration and a lossof competitive pressure and a loss inconsumer welfare.

Part IV: Assessing verticallyintegrated markets

The importance of vertical restraints ontrade has waxed and waned in the post-warregulatory world. The pattern of leniencetoward vertical restraints seen throughoutthe 1980s and ‘90s seems to be fading.Consumer organisations have long takenthe view that the anti-competitive effects of vertical restraints have beenunderestimated. For consumers, one of thedefining problems for regulation of the 'neweconomy' rests on a more sceptical view ofthe operation of vertical restraints acrosscurrently diverse markets. We propose that regulators should dust off themethodologies developed (but repudiated)in the US and apply them rigorously incompetition cases.


Vertical integration is increasingly being seenby regulators in a relatively benign light. Therecent European Commission reforms to therules on vertical restraints have set thefollowing twofold definition.

1. Vertical restraints in a firm with less than30% market share will be presumed not to beharmful, unless;

2. More than 50 % of an industry operates thesame pattern of vertical restraint.

This combines the ‘market share’ test with the ‘network’ effect test and is remarkablysimilar to efforts in the US to deal with theproblem of vertical market restraints manyyears ago.

44


The US developed a series of guidelines forassessing vertical restraints on a similar basis tothe horizontal merger guidelines. However,these were repudiated under the Bushadministration and failed to make a comebackin the Clinton administration. The repudiatedguidelines were nonetheless extremely useful inoutlining the consensus view of how to dealwith vertical restraints under competition law.The view enunciated in the guidelines waspositive about the potential benefits of verticalrestraints.

The relationship between vertical restraints andcollusion were neatly bounded by theguidelines by placing upper and lower limitson the likelihood that collusion would arisefrom vertical restraints. The guidelines statedthat vertical restraints were unlikely to aidcollusion unless three conditions were found:

• concentration is high in the primary market; • firms in the secondary market using the

restraint, account for a large portion of salesin that market; and

• entry into the primary market is difficult.

Conversely, for exclusive dealing to lead toanti-competitive exclusion, the followingmarket conditions must normally be found:

• the ‘non-foreclosed market’ is concentratedand leading firms in the market use therestraint;

• firms subject to the restraint control a largeshare of the ‘foreclosed market’; and

• entry into the "foreclosed market" is difficult.’

The two stage process

Stage oneThe US Department of Justice (DoJ) guidelinesfollowed a two-stage process that involved‘screening’ a market for likely problems causedby vertical restraints and then applying a morerigorous approach to those that did indicate aproblem. In stage one, the DoJ was to apply themarket structure screen aimed at excluding thefollowing circumstances:

• the firm employing the restraint has a shareof the relevant market of 10 % or less; or

• the Vertical Restraints Index (VRI) is under1,200 and the coverage ratio is below 60 % inthe same (e.g. supplier or dealer) relevantmarket; or

• the VRI is under 1,200 in both relevantmarkets;

• the coverage ratio is below 60 % in bothrelevant markets.

Vertical Restraints Index and theCoverage RatioThe Vertical Restraints Index is the verticalcounterpart of the Herfindahl-Hirschman Indexand was developed in the guidelines as asimilar rule-of-thumb system. With therepudiation of the guidelines, the VRI died as atool of analysis for the DoJ. However, the toolitself is quite useful and there is no reason whyit cannot be applied in competitioninvestigations involving vertical restraints.

To quote from the guidelines:

The Vertical Restraints Index (VRI) is calculatedby squaring the market share of each firm inthe market that is a party to a contract or otherarrangement that contains the vertical restraintand then summing the values obtained forfirms at the same level of operations.

For example, if only two firms in a dealermarket employ a restraint, one with a 5% andone with a 20% market share, the dealer marketVRI equals:

52

+ 202

= 25

+ 400 = 425

If four suppliers, each with a 25% market share, employ a restraint, the supplier marketVRI equals:

252

+ 252

+ 252

+ 252

= 625 + 625 + 625 + 625 =2,500

If all firms in the relevant market use therestraint, the VRI is equal to the HHI used inmerger analysis. The maximum possible valueof the VRI is 10,000, achieved when there isonly one firm in a market and that firmemploys a vertical restraint. The VRI reflectsboth the distribution of the market shares offirms using a vertical restraint and the extent towhich it is used in the relevant market.

In essence, the VRI is the HHI in that itassumes that the vertical restraint employed bya firm is uniform and comprehensive. Like themobility measurements outlined above, there isnothing to stop analysts from applying the VRIto different vertical restraints in a market andassessing the likely impact. It would beinteresting to assess the partial application of avertical restraint using the VRI. For example, ifa jeans manufacturer has 25% of a market butonly operates a vertical restraint on 50% ofdistribution, this would indicate a problem forthe VRI. The problem relates to intrabrandcompetition (competition between suppliers ofa single brand). The problem arises because theDoJ (and most other competition regulators)have an ideological disposition to discount theneed to encourage intrabrand competition andto concentrate instead on interbrandcompetition (competition between suppliers ofdifferent brands).

The problem of dealing with the VRI on itsown was partially solved by tying it to theCoverage Ratio. This involved looking at theVRI for the companies in question and thenassessing the resultant figure in relation to theCoverage Ratio. The Coverage Ratio was asimple measure of the share of the market thatwas covered by the same, or similar, verticalrestraint. Thus, if two firms with 25% of themarket each were merging and seeking toapply the same vertical restraint, but they werethe only firms applying such a restraint, thecoverage ratio would be the same. However, iftwo firms with 10% of the market each soughtto merge and apply the same vertical restraint,but the remainder of the market also applied arestraint, the coverage ratio would be equal to100%. This would easily fail on the network testapplied by both the EU and the US.

The car market in Europe is a classic exampleof where the application of the VRI andCoverage Ratio would work. No one firm hassufficient market share to overly worry anyoneunder a VRI assessment. However, theCoverage Ratio indicates that the entire marketis covered by the same restraint – a muchclearer indication of a problem. What isinteresting is the effect that any one merger oralliance has on the application of theVRI/Coverage Ratio.

45

Part IV: Assessing vertically integrated markets

Stage twoIf an agreement fails to pass muster under themarket structure screen, the market wouldproceed to Stage Two. This stage involved whatthe DoJ referred to as a ‘structured rule ofreason’ approach. This was a clear signal thatthe DoJ were soft-pedalling on vertical restraints,given the fear of business lobbyists that a ‘per se’prohibition rule would be applied. The StageTwo process involved a balance of benefits anddisbenefits from the vertical restraint in themarket. In a strange twist of fate, this approachcame perilously close to the operation of Article81(3) under European competition law – wherean otherwise anti-competitive agreement couldbe allowed, provided it met some fairlyanodyne, and rarely applied, criteria.

The key criteria to be applied in Stage Twoinvolved an assessment of:

• Barriers to entry in both the upstream anddownstream markets;

• Is the product homogenous or heterogeneous?what is the nature of product competition? (inthe former, collusion is more likely)

• What is the history of collusion in the market?• How exclusionary is the restraint? Is it an

exclusive supply agreement?• Is there an indication of the intention of the

parties applying the restraint?• Are the restraints being applied by small

firms or new entrants as a means of getting atoehold in the market?

• Are there easily identifiable pro-competitiveefficiency gains from the restraint?

Tying

The DoJ also developed guidelines for theanalysis of tying agreements. This is anagreement in which a supplier only suppliesone product in return for another product beingcarried as well (e.g. only supply 501s if 602s aresold as well). This can also be referred to as‘full line forcing’. The judicial treatment of suchrestraints has tended to be lenient (as with mostother vertical restraints). The DoJ guidelinesproposed that tying arrangements be subjectedto the following tests:

• Does the firm have market power? (e.g. over30% of the market)

• Does this power translate into dominance?

46


Part V: Mitigating factorsin assessments ofmarket power

Simple snapshots of market power arelimited indicators of market structure. A keyweakness is the lack of dynamism in theresults that they arrive at. To more fullycapture the essence of a market, one has tobuild in ideas of movement and innovation.This can be done in a number of ways,including basic measurements of mobilitywithin markets. The potential for otherplayers in the market exercising a form ofcountervailing power must also be takeninto account. This section maps the majortools available to take account of mobilityand countervailing power and highlightsissues involved in assessing technologicalinnovation in a market.


Once the market under investigation and itsdegree of concentration has been identified, it isnecessary to look at the factors that help todefine that market and the likelihood that themarket will change. In particular, one mustlook at:

• sources of countervailing power• technological innovation in the market• threat of successful entry by new players

There is a definitional problem awaitingcompetition analysts in looking at barriers toentry in isolation. As we will see below,barriers to entry can also be observed asbarriers to establishment. As John Baldwin 14

has noted: ‘Analysts in the field of industrialorganisation have often devoted specialattention to modelling the effect of entry.Implicit in these exercises is the notion that the

disciplining influence exerted by entry isdifferent from internal pressures that comefrom incumbent firms. If entry and exit are justmarginal manifestations of a general turnoverprocess at work, then there is little need to treatthem separately.’

What is probably more useful is to analysemarket mobility.

The relative stasis of concentration measureshas prompted many economists to argue thatconcentration measures on their own do notindicate industry mobility. If, as we arguebelow, mobility is preferable to entry and exitas a means of assessing markets, there must besome means of assessing mobility. The problemwe face is access to data. Dependent onavailability of data, at least four approaches are possible to enable the measurement ofmobility:

• changes in market share between incumbents• degree of change in market rankings• extent to which market share for firms

regress to the mean market share• extent of inter-group mobility.

Using the Dissimilarity Index, one cancalculate the redistribution of market share forfirms over a time period. The index can be usedin complex monopoly cases where long-runtrends are needed or where incumbents aretrying to indicate how much market share theyhave lost to new entrants. The Index iscalculated by taking two years some distanceapart (e.g. 1999 and 1989). The market share ofeach firm in the latest year is taken and themarket share in the first year is then subtractedfrom that figure. The resulting figure is dividedby two. When all firms are included, the totalfigure will vary between 0 and 100 and willgive an idea of how much movement there hasbeen in the distribution of market shares. TheIndex can be further nuanced by extracting outthe market share changes attributable to newentrants, exits and mergers.

Dissimilarity indices can also be adapted forgroups or sub-groups of firms. For example,dissimilarity indices can be calculated forgroups of firms ranked according to size (adissimilarity index for CR4, CR5-8 etc). This

47

downstream market power. The authors (16) ofthis report argue that buyer power can bemanifested in the following forms:

Slotting allowances: These are payments, forexample, to a supermarket for the right to haveone’s goods on display in a particular place onthe shelves, or even to have them on display at all.

Exclusive distribution: Exclusive or soledistribution agreements can be used byimportant buyers to extract concessions fromtheir suppliers.

Conditional purchase behaviour: This is thepurchase of goods only on condition thatsignificant concessions are made by thesupplier of such goods. Two sub-cases spring tomind. The first is where a purchaser gets into adominant position with a supplier such that thepurchaser’s business becomes of vitalimportance to the supplier, at least in the shortto medium term ... The second sub-case iswhere a purchaser will only buy on conditionthat other outlets are not supplied with theproduct, or not supplied with a precise versionof the product.

Exclusivity contracts: …retailers …mayconcern the desire to foreclose the market (andthereby allow for higher downstream prices tothe potential benefit of both parties)…Alternatively, the explanation may be in termsof dealing with vertical externalities whichresult from successive independent behaviour.

Cloning behaviour: behaviour aimed at ‘free-riding’ on the investment of established playersin the market

Joint marketing: A powerful retail chain mayencourage a manufacturer to engage in a jointpromotion exercise whereby the manufactureroffers concessions only to purchasers at aparticular store chain.

Predatory buying of inputs: …it is conceivablethat a dominant buyer may seek to expand itspurchases, driving up factor prices to the pointwhere a rival is unable to continue sufferinglosses, due to high costs, and so leaves themarket.

Part V: Mitigating factors in assessments of market power

can be used to identify the degree of marketshare reallocation within different groups of firms.

A further measure of mobility can be calculatedby dividing the ranked list of firms (by marketshare) into quintiles (five groups) and assessingthe degree to which firms moved betweenquintiles over a set period. This gives areasonably clear picture of firm growth/declinein an industry and can aid the assessment ofmarket stasis or vibrancy.

Despite the relatively straightforward nature ofthe dissimilarity index as a first step in themeasure of mobility, it is not frequently used incompetition investigations. This is a shame,given the relative stasis of the HHI calculation.

Countervailing power

The study of markets at the end of the 20thcentury made great play about the emergence ofcountervailing powers in consumer productmarkets. In particular, the strength of mass-market retailers became a subject of muchanalysis. In terms of understanding specificmarkets, one has to ask the following questions:

• How many levels does the market operatethrough? How many linked markets arethere between first producer and finalconsumer? (While similar to the chain of sub-stitutability, this focuses more on the chain ofproduction, distribution and consumption.)

• At what stage of the chain is the marketlocated? Is it upstream, downstream orintermediate to other markets?

• What other players exist in the market atother stages of the production, distributionand consumption chain? What degree ofconcentration exists at each stage of thisprocess?

Recent work on the emergence ofcountervailing power is relatively slim.However, one study for the OFT has provided apotential checklist for practices in sectors whereretail countervailing power might exist. Thedegree to which the following factors operate isthe degree to which upstream agglomerationsof market power may be counteracted by

48


Strategic purchasing of facilities: Control of anessential facility for distribution is a generalfeature of privatised utilities. However, keyfacilities for distribution arise in otherunregulated industries which may allow forbuyers to exercise power by controlling market access.

Reciprocal dealing: Reciprocal dealing involvesa monopsonistic buyer of some productagreeing to purchase from a specific seller oncondition that the seller also buys a productfrom the buyer.

Terms of business: …potential strategicpractices which accommodate rivals wherebybuyers act together to improve their mutualposition against their suppliers… a more likelycandidate for such behaviour would be‘standard terms of business’ which, while notbeing mutually agreed in any specific way,nevertheless become broadly adopted asindustry practice and therefore tacitly agreedbetween buyers.

Technological innovation

The issue of technological innovation is adifficult one to map for competition analysis. In high technology markets, the lure of the‘innovation’ defence is a major one whenattempting to accrue market power. One of themajor drawbacks in this field is that thedefence/analysis rests on an assessment offuture developments in a sector. As such, it isprone to inaccuracy and crystal ball gazing. It isalso a market in which entry costs are very low– one of the reasons that industry journalistsare prone to criticise competition regulators formeddling in markets that they ‘do notunderstand’. The criticism is often characterisedby the potential impact of the ‘next big thing’and is usually wrong-headed. In analyticalterms, one must try to pin innovation in amarket down to some form of measurableoutcome. The following questions might help:

• Is the market a mature one? (e.g. cable TV)• Is the market part of a wider market? (e.g.

cable telephony/Internet access)• Is the market part of a wider market that is

converging in delivery and content?

• Is the market characterised by high R&Dspending? (need to be careful with this one)

• Is the market highly dependent onintellectual property protection?

Issues of barriers to entry, establishmentand exit are of central importance to anyunderstanding of a market, as they are keyfactors in turning a static market analysisinto a dynamic cognisance of the marketunder analysis. Preferable to viewing marketentry on its own is analysis of the entry-establishment-exit sequence as part of theprocess of market mobility. This sectiondescribes various studies of entry,establishment and exit barriers andsuggests a way to calculate the impact ofsuch barriers on markets.


One of the key concepts used in the recentassessment of markets is the contestabilityapproach pioneered by Baumol, Panzar andWillig (17) and adopted by a number ofregulatory agencies. The approach has beenused to undermine straightforward staticmodels of market concentration that simplyequate concentration with a lack ofcompetition. However, the contestabilityapproach itself has tended to be underminedby the actual operation of markets and theability of incumbents to fend off competition.Contestability essentially rests on the threat ofcompetition, rather than actual competition,being a significant constraint on the abuse ofmarket power by incumbent firms. In essence,incumbents will not utilise their market powerfor fear of triggering market entry. While theargument is fairly weak for the application ofcontestability theory, the grounds on which itcan be applied are a useful test for a market.The three conditions that have to be met for amarket to be deemed contestable are:

49

Part VI: Barriers to entry, establishment and exit

Part VI: Barriers to entry,establishment and exit

• there are no barriers to entry or exit;• all firms, both incumbent and potential

entrants, have access to the same productiontechnology and there is perfect informationon prices, available to all consumers and firms;

• entrants can enter and exit beforeincumbents can adjust prices.

In reality, few markets fulfil this checklist but itis a useful one to bear in mind when assessingmarkets.

A useful model of market entry that takesaccount of a large number of factors that firmswill take into account when assessing whetheror not to enter a market was carried out by JohnBaldwin,(18). He argued that firms will normallytake into account the following factors:

• perceived post-entry profit: both existingprofit and a likely growth in profit

• barriers to entry: economies of scale,concentration, advertising intensity and R&D

• market risk: the volatility of market growth• number of firms and market size• industry growth

To this list of factors one can add:

• existence and size of sunk costs• reaction time of incumbents• nature of past incumbent firm behaviour

(reputation)• existing market regulations

The shopping list of factors governing firm entrystrategy should be placed under three headings:

• barriers to entry• barriers to establishment• barriers to exit

Barriers to entry

The study of barriers to entry is the bestdeveloped of these areas. Barriers that a firmmay face include the following:

Pre-existing exclusive supplycontractsIncumbent firms may already have access to supply contracts to firms or government.

50


Such contracts can help to offset activities inother sectors (effective subsidy) or entail asewing up of a significant part of the market.An example of the former can be seen in therelationship in the supply of military aircraftbetween the US government and Boeing.

Tests:

• Is the market characterised by a smallnumber of large contracts?

• Is the government a big purchaser of goods?

Governmental regulations andrelationshipsIncumbent firms have often been deeplyinvolved in the establishment of regulationsthat govern their own behaviour. Suchregulations can be set at an artificially highlevel in the knowledge that lower costoperators will not be able to fund such costs.There is thus an incentive on incumbents tochampion ever-higher technical andperformance standards as a means of deterring entry.

On a similar tack, incumbent firms often havewell-established relationships with governmentand regulators that makes the ability of newentrants to operate in that market difficult. Theability of British Airways to effectively haveUK international aviation policy written for itssole benefit is a clear example of this form ofregulatory capture.

Tests:

• Is the incumbent an ex or current nationalchampion?

• Are government regulations important inshaping the market?

Vertical restraintsTying relationships between existing suppliersand manufacturers can make it difficult to enteran established market. There are a largenumber of potential vertical restraints that canlimit market entry. For example, the freezerexclusivity issue still in published text was aclear effort at restraining market entry.Similarly, exclusivity links in the brewingindustry clearly acted as a barrier to entry.

Tests:

• Are vertical restraints a structural feature ofthe market?

• What is the rationale for the operation ofvertical restraints?

Access to essential facilitiesThe issue of essential facilities is complex andcontroversial. An essential facility is a one thatis needed by a firm to enable it to carry out itsnormal business activity. Such facilities caninclude such things like ports, gas pipelinesand airport slots. In any industry in whichessential facilities are present, one has toconsider:

• Is access is provided at fair and reasonablerates?

• Is capacity artificially restrained to limitentry?

• Is redistribution of existing assets necessaryas a pre-requisite to market entry?

Economies of scaleIn comparison to the more strategic factorsoutlined above, the existence of scaleeconomies is relatively straightforward. Scaleeconomies occur when a firm, through the sizeof its production, attains economic efficienciesthat allow it to undercut other firms. If anincumbent firm has attained such economies,the likelihood of successful entry by a new firmis reduced. At the least, entry is only likely byfirms either with their own economies of scaleor with deep enough pockets to allow losses tobe attained in order to attain economies of scaleover time.

Tests:

• Do incumbents possess economies of scale?• Are any potential foreign competitors that

also possess economies of scale able to enterthe market ?

• Are there potential new entrants from otherindustries with sufficiently deep pockets toattain scale economies within a reasonableperiod of time?

ConcentrationAn industry that is highly concentrated is likelyto act as a barrier to entry to a new firm.However, this barrier is potentially overcomeby two factors, both strategically driven:

51

Part VI: Barriers to entry, establishment and exit

Judo economics: this wonderfully namedtheory that rightly points out that smaller firmscan be fleeter of foot than their largercounterparts and so can gain market share, orsuccessful niches, relatively quickly.

Niche operation: to avoid undue attentionfrom competition regulators, entry may befacilitated but constrained as a means ofshowing good faith and behaviour.

Advertising intensityIt is interesting to analyse advertisingexpenditure in markets where significant entryhas not occurred and trace patterns relatingexpenditure to the potential for entry.Advertising expenditure can operate either as adefensive move (protecting brand identity frominsurgence) or offensive (flooding newermessages).

Tests:

• How important is branding in the market?• What role does advertising play in the

market?

R&DIndustries with high spending on R&D arealways difficult to break into. R&D expenditurerequires two things: innovation and deeppockets.

A close look at recent developments in thepharmaceutical industry bears this analysis out.In the 1990s, the main source of innovation andpotential innovation came from biotechnologycompanies seeking genetic therapies andmedicines. Such firms had access to significantamounts of innovation and research. Incontrast, established pharmaceutical companieshad access to little biotechnological researchbut plenty of money. When the former firmsfound that their R&D was not leading to fast-enough medical breakthroughs, they found thatpharmaceutical companies were only too eagerto offer their deep pockets as purchasers or‘strategic allies’.

Tests:

• Is this an industry that requires largeexpenditures on R&D

• Are there potential new entrants with deep pockets?

Sunk costsSunk costs are those costs incurred in enteringa market that are irrecoverable on exit. If a firmestablishes a factory with specialist equipment,it will not be able to sell this equipment onexiting the industry. Sunk costs are importantbecause they are part of the fabric of an entrydecision but have already been factored intocosts by an incumbent. A difference in costbases is thus likely to occur. However, this canbe overstated because entry is usually presagedon being a lower cost operator than anincumbent. Sunk costs are also importantbecause they have an impact on potentialprofitability and market risk.

Tests:

• Does this industry require a largeirrecoverable cost? (e.g. cable telephony)

• Have incumbents shown a tendency to actaggressively against new entrants?

• Have potential entrants indicated that theypossess deep pockets?

The summary calculation for the likelihood ofmarket entry has been summarised as:

Market entry = sum (market size+marketgrowth+expected post entry profit+barriers to entry+market risk)

Barriers to entry and market risk are bothdisincentives and the former three incentives.

Barriers to entry can be summed as the barrierslisted above but are more usually summed as:economies of scale+market concentration+advertising intensity+R&D costs.

Expected post-entry profit is modelled usuallyas the average profit of the sector expectingentry. However, it can also be modelled by thefirm as the variance between pre-entry profitand expected post-entry profit.

The market risk is usually modelled by thevolatility in market growth.

52

Identification of strategic behaviour by firmsis the most theoretically dense element ofour work. The evolution of thinking aboutfirm behaviour owes a good deal todevelopments in game theory andmathematical modelling and is thuscomplex. The importance of reputation andinformation cannot be digested at onesitting. This section looks at existingresearch and places it within the frameworkof a possible investigation, suggesting aframework for thinking and a set of targetedquestions to help identify key problems and approaches.


Strategic behaviour of firms is at the heart ofcompetition law and policy and has been adriving force behind changes in the nature ofthat law. However, the manner in which alegal/economic system of regulation can assess strategic behaviour is less clear, except in specific cases of laws aimed at excesses ofsuch activities. For example, the Robinson-Patman Act (US) is specifically aimed atrestraining price discrimination by suppliers if this discrimination is aimed at restrictingcompetition. Similarly the treatment ofpredatory behaviour has a specific behavioural aspect to the law, in its assessment of intent.

While the focus on intent provides a way in forthe study of strategic behaviour, the real abilityof regulators to separate Type I from Type IIerrors (see below) is limited by the ‘fog of war’that occurs in normal markets.


Structural barriers andmarket entry

It should be noted in dealing with realindustries that the existence of structuralbarriers may well deter entry to the market interms of number and frequency. However, it isunlikely to affect the size of the likely entrantnegatively. If anything, structural barriers makethe entry of larger firms more likely than smallones. There is also limited evidence that thelikelihood of market entry is determined by theprofitability of the smaller players in themarket. This is simply because it is likely thatany new entrant would enter the market as asmall player. This is an important practicalissue when assessing the likelihood of market entry.

Barriers to establishment

The key problem with looking at barriers toestablishment is separating out the normal cutand thrust of competitive behaviour from theclearly anti-competitive behaviour or abuse of a dominant position. The following rules ofthumb are useful:

• If the market is composed of a small numberof relatively evenly divided firms, evidenceof collusion should be looked for. Thisincludes:– outright collusion/cartelisation– tacit collusion/ unwillingness to compete

• If the market has a dominant player, oneneeds to look for abuse of that position. Thiscan take many forms, many of which areincluded above in the outline of barriers toentry. Specifically one can identify:– strategic behaviour, and particularly

predatory behaviour – tilting the playing field– market foreclosure

• If the market has a large number of smallsuppliers, collusion is less likely. (Cartelsgenerally rely on few members with areduced incentive to cheat).

Part VII: Strategic behaviour

53


When faced with the potential entry of acompetitor into a market, an incumbent firmhas a number of options. The two centralchoices are to fight all entrants or accommodatethem (see the ‘chain store paradox’, below).However, Jean Tirole and Drew Fudenberghave identified a four-fold typology of firmsand likely response (19) in a study ofinvestment in R&D and advertising as aresponse to threatened entry or expansion.

The key findings of the Fudenberg and Tirolemodel are as follows:

• Lean and Hungry: An incumbent may reduceadvertising to deter entry by signalling to acompetitor that it has reduced its ‘goodwill’in the market with consumers and isprepared to fight on price.

• Fat-Cat Effect: An incumbent mayaccommodate market entry but over-invest inadvertising as a means of softening up thenew entrants pricing behaviour and relaxinginvestment post-entry.

• Puppy Dog Ploy: A firm accommodates entryand reduces investment as a way of lookingquiescent.

• Top Dog: An incumbent increases investmentas a means of deterring entry.

The question for each market is to determinethe profile of the incumbent firm. While thetypology is not designed to be generallyapplicable, its models can be used moregenerally if they are viewed as characteristics ofmarkets rather than strategies of firms. Forexample, one can see that the Fat Cat effect iscommon in the music market as firms over-invest in advertising/PR to increase the costs ofentry of new players in order to make them lessable to compete aggressively on price.

Bertrand or Cournot?

Another useful tool for an assessment of amarket is to look at the market in relation to themodels developed in the 19th century byCournot and Bertrand. The Cournot-BertrandParadox is useful as a preliminary step inassessing markets and strategic behaviour. Itstwo approaches are:

• Cournot market: Each (of the two) firm(s)chooses its output level and the combinedoutput of the two sets the price. The oilmarket is the clearest example of a Cournotoligopoly.

• Bertrand market: Each (of the two) firm(s)uses price as the key variable and adjustsoutput accordingly.

The paradox of the original models was that aCournot market led to prices below themonopoly price but above the perfectlycompetitive market price. A Bertrand market,however, led to prices the same as thosemodelled in a perfectly competitive market.The ‘paradox’ has tended to be solved by agame theoretic application involving eitherdifferent costs or information problems. In bothcases, Bertrand markets develop Cournotsolutions: that is, markets originally developedon a price basis end up not setting perfectlycompetitive prices but prices betweenmonopoly and competition.

The key question for analysts in this area is:‘what sort of market am I looking at: Bertrandor Cournot?’ ‘Is output or price the primaryvariable?’ Answering these questions will focus down the chain at the mechanisms forsetting output and prices and the speed withwhich these decisions can be altered. Think ofthe oil market, where enormous sums areinvested in getting output organised — theprice of oil is a secondary driver to this.However, the volatility of Cournot marketprices leads to collusion to restrict supply andcontrol price. The oil market is naturally aCournot market but the OPEC cartel tries toapply a Bertrand model – and the tension is apparent.

Key questions in Bertrand-Cournot:

• Is this market predominantly a Bertrand orCournot market?

• Is this market naturally one or the other?• How quickly can the market adjust either

price or output?• If output is sticky, how likely is a collusive

attempt at Cournot by incumbents?

54


Predatory pricing

Identification of predatory pricing isnotoriously difficult. Roger Fones (19), of theDepartment of Justice, outlines a number ofprinciples that the DoJ has applied to cases ofpredation: 20

• It is not illegal predation unless consumersare worse off in the long run as a result.

• A pricing strategy by a suspected predatorharms consumers when the strategy isrational only if the victim exits the market.

• The incumbent’s prices must be ‘below anappropriate measure’ of its own costs.– Corollary A: An appropriate measure of

costs should not establish a price umbrellafor inefficient firms.

– Corollary B: An appropriate measure ofcosts should minimise the risks ofcondemning legitimate competitivebehaviour.

– Corollary C: An appropriate measure ofcost should be reasonably measurable witha high degree of confidence andpredictability.

The UK’s OFT has identified factors to be takeninto account in predation cases. 21 Although theparticular case in question involved no findingof predatory pricing, the then-Director Generalof Fair Trading did spell out the test todetermine the existence of predation. Threefactors needed to be taken into account:

• The relationship between prices and costs;• The structure and other characteristics of the

market for the product in question;• Any evidence on the motives and intention

of the firm and any relevant evidence fromits behaviour in other markets’ 22

Both definitions satisfy the classic three-stageprocess (pricing below cost, exit andrecoupment) and add a degree of commonsense to the cost measurement problem.

Identifying predation is important because ofits impact on the effective functioning ofmarkets. The great Alfred Kahn summed upthe importance of predatory behaviour to thechilling of competition: ‘The extent to whichmarkets are effectively contestable cannot be

independent of the ways in which the rich,dominant incumbents responded in the past toprevious entrants. As Irwin Seltzer once put it,a No Trespassing sign alone may not deter ahiker from walking on another’s property butwhen, just beyond the sign, the field is litteredwith bodies of previous trespassers — and allthe more when other fields, owned by otherpeople, are similarly littered — the lesson islikely to sink in.’ 23

The key problem in identifying predation ispicking normal competitive responses frompredatory behaviour. This has lead Joskow andKlevorick 24 to identify two types of error 25 thatcan be made in predatory pricing claims:

• Type I error: the identification of acompetitive price cut as a predatory price cut.

• Type II error: the failure to detect predatorypricing.

The Type I and II errors can actually be appliedto almost all competition cases — eithermissing behaviour that is damaging ormisidentifying behaviour that is benign.

For predation to be found in a market, Ordoverand Willig identified three basic conditions thatneed to be satisfied:

• The market must be horizontallyconcentrated. In an un-concentrated market,there is sufficient competition discipline fromthe remaining rivals to preclude monopolypower, irrespective of the exit of one firm.

• The market must be protected by a form ofentry barriers that we term entry hurdles.These hurdles exist whenever the prospectiveentrant is cost-disadvantaged relative to theincumbent solely because the incumbent isalready functioning as a going concern, andthe entrant has not yet committed therequisite resources. In general, entry hurdlesarise when investments are not fullyreversible. The need to incur the irreversibleportion of the investment, and thereby to putthat amount at risk, confronts the prospectiveentrant with a cost disadvantage relative tothe incumbent whose resources are alreadycommitted.

55


• The presence of re-entry barriers. A re-entrybarrier may be defined as the cost that a firmthat has exited a market must incur toresume production.’ 26

The key problem with legal tests for predatorybehaviour is that it is out-of-kilter withdevelopments in competition economics andlimits the focus of authorities on a limitednumber of potential abuses. In particular,predatory behaviour laws underestimate:

• the fact that information and the signalling ofintention are central to predatory behaviour;

• the creation and maintenance of a reputationis at the core of much predatory behaviour;

• the relationship between costs and prices isless important than previously thought;

• the static ‘bright lines’ approach of theAreeda/Turner rule 27 and the legal approachthat this bolstered are not as relevant as theyonce were.

The problems with identifying and prosecutingpredatory behaviour has led many to recastpredatory behaviour within a broader area,entitled exclusionary behaviour.

Fighting brands, fightingships and reputation

One of the key drawbacks to traditionalcompetition analysis of predation is the lack offocus on:

• Reputation: activities carried out to buildreputation;

• Multi-market contact: competition occurringin more than one market where activitiesdesigned to protect the position of the firm inone market leads to predation in another;

• Inter-temporal multi-market competition:activities in one market designed to ‘warnoff’ competitors from entering another in the future.

The use of ‘fighting brands’ or a ‘fightingcompany’ is particularly interesting to analyse.Yamey explains a fighting brand thus: ‘Its saleis confined to the affected areas; the quantitiesoffered are controlled so as not to makeunnecessary sacrifices of profit; and it is

withdrawn as soon as the objective has beenattained, namely the acquisition of theindependent by the monopolist, or thewithdrawal of the independent, or itsabandonment of plans of enlarging its share of the market.’ 28

He quotes examples of fighting brands from theCanadian match industry and the use of‘fighting ships’ by the UK shipping cartels inthe late 19th century when a cartel of shippingcompanies responded to independent shipsentering the lucrative China trade market bydumping capacity on the route to destroy themarket for the independents, drive them fromthe route and then re-build the cartel. 29

The chain store paradox and signalling

The chain store paradox 30 identified a paradoxbetween a formal game theoretical model andreal-world experience. The model operatedwith a 20-store incumbent chain store facingentry by a new firm. The model suggests thatthe incumbent would never fight entry. Inreality, that incumbent would fight the firstentrant as a means of sending a signal to allpotential entrants.

The paradox was solved by introducing:

• Imperfect information• Reputation

Kreps and Wilson 31 introduced the concept of‘imperfect information’ or uncertainty, Theyfound that ‘if rivals perceive the slightestchance that an incumbent firm might employ‘rapacious responses,’ then the incumbent’soptimal strategy is to employ such behaviouragainst its rivals in all, except possibly the lastfew, in a long string of encounters. For theincumbent, the immediate cost of predation is aworthwhile investment to sustain or enhanceits reputation, thereby deterring subsequentchallenges.’ 32 Yun Joo Jung, John H Kagel andDan Levin took the chain store paradox,remodelled it along the lines developed byKreps and Wilson and found that ‘predatorypricing is alive and well’. 33

56

Conduct criteria• Some uncertainty should exist in the minds

of rivals as to whether price initiatives will befollowed.

• Firms should strive to attain their goalsindependently, without collusion.

• There should be no unfair, exclusionary,predatory or coercive tactics.

• Inefficient suppliers and customers shouldnot be shielded permanently.

• Sales promotion should be informative, or atleast not misleading.

• There should be no persistent, harmful pricediscrimination.

Performance criteria• Firms’ production and distribution

operations should be efficient and notwasteful of resources.

• Output levels and product quality (e.g.variety, durability, safety, reliability and soforth) should be responsive to consumerdemands.

• Profits should be at levels just sufficient toreward investment, efficiency andinnovation.

• Prices should encourage rational choice,guide markets toward equilibrium and notintensify cyclical instability.

• Opportunities for introducing technicallysuperior new products and processes shouldbe exploited.

• Promotional expenses should not beexcessive.

• Success should accrue to sellers who bestserve consumer wants.


Part VIII: Aiming forworkablecompetition

The established models of competition(perfect, monopolistic, oligopolistic, etc.)and the interpretations discussed in thepreceding sections are theoretical models.While they are useful as constructs uponwhich to judge actual markets, a definitionof ‘workable competition’ can be valuable.

A useful definition of the workable competitiontarget can be arrived at by combining ourthinking on consumer behaviour with moreestablished work on Structure-Conduct-Performance. 34 We propose that the targetssought by competition regulators besupplemented as follows:

Consumer behaviour criteria• Information should be transparent and as

independent as possible.• Consumers must be provided with, or at least

not denied, opportunities to learn in markets.• Unnecessary sunk costs by the consumer

should be avoided.• Pressure-free re-evaluations should be

afforded consumers wherever possible.• Search costs should not be artificially

increased.• Feedback loops should be as timely and

accurate as possible.• Losses and gains should be de-linked

wherever practicable to allow consumers tobundle and re-bundle gains and losses.

Structural criteria• The number of traders should be at least as

large as scale economies permit.• There should be no artificial inhibitions on

mobility and entry.• There should be moderate and price-sensitive

quality differentials in the products offered.

57

Consumers and regulation


Two key questions need to be answered inthe study of regulation: when doesregulation occur and when shouldregulation occur? While both questions areoften lumped together into a single query, itis important to split them apart and identifyboth the drivers for regulation and the timewhen regulation is an appropriate response.

What exactly is regulation?

While the topic of regulation is a key one in somany contemporary debates, it is not alwaysclear what is being talked about. Commentatorsin the debate range in their definitions ofregulation across all forms of legal control, allforms of economic control and all forms ofregulation aimed at natural monopolyproblems. We are presented with two clearproblems in this regard. Firstly, how does onedefine regulation as a category of activity andsecondly, how does one identify the areaswithin which regulation occurs.

Anthony Ogus quotes Roger Noll 35 as arguingthat the ‘central meaning’ of regulation is as a‘sustained and focused control exercised by apublic agency over activities that are valued bya community’. This characterisation indicatesthat the focus of regulation is on those activitiesof general value to the community.

Ogus argues for two models of the form ofregulation: the collectivist and market models.

Under the collectivist system:• Regulation has a directive function. ‘To

achieve the desired ends, individuals are

58

compelled by a superior authority – the State– to behave in particular ways with the threatof sanctions if they do not comply.’

• It is public law in the sense that in general itis for the State (or its agents) to enforce theobligations which cannot be over-reached byprivate agreement between the partiesconcerned.

• Because the State plays a fundamental role inthe formulation as well as enforcement of thelaw, it is typically centralised.

Under the market model:• Regulation has a facilitative function, offering

a set of formalised arrangements with whichindividuals can ‘clothe’ their welfare-seekingactivities and relationships.

• Private nature differs from regulation in that:– it is left to individuals to enforce rights;– obligations are incurred voluntarily, in the

sense that they can be displaced byagreements between the affected parties iffound to be inappropriate:

• Private law is largely decentralised.

What is clear from these two models is that inmost market economies, the model ofregulation that applies most directly is amixture of the two. The exact balance of themixture will depend on the area of law and theeconomy subject to the regulation. For example,public utilities regulation is considerably morecollectivist in nature than the regulation ofcommercial contracts.

Ogus further divides social from economicregulation and identifies the two main driversfor social regulation as arising from the classicdefinition of market failure:

• Individuals in an existing or potentialcontractual relationship with firms supplyinggoods or services have inadequateinformation concerning the quality offered bysuppliers. In consequence, the unregulatedmarket may fail to meet their preferences.

• Even if this information problem does notexist, market transactions may have spillovereffects (or externalities) which adverselyaffect individuals not involved in thetransactions.


Part I: What it is and whyit has grown

Using the classic model of information andexternality based market failure, Ogus providesa useful typology of potential regulatorysolutions. He posits three central elements tothe regulatory spectrum in dealing with classicsocial regulation-based mechanisms foraddressing market failures:

Low end of spectrum:• Information regulation: forcing suppliers to

disclose details concerning the quality oftheir goods and services:

• Private regulation: imposing obligationswhich can be enforced only by individualsfor whose benefit they have been created;

• Economic instruments: these are not coercivebut rather induce desirable behaviour byfinancial incentives.

Middle of the spectrum: • ‘Command-and-control’ in which standards

backed by criminal sanctions are imposed onsuppliers.

High end of the spectrum• Highly interventionist approach of prior

approval which prohibits the undertaking ofan activity without a license or authorisationissued by an agency.

The threefold characterisation of choices inregulation is matched by the approach used byFriedman 36 where the protection offered bylicensing is only one of three possibleapproaches:

• Registration (e.g. of automobiles, pedigreedogs);

• Certification: does not directly involve thepolice power of the State in economictransactions (e.g. common in medicalspecialities);

• Licensure: directly involves the police powerof the State (e.g. only licensed medicaldoctors allowed to practice medicine)

The approach taken by Ogus is echoed in thework of Buchanan and Tullock, who argue thatthere are essentially three choices for decision-makers in responding to a market failure. These are to:

59

1. Rely on individual or private action2. Organise voluntary exchange3. Organise a collective choice

As the outline suggests, the key problem withchoices two and three are that they imposeconsiderable organisational costs on thoseseeking to use them. While private orindividual action imposes costs on theindividual undertaking the action, theorganisation of voluntary exchanges orcollective choice creates organisational and co-ordination problems, the latter often being heldup as an example of market failure.

It is interesting to compare the relativepopularity of the three approaches across timeand across industries. For many years, theregulation of many retail financial servicesproducts have relied on the low-end spectrumof regulation, most notably informationregulation. In contrast, privatised industrieshave tended to operate a licensing system, themost interventionist approach. As the markethas evolved over time, this has moved towardthe middle element of the spectrum. Thespectrum reflects the middle ground betweenthe more collectivist high-end intervention andthe more classically market-oriented low-end ofthe spectrum. The fact that the middle of thespectrum is most common in regulatoryinterventions is further evidence that the mixedapproach is most appropriate in what is still amixed market economy.

Why regulation has grown

Defining regulation can be difficult because ofthe range of options available to the regulatorand the regulated. What is clear is that the roleof regulation has grown in importance sinceWorld War II. This is not to argue that the post-war world has become obsessed withregulation of everyday life. It is more areflection of the evolution of politics in manycountries. The regulation of industry movedfrom being an issue of significant politicalrevolt and debate to one of bureaucraticefficiency and societal equity. The growth inregulation has emerged from a number ofsources:

Part I: What it is and why it has grown

• A significant change in the perception of therole of the State in the economy. In the UK,the immediate post-war election of a Labourgovernment and the creation of the NHS anda wave of nationalisation was evidence of amarked change in the attitude of the publicand political elite to the role of the State.Much of the later political debate has centred on the importance and diminution of that role.

• Increased public awareness of marketshortcomings. Political demands forregulation have grown with the growingtransparency of political debates and thehighlighting of failures in market provision.The nationalisation and privatisation of therail industry is a good example of anindustry whose shortcomings are wellpublicised.

• Universal suffrage which placedgovernments under pressure to adoptpolicies that appealed to a broader section ofpopulation and a process of politicalorganisation centred around regulatorydemands. The cosy world of elite politicswhich directed regulation toward specificgoals has been replaced by a broader basefrom which demands for regulation emerge.

• The structure of political rewards. Thecreation of regulation and regulatoryagencies creates a group of individualswhose interests are best served bymaintaining influence and apportioningbenefits to maintain their position.

• Rapid advances in technology constitutedgreater threats to health and safety: 19th andearly 20th century politics were largely aboutrisk, health and safety. The post-war worldsaw regulatory agencies internalise thatdebate. This has lead to a bureaucratisationof the means by which increasingly complextechnology-based risks have emerged.

• The ‘rights revolution’ in post-war developedeconomies: The emergence of a politicaldebate centred, in part, on the idea thatcitizens have rights to certain services andproducts has altered the regulatory contract.

60

• The tendency of politicians to add to theregulatory burden is great, given their shortaccountability horizon.

• Decoupling between the burdens of andbenefits of regulation: This process takes two forms: – micro-decoupling; where benefits from an

existing or prospective governmentprogramme are concentrated in a particulargroup while costs are broadly dispersedamong the public as taxpayers orconsumers;

– macro-decoupling; where political powerrests with the voting majority, while aminority provides most of the tax base.

Dicey: ‘[T]he beneficial effect of Stateintervention, especially in the form oflegislation, is direct, immediate and visible,whereas its evil effects are gradual and indirectand lie out of sight’.


Why regulation should happen

The classic case for regulation rests on the ideathat markets fail in the delivery of certaingoods or services, or produce unacceptable sideeffects in delivering the good or service.However, the definition of when regulationshould occur is hotly debated. The minimalistapproach argues that ‘(t)he most fully acceptedargument for intervention relates to what isvariously known as the public good, collective-consumption or non-exclusivity problem.Situations arise in which it is impossible or atleast infeasible or undesirable to excludeanyone from the consumption of a commodity.’37 In this conception, intervention can only beargued to be positive if it addresses the classicpublic goods problem. Interestingly, theminimalist approach does not includeexternalities, commonly assumed to be a triggerpoint for intervention. Ogus has a more fulsome list of trigger points forregulation:

Key failures:• monopolies and natural monopolies;

• public goods: consumption by one persondoes not leave less for others to consume; itis impossible or too costly for the supplier toexclude those who do not pay from thebenefit;

• Other externalities: problems of attempts tocorrect externalities:– the third party on whom the cost is

imposed may have received ex ante, or willreceive ex post, indirect compensation forthe loss (e.g. temporary road closure toresurface and increase traffic);

61

– externalities are not unilaterally imposed.There is friction from the competing andconflicting claim of two parties for use of asingle resource; the burden of avoiding oreliminating the friction should be imposedon whichever party can achieve this atlowest cost;

– Not appropriate to eliminate ‘pecuniaryexternalities’: pure value (financial)changes borne by third parties which resultfrom changes in technology or in consumerpreferences. They involve indirect effectswhich alter the demand faced by theharmed or benefited third party. Pecuniaryexternalities are the result of the naturalplay of market forces. They involve wealthtransfers which cancel out and do notincrease the costs faced by society. 38

– Transaction costs: externality may give riseto a misallocation but the administrative andother costs of correcting it may outweigh thesocial benefits arising from such actions;

• Information deficits and bounded rationality;

• Co-ordination problems:’…the private lawwas used to facilitate the co-ordination ofutility-maximising activities. Contracts andlegal forms of organisation, notablycorporations, are used overtly for thispurpose. Such specific legal arrangementsgive rise, of course, to significant transactionscosts; and other private-law concepts,particularly torts and property rights, can betreated as transactions costs savings devicesto the extent that they lay down standards ofbehaviour that it is assumed the partieswould have agreed to in contracts iftransactions costs had not inhibitedthem….although coercion is invariablyinvolved, regulation justified on this basis isnot really forcing people to do what theydon’t want to do, but rather enabling them to do what they want to do by forcing themto do it’;

• Exceptional market conditions and macro-economic considerations.’ 39s

Ogus also ties market failure to a failure of law.He argues that a prima facie case exists forregulation in the public interest where the‘market failure’ is accompanied by a ‘private

Part II: Regulation and market failure

Part II: Regulation andmarket failure

law failure’. He further points to non-economicgoals of regulation in seeking to furtherdistributional justice, community values orsome form of paternalism, which he defines as‘the interference with a person’s liberty ofaction justified by reasons referring exclusivelyto the welfare, good, happiness, needs, interestsor values of the person being coerced.’ 40

The tendency of economists to react against thesocial aspects of regulation has been temperedin some areas by the experience of the fallibilityof markets and of consumers in them. A paperby Joseph Stiglitz, Peter and Jonathan Orszagcalled ‘The Role of Government in a DigitalAge’ provides a slightly broader list of criteriaunder which reasons for regulation can befound. The criteria they list are:

• ‘Failure of competition…(t)he governmenttherefore has a role to play in ensuringeffective competition in private markets.

• Public goods…(I)n general, private marketswill not supply public good (41) – or notsupply them in sufficient quantities – andtherefore the government has a role to playin providing them.

• Externalities.…In general, the governmenthas a role to play in correcting negativeexternalities or promoting positiveexternalities. Without governmentinvolvement, private markets will typicallyunder-produce goods with positiveexternalities and over-produce goods withnegative externalities.

• Incomplete markets. A possible justificationfor government activity is incomplete markets.For example, imperfections in capital andinsurance markets – such as the absence ofinsurance coverage for certain types of risks –may warrant government involvement. Aclassic example of an imperfect capital marketis the inability to borrow against higher futureearnings, which justifies a government role inproviding loans or loan guarantees for post-secondary education expenses. In addition,certain types of goods or services may requirelarge-scale co-ordination, which may bepossible but difficult to achieve withoutgovernmental assistance.

62

• Information failures….information is in someways a public good, and therefore thisrationale for government is similar to thesecond rationale.

• Macroeconomic fluctuations. Thegovernment has a role to play in correctingmacroeconomic imbalances, such as thosethat lead to periodic problems with highunemployment, inflation, or recession.

• Redistribution. Even if private marketsproduce goods and services efficiently, societymay not like the distribution of income thatresults. The government may therefore have arole in redistributing income – for example,through a progressive tax system – to producea more equal distribution of income.

• Merit goods. There may be cases in whichindividuals would make ‘bad’ decisions ifleft to their own devices and in whichgovernment paternalism is thereforewarranted. For example, the governmentcompels individuals to attend school or wearseat belts largely because it is concerned thatpeople will not do ‘what’s best’ in theabsence of such mandates. The governmentmay sometimes be justified in compellingindividuals to consume ‘merit goods’ (e.g.elementary education).’

Stiglitz et al thus explicitly recognise thatmarket failure can extend beyond the simplymonopolisation and public goods problem andgeneration of externalities and extend intoincomplete markets, information problems andpoor decision-making by individuals. However,the most interesting inclusion is that ofredistribution, a factor often left outside ofclassical economic approaches to regulation.While most discussion of redistribution hastended to focus on taxation systemsrebalancing income distribution resulting fromthe interplay of market forces, regulation has amuch broader role to play in redistributivepolicies. This is particularly true in the area ofutility regulation, where redistribution hasalways been an implicit part of the system(through universal service obligations) butwhere the introduction of marginal cost pricinghas exposed the redistributive policies togreater pressure and scrutiny than before.


Why regulation does happen

The enormous literature that exists on theinstance of regulation can be seen as part of thepositive approach to regulation: the study ofwhen regulation does occur. The literature onwhen regulation should occur is lessvoluminous. This more normative approachtends less to advocacy and more towardidentifying factors common to regulatoryinterventions and attempts to predict whensuch circumstances will lead to intervention.Separating out the normative from the positiveis never a straightforward process.

One such approach (Normative Analysis as aPositive Theory, or NPT) argues that 'regulationis supplied in response to the public’s demandfor the correction of a market failure or for thecorrection of highly inequitable practices (e.g.price discrimination or firms receiving windfallprofits due to some change in industryconditions.) According to this theory, if amarket is a natural monopoly, the public willdemand that the industry be regulated becausea first best solution is not achieved in theabsence of regulation. Unfettered competitionwill result in either too many firms producingand/or price exceeding the socially optimallevel. By regulating the industry, net welfaregains result. It is this potential for welfare gainsthat generates the public demand forregulation. In this way, the public interesttheory uses normative analysis (when shouldregulation occur) to produce a positive theory(when does regulation occur).'

Viscusi et al argue that the NPT approach is bothrefuted by evidence and ignores the process bywhich regulation occurs. It is thus a rather flat

63

approach to dealing with the development ofregulation. They choose to reformulate NPT in apotentially more cynical manner as: ‘regulationis originally put in place to correct a marketfailure but then is mismanaged by the regulatoryagency.’ But even this approach fails tosufficiently stack up in the real world ofregulation, both for not dealing with industriessuffering regulation despite a lack of marketfailure (as it is classically defined) nor industriesthat actually seek regulation.

Theory of regulation: Stigler-Peltzman

Perhaps the most significant advance in thestudy of regulation came in the work of Stiglerwho attempted to develop a general economictheory of regulation. In a now classic article 42,Stigler argued that interest groups will seek tomaximise their income by seeking to persuadethe State to use its monopoly of coercion totheir benefit. Regulation is a key tool forinterest groups to persuade the State toredistribute income from one group to another.

The basic argument that interest groups willseek to persuade the State to redistributeresources for their benefit was furtheradvanced by Peltzman 43. His key assumption isthat the individual who controls regulatorypolicy chooses policy to maximise politicalsupport. Most US academics assume thisregulator to be, in the first instance, a politician.While this assumption does not always holdtrue, it is a useful starting point. Other work,most notably from the Public Choice schooland from capture theory, deals more directlywith regulatory bureaucracies and agencies.

Peltzman argues that in deciding governmentpolicies, a politician will decide the size of thegroup that will gain from regulation and howmuch wealth should be transferred to them.This calculation will be made as part of a desireto be re-elected and thus involves assessmentsof political support (both in terms of votes andfinancing). Naturally a regulatory policy thatinvolves a transfer to a group also involves atransfer from a group. Here the politician willneed to assess the opposition likely to emergefrom the group suffering a loss. As a rule-of-

Part III: Why regulation happens

Part III: Why regulationhappens

thumb, the opposition to a measure willincrease along two vectors: the size of thepopulation likely to lose out and the intensityof this group. On balance, a regulator willchoose a policy (defined by the size of thegroup that gains and the amount of the transferto them) to maximise political support minuspolitical opposition. The object is thus netpolitical support. There is no mention, ornecessary place, for assessments of overallsocial welfare.

One of the main benefits of theStigler/Peltzman approach is that it helps toidentify both those industries where regulationis to occur and those interests most likely tolose out in regulatory debates. In short, themodel argues that regulation is likely to bebiased toward the benefit of a small,concentrated interest group with strongly feltpreferences at a loss to a large interest groupwith weakly felt preferences (e.g. consumers.)

The weakness of the larger social interest is keyin this analysis, both as a demandeur ofregulation and as a recipient of regulatorycosts. In the case of demanding regulation, themodel argues that a large beneficiary groupwill be weak because of the rising marginal costof organisation and its tendency to consumethe benefits sought. Keeping per capita benefitshigh requires that the interest group be keptsmall. Keeping the costs of regulation dispersedlimits opposition as the rising marginal cost oforganisation easily wipes out the marginal lossfrom a policy.

A second key conclusion from theStigler/Peltzman model is that regulation willalso be sub-optimal in advancing the wealthand social interest of the community. This isbecause regulators will seek to maximise theirown net benefit rather than the maximumbenefit possible. Trade-offs and middle groundwill always be sought to maximise benefitsminus losses. Following the first conclusion, itis relatively straightforward to reduce benefitsby a small amount to each gainer. Such a smallreduction in gain will actually result in a largereduction in loss for each individual loser (asthe loss is spread out across such a largegroup). Such a strategy will minimiseopposition (raising the marginal cost of

64

organising such opposition). Thus regulation islikely to regulate both price and entry to thebenefit of incumbents (as this directlypositively affects profits) but will not reachmonopoly pricing levels as a means of limitingopposition.

The third major advance of the theory is theidentification of industries that are likely to beprone to regulation. This argument can besummed up as follows. Industries where theprice achieved is similar to that likely underregulation (not the full monopoly price, buthigher than under competition) are unlikely tosee much demand for regulation. This isbecause neither party (industry or consumer)has much interest in organising to see change.However, industries where the price achievedis significantly lower than that seen in thepotential regulated industry will see a demandfor regulation from the firms already in it. Inthe real world, this is evidenced by firmsdeveloping entry requirements, codes ofconduct, etc. that raise entry barriers andrestrict competition. Likewise, in industrieswhere prices are significantly above that likelyunder regulation, consumers are likely todemand regulation to lower payments. Thusdemands for regulation are most likely to occurboth in relatively competitive industries (whereincumbents will gain from restricting entry)and in monopolistic industries (whereconsumers will gain from seeking regulation).

Role of interest groups

While much of the literature on regulationfocuses on the market failure justification forregulation, a second approach, focused on therole of interest groups, has risen to prominence.The Stigler/Peltzman approach is an exampleof a hybrid approach modelling both marketsand interest group lobbying. The importance ofthe interest group in seeking gains fromregulation has been recognised officially in theUS Annual Report of the Council of EconomicAdvisers (1994); ‘As recognised by both theframers of the Constitution and modernscholars of public choice, all political systemsprovide interest groups with an incentive for‘rent seeking’ that is, manipulation of collectiveaction for private benefit…[rent seeking] can


lead government agencies to make decisionsthat benefit a particular interest group eventhough they are costly to society as a whole’.44

Gary Becker models interest group lobbying asa zero sum game: no one group can gain unlessanother one loses. Aggregate influence is fixed.The implication is that what is important fordetermining the amount of regulatory activity(as measured by the wealth transfer) is theinfluence of one group relative to the influenceof another group. For Becker, the politicalprocess is one in which politicians transmit theinterests of interest groups (e.g. firms,consumers or voters). What matters in thisprocess is the relative balance of powerbetween the groups. While Stigler/Peltzmanargued that the lobbying of some groups wasalways more likely to be stronger than others,Becker takes the problem of free-riding (interestgroup members gaining without any effort inthe cause) as a problem shared by all interestgroups. What matters for Becker is that theoptimal pressure to be applied by one groupdepends almost entirely on the pressure beingapplied by the opposition group. The familiarpatter of lobbying (e.g. one side hires lawyers,the other does the same) escalates into alobbying arms race. This arms race uses upresources that could more effectively used inpromoting development and efficiency, ratherthan wasted in lobbying battles.

The disdain for the wasteful activities ofinterest groups is shared by the Public Choiceschool. One of its major lines of inquiry is onthe process of political 'rent-seeking'. Rent-seeking is defined as ‘the resource-wastingactivities of individuals in seeking transfers ofwealth through the aegis of the State’. 45

Interestingly for the Public Choice school, thewaste of rent-seeking is not identified as

65

existing in the creation of monopoly positionsthrough government action, but the use ofresources to maintain that position. As David B.Johnson 46 puts it: ‘The nation’s resources arewithdrawn from productive activity and put torent-seeking activity by the monopolist becausehe is willing to pay a higher price. Themonopolist gains a monopoly right, but citizenslose because the resources which could havebeen producing real goods and services areused to chase rents.’

The activity and relative strength of interestgroups was rather effectively modelled byJames Q. Wilson. His typology of politicallobbying activity is reproduced in tabular form below.

For Wilson, normal regulatory politics dividealong axes of benefits and costs. Where costsare concentrated on a particular group butbenefits are also concentrated on a particulargroup, we tend to see a pattern of interestgroup politics. Such interest groups are rangedequally against each other as, to borrow fromStigler/Peltzman, the marginal costs/benefitsof organisation are similar for each group. Forexample, one can see a situation where a steelindustry seeking a protective anti-dumpingtariff seeks to gain from such a policy but findsitself met by auto makers who will lose a gooddeal if the steel makers gain. In both instances,one finds a small numbers group of highlyconcentrated interests with much to gain andlose. Wilson's model posits that governmentaction is uncertain in this area. Given theStigler/Peltzman calculation, the net politicalbenefit calculation is difficult to make.

In contrast, where the benefits of a policy areconcentrated on one interest group but thelosses are spread out over a larger number, one


Models of politics

COSTS

BENEFITS Concentrated Diffuse

Concentrated CONTROVERSY GOVT ACTIONUNCERTAIN GOVT ACTION Client politicsInterest Group politics

Diffuse GOVT ACTION UNCERTAIN GOVT ACTIONEntrepreneurial politics Majoritarian politics

tends to find a model of client politics. Here thegroup seeking to gain from a regulatorysolution will try to impose that cost on widersociety. Examples of this model abound andregularly occur in the agricultural sector.Government action is likely to occur insituations of client politics as theStigler/Peltzman net benefit swings in favourclearly of those seeking to extract benefits tothe cost of society. Such benefits will never bethe maximum possible, given the need tomaximise net benefits to the regulator.

Where both costs and benefits are diffuse, onefinds almost the model of representativedemocracy that western civilisation is supposedto aspire to. Indeed, a strand of regulatorythought – that of civic republicanism – capturesthe approach of the Majoritarian politicaldiscourse. In Cass R Sunstein’s formulation, therepublican civic tradition is characterised byfour key principles: 47

1. Deliberation in politics is made possible by afocus on civic virtue;

2. There is an equality among political actorstogether with a commitment to eliminatedisparities in political participation andunbalanced influence among individuals orgroups;

3. There is a belief in universalism as a processof mediation based on practical reasoningthat yields correct outcomes that benefit thecommon good;

4. Citizenship, as indicated by a broad guaranteeof rights of participation.

The civic republican tradition thus appears,under the Wilson model, to only apply in thosesituations where benefits and costs are bothdiffuse. Like the model of interest grouppolitics, it is unclear whether or notgovernment action will arise when both costsand benefits are diffuse. As under theStigler/Peltzman model, the net politicalbenefit calculation is difficult to make. It is alsonot clear that making the calculation is likely tobe an urgent requirement.

Perhaps the most interesting area of the modelis the area where benefits are diffuse but costsare concentrated. Here we enter the realm ofthe entrepreneurial politician. The entrepre-

66

neurial politician needs to mobilise a diffusegroup of interests, each with a relatively smallbenefit to be sought, to counteract aconcentrated group with a good deal to lose.Wilson argues that the only way in which sucha campaign can succeed in bringing forwardgovernment action is for the policyentrepreneur to utilise a broad, sociallyaccepted, moral discourse to challenge theconcentrated group that is likely to bear thecost of action. In essence the body seeking toimpose a concentrated cost to the benefit ofmany must appeal to societal norms of fairnessand equity and target those bearing the cost asnot conforming to those norms. If the grouprepresenting the diffuse interest fails in its task ,the body bearing the cost will seek governmentaction to minimise that cost and limit thebenefits of the wider group. If it succeeds it can(in Stigler/Peltzman terms) minimise theirmarginal cost of organisation by tying everyoneinto a shared vision of change. For interestgroups seeking to impose a cost on an industry,this area of politics is perhaps the most difficultarena to operate in while offering thepotentially greatest benefit. One can seeexamples of this approach in the environmentalmovement.

Watching out for thebootleggers

One of the most interesting and colourfulapproaches to regulation was put forward byBruce Yandle in ‘Bootleggers and Baptists in theTheory of Regulation,’ which sought to add tothe Stigler (which interest groups wouldsucceed) and Peltzman (no one group woulddominate and a balance would be struck)approaches. The Bootlegger-Baptist approachoffers a terribly apposite description of the realworld of interest group lobbying. The approachhas two starting points. Firstly, it starts fromthe premise that alliances can be formedaround issues for totally contradictory reasons.Secondly, it argues that rhetoric can be just asimportant as campaign finance. No lobbyingwill succeed unless it combines the two. TheBootlegger-Baptist alliance centred on thecampaign to maintain the policy of prohibitionin the US in the 1930s. The Baptists wanted tokeep the policy as a moral effort to stop


Americans from succumbing to the demondrink. Bootleggers, on the other hand,supported prohibition because it guaranteedthem enormous profits. An unholy alliance wasthus formed to keep a policy in place; Baptistsproviding the moral rhetoric to stir the spirit ofthe American electorate and bootleggersproviding funds to ensure their market was notthreatened.

As Yandle himself argues: 'B&B theory helps toexplain how leaders of consumer groups helpmajor pharmaceutical companies (the ones withapproved chemical entities) by valiantlysupporting a cautious FDA approval process.The theory explains why holders of permits toproduce and market EPA-approved insecticidesvalue the efforts of environmental groups whooppose rule changes that facilitate the entry ofnew and sometimes less risky substitutes.Indeed, once the theory is explained,bootleggers and Baptists seem to come out ofthe woodwork.

Perhaps we should we expect no less. Politicalaction, which by definition always serves someinterest groups, requires politicians to appeal topopular icons. By making a “Baptist” appeal,the canny politician enables voters to feel betterby endorsing socially accepted values in thevoting booth. The same politician, if he isadroit, also can enjoy the support ofappreciative bootleggers in the costly struggleto hold office. Bootleggers and Baptists are partof the glue that binds the body politic.47

Of course, the Bootlegger-Baptist theory doeshelp to identify potentially dangerous alliances.However, it can lead to an overly cynical viewof the world and one that often ignores thereality of interest group position forming. It iseasy to assume that common positions arearrived at in a concerted and collusive way. It isalso easy to assume that interest groups areintelligent enough, or aware enough, to knowthat an alliance might exist. The reality isconsiderably more complex. However, theapproach does provide a salutary warning topublic interest groups that find themselves inbed with unlikely allies – all Baptists must beon the lookout for bootleggers, as they will seekto impose costs on the cleric's flock.

67

Government failure and therole of regulators

The importance of the role of the regulator wasrecognised by John Landis, former dean of theHarvard Law School, in his report to PresidentKennedy: ‘The prime key to the improvementof the administrative process is the selection ofqualified personnel. Good men make poor lawsworkable; poor men will wreak havoc withgood laws.’ 48

Landis chose to focus on the need for goodadministrators; however, Wilson 49 hasquestioned this categorisation of the goodadministrator by outlining essentially threecategories of administrator in a regulatoryagency:

1. The Careerist: An employee aiming at long-term employment in the agency. The majorconcern of the Careerist is that the agencycontinues to exist and grows. A Careeristtends to look down on deregulation.

2. The Politician: An employee who see theagency as a stepping stone to bigger, betterthings. This category of employee tends to beboard member or commissioner level.

3. The Professional: This employee identifiesmore with holder of the same skills (e.g. law,economics) than with the regulatory agency.What is important for the professional is themaintenance of the professional esteem of peers.

What Wilson's categorisation brings out is thefocus of a school of academia on the role of theregulator and the government itself.


Part IV: Government failure,regulators andregulation

The development of Public Choice Theoryreally came to the fore in the 1970-80s. ThePublic Choice approach is not a uniform oneand involves a number of different targets. Ingeneral, positivist Public Choice theoristsanalyse the political process underlying thelegislative activity, rules of bureaucraticdecision-making; actual decisions of thebureaucracy; the regulatory process and itsrules and regulations, and the constitutionalprocess, particularly the rules for making rules.The more normative members of the publicchoice community analyse the same sorts ofareas but do so in an effort to ensure that themost efficiency enhancing solution is arrived at.Unlike the approach adopted by Posner, thepublic choice approach focuses on non-judicialrule-making and tends to look mostly atregulatory and bureaucratic decision-making.

The focus on the government and thebureaucracy of regulation violates one of thetenets of classical and neo-classical economics— namely that the State is a black box and isexterior to the operation of the economy. Forneo-classicists, the institutions and decisions ofgovernment are exogenous (external) togovernment, while for public choice theoriststhe decisions of institutions are endogenous(internal) to the economy. James M. Buchananhas described this approach as follows:

‘The critically important bridge between thebehaviour of persons who act in themarketplace and the behaviour of persons whoact in political process must be analysed. The‘theory of public choice’ can be interpreted asthe construction of such a bridge. The approachrequires only the simple assumption that thesame individuals act in both relationships.Political decisions are not handed down fromon high by omniscient beings who cannot err.Individuals behave in market interactions, inpolitical-government interactions, in co-operative-governmental interactions, and inother arrangements. Closure of the behaviouralsystem, as I am using the term, means only thatanalysis must be extended to the actions ofpersons in their several separate capacities.’ 50

The focus on efficiency starts from the premisethat all decisions involve a two-fold cost; firstlythe external costs of decision-making (those

68

costs borne by those who disapprove of adecision) and the internal decision-makingcosts (the bargaining costs associated withmaking a decision).

In a similar vein to the Stigler/Peltzmanapproach, the public choice theorists argue thatbureaucrats do not make decisions that aresocial welfare enhancing. In terms of economictheory, the utility-maximisation process of thebureaucrat is not the same as the maximisationof social welfare: 'other motives' driveregulators. Some of these motives form thecharacterisation put forward by Wilson above.Other include 'power, prestige, the size of thebureau’s budget, job security, perquisites,future salary and working conditions.'

Public choice theorists argue that legislatorshave few incentives to reform bureaucracies toensure that they do maximise social welfare.This lack of incentive comes from acombination of factors:

• Bureaucracies are difficult to monitor theperformance of (i.e. no profit/loss);

• The information that does exist is biased, as it is typically held by the body itself;

• Politics generally means that inefficiencyrarely leads to the termination of thebureaucracy;

• The bureaucrats and clients can formpowerful coalitions against change.

The focus on the government and bureaucracyoften mixes two different areas of inquiry. AsSpence and Gopalakrishnan 51 maintain, theargument has 'coalesced along two distinguish-able (but not always distinguished)dimensions: one that focuses on the substantiveefficiency of regulatory policies, and anotherthat focuses on procedural efficiency, or ratherthe inefficiencies that stem from the process bywhich agencies make policy'.

Identifying government failure

While market failure is often claimed as theprimary focus of many analysts, there is astrong train of academic analysis that suggeststhat the failure of government or regulators tocarry out their tasks is a key element in the


understanding of regulation. Before looking atthe theories around the failings of regulators tocarry out their tasks effectively, we need toidentify what it is that a government does andhow this can be balanced in comparison to therole of the market.

It has been argued that the activities providedby non-market players (e.g. government andregulatory agencies) are essentially of fourtypes:

1. regulatory services (e.g., environmentalregulation, radio and TV licensing, interstatecommerce regulation, food and drug control);

2. ‘pure’ public goods (national defence, spaceR&D);

3. quasi public goods (education, postalservices, health research);

4. administering transfer payments (federalstate, and local welfare programmes, socialsecurity etc). 52

Because of the nature of these products, thedemands for them and the requirement fordistribution means that there may well befailures in their provision. These failures areidentified as the government failureequivalents of market failures. It is argued thatgovernment failure occurs not just on thedemand-side of the equation. The demand sideof government regulation is a key attribute ofthe Stigler/Peltzman approach. It tends tofocus on the position of firms in themarketplace seeking regulation fromgovernment as a means of restricting theoperation of the market. However, thegovernment failure approach identifies supply-side problems inherent in the regulatorycontract. In particular, it argues that non-market goods are beset with problems. Mostnotably the following characteristics make theproper functioning of the non-market difficult:

1. difficulty in defining and measuring output2. single source production3. uncertainty of production technology (e.g.

education)4. absence of bottom line and termination

mechanism 53

The difficulty on the non-market supply-side isevidenced in UK experience. The recent

69

arguments over the setting of targets for publicofficials and in health and education are clearattempts to deal with the problems identifiedabove; namely the units of measurement andthe operation of a single supplier. The problemof performance in an environment where thebottom line is not a primary driver of corporateperformance is also one that most regulatorsand government agencies would understand.

The problems inherent in the nature of non-market goods can be overplayed. As with allprocesses of identification, a number of‘products’ of the non-market have becomeproducts of the market, such as those in utilitiesand in some healthcare systems and lawenforcement operations. However, this does notinvalidate the argument that there are problemsinherent in the production and supply of thosegoods. What it does indicate is that the‘solution’ to the problem of the non-marketgood must be continually monitored to identifythe degree to which it has delivered a marketsolution to the failure of the non-market todeliver the desired result.

Wolf 54 has identified four main types andsources of failure that arise in the provision ofnon-market goods and services:

1. ‘the disjuncture between costs and revenues:redundant and rising costs: the tendency ofgovernment agencies to continually expandat the expense of taxpayers/license payersbecause the supply of resources is atsomeone else’s expense;

2. internalities and organisational goals: thetendency of government agencies to havetheir own agendas and desires to internaliseactivities and power, including:

• specific agency maximands • budget growth: ‘more is better’ – the

continual desire of agencies to expand. Theannual battle between the ‘spending’departments in the UK for resources is anexample of this;

• technological advance: ‘new and complex isbetter’ – this problem tends to arise inmilitary departments who continually cravethe latest gadget;

• information acquisition and control(‘knowing what others don’t know is better’).

Part IV: Government failure, regulators and regulation

3. Derived externalities – generation ofunintended effects: a perennial problem in allmarket and non-market activity. Everydecision and action taken will have effectsbeyond those intended. The reason it cropsup as a non-market failure is tied into theproblem of accountability and budgeting – ifthe agency making the decision is unlikely tobe concerned about its accountability,unintended side effects are unlikely to beminimised.

4. Distributional inequity: distributional issuesare often thought of as external to theregulatory process. Indeed many policies aredesigned on the basis that distributionalinequalities that may result from the policywill be picked up elsewhere.

Comparing market with non-market failure

The preceding analysis suggests that bothmarket and non-market failures exist and thatboth market and non-market demands andsupply of regulatory behaviour is subject torent-seeking behaviour of one sort or another(be it firms or bureaucrats). Given this fact, itmight be useful to compare the two approacheswhen proposing one or the other as a ‘solution’to a ‘problem.’

Charles Wolf has done us a service bytabulating the main characteristics of marketand non-market failure. 55

Wolf’s characterisation is a useful rule-of-thumbin comparing the potential failures from a policyproposal. Thus one can ask some basic questionsabout the policy being proposed and the‘problem’ being ‘solved’. For example, where

70

significant externalities are being generated byan industry in terms of pollution, a logicalsolution may be to regulate the output.However, if this is done through the non-market,one may run up against the desire of the agencyto accrue more power and impose greater costson the industry. The ‘solution’ to the ‘problem’of pollution may thus have been to create anunwieldy and excessive bureaucracy.

Of course, Wolf’s characterisation of the polarchoices in failures is not a hard-and-fast rule.However, it is an indicator of potentialproblems and a useful means of designingregulatory solutions that minimise the failuresinherent in the choices available.

Wolf also provided a handy matrix of demandand supply problems in the non-market failure arena and tried to tie them together (see opposite page).

This is a useful tool to identify both potentialproblems within the non-market sector andsome potential warning signs to look out for, orfor policy solutions to be proposed to minimiseexpected failures.

Wolf draws the following conclusions from the matrix:

• The typical miscarriages of the non-market(i.e. of government) are no less identifiable,characteristic or predictable than thosecommonly attributed to the market;

• The typology of these characteristic non-market failures suggests that they are bothformidable and relatively neglected;

• Whether they are more or less formidablethan the failures of the market may beascertainable and demonstrable in somecontexts but likely to be debatable in others;


Market and non-market failure compared

Market failures Non-market failures

Externalities and public goods Disjunction between costs and revenues: redundant and rising costs

Increasing returns Internalities and organisational goals

Market imperfections Derived externalities

Distributional inequity (income and wealth) Distributional inequity (power and privilege)

• The choice between markets andgovernments is not a choice betweenperfection and imperfection but betweendegrees and types of imperfection, betweendegrees and types of failure. In manyinstances, it may be simply a choice betweenthe disagreeable and the intolerable. 56

The result of combining market and non-market failures presents a difficult balancingact. On the one hand, market failures clearlyarise and require action. However, the mannerin which society deals with them mustrecognise the existence of non-market failure.This presents the political agency with adilemma not of regulation but of balance. Inregulatory policy, the choice is not betweenleaving the market to itself or regulating everyelement of the market. The political actor mustrecognise that in a large number of cases themarket will not want to be left to itself, it willdemand regulation as a means of restrictingcompetitors and maximising profits. Similarly,the agency must recognise the problems of non-market intervention not just in terms of theability of the State to deliver, but in terms of thetendency for non-market failure to emerge.

Each regulatory quandary is thus likely to bedealt with in slightly different way. The endresult of any such analysis will be a balancebetween market and non-market action. Whereon the continuum between the two the resultsits is likely to depend on a number of keyfactors:

71

• size and extensiveness of the externalitiesresulting from market activities;

• degree of monopoly power associated withthe production and sale of market output;

• extent and visibility of imperfections in the market;

• degree to which distributional benefits whicharise from the market activity are inequitablydistributed. 57

Taking a leaf out of the Public Choice school ofthought can be useful when proposing a non-market solution: ‘the key is to designinstitutions that facilitate competition for thoserents that accompany newly created surplusesor new wealth creation and that discourage thewasteful competition for existing rents.’ 58 In allcircumstances, a recognition of the failures ofboth market and government must be madeand a recognition that the demand forregulation can occur from within industriesrather than simply from without. The PublicChoice school recommends that in designingregulatory solutions:

• the role of government regulation iscircumscribed quite tightly;

• rent-creating government institutions are avoided;

• constitutional reforms are adopted thatrequire super majorities (two-third or three-fourth majorities) to pass rent-seeking-type legislation.

Part IV: Government failure, regulators and regulation

Typology of combined demand and supply problems in non-market failure

Demand/supply Non-market Non market conditions demand conditions supply conditions

Non market Increased Political Decoupling Output Single Uncertain Absence offailures awareness/ rewards/ measure- source technology termination

political time ment production mechanismorganisation discounts

Disjunction between costs/revenues; redundant and rising costs X X X X X X

Internalities and organisational goals X X X X

Derived externalities X X X X

Distributional inequities X X X X

In the field of trade policy many of these PublicChoice recommendations have been part of theregulatory landscape for many years. Followingthe disaster of the Great Depression in the USin the 1930s, trade policy (which constitutional-ly resides in Congress), was increasingly passedto the Executive because the politicians wereaware that they could not trust themselves toavoid coming to the aid of supplicantindustries. US trade policy came to be designedaround insulating the trade policy-makingmachinery from rent-seekers; the constitutionalbalance of power circumscribed the role of theExecutive agency and the Fast-Trackprocedures developed after 1974 ensured thatrent-seeking changes to trade pacts were almostimpossible after the agreement had beensigned. While less well-documented, the meansby which trade policy is made in the EU issimilar, if less successful, in terms of its attemptto isolate itself from rent-seekers. While notserving as a model for democraticinclusiveness, it can be argued that trade policyhas been a more effective model of efficiency.Here lies a much broader problem in regulatorypolicy: the balance between efficiency andaccountability. Here the Public Choice school isrelatively silent.

72

When will regulation be successful?

A key challenge for regulators is to identifysolutions to problems that deliver efficiencyimprovements or do not make situationsdemonstrably worse. The approach oftenfavoured by economists when studying marketinterventions is to look for the Pareto optimalsolution. This solution seeks a position whereany solution creates a situation where no onecan be made better off without someone elsebeing made worse off. While Pareto optimalsolutions are frequently discussed in economictextbooks, it is not clear that the optimalsolution can ever be fruitfully and effectivelyachieved. It is for this reason that the study ofpractical regulation has lead to thedevelopment of two key yardstick targets forregulation. The targets are generally referred toas Kaldor-Hicks efficiency and the wealthmaximisation hypothesis.

The need to ensure that a policy is sociallybeneficial has been a concern of to economistsand social commentators for years. NicholasKaldor argued in 1939 59 that the repeal of theCorn Laws was beneficial to society because ithad so benefited consumers that even iffarmers were compensated for their loss,society would still be in a better position thanbefore the regime had changed. This benefitwas assessed as existing even if consumers hadchosen not to compensate those who had lostout as a result of the law change. This is theKaldor criterion: a policy is deemed to beefficient if it results in benefits for those whogain on such a scale that they ‘potentially’ cancompensate fully all those who have lost outand still remain better off. Importantly, the


Part V: Getting regulationto work forconsumers

gainers are not ‘required’ to compensate thelosers for a policy to be deemed efficient. Theapproach thus means that the criterion isachieved when the gains to one group exceedthe losses to another.

The 'compensation' principle approach ofKaldor was initially a fairly one-dimensionalapproach. However, in combination with thework of Hicks, 60 we have a policy that statesthat a legal change can be argued to haveimproved societal well-being if and only it,both the beneficiaries of a change could fullycompensate the losers and remain better offthemselves, and the losers could not havecompensated the beneficiaries sufficiently toget them to forego their benefits withoutthemselves being worse off than in theiroriginal position. The compensation principlethus has a forward and backward inductionprocess and forms the basis for much of thecost-benefit analysis that has followed it.

The relative starkness of the Kaldor-Hicksapproach has been finessed by Richard Posner.In many ways, the Posner ‘wealthmaximisation’ approach is even more basicthan the compensation principle. Posner hasargued that ‘(a) second meaning of ‘justice’,and the most common I would argue, is simply‘efficiency’. When we describe as ‘unjust’convicting a person without a trial, takingproperty without just compensation or failingto require a negligent automobile driver toanswer in damages to the victim of hiscarelessness, we can be interpreted as meaningsimply that the conduct or practice in questionwastes resources.’ 61 Posner places the pursuitof wealth maximisation as the key driver inassessments of justice in legal cases. It has to benoted that Posner is primarily thinking ofcommon law judgements rather thanadministrative law regulatory solutions.However, his approach is also influential in this setting.

Posner contrasts the wealth maximisationprinciple with the more philosophical positionsof utilitarianism and Kantianism (which tend toemphasis autonomy and human respect). Hiscontrast must be placed within the debates ofthe legal community around justice andfairness in dealing with real-world cases.

73

Posner, after all, is a judge. The advantages ofthe wealth maximisation approach for Posnerare five-fold:

1. ‘the pursuit of wealth, based as it is on themodel of the voluntary market transaction,involves greater respect for individual choicethan in classical utilitarianism’ 62

2. ‘economic liberty…can be grounded morefirmly in wealth maximisation than inutilitarianism’ 63

3. ‘the wealth maximisation principleencourages and rewards the ‘Calvinist’ or‘Protestant’ virtues and capacities associatedwith economic progress’ 64

4. ‘wealth maximisation is a more defensiblemoral principle also in that it provides afirmer foundation for a theory of distributiveand corrective justice,’ along with a firmercommitment to the principle of rights than isevident in utilitarian and Kantian thinking’ 65

5. pragmatic: ‘we look around the world andsee that in general people who live insocieties in which markets are allowed tofunction more or less freely not only arewealthier than people in other societies buthave more political rights, more liberty anddignity, are more content…so that wealthmaximisation may be the most direct route to a variety of moral ends.’ 66

Posner makes a profoundly political judgementabout the benefits of the market solution fordealing with issues of justice. In essence heargues that arguments about the division of the'cake' must come second to efforts to increasethe size of the 'cake'. Interestingly, he does not,however, reject entirely the argument thatdistributional issues have a role in judgementsof fairness; it is more of an argument aboutranking positions. For Posner, the creation ofwealth comes first; the distribution of thatwealth comes second. First efficiency, then equity.

Of course there are significant drawbacks toboth Kaldor-Hicks and Posner in theirapproaches to assessment of validity forregulations. Firstly, the approaches balance

Part V: Getting regulation to work for consumers

gains and losses in terms of money. Thus onlythose benefits and losses that can be quantifiedin monetary terms can be effectively taken intoaccount when dealing with a proposedregulation or legal decision. Secondly, theapproaches take no explicit account ofdistributional justice. The approaches thusignore a significant characteristic in manyregulatory systems and approaches. While suchan avoidance of distributional equity as anissue may make life simpler in analytical terms,it provides problems for regulators attemptingto deal with real-world problems. Finally, iteffectively allows the imposition of losses onindividuals, provided that it is more thanbalanced by the gains afforded to beneficiaries.This problem is not far from the argumentabout distributional equity and is a factor inmany of the arguments made about equitableregulation.

Distributional equity is a broader social concerndriven by a desire to strive towards what Tobin67 has called 'specific egalitarianism'. This is theconcern to ensure that the distribution of aresource, or service is done in a manner that isless unequal than a simple ability to paysolution would provide for. However, adistributional equity solution may involvemeasures outside of an assessment of allocativeefficiency. As we discuss below, the decision-making process should first assess and aim forthe allocative efficiency position before thenconsidering what the distributional impactsmay be.

Attempts at reformingregulation

There have been numerous attempts atreforming the way regulation works. Muchimpetus for the reform of regulation has comefrom the business world and its perception thatregulation is imposed upon them. The impetuscan almost be said to be a result of the triumphof the argument that regulation arises as aresult of government failure and the desire ofthe bureaucracy to grow at the expense ofbusiness.

Two of the most far reaching examples of thederegulatory push have come from the US and

74

the UK. One of the first codified efforts to reinin regulation was the introduction by RonaldReagan of a cost-benefit analysis element to theexecutive order 12291. In Section 2 (GeneralRequirements) it stated that: ‘In promulgatingnew regulations, reviewing existingregulations, and developing legislativeproposals concerning regulation, all agencies tothe extent permitted by law shall adhere to thefollowing requirements:

1. Administrative decisions shall be based onadequate information concerning the needfor, and consequences of, proposedgovernment action;

2. Regulatory action shall not be undertakenunless the potential benefits to society for the regulation outweigh the potential costs to society;

3. Regulatory objectives shall be chosen tomaximise the benefits to society;

4. Among alternative approaches to any givenregulatory objective, the alternative involvingthe least net costs to society shall be chosen;and

5. Agencies shall set regulatory priorities withthe aim of maximising the aggregate netbenefits to society, taking into account thecondition of the particular industries affectedby regulations, the condition of the nationaleconomy, and other regulatory actionscontemplated for the future.68

While the cost-benefit-analysis approach wasradical for its time, the order was vagueenough to allow little progress to be made inreducing the burden of regulation. There wassome irony, for the proponents of governmentfailure to see a government trying to regulateitself against its best instincts.

In typically more muted fashion, the UKdeveloped its deregulation initiative in theearly 1990s. Instead of promulgating anexecutive order, the UK opted for guidelinescalled ‘Thinking about regulation: a guide togood regulation (1993)’. Within the guidelineswas a 10-point plan to achieving goodregulation. The 10 points for regulators to‘think about’ were:

1. Identify the issue… keep the regulation inproportion to the problem.


2. Keep it simple… go for goal-basedregulation

3. Provide flexibility for the future…set theobjective rather than the detailed way ofmaking sure the regulation is kept to.

4. Keep it short5. Try to anticipate the effects on competition

or trade.6. Minimise costs of compliance7. Integrate with previous regulations8. Make sure the regulation can be effectively

managed and enforced9. Make sure that the regulation will work and

that you will know if it does not.10.Allow enough time.

Much of the 10-point guideline reflectsestablished economic thinking about regulationbut falls short of imposing a time limit onregulation. The idea of ‘sunset’ clauses onregulation is not new and indeed someelements of trade regulation have sunsetclauses for trade protection measures.

75

Part V: Getting regulation to work for consumers

76


Conclusions

77

Conclusions

General understanding of the economics ofregulation too often focus on a simplisticreading of classical economics or smallgovernment. This tends to paint opponentsof a particular regulation into attacking allregulation and the proponents of regulationinto defending all forms of regulation. Whatis needed – and what must be the role of theconsumer organisation – is to balancelegitimate demands for regulation with aclear-headed understanding of the realdynamics of regulatory policy and its knock-on effects.

Most commentators on regulation have aclichéd view of the main developments andtend to fall back on hackneyed short-handversions of the theories. Most claims of theinefficiency and inappropriateness of regulationrely on a simplistic interpretation of theoriesthat the developers of the theory would notrecognise. Thus we find that opponents ofuniversal service or social goals rely on asimplistic reading of the Atkinson-Stiglitztheorem (which roughly argues that undercertain assumptions, income is betterredistributed through taxation than otherregulatory policy). While the theorem is auseful guide to understanding the efficiency ofregulation, it relies on assumptions that almostnever apply in reality (such as a lack ofconsumption externalities and a tax system thatpicks up all income). As with all theorems, thelessons from Atkinson-Stiglitz can be taken toofar and often are. When the assumptions arerelaxed, we find that there emerges anappropriate realm for regulation in dealingwith redistributive problems.

The idea that redistribution is not a properrealm for regulation has become a popularrefrain in business circles. There is a degree ofirony in this, as the Stigler/Peltzman approachclearly recognises that a major driver ofregulation is the tendency of firms to engage inbattles for rent (and thus redistribution of thegains of regulation). A recent article in theFinancial Times 69 provided a good example oflarge industries in monopoly positions seekingto remove regulation from their activities (onspurious ‘competitiveness’ and ‘foreigncompetition’ arguments) as a means ofmaximising their rent at the expense of other

78

businesses (their customers) and consumers.

The issue is not about whether redistributiveelements do exist in regulation but how theyshould be managed. We can learn somethingfrom the work on progressive taxation tounderstand where redistribution can play itselfout in regulation. Seligman 70 identified anumber of justifications for progressive taxation:

1. ‘Benefit Theory (also known as the ‘give andtake’ or ‘quid pro quo’ theory.) Protection isthe chief function of the State, and taxes canbe looked upon as the premiums one pays toa collective insurance agency that guaranteespeace and order. But the greater one’sproperty or income, the greater the benefitsreceived. However, the benefits areproportional to ‘clear income’, income lessexpenses. Hence the tax must be graduatedwith respect to total income in order tocorrespond to clear income. A variant to thistheory is that one should pay according tothe cost of the service that the State renders,and this cost increases with clear income.

2. Equal sacrifice theory. A tax is a sacrifice andwe should all sacrifice equally. However, a $1tax for someone making $10,000 yearly ofincome is not the same as a $1 tax forsomeone making $1mn yearly. To equalisethe sacrifice, the wealthy should pay a higherfraction of their income in taxes.

3. Faculty theory. Individuals in society shouldcontribute to society in proportion to theirfaculties or abilities. Income is a goodsurrogate for faculty or ability.

4. Tax ill-gotten gains. The rich get richprimarily by exploiting imperfections in thegovernment (e.g. bribes, hiring expensivelawyers to bypass taxes, getting insideinformation). They should pay more in taxesto offset this ill-gotten advantage.’

Such an approach works well in general socialregulation; the more pressing question is thedegree to which it applies in economicregulation. Here a distinction must be madebetween competition policy and sectoralregulation. In the case of the former, the policyobjectives and remits have become more

Conclusions

narrow over time. This focus has led to asituation where modern competition policyalmost conforms to the cleanest model ofregulation. It also conforms to the Atkinson-Stiglitz view of life; namely that markets arebest left to themselves and redistribution issuesare best left to politicians and governments.

In contrast, sectorally regulated industries havetended to either be created with specific socialand environmental goals in mind or had theseadded to them over time. For sectoralregulators, arguments for progressive taxationhave their sectoral corollaries. The tabulationbelow indicates the possible parallels:

One can see that sectoral regulators haveredistributive powers built into their remitsbecause the ‘market’ they are regulating issubject to failure, already provides a specificdistributional result and, left to its ownresources, would rebalance this distribution.Such a recognition makes the comparison ofcompetition and sectoral regulation interesting.It would tend to suggest that competitionpolicy is likely to have a problem in dealingwith sectors where the distribution of benefitshas been left unchallenged for a long time. Insuch sectors those that gain from thedistribution tend to create complexjustifications for their seizing of the rents ofregulation and usually wrap their arguments inthe claim of consumer protection. The recentarguments over Resale Price Maintenance ofover-the-counter medicines is a good exampleof this. Here the pharmaceutical industry andsome pharmacists justified price -fixing (andthus the misappropriation of the consumersurplus) on the basis that the consumer interestwas being protected. Much of the argument

79

against change was about the distribution ofrents; namely that competition wouldredistribute rent to retailers and away frommanufacturers. This is a familiar argument tojustify restrictions on parallel trade and theproper operation of the single market. It is alsoalways done in the name, rather than theinterest, of the consumer.

If redistribution is likely to be a part ofregulation, we need to address two issues: 1)what result is likely to emerge from this, and 2)how should this process be managed.

The result of incorporating redistributionfactors into the system of regulation, acceptingboth government and market failure, is likely tobring:

• ‘Public policy in social regulation willconform neither to a strict Rawlsianmaximin’ or ‘least worst’ policy nor to a strictutilitarian policy that would maximiseexpected-benefits-less costs. Rather it willconform to a perception-driven benefits-less-costs policy, with perceptions influenced bythe publicity given to rare events, thepoliticking of interest groups, and thedifficulty of formulating a public philosophyof dealing with uncertain, potentially fatalevents. We can expect politicians to try tocapitalise on this confusion of concepts,perceptions of reality, and special interests.

• As a corollary to point 1, economists who arefrustrated because policy makers to notadopt a straightforward ‘maximise expected-benefits-minus-expected-costs’ approach tosocial regulation will continue to befrustrated.

Conclusions

Progressive taxation and sectoral regulation compared

Personal taxation Sectoral regulation

Benefit theory Payment for collective RPI-X – paying the cost of benefits of society being granted monopoly rights

Equal sacrifice theory £1 to the poor is worth more Regulatory focus on consumersthan £1 to the rich

Faculty theory Tax should be paid on ability Distribution of pain in rebalancing to pay pricing formulae

Tax ill-gotten gains Punishment for beating system Windfall taxes

• ‘Grandfathering’ of status quo practice willbe the more frequent, the higher the cost ofchanging the status quo.

• Up to a certain, critical size, large societiesand organisations are more likely to enactsocial regulations or to obey enacted socialregulations than small ones. After the criticalsize is reached the probability of enactmentor obeying will be only very weakly sizedependant.

• Although policy makers will not adopt asimple ‘expected benefits minus expectedcosts’ formula, costs will enter into theperception driven, modified version referredto in point 1. Inventors of a safety devicecannot expect its adoption because ‘the valueof a human life is infinite’. It will be adoptedonly if it is cheap enough.

• Publicity will continue to play a large role ininfluencing perceptions of benefits and costsof social regulation.

• As a corollary of point 6, social regulationcan be expected to be self-limiting. Socialregulations with large cost burdens but smallbenefits will induce political action to repealor modify them.’ 71

Given that social and economic regulation islikely to be a messy and contested affair, onemust seek to identify the best way of managingthis problem. At the risk of adding to theinterminable lists of do’s and don’ts in the area,we recommend the following:

Targets for regulation

Pick on the obvious targetsGo for the most egregious examples ofproblems rather than regulate everything in sight.

Don’t over-extend or take things too farThe ‘final 10 %’ rule should always be borne in mind: Is it really worthwhile to extendcoverage of the regulation to everyoneirrespective of cost or is there a cheaper way of achieving the ends?

80

Process for regulating innormal markets

Start with efficiency and thecompensation principleAlways start clean before complicating matters.Use Kaldor-Hicks as the starting point: Is thechange going to generate sufficient benefits forthe winners to compensate the losers? Questionone should always be about whether this policyor regulation is efficiency enhancing.

Analyse distributional equityThe compensation principle is fine, as far as itgoes, but it does not go far enough. Regulatorsmust look at the distributional equity problemsthat arise from their ‘clean’ analysis of theproblem. Where that analysis takes theregulatory decision-making process cannot bepredetermined. Instead ,it should move to:

Look for ‘unfair’ redistribution ornegative externalitiesThis is the ‘should we care’ part of the analysis.As every regulatory decision involves someform of redistribution both from and tosomeone, one has to ask if it is worth worrying about. There are two clear exampleswhere we should; firstly, where theredistribution is unfair and where it results innegative externalities for some consumers.Unfairness is a subjective assessment and isoften best left to politicians. However, eachlegal system is based on culturally-rootedunderstanding of fairness, so it is legitimate tobring it in as a question. The issue of negativeexternalities is more complex. Just as allregulation should look toward the possibility ofunintended side effects, this question should bea part of the assessment process. This helps tonarrow down the distributional question tofairness and unintended losers.

Should a balance be struck and if sohow best can it be done?Given that every regulatory intervention willinvolve some redistribution of benefits,recognition must be made of the best means ofachieving that rebalancing. For example, in acomplex monopoly investigation into new carpricing, the rebalancing was betweenmanufacturers, dealers and consumers.

Conclusions

The manufacturers losing out could not be areason not to move. The better question washow best that rebalancing could be carried out.That can only be answered by asking;

Who can best deal with the spillover effects?

The division between how and who is aartificial, as they are inextricably linked. Onceone has analysed distributional implications,one must then identify how best to deal withthose implications. Here we have developed auseful rubric originally put forward by JusticeBreyer in the US about regulation in general(see above).

How will improvement inefficiency and redistributionaffect rent seeking?

What is clear from the literature is that firmsand interest groups will seek to redistributerents in their favour and at the cost of someoneelse. In carrying out this analysis, one mustalways be aware that rent-seeking activity isalways likely to be displaced rather thanremoved. The balance of power between thoserent seekers may be changed but such activitywill simply find a new home. If it becomesclear that rent-seeking is likely to be displacedto a government department, efforts should bemade to ensure that the department is

81

insulated as far as possible and givendemocratic accountability mechanisms, fromthe increased thrust of rent-seeking.

As can be seen from the typology ofdistributional problems, solution, actors andactions, the most common response to the‘problem’ posed by relying on this particularordering of questions is to ensure sufficientsafety nets in the marketplace. This wouldsuggest a split set of answers. In the regulatedindustries, the current distribution of powers issuch that the regulator itself is likely to have todeal with all aspects of the regulation,including its after-effects. However, incompetition policy, the spillover effects arealmost always going to be better dealt with byother regulators, namely governmentdepartments and politicians.

This characterisation of the problem poses apeculiar difficulty for the future direction ofregulatory policy. If one accepts that marketsand regulatory systems are prone to failure, beit government or market, one must accept thatsocietally designed balancing mechanisms forallocative efficiency and distributional equityhave to be acknowledged. What must also beacknowledged is that any attempt to ‘clean up’regulatory policy by making it focus moreclearly on allocative efficiency issues is likely tohave the effect of redirecting rent-seekingactivities to other regulatory agencies. Thus,attempts to make merger policy more focusedon competition and consumer welfare will

Conclusions

Matches of regulatory ends, means and player

Problem Tentative solution Actor Direction of regulation

Rent control Taxes; deregulation Government/ Government or market(excess profits) mixed market

Spillovers Marketable rights; Consumer Safety netted bargaining standards empowerment/ market

Government as ringmaster

Inadequate information Disclosure; screening; Governmental Safety netted marketstandards setting; standard settingbargaining

Other (moral hazard, Incentive-based Regulator/market Safety netted marketunequal bargaining regulation; standardspower, paternalism)

redirect rent-seeking by large firms towardinsulating themselves through industrial policy changes.

Regulation is neither a clean nor an orderlyprocess. The literature on regulation shouldmake us sceptical of both industry andregulatory claims for legitimacy in seeking newor extended regulation. While it is true thatconsumers almost always pay the cost of over-or under-regulation, this is not on its own ajustification for total deregulation. Markets failpure and simple and relying on somedoctrinaire belief in their infallibility is assimplistic as relying on regulators to order ourlives for us. Firms will seek regulation and willseek to impose costs on consumers;government departments and regulators willseek to extend their power and budgets beyondthe reasonable. All this is clear from theliterature on the issue. What is not clear is howto deal with it.

Most economists have tended to argue for thestricter application of efficiency based tests.This is an important first step. However, it cannot be the only step. Recognition has to bemade that relying solely on efficiencyarguments will generate problems ofdistributional equity and externalities of theirown. A more explicit ordering of the problemmust recognise that pure solutions cannot beaimed for. Above all, regulatory discussionshould recognise that the choice is not simplybetween State and market nor betweenregulation and deregulation but between the‘disagreeable and the intolerable’. 72 While it istrue that market-based mechanisms arepreferable for improving the allocation ofresources in a normal market, it does not followthat distributional consequences can beignored. Thus there will always be someregulatory role even in the most vigorousmarket. As we signalled above, this role islikely to be the operation of effective ‘safetynet’ style regulation – the classic domain ofconsumer protection and promotion laws andpolicies. It follows that whichever solution isaimed at, it can only work if the consumerengaged in the market in question issufficiently empowered and protected to makeit work effectively.

82

Conclusions

Appendices

83

Appendices

The table below gives a general sketch ofelements to look for in a market assessment.

84

Appendices

Appendix I:behaviour,structure,strategy/conduct,performance, policy

Element Evidence Drivers for change

Consumer behaviour Framing and information

Cultural aspects to decision making

Consumer learning

Endowment effects

Sunk costs and the theory of momentum

Relative search costs

Costs of regret

Self-control and pre-commitment

The treatment of losses and gains

Structural elements Number of buyers

Number of firms

Size of firms

Barriers to entry

Cost conditions and structure

Diversification

Integration

Concentration

Product characteristics

Conduct Pricing behaviour

Advertising

Production/marketing strategy

Cartel/collusion

Innovation

Legal arrangements

Performance Consumer satisfaction

Efficiency

Company growth

Market share

Profitability

Public policy regulation

Competition laws

85

something until you’ve asked five times: ‘why?’This rule-of-thumb helps to separate meansobjectives from fundamental objectives. Forinstance, if one is dealing with a consultationon Universal Service Obligations, examinationof “why?” will move the issue from merelyprotecting the tool of the USO to assessing howto best deal with social exclusion (or thepertinent fundamental objective). This takes theissue from a ‘means to an end’ to the desiredend itself and allows for a wider mapping ofpossible solutions. While it does not, however,necessarily provide the right (or different)answer, it does encourage a broader view.

If we can get the decision-making process morefinely tuned, we can hope to avoid the trapsidentified below, many of which are common togeneral decision-making processes. Raiffa et alidentify the eight most common errors indecision-making, all of which are valid fordealing with understanding markets:

• working on the wrong problem• failing to identify your key objectives• failing to develop a range of good, creative

alternatives• overlooking crucial consequences of your

alternatives• giving inadequate thought to tradeoffs• disregarding uncertainty• failing to account for your risk tolerance• failing to plan ahead when decisions are

linked over time.

A number of more pernicious problems alsohamper incisive analysis. These are morecommonly thought of as psychological traps:

• The Anchoring Thoughts Trap: over-relianceon the first information received about aproblem/issue

• The Status Quo Trap: keeping on doing whathas always been done

• The Sunk Cost Trap: current choices tend tobe biased by past choices, the tendency tojustify past decisions with current decisions

• The Confirming-Evidence Trap: seeking outthat evidence that confirms your originalview and under-weighting evidence thatindicates the opposite

• The Framing Trap: how is the questionposed? ‘Is it right that 5 million people will

Appendix II: Hints on decision-making and the analytical approach

Appendix II:Hints on decision-making and theanalytical approach

To borrow from Howard Raiffa 73, a gooddecision-making process:

• focuses on what’s important• is logical and consistent• acknowledges both subjective and objective

factors and blends analytical with intuitivethinking;

• requires only as much information andanalysis as is necessary to resolve a particulardilemma

• encourages and guides the gathering ofrelevant information and informed opinion

• is straightforward, reliable, easy-to-use andflexible

Raiffa et al describe the decision-makingprocess by dividing it into six stages (with threeadded variables). As with all goodmanagement techniques, the approach issummed up as an mnemonic: PrOACT.

• Problem: what is the real problem beingassessed?

• Objectives: what are the real objectives adecision?

• Alternatives: map all alternatives• Consequences: try to map all consequences

for all possible alternatives• Tradeoffs: recognise and deal with all

possible tradeoffs:• Uncertainty: do some crystal-ball gazing;• Risk tolerance: consider risk tolerance in

relation to likely outcomes• Linked decisions: what are the side effects of

a decision likely to be?

Raiffa also points to another handy tool foranalysts, as embedded in the wisdom of theJapanese saying: ‘You don’t truly understand

86

Competition analysis produces results that canbe interpreted in multiple ways. For example, amarket with largely uniform prices can be aresult of fierce competition or tight collusion.Steven Salop 74 identified a number ofcommon errors made in the definition ofmarket power. Salop characterised these errorsas following analytical traps:

• The Marginal Cost Trap: Mistaking a firm'sinability to profitably raise price above itsmarginal cost for an inability to exercisemarket power by excluding rivals.Conversely, mistaking a firm's ability toprofitably raise price above its marginal costfor an ability to exercise additional marketpower by adopting alleged anti-competitivebehaviour.

• The Cellophane Trap: Mistaking a firm'sinability to exercise market power by raisingprice above the current price for an inabilityto have already exercised market power byraising price up to the current level, therebymislabelling a completed anti-competitive actas a lack of market power.

• The Price-Up Trap: Mistaking a firm'sinability to profitably raise price above thecurrent level for an inability to exercisemarket power by preventing competitor'sconduct that otherwise would reduce pricebelow the current level, thereby mislabellinga maintenance of market power as a lack ofmarket power.

• The Threshold Test Trap: Mistaking a firm'sinability to profitably raise price above thecurrent level because of current competitiveconstraints from certain rivals for an inabilityto exercise market power even after thoserivals are excluded.

Appendices

Appendix III:Possible errors in marketassessments

not gain from this market liberalisation?’ or‘Is it right that 50 million people will gainfrom this market liberalisation?’ (same issue,different questions)

• The Overconfidence Trap (tied to theanchoring problem): tendency to forecastmarket behaviour based on initial forecastswithout dealing properly with all availableinformation

• The Recallability Trap: dramatic events skewviews of future probabilities (e.g. planecrashes vs. car crashes)

• The Base Rate Trap: what are the underlyingassumptions about probabilities and are they robust?

• The Prudence Trap: opting for worst casescenarios ‘just to be safe’

• The Outguessing Randomness Trap:mistaking luck for a pattern – random thingshappen in a random manner;

• The Surprised-by-Surprises Trap:coincidences happen

Conclusion

Markets are extremely complex entities and itwould do well to follow these simple decision-making rules when dealing with them:

• Follow decision making guidelines;• Avoid common traps and retain objectivity

wherever possible;• Break complex markets into comprehensible

bundles;• Continually zoom in and out of the bundles

to ensure that you can see the whole from thesum of its parts.

87

Ideas discussed in Chapter I rest on workcarried out in behavioural and experimentaleconomics and psychology, as summarised below.

The classical model of consumer behaviourrests on the belief that consumers act tomaximise their utility: ‘All human behaviour canbe viewed as involving participants who [1]maximise their utility [2] from a stable set ofpreferences and [3] accumulate an optimumamount of information and other inputs in avariety of markets.’ 78

Over time this assumption has come underincreasing challenge. In competition analyses,the classical view of the utility-maximisingconsumer can lead to demands that consumerbe given ever more information on which tomake rational judgements.

As Kahneman et al argue, there are two waysin which the traditional approach to consumerbehaviour in economics can be altered. One isto change the external environment facing theconsumer as decision-maker by introducingsuch factors as information imperfections,search and transactions costs, and risk. Thesecond, more controversial approach is to alterthe behavioural assumptions underlyingconsumer behaviour by adding in motives orincluding the possibility of apparentlyirrational behaviour. However, the erosion ofthe traditional approach has not been complete,for at least two simple reasons: ‘First, addingcomplexity to the model of the agent generallymakes it more difficult to derive unequivocalpredictions of behaviour from a specification of theenvironment. Second, there is a threat of a slipperyslope. It appears all too easy to lengthen the lists of

Appendix III: Possible errors in market assessments

Appendix IV:Factoring inconsumerbehaviour

• The Unilateral SSNIP (75) Trap: Mistaking afirm's inability to profitably raise price abovethe current level unilaterally (assuming thatrivals do not change their prices or outputs)for an inability to exercise market power byconduct that affects rivals' output and priceresponses. The Unilateral SSNIP Trap isclosely related to the Threshold Test Trap.The two traps are distinguished because theerrors to which they refer occur in differentcontexts. For example, the unilateral SSNIPTrap could arise in a horizontal agreementcase as well as in an exclusion case. TheUnilateral SSNIP Trap focuses on the marketdefinition methodology, whereas theThreshold Test Trap focuses on an erroneousfinding of market power.

A further problem for market analysis is thedefinition of errors related to predation. Thekey problem in identifying predation is pickingnormal competitive responses from predatorybehaviour. This has lead Joskow and Klevorick76 to identify two types of errors 77. that can bemade in predatory pricing claims:

• Type I error: the identification of acompetitive price cut as a predatory price cut.

• Type II error: the failure to detect predatorypricing

Avoiding commission of Type I or Type II erroroften comes down to a matter of judgementand experience as much as analytical rigourand data. This is often due to a lack of data thatwould allow a proper analysis of a market event.

non-economic motives or cognitive errors thatmight affect economic behaviour.’ 79

Three key caveats need to be placed on theclassical approach:

• Bounded rationality: human cognitiveabilities are not infinite and we all have bothlimited computational skills and flawedmemories;

• Bounded willpower: people often takeactions in the short term that they know to bein conflict with their own long-term interests;

• Bounded self-interest: people generally care,or act as if they care, about others, evenstrangers, in some circumstances. 80

Bounded rationalityThe theory of bounded rationality was firstespoused by the late Herbert Simon. The theoryargues that: ‘The capacity of the human mindfor formulating and solving complex problemsis very small compared with the size of theproblems whose solution is required forobjectively rational behaviour in the real world– or even for a reasonable approximation tosuch objective rationality.’ 81

Framing and informationIn their introduction to their edited collection of essays, Hogarth and Reder point out that:‘…in discussing choice anomalies that could beattributed to ‘framing’ effects, Tversky andKahneman make a distinction between what theyterm ‘transparent’ and ‘opaque’ versions of choiceproblems. Briefly stated, when a problem ispresented in transparent form, choice behaviourdoes not violate basic tents of rationality. Whenchoice problems are formulated in an opaquemanner, however, people may well violate basicprinciples … because of the effect of ‘framing’ and so on.’ 82

Cultural aspects to decision makingKey elements to the understanding of culturalelements to decision-making include:• What is the culture of the market

in question?• In what sub-culture does the market sit?• How will changes to that market affect the

cultural interactions of participants?

88

Consumer learningAs Tversky and Kahneman state: ‘… effectivelearning takes place only under certain conditions;it requires accurate and immediate feedback aboutthe relation between the situational conditions andthe appropriate response. The necessary feedback isoften lacking for the decisions faced by managers,entrepreneurs, and politicians because (i) outcomesare commonly delayed and not easily attributableto a particular action; (ii) variability in theenvironment degrades the reliability of the feed-back, especially where outcomes of low probabilityare involved; (iii) there is often no informationabout what the outcome would have been if anotherdecision had been taken; and (iv) most importantdecisions are unique and therefore provide littleopportunity for learning ….’ 83

Bayes Theorem is useful in assessing learning.The theorem states that when an individual hasto make a decision about the likelihood of aparticular outcome, he or she will start with aset of beliefs about the probability based onpast experience and estimations. The newdecision will apply these prior beliefs to thenew situation and update the probability of anevent occurring. In other words, the BayesTheorem tells us that people learn from theirpast experiences and use this to update theirbeliefs in similar situations.

In addition to the factors in informationtransmission outlined above, the distance intime between particular decisions will have asignificant effect, as will the type of issue aboutwhich decisions are made.

Prospect Theory established a set ofexperiments to test violations of the theory.Tversky and Kahneman came to three generalconclusions:

• The structure of a problem may affect thechoices that are made. The same problempresented in different ways may influencethe decisions of participants.

• Outcomes received with certainty are over-weighted compared to outcomes that areuncertain.

• Gains get treated differently to losses. Lossesgenerate a risk seeking response while gainsproduce a risk averse response.

Appendices

Richard A Thaler 84 generated a number ofexperiments that produced the followingresults:

Endowment Effect: Any product that is part ofthe already existing endowment of theindividual will be more highly regarded than aproduct that is not. Individuals thus tend torate what they already own more than aproduct that they do not.

Sunk Cost Effect: A sunk cost is an alreadyborne cost that is not easily recoverable.Traditional economic theory argues that sunkcosts are irrelevant to current decisions,focusing instead on incremental costs andbenefits. Thaler argued that individual sunkcosts did in fact affect decision-making. He alsolinked it to:

Theory of Momentum has been important inthe study of international negotiation andargues that individuals will complete a taskonce work has begun, irrespective of thecontinuing validity of the decision.

Search costs are relative: Any difference inprice between goods is seen in relation to thetotal price of the goods.

Psychic costs of regret are large: Presentdecisions can often be limited on the basis ofthe individual not being able to trustthemselves to make the right decision in the future.

Consumers can show self-control and pre-commitment: Consumers often recognisethat their existing consumption patterns areincapable of meeting certain future needs. Thisprompts saving and tying the consumer intopatterns of committed expenditure.

Losses and gains are treated differently:Thaler identified four essential options fordealing with combinations of losses and gains:

• segregate gains: individuals prefer to treatmultiple gains as a series of individual gains.

• integrate losses: individuals like to place alltheir losses in one basket.

89

• let big gains cancel small losses: if the overallbalance of gains and losses is toward thegain, then the losses should be pooled withthe gains to cancel them out;

• segregate ‘silver linings’: when large lossesout-weigh small gains, the gains may beseparated out as a ‘silver lining’ to the cloudof the large loss.

Appendix IV: Factoring in consumer behaviour

90

Appendices

Appendix V:Structure-conduct-performanceparadigm

SupplyRaw materials

TechnologyUnionisation

Product durabilityValue/weight

Business attitudesLegal framework

DemandPrice elasticity

SubstitutesRate of growth

Cyclical and seasonalcharacter

Purchase methodMarketing type

Market structureNumber of buyers and sellers

Product differentiation Barriers to entry

Vertical integrationDiversification

ConductPricing behaviour

Product strategy and advertisingResearch and innovation

Plant investmentLegal tactics

PerformanceProduction and allocative efficiency

ProgressFull employment

Equity

Public policyInternational trade

rulesRegulation

Price controlsAntitrust

Information provision

1 Ada Leverson. Letters to the Sphinx (1930) pt. 1. The Importance of Being Oscar.

2 Adam Smith, The Wealth of Nations, 1776.Book 4, chapter 8.

3 John Maynard Keynes. 1936. The GeneralTheory of Employment, Interest and Money.Macmillan and Co. London. Quoted in GeraldP O’Driscoll and Mario J Rizzo. 1985. TheEconomics of Time and Ignorance. Routledge.London.

4 Gerald P O’Driscoll and Mario J Rizzo. 1985.The Economics of Time and Ignorance. Routledge.London. p9.

5 Adam Smith. 1991. The Wealth of Nations,Book I, Chapter X, p130. Everyman's Library.

6 See Harold Demsetz. 1974. ‘Two Systems ofBelief About Monopoly.’ In Oliver EWilliamson (ed). 1990. Industrial Organisation.Edward Elgar. London.

7 Tobin J. 1970. 'On limiting the domain ofinequality'. Journal of Law and Economics. Vol13 (October, pp 263-77. Quoted and discussedin Blank, RM. 2000. 'When Can Public PolicyMakers Rely on Private Markets? The EffectiveProvision of Social Services.' The EconomicJournal (March). Blackwell Publishers, Oxford.

8 FM Scherer. 1990. Industrial Market Structureand Economic Performance. 3rd Edition.Houghton Mifflin. New York.

9 Gary S Becker. 1976. The Economic Approach toHuman Behaviour.

91

10 Christine Jolls , Cass R Sunstein, and RichardThaler. 1999. A Behavioural Approach to Law andEconomics. University of Chicago Law School.Working Paper.

11 In United Brands v Commission. The EuropeanCourt of Justice found that in relation to Article86 (on abuse of dominant position) cases mustbe considered “with reference to a clearlydefined geographic area in which [the productor service] is marketed and where theconditions are sufficiently homogeneous for theeffect of the economic power of theundertaking concerned to be able to beevaluated” (Case 27/76 [1978] ECR 207, [1978]CMLR 429 at paras 10 and 11).

12 This is a more nuanced and potentiallyforgiving model than that adopted in the EU.The classic statement of this issue in Europeanjurisprudence comes in Tetra Pak II [OJ (1992) L72/1, (1992) 4 CMLR 551]. This case concernedthe creation of a dominant position in aparticular type of packaging. The ECJ held thatsubstitutability between products should be lookedat over a short period of time. The Courtsuggested that the substitution through shiftingfrom one form of packaging to another wouldinvolve changes in consumer habits and wouldonly occur over the long term. It held that anysignificant change in consumer/customerbuying habits that would require a long timewas not relevant to the facts of the case, andthat only current patterns of consumer/customer behaviour could be influential indeciding the case and that speculative futurepatterns should be discounted.

13 See above for similar assessment for CR4cases.

14 John R Baldwin. 1995. The Dynamics ofIndustrial Competition: A North AmericanPerspective. Cambridge University Press.Cambridge.

15 Baldwin. op cit p160

16 Paul Dobson, Michael Waterson, and AlexChu. September 1998. The Welfare Consequencesof the Exercise of Buyer Power. Office of FairTrading. London.

Endnotes

Endnotes

17 W.J. Baumol, J.C. Panzar and R.D. Willig,Contestable Markets and the Theory of IndustryStructure, Harcourt, Brace, Jovanovich, NewYork, 1982, revised 1988.

18 Baldwin.1995. op cit. p81.

19 Drew Fudenberg and Jean Tirole. 1984. ‘TheFat Cat Effect, The Puppy-Dog Ploy, and theLean and Hungry Look’. Reproduced in OliverE Williamson (ed). 1990. Industrial Organisation.Edward Elgar. London.

20 ‘Predation in the Airline Industry’. Remarksby Roger W. Fones Chief Transportation,Energy, and Agricultural Section AntitrustDivision U.S. Department of Justice Before theAmerican Bar Association Forum on Air andSpace Law Seattle, Washington June 12, 1997.

21 Becton Dickinson UK Ltd. OFT, 15 June 1988.

22 OFT, 15 June 1988. Quoted in Whish op cit,1993.

23 Alfred Kahn. 1985. ‘The MacroeconomicConsequences of Sensible MicroeconomicPolicies.’ Society of Government EconomistsNewsletter. May 1985, p2.

24 P. Joskow and A Klevorick. 1979. ‘AFramework for Analysing Predatory PricingPolicy. Yale Law Journal. Vol 89. pp213-270.

25 For more on errors, see above.

26 Ordover and Willig. op cit. pp10-12.

27 Areeda/Turner argued that the ‘Treatment ofpredatory pricing in the cases and the literaturehowever, has commonly suffered from twointerrelated defects: failure to delineate clearlyand correctly what practices should constitutethe offence, and exaggerated fears that largefirms will be inclined to engage in it.’ Theiranalysis of predation centred on the belief that‘...predation in any meaningful sense cannotexist unless there is a temporary sacrifice of netrevenues in the expectation of greater futuregains.' Such a line of behaviour would notmake sense unless the predatory firm had ‘(1)greater financial staying power than his rivals,and (2) a very substantial prospect that the

92

losses he incurs in the predatory campaign willbe exceeded by the profits to be earned after hisrivals have been destroyed.’ P. Areeda and D. FTurner. ‘Predatory Pricing and Related Practicesunder Section 2 of the Sherman Act.’ HarvardLaw Review Vol 88, p697.

28 B.S Yamey. 1972. ‘Predatory Price Cutting:Notes and Comments,’ Journal of Law andEconomics, Vol.15 pp129-42. Quoted in LouisPhilips. 1995. Competition Policy: A GameTheoretic Perspective. Cambridge UniversityPress. Cambridge.

29 Philips. op cit. p220.

30 R Selten. 1978. ‘The Chain Store Paradox’,Theory and Decision. 9, pp127-159.

31 D.M Kreps and R. Wilson. 1982. ‘Reputationand Imperfect Information.’ Journal of EconomicTheory. Vol. 27, pp 253-279.

32 Kreps and Wilson. 1982. op cit. p254.

33 Yun Joo Jung, John H Kagel and Dan Levin.‘On the Existence of Predatory Pricing: AnExperimental Study of Reputation and EntryDeterrence in the Chain Store Game.’ p73.

34 FM Scherer and D Ross. 1990. IndustrialMarket Structure and Economic Performance (3rdedition). Houghton Mifflin. New York. p53

35 P Selznick. Focusing Organisational Researchon Regulation, in Roger Noll (ed) RegulatoryPolicy and the Social Sciences 1985 p363, in Ogus

36 Friedman. 1982. Quoted in Edward E Zajac,Political Economy of Fairness. 1995. MIT Press

37 Richard I Gordon. Regulation and Economicanalysis: a critique over two centuries. KluwerAcademic Press 1994. p22

38 DR Harris and C Veljanovski ‘Liability forEconomic Loss in Tort’ in Furmston p48

39 Ogus op cit. p30.

40 Dworkin Paternalism, in R Wasserstrom (ed)Morality and the Law. 1971 p108. Quoted inOgus op cit.

Endnotes

41 Public goods defined as having ‘two criticalproperties: First, no additional costs areinvolved in providing the good to an additionalperson (formally, the good has zero marginalcosts and is referred to as being“nonrivalrous”). Second, it is impossible toexclude individuals from benefiting from thegood (formally, the good is “nonexcludable”).A classic example of a public good is nationaldefence: Defending 270 million people does notnecessarily cost more than defending 260million people, and it is generally not possibleto exclude anyone from the benefit of nationaldefence.’

42 Stigler, The theory of economic regulation.

43 Peltzman. Toward a more general theory ofregulation. Journal of Law and Economics 19 Aug1976 211-40)

44 Edward E Zajac. Political Economy of Fairness.1995. MIT Press.

45 Buchanan, Tollison, and Tullock, 1980 ix

* David B. Johnson (1991, pp329-30)

46 Nicholas Mercuro and Steven G MedemaEconomics and the Law: from Posner to Post-Modernism. 1997. Princeton University Press.p98.

47 Bootleggers and Baptists in Retrospect: Themarriage of high-flown values and narrowinterests continues to thrive. By Bruce Yandle.Regulation Magazine. 1983

48 Stephen Breyer: Regulation and its Reform1982. Harvard Univ Press. p342

49 James Q Wilson. 1980. The Politics ofRegulation, in James Q Wilson (ed), The Politicsof Regulation New York Basic Books, 1980.

50 Mercuro and Medema op cit. P86/87,(Buchanan, 1978, p12).

51 David B. Spence and Lekha Gopalakrishnan.The New Political Economy of Regulation: Lookingfor Positive Sum Change in a Zero Sum World.Working Paper 990415. University of Texas atAustin.

93

52 Charles Wolf Jr. 1988. Markets or governments:Choosing between imperfect alternatives. MITPress. p37-38

53 Wolf op cit p51.

54 Wolf op cit p62.

55 Wolf op cit. p85.

56 Wolf op cit. p87.

57 Wolf, op cit p58.

58 Nicholas Mercuro and Steven G MedemaEconomics and the Law: from Posner to Post-Modernism. 1997. Princeton UniversityPress. p97

59 Mercuro and Medema, op cit, p19.

60 Mercuro and Medema, op cit, p50.

61 Posner 1975, p777; quoted in Mercuro op cit).

62 the pursuit of wealth .. Posner, 1983, p66; inMercuro op cit).

63 ‘economic liberty…Posner, 1983, p67; inMercuro op cit).

64 ‘the wealth maximisation principle (Posner,1983, p68; in Mercuro op cit).

65 ‘wealth maximisation ’ (Posner, 1983, p69; inMercuro op cit).

66 pragmatic: (Posner, 1990, p382; in Mercuro op cit).

67 Tobin J. 1970. 'On limiting the domain ofinequality'. Journal of Law and Economics. Vol13 (October, pp 263-77. Quoted and discussedin Blank, RM. 2000. 'When Can Public PolicyMakers Rely on Private Markets? The EffectiveProvision of Social Services.' The EconomicJournal (March). Blackwell Publishers, Oxford.

68 Viscusi, Vernon, Harrington . 1995.Economics of regulation and antitrust. p25

Endnotes

69 CBI wants watchdogs' powers curbed ByKevin Brown and Brian Groom, FinancialTimes, May 22 2001

70 Seligman, quoted in Edward E Zajac PoliticalEconomy of Fairness. 1995. MIT Press.

71 Zajac op cit. p158.

72 Wolf op cit. P87.

73 John S Hammond, Ralph L Keeney andHoward Raiffa, 1999. Smart Choices: A PracticalGuide to Making Better Decisions. HBS Press.Boston, Mass.

74 Steven C Salop. 2000. The First PrinciplesApproach to Antitrust, Kodak and Antitrust at theMillennium. Georgetown University LawCenter.http://papers.ssrn.com/paper.taf?abstract_id=195490

75 SSNIP is a Small but Significant non-transitory Increase in Price. These arise in theUS merger guidelines documents as anindication of market power. A SSNIP indicatesmarket power; absence of an SSNIP indicates alack of market power.

76 P Joskow and A Klevorick. 1979. ‘AFramework for Analysing Predatory PricingPolicy. Yale Law Journal. Vol 89. pp213-270.

77 For more on errors, see above.

78 Gary S. Becker. 1976. The Economic Approachto Human Behaviour.

79 Kahneman, Daniel, Knetsch, J L and Thaler, RH. ‘Fairness and the Assumptions ofEconomics,’ in Hogarth, Richard M and Reder,M.W (eds). Rational Choice: The ContrastBetween Economics and Psychology. University ofChicago Press. London. 1987.

80 Christine Jolls, Cass R Sunstein, and RichardThaler. 1999. A Behavioural Approach to Law andEconomics. University of Chicago Law School.Working Paper.

94

81 Simon, H. 1957. Models of Man. Wiley. NewYork. p198.

82 Hogarth, Robin M and Reder, M.W.‘Introduction: Perspectives from Economics andPsychology,’ in Hogarth, Richard M and Reder,M.W (eds). Rational Choice: The ContrastBetween Economics and Psychology. 1987.University of Chicago Press. London. p7.

83 Tversky, Amos and Kahneman, D. ‘RationalChoice and the Framing of Decisions,’ inHogarth, Richard M and Reder, M.W (eds).Rational Choice: The Contrast Between Economicsand Psychology. University of Chicago Press.London. 1987. pp90-91.

84 Thaler, R A. Quasi Rational Economics. RusselSage Foundation. New York. 1991.

Endnotes

Consumers International (CI) supports, linksand represents consumer groups and agenciesall over the world. It has a membership of over250 organisations in 115 countries. It strives topromote a fairer society through defending therights of all consumers, especially the poor,marginalised and disadvantaged, by:

• supporting and strengthening memberorganisations and the consumer movementin general

• campaigning at the international level forpolicies which respect consumer concerns.

Consumers International was founded in 1960as the International Organisation of ConsumersUnions (IOCU) by a group of nationalconsumer organisations. The group recognisedthat they could build upon their individualstrengths by working across national borders.The organisation rapidly grew and soonbecame established as the voice of theinternational consumer movement on issuessuch as product and food standards, health andpatients’ rights, the environment andsustainable consumption, and the regulation ofinternational trade and public utilities.

Consumers International is an independent,non-profit organisation. It is not aligned with orsupported by any political party or industry. Itis funded by fees from member organisationsand by grants from foundations, governmentsand multilateral agencies.

Consumers International’s Head Office is basedin London. It has Regional Offices in KualaLumpur (Malaysia), Santiago (Chile) andHarare (Zimbabwe), and in London.

95

What is Consumers International?

What is ConsumersInternational?

Head Office24 Highbury Crescent, London N5 1RX, UKTel: +44 20 7226 6663 Fax: +44 20 7354 0607E-mail: [email protected]

Office for Asia & the Pacific 5th Floor Wisma WIM, 7 Jalan Abang Haji Openg, Taman Tun Dr. Ismail, 60000 Kuala Lumpur, MalaysiaTel: (+603) 7726 1599 Fax: (+603) 7726 8599E-mail: [email protected]

Office for Latin America & the Caribbean Casilla 9635, Santiago, ChileTel: (+56-2) 335 1695 Fax: (+56-2) 231 0773E-mail: [email protected]

Office for Africa Private Bag A6215, Avondale, Harare, ZimbabweTel: (+263-4) 302 283 Fax: (+263-4) 303 092E-mail: [email protected]

Office for Developed and Transition Economies 24 Highbury Crescent, London N5 1RX, UKTel: (+44 20) 7226 6663 Fax: (+44 20) 7354 0607E-mail: [email protected]

For further information please visit our website: http://www.consumersinternational.org

The consumer guide to competition - International Development

Documents

Transcript of The consumer guide to competition - International Development