IIIIIIII

202 .2 99CR

Creating Performance IncentivesThrough Regulation andBenchmarking

.**»M0É0^

1999

maËÈÈmmzssm

Creating Performance Incentives ThroughRegulation and Benchmarking

Moderator:

Speakers:

Penelope Brook Cowen, Senior Private SectorDevelopment Specialist, Prívate Sector DevelopmentDepartment, The Woríd Bank

Tony Ballance, Chief Economist, Office of WaterServices, (OFWAT), (United Kingdom)

Warríck Smith, Manager, Private Sector DevelopmentDepartment, The World Bank

1999 Water Supply & Sanitation ForumApril 8 -9 ,1999

Water & Sanitation DivisionThe World Bank

LIBRARY IRCDox 9Q190, 2509 AP TMC HAGUE

T lTel.: +31 70 30 689 80Fax: +31 70 35 899 64

BARCODE: /

IIIIIIIIIIIIIIIIIII

Table of Contents

Introductory Notes iii

Biographies v

Speaker Handouts:

Creating Performance Incentives Through Regulation

And Benchmarking - The UK Water Model 1

Additional Readings:

Public Policy for the Private Sector:Regulating Water Companies 21Public Policy for the Private Sector:Utility Regulators - The Independence Debate 25

Public Policy for the Private Sector:Utility Regulators - Roles and Responsibilities 29

Public Policy for the Private Sector:Utility Regulators - Decisionmaking Structures,Resources, and Start-up Strategy 33

Economic Notes: Regulating Brazil's Infrastructure -Perspectives on Decentralization 37

The CATO Journal - Selected articles 49

Sixth International Trainingram on Utility RegulationAnd Strategy 71

IIIIIIIIIIIIIIIIIII

Introductory Notes

More and more governments are turning to the private sector for assistance inimproving the efficiency and coverage of water and sanitation services. Butsuccessfully engaging the private sector - and, in particular, persuading the privatesector to take on commercial and investment risk in water and sanitation provision- depends on the broader policy and institutional environment in a country. Thissession will focus on a central element of this environment: the institutions put inplace to monitor and regulate service providers.

Investments in water and sanitation networks and production/treatment facilities arehighly capital-intensive, and typically amortize over long periods. They are also inlarge part what economists refer to as "sunk" investments - they cannot easily bepicked up and taken elsewhere if business goes poorly, or if politicians intervene andchange the rules of the game. Accordingly, private companies are unwilling to makesubstantial investments in water and sanitation unless they have confidence ofmaking a reasonable return on these assets over their usefiil life. But this is a realitythat combines poorly with traditional approaches to tariff-setting for water andsanitation in most developing countries. On average, tariffs in the sector indeveloping countries cover less than half the costs of delivering water services, andpoliticians are often highly resistant to increasing tariffs to cover costs, and tomaintaining cost recovering tariffs over time. Governments seeking to encourageprivate sector participation in water and sanitation - and, in particular, privateinvestments in the upgrading and expansion of services - thus often face the dualproblem of raising tariffs to cost recovery levels before contracting, and putting inplace regulatory arrangements that will credibly yield sufficient tariffs over the lifeof a private sector contract.

Giving investors a credible assurance that they will earn a reasonable return on theirassets is, of course, just one side of the story. Regulators must balance the interestsof investors against the interests of consumers - seeking both to protect the latterfrom potential abuses of monopoly power by service providers, and to create andmaintain incentives on providers to pursue efficiency improvements in operationsand in new investments.

In most developing countries, governments face serious problems in designing andimplementing regulatory arrangements that are (a) endowed with sufficient capacityto operate effectively, and (b) viewed by both prospective investors and consumersas fair and credible. These problems are often particularly pressing in the water andsanitation sector (by comparison with other infrastructure sectors), because of suchfactors as the relatively high degree of decentralization of the sector, and frequentlypoor information about the state of sector assets at the beginning of PSP processes.

ill

Governments have two main tools at their disposal in seeking to design regulatoryregimes to achieve these broad objectives:

The powers that are conferred on regulators and the rules that are set out to guideregulatory decisions; andThe design of the regulatory institutions themselves.

This session will focus on possible approaches to these two issues in the water andsanitation sector in developing countries. The first speaker, Dr Anthony Balance(Chief Economist of OFWAT), will talk about ways of designing and implementingregulatory rules to create strong performance incentives for service providers,including the use of benchmarking. The second speaker, Warrick Smith (Managerof the Private Participation in Infrastructure Group at the World Bank) will talkabout options for designing regulatory agencies so as to reinforce their independencefrom political intervention, and in particular, options for developing credibleregulatory regimes in decentralized systems. Commentary will be provided by asmall panel of regulators from developing countries.

Penelope Brook CowenSession Leader

IV

Biographies

PENELOPE J. BROOK COWEN

Ms. Brook Cowen is a Senior Private Sector Development Specialist with thePrivate Participation in Infrastructure Group at the World Bank. She has beeninvolved in advising on the design and implementation of private sector contracts ina number of countries, including Albania, Bolivia, Ghana, the Philippines,Slovenia, and South Africa. Her interests include the design of private sectorparticipation to accommodate the needs of countries with severe governanceproblems, and the tailoring of private contracts and regulatory structures to ensureimproved access for low-income households. She led the preparation of theToolkits for Private Participation in Water and Sanitation, published by the WorldBank in 1997. Before joining the World Bank, Penelope was an Associate Directorin the Investment Banking Division of CS First Boston New Zealand. There, shewas responsible for economic and commercial advice to government and privatesector clients on regulatory reform and privatization, across a range of sectors.From 1988 to 1990, she served as a founding director of Trans Power NewZealand, a state-owned enterprise responsible for New Zealand's electricitytransmission system.

Ms. Brook Cowen holds a masters degree in economics (first class honours) fromthe University of Auckland, New Zealand, and a doctorate in economics,specializing in public policy and industrial organization, from the University ofOxford. She is an Honorary Associate of the Centre for Energy, Petroleum andMineral Law and Policy at the University of Dundee.

TONYBALLANCE

Tony Ballance joined Ofwat in May 1993. He had previously worked as a senioreconomist at the International Wool Secretariat, the organization responsible forpromoting the wool mark.

Mr. Ballance is an economist by background and was educated at the University ofWales, Aberystwyth, the University of Kansas and the University of Manchester(where he completed his Masters and Ph.D. in Agricultural Economics). Hebecame Ofwat's Chief Economist in October 1996 and is responsible for the areasof competition policy, charging policy, and the supply/demand balance. He is amember of the Society of Business Economists.

WARRICK SMITH

Mr. Smith is Manager of the Private Participation in Infrastructure (PPI) Group inPSD, and Co-chair of the PPI Thematic Group. Since joining the Bank in 1993, hehas advised on a wide range of PPI issues in all infrastructure sectors and in allregions of the world. His work typically focuses on market reforms, the design ofregulatory and institutional frameworks, and strategies for dealing with political

I and regulatory risks. He has written widely on these topics, with a particularemphasis on approaches in countries with weak institutional capacity.

Before joining the Bank, Mr. Smith held several senior posts with the AustralianGovernment. He was Secretary of the inquiry established by the Prime Minister todevelop a policy, regulatory and institutional framework for introducingcompetition into Australia's infrastructure sectors; Principal Counsel responsiblefor international business law in the Attorney-General's Department; and Directorresponsible for international markets in the Department of Energy. He also spentseveral years working on international trade policy matters in the Department ofTrade, as well as a period in private legal practice in the US and Australia. Mr.Smith is trained in economics, law and public administration, with degrees from theAustralian National University and Harvard University.

VI

IIIIIIII Creating Performance Incentives

I Through

Regulation & Benchmarking -- The UK Water Model

I

• Dr. A. J. Ballance| Ofwat

Creating Performance IncentivesThrough

Regulation & Benchmarking-The UK Water Model

Dr A J BallanceOf wat, UK

Friday 9 April 1999

IIIIIIIIIIIIIIIIIII

AgendaWater regulatory framework (England & Wales)

Price cap regulation- a bit of theory

- incentives

Comparative competitionTechniques for assessing efficiency- cost drivers

- techniques

- trade-offs

Improving service performanceThe Ofwat package

1. The water regulatory framework(for England and Wales)

27 water (and sewerage) companies in England andWalesPrivatisation in 1989Licences of appointment (for 25 years)Geographic monopoly within specified areasRole of the Director GeneralCustomer Service CommitteesQuality regulators (EC, Government, DWI andEnvironment Agency)

Water companies in England & Wales

Role of the Director GeneralStatutory duties

To ensure that companies carry outtheir functionsTo ensure that companies can financetheir functionsTo protect customersTo promote economy and efficiencyTo facilitate competition

IIIIIIIIIII

IIIIIIIIIIIIIIIIIII

The 1999 Periodic Review

The K factor (5 year) price capComparative competitionCost of qualityAppeals to the Monopolies and MergersCommission (MMC)The business planning processMethodology

2. Price cap regulationA bit of theory ....

RPI-X is medium-term, incentive-based, price-cap regulationHigher profits are the engine of lower pricesUsing comparative competition to simulate acompetitive marketStrong incentives for improving cost efficiencyand service performanceIncentives are a mixture of rewards andpenalties

Medium term etc

STABILITY - Price limits set for fiveyearsOUTPUTS - Companies must know whathas to be delivered and by whenRISKS - Lion's share with company (butsome carried by the customer as this ischeaper)OUT-PERFORMANCE - Goes tocustomers in the longer term 8

Strong incentives ...

Beat the Regulator's assumptions oncosts and you keep the difference (atleast until the next review)Lower cost base feeds through to lowerprices periodic review to periodic reviewBrownie points for good service!BUT - PR timetables promote gameplayingCost savings must be possible

IIIIIIIIIIIIIIIIiII

IIIIIIIIIIIIIIIIIII

Rewards and penalties...

Must deliver service & outputs or skyfalls in (....Yorkshire Water in 1995!)

Try to get incentive balance right -- high performance = high returns- good performance earns cost of capital- low performance = trouble on all fronts

fYS

Track record so far....

1989 settlement too easy - a one waybet for owners and companies!

1994 determination- corrected imbalances- turned out rather softer than we thought- but did deliver on cost reductions &

improved service!

1999?

Incentives are about...

Creating the right climate for futureefficiency improvements in allcompanies

Removing barriers and distortions thatmight affect cost-effective decisions- capex / opex trade-offs- reduce potential for regulatory gaming- minimising the PR timetable hiatusClear & consistent regulatory rules 12

Incentives not helped by...

Feather-bedding

Leaving barriers and distortions thataffect proper decisionsChanging the goalpostsInconsistency policies review to reviewToo much uncertainty about how the nextreview will be carried out

13

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

en

So it is all aboutcarrots and sticks

Carrots & sticks?

Possible butlikely to be too tight

•I -.„ •;>"> • -

Back in the _^USA & rate of ••• * ;

return

Possibleand mightbe best for

ali!

Possible butlikely to be too soft

Bigger carrots

15

3. Comparative competition

Where market competition is limited(eg the water industry in the UK)Comparisons cover costs (allowing fordifferences in operating environments)and performance (particularly service tocustomers)Publication of league tables of costs andperformance

16

Comparative competitioncontinued

Role of the CitySignificant number of independentcomparatorsRole of mergersCost/output matrix

17

10

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

4. Techniques for assessingefficiency - cost drivers

Identification of cost drivers:- base service level- quality- balancing supply & demand- enhancing service levelsCost elements:- o p e x- capital maintenance (depreciation)- return on capital- t a x 18

The framework for settingprice limits

K - > P 0 - X +

Where,Po - passed out-performanceX - future efficiency gainsQ - quality standardsV - enhancements to the security of

supply (balancing supply & demand)S - enhanced service levels

19

11

The price cap

Incentive regulation and efficiencyMeasuring past efficiency gains - Po (ora "glide path")Estimating future efficiency - XProfit sharing and benefit sharing- formal- voluntary

The period between price reviews20

Techniques for assessingefficiency

Econometric analysis applied to:-operating costs-capital maintenance expenditure

Unit cost analysis-cost base analysis-applied to capital maintenance

expenditure and capex enhancement21

12

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

Techniques for assessingefficiency continued

Benchmarking- metric- process

Total factor productivity

Capping of allowable costs:- use of the regulatory capital value

Economic appraisal techniques- least cost approach to balancing supply

and demand 22

The use of econometrics

Explaining differences in costs

Differences in operating environments- climate- geology- topography

Positive correlation between efficiencyrankings and subsequent efficiencysavings

23

13

Procedure fordeveloping models

Expert review of potential cost driversData collection and validationIdentification of atypical expenditure and exceptionalitemsRevised data for statistical analysis

Generate plausible conceptual modelsStatistical analysis to develop robust relationshipsExpert (external) reviewPreliminary assessment of relative rankingsReview of further company specific argumentsFinal models 24

Prospects for PricesOperating costs

£ billions4

3 -.

2 ..

1 ..

\AdmltAlopmtlng

dK

Financial year to 31 March

25

14

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

The use of the cost base

Identification of standardised costsData collectionCheck for consistency between standardcosts and company estimates of realcosts

Use standard costs to adjust real costsFrontier or lower quartile?

26

The frontier -v- catch-up

Distinctions between movements in thefrontier and the speed of catch-up areimportant ones to recogniseMovements in average efficiency are acombination of the twoIssues of fairness and incentivesSpeed and size of catch-up areimportant

27

15

Relative efficiency -v-yardstick competition

Ofwat has applied relative efficiency modelsOffer has primarily adopted a yardstick modelapproachCould yield similar short-term benefits tocustomersFinancing of companies' functions?Excess returns?Incentives?

Stimulating the market? 28

Trade-offs

Capex-opex trade-offCapital maintenance/qualityLower-higher serviceHeadroom (judgement about overallprice cap)

29

16

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

Comparators

Number of comparatorsUse of international comparisonsUse of other industries

30

5. Improving serviceperformance

Annual monitoringPublication of league tablesSetting service standardsGuaranteed standardsCustomer chartersCompensation arrangementsMarket competition?

31

17

6. The Ofwat 1999Periodic Review Package

Po adjustment

Returns at (market) cost of capital

Regulatory capital value - actualexpenditureRoll-out of cost savings (8c incentives)- capex-opex

32

The Ofwat 1999 PeriodicReview Package continued

Movement to efficiency frontierCost of qualityThe profile of prices ( V or T1)- customer viewsBalancing supply and demandService performance adjustment

33

18

IIIII1IIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

The 1999 solutionIs a combination "carrots" and "sticks":- right approach to past out-performance- challenging assessment of scope for

improvements reflecting relative efficiency-take account of service performance- adequate certainty about PR2004- clear minimum outputs & no deterioration in

service

34

And our sensible package?

Po evolving into a nlow altitude'' glidepath for both capex and opex out-performanceTough and challenging XAX" factorsClarity of "Q" [& "S"] outputs butsubject to tough cost reduction targetsPaying for growth rules OK re "V" factorLower cost of capital but will bebankable! 3f

19

Website: http://www.open.gov.uk/ofwattballance@ofwat.gtnet.gov.uk

20

IIIIIIIIIIIIIIIIIII

1 M ^ ^ p U B L I C P O L I C Y F O R T H E

rnvatesectorThe World Bank

Note No. 77

Michael Kleinand TimothyIrwin

May 1996

Regulating Water Companies

The water industry differs in two key respectsfrom other network industries, such as gas, elec-tricity, and telecommunications. First, there arefewer opportunities for introducing competi-tion among suppliers, since the network ofpipes is a major element of the total cost ofwater and can be operated efficiently only as amonopoly. Second, the quality of water is cru-cial, but hard for consumers to check. Together,these problems mean that getting the best per-formance out of water companies requires regu-lation by the government of the price andquality of water.

To regulate well, however, the regulator needsto have an idea of how much it would cost anefficient company to supply high-quality water.One way of generating that information is toauction the right to supply water every twentyyears or so. Firms state the price at which theywould be willing to supply water of a specifiedquality, and the firm offering the lowest pricewins the contract. In between auctions, how-ever, regulators need to use other methods toadjust the price in response to changing circum-stances. No method is perfect; the best may beto increase the price every year by the rate ofinflation, perhaps with an adjustment for ex-pected productivity changes, and review theprice every three to five years to ensure that thewater company's profits are reasonable. The im-portance of investments to maintain the qualityof water means that regulators should be care-ful, when reviewing prices, to allow the firm tocover the costs of such investments.

Why regulate water companies?

When water is sold by street vendors, consum-ers have a choice of suppliers. As a result, wa-

Private Sector Development Department • Vice Presidency for Finance and Private Sector Development

21

ter sellers have an incentive to sell water at aprice not much higher than its cost and to takesteps to show that the water is safe to drink.But the arrival of piped water changes every-thing. It is much, much cheaper than watersold by vendors, as table 1 suggests. In theAsian cities in the table, these lower prices aredue in part to government subsidies. But evenwhen the subsidies are taken into account,piped water is still at least 50 percent, and usu-ally 75 percent, cheaper. At the same time,however, consumers lose the choice of suppli-ers that they used to have.

In the nineteenth century, water companies laidcompeting pipelines in towns in Canada, theUnited Kingdom, and elsewhere. But it is usu-ally efficient to have just one network of pipes,and as a result of either free competition ormunicipal regulation, the competing networks

TABLE 1 PRICE OF VENDED AND PIPEDWATER(OS. cants percabic meter)

City Vended Piped

Bandung

Jakarta

Manila

Karachi

Ho Chi Minn

616 .

185

187

175

. 151

1017

11

8

8

vrUtiOm

Regulating Water CompaniesIIIIIIIIIIIIIIIIIII

of the nineteenth century soon turned into mo-nopolies. Technically, the water supply systemis a natural monopoly: the cheapest way tosupply water involves just one firm owning anetwork of pipes. Water monopolies, of course,can and do exploit their privileged position. Inthe worst case, they may even be able to chargeas much for water as the street vendors, inwhich case all the benefits of piped water ac-crue to the monopoly.

FIGURE 1 MAIN COMPONENTS OF WATER AND SEWAGE SYSTEMS

Servicestorage

rBSBiVDirs

Distributionto castraientand waste-

collection

Source.-North Wait Waur, Iba A

In some industries in which networks are impor-tant—gas, electricity, and telecommunications—governments have limited the scope of the naturalmonopoly problem by separating productionfrom transmission through the network. Thus,competing electricity generators, for example, cansend power to consumers using one network.Theoretically, this is possible in water too.

Competing water "production" firms that ownthe bulk storage reservoirs and water treatment

22

plants (figure 1) could sell water to a companythat distributed it to consumers through onenetwork of pipes. Although such a system hasrecently been proposed in Chile, no one hasyet succeeded in implementing this sort of com-petition. The reason is probably that network-related costs are a larger proportion of totalcosts in the water industry than in gas, elec-tricity, and telecommunications. The gains tobe made from introducing competition in, say,water collection and water treatment are thusrelatively small, and they have to be weighedagainst the coordination problems introducedby splitting up ownership of the system.

Competitive water supply may be efficient nearthe boundary of two water companies' territo-ries or in regions where water is very scarceand therefore the cost of the network is lowerrelative to the cost of the water. Competition isalso possible for services peripheral to the mainservice, such as connecting new users to thesystem. But for the time being, most water willbe supplied monopolistically, and society needssome way to encourage efficiency despite themonopoly.

The difficulty of regulating well

In villages, consumers can form cooperatives torun the water system themselves; since the pro-ducers are also the consumers in such a system,they have good reason not to charge too muchfor water or to be careless about its quality. Butin larger regions, consumers need to delegatethe problem of setting prices and quality stan-dards to someone else. The traditional option isto delegate it to the government. Governmentownership doesn't automatically solve the prob-lem, however. Monopoly suppliers of all typesare tempted to charge high prices or to lowerquality. And government ownership introducesits own problems, since the government, as anowner, usually exerts relatively weak pressureon firms to lower their costs.

Whether the water firm is publicly or privatelyowned, the key to achieving efficiency lies inthe choice of a regulatory mechanism to over-

IIIIIIIIIIIIIIIIIII

see the firm's performance. Good mechanismsprotect consumers from high prices and lowquality. But they also safeguard the legitímateinterests of the water companies, since, if thecompanies are to invest, they need to believethat the regulators will let them earn enoughrevenue to make a reasonable profit.

If the regulator had enough information—in par-ticular, if it knew what it would cost an efficientwater company to produce water of differentqualities—it could simply rule that the actualwater company had to sell water of a certainquality for a price equal to the efficient firm'scost of production. That price would be just highenough to allow an efficient water company tomake a reasonable profit, but no higher. Nei-ther the company nor the consumer would beexploited. And, as technology and demandchanged, in this perfect system, the regulatorwould revise the price and the quality standardso that they were always at the right levels.

In fact, of course, the regulator cannot easilytell how much it would cost an efficient firmto produce water. At best, it can observe ac-

. tual firms' costs, but these can be concealedby clever accountants. Moreover, an importantpart of a water firm's cost is the cost of thefinancial capital tied up in the firm. Estimatingthe cost of that capital requires an estimate ofthe riskiness of the investment, complicatingthe regulator's information problem yet further.With imprecise cost estimates, there's always arisk that the regulator will set the price toohigh, hurting consumers and unnecessarily dis-couraging water use, or too low, encouragingthe wasteful use of water and discouraging in-vestment by water companies.

In addition, because the regulator probablyguesses what it would cost an efficient firm toproduce water partly by observing the actualwater company's costs, the water company nolonger has such a strong incentive to produceefficiently. Since lower costs would lead theregulator to lower the price the company cancharge, the company would not get all the ben-efits of cutting costs.

A big part of the regulatory problem, there-fore, is to design rules that give the regulatoraccess to better information about the appro-priate price of water.

How to generate good information

Probably the best way of discovering theappropriate price is to establish a competitivesystem of tendering—or "auctioning"—the rightto supply water. The regulator says, for example,

A small town in France managed tocut the price of water from 3-0francsper cubic meter to 1.7francs when itdecided in 1994 to auction the rightto supply water

that it wants a firm to provide water of a speci-fied quality. It then asks firms to propose a pricefor supplying the water. The firm that proposesthe lowest price wins the right to supply thedistrict at that price Cor perhaps at the price ofthe next-lowest bidder—the details of the auc-tion can vary). In principle, the most efficientsupplier of water will win the auction, and theresulting price will be appropriate.

Experience confirms the value of auctions. InBuenos Aires in 1993, for example, the win-ning bidder offered to deliver water at a priceabout 27 percent lower than the price understate ownership. Although the price later in-creased, it remained lower than it had been.What's more, the new supplier agreed to in-vest US$200 million a year for the first five years,compared with annual investment of US$20 mil-lion to US$40 million in the preceding years.In another example, a small town in Francemanaged to cut the price of water from 3.0francs per cubic meter to 1.7 francs when itdecided in 1994 to auction the right to supplywater.

23

Regulating Water Companies III

The Nots series is anopen forum intended toencourage dissemina-tion of and debite onideas, innovations, andbest practices forexpanding the privatesector. The viewspublished are those ofthe authors and shouldnot be attributed to theWorld Bank or any of itsaffiliated organizations.Nor do any of the con-clusions representofficial policy of theWorld Bank or of itsExecutive Directorsor the countries theyrepresent.

Comments are welcome.Please call the FPDNote line to leave amessage (202-458-1111)or contact SuzanneSmith, editor. RoomG8105, The World Bank,1818 H Street NW.Washington, D.C. 20433,or Internet addressssmith79worldbank.org.

»Printed on recycled

paper.

Yet auctions are no panacea. To keep up withchanges in technology and demand would re-quire repeating the auctions every couple ofyears—which is what happens, for example,with garbage collection in many cities. Watercompanies, however, must make investmentswith a life of decades that have little value inother uses. Pipes, once laid, will last for years,and digging them up later to move them to anew site is prohibitively costly. A water com-pany that could easily lose its contract in anauction next year would therefore be justifi-ably cautious about long-term investments.

The problem is partially addressed by requir-ing a new winning firm to pay the old firm forthe pipes and other immovable assets. Butworking out the price the new firm should payis difficult. For one thing, the pipes are under-ground and their condition is hard to assess.To encourage valuable investments, then, auc-tions must be repeated only infrequently (ev-ery twenty years perhaps), or the incumbentmust be given an advantage over other bid-ders. But either way some of the benefits ofthe auction are lost. First, an incumbent with aprivileged position has weaker incentives tooffer the lowest possible price at the next auc-tion. Second, technology and demand—andtherefore the appropriate water price—changeduring the term of a twenty-year contract. Be-tween auctions, the regulator must again try toestimate how the right price has changed.

How to adjust pricesbetween auctions

How should regulators adjust prices betweenauctions? Over three- to five-year periods, thebest option is probably to adjust them in amechanical way. Traditionally, regulators in theUnited States have adjusted prices so as to keepthe company's rate of return on capital at aconstant level: if the company's rate of returnfalls below that level, the regulator allows pricesto rise. The problem with this method is that itgives the company little incentive to limit itscosts and, when the target rate of return ishigher than the cost of capital, it gives the com-

pany a strong incentive to invest more—in any-thing at all.

More recently, therefore, the United Kingdomhas chosen to change the price by means of aformula, known as RPI-X that increases thewater price by the increase in the retail priceindex adjusted by a factor, X, to account forexpected productivity gains and other changes.Under this method, the company has incen-tives to lower costs, since it keeps the result-ing profits. The method can also be refined bychoosing a price index that relates more spe-cifically to the input price inflation experiencedby water companies. Care needs to be taken,however, to avoid re-creating the problem ofcompensating the company for cost increasesit could have avoided.

RPI-X price adjustments are probably better thanrate-of-return price adjustments, but the differ-ence between them is not as big as it mightseem. RPI-X formulas need to be reviewed ev-ery three to five years or so, since the regulatordoes not know exactly how large X should beand, in reviewing whether X was set appropri-ately, will take into account the profits beingmade by the firm: for example, if they are verylarge, X is probably too small. In addition, theimportance of quality means that regulatorsshould allow firms to pass on the costs of rea-sonable investments that maintain water quality.

The undesirable incentive effects of both RPI-Xand rate-of-return adjustments can be reducedby comparing the prices charged by other wa-ter companies in different, but sufficiendy simi-lar locations, as happens in the United Kingdom.If comparable companies can profitably sell wa-ter at lower prices than the company under ex-amination, the regulator may be justified inkeeping prices low despite low profits.

Michael Klein, Manager, Private SectorDevelopment Department (email: mklein®woridbank.org), and Timothy Inuin, PrivateSector Development Department (email:tirwin@tvorldbank.org)

24

IIIIIIIIIIIIIIIIIII

M ^ P U B L I C P O L I C Y F O R T H E

PnvatesectorNote No. 127 October 1997

Warrick Smith

Utility Regulators—The IndependenceDebateThe global trend of utility privatization haspushed regulatory issues to the fore, amongthem the role of regulatory agencies. Theseagencies have a long history in the UnitedStates, and creating or strengthening them hasbecome a central goal of reforms around theworld. But many issues remain contentious,particularly the notion of agency indepen-dence. Some governments are reluctant tosurrender political control over regulatorydecisions. And even those who agree on thedesirability of independent agencies may ques-tion whether they are feasible or appropriatein all country settings. This Note considers thedebate over the independence of utility regu-lators, focusing on the position of developingcountries.

Independence—What and why?

Independence is subject to different interpre-tations. Some use it interchangeably with au-tonomy; others perceive greater or lesserdifferences in meaning between the terms. ThisNote defines independence for utility regula-tors as consisting of three elements:• An arm's-length relationship with regulated

firms, consumers, and other private interests.• An arm's-length relationship with political

authorities.• The attributes of organizational autonomy—

such as earmarked funding and exemptionfrom restrictive civil service salary rules—necessary to foster the requisite expertise andto underpin those arm's-length relationships.

The rationale for giving regulators independenceas broadly defined here lies in the special chal-lenges posed by utility regulation, including thecritical role of regulatory discretion.

Regulatory challenges

Utility regulation has three main aims: to protectconsumers from abuse by firms with substan-tial market power, to support investment byprotecting investors from arbitrary action bygovernment, and to promote economic effi-ciency. While there is growing recognition thatcompetition can reduce the need for regula-tion in utility industries, most industries containsome areas of monopoly where the benefits ofregulation potentially outweigh the costs.

Regulating utilities is complicated by three re-lated considerations. First, prices for utility ser-vices are usually political. There are no votesin raising utility prices, and history is repletewith examples of justifiable price increasesbeing withheld at the expense of investors andthe long-term interests of consumers.

Second, investors are aware of these pressuresand of the vulnerability of their usually large,long-term, and immobile investments. Unlessa government has made a credible commitmentto rules that ensure an opportunity to earn rea-sonable returns, private investment will notflow. Weak credibility will be reflected in highercapital costs and thus higher tariffs. In privat-ization, this translates into smaller proceedsfrom sales of existing enterprises and higherfinancing costs for new projects.

Third, the long-term nature of most infrastruc-ture investments makes creating credible com-mitments difficult. Highly specific rules, ifconsidered sustainable, can provide assuranceto investors and lower the cost of capital. Butthey make it difficult to adjust regulation tounforeseen developments, including changes

The World Bank Group • Finance, Private Sector, and Infrastructure Network

25

Utility Regulators—The Independence Debate

in technology and market conditions. They alsomake it difficult to tailor responses to situa-tions and to provide incentives for efficiency.There is thus an important tradeoff betweenreducing the risk of expropriation, and with itthe cost of capital, and retaining the flexibilityto pursue efficiency and other objectives.

In designing regulatory systems, then, policy-makers need to resolve two fundamental chal-lenges: How much discretion should regulatorysystems contain? And how should that discre-tion be managed to reduce the risk of misuseand thus the cost of capital?

How much discretion?

The discretion in regulatory systems differswidely among countries and industries. At oneextreme, U.S. laws typically delegate broad dis-cretion to regulators, often vaguely definingpricing standards as "just and reasonable" andlimiting other powers only by reference tobroad public interest criteria. At the other endof the spectrum, some countries implementregulation through tightly specified laws or con-tracts that seek to eliminate discretion. Theyattempt to deal with all contingencies foreseenat the time an arrangement is finalized, usuallyrelying on detailed cost-based formulas for tariffadjustments. This approach—sometimes called"regulation by contract"—is often favored byinvestors who perceive a high risk of misuseof discretion by the government or regulator.Adjustments to the initial arrangement will re-quire renegotiation, which can be difficult ifthe bargaining power of the parties changesonce the investment is made.

Most regulatory systems lie somewhere be-tween these extremes. Key policies and prin-ciples tend to be defined in laws, licenses, orcontracts, which carefully delimit residual dis-cretion through reference to criteria, factors,and objectives. Greater flexibility and discre-tion are usually more important in industriesin which there is rapid technological change,in which the introduction of competition re-quires continuous adaptation of rules to chang-ing market conditions, and in which highpriority is placed on providing incentives for

efficient operation. Discretion is thus typicallymore important for telecommunications thanfor toll roads. Another consideration is acountry's stability and reputation for respect-ing private property rights: the higher a coun-try scores on these criteria, the more discretionit can retain without significantly increasing thecost of capital. This consideration is especiallyrelevant for reforming and developing coun-tries, many of which lack a long track recordof good performance in these areas.

How to manage discretion?

When discretion is retained on tariffs or otherissues of concern to investors, the challenge isto manage it in a way that minimizes the riskof misuse. The exercise of discretion needs tobe insulated from short-term political pressuresand other improper influences and to be basedon competent analysis.

Entrusting discretion to ministers will not meetthese tests, particularly when the state contin-ues to own utility enterprises. In this case, therewill be no arm's-length relationship betweenthe regulator and the firm, and there may beconcerns that, in exercising discretion, minis-ters will favor the state enterprise over rivalprivate firms. But even if the state has no own-ership role, ministers will still be subject toshort-term political pressures, and changes ingovernment can lead to abrupt changes in regu-latory policy. Restrictive civil service salary rulesin many countries also make it difficult forministries to attract and retain well-qualifiedprofessional staff. What is required is an agentat arm's length from political authorities, utili-ties, and consumers. Organizational autonomyhelps to foster the requisite expertise and pre-serve those arm's-length relationships.

The quest for independence

Creating an independent agency, no easy taskin any setting, is even more challenging incountries with a limited tradition of indepen-dent public institutions and limited regulatoryexperience and capacity. The two main ele-ments of independence—insulation from im-proper influences and measures to foster the

26

development and application of technicalexpertise—are mutually supporting: technicalexpertise can be a source of resistance to im-proper influences, and organizational autonomyhelps in fostering (and applying) technicalexpertise.

There is strong consensus on the formal safe-guards required:• Providing the regulator with a distinct legal

mandate, free of ministerial control.• Prescribing professional criteria for appoint-

ment.• Involving both the executive and the legisla-

tive branches in the appointment process.• Appointing regulators for fixed terms and

protecting them from arbitrary removal.• Staggering terms so that they do not coin-

cide with the election cycle, and, for a boardor commission, staggering the terms of themembers.

• Exempting the agency from civil service sal-ary rules that make it difficult to attract andretain well-qualified staff.

• Providing the agency with a reliable sourceof funding, usually earmarked levies on regu-lated firms or consumers.

Formal safeguards of this kind are especiallyimportant in countries with a limited traditionof independent public institutions. But they arenot enough. Persons appointed to these posi-tions must have personal qualities to resistimproper pressures and inducements. And theymust exercise their authority with skill to winthe respect of key stakeholders, enhance thelegitimacy of their role and decisions, and builda constituency for their independence.

Some argue that governance traditions in somecountries make independence illusory—"If thePalace calls, the regulator will comply." Certainly,adopting even the most sophisticated law willnot magically transform the basic institutionalenvironment. Nevertheless, for several reasons,creating such agencies is worth the effort, evenin more challenging environments.

First and foremost, independence must be un-derstood as a relative rather than an absoluteconcept. In any system, the goal can only be

to reduce the risk of improper political inter-ference, not to provide ironclad guarantees.Progress must be measured at the margin—and relative to the outcome of ministers re-taining direct control over regulatory decision-making. Second, the ability of independentagencies to sidestep civil service salary restric-tions and to have access to earmarked fundingmakes it possible to recruit and retain better-qualified staff and to hire external consultants.This can improve the technical quality of deci-sions and thus enhance the agency's authority.Adequate salaries can also help to reduce con-cerns about corruption. Finally, even if thereare reasons to doubt that an agency will exer-cise truly independent judgment in the shortterm, that may change in the longer term. Con-centrating expertise in a body with a specialistmandate sharpens commitment to professionalnorms, which can be an important source ofresistance to improper influences. And as theregulator enters the fray, it will have the op-portunity to build a constituency of its own,increasing insulation from political interference.

Reconciling independence withaccountability

Independence needs to be reconciled withmeasures to ensure that the regulator is ac-countable for its actions. Checks and balancesare required to ensure that the regulator doesnot stray from its mandate, engage in corruptpractices, or become grossly inefficient. Strik-ing the proper balance between independenceand accountability is notoriously difficult, butthe following measures to do so have beenadopted by a growing number of countries:• Mandating rigorous transparency, including

open decisionmaking and publication of deci-sions and the reasons for those decisions.

» Prohibiting conflicts of interest.• Providing effective arrangements for appeal-

ing the agency's decisions.• Providing for scrutiny of the agency's bud-

get, usually by the legislature.• Subjecting the regulator's conduct and effi-

ciency to scrutiny by external auditors orother public watchdogs.

• Permitting the regulator's removal from officein cases of proven misconduct or incapacity.

27

Utility Regulator»—The independence Debate

Viewpoint is an openforum intended toencourage dissemina-tion of and debate onideas, innovations, andbest practices for ex-panding the privatesector. The views pub-lished are those of theauthors and should notbe attributed to theWorld Bank or any ofits affiliated organiza-tions. Nor do any of theconclusions representofficial policy of theWorld Bank or of itsExecutiva Directorsor the countries theyrepresent.

To order additionalcopies please call202-458-1111 or contactSuzanne Smith, editor.Room F6P-188,The World Bank.1818 H Street, NW,Washington, D.C. 20433,or Internet addressssmith7ttwDrldbank.org.The series is alsoavailable on-line(www.worldbank.org/html/fpd/notes/notelist, html).

® Printed on recycledpaper.

Possible paths of transition

Resistance to independent agencies is break-ing down. Ministers once adamant about main-taining political control over tariffs and otherregulatory matters increasingly see the bene-fits of creating such agencies, which includeimproving offers from investors, helping to sus-tain reforms, and shifting responsibility for un-popular decisions to someone else. But whatif the government resists?

The choice can be stark. Governments can re-duce discretion by adopting highly specificrules, forfeiting flexibility and other advantages.Or they can retain discretion, pay investors riskpremiums, and accept reduced proceeds fromprivatization, higher tariffs or both. In eithercase, ministerial structures will usually make itdifficult to develop expertise to deal with regu-latory problems arising after privatization.

But several options lie between the traditionalministerial model and the delegation of broaddiscretionary authority to a fully independentagency. These options can form a path of tran-sition to greater independence and delegationof discretionary authority. First, a dedicatedregulatory unit can be created within a minis-try, to coordinate regulatory activity and fosterthe development of technical skills and pro-fessional norms. The autonomy of the unit canoften be enhanced by placing it under the re-sponsibility of a minister other than a sectoralminister—particularly important if there is po-tential for conflict between private firms andstate enterprises under the purview of thesectoral minister. Once such a unit has beencreated, governments can increase the trans-parency of regulatory processes and approxi-mate an independent agency in other ways.Exempting staff from civil service salary ruleswill usually be more problematic, but concernsabout technical competence can be addressedby contracting out certain tasks to consultants.

Second, an agency can be created with manyof the attributes of an independent agency, butwith one or more ministers taking part in

decisionmaking (as in Colombia). This ap-proach can improve the technical quality ofregulatory decisionmaking, particularly com-pared with the first option. But as long as min-isters retain significant influence, the risk ofmisuse of regulatory discretion remains.

Third, a more truly independent agency canbe created, but with some or all of its powerslimited to making recommendations to theminister (as in Hungary). A variant is to givethe agency decisionmaking authority but haveappeals go to the minister rather than anotherindependent authority (as in Argentina). Thisapproach reinforces the separation of profes-sional and political considerations in decision-making and usually provides the agency withgreater insulation than under the second op-tion. Political considerations are not excludedfrom the regulatory process, but the agencycan build a reputation for professionalism andbalanced judgment, enhancing its authority andreducing the likelihood of being overruled.Models can also be devised in which the min-ister is permitted to depart from the agency'srecommendations or decisions only in narrowlydefined circumstances.

Even where the minister has withdrawn com-pletely from regulatory decisions, a transitionstrategy may still be appropriate. Delegatingbroad discretionary powers to an untestedagency poses risks, particularly in countrieswith limited regulatory experience and capac-ity. The broader the agency's authority, themore enticing a target it will be for those withincentives to undermine its independence. Andlack of detailed standards—like those that havetaken more than a century to develop in theUnited States—can create uncertainty and riskfor investors. The prudent course is to take thetime to carefully define a new agency and en-sure that it has access to adequate resourcesand other support. These issues are examinedin two companion Notes.

Warrick Smith Cwsmitb3@tvoridbank.org),Private Participation in Infrastructure Group

28

• M • ^ P U B L I C P O L I C Y F O R T H E

PnvatesectorNote No. 128

Warrick Smith

October 1997

Utility Regulators—Rôles andResponsibilitiesCreating independent regulatory agencies hasbecome a key element of utility sector reformsaround the world. As discussed in a companionNote, these agencies are intended to insulatedecisionmaking from improper pressures andfoster technical expertise. This Note focuses onthe defining of responsibilities of such agen-cies, particularly in developing countries. It con-siders the scope of agencies' industry coverage,their role relative to ministers, and their rolerelative to other regulatory objectives and bodies.

Industry coverage

Specialist utility regulators can be organizedon three main bases:• Industry-specific, in which there is a sepa-

rate agency for each industry—such as gas,power, water, and telecommunications—asin the United Kingdom.

• Sectorwide, in which there is an agency foreach more broadly defined sector, such asthe energy regulator in Colombia and thetransport regulator in Canada.

• Multisector, in which there is a single agencyfor all or most utility industries, such as thestate-level regulators in Brazil and the UnitedStates, and the national regulators in CostaRica and Jamaica.

Advantages of multi-industry agencies

Making an agency responsible for more thanone industry offers several potential advantages.

Sharing resources. Economists, financial ana-lysts, and other professionals can work acrossindustries, and administrative staff and facilitiescan be shared. This is particularly important incountries where regulatory expertise is scarce.

Facilitating learning across industries. All util-ity industries have unique features, but the mainissues in their economic regulation are sub-stantially the same: administering tariff adjust-ment rules, managing the introduction ofcompetition into traditionally monopolistic in-dustries, and managing relationships with stake-holders. Having a single agency aids the transferof insights and experience between industries.

Reducing the risk of industry capture. A keychallenge in utility regulation is to guard againstthe agency's capture by the regulated indus-try. If the industry and the regulator developtoo close a relationship, the industry may beable to divert regulatory effort to promote itsown interests rather than the public's. Thebroader responsibilities of a multi-industryagency help to reduce this risk.

Reducing the risk of political capture. Agenciesintended to operate at arm's length from po-litical authorities remain vulnerable to inter-ference from them. Placing responsibility forseveral industries in one agency may make it amore attractive prize for political authorities.But there are two reasons why a multi-indus-try agency might be exposed to less risk ofpolitical capture rather than more. First, theagency's broader constituency raises the stakesof political interference: interfering in a deci-sion on, say, water tariffs will be seen as athreat to all industries regulated by the agency.Second, an agency responsible for more thanone industry can develop greater independencefrom sectoral ministries. Political pressures areunlikely to have effect unless they come fromhigher-level authorities, who can consider therepercussions of short-term actions from abroader perspective.

The World Bank Group • Finance, Private Sector, and Infrastructure Network

29

Utility Regulators—Roles and Responsibilities

FIGURE 1 MULTISECTORAL AGENCY-ILLUSTRATIVE ORGANIZATION

Commission

ectricity | &?,

Tarife i Tariffs! Compliance I ConpNance

(Standards I.Tariff* •: Complane» i Exttmal

I relation*

Reducing the risk of economic distortions. Allindustries compete for investment capital, andthere is direct competition between some util-ity industries in meeting consumer needs, suchas between gas and power or among differenttransport modes. Some regulatory issues areunique to specific industries and thus warrantdifferent approaches. But many issues, suchas the valuation of capital and the treatment ofinflation, are common to all industries. Incon-sistent approaches to these issues in compet-ing industries can create economic distortions.Having a single agency makes it easier to adoptconsistent approaches.

Dealing with blurred industry boundaries. Tra-ditional boundaries between utility industriesare rapidly blurring. Gas, power, water, andrailway firms are entering telecommunicationsmarkets. Gas utilities are entering the powerindustry, and water and power utilities aremerging. Such developments can pose impor-tant regulatory challenges. A firm involved inseveral industries may be able to exploit dif-ferences in the rules that apply to its activitiesin different industries. And regulatory decisionson one industry can affect other industries.Multi-industry agencies can deal with thesechallenges in a coordinated way.

Offsetting disadvantages?

Proponents of industry-specific agencies oftenargue that multi-industry agencies have weak-nesses or limitations that offset their advan-tages. One concern is that a multi-industryagency may lack sufficient industry-specificexpertise or focus. This concern can be ad-

dressed in several ways. Industry-specific de-partments can be created within the agency,but with a cross-sectoral decisionmaking bodyand cross-sectoral departments for poolingexpertise and managing shared resources (fig-ure 1). The agency can also draw on advicefrom industry-specific advisory groups.

A second concern is that placing responsibilityfor several industries in one agency is tanta-mount to "putting all your eggs in one basket"—the agency's failure would have costs for allindustries. Although industry-specific agencieshelp to diversify this risk, they do so at theexpense of the strength of a single agency, in-creasing the risk of failure.

A third argument is that having a number ofagencies allows experimentation with differentapproaches. However, industry-specific experi-ments are still possible in multi-industry agencies.

Finally, it is sometimes suggested that multi-industry agencies are appropriate only for verysmall economies. Certainly, the arguments forsuch agencies are especially strong in thesecases. Yet California's Public Utilities Commis-sion is responsible for gas, power, water, trans-port, and telephony in an economy with apopulation of more than 30 million, a GDP andutilities that dwarf those of most countries, andno evident shortage of trained professionals.

Creating multi-industry agencies

The preferred approach to creating a multi-in-dustry agency is usually to set it up as one fromthe outset, adding industries to its jurisdiction asthey undergo reform. If an industry-specificagency already exists, it may be possible to ex-pand its mandate to cover additional industries.

The alternative strategy—creating a series ofindustry-specific agencies and later mergingthem—has disadvantages. It delays such ben-efits of a multi-industry agency as fosteringlearning between industries, which are particu-larly important during an agency's early years.And the obstacles to later merger should not

30

be underestimated. Industry-specific regulatorswill have incentives to resist merger, not leastbecause of the implications for their jobs. Regu-lated firms may also resist, often out of con-cern that they will have less influence over amulti-industry agency. Mergers thus usuallyrequire substantial political will and effort.

The main challenge in creating multi-industryagencies is to ensure an effective coordinatingmechanism during their design and establish-ment. Because advisers with industry-specificresponsibilities have little incentive to proposemulti-industry approaches, leadership usuallymust come from a central ministry.

Role relative to ministers

One of the most sensitive relationships for aregulatory agency is that with the relevant min-isters. It is sometimes suggested that the min-istry is responsible for policy and the agencyfor regulation. But this distinction is unhelpfulin practice, because the dividing line betweenthe concepts is fuzzy at best, and agencies withsignificant discretion clearly have a policy role.

Four main considerations generally determinethe allocation of responsibilities between agen-cies and ministries. The first is whether the mat-ter in question is judged to be appropriate fordecision on political or technical criteria. Suchjudgments can change over time. For example,while political control over tariffs was once con-sidered the norm, there is now growing recog-nition that, once the key policy principles orrules are established, society's interests are bestserved by delegating responsibility to an inde-pendent agency. Tax and subsidy issues, bycontrast, are still widely regarded as the provinceof political rather than independent bodies.

The second consideration is whether colocationof particular functions could create significantconflicts of interest. For example, responsibil-ity for actively promoting investment in a sec-tor often conflicts with a regulatory agency'srole as an impartial arbiter of investor and con-sumer interests, as well as dilute its focus.

The third consideration is which body has theexpertise for a task and whether having relatedtasks performed by the same body yields anyeconomies. Once created, an agency usually be-comes the main repository of public sector ex-pertise on the industries it regulates. If theministry is subject to restrictive civil service sal-ary rules and the agency is not, the ministrymay find it difficult to maintain expertise. Thisoften justifies giving the agency an advisory roleon matters remaining under ministerial control.

The fourth consideration is the degree of con-fidence political authorities have in the agency.Agencies tend to be given greater authorityonce they have proved their reliability.

Based on these considerations, there is generalconsensus that ministers should retain respon-sibility for broad sector policy, including publicinvestment, privatization, sector restructuring,taxation, subsidies, intergovernmental relations,and the legislative framework. But even in theseareas, agencies may be given advisory roles.

There is less consensus on where responsibil-ity for granting licenses or concessions shouldlie. Much depends on the criteria governingthe award of licenses: the more objective andtechnical the criteria, the stronger the case fordelegating the responsibility to an agency.

Most systems give agencies responsibility foradministering tariff adjustment rules, elaborat-ing detailed standards, monitoring compliancewith norms, and facilitating the settlement ofdisputes. In some systems, the power to im-pose sanctions for noncompliance with normsis reserved for the courts. In most, however,the agency performs this role, although majorsanctions—such as cancellation of licenses—may require ministerial decision.

Role relative to other regulators

Utility regulators' main focus is economic regu-lation of firms with monopoly power. But utili-ties, like other firms, are subject to regulation tomeet a raft of other objectives, including safety,

31

Utility Regulators—Roles and Responsibilities

Viewpoint is an openforum intended toencourage dissemina-tion of and debate onideas, innovations, andbest practices for ex-panding the privatesector. The views pub-lished are those of theauthors and should notbe attributed to theWorld Bank or eny ofits affiliated organiza-tions. Nor do any of theconclusions representofficial policy of theWorld Bank or of itsExecutive Directorsor the countries theyrepresent.

To order additionalcopies please call202-458-1111 or contactSuzanne Smith, editor,Room F6P-188,The World Bank,1818 H Street NW,Washington, D.C. 20433,or Internet addressssmith7@worldbank.org.The series is alsoavailable on-line(www.worldbank.org/html/fpd/notes/notelist.html].

® Printed on recycled

paper.

antitrust, and environmental aims. How shoulda utility regulator's role be defined in relation tothese objectives and to other regulators?

A sound general rule is to avoid a proliferationof agencies. Creating numerous agencies candissipate expertise, forgo the economies inhaving one entity perform related tasks, createcoordination demands, and introduce addi-tional complexity. But as with most generalrules, there are exceptions. Separate agenciesmay be required to avoid significant conflictsin the mandate of a single agency. When anexisting agency responsible for, say, environ-mental regulation is performing well, immedi-ately transferring its responsibilities in utilitiesto a new utility regulator is usually inadvis-able. And there are inescapable tradeoffs be-tween cultivating expertise, economies of scale,and coordination in utility regulation and do-ing the same in environmental or other regula-tion for the economy as a whole.

There is one rule that should have no excep-tions: If more than one agency is involved inregulating utilities, the role of each should bedefined as clearly as possible to avoid duplica-tion, jurisdictional uncertainty, and turf disputes.

Service quality issues

Customer service standards are usually theprovince of the utility regulator. The allocationof responsibility for safety and environmentalregulation can vary widely, even between sec-tors in a single country. Two main issues war-rant consideration.

Standard setting. Quality standards have a di-rect impact on utilities' costs and thus on prices.If the utility regulator is not responsible for de-termining standards, it may have a role in pro-viding advice to the agency that is responsible.

Tariff adjustment. Because changes in qualitystandards affect costs, they may require tariffadjustments. When different agencies are re-sponsible for regulating tariff and quality pa-rameters, coordination issues can arise. These

issues can be addressed in several ways, in-cluding through tariff rules that permit certaincost increases to be passed on automatically.

Antitrust matters

Antitrust regulation includes prohibitions oncertain anticompetitive agreements and merg-ers and on the misuse of market power. Incountries with modern antitrust regimes, thesematters are usually entrusted to a specialistagency with economywide jurisdiction. Howshould a specialist utility regulator's role bedefined relative to the antitrust agency? Thereare two main issues.

Clarifying the interaction between regimes.There may be overlap between utility and an-titrust regulation in some areas—for example,between industry-specific regimes governingaccess to networks and economywide rulesgoverning the misuse of market power. Theinteraction between the two regimes shouldbe defined clearly from the outset.

Exploiting complementary expertise. Utilityregulators and antitrust agencies have comple-mentary expertise. Both agencies may be in-volved in reviewing proposed mergers orallegations of anticompetitive conduct in utilityindustries. In some cases, a member of the anti-trust agency is also made a member of the util-ity agency, or the agencies make formalsubmissions to proceedings conducted by theother. Antitrust agencies may also be given spe-cial roles in utility regulation, such as hearingappeals of decisions by the utility regulator.

Decisions on the responsibilities of a utilityregulator have important implications for otheraspects of the agency's design, including itsdecisionmaking structure, its resources, and thestart-up strategy. These and related issues areconsidered in a companion Note.

Warrick Smith (%vsmith3®worldbank.org),Private Participation in Infrastructure Group

IIIIIIIIIIIIIIIII

32

• M - ^ 0 U B L I C P O L I C Y F O R T H E

PrivatesectorNote No. 129 October 1997

Utility Regulators—DecisionmakingStructures, Resources, and Start-up Strategy

Warrick Smith

TABLE 1

Governments creating specialized regulatoryagencies must make decisions on a wide rangeof issues. Questions relating to the indepen-dence and responsibilities of such agencies areconsidered in two companion Notes.1 This Notefocuses on a third set of issues, relating todecisionmaking structures, resources, and start-up strategy. Like the other two Notes, it em-phasizes the situation of developing countries.

Decisionmaking structure

The design of an agency's decisionmakingstructure encompasses issues relating to thenumber of decisionmakers, the basis for select-ing them, the roles accorded to stakeholders,and the regulatory and appeals processes.

DECISIONMAKING STRUCTURES—INDIVIDUAL

VERSUS COMMISSION

Strengths and weaknesses

Characteristic.

/KM».-The j w

Number of decisionmakers

Many countries entrust decisionmaking author-ity to a commission or board of three to fivemembers; others prefer a single individual. Eachapproach has its strengths and weaknesses, andthe choice often depends on a country's tradi-tions and conditions (table 1). Agencies respon-sible for several industries usually choose acommission.

Selection of regulators

When agencies are to be independent, the goalshould be to select regulators with the personalqualities needed to exercise independent judg-ment and resist improper pressures or induce-ments. The selection is critical, particularly fornew agencies that have yet to establish a repu-tation for competence and reliability.

Qualifications and disqualifications for appoint-ment are usually set out in the law establish-ing the agency. Disqualifying factors generallyinclude having a financial interest in regulatedfirms, which creates a conflict of interest and,in some countries, being related to the presi-dent or ministers. A common qualification re-quired is significant experience or training ineconomics, finance, law, public administration,or industry.

It is sometimes suggested that some or all ap-pointees should have industry-specific techni-cal expertise or long experience in the regulatedindustry. But this requirement is unnecessaryand in some cases undesirable. It is unnecessarybecause such technical expertise will be avail-able from agency staff or consultants. It is un-desirable if it ends up excluding professionals

The World Bank Group • Finance, Private Sector, and Infrastructure Network

33

Individual

Speed of decisionniakina E

AccountabilityfordBCisians - - 0

Resource deraonfs ' 121

Predictability of decisions 0

Invulnerabilttyto individual preoccupations G

Invulnerability to improper influences C

Potential to reflect multiple perspectives G

Potential tostagger termsto enhance

stability and wealran links with particulargovernments G

• •DQ-G0m

» is aronyvr on each characteristic.

Utility Regulators—Decisionmaking Structures, Resources, and Start-up Strategy

with broader perspectives relevant to economicregulation or if it unduly restricts the pool ofcandidates. It is particularly inappropriate formulti-industry regulators, because requiring ex-pertise in each industry be represented on thecommission could crowd out appointees withbroader perspectives. It could also result in theexpert for each industry becoming the de factoregulator for that industry and thus the loss ofthe potential benefits of a commission approach.

Another view is that the decisionmaking bodyshould be composed of representatives ofconsumers and regulated firms rather than tech-nical experts. Although it is important for stake-holders to participate in the regulatory process,there are several reasons why including themon the decisionmaking body is inadvisable:• In most industries, attempting to identify

single representatives of consumers and theindustry is not feasible. Residential, indus-trial, and rural consumers all have differentand sometimes conflicting interests, and in-terests are likely to vary within these groupsacross regions or income classes. Regulatedfirms can also have different and sometimesconflicting interests in regulatory decisions.So, a representative approach can result inpressures to create very large decisionmakingbodies, which would increase delays andreduce individual accountability.

• Decisions of representative bodies hinge ontheir composition and voting rules. If thecomposition and voting rules favor one in-terest over another, decisions can be expectedto be biased accordingly. If the interests ofconsumers and utilities are equally balanced,and the casting vote is left to a representa-tive of the government, short-term politicalconsiderations can be expected to dominateregulatory decisionmaking.

• Representative bodies internalize bargainingand the exchange of concessions betweeninterests, at the expense of a more open andtransparent evaluation of competing socialinterests.

The executive branch usually plays the domi-nant role in the appointment process, but the

34

legislature often also has a role, such as in con-firming appointments. Involving both branchesof government is especially important in sys-tems in which the executive does not neces-sarily control the legislature; it provides a checkagainst partisan appointments and helps to le-gitimize regulators' authority.

Stakeholders' roles

To ensure that a regulatory agency makes de-cisions that are well informed and accepted asfair and legitimate, consumers, regulated firmsand other stakeholders must have the oppor-tunity to present their views. For the reasonsnoted above, their participation in the decision-making body is inadvisable. But there are sev-eral other options.

Open regulatory processes. Those with a sig-nificant interest in a regulatory decision are usu-ally permitted to present their views to theagency before the decision is made. In theUnited States, the process for doing so is usu-ally formal hearings, often criticized for beingtoo legalistic, costly, and slow. Regulators in theUnited Kingdom initially adopted much moreinformal processes, but die trend now is towardgreater formality. Countries such as Argentinaand Bolivia are experimenting with open pro-cesses that more closely reflect local traditions.

Consultative or advisory bodies. Some countrieshave created special consultative or advisorybodies, usually organized on an industry-specificbasis, to advise the regulator and other publicauthorities. These bodies are usually part-timeand composed of representatives of consum-ers, utilities, and industry experts. Special con-sumer councils can be especially important incountries that lack effective advocacy of con-sumer interests.

Regulatory process

Decisionmaking processes range from formalhearings, as in the United States, to more in-formal processes, such as those in the UnitedKingdom. Wherever the balance is struck, the

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

focus should be on transparency in decision-making, which reduces opportunities for im-proper influences and underscores the fairnessand legitimacy of decisions.

The regulatory process usually involves threemain steps: providing people with an interestin a decision opportunity to present their views,publishing the decision and the detailed rea-sons for reaching that decision, and providingstakeholders an opportunity to challenge thedecision through an appeals process.

The appeals process is important to ensure thatthe regulator does not stray from its mandateand that it remains accountable. Two closelyrelated issues need to be considered in design-ing an appeals process.

Appellate body. If the regulatory agency is tobe independent, the appellate body should alsobe independent. In most countries, appeals ofregulatory decisions go straight to the courts.But, in some countries, there is an intermediatestep in which appeals go to a body that is ex-pected to have more technical expertise than thecourts and that may also be able to respond morequickly. In the United Kingdom, for example,the antitrust agency hears appeals relating tolicense amendments. In Bolivia, a special super-intendency hears appeals from sector regulators.

Grounds of appeal. The grounds of appeal areusually limited to errors of fact or of law, in-cluding failure to follow a required process.Appellate bodies are generally not permittedto reconsider the merits of the decision andsubstitute their own judgment.

Resources

An agency's effectiveness is determined largelyby the adequacy of its resources, both humanand financial.

Human resources

Utility regulation requires personnel with a mixof skills in such fields as economics, finance,

law, and engineering, and the character andintegrity to resist improper pressures and in-ducements. People with these attributes arescarce in many reforming countries, and thosewho do have them will often receive attractivejob offers from privatized utilities. So, to at-tract and retain well-qualified staff often re-quires exempting agency staff from restrictivecivil service salary rules.

There is no magic formula for determining thenumber of staff required by an agency. It alldepends on the responsibilities of the agency,the climate in which it must discharge those re-sponsibilities, and its strategies for performingthose tasks. In the United States, staff size rangesfrom less than 40 in the public utilities commis-sions responsible for multiple industries in thesmaller states to more than 1,000 in the FederalEnergy Regulatory Commission. As a generalproposition, "small is beautiful." Overstaffing candilute an agency's professional focus and in-crease the direct costs of regulation. It can alsoincrease the indirect costs of regulation if staffmake unnecessary demands on utilities to jus-tify their jobs. For these reasons, a sound gen-eral principle is to keep the permanent agencystaff as small as possible, engaging consultantsto assist with specialized tasks.

Regulatory agencies increasingly contract outtasks to private firms or consultants, such asthe analytical work underpinning tariff adjust-ment and similar decisions and the complianceaudits of regulated firms. But the agency mustretain—and be seen to retain—responsibilityfor its decisions, to avoid undermining the legit-imacy of its actions. It must also ensure thatthe contractor is not subject to improper influ-ences or inducements from regulated firms orother sources.

Funding

Regulatory tasks, like other government func-tions, were traditionally funded from general taxrevenues. Now, most regulatory agencies ob-tain their income from levies on consumers.These levies may be charged to consumers di-

35

Utility Regulators—Decisionmaking Structures, Resources, and Start-up Strategy

Viewpoint is an openforum intended toencourage dissemina-tion of and debate onideas, innovations, andbest practices for ex-panding the privatesector. The views pub-lished are those of theauthors and should notbe attributed to theWorld Bank or any ofits affiliated organiza-tions. Nor do any of theconclusions representofficial policy of theWorld Bank or of itsExecutive Directorsor the countries theyrepresent.

To order additionalcopies please call202-458-1111 or contactSuzanne Smith, editor.Room F6P-188,The World Bank,1818 H Street, NW,Washington, O.C. 20433,or Internet addressssmith70worldb8nk.org.The series is alsoavailable on-line(www.worldbank.org/html/fpd/notes/notelist html).

® Printed on recycledpaper.

rectly, but are more often collected indirectlyby imposing a levy or license fee on regulatedfirms and allowing them to pass the cost on toconsumers through tariffs. In OECD countries,this approach is usually seen as part of a cost-recovery strategy: it reduces demands on gen-eral tax revenue and imposes the financial costsof regulation on the primary beneficiaries (con-sumers). In many developing and transitioneconomies, by contrast, earmarked funding isoften viewed primarily as a means of ensuringthat agencies have a reliable source of incomeand thus as a safeguard of agency independence.

To prevent levies from becoming too burden-some, the law establishing the agency usuallysets a cap on levies, often defined by referenceto industry turnover or some other indicator.The cap is 0.5 percent for telecommunicationsregulators in Argentina, Peru, and Venezuela;1.0 percent for the energy regulator in Colom-bia; and 2.0 percent for the water regulator inPeru. The cap establishes the maximum levy,and actual levies are set each year to cover abudget approved by the legislature. When anagency is responsible for more than oneindustry, a different levy is usually set for eachindustry that covers the costs of its own regu-lation and contributes to costs shared acrossindustries.

Start-up strategy

Utility regulators should be established as longbefore privatization as possible, even if theirformal powers do not come into effect imme-diately. This allows regulators time to familiar-ize themselves with their new responsibilities,to establish their offices, and to undertake anynecessary training. It also provides assuranceto consumers that their interests will be pro-tected after privatization and gives potential in-vestors an opportunity to assess the regulatorysystem before formulating proposals.

Most new regulatory agencies can expect achallenging infancy. Besides mastering com-plex technical issues, regulators must definenew and often difficult working relationships

36

with political authorities, regulated firms, con-sumers, and other stakeholders. In countriesin which the requisite skills are scarce, regula-tory experience is limited, and there is littletradition of independent public institutions, thechallenges can be daunting. And life is notmade easier for a regulator if privatization re-mains politically contentious and if the first pub-lic evidence of its effects is a price increaseallowed by the regulator.

To meet these challenges, regulators must haveadequate training—not only in such traditionaldisciplines as law, finance, and economics, butalso in negotiation analysis, media relations,and the like. Regulators may also need techni-cal support during the first months in office.Such support is often provided by consultantsacting to some degree as "shadow" regulators.

No less important, newly appointed regulatorsbenefit from contacts and exchanges with moreexperienced regulators from other countries.Some of these contacts occur on an ad hocbasis, through visits and participation in con-ferences. But there is also an encouraging trendtoward systematizing such contacts, for ex-ample, through a "twinning" arrangement be-tween a new regulator and a more experiencedforeign regulator. These arrangements can pro-vide a basis for exchanging staff and materialsor providing other forms of support and ad-vice. There has also been a recent trend to-ward creating "networks" of regulators, suchas the International Forum for Utility Regula-tion sponsored by the World Bank.

Warrick Smith, "Utility Regulators—The Independence Debate"

(Viewpoint » 127, October 1997) and "Utility Regulators—Roles

and Responsibilities" (Viewpoint * 128, October 1997).

Warrick Smith (wsmith3@wortdbank.org),Private Participation in Infrastructure Group

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

Country Dept. ILatin America and the Caribbean Region

Economic Notes

Regulating Brazil's InfrastructurePerspectives on Decentralization

Warrick SmithBen Shin

September 1995

The World Bank

This series presents the findings of work in progress. They are prepared or sponsored by LA1 staff toaddress immediate policy concerns identified through the dialogue with governments and they reflect theviews of the LA1 Department. This note was prepared by Warrick Smith and Ben Shin (PSD), under thegeneral direction of Antonio Estache (LA1IN). Questions and concerns should be addressed to Mr.Estache, telephone number (202) 458-1442, email address: aestache@worldbank.org.

37

REGULATING BRAZIL'S INFRASTRUCTURE:PERSPECTIVES ON DECENTRALIZATION

Warrick Smith and Ben Shin

Brazil is currently undertaking a fundamentalshift in the role of government in infrastructureprovision. Governments will retreat from therole of owner and operator of infrastructureservices and place emphasis on their role asregulator of services provided by private firms.This shift will require a strategy for definingand implementing regulatory frameworks forBrazil's infrastructure industries. As in thecase of other countries with federal systems, aspecial challenge in Brazil will be theestablishment of regulatory and relatedinstitutional arrangements that reflect theproper roles of national state and municipaltiers of government.

The purpose of this note is to providedecisionmakers in Brazil with a framework forthinking about the important connectionsbetween decentralization and regulation, witha specific emphasis on the implementation ofregulatory arrangements at sub-national tiersof government. Section 1 provides a briefoverview of the key tasks and challenges ofinfrastructure regulation. Section 2 reviewsthe main considerations affecting theassignment of regulatory responsibilitiesbetween tiers of government. Section 3considers the special challenges associatedwith designing and implementing regulatoryarrangements at sub-national tiers ofgovernment. Section 4 presents a briefconclusion.

1. Regulating Infrastructure: Key Tasksand Challenges

Regulation affects most aspects of life inmodem societies, and has many economic,social, and political objectives. This notefocuses on regulating infrastructure -predominantly power, gas, water,telecommunications, and transport - where

regulation seeks to avoid the potentialinefficiency and other ills arising from thenatural monopoly characteristics of segments ofthese industries.1 The central - and mostcontroversial - regulatory task is controllingprices or profits in monopolistic segments,whether these be final prices to consumers orprices and other conditions affectingcompetitors' access to networic facilities. Otherimportant regulatory tasks are defining andenforcing compliance with concessions orsimilar arrangements, including adherence toservice quality, investment levels, and technicalstandards. Environmental and antitrustregulations affecting these industries may alsorequire definition and enforcement.

Designing and implementing regulatoryarrangements for these sectors pose a numberof challenges for policymakers.2 On the onehand, regulated firms have incentives tomanipulate the regulatory process to increasetheir profits and are assisted in doing so by theinevitable information asymmetry betweenfirms and regulators. On the other hand, settingprices for most infrastructure services has apolitical dimension, and governments âcestrong political pressure to use regulation tokeep prices below the long-run costs of supply.Potential investors in infrastructure activitiesare aware of this risk and of the vulnerability oftheir usually large, long-term and highly-specific investments once they have been made.Governments need to be able to commitcredibly to regulatory policies to provide

1 An activity can be characterized as "naturally"monopolistic if a single supplier can meet all marketdemand at least cost. It is now recognized that agrowing number of infrastructure activities, includingpower and gas production and marketing and cellularand long-distance telephony, do not exhibit thisfeature, and a growing number of countries areintroducing competition in these activities.

2 See Smith & Klein (1994).

38

IIIIIIIIIIIIIIIIIII

Regulating Brazil's Infrastructure: Perspectives on Decentralization

investors with assurance of a fair return on theirinvestment. This condition is necessary bothfor attracting initial investment at reasonablecost and for encouraging efficient operation ofthat investment once it has been made.

Increasingly, the main strategy governments areadopting to commit to their regulatoryundertakings is to establish a regulatoryframework that limits discretion in regulatoryrules and places responsibility for the exerciseof residual discretion in an agency with therequisite expertise and independence."Independence" in this context includesmeasures to protect the regulatorydecisionmaking process arom being "captured"by short-term political pressures and by theregulated industry or other special interestgroups.3

The most appropriate means of implementingthis strategy will vary from jurisdiction tojurisdiction. In jurisdictions with stablepolitical and legal institutions and anestablished reputation for treating private firmsfairly, perceptions of lower regulatory risk mayreduce investors' demands for safeguardsagainst the misuse of regulatory discretion, suchas rigid and specific rules and independentagencies. More flexible rules will, in turn,enhance regulators' capacity to pursueefficiency goals and to adapt regulation tochanging economic and technologicalconditions.

In jurisdictions with less conducive conditions —which include the overwhelming majority ofdeveloping and reforming countries - the trade-off can be stark; investors respond toincreased risks by insisting on greatersafeguards or demanding higher returns oncapital. The less comfort provided by theregulatory system, the higher the risk premiumrequired to attract investment. Another trade-

3 For a useful discussion of the theory ofregulatory capture and its implications for the designof regulatory policies, see Neven, Nuttall & Seabright(1993).

off also arises: the lower the confidence in theregulator's independence, the greater theemphasis on specific, rigid rules and hence theconstraints on using regulation to pursueefficiency goals.

2. Assigning Regulatory Responsibility

The basic challenges and tradeoffs involvedin infrastructure regulation are inescapable,regardless of whether regulation is a national,state or municipal responsibility. However,several factors will influence the mostappropriate assignment of responsibilities.

Arguments for DecentralizingResponsibility

Four main arguments can be made forassigning regulatory responsibility forinfrastructure to lower tiers of government:

• It enables regulatory objectives andapproaches to be shaped by localconditions, priorities and preferences. Forexample, the "optimal" form and content ofregulation for Para may be entirelyinappropriate for Parana.

• It can reduce the information asymmetrybetween regulators and firms by bringingthe regulatory authority closer to affectedfirms. For example, a regulator based inRecite will be much better placed tomonitor a local concessionaire than aregulator located in Brasilia.

• It can improve the responsiveness andaccountability of the regulator by bringingthe regulatory authority closer to users.For example, consumers in Recife may feelmore confident that a local regulator willtake their concerns seriously than if theyhad to petition a more remote authority.

• It can foster experimentation with moreinnovative approaches to regulatoryproblems. For example, allowing sub-

39

Regulating Brazil's Infrastructure: Perspectives on Decentralization

national jurisdictions the freedom todevelop more flexible arrangements maylead to more positive social outcomes thanif a single template were imposed bynational authorities.

The last argument is sometimes characterizedas a benefit of "regulatory compétition."According to this view, governments competeagainst each other to attract mobile factors ofproduction — including workers and privateinvestment in infrastructure - through theirregulatory regimes. This competition createsincentives for governments to improve thequality of their regulation and to emulate theapproaches of successful governments.4

Possible Limits to Decentralization

Arguments of the kind outlined above areoften considered sufficient to support astrong presumption in favor of decentralizingregulatory responsibility. Depending on theindustry, jurisdictional units, and regulatoryissue in question, however, decentralizedapproaches may have weaknesses.

Spillover Effects. Decentralized regulatorshave weak incentives to take account ofspillover effects on other jurisdictions. Thismay result in insufficient production of a goodor service that generates positive externalities orexcessive production of a good or service thatgenerates negative externalities. A classicexample of the latter is inadequateenvironmental regulation of water systems thataffect downstream users in another jurisdiction.

In principle, affected jurisdictions can negotiatecoordinated responses to these problemswithout transferring regulatory responsibility toa higher tier of government; some countries

4 The notion of competing governments andjurisdictions was introduced by Tiebout and has beena popular idea ever since. For strong proponents ofthis view see Siebert & Koop (1993) and Easterbrook(1983). For a critical view of the application of thistheory to the regulation of different utility sectorswithin one tier of government see Helm (1994).

have established special mechanisms tofacilitate approaches of this kind In practice,however, such negotiations tend to be slow andlaborious, and incentives to comply with theresultant agreement may be weak.5 In thesecircumstances some centralization of regulatoryauthority may be desirable. Many constitutionsin federal systems, such as those in Argentina,Australia, Canada, and the U.S., address theseconcerns by assigning responsibility for"interstate" matters to the national government,although it often remains contentious as towhat degree of interstate impact is requiredto invoke national jurisdiction.

Inter-jurisdictional Trade. While not strictlyspillovers, a number of regulatory issues mayaffect inter-jurisdictional trade and haveimpacts beyond a single jurisdiction in this way.In some cases, efficient inter-jurisdictional trademay require harmonization of key technicalstandards. For example, adoption of differentrailway gauges between jurisdictions canimpose practical restrictions on interstate railtransport, as can some standards affecting tradein gas, electricity, waier, andtelecommunications. Similarly, regulatorypolicies that create subsidies for local producersor maintain local monopolies may distort orrestrict inter-jurisdictional trade. For thesereasons, matters affecting interstate trade aretypically regulated at the national level,including the terms and conditions of access tointer-state networks such as pipelines andtransmission grids.

Jurisdictional Size & Economies of Scale.Many infrastructure activities require largecapital investments and exhibit significant scaleeconomies. Some jurisdictions may be smallerthan the minimum efficient size for particularactivities. For example, it may not be efficientfor small jurisdictions to procure and regulatetheir own power systems, and some municipal-based water systems may be too small to attractprivate investors. In these cases, severaljurisdictions may need to collaborate in

5 See Neven, Nuttall & Seabright (1993).

40

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

Regulating Brazil's Infrastructure: Perspectives on Decentralization

elaborating and administering a commonregulatory framework, which may increasecosts and weaken the credibility andeffectiveness of the regulatory regime.Some of the potential difficulties in this area areillustrated by the attempt to grant a privateconcession for a water system in Caracas.Twenty-three municipalities had to cooperate toform a concession area of sufficient size toattract private investment. The difficulty ofestablishing a credible regulatory frameworkunder these circumstances has been identified asone of the reasons that no responsive bids werereceived.6

"Destructive" Competition. It is sometimessuggested that regulatory competitionbetween decentralized jurisdictions might bedestructive. One concern is that strongbargaining pressure from foreign investors,coupled with limited mobility of citizens,may lead sub-national jurisdictions to bid-upsubsidies or regulated rates of return or bid-down taxes or other obligations to attractinvestment. Evidence on this issue is mixed,but concerns of this kind may contribute topressures for jurisdictions to adoptcoordinated approaches.7

Constrained Regulatory Capacity.Regulation of infrastructure industries is acomplex and demanding task. Depending onthe detail of the regulatory regime, staff withspecialized economic, financial, and legal skillsmay be required. Moreover, if the regulatoryentity is intended to be independent,decisionmakers must be able to resist improperinducements or pressures from regulated firms,political authorities, and other interest groups.Such resources are scarce in many reformingand developing countries and often have a highopportunity cost. The resource pool typicallybecomes much shallower as one progresses tolower tiers of government, while closer

* Sec Triche, Mejia & Idelovitch (1993).7 See Siebert & Koop (1993) for an evaluation

of the findings of empirical studies on the impact ofinter-jurisdictional tax competition.

proximity to firms and to consumers mayincrease the risk of capture. The specialchallenges this may pose tor establishing andmaintaining effective regulatory arrangementsat sub-national tiers of government isconsidered in Section 3.

Striking a Balance

The optimal balance between national, stateand municipal regulation cannot bedetermined in the abstract. Much depends onthe characteristics of the industry,jurisdictional units, and the regulatory issuein question, as well as the broader legal andpolitical environment. In the latter regard, itis important to acknowledge that in manycases the assignment of regulatoryresponsibilities will be established inconstitutions or other political compacts, andhence may not allow much flexibility in thenear term. It is nevertheless instructive toexamine some of the main mechanisms thatcan be employed to manage the policy trade-offs involved.

Division of Policy Determination andImplementation. In the simplest model,regulatory policy is determined andimplemented by a single tier of government Inpractice, however, it is quite common to findregulatory policy set by one tier of governmentand for implementation to be delegated to alower tier. The degree of discretion conferredon lower tiers of government can varyconsiderably.

At one extreme, policy administration can bedelegated to a subordinate entity that has littleindependent discretion. For example, nationalantitrust policy in countries such as Australiaand Canada is administered in part by regionaloffices of national regulatory agencies; theyoperate according to identical rules and policiesand are expected to adopt a consistent nationalapproach. This limited form of decentralizationmay still improve responsiveness to consumersand help to ameliorate information asymmetriesbetween regulators and firms. At the other

41

Regulating Brazil's Infrastructure: Perspectives on Decentralization

extreme, broad regulatory goals and objectivesmay be determined at the national level butimplementation details left largely to lhediscretion of lower tiers of government. Oneexample is product standards in the EuropeanUnion: Member States often have substantialdiscretion over how to achieve the minimalrequirements specified by European UnionDirectives.8

Between the two extremes, nationalgovernments may delegate certainresponsibilities to sub-oational governments,but do so subject to certain conditions intendedto achieve a degree of national consistency andcoherence. A variation on this approach is forregulatory policy to be determined at thenational level, but with provision for sub-national jurisdictions to create exemptions. Forinstance, federal antitrust laws in Australia,Canada, and the US. permit states to createlimited exemptions in accordance with localpolicy preferences.9

Division of Different Regulatory Tasks. Inthe simplest model, all questions of regulatorypolicy concerning a particular industry are setby a single tier of government In practice,however, different aspects of one industry areoften regulated by diffèrent tiers. In Canada,for example, primary responsibility forregulating water utilities falls on municipalauthorities; environmental regulation is largelythe responsibility of the Provinces; and antitrustregulation is enforced at the Federal level.Similarly, electric utilities in Germany aresubject to economic regulation bymunicipalities (granting of concessions), Lander(rates), and the Federal government (antitrustlaws). In many countries, a single firm may besubject to environmental regulation at themunicipal, state, and federal levels.Approaches of this kind place a premium on

mechanisms to coordinate regulatory10

Formal Responsibilities vs. Induced orCooperative Approaches. In the simplestmodel, regulatory responsibilities for particularactivities and objectives are rigidly assignedbetween tiers of government by a constitution orsome other political compact In practice,however, constitutions and other arrangementsare subject to differing interpretations andapplications over time. No less important, theassignment of regulatory responsibility is oftenblurred by induced or cooperative approachesbetween jurisdictions. In many cases, nationalgovernments induce sub-national authorities topursue national goals or policies throughfunding arrangements. In the area of roads, forexample, the Australian and U.S. federalgovernments have a long history of usingconditions imposed on federal road funding toinfluence decisions on planning and regulatorypolicy.

As noted above, sub-national authorities mayalso work together to establish commonregulatory approaches to address sharedproblems, even without inducement fromfederal authorities. In some cases, specialintergovernmental forums have been establishedto facilitate cooperative arrangements of thiskind.11 The European Union can be seen as amore advanced model of the same approach,where national sovereignty is pooled accordingto a formal treaty with an elaborate institutionalapparatus to facilítate joint decisionmaking.12

8 See Gatsios & Scabright (1989).9 See Government of Australia (1993).

10 For a criticism of existing arrangements in theU.S., see Phillips (1993).

11 In Australia, for example, the Council ofAustralian Governments comprises the heads of theFederal, State and Territory Governments. It meetsapproximately twice a year with an agenda advancedby working groups of officials. There are also anumber of other formal and informal inter-governmental groups dealing with policy coordinationon specific issues.

12 See Keohane& Hoffmann (1991).

42

IIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIII

Regulating Brazil's Infrastructure: Perspectives on Decentralization

3. Confronting Constrained RegulatoryCapacity

Most of the policy considerations relevant toassigning regulatory responsibilities betweentiers of government are identical acrosscountries, developed and developing alike.However, constrained regulatory capacity isespecially important in many developing andreforming countries, and is likely to be a keyissue in Brazil, The principal concern is therelative scarcity (and high opportunity costs)of qualified regulatory personnel at lowertiers of government, with this concernexacerbated by the greater risk of regulatorycapture flowing from closer proximity toconsumers (voters) and firms. Theseconcerns affect the design of regulatoryarrangements and of strategies to strengthenthe regulatory capacity of decentralizedregulators.

Designing Regulatory Arrangements

As noted in Section 1, the main strategy forhandling the contracting problems associatedwith infrastructure regulation is tocircumscribe the discretion contained inregulatory rules and to allow an entity withthe requisite expertise and independence toexercise residual discretion. The scarcity ofskilled regulatory personnel and greaterexposure to industry and consumer pressuresmust be taken into account in designingregulatory arrangements. Some of the maindesign issues are considered below.

Discretion and Independence. In thisenvironment, potential investors are likely to beespecially sensitive to discretion in regulatoryarrangements, and have a strong preference fordetailed and rigid rules and standards. To theextent that discretion is tolerated - and as apractical matter, it is virtually impossible toremove all discretion - a premium will beplaced on measures that safeguard theindependence of the regulator. These measuresmay range from procedures affecting theappointment process and security of tenure to

procedures ensuring transparentdecisionmaking and the review of decisions. Tothe extent that these conditions are not met,investors will demand higher rates of return tooffset the regulatory risks they perceive.

Sectoral Breadth of Authority. Entitiescharged with regulating infrastructure can beorganized on one of three main bases,

depending on the sectoral breadth of theirauthority.

• Industry-specific. Separate agencies areresponsible for telecommunications, gas,water, electricity, rail, and so on.Examples of this approach include theUnited Kingdom and national levelregulators in Argentina.

• Sector-specific. Separate agencies areresponsible for sectors comprising a groupof related industries, such as an energyregulator for electricity and gas, a transportregulator for rail, roads, and ports and acommunications regulator fortelecommunications, post, andbroadcasting. Countries which use sectorregulators for energy include Colombia,Hungary, and the national-level regulatorsin the United States and Canada. Sectoralregulators are being established in Russiaand are being advocated as a reform optionin the United Kingdom.13.

• Multi-sectoral. A single regulatory agencyis responsible for all or most infrastructuresectors. Examples include state-levelregulators in the U.S., Canada andAustralia, as well as national-levelarrangements being implemented inJamaica and Bolivia.

Establishing a regulatory agency with a broadindustry base offers a number of advantagesover industry-specific agencies, particularly in

13 See Helm (1994).

43

Regulating Brazil's Infrastructure: Perspectives on Decentralization

countries with limited regulatory capacity andexperience. These advantages are even moreimportant when sub-national regulators areinvolved. The advantages include:

• Opportunities to share regulatory resources— for example, economists, financialanalysts, and lawyers can work acrosssectors, and administrative personnel canbe shared.

• Greater resistance to industry capture - theagency's broader constituency weakens theinfluence of any one industry.

• Greater resistance to improper politicalinterference - the broader constituency andhigher political profile increase the stakesof inappropriate political intervention.

• Application of consistent approachesacross industries - this has two mainbenefits. First, it contributes to greaterpredictability (and hence reduced risk) forinvestors. Second, it reduces the risk ofmarket distortions that can arise whenservices that are subject to substitutecompetition - such as gas and electricity ordifferent transport modes - are subject toinconsistent regulatory treatment.14

• Easy translation of lessons and experiencegained in one sector to other sectors - thiscan be particularly important in relation tothe design of efficient tariff structures andsome of the new regulatory challengesassociated with managing competition innetwork services such astelecommunications, electricity and gas.

Three broad strategies can be adopted forestablishing multi-sectoral regulatoryframeworks. First, a multi-sectoral frameworkcan be established at the outset, with eachsector brought within the regime at the time ofor before privatization. This will usually be thepreferred approach, and has been adopted in

14 This phenomenon is referred to "regulatoryarbitrage"; see Helm (1994).

countries including Bolivia. Second, if anindustry-specific regulator already exists, newindustries may be brought within its jurisdictionrather than creating additional entities. Thefeasibility of this approach will depend on howeasily the structure and operation of the initialinstitution can be modified to meet a broadermandate. Third, regulators can be establishedon an industry-specific basis but consolidatedover time through mergers. This happened inmany states in the U.S. and pressure ismounting to merge industry-specific regulatorsin the U.K., Argentina and Chile. This strategyhas a number of weaknesses, however,including the likelihood that existing entities willresist merger and that the benefits of a broaderperspective will not be available during thecritical early phases of a new regulatory system.

One of the main challenges in establishingmulti-sectoral agencies is to ensure thatspecialized knowledge or understanding ofparticular industries is not unduly sacrificed.This challenge can be addressed by ensuringthat internal organizational arrangements fosterthe development of industry-specific expertise,such as by establishing strong sector-specificdepartments. This approach has been adoptedby many state-level regulators in the U. S. andis being emulated in Bolivia.

Functional Breadth of Authority. Thequestion may arise whether the entityresponsible for economic regulation of anindustry should also have responsibility forsafety, environmental, anti-trust and otherregulation for the infrastructure industry.While international experience is mixed, threegeneral principles are widely accepted:

• Where regulatory capacity is limited, thefewer the regulatory agencies involved insupervising the industry the better.

• Where several regulators are involved, thejurisdictions of each should be defined asprecisely as possible to reduce uncertainty,duplication, or conflict.

44

III

IIIIIIIIIIIIIIIIIII

Regulating Brazil's Infrastructure: Perspectives on Decentralization

• Where two or more regulators areresponsible for closely related aspects ofthe same industry — such as water qualitystandards set by an environmental regulatorand water rates set by an economicregulator - effective coordination will beessential to ensure proper regulation.

Minimization of Regulatory Demands. Ifregulatory capacity is limited, this limitationmust be taken into account in designing allaspects of the regulatory system. Regulatorydiscretion can be limited and residualdiscretion can be tempered by clearlyarticulated criteria. Administrativeprocedures intended to make regulatorydecisionmaking transparent should avoidexcessive legalism, a criticism often leveledat regulatory processes in the U.S. Theburden of supervising compliance with rulescan be reduced by ensuring that penalties areclearly defined and set at a level that willprovide appropriate incentives for firms. Itmay also be feasible to enhance the role ofconsumers in monitoring firm behavior byproviding incentives to report non-compliance. For example, in addition tofines, consumers may receive rebates if afirm has foiled to meet clearly definedperformance obligations: New Zealand andthe U.K. are adopting this approach intelecommunications.

Contr acting-out Regulatory Tasks.Another way to reduce demands onregulators is to allow them to contract-outparticular tasks. For example, some of theanalytical work associated with priceregulation could be contracted-out toconsultants, and external arbitrators could beused to settle certain types of regulatorydisputes. Similarly, some regulatory tasksmay be delegated to regional or nationalagencies. In each case it will be important toensure that the proposed arrangements do notraise concerns over capture or otherwiseweaken the legitimacy or effectiveness ofregulatory decisionmaking.

Strengthening Regulatory Capacity

Even where regulatory arrangements havebeen designed to reduce regulatory demands,nascent regulators will typically requireassistance in a number of areas, includingtraining in specific regulatory functions.This type of assistance will usually exhibitsignificant scale economies, as trainingprograms and other materials may be sharedby several decentralized regulators. Anumber of strategies exist for deliveringassistance of this kind.

Assistance Through National-levelAgencies. A national policy favoringdecentralization of regulatory responsibilitymay be complemented by technicalassistance channeled through a national-levelagency. For example, such an agency couldorganize training programs for regulators,provide model contracts for concessions, orprovide technical advice on specificregulatory problems. It may also act as afocal point for disseminating lessons ofexperience among sub-national regulators.

Mutual Assistance Through a RegulatoryAssociation. An alternative orcomplementary strategy is to establish anassociation of regulators. Sub-national (andnational) regulators can interact in this forumand share lessons of experience. Theassociation may also reinforce professionalnorms (and hence reduce vulnerability toimproper influences) while providing a group ofpeers to respond to charges of improperpolitical interference. It could also provide aframework for the joint development of trainingprograms and research on common regulatoryproblems. The National Association ofRegulatory Utility Commissioners in the U.S.provides a possible model.

Twinning Arrangements. A third vehicle forsupporting nascent regulators is establishing"twinning" relationships between sub-nationalregulators and more experienced regulators,whether from foreign or national jurisdictions.

45

IRegulating Brazil's Infrastructure: Perspectives on Decentralization

Arrangements of this kind can provide a sourceof ongoing institutional support, technicaladvice and exchange and training opportunities.A growing number of U.S. and other OECDregulators have experience in participating inthese relationships with regulators in reformingand developing economies.

4. Conclusions

IIIIII

Selecting the most appropriate assignment ofregulatory responsibility for infrastructurebetween tiers of government in Brazil willcall for complex policy judgments, oftenmade in a sensitive political environment.This choice cannot be made in the abstract orby appeal to general slogans: carefulconsideration of the particular industry,jurisdictional units, and regulatory issues isessential. In many cases the most «appropriate response may involve a division Iof responsibilities between policyformulation and implementation, and _between different regulator)- tasks. I

While most policy considerations are •identical across countries, constrained Iregulatory capacity is of special concern in areforming country such as Brazil. This _constraint must be given due weight in Iframing any privatization anddecentralization strategy. and when _considering the detail of regulatory and •institutional arrangements. In particular,sub-national governments may need to find _ways to resist the pressures that can arise for Ieach industry to have a separate industry-specific regulatory agency. Thought willalso need to be given to strategies for Istrengthening regulatory capacity at all tiersof government.

46

IIII

I• Bibliography

I Easterbrook, F.H. (1983), "Antitrust and the Economics of Federalism", Journal of Law& Economics, 26, 23-50.

| Gatsios K. & P. Seafaright (1989), "Regulation in the European Community", OxfordReview of Economic Policy 5, 37-60.

I Government of Australia (1993), National Competition Policy : Report of theIndependent Committee of Inquiry (Australian Government Publishing Service,

• Canberra).

Helm, D. (1994), "British Utility Regulation : Theory, Practice & Reform", Oxford Review• of Economic Policy, 10, 17-39.

Keohane, R.O. & S. Hoffmann (eds) (1991), The New European Community :I Decisionmaking & Institutional Change (Westview Press; Boulder, Co).

INeven, D.s R. Nuttall & P. Seabright (1993), Merger in Daylight : The Economics &

Politics of European Merger Control (Centre of Economic Policy Research, London).

- Phillips, CF. (1993), The Regulation of Public Utilities : Theory and Practice (PublicJ Utilities Reports; Arlington, Va).

I Siebert, H. & M. J. Koop (1993), "Institutional Competition versus Centralization : QuoVadis Europe?", Oxford Review of Economic Policy, 9, 15-30.

I Smith, W. & M. Klein (1994), "Infrastructure Regulation : Issues & Options", (mimeo,

World Bank).

I Triche, T., A. Mejia & E. Idelovitch (1993), "Arranging Concessions For Water Supply &

Sewerage Services : Lessons From Buenos Aires & Caracas", World BankInfrastructure Notes (May).

IIIII 47

48

AN INTERDISCIPLINARY JOURNAL OF PUBLIC POLICY ANALYSIS

ARTICLES

Peter Bauer B. R. Shenoy: Stature and Impact

The Côte d'lvoire's Troubled Economy:Why World Bank Intervention Failed

Deregulated Private Water Supply: APolicy Option for Developing Countries

Harmonizing Ethnic Claims in Africa:A Proposal for Ethnic-Based Federalism

The Unintended Consequences of TradeAdjustment Assistance

The Problem of Lemons and Why We MustRetain Juvenile Crime Records

A Vote for Perot Was a Vote for the Status Quo

Securities Units of Banking Conglomerates:Should Their Location Be Regulated?

Legal Institutions and Abortion Ratesin Mississippi

China's Future: Market Socialism orMarket Taoism?

BOOK REVIEWS

49 .VOLUME 18 NUMBER 1 • PUBLISHED BY THE CATO INSTITUTE

Pascal Wickand Jane S. Shaw

Penelope Brook Cowenand Tyler Cowen

Mwangi S. Kimenyi

Cecil E. Bohanonand Marilyn Flowers

T. Markus Funkand Daniel D. Polsby

William A. Niskanen

João A. C. Santos

Charles A. Barbourand W. F. Shughart II

James A. Dorn

III

DEREGULATED PRIVATE WATER SUPPLY-.A POLICY OPTION FOR DEVELOPING I

COUNTRIESPenelope Brook Cowen and Tyler Cowen I

The Privatization Alternative IMany citizens in developing and transition economies are excluded g

from enjoying safe and reliable water supply. In many cities, 30 to |60 percent of the population has no formal water hook-up at all, butrather must resort to wells, buckets, supply by tanker-trucks, and •physical transport of water through human labor and beasts of burden. •

A few simple facts illustrate the serious nature of this problem. InJakarta, 75 percent of the population has no formal connection; in IMaputo 65 percent. In Madras, the percentage served is around 50percent; and even in relatively prosperous Manila, 29 percent of the •citizenry has no connection. When individuals must resort to nonpiped |water sources, prices are often at least 10 to 20 times higher. InLuanda, where the price for piped supply is around nine cents per •cubic meter, households can pay as much as $16.00 per cubic meter •for tanker supply. Table 1 portrays some connection rates and price _differentials. I

The fundamental problem is institutional rather than technological.Tariffs set by governments at levels below cost recovery fail to encour- •age inclusion. In developing countries, water utilities recover on aver- Iage around 30 percent of their total costs (World Bank 1994). As a

Cato Journal, Vol. 18, No. 1 (Spring/Summer 1998). Copyright © Cato Institute. Allrights reserved. •

Penelope Brook Cowen is a Private Sector Development Specialist at the World Bank, |and Tyler Cowen is Professor of Economics at George Mason University. This paper hadits beginnings in a conversation with Vincent Gouarne in Tirana, Albania, in June 1995. _The authors would also like to thank Meyer Burstein, David Ehrhardt, Timothy Irwin, •Daniel Klein, Michael Klein, Warrick Smith, Alex Tabarrok, Katarina Zajc and an anonymous ™reviewer for their suggestions and comments. The interpretations and conclusions expressedin this paper are entirely those of the authors. They do not necessarily represent the viewof the World Bank, its Executive Directors, or the countries they represent.

50

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

TABLE 1PRICES FOR PIPED AND VENDOR SUPPLIES IN SELECTED

DEVELOPING COUNTRIES

Percent Price for Price forwithout Piped Water Vendors

City and Year Connections (US cents/m3) (US cents/m3)

Bandung (1991) 61 9̂ 9 616Jakarta (1991) 75 17.2 185Manila (1992) 29 10.5 187Karachi (1992) 17 7.5 175Ho Chi Minh (1991) 35 7.6 151

SOURCE: Asian Development Bank (1993).

result, Utilities have no incentive to deliver services to large sectorsof the population, almost always low-income households. Incentivesfor research and development are similarly weak, given that the price-controlled monopolist cannot capture the full benefits of a new prod-uct idea.

The human costs of these institutional arrangements have beenvery high. According to one estimate (the World Health Organization,cited in Cooper 1997), contaminated drinking water accounts for 80percent of disease in India, including a sizable share of the 500,000Indian children who die each year from diarrhea. Around the world,diarrheal diseases kill more than 3 million people annually, and causeapproximately 900 million episodes of illness (World Bank 1992).

We propose that unregulated privatization be considered as onemeans of limiting these tragedies. To date, the world has experimentedwith four kinds of institutional regimes: outright public provision ofwater (common throughout the world), government-supported naturalmonopoly with regulated price (the English model), government-supported natural monopoly with regulated rate of return (the Ameri-can model), or a government-controlled franchise, lease, or concessionagreement (the French model and its variants). We add a fifth possibil-ity: complete privatization of water assets and unregulated naturalmonopoly. This scenario involves no price regulation, no rate of returnregulation, no residual government ownership of assets, and no surrep-titious regulation through antitrust law.

The rationale for unregulated privatization is straightforward. Anunregulated private monopoly would have an incentive to bring asmany potential buyers into the system as possible, so as to maximizeprofit. Unregulated private monopolies could thus significantly

51

IDEREGULATED PRIVATE WATER SUPPLY

increase the number of water connections in developing countries.If unregulated privatization could produce hook-ups for currently •neglected low-income customers, the poor would end up with higher •real incomes, better water service, more time for other endeavors,and a greater probability of a long Ufe. London water supply in the Iearly 18th and 19th centuries, which was private and relatively unregu-lated, had a favorable record for extending the number of connections •(see Dickinson 1954: 102-3). I

While standard theory emphasizes the output-restricting nature ofmonopoly, water utüities will use price discrimination and fixed hook- 8up fees (Oi 1971) to capture as much profit as possible, thereby ™increasing supply in the process. UnUke the governmental and regu- _lated alternatives, a private unregulated monopoly also would have |strong incentives to hold down costs and supply an optimal quality ofproduct. Our main point is that this monopoUstic alternative deserves •serious consideration. I

In a comparative institutional context, the more heavily regulatedalternatives may end up excluding more potential buyers. Developing Ior transition economies, regardless of their historical background orgeographic locale, tend to share common problems with their govern- «ments. These governments have relatively low levels of credibiUty, |weak track records, and very short time horizons. The governmentperforms especially poorly as an owner or regulator, partly through lack •of experience and partly through improper incentives and corruption.ï •

In Guinea, for example, progress under a water lease has beenhampered by problems in defining and implementing the regulatory Ifunction, and by continuing disputes between the government-ownedwater holding company and the private water operator over who is •responsible for failures in service expansions and water loss reductions. 8While the lease contract has increased the number of connectionsand improved water flows, progress has fallen short of expectations B(Brook Cowen 1996). •

Unregulated privatization has received Uttle serious attention, and _has generally been rejected or dismissed, albeit without serious analyti- |cal consideration. Few sectors have been classified as market failuremore universally than the supply of water to households and resi-dences. Throughout the world, water systems are characterized byextensive government ownership or thorough regulation and control.Water supply typically is regarded as a natural monopoly, and thereforea poor candidate for unregulated market provision. In the technical

'See, for example, Zajc (1996) on the incidence of these problems in transition countriesand their implications for water privatization.

52

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

literature, Guislain and Kerf (1996) restrict their discussion of infra-structure privatization to concessions and divestitures accompaniedby a regulatory license. Klein (1996) takes as a starting assumptionthat the weakness of competitive pressures in the water sector createsa need for at least some form of regulation. Breyer (1982:17) considersa price-discriminating natural monopolist briefly, but does not explorethe unregulated alternative in depth. Loeb and Magat (1979) attemptto replicate private price discrimination through regulation; theirscheme has the government award monopolists for the consumersurplus they generate. Armstrong, Cowan and Vickers (1994) essen-tially argue that if competition is not possible, regulation will benecessary.

In the remainder of the paper , we provide an analysis of potentialsources of inefficiency in water markets, explain why governmentownership and regulation have failed to provide fully satisfactoryresults, discuss how unregulated privatization might resolve the effi-ciency problems that plague regulation and government ownership,consider the ability of an unregulated monopolist to price discriminatewhen selling water to residential users, and focus on some residualproblems with the proposed policy option of unregulated privatenatural monopoly. We find that the case for unregulated privatizationis not conclusive, but that the proposal deserves serious consideration.

Sources of Efficiency LossThe relevant natural monopoly problem comes from the distribution

of water rather than from water itself. Water, considered apart fromthe problem of distribution, satisfies the traditional definition of aprivate good—nonrivalry in consumption. If one person consumessome water, another person cannot use the same water. While thereare public health benefits to a clean water supply, the private benefitsof clean water are high as well, giving individuals a strong incentiveto pay for water quality. Consistent with the private-good nature ofwater, we observe the efficient private supply of water in a variety ofcircumstances. We buy bottled water at the supermarket, and private,for-profit car washes supply water to clean our cars. The private sectorhas had considerable success in supplying and running wells, at leastwhere wells are a reasonably efficient means of water delivery.

The construction and maintenance of water distribution networkspresents the difficult problem, and the potential cause of marketimperfection. Once a system of water pipes is built, the owner of thesystem has a monopoly advantage in the market. If only one set ofpipes exists, the owner of those pipes can exercise market power and

53

IDEREGULATED PRIVATE WATER SUPPLY

charge a price for water in excess of its marginal cost. As discussedabove, other means of obtaining water, such as water delivered by fttanker-vendors, typically involve costs from 10 to 20 times higher than Ibuying water through a piping system. Alternatively, we might imaginea system of competing pipes, or competing pipe systems. Such systems Idid occur in Canada and the United States in the 19th century, anddo still occasionally arise when water of different qualities is being msupplied- For example, in Hong Kong seawater pipes supply flushing |water (Klein 1996). In this case, however, consumers ultimately mustfinance both piping systems. The high prices needed to recoup the •costs of multiple piping systems will imply a restriction of water •output, just as the monopoly did. Furthermore, the stability of market m

equilibrium is problematic when multiple, competing suppliers own Inetworks with high fixed costs (Bittlingmayer 1982). Competition willtend to force prices back down to marginal cost, but at marginal cost •no supplier can break even and recover the fixed costs spent on 8constructing the piping system.

Some treatments emphasize market means of overcoming the natu- Iral monopoly problem. Under one proposal, water is distributed "through a club-owned network, with different suppliers competing gagainst each other to win contracts with consumers. Either the suppli- Jers or the consumers themselves own or control the club. We seemerit in this idea, but for the purposes of exposition we assume that •the natural monopoly problem cannot be overcome so easily. The ™potential ability to make the market competitive, however, would onlyfavor our basic proposal.2 I

Assuming that competition is not possible, the fundamental probleminvolves the construction of a distribution network with fixed costs mthat are high relative to marginal costs of supply. The problem behind |the private provision of water thus resembles analogous problemswith the sale of cable television services, electricity, and natural gas.

Institutional regimes for water provision face three kinds of effi-ciency problems: inefficient levels of output, inefficient levels of cost,and inefficient levels of product quality. A non-price-discriminatingmonopolist, in the absence of regulation, will set price above marginal

also see some problems with the club proposal. Even if many suppliers competeby selling water services through a single pipeline, the fixed costs of the pipeline still mustbe covered somehow. Competitive pricing allows no means of financing the pipeline andallowing each company to break even. Presumably some kind of Ramsey pricing is necessary,where inelastic demandeis face the highest mark-ups, but this introduces some of the welfarelosses of market power. Along other lines, Demsetz (1968) analyzed ex ante competition "forthe market." In this proposal utilities offer competitive bids to communities for watersupply; see below for a discussion of franchising.

54

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

cost and restrict output, compared to a first-best social optimum. Thatis, consumers would be willing to pay more for additional units ofoutput than it would cost society to produce them. The monopolistdoes not expand output, however, because the extra units of outputcould be sold only by lowering the price for all units and thus reduc-ing profits.

Regulators have found it difficult to address this problem of monop-oly without inducing other distortions. One approach grants privateownership but places a cap on price (the British model for waterprovision). Placing a cap on price, however, gives the private supplieran incentive to skimp on service and product quality. As with all pricecontrols, the supplier will raise the real price to its desired level bylowering the quality of the product. Not only will quality decline inthe short run, but long-run investments in system maintenance willbe suboptimal. As discussed above, this problem is particularly drasticin developing and transition economies.

An alternative method of regulation, common in the United States,uses rate of return caps to limit the profits of the private monopolist.In practice, rate of return regulation usually involves price caps aswell, whether implicitly or explicitly, and in that regard also leads toskimping on service and product quality. Furthermore, rate of returnregulation brings a new set of distortions in the form of higher costs.As rate of return regulation is practiced, firms typically are guaranteeda minimum as well as a maximum rate of return. Without the minimumguarantee firms would not participate in the arrangement, given thatthey have sacrificed upside potential for profit. Firms therefore canuse cost increases as a justification for price increases; not surprisingly,the incentive to keep down costs is low. The end result is high costsand a lower level of water consumption than is optimal.

Leasing and concession agreements, in their various forms, provideyet another attempt to overcome the basic problem with naturalmonopoly. These institutional arrangements, however, do not avoidthe fundamental problems associated with regulation. Leasing andconcession agreements typically regulate prices and rates of returnto various degrees, either implicitly or explicitly. In this regard theyinvolve the welfare losses associated with price and rate of returnregulation. Leasing and concessions may provide for a looser or moreinformal kind of regulation, given the ongoing relationships betweenthe water company and the relevant government, but in the finalanalysis either the supplier is free to adjust its prices or it is not. Ifprices can be set freely, we return to de facto unregulated naturalmonopoly (of course this may be an advantage of concessions, as willbe discussed below). If prices and rates of return are not free to adjust,

55

IDEREGULATED PRIVATE WATER SUPPLY

Iwe return to the distortions of regulation and the weak incentives toexpand the number of system hook-ups. Particular problems arise •where concession contracts mandate expansions into low-income •areas, while also mandating "life-line" (below-cost) tariffs for low-income consumers. I

Leasing and concession agreements involve further distortionsthrough the government's role as residual asset owner. As the leasing mor concession agreement nears an end, the private concessionaire has |an incentive to cease maintenance or even strip the water assets.Leasing works poorly when the company faces a short time horizon. IThe government can alleviate these problems by promising a forth- ™coming renewal of the lease, or by offering comprehensive provisions —

for compensation upon contract termination, but if these promises Iare credible, leasing and concessions do not differ greatly from assetprivatization with regulation. If the promise is not credible, we return •to poor incentives for maintenance3. I

A fourth proposal involves outright government ownership of waterassets and full governmental control. The record of governmental Iprovision in this sector, however, is extremely poor. In developing "countries, where government ownership has been the norm, tariffs gare routinely set well below cost recovery levels, routinely less than Jhalf of supplied water is actually paid for, and large segments of thepopulation go without formal services (World Bank 1994). For political Mreasons, governments have weak incentives to reduce costs, price Iwater at marginal cost, maintain water systems, introduce innovations,and cut staffing to efficient levels. Not surprisingly, countries around Ithe world are moving away from the government ownership option,and embracing various forms of private sector participation (Rivera g1996). Table 2 shows examples of private sector contracts that are |now in place.

Sketch of an Unregulated Natural Monopolyfor Water

Consider a scenario where a government allows complete privatesector ownership of all water system assets, including the impoundingof bulk water, water treatment, and distribution. The private sectorwould own all water system assets (which may or may not be verticallyintegrated) just as the private sector owns the assets in the automobileindustry or the computer industry. Furthermore, suppose that the

^Zajc ( 1996) provides a comprehensive survey of the efficiency problems with concessionsand leases.

56

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

EXAMPLES OF

ContractualType

ManagementContract

Lease

Build-Operate-Transfer

Concession

Divestiture

TABLE 2

PRIVATE SECTOR ARRANGEMENTS

SANITATION

Water

ColombiaGazaMalaysiaTurkeyGuineaItalySenegalSpainAustraliaChinaMalaysiaThailandCôte

d'IvoireMacaoSpainEngland and

Wales

Sanitation

UnitedStates

ChüeMexicoNew

ZealandMalaysia

IN WATER AND

Water andSanitation

Puerto RicoTrinidad and

Tobago

CzechRepublicFrancePoland

ArgentinaFrancePhilippines

England andWales

SOURCE: World Bank (1997).

owner of the water assets could set prices and quantities withoutregulatory interference. Water suppliers and customers would relysolely on contract to set the terms and conditions of water delivery,and the courts would agree to uphold any contracts which are written.The absence of regulation, as defined in this paper, also implies acredible laissez-faire antitrust policy with regard to pricing and outputdecisions. If water companies set their prices with an eye to avoidingcharges of "anticompetitive behavior," or "price gouging," we wouldreturn to an implicit form of price regulation. The laissez-faire antitrustpolicy also would allow complete freedom of merger and cooperativerelations across differing firms.

Our use of the term "unregulated" refers to the absence of aset of government regulations found in today's regimes—specifically,restrictions on asset ownership, pricing, service delivery, etc., andexclusivity arrangements. Under laissez faire, the provision of servicesis regulated by market forces and economic incentives. In this sense

57

IDEREGULATED PRIVATE WATER SUPPLY

our analysis compares one kind of regulation to another, rather thancomparing regulation to an unregulated state of affairs. Furthermore, Iall regimes possess an implicit form of government regulation through "ex post liability law. In the scenario we consider, private water suppliers _would remain liable to lawsuits for breach of contract, fraud, or provi- |sion of water of dangerous quality. Nonetheless we continue to usethe word "unregulated" for purposes of expositional simplicity and •for lack of a more accurate descriptive term. I

The forces for natural monopoly within a single geographic areawould be strong in an unregulated environment. Experience suggests Ieconomies of scale in the operation of distribution networks for popula-tions of at least 50,000 to 100,000 people. Economies of scale in _system management as a whole are more extensive (evidence from |Britain indicates that managerial economies of scale are exhausted atpopulations of 500,000 to 1 million). The natural monopoly may be •limited at certain margins, such as when industrial users develop their •own wells. At the residential level, some households may find it moreprofitable to dig wells, or to collect and store rainwater for at least Isome uses. For the typical residential user, however, we envisage asituation where water can be obtained at lowest cost from a single «dominant supplier within that geographic region. |

We expect suppliers to offer standard packages to their consumers.If an individual is building a house, the water supplier will offer to •outfit the house with pipes for some fixed sum, perhaps based on the •value of the house and the neighborhood. Where real estate developersare responsible for installing household connections, they routinely Iuse this approach. If a house is already in place and already possessesa hook-up (perhaps as a legacy from a previous, regulated regime), mthe water supplier will offer so many units of water at a given price, |so many more units at another price, and so on. Households willeither accept or reject these offers, depending on the promised bundle Iof price and service. •

The water company has strong incentives to set initial offers thatwill be accepted. The company will try to capture as much surplus Ifrom each household as possible, but the company also wishes toensure that each household signs up to purchase water. Given the •initial assumption of natural monopoly, the company can serve subse- |quent households at relatively low marginal cost. Note that in thepolar case where the company has perfect knowledge of household Idemands, and can precommit to a series of price offers, a "first-best" •result will obtain. The water company will extract all of the consumer ~surplus associated with water purchases. We do not present the first- |best as an attainable real world outcome; the relevant comparison is

58 I

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

between imperfect markets and imperfect government regulations.Nonetheless, presentation of the first-best illustrates some basic incen-tives behind unregulated monopoly and also serves as a foil, by contrastallowing us to see ways in which unregulated monopoly falls short ofan ideal outcome.

The relevant consumer surplus can be extracted in either of twoways. The company may charge a fixed fee for a hook-up, and thensell remaining water units at marginal cost over some specified periodof time. Both the fixed fee and the subsequent per unit prices wouldbe determined by initial contract; Oi (1971) has analyzed the efficiencyof this arrangement. Alternatively, if the hook-up is already in place,or if it is too costly to bargain over the hook-up fee, companies willsimply supply the hook-up and then sell water at some price aboveits marginal cost of production.

This situation, if it can obtain, solves all three of the efficiencyproblems discussed above. First, the supplier will produce a sociallyoptimal amount of output. For any unit whose value exceeds itsmarginal cost, the supplier will produce it and offer it on the market.With perfect price discrimination, a supplier never increases profitby withholding output from the market. Second, the supplier has first-best incentives to engage in cost reduction. Any reduction in coststranslates into a one-to-one increase in profits. Suppliers thereforewill reduce their costs to the point where the social benefits of costreduction equal the social costs. High costs cannot be socialized butrather eat directly into profits. Third, a perfect price-discriminatingmonopolist has first-best incentives with regard to product quality.The supplier captures all of the consumer surplus in the form ofprofits. That same supplier will therefore offer tibe product qualitiesthat maximize consumer surplus, net of the cost of production.4

The ability of a natural monopolist to perfectly price discriminatemay be problematic, under a variety of assumptions. For that reasonthe first-best results may not strictly hold. Nonetheless an unregulated,privatized natural monopoly obtains first-best results under the basicassumption that the water company succeeds in maximizing its profit.Even in a second-best setting, the monopolist may produce a greaterquantity and quality of water outputs than do today's highly regulatedalternatives. In most developing and transition economies the keyproblem is to get users some minimal amount of clean water, not tosatisfy all optimality conditions.

Note that the potential efficiency of price discrimination also indi-cates why leasing and concession agreements, and divestitures with

4PhIips (1983) provides an overview and survey of the economics of price discrimination.

59

IDEREGULATED PRIVATE WATER SUPPLY

a license, may sometimes result in first-best or near first-best outcomes.If the company holding the concession has sufficiently cozy relations •with the host government, that company may be given latitude to 'replicate the efficient price-discriminating natural monopoly outcome.Quantity and quality decisions will again be optimal, if the "unregu- Ilated" outcome can be obtained under the guise of regulation. Inpractice, however, governments frequently impose uniform tariff •rules, or otherwise restrict price discrimination by regulated private Iwater companies.

The Feasibility of Price Discrimination for WaterPrice discrimination is most feasible when four primary conditions ^

hold. First, the seller must hold some degree of market power. Second,the product cannot be vulnerable to low-cost resale from low-price Ibuyers to high-price buyers. Third, the seller must be able to makegood estimates of buyer demands. Fourth, the supplier must be able •to commit to initial price offers. Each of these assumptions character- Iizes the water market to some degree.

We take the presence of market power as given, and as following Ifrom the natural monopoly assumption. If somehow no market powerwere present, price discrimination would be impossible, but a regime •of unregulated private water supply would in any case prove effective. |

The absence of cheap resale from low-cost to high-cost buyers alsofollows from the natural monopoly assumption. By construction of Bthe example, it is much cheaper to sell the water through a system ™of pipes than through bottles, wells, and buckets. Even if some resalewere possible, however, market demands would shift without changingthe fundamental nature of the problem. Assume, for instance, that inthe absence of resale low-valuation buyers would be charged $20 andhigh-valuation buyers would be charged $100. Now consider resale,which is profitable at any price above $80 to the high-valuation buyers.The high-valuation buyers will refuse to pay more than $80, andthe price-discriminating monopolist must lower prices accordingly,presumably to just below $80. Even at this lower price an optimalquantity of output is still produced, and the monopolist still has fullincentives to economize on costs at the margin. Optimal quality cannotbe guaranteed, since the monopolist cannot necessarily reap the fullbenefits of a quality improvement (higher prices for quality improve-ments may be undercut, implying that the innovator cannot reap allof the new consumer surplus that is produced), but some incentivesfor quality improvement remain nonetheless.

60

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

The third and perhaps most problematic condition for effectiveprice discrimination is whether the seller can predict the marketdemands of the buyers. The water supplier will estimate two differingfeatures of water demand: how much a given buyer values having anywater connection at all, and how much a given buyer values subsequentunits of water. We envisage a market where the water supplier setsprice by examining the previous use patterns of the water buyer, thevalue of the water buyer's property, and the wealth of the neighbor-hood. In wealthier areas the supplier may consider the number ofbathrooms in the house, whether the water buyer has a lawn, andother pieces of ancillary information, such as the water buyer's age,job, or credit record.5 Insofar as water demand is closely correlatedwith observable characteristics of the property and the buyer, effectiveprice discrimination will be relatively easy. The water supplier willrun information on the buyer and the property through its "pricingoffice," which will respond with a suggested price offer, both for initialservice and for successive units of water use. In Los Angeles, forexample, the water department has the capacity to customize basetariffs across consumers, according to such factors as lot size, tempera-ture zone, and size of household (Mann 1996).

Price discrimination will inevitably be imperfect in practice. Pricessometimes will be set too high, thereby excluding buyers from eitherparticipation in the piping network or from the purchase of additionalunits, even when the social benefits of added output would exceedthe social costs. While some inefficient exclusion will occur, watersupply may well be higher than under most current regimes in develop-ing economies.

Even when suppliers make pricing mistakes, they need not excludebuyers altogether. To the extent that monopoly power is considerable,price will exceed marginal cost by a large amount, and the profits ofwater sales will be large. Each excluded buyer represents a chunk offoregone profit. Consider the position of a water company whichbelieves that a given buyer values regular water use at, say, $1,000,and where the company can produce those same water services at acost of $300. If the company knows that the buyer's valuation is inthe neighborhood of $1,000, but the company is not sure about theexact valuation, the company will more likely price the services toolow rather than too high. If the company charges $1,001, it loses $700of potential profit. The expected return to guessing low will tend toexceed the expected return to guessing high. The microeconomic

'Since U. S. public utilities routinely run credit checks, this need not involve a significantloss of privacy.

61

DEREGULATED PRIVATE WATER SUPPLY

intuition here is simple: individuals or institutions which face a goodchance of capturing a significant prize will behave cautiously when theyare within range of winning the prize. For similar reasons, monopolisticfirms in other contexts will choose high levels of product safety, reliableservice, and easy access to their product, all in the desire to protecttheir monopoly profits (Klein and Lejffler 1981).

The excluded buyers will tend to be those whose valuations do notmuch exceed the marginal cost of producing water services. If themarginal cost of production is $300, and the buyer values service at$320, the firm has less marginal profit to lose by trying to squeezeout all of the buyer surplus. Some of these buyers may end up excluded,since the firm will sometimes guess incorrectly and offer a take-it-or-leave-it price above $320. Even when exclusion results, however, thewelfare costs of this exclusion tend to be relatively low. In the example,the buyer valued the product only slightly more than its marginal costof production. When expected profit, and expected social surplus, arelow, fewer resources will be spent trying to capture that profit andsome potential gains from trade may be foregone.

If such resulting instances of exclusion prove unacceptable, perhapsfor reasons of fairness or equity, a government may decide to intervenein the market and require service to low-income buyers at pricesthey can afford. In this case our proposal would cease to be purelyunregulated, and would involve the costs of price controls, at leastfor some buyers or some neighborhoods. This outcome, however,represents a worst case scenario for our proposal, which still appearsto provide superior overall performance, compared to a regime withfull regulation across all contracts and all buyers.

Many cases of harmful exclusion will come in the form of overpricedmarginal units, rather than overpriced fees for basic hook-ups. Compa-nies often will choose price discrimination in the form of a schedule,where the prices for water services vary with the quantity consumed.Assume that a buyer values the first unit of water services at $100,the second unit of services at $60, and the third unit at $30. Thecompany will try to offer a price schedule that matches these demandsexactly, but if the company calculates demand incorrectly, it mayoffer, for instance, a schedule of $100-$60-$40, thus excluding thebuyer from the third unit of water services. The buyer will take shortershowers than would be socially optimal, but some amount of safewater will still be supplied.

Fragmentary data and lack of experience with unregulated privatiza-tion prevent us from offering an empirical assessment of the relativemagnitudes of these exclusion costs across institutional regimes. None-theless we see no prima facie case for dismissing the unregulated

62

CATO JOURNAL

alternative. The unregulated monopoly has a continual incentive toreduce exclusion problems, whereas the regulated monopoly doesnot, and may even have an incentive to increase costs and thereforeprices, such as under rate of return regulation.

The foregoing discussion has assumed that water companies makesingle, take-it-or-leave-it offers, which customers must either rejector accept. The analysis becomes more complex if the company mustengage in bargaining with its customers.

Bargaining with customers may have either positive or negativeeffects on welfare, compared to the take-it-or-leave-it alternative.When bargaining is present, some of the initially excluded customersmay receive price reductions until they are no longer excluded. Low-valuation buyers face a lesser danger of complete exclusion. On thenegative side, consumers may hold out for excessively low prices, ifthey cannot observe the marginal cost of the firm. If the marginalcost is $30 and an individual values the service at $40, the individualmay nonetheless hold out for a price of $20, in the mistaken beliefthat marginal cost is $19. Since customers probably cannot observethe marginal cost of the firm with great ease, the potential for suchlosses exists. Furthermore, some quantity of real resources will beconsumed in the bargaining process. Customers may delay buyinghook-ups or may try to masquerade as low-valuation buyers, forinstance, or the company may invest in signaling its resoluteness asa bargainer. All of these real resource investments are made for thepurpose of receiving transfers, and thus violate first-best efficiency.

We expect that bargaining costs will be a significant issue only forvery large users, such as large businesses or perhaps condominialdevelopments that buy their water services collectively. We envisagethe water company as being able to commit to a price offer to individualusers, rather than having to bargain on a house-to-house basis. Mostunregulated large-scale suppliers of household services offer theirwares on precisely such terms. If a city has only a single newspaper,for instance, that newspaper may be sold at a price above marginal cost.Yet the newspaper company does not bargain with each household, butrather can precommit to a given schedule of prices, and then sell papersto interested subscribers. We expect a similar practice to develop withwater. Bargaining over prices is most likely when the purchase isoccasional, rather than repeated, and when the item has significantvalue, such as an automobile, a home, or an expensive painting. Evenin these cases, such as with automobiles, bargaining is often largelya ritual of convergence on a publicly available "book price."

Those institutions that can bargain with the water company, suchas large businesses or developments, will consume some resources in

63

DEREGULATED PRIVATE WATER SUPPLY

the form of bargaining costs. Longer-term rent-seeking costs mayarise as well. Individuals will be more likely to live in large condominialdevelopments, for instance, if such decisions hold the promise ofreducing their water bill. Residential decisions will be made ineffi-ciently, as the search for transfers from the water company will leadto too many cooperative developments and too few stand-alone houses.In these regards an unregulated privatized monopoly will again fallshort of a first-best optimum.

Further IssuesWe see three other potential problems with unregulated privatized

monopolies in the water sector: equity and distributional objectives,rent-seeking costs, and the imperfect ability of governments to pre-commit to a laissez-faire regime. We consider each problem in turn,and how privatization might be structured to overcome the relevantobjections.

EquityCommentators often find the distributional implications of perfect

price discrimination to be disagreeable. If the water company succeedsin price discriminating, it will capture all of the produced social surplusfor itself, and leave consumers with very little benefit. We do notregard this as a decisive objection to unregulated privatization for tworeasons. First, it is possible to structure privatization in such a wayas to prevent wealth transfers away from consumers. Second, waterpolicy may be an inefficient means of realizing distributionalobjectives.

If the distributional implications of price discrimination were objec-tionable, the income transfer could be reversed by giving water cus-tomers an equity stake in the water company itself. The governmentcould privatize water company assets using a Czech-style voucherplan, and send the vouchers to potential water customers. High com-pany profits would then imply high values for the shares, thus reversingthe initial transfer of income or social surplus. As long as the companycontinued to maximize profit, an efficient quantity and quality ofwater would be produced, without objectionable distributionalconsequences.6

"The firm may deviate from profit maximization if enough of its shareholders are customersas well. The customers, if they have enough voting power, may eschew direct profit maximiza-tion and instruct the company to mimic the price and quantities of a perfectly competitivefirm. Even in this (unlikely) case, however, the water monopolist will produce a first-best outcome.

64

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

The water market could even be used to redistribute income towardthe poor, if the government distributed especially high numbers ofshares to the poor (we are not necessarily recommending this policy,however). Even if a foreign company were supplying water, the govern-ment could require that company to set up a local subsidiary, and thegovernment could then purchase shares in that subsidiary for itspoor. Alternatively, the government could demand that the companydistribute such shares for free, as part of the payment for being allowedto market water in the country. The government also could chargethe foreign company an entry fee, up to the size of the expectedprofits (adjusting for risk), and rebate these funds to disadvantagedgroups. Even in the absence of rebates or voucher-style privatization,the distributional consequences of unregulated privatization areunlikely to be strongly negative, and may even be positive.

To the extent that clean, potable water brings external benefits, thecommunity will gain under price discrimination, even if the monopolistwater company extracts the full consumer surplus for each individual.Each individual would fail to reap surplus from his or her waterpurchase decision, but the community as a whole would receive theexternal benefits of the additional supply. The widespread provisionof clean water would help break the well-known cycle of disease,poverty, and poor sanitation that plagues so many parts of the world.From the community's point of view, the potential status of cleanwater as a good with positive externalities strengthens the case forunregulated natural monopoly. If water is a public good, from thecommunity's point of view it becomes less important how much con-sumer surplus is retained by buyers, and more important to increasethe absolute number of hook-ups as rapidly as possible.

Developed countries also are unlikely to experience significant dis-tributional problems with unregulated natural monopoly. Householdscurrently purchasing water from tankers are likely to face lower perunit prices once they receive a piped connection, even with pricediscrimination. A government also could offset any undesired distribu-tional consequences of its water policy by changing tax rates or byusing the numerous other policy instruments that influence the distri-bution of wealth.

Using water policy to implement distributional objectives has hadan undistinguished track record. Governments often have requiredwater companies to set price below marginal cost to achieve distribu-tional objectives. Fortunately, such practices are now almost univer-sally discredited, even though they continue in practice. Using pricingto achieve distributional objectives has caused many water utilities tobe insolvent, and has brought unfavorable long-run distributional

65

IDEREGULATED PRIVATE WATER SUPPLY

consequences as well, again as discussed above. For the same reasonsthat we reject the distributional argument for pricing below marginal Icost, we do not accept the distributional critique of unregulated priva-tized monopoly. _

Rent SeekingA regime of unregulated privatized monopoly may involve signifi- •

cant rent-seeking costs if firms can compete for that monopoly position. •In traditional rent-seeking models, the resources expended on captur-ing a monopoly position are exactly equal to the monopoly profits at Istake. If the water company would earn an expected $500 million inprofits (in present value), companies would be willing to invest up to _$500 million dollars to earn that position (Tullock 1967). The more |successfully a monopolist could price discriminate, the greater thecorresponding rent-seeking costs. Unregulated privatized monopoly •could cease to serve as a first-best optimum. We see rent-seeking •costs as a potential problem for unregulated privatized monopoly.Nonetheless the transition to privatization could be structured to keep Irent-seeking costs to a minimum. The theory of rent seeking impliesonly that the would-be monopolist will pay a sum equal to the available «rents; this sum make take the form of a transfer rather than the |consumption of real resources. Assume, for instance, that the govern-ment is selling or auctioning off existing water assets to private compa-nies. The winning company will be willing to bid a sum up to theexpected profit, adjusted for risk. So if expected profits are $500million, companies will bid some sum just short of this amount (againadjusting for risk), and transfer the funds to the government. Rentseeking takes the form of a pure cash transfer and consumes no realresources. In fact, the transferred funds could be used to satisfydistributional objectives, such as cash rebates to low-income watercustomers, as discussed above.

Rent seeking for monopoly positions will consume real resourcesonly when cash transfers are not available. We can imagine watercompanies which court the local politicians, engage in expensive adver-using campaigns, and send costly signals of their trustworthiness. Inall these cases the search for a monopoly position will lead to realresource consumption, and in fact we do observe all of these phenom-ena in the contracting process. Nonetheless, both the governmentand the water company will attempt to replace costly signals andinvestments with pure cash transfers, simply because the latter areboth cheaper and of greater value to the recipient. Rent-seeking costsalso can be limited by noncompetitive procurement practices. If onecompany stands in a favored position to win a given contract, that

66

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

company need not invest large sums of real resources to capturethe subsequent rents. When rent-seeking costs are potentially high,governments may obtain superior results by limiting entry into theprofitable activity. The winning company will still be able to serve theentire market, and other companies will be dissuaded from investingresources to capture that position. In sum, we see rent-seeking costs asa potential problem for unregulated privatization, but not necessarily adecisive problem. A comparative analysis also must consider the rent-seeking costs involved with various forms of government ownershipand regulation. These costs may be quite high, given the profits at stake.

Government PrecommitmentThe imperfect ability of governments to precommit provides per-

haps the most serious problem for the unregulated privatization ofwater. By construction of our policy proposal, water companies andcustomers are free to set whatever prices and quantities they canagree to. The analysis so far has simply assumed that governmentswould honor and enforce these contracts with credibility. In reality,governments often do a poor job of enforcing contracts. Many govern-ments are too incompetent to enforce contracts efficiently, or politicalpressures intervene and the government deliberately voids or rewritescertain contracts. Even in developed countries governmental interfer-ence into the contracting process is common. In the context of anunregulated water market, we can imagine the government rewritinga contract where buyers promise to pay high prices in return for anexpansion of capacity or additional hook-ups. Once the hook-upshave been made, political pressures might induce the government toregulate or cap prices. Knowing this in advance, the water companymight be reluctant to conclude certain kinds of contracts with potentialwater buyers. In particular, they will be reluctant to conclude contractsthat require them to sink significant amounts of capital. (The watersector typically is the most capital intensive of the infrastructuresectors.) The absence of government credibility will limit the gainsfrom trade.

To a considerable degree, imperfect government credibility simplymimics or recreates the costs of regulation. The costs of forthcomingregulation resemble the costs of having regulation now. In this regardthe initially unregulated alternative should not produce inferior perfor-mance, compared to regulation. In some cases, however, the initialabsence of regulation may create more risks for companies than ifregulation were already in place. If a water company knows that futureregulation is forthcoming in any case, the company may prefer toknow the nature and extent of regulation upfront. Transactions costs

67

DEREGULATED PRIVATE WATER SUPPLY

may be lower if regulation is present from the onset. Although weregard this problem as a serious one, we do not see regulatory riskas a decisive argument against unregulated privatization. First, aninitially unregulated system will not necessarily imply more regulatoryrisk than a system with initial regulation. Even when initial regulationis present, die water company and its customers always face the riskof additional regulation. A non-credible government cannot makepolicy risk disappear or even diminish by instituting regulations today.In fact the appearance of regulation may be a signal that more regula-tion is forthcoming in the future. Typically we expect greater credibilityfrom governments which are willing to experiment with market solu-tions, even if those governments cannot precommit in absolute terms-Today's world exhibits a significant positive correlation between agovernment's willingness to allow the private sector to operate andthe credibility of that government. Starting with a laissez-faire experi-ment may increase rather than decrease a government's credibility,as it has in Singapore, New Zealand, Chile, and other countries ina variety of (non-water) contexts. Experimenting with unregulatedprivatization thus might lower regulatory risk, rather than increase it.

The regulatory risk argument also proves too much. We could, forsimilar reasons, argue that the government should regulate everyeconomic sector immediately, to reduce the uncertainty about subse-quent regulation. Yet successful economies do not typically approachregulation in this fashion. Rather, a responsible government firstattempts to discover what a good policy might be and then implementsthat policy. It should not shy away from good policies for fear thatthe policy might later be abandoned.

Furthermore, a policy "proposal" is precisely that—a proposal aboutwhat would work, not a prediction about what will be adopted. Govern-ments might be unwilling to embrace credible commitments to favor-able policies, but policy analysts nonetheless should continue to holdsuch commitments as an ideal or aspiration (Philbrook 1953). Credibil-ity is, in part, a function of what a government, its citizens, and itsadvisors believe. By attempting to persuade and to change beliefsabout what will work, policy analysts themselves manufacture credibil-ity for policies. To argue that a policy will not have credibility is toassume what is at stake in the policy debate itself.

ConclusionThe need for water policy reform is pressing, given the stakes in

terms of economic development and human health. The lack or veryhigh cost of access by the poor to safe sources of water has devastating

68

IIIIIIIIIIIIIIIIIII

CATO JOURNAL

social and economic consequences. We have considered unregulated,privatized monopoly as a potential policy improvement. Under someconditions, this policy can approximate a first-best solution across thequantity and quality of output. While we do not expect this first-best result to hold, laissez-faire in water may nonetheless result in asignificant increase in the number of water hook-ups. Given the num-ber of individuals who have no access to clean, safe water, this factorshould weigh heavily in our evaluation of the policy. The unregulatednatural monopoly will bring problems of partial exclusion, bargainingcosts, rent-seeking costs, and imperfect government credibility, butin comparative terms we do not see a knock-down argument againstunregulated private provision in this context. Unregulated privatizationshould join the roster of plausible policy alternatives for the watersector.

ReferencesArmstrong, M.; Cowan, S.; and Vickers, J. (1994) Regulatory Reform: Eco-

nomic Analysis and British Experience. Cambridge: MIT Press.Asian Development Bank (1993) Water Utilities Data Book. Manila: Asian

Development Bank.Bittlingmayer, G. (1982) "Decreasing Average Cost and Competition: A New

Look at the Addyston Pipe Case." Journal of Law and Economics 25(October): 201-29.

Breyer, S. (1982) Regulation and Its Reform. Cambridge, Mass.: HarvardUniversity Press.

Brook Cowen, P. (1996) "The Guinea Water Lease—Five Years on: Lessonsin Private Sector Participation." World Bank Viewpoint Note No. 78 (May).

Cooper, KJ. (1997) "Battling Waterborne Ills in a Sea of 950 Million."Washington Post, 17 February: A27, A30.

Demsetz, H. (1968) "Why Regulate Utilities?" Journal of Law and Economics11 (April): 55-66.

Dickinson, H.W. (1954) Water Supply of Greater London. London: Newco-men Society.

Guislain, P., and Kerf, M. (1996) "Concessions: The Way to Privatize Infra-structure Sector Monopolies." Public Policy for the Private Sector: Infra-structure ('Special ed.) (June): 21-24.

Klein, B. and Leffler, K. (1981) "The Role of Market Forces in AssuringContractual Performance." Journal of Political Economy 89 (August):615-41.

Klein, M. (1996) "Economic Regulation of Water Companies." PolicyResearch Working Paper No. 1649, World Bank.

Loeb, M. and Magat, W.A. (1979) "A Decentralized Method for UtilityRegulation." Journal of Law and Economics 22 (October): 399-407.

Mann, P. (1996) World Bank Seminar, 17 December.Oi, W. Y. (1971) "A Disneyland Dilemma: Two-Part Tariffs for a Mickey

Mouse Monopoly." Quarterly Journal of Economics 85 (February): 77-90.

69

IDEREGULATED PRIVATE WATER SUPPLY

Philbrook, C. (1953) " 'Realism' in Policy Espousal." American EconomicReview 43 (December): 846-59. •

Phlips, L. (1983) The Economics of Price Discrimination. Cambridge: Cam- |bridge University Press.

Rivera, D. (1996) Private Sector Participation in the Water Supply and MWastewater Sector. Washington D.C.: World Bank. |

Tullock, G. (1967) "The Welfare Cost of Tariffs, Monopoly, and Theft."Western Economic Journal 5 (June): 224-32. ft

World Bank ( 1992) World Development Report 1992: Development and the |Environment. New York: Oxford University Press.

World Bank. (1994) World Development Report 1994: Infrastructure for •Development. New York: Oxford University Press. |

World Bank. (1997) Toolkits for Private Participation in Water and Sanitation.Washington D.C.: World Bank. m

Zajc, K. (1996) "Private Sector Participation Options in the Water Sector in |Transition Economies." Ph.D. dissertation, George Mason University.

70

IIIIIIIII

Faculty - Over thirty international faculty is assembled from regulatory institutions, leading• universities. The World Bank, infrastructure companies, financial institutions and other' international organizations. Faculty from recent presentations included: Sanford Berg,

IUniversity of Florida; Anthony Ballance, OFWAT (UK); Geoffrey Cannock, OSIPTEL (Peru);

Sally Hunt, NERA; Laurie Mahon, Chase Securities, Inc.; Marie Rounding, Canadian GasAssociation; Hector Olea, Comisión Reguladora de Energía (Mexico); and Warrick Smith, WorldBank. A comparable faculty will make presentations in June.

m Teaching Methods — 50 sessions, including lectures, sector-specific sessions, casestudies, practical exercises, and panel discussions with leading experts.I

• Opportunity - The program provides an exciting opportunity for learning problem-solvingtechniques and promoting international exchanges across sectors and countries.

III

SIXTH INTERNATIONAL TRAINING PROGRAM ONUTILITY REGULATION SC STRATEGY

A Collaboration Between The World Bank andThe Public Utility Research Center at the University of Florida

June 14-25,1999Gainesville, Florida USA

Purpose - To enhance the economic, technical, and policy skills required for designing andmanaging sustainable regulatory systems for infrastructure sectors. To respond to theprincipal areas of concern to utility regulators around a series of cross-sectoral and sector-specific topics.

St3te-0f-tfl&Art - An intensive two-week program specifically tailored to the professionalneeds of utility regulators and regulatory strategy staff in the private sector.

Participants - A select group of 85 senior- and mid-level utility regulators (electricity, gas,telecommunications,and water) from OECD and non-OECD countries undertaking infrastructurereforms as well as regulatory strategy executives from private infrastructure companies.Participants will share expertise and experience with one another through structured cases.Over 30 countries were represented in 1997, 1998, and in January 1999. The sessicns willbe conducted in English. Application deadline is April 5, 1999.

Tuition — US$4,400 for utility regulators and US$6,000 for executives from private orpublic infrastructure companies (includes materials, lodging, and most meals).

For a brochure, program calendar and an application form, contact:Public Utility Research Canter, University of FloridaTel:+1-352-392-6148 - Fax:+1-352-392-7796E-mail: purcecon@dale. cba.ufl.edu -Web: http://www.cba.ufl.edu/eco/purcNote: You can download the brochure and application form from our website.

UNIVERSITY OF fl?Fr^l WorldF L O R KSbä Bank71 m*^m

Topics and Case Studies - Seven broad topics will be covered during the training course:

I. Market Structure Reform and Regulation of Network Industries: Why are countries reformingtheir utility sectors? What drives the search for market solutions across infrastructure sectors?What are the constraints involved for introducing competition in network industries? What arekey economic and legal principles for ensuring their sound implementation? How should theinterface between monopoly and competition be regulated? What are the impacts < differentforms of vertical separation and service unbundling on competition and regulation?

II. Financial Analysis For Utility Regulation: What principles and practices of cost ountingcan be applied to the treatment of operating costs, capital expenditures, depreciatir nd taxesof utility companies? How can regulators determine the cost of capital and ass? projects,particularly in countries with scarce or unreliable cost information? What are the ...formationrequirements for regulators? How can regulators improve data quality and minimize informationrents?

HI. Principles and Application of Incentive Regulation: What should be the extent of regulation?What are the trade-offs between flexibility and predictability of regulatory arrangements? Whathas been the experience with alternative schemes of incentive regulation? What incentive rulespromote competition, efficiency, and innovation? What are the strengths and limitations ofalternative forms of price regulation? How does the choice of price regulation affect thesystem's overall credibility, efficiency, and legitimacy? What has been the experience withconducting price reviews under alternative incentive systems?

IV. Non*Price Aspects of Utility Regulation: What are the rationale and methods for introducingperformance standards and incentives related to quality of service, health, safety, andenvironmental factors? What are effective regulatory strategies for monitoring performance andenforcing compliance?

V. Managing the introduction of Competition in and for the Market: Where can competitiveforces be introduced or strengthened? What policies hinder competition and what policiespromote competition? When should regulators intervene in market structure? What has beenthe experience with different types of market mechanisms for unbundled utility services? Howshould regulators apply competition rules and antitrust principles?

VI. Rate Structure: What are key considerations in rate design? How do the joint and commoncosts associated with network industries affect pricing rules? How does the introduction ofcompetition affect decisions about tariff re-balancing, cross-subsidization, and funding of socialobligations? How does regulation affect providers' investment and service strategies? Howshould universal service obligations be developed and funded?

VII. Managing the Regulatory Process: What are key considerations for the establishment andfunctioning of regulatory institutions? How can the regulatory process promote legitimacy andcredibility of regulatory decisions? What strategies are at the disposal of regulators toeffectively manage complex and often politically sensitive negotiations involving government,investors, consumers, and other interest groups? What has been the experience with alternativemechanisms for consensus building and dispute resolution? What strategies can regulators-useto effectively communicate with the public? How can regulators become more efficient andeffective in accomplishing their tasks in different institutional settings?

PLEASE DISTRIBUTE THIS INFORMATION TO YOUR AGENCY

72