The Economics of Public and Private Roles in Health Care:Insights from Institutional Economics and Organizational Theory

Alexander S. PrekerLead EconomistThe World Bank

April HardingPrivate Sector Economist

The World Bank

THE ECONOMICS OF PUBLIC AND PRIVATE ROLES IN HEALTH CARE:INSIGHTS FROM INSTITUTIONAL ECONOMICS AND ORGANIZATIONAL THEORY

1

A. A HISTORICAL SNAPSHOT

Advances in health during the past few decadesare impressive. The increase in life expectancy and thedecrease in fertility throughout the world have beengreater in the past 40 years than during the previous4,000 years (Figure 1). Life expectancy is almost 25years longer today than at similar income levels in 1900.

Unparalleled Improvements inLife Expectancy and Fertility Rates

Figure 1

30

35

40

45

50

55

60

65

70

75

80

1950 60 70 80 90 2000 10 20 30 40 20502.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

Fert

ility

Rat

es

Life

Exp

ecta

ncy

in Y

ears

Life Expectancy

Total Fertility Rates

These gains in health are partly the result ofimprovements in income and education, withaccompanying improvements in nutrition, access tocontraceptives, hygiene, housing, water supplies, andsanitation. Asis described by the World HealthOrganization (WHO) in its 2000 World Health Report,the achievements in health during the twentieth centuryare also a result of new knowledge about the causes,prevention, and treatment of disease, and policies thatmake known interventions more accessible.

International experience indicates that theunderlying cause of most threats to good health are wellknown today, and affordable drugs, surgeries and otherinterventions are often available, even in low-incomecountries. But, because of weakness in one or more ofthree core functions of health systems—financing,generation of inputs, and provision of services—potentially effective policies and programs often fail toreach the poor.

The financing function includes the collectionand pooling of revenues and the use of these revenues,through purchasing or budget transfers to serviceproviders. The resource-generation function includes theproduction, import, export, distribution, and retail ofhuman resources, knowledge, pharmaceuticals, medical

equipment, other consumables, and capital. The service-delivery function includes both population-base andpersonal clinical services provided by the public sectorand private sector (nonprofit and for-profit)

These core functions are influenced bygovernments through their stewardship function and bythe population through demand and markets. Thecombined effects of these five factors lead to either goodor poor performance in health outcomes, financialprotection, and responsiveness to consumer expectations(Figure 2).1

Health systems are dysfunctional sometimesbecause of uneven development among the corefunctions and sometimes because of poor coordinationand complimentarity between the public and privatesector.

Core Functional Components andPerformance Measures

The State

People

Demand and Markets

Stewardship

Serv

ice

Del

iver

y O

rgan

izat

ion

& M

anag

emen

t

Prod

uctio

n Pr

oces

s

Financing Collection Pooling Purchasing

Other Inputs Knowledge People Pharmaceuticals Equipment Consumables Capital

Health Outcome Level and Distribution

Financial Protection Level and Distribution

Consumer Quality Level and Distribution

Figure 2

We argue in favor of greater private sectorparticipation in generating inputs and providing healthservices and a strong government engagement insecuring equitable and sustainable financing as well asexecuting the “stewardship function.” In too manycountries, these roles are reversed, with adverse effectson equity and efficiency.

From Centuries of Minimalism …

Ideological views on the roles of the state andthe private sector belong to a long list of false antithesesin the field of medicine and health care.2 Since thebeginning of written history, the pendulum has swungback and forth between minimalist and heavy-handedstate involvement in the health sector.

2

During antiquity, people used home remediesand private healers when they were ill. Yet, as early asthe second millennium B.C., the papyri give fascinatingevidence that Imhotep, archetypal physician, priest, andcourt official in ancient Egypt, introduced a system ofpublicly provided health care with healers who werepaid by the community.3

This early experiment in organized health caredid not survive the test of time. The Code ofHammurabi (1792–50 B.C.) laid down a system ofdirect fee-for-service payment, based on the nature ofservices rendered and the patient’s ability to pay.4 Forthe next three thousand years, the state’s involvement inhealth care revolved mainly around enforcing the rulesof compensation for personal injury and protection ofthe self-governing medical guild.5

At best, financing, organization, and provisionof health care was limited to the royal courts of kings,emperors, and other nobility who might have a physicianfor their personal use and for their troops at the time ofbattle. The masses got by with local healers, midwives,natural remedies, apothecaries, and quacks.

… To Heavy-Handed State Involvement

Unlike this early private participation in healthcare, during the twentieth century, governments of mostcountries have become central to health policy, oftenboth financing and delivering a wide range of care.Today, most industrial countries have achieveduniversal access to health care through a mix of publicand private financing arrangements and providers.6

Proponents of such public sector involvement inhealth care have argued their case on both philosophicaland technical grounds. In most societies, care for thesick and disabled is considered an expression ofhumanitarian and philosophical aspirations.

But one does not have to resort to moralprinciples or arguments about the welfare state towarrant collective intervention in health. The pastcentury is rich in examples of how the privates sectorand market forces alone failed to secure efficiency andequity in the health sector.

Economic theory provides ample justificationfor such an engagement on both theoretical and practicalgrounds to secure:

• efficiency—since significant market failure exists inthe health sector (information asymmetries; publicgoods; positive and negative externalities; distortingor monopolistic market power of many providersand producers; absence of functioning markets insome areas; and frequent occurrence of hightransaction costs)7

• equity—since individuals and families often fail toprotect themselves adequately against the risks ofillness and disability on a voluntary basis due toshort-sightedness (free-riding) and characteristicshortcomings of private health insurance (moralhazard and adverse selection).8

Largely inspired by western welfare stateexperiences such as the British National Health Service(NHS) and the problems of market failure, during thepast 50 years, many low- and middle-income countriesestablished state-funded health care systems withservices produced by a vertically integrated publicbureaucracy.

… Back to Neoliberalism of the 1990s

During the 1980s and 1990s, the pendulumbegan to swing back in the opposite direction. Duringthe Reagan and Thatcher Era,9 the world witnessed agrowing willingness to experiment with marketapproaches in the social sectors (health, education, andsocial protection). This was true even in countries suchas Great Britain, New Zealand, and Australia—historical bastions of the welfare state.

As in the ascendancy of state involvement, therecent cooling toward state involvement in health careand enthusiasm for private solutions has been motivatedby both ideological and technical arguments.

The political imperative that has accompaniedliberalization in many former socialist states and theeconomic shocks in East Asia and Latin Americacertainly contributed to a global sense of urgency toreform inefficient and bloated bureaucracies and toestablish smaller governments with greateraccountability.10

Yet, it would be too easy to blame ideology andeconomic crisis for the recent surge in attempts toreform health care systems by exposing public services

3

to competitive market forces, downsizing the publicsector, and increasing private sector participation.11

In reality, the welfare-state approach failed toaddress many of the health needs of populations acrossthe world.12 Hence, the dilemma of policymakersworldwide: although state involvement in the healthsector is clearly needed, it is typically beset by publicsector production failure.13

Toward a New Stewardship Role of the State …

Today, governments everywhere are reassessingwhen, where, how, and how much to intervene orwhether to leave things to the market forces of patients’demand.

The growing consensus is that to address thisproblem requires a better match between the roles of thestate and the private sector, and their respectivecapabilities—getting the fundamentals rights. In mostcountries, this means rebalancing an already complexmix of public and private roles in the health sector.14

To improve efficiency or equity, governmentscan choose from an extensive range of actions—fromleast to most intrusive. These include:

• providing information to encourage behavioralchanges needed to improve health outcomes

• developing and enforcing policies andregulations to influence public and privatesector activities

• issuing mandates or purchasing services frompublic and private providers

• providing subsidies to pay for services directlyor indirectly

• producing (in-house) preventive and curativeservices.In many countries, for reasons of both

ideological views and weak public capacity to deal withinformation asymmetry, contracting, and regulatoryproblems, governments often try to do too much—especially in terms of in-house service delivery—withtoo few resources and little capability.

Parallel to such public production, the samewell-intending governments often fail to:

• develop effective policies and make availableinformation about personal hygiene, healthy

lifestyles, and appropriate use of health careregulate and contract with available privatesector providers

• ensure that adequate financing arrangements areavailable for the whole population

• secure access to public goods with largeexternalities for the whole population.

The next section will present a discussion of themost significant sources of government productionfailure to which market-based solutions are beingapplied and the market imperfections that must beaddressed to optimize complementarity between the twosectors.

B. THE NATURE OF GOVERNMENT FAILURE

Many attempts have been made in recent yearsto reinvigorate the public sector through “best-practice”management techniques. Borrowed from the privatesector and organizational reforms, these tools attempt toreplicate the private incentive environment. 15

These reforms have included efforts tostrengthen the managerial expertise of health sectorstaff, both through training and recruitment policies.Frequently, attempts are also made to use businessprocess reengineering, patient-focused care, and quality-improvement techniques. Such efforts have alsoincluded setting up clinical directorates, introducingimproved information systems to facilitate effectivedecision making, and performance benchmarking.16

Why has the public sector been so impervious tothese types of management and organizationalreforms?17 A review of theories regarding governments’performance of their multiple functions is needed toshed light on the profound nature of the structuralproblems involved. This review complements the well-developed theories of market failure provided by healtheconomists.18 Below we explore the problems of poorpublic accountability, information asymmetry, abuse ofmonopoly power, failure to provide public goods, andloss in strategic policy formulation that have parallels inmarket failure.

Problems Relating to Public Accountability

The first set of problems relates to the difficulttask of translating individuals’ preferences into publicpolity and getting that policy implemented. As we know,

4

all public interventions involve transfers of benefits tosome people and costs to others, leaving both winnersand losers.19 Accountability means that governmentaction accords with the will of the people it represents.Yet, since people’s values are never perfectlyhomogeneous in any society, accountability will alwaysbe based on some imperfect rule about the aggregationof individual values or respect for minority interests.20

This raises several intractable procedural issuesrelating to accountability in the electoral process,taxation policies, content of public spending programs,and vested bureaucratic interests.

Ballots are blunt instruments that cannot capturethe full range of issues that may be bundled together atelection time. The intensity of views on any one issuecannot be reflected. And, election promises are often notkept. So at best, public spending policies are animprecise reflection of social values.

Majority rule in itself can be a form of tyrannyif applied strictly without constraints. Most democraticsocieties safeguard minority interests to some extent, buteven with good intentions there are limits to bothpracticality and desirability in this respect.

Finally, public servants may have strongconflicts between their own interests and their assignedresponsibilities to execute the collective will of thesociety they represent. And, their political overseersmay also have strong vested interests that are differentfrom those of the society that they represent.21

In an ideal setting, good public accountabilitywould be secured through a large intersection(authorizing environment) between fairly homogeneoussocial values, a political agenda that reflects suchvalues, and vested bureaucratic interests (Figure 3, leftbox).22 For example, there may be general socialagreement that the population has to be protected againstthe financial consequences of illness through some sortof health insurance system. When the policies of thepolitical party in power are consistent with such valuesand bureaucrats have the capacity to implement them,the intersection will be large.

Ideal Real

Social Values

PoliticalAgenda

BureaucraticInterests

The Authorizing Environment Neededfor Good Public Sector Accountability

Social Values

PoliticalAgenda

BureaucraticInterests

Figure 3

In the health sector, a further tension existsbetween the authorizing environment needed for goodpublic accountability and most individuals’ desire forsome sovereignty over their own health care. This leadsto difficult dilemmas during the rationing of care ordesign of compulsory programs, based on theapplication of some sort of majority rule that infringeson either perceived minority group rights or individualsurvival.23 Normal rational individuals who may sharesocial values about the need to ration health careresources often lose their commitment to such valueswhen confronted by resource constraints in the face ofserious illness or death.

Information Asymmetry in the Public Sector

Information asymmetry can occur in three majorways in the health sector—between patient (public) andhealth care professional; between patient andadministrator or health care system; and between healthcare professional and administrator or health caresystem.

Patients know more about their symptoms thandoctors but may—unwittingly or deliberately—notarticulate this clearly. Doctors know more about thecauses, prognosis, and effectiveness of availabletreatments but may not communicate this clearly to thepatient. Or, the patient may not understand theimplications of what he/she is being told. Forpaternalistic reasons, the doctor may deliberatelyconceal information, based on a judgement that thepatient will not be able to cope with the full knowledgeof his/or her ailment (e.g., terminal cancer). Throughself-selection, these problems are typically worse in the

5

public sector, which has to deal with large volumes ofpoorer and less educated patients.

When it comes to interactions between patients(public) and health care administrators, patients may tryto avoid being excluded or paying higher health carepremiums by choosing to conceal pre-existingconditions. Similarly, there is often have a lack ofinformation and understanding of available publicprograms. The benefit of gathering useful data aboutsuch programs is usually undervalued compared with itscost. Some of these programs may be too complex tofully understand even if information were available.Frequently, the lack of transparency in the rationing ofscarce resources is deliberate.

Finally, some serious information asymmetriesexist between health care providers and theadministrators for whom they work (or betweenadministrators and owners). Health care providers—asadvocates for patients—have a much betterunderstanding of legitimate needs or demands.Administrators have a much better understanding aboutsupply and cost of available resources but know littleabout a selected intervention’s appropriateness oreffectiveness. The doctor’s well-know informationadvantage over the patient is not solved by the existenceof a public employer or administrator who knows evenless about interactions between patients and providers.

Associated Higher Transaction Costs

Such information asymmetries add to agencycosts in terms of structuring, monitoring, and bondingcontracts among agents and principals with conflictinginterests.24 Private firms, concerned about profits, have astrong incentive to limit agency costs related toinformation asymmetry.25 Public agencies that are notheld to a clear “bottom line” due to unspecified socialfunctions and many complex sources of subsidies do notreceive clear signals about the agency cost of suchinformation asymmetry.

Not surprisingly, many public health carefacilities in low- and middle-income countries do notkeep detailed patient records, do not know through-putstatistics, and do not know the unit costs of theprocedures they are using or illnesses they treat.Hospitals are often not doing what policymakers oradministrators would like, or what they think theyshould be doing. And, doctors working in such public

hospitals maintain an information advantage that givesthem a great deal of latitude to pursue their owninterests. Paradoxically, since public systems often givetheir beneficiaries less leverage than private systemsgive their clients, patients may be less able to use theinformation they do possess to influence their owntreatment.

Potential for Corruption

Worse than this “ignorance by default,” due toinformation asymmetry, are the information problemsdeliberately engineered by politicians, bureaucrats,organizations, and health care providers entrusted withpublic accountability.

Although deliberate deception and fiscal fraud isusually sanctioned severely, it is much harder to hold thepublic sector accountable for petty abuses, avoidance,and obfuscation. Such deception may take the form ofhidden costs, subsidies, cost shifting, or inflation. Or itmay occur as an amplification of benefits, exaggerationof the consequences of alternatives, or claiming creditfor activities that originate elsewhere.

Given the complex nature of health care, it isnot hard to imagine the considerable scope formisleading or defrauding patients and the public in thehealth sector.

Abuses of Public Monopoly Power

Monopoly power occurs for four reasons whenthe public sector gets involved in producing healthservices. This may be due to: (a) legal restrictions oncompetition; (b) access to subsidized capital andrevenues, creating an uneven “playing field”; (c) below-cost distribution of goods and services to achieve equitygoals; or (d) production of public goods or goods wheremarkets are not viable.

When—in addition to the above accountabilityand information problems—the public sector enjoysmonopoly power, people who work for it are given widescope for abusing this power through the extraction ofrents, internal distribution of “slack” to employees, andlowering of quality.

Public monopolies exhibit the usual negativefeatures. First, monopoly suppliers often reduce outputand quality, while raising prices. The excess in pricesover and above what the market would normally bear—

6

rents—leads to allocative inefficiency or a netdeadweight welfare loss to consumers who have to forgothe consumption of other goods.

A manifestation of such rents is the informaluser charges that are commonly levied on patients andtheir families in public health facilities. In manycountries, these rents are not limited to doctors’accepting bribes or peddling influence (allowingprivileged patients to circumvent the usual rules onresource allocation, to receive preferential treatment,and to cut waiting time). It also includes charges leviedby other salaried workers for items ranging from toiletpaper and clean linen to food, drugs, and medicalsupplies.

In a recent study on corruption around theworld, such abuses in publicly run health servicesranked number one in terms of the burden placed onhouseholds.26 Patients who can afford to pay formalcharges in the private sector often prefer it to payingsuch rents in the public sector. Taxpayers charged twicefor low-quality services and such abuses have littlerecourse but the blunt and often ineffective instrumentof voting power.

Second, monopoly suppliers have strongincentives to lower expenditures through decreasedoutput when staff members benefit from the financialresiduals. Although public organizations cannot legallydistribute such residuals outside the organization toshareholders, they can be internally consumed in severalways. First, executives often receive generous socialbenefits and travel allowances (perks). Second, time-keeping is often not enforced rigorously (doctors oftenwork short hours in public institutions). Third, some ofthe residual may be used to pursue personal agendas(discretionary spending on special projects andresearch).Failure of Critical Policy Formulation

The most frequently cited reason for greatergovernment involvement in the health sector is that,when left to competitive forces and prices alone, inseveral other critical areas the market does not lead towelfare-enhancing production and allocation of anumber of health care goods and services. Theseinclude:

• public goods (policymaking and information)• goods with large externalities (disease prevention)

• goods with intractable market failure (insurance).

It is therefore surprising that governments in thehealth sector often neglect the exact same three areaswhile they are busy producing curative services that theprivate sector could easily provide. Furthermore, whenmost public funds are spent on poorly targeted publicproduction, few or no resources are left for strategicpurchasing of services for the poor fromnongovernmental providers.

The next section presents a discussion of someof the key theoretical underpinnings for a new approachto optimize complementarity between the public andprivate sector.

C. THEORETICAL UNDERPINNINGS

The current trend worldwide is to use threetypes of approaches to address the public sector failuresin service delivery described above in descending orderof importance. They include increased:27

• exit possibilities (market consumer choice)• voice (client participation)• loyalty (hierarchical sense of responsibility).

When possible, one would always use exit,unless forced to use the weaker variants because thegoods and services involved are not “marketable.”

We will focus mainly on the first option—theexit option. This approach relies on greater privatesector participation, allowing clients and patients achoice or alternative to publicly provided services. Suchexit options can be implemented in parallel with otherpublic sector management reforms that increase voiceand loyalty.From Neoclassical Economics …

One of the central tenets of neoclassicaleconomics is that, in an optimally functioning market,competitive forces will lead to a more efficientallocation of resources—Pareto-optimal competitiveequilibrium—than nonmarket solutions.

According to the neoclassical model, when thereare many firms and consumers—and prices are allowedto respond to the forces of supply and demand—competition will result in an equilibrium situation whereit is impossible to make someone better off without

7

making someone else worse off. This will result in awelfare-maximizing situation.

The perfectly competitive Walrasian model, asit is sometimes called, requires a number of assumptionsto be met. These include:

• The goods involved behave like private goods (i.e.,rivalry, excludability, and rejectability (Box 1).

• Rights can be perfectly delineated.• Transaction costs are zero.

According to neoclassical theory, a breakdownoccurs in both efficiency and equity when public goodsor services with significant externalities are allocatedthrough competitive markets. Likewise, as describedearlier, significant problems occur in efficiency andequity when private goods are produced or provided bya public sector monopoly.

Unfortunately, most health care goods andservices do not behave like perfect private or publicgoods. Many are not perfectly excludable but areassociated with complex externalities. Rights are oftendifficult to delineate, leaving residual claimants. Andtransaction costs are often high. Therefore, though in aPareto-optimal state, it would be impossible to makesomeone better off without making someone else worseoff, few situations meet such criteria in the health sector.

On one hand, although many public healthactivities (e.g., sanitation services, control andprevention of communicable diseases, and healthpromotion) generate significant externalities, they arenot pure public goods. All have some element ofexcludability, rejectability, and rivalry. For example, avaccine given to one patient cannot simultaneously beconsumed by other patients. At will, patients can choosenot to be vaccinated. And, vaccination programs can, inprinciple, compete with each other for market share.

Box 1. Private, Mixed, and Public Goods

The neoclassical model classifies goods and service asprivate, mixed, or public. Private goods exhibit excludability(consumption by one individual prevents consumption by another—no positive or negative externalities); rivalry (there is competitionamong goods based on price); and rejectability (individuals canchoose to forgo consumption). True public goods have significantelements of nonexcludability, nonrivalry, and nonrejectability.Mixed goods have some but not all of the characteristics of privategoods (see Figure 4).

The Nature of GoodsBased on Neo-Classical Economics

Nature of Economic Good

Properties Public Mixed Private

Excludability -- + +

Rivalry -- + +

Rejectability -- + +

Consumption Goods

Medical ClinicsHospitalsMedical SuppliersPharmaceuticals

Consumer Protection

PolicymakingRegulationsSetting StandardsQuality ControlFigure 4

Likewise, even expensive diagnostic andtherapeutic care—though often provided in publiclyowned inpatient facilities at highly subsidized rates—isreally private goods and hence marketable. The same istrue for ambulatory and community-based care. Evenwhen governments try to fully control the market forsuch services, preventing their sale in the informaleconomy is often difficult.

Therefore, although neoclassical theory is ofteninvoked by mainstream economists to justify public andprivate roles in the health sector, consumptioncharacteristics alone almost never indicate anythingabout the specific production processes needed to securetechnical efficiency and equity. Neoclassical theorycontributes little to the understanding of optimalorganizational arrangements for service production. It isessentially “institution free.”28 Other theories are neededto fill this vacuum in understanding productioncharacteristics.…to the Economics of Organizations

Recently, much progress has been made inidentifying the key factors causing wide variations inorganizations’ performance. The developments mostrelevant for understanding the advantages anddisadvantages of different arrangements for servicedelivery come from principal-agent theory, transactioncost economics, property rights, and public choicetheory. These fields are often grouped together under thetitle “institutional economics.” Institutional economicsdirectly addresses the issue of how best to structureorganizations that consist of individuals pursuingmultiple and often conflicting interests.29

8

Agency Theory

This framework highlights that social andpolitical objectives may be more readily achievedthrough a series of explicit and transparent “contracts”for labor/services between an “agent” that undertakes toperform various tasks in an acceptable way on behalf ofa “principal” in exchange for a mutually agreed award.Usually, the principal needs the agent’s efforts andexpertise but has only limited ability to monitor theagent’s actions or evaluate whether the final outcome issatisfactory.

The agency literature surveys the range ofcontracts (e.g., payment and monitoring arrangements)observed in the economy as attempts to align incentivesand reward cooperation between self-interested butinterdependent individuals.30 Several studies havegeneralized the agency insight from the employmentcontext to the full range of relationships that make upthe firm—now conceptualized as a nexus of manycontracts.31 The need for incentive alignment ispervasive in the health sector: relations between patientand physician or governments and contracting agenciesare classical examples of principal-agent structure.

Transaction Cost Economics

Transaction cost economics emphasizes thelimitations of contracts and the need for flexible meansof coordinating activities. Principals and agents are bothopportunistic. Agents will seek to minimize theaggregate production and transaction costs andmaximize the benefits (unless closely monitored, agentsmight be unreliable, engaging in behavior such as rentseeking, cheating, breach of contract, incompletedisclosure). Principals will try to maximize their benefitto the extent that the relationship could become unviablefor the agent.

The extent of such opportunism variesdrastically from country to country and from onecultural setting to another. In some settings, such asmonopolistic national health services, opportunism maybe less apparent than in other settings where providersare more accustomed to competing with each other.Although opportunism may appear to be greater incountries such as Chile, India, and the United States,there is good evidence that principal-agent relationshipswithin national health services such as Costa Rica, New

Zealand, Scandinavia, Sri Lanka, and the UnitedKingdom are also vulnerable to opportunistic behavior.

This theory sheds most light on firm boundariesand the conditions under which activities are bestarranged within a hierarchy instead of through marketinteractions with suppliers or other contractors. Moregenerally, vertically integrated organizations, simple“spot” contracts, franchises, or joint ventures areinterpreted as discrete structural alternatives—eachoffering different advantages and disadvantages foreffective governance.32 Governance arrangements areevaluated by comparing the patterns of costs generatedfor planning, adapting, and monitoring production andexchange.33

Unlike public organizations, private firms havethe flexibility (indeed the requirement) to adjust theirgovernance structure to changes in the marketenvironment—making them fruitful sources of “betterpractices” for governance arrangements. Public agenciesthat have tried to adjust public organizations to changesin market environment (e.g., formation of NHS Trusts inthe United Kingdom, establishment of the HospitalAuthority in Hong Kong, and corporatization of publiclyowned hospitals in New Zealand) have often run intoproblems with the underlying structure of incentives andits sustainability. Major policy reversals occurredrecently in both the United Kingdom and New Zealand,adding weight to the argument of some critics of theoriginal reforms that they would have been better off toprivatize instead of settling for the imperfect middleground of public sector corporatization.

For example, vertically integrated (within firm)organizations arise as a response to problems withmarket contracting. The firm substitutes low-poweredincentives, like salaried employment, for the markets’high-powered incentives of profit and loss. Verticalintegration permits the details of future relationsbetween suppliers (including employees), producers,and distributors to remain unspecified; differences canbe adjudicated as events unfold. Vertical integration (orunified ownership) pools the risks and rewards of theorganization’s activities and can facilitate the sharing ofinformation, the pursuit of innovation, and a culture ofcooperation.

Box 2: Influence Activities

An important issue related to moral hazard and the structure oforganizations is influence activities and the associated costs, known

9

as influence costs.34 analysis has shed much light on the propensity ofpublicly owned service delivery organizations to capture inordinateportions of the sector budget as well as on their ability to influencesector policy to their benefit—often at the expense of the publicinterest.

In the health sector, provider organizations expend effort toaffect decisions regarding the distribution of resources or otherbenefits among providers to their benefit. These influence activitiesoccur in all organizations, but countervailing forces are particularlyweak in public service delivery structures—and influence costs areone of the most important costs of centralized control. Evidence ofinfluence activities is seen in public utilities where monopolies areoften maintained to protect low productivity of state-ownedenterprises from competition from more efficient producers. In thehealth sector, the tendency to allocate resources to tertiary andcurative care at the expense of primary, preventative, and publichealth is evidence of similar capture.

The cost of these activities includes both the lossesassociated with poor resource-allocation decisions as well as the lossassociated with efforts to capture rents. These costs can be reducedwhen no decisionmaker has the authority to make decisions thatservice providers can easily influence. This condition can sometimesbe brought about by creating legal or other boundaries between thepolicymaker, the funder, and the service provider unit. Manyorganizational reforms have attempted to diminish these activities.Examples include reforms separating the policymaker from the payerfrom the provider in public service delivery as well as privatization ofutilities.

Despite these positive features, verticalintegration suffers from characteristic weaknesses as amechanism of governance. The two most prominent arethe weakening of incentives for productivity and theproliferation of influence activities (Box 2). The weakincentives come as individuals capture less and less ofthe gains of their own efforts as rewards and their lossesare spread throughout the organization. Despite its focuson the contracting problems that motivate internalorganization, transaction cost economics views verticalintegration as the governance mechanism of last resort.Even in the many instances where policy objectivesimply that spot market transactions are not desirable,contractual networks, virtual integration, franchising, orconcessions will outperform unified ownershiparrangements.

Box 3: High-Powered Incentives of Ownership

Suppose a transaction involves several people supplying labor,physical inputs, and so on. If all but one party have contractedto receive fixed amounts, there is only one residual claimant.In that case, maximizing the value received by the residualclaimant is the same as maximizing the total value received byall parties. If the residual claimant also has residual control,just by pursuing his own interests and maximizing his own

returns the claimant will be led to make efficient decisions.The combination of residual control and residual claimsprovides strong incentives and capacity for an owner tomaintain and increase an asset’s value. Firms often attempt toreproduce these high-powered incentives by allocating residualclaims in the form of bonuses or shares to key decisionmakersin their firm.

Misalignment of residual rights and returns causesserious problems. The residual claimant to the returns from astate-owned enterprise is the public purse, but the residualdecisionmakers are effectively the enterprise manager, theworkers, and the bureaucrats in the supervising ministry. Noneof these has any great personal stake in the value of theenterprise. The resulting low productivity is well documented.Another example of misalignment comes from the U.S.Savings and Loan (S&L) industry. Those who had the right tocontrol the S&L’s investment also had the right to keep anyprofits earned but were not obligated to make good on losses.That combination of rights and obligations created anincentive for risk taking and fraud that was not effectivelycountered by other devices during most of the 1980s.35

These fields of analysis have led to betterunderstanding of the institutional sources of governmentfailure. The framework has been used to design organizationalreforms that seek to allocate to the holders of criticalinformation the authority to make relevant decisions and thefinancial incentive to do so (in the form of residual claims onthe outcome of the decision).

Property Rights Theory

Property rights theory looks at the sameincentive issues from a slightly different perspective.Since private ownership appears to have strong positiveincentives for efficiency, property rights theorists haveattempted to find out why (Box 3). Explanations havefocused on two issues: the possession of residualdecision rights and the allocation of residual returns.36

Residual rights of control are the rights to make anydecisions regarding an asset’s use not explicitlycontracted by law or assigned to another by contract.The owner of an asset usually holds these rightsalthough they may be allocated to others.37

The notion of ownership as residual control isrelatively clear for a simple asset like a car. It gets muchmore complicated when applied to an organization suchas a firm. Large organizations bundle together manyassets, and who has which decision rights may be

10

ambiguous. In addition to residual decision rights, anowner holds the rights to residual revenue flows fromhis assets. That is, the owner has the right to whateverrevenue remains after all funds have been collected, andall debts, expenses, and other contractual obligationshave been paid out.

Political Choice Theory

This field focuses on the self-interested behaviorof politicians, interest groups, and bureaucrats, andstudies its implications for effective government and thesize of government. Individuals are viewed as rationalutility maximizers. Bureaucrats, attempting to maximizetheir budgets, will acquire an increasing share ofnational income. As a result, the state will grow muchbigger than necessary to deliver core functions.Powerful interest groups will capture increasing portionsof resources. Institutional rigidities develop, reducingeconomic growth.38 This analysis has led public choicetheorists to support conservative political agendas(minimizing the role of the state).

Production Characteristics of Goods and Services …

The principles of institutional economics lead toa much more refined and useful understanding of thedifferent kinds of institutional arrangements required forefficient and effective production of goods and services.A model along these lines can be developed, based ontwo different and often ignored goods characteristics—contestability and measurability.39

A market can be said to be perfectly contestableif firms can enter it freely (without any resistance fromexisting firms) and then exit without losing any of theirinvestments, while having equal access to technology(no asset specificity).40 Contestability allowscompetition for the market to substitute for competitionin the market.

Contestable goods are characterized by lowbarriers to entry and exit from the market, whereas non-contestable goods have high barriers such as sunk cost,monopoly market power, geographic advantages, andasset specificity. Investments in specific assets representa “sunk cost” since its value cannot be recoveredelsewhere.41 Two specific assets that are especiallyrelevant in the health sector are expertise and reputation.Once incumbents have invested in activities that resultin expertise or generate trust, they enjoy a significant

barrier to entry for other potential suppliers, therebylowering the degree of contestability. Opportunism, onthe other hand, will lower such trust or barriers to entry.The degree of such opportunism will vary from onecountry to another and in different cultural settings.

Measurability in the health sector, as in othersectors, is the precision with which inputs, processes,outputs, and outcomes of a good or service can bemeasured. By definition, it is difficult to measure withprecision the output and outcome of health servicescharacterized by a high degree of informationasymmetry. Information asymmetry is the extent towhich information about the performance of an activityis available to users, beneficiaries, and contractingpurchasing agencies (see discussion above under “TheNature of Government Failure”).

These theoretical underpinnings are used in thefollowing section to understand some recent reforms inthe public sector.

D. TRENDS IN PUBLIC SECTOR REFORMS

Rarely does technical or systemic analysis aloneinfluence the extent of public and private involvement inthe health sector. In actuality, service deliveryarrangements are the product of complex economic,institutional, and political factors.

Extensive reforms of public sector organizationsand state-owned enterprises, implemented over the past15 years, addressed the same problems encountered indelivering public health services. In the realization thatthese organizational problems were structural in nature,using the analytical tools of organizational economics,these reforms have focused on altering the institutionalarrangements for service. These developments shed lighton similar problems in delivering public health services.

One way to understand organizational reformsin service delivery is to view the different incentiveenvironments within which government’s tasks can beperformed (Figure 5).42 The civil or core public servicelies at the center (usually constitutional control bodies,line ministries), and the activities of the staff are highlydetermined. Job tenure is also strong. Budgetary units(government departments), autonomous units,corporatized units, and privatized units are four commonorganizational modalities that straddle these incentiveenvironments in the health sector.43

11

Markets\PrivateSector

Broader Public Sector

Core Public Sector

Incentive Environments

Figure 5

The broader public sector is distinguished by therelative flexibility of the financial management regimeand by managerial freedom in recruitment andpromotion. This sector may include special purposeagencies, autonomous agencies and, on the outer limits,state-owned enterprises. Beyond the public sector liesthe domain of the market and civil society. Services maybe delivered by for-profit, nonprofit, or communityorganizations. The incentives for efficient productionare higher moving toward the periphery, where servicedelivery is often better.

Many reforms throughout the world have soughtto move delivery away from the center of the circle tomore arm’s-length contracts with public and privateorganizations. However, the nature of the outputs andthe existence of mechanisms for public sectormanagement of their delivery constrains movingdelivery outward.

Increased autonomy or corporatization—movingfrom the center of the circle to the outer limits—requiresaccountability mechanisms not tied to direct control.These controls (e.g., contracts) take considerablecapacity to write and enforce, especially for healthservices where outputs and outcomes are difficult tospecify.

How far countries may go in pushing activitiesto incentive environments in the outer circles dependson the nature of the outputs (the services involved) andtheir capacity to create accountability for publicobjectives through indirect mechanisms such asregulation and contracting.

E. APPLICATION TO HEALTH SECTOR

Looking at the health sector across the world, allhealth care goods and services can be categorized on acontinuum from high-contestability and high-measurability services to low-contestability and low-measurability services, and significant informationasymmetry.

Factor markets and product markets will bediscussed separately, since each has uniquecharacteristics. Although the following discussionrefers mainly to curative and public health services, theanalysis could be extended to some of the broaderintersectoral determinants of good health such as water,sanitation, education, healthy lifestyle policies, and goodnutrition.

Production Characteristics of Factor Markets

For the production of inputs, this contestabilityand measurability matrix would look as follows (Figure6).

Type I Type III

• Production of Consumables• Retail of

• Drugs & Equipment• Other Consumables

• Unskilled Labor

• Production• Pharmaceuticals • High Technology

• Large Capital Stock

•

Type IV Type VI

Type VII Type IX

• Research• Knowledge

• Higher Education• High Skilled Labor

••

Production Characteristics of Inputs(Factor Markets)

High Contestibility Low ContestibilityMedium Contestibility

Type II

• Production of Equipment• Wholesale

• Drugs & Equipment• Other Consumables

• Small Capital Stock

• Basic Training• Skilled Labor

Type V

Type VIII

•

Hig

hM

easu

rabi

lity

Med

ium

Mea

sura

bilit

yLo

wM

easu

rabi

lity

Figure 6

The production of consumable items and theretail of drugs, medical supplies, and other consumableswould be the best example of highly contestable goodswhere outputs are also easy to measure (Type I). Manycompanies usually jostle for a share of the market, andbarriers to entry are few (the initial investment capital ismodest and there are few requirements for specializedlicensing or skills). Unskilled labor also belongs in thiscategory.

As we move across the first row, a number offactors begin to contribute to raising the barriers to

12

entry, thereby reducing the contestability of the goods orservices in question (Type I). Investment cost (sunkcost) and increasing technical specifications createmoderate barriers to entry in the manufacture ofspecialized equipment and supplies. Wholesale trade indrugs, medical supplies, and medical equipment hassome entry barriers because of the larger investmentrequirements and more limited supply and distributionchains. The specialization and licensing of pharmacistscontribute to these entry barriers. In the case of smallcapital stock (e.g., clinics and diagnostic centers), entrybarriers are created mainly though certification andlicensing.

Finally, as we move across to Type III activities,entry barriers are much higher, as in the manufacture ofpharmaceuticals and high-technology medicalequipment, due to large up-front investment costs thatcannot be recovered later during sale of the assets (sunk-costs). These production activities are also associatedwith costly and long lead-time for research,development, and registration of new products. Otherbarriers to entry under this category include productdifferentiation (specialized medical equipment) andcopyright protection (brand-name drugs). Furthermore,because of the benefits conferred through economies ofscope and scale, over time a significant globalconcentration of pharmaceutical and high-technologyindustries has occurred, giving them considerablemonopoly power.

For all the activities in the first row (Type Ithrough III), measurability of outputs remains high.There is little information asymmetry.

As we move to the second row, measurement ofthe outputs and outcomes become more problematic.Outputs and outcomes can be measured, but it is moredifficult than in the case of activities in the first row.

Various barriers to entry reduce contestability.Training is almost always associated with speciallicensing and long lead times (Type V). The specializedlabor market is usually associate with many professionalbarriers as well as subsequent restrictions in scope ofpractice and labor mobility. Contestability is even lowerin the Type-VI—?why parens here) category. Mostresearch and other knowledge-generating activitieswould fall under this category. So does the training ofhighly specialized staff in universities and other highereducation centers.

There are no good examples of inputs for thehealth sector that would fit into the last row, withsignificant information asymmetry in addition tomeasurement problems.

Production Characteristics of Product Markets

Interventions and services can also becategorized along a similar continuum from highcontestability and high measurability through tointerventions and other outputs with low contestability,low measurability, and significant informationasymmetry (Figure 7). Whereas reduced contestabilitydue to market concentration is one of the main problemsencountered in factor markets (production of inputs), akey problem with interventions and other outputs(product markets) has to do with difficulties inspecifying and measuring outputs and outcomes.

13

Type I Type III

• •

• Non Clinical Activities• Management Support• Laundry & Catering

• Routine Diagnostics

Type IV Type VI

Type VII Type IX

•

• Policymaking• Monitoring/Evaluation

• Ambulatory Care• Medical• Nursing• Dental

Production Characteristics of Outputs(Product Markets)

High Contestibility Low ContestibilityMedium Contestibility

Type II

•

• Clinical Interventions• High Tech Diagnostics

Type V

Type VIII

• Public Health Interventions• Intersectoral Action• In-Patient Care

Hig

hM

easu

rabi

lity

Med

ium

Mea

sura

bilit

yLo

wM

easu

rabi

lity

Figure 7

As we move to the second row (Type IV toType VI), measurement of the outputs and outcomesbecomes more problematic. Although routinediagnostics such as laboratory tests may be highlycontestable (many players in a competitive market withfew barriers to entry), monitoring their performance interms of effectiveness and quality of the activitiesundertaken is much harder (Type IV). The same is truefor various nonclinical hospital activities.

As we move across to the Type V category,contestability is reduced by various barriers to entry.High-tech diagnostics usually requires specialization,licensing, and large sunk costs, giving establishedplayers a marked advantage over new entrants. A furtherbarrier to entry for these activities would be governmentpolicies that control or restrict the introduction of somenew technologies (CAT or NMR scanners). Clinicalinterventions are usually outsourced only to certifiedproviders. In each of these cases, outputs and outcomescan be measured, but it is more difficult that in the caseof activities in the first row.

In addition to difficulties in measuring outputand outcomes, most clinical interventions arecharacterized by an additional constraint of informationasymmetry. At times, information may be readilyapparent to patients (e.g., the quality of “hotel services”such as courtesy of clinical staff, the length of waitingperiods, the cleanliness of linens, the palatability offood, and privacy. Without survey techniques, however,such information may not be readily available to thecontracting policymakers or administrative staff.

For these reasons, ambulatory clinical care

would fall under the Type VI category (relatively lowbarriers to entry other than professionalqualifications/certification of staff) but high informationasymmetry and difficulties measuring outputs andoutcomes. As we move across the third row,contestability diminishes due to specialization and cost,in addition to measurement problems. For these reasons,public health interventions, intersectoral actionprograms, and inpatient clinical care belong to Type VIIactivities.

This leaves a few clear-cut activities such aspolicymaking, monitoring, and evaluation under theType IX category. The contestability and measurabilityof these activities is extremely low. These activities aretherefore usually retained as a core part of an integratedbureaucracy.

F. “MAKE OR BUY” DECISION GRID

Set Priorities First …

In many countries “make or buy” decision aremade before policymakers and providers have gonethrough an explicit priority-setting process. Prioritiesspecified should include the range of interventions tofinance through public resources (including preventiveservices) and should ensure that public subsidies areappropriately targeted (e.g., to the poor and othervulnerable groups).44 Countries often rush into “make orbuy” decisions before setting priorities for needed andaffordable interventions. Such prioritization iscomplicated by the fact that the cost of treating differentillnesses varies greatly and often bears little relation tothe effectiveness of available interventions.45

Furthermore, for a whole rage of activities, informationdisclosure and coordination through a strongstewardship function may be sufficient.

… Then Decide Who Can Produce What

Based on the above discussion, it is now easy tomap the goods and services that can be bought, thosewhere coordination is enough, and those that are betterproduced in-house by the public sector itself (Figure 8).The size of the “make” in-house production triangle willdepend largely on the effectiveness of policyinstruments to deal with contestability and measurabilityproblems. See section G for discussion..

14

Hig

hM

easu

rabi

lity

Med

ium

Mea

sura

bilit

yLo

wM

easu

rabi

lity

High Contestibility Low ContestibilityMedium Contestibility

“Make or Buy” Decision Grid

Option toBuy InputsOption to

Buy Services

Leave to Market

Retain Hierarchical

ControlFigure 8

… Finally Decide From Whom to Buy and How

Once “make or buy” options have been settled,the next questions relate to: (b) whom to buy from; and(c) how to structure the purchase.

Whom to buy from

• Consider all possible producers (public,nongovernmental and private-for-profit; domestic,and international).

• Base purchase decision on best product at the lowestprice responsive to specific needs—type of good,price, quantity, after acquisition support, timeliness,and so forth (consider international competitivebidding when possible).

• If there is currently no market, consider stimulatingdemand rather than in-house production.

• If contestability is low and there is no competitivemarket, consider using benchmark purchasing(based on estimated reference costs) so thatsuppliers have to compete for rather than in themarket.

• If there is a market, but it is dysfunctional, considerimproving its function through appropriateincentives (strategic subsidies) or regulations(antitrust).

And how to buy

• Choose the contractual arrangement most suitablefor a given purchase (spot market for unpredictableitems, medium-term supply contract for predictableitems, franchise arrangements for standardizedneeds at multiple locations, and relational contractfor difficult-to-monitor purchases).46

Once a buying decision has been made, allpotential producers must be treated alike by creating alevel playing field. This includes ensuring that there areno hidden competitive advantages such as taxconcessions or access to subsidized capital. And itmeans ensuring that no unfair competitive advantage isgiven to any producer through privileged access toinformation.

G. POLICY LEVERS AVAILABLE TO GOVERNMENTS

From the previous discussion, we can see thatmost goods and services have some degree of marketimperfection in terms of reduced contestability andmeasurability. Governments have at their disposal avariety of instruments that they can use to address theseproblems. A few of these instruments—from least tomost intrusive—include requiring informationdisclosure, introducing regulations, contracting forservices, providing subsidies or direct financing, andbeginning public production. In this section we discussthese instruments as they relate to “make or buy”decisions.

Standard Policy Instruments

Factor markets. For some inputs—theproduction of consumables, unskilled labor, and theretail of drugs, medical equipment, and consumables—there are few serious market imperfections such asreduced contestability and low measurability (Figure 9,upper left corner of matrix). With minimal governmentintervention such as good information disclosure andsome quality or safety standards, competitive marketsare best at producing these inputs. Public production ofthese inputs usually leads to low quality, lack ofinnovation, and inefficient production modalities.

15

Policies to Deal with ReducedContestibility and Measurability

Information Disclosure

Regulations and Contracting

Public Financing

Production

Hig

hM

easu

rabi

lity

Med

ium

Mea

sura

bilit

yLo

wM

easu

rabi

lity

High Contestibility Low ContestibilityMedium Contestibility

Figure 9

The other extreme—training very specializedlabor and generating knowledge about rare healthconditions and their treatment—is characterized byconsiderable market imperfections due to reducedcontestability and low measurability. A mix of strongregulation and in-house production is often needed toensure adequate generation of these inputs.

Most other inputs can be bought. However,markets often give the wrong signal about the level(surpluses and shortages), mix, and distribution of theseinputs. This is especially true of human resources andthe production of pharmaceuticals and medicalequipment with a long development or training phase.Skilled use of regulations and contracting mechanisms istherefore needed when purchasing inputs that havemoderate contestability and measurability problems.

Large producers may try to severely reducecontestability by erecting strong barriers to entrythrough protective policies (patents and licensingrequirements), benchmarking (manufacturing standards),large sunk cost requirements, collusion, and a highdegree of specialization (R&D). For these inputs,stronger policy measures may be needed such asmonopsony purchasing power and long term contracts.

Despite this complex landscape in goodscharacteristics, in many areas of reduced contestabilityand measurability, governments could achieve mostequity, efficiency, and quality objectives throughregulations and contracting.47

Product markets. As in the case of inputs, theproduction of interventions and other outputs can be

“contracted out” (purchased) and do not in principlehave to be produced in-house.

In practice, decisions about which interventionsto make in-house and which to contract out arecomplicated by a number of factors. First, for someoutputs (e.g., clinical interventions) what is to bedelivered is much harder to specify than the inputs. Thismakes it difficult to manage the resulting contracts andprevent opportunistic behavior by providers (privatehealth insurance is especially vulnerable to opportunisticbehavior). Second, contestability is often reduced for thereasons described earlier in our discussion on inputs.Finally, complex health problems often require strategiccoordination among different interventions and otheroutputs (integrated care, continuity of care, appropriateand timely referrals, and the like).

In the case of outputs, policymakers need toexamine two critical questions in addition to the degreeof contestability and measurability before arriving atgood “make or buy” decisions. Is a strategic coordinatedresponse needed? To what degree do the goods andservices benefit from ongoing innovation andadaptability?

For example, nonclinical activities such ascustodial services, catering, laundry, and managementdo not require special strategic coordination. They canusually be “unbundled” and “contracted out” as astandard service to firms that specialize in theseactivities without too much customization. In contrast,clinical and public health interventions often do need tobe coordinated and tailored to the individuals andpopulations receiving them and the organizationsproviding them. Experience has shown that“unbundling” these activities often leads to manyproblems such as cost shifting, discontinuity of care, andpoor quality.48

Other Often Forgotten Policy Levers

The contestability and measurability of goods andservices is not static but influenced by elements of thesystemic environment. Government policies directlyinfluence this environment and the “nature of the good,”yielding alternative levers to take them closer to (orfarther away from) the ability to use the indirect tools ofcontracting and regulation. These alternative leversinclude:

16

• Governance: relationship between owner(governments) and health care organizations

• Market environment: competition in or for goodsand services markets

• Purchasing mechanisms: funding or paymentsarrangements for the goods or services.

These three factors exert a powerful influenceon the “nature of the goods” and hence on the ability toensure delivery through indirect mechanisms. In the nextthree sections, we will see how these factors combine todetermine the level of contestability in the market ormeasurability of a good. This will include a discussionof which instruments are effective in dealing with therelated market and government failures.

Governance and Internal Incentive Regime

Changes made in governance— the relationshipbetween government and organizations— influence thegoods characteristics of the health care goods andservices in question. This relationship can be modifiedsubstantially in five different dimensions: (a) thedecision rights given managers; (b) the residualclaimant status; (c) the degree of market exposure; (d)accountability arrangements; and (e) adequacy ofsubsidies to cover social functions.49

Contestability may be enhanced by:

• unbundling large bureaucratic structures(modification in governance)

• outsourcing other functions to specialized providers(modification in payment system)

• leveling the playing field by exposing all the actorsto the same potential benefits and losses due tomarket exposure (modification payment system)

• decreasing barriers to entry due to politicalinterference or unwarranted trust in publicproduction (modification in market structure)

• explicitly separating contestable commercialfunctions and noncommercial social objectives(modification in governance and stewardship).

Measurability may be enhanced by.

• relying on quantifiable results (output or outcomemeasures) for accountability and performancetargets rather than process (inputs and bureaucraticprocedures) (modification in governance andpayment system)

• shifting from difficult-to-define, long-termrelationships (employment or service arrangements)to shorter term, more specific contractualarrangements (modification in payment system)

• using quantifiable monetary incentives instead ofmore difficult to track nonmonetary incentivepayments such as ethics, ethos, and status(modification in payment system)

• tightening reporting, monitoring, and accountabilitymechanisms (modification in governance andpayment system).

For example, by removing restrictivegovernment monopolies from vaccination services(governance/market), such programs could be shiftedinto a Type II or even Type I position. It is easy tomeasure the number of children vaccinated, whocontracts a given disease, and entry barriers for firmsthat want to provide vaccination services on behalf ofthe government. Similar action applied to other servicescould shift many away from the lower right corner of thegrid toward the upper left corner (Figure 10).

Type I Type III

Type IV Type VI

Type VII Type IX

The Nature of Health Care GoodsBased on Organizational Economics

High Contestibility Low ContestibilityMedium Contestibility

Type II

Type V

Type VIII

Hig

hM

easu

rabi

lity

Med

ium

Mea

sura

bilit

yLo

wM

easu

rabi

lity

Exp e

n siv

eEq

u ip m

e nt

and

Dru

gs University

Hospitals

Ambulatory Care Figure 10

Likewise, tertiary and quaternary care providedin university hospitals could be shifted from a low-contestability/measurability grid to a medium-contestability/measurability position through betterinformation on outcome, policies that favor clearlydefined contracts, performance benchmarks, and atightening of reporting, monitoring, and accountabilitymechanisms. The same would be true for public healthservices and activities such as vaccination that are oftenpart of the responsibilities of ambulatory care providers.

Several factors may also alter the goodscharacteristics of pharmaceuticals, medical equipment,

17

and consumable supplies. As recently as 10 years agodevelopment costs, patent protection, and a small marketshare may have made highly specialized medicalequipment or drugs (Type III goods) very expensive.Today, they may behave like ordinary goods (Type II orI). Examples include the quick production of genericdrugs by many companies once patent protection expiresor the rapid increase in use of sigmoidoscopes andtranscutaneous surgical instruments once the technologywas no longer new and prices dropped.

This shift in goods characteristics is not a one-way street. The goods properties can also become lesscontestable and more difficult to measure.Organizational reforms do not always lead to increaseddecision rights, residual claimant status, marketexposure, accountability arrangements, and explicitsubsidies to cover social functions. In fact, during thepast 50 years, many national health systems deliberatelyshifted goods and services in the opposite direction bynationalizing ownership and production.

And market imperfections may contribute toentry barriers instead of lowering them. Doctors,dentists, and pharmacists can and do collude to restrictentry by potential competitors. Hospitals have a naturalmonopoly for their services for patients living nearbyand can create monopoly power through relations withother hospitals and referring doctors. Medical equipmentdistributors with licensing agreements for the topinternational companies can easily monopolize adomestic market. Pharmaceutical retailers can controltheir mark-up by forming professional cartels. Thepublic and nongovernmental sectors have a competitiveadvantage over the private sector due to their access tosubsidized or free capital from domestic and foreigndonors.

Market Environment

A central argument in favor of exposing providersto market forces is that, in a functioning market,competitive forces will lead to a more efficientallocation of resources than a command economy ornonmarket solutions. The structure of the market towhich organizations are exposed, therefore, has a criticalinfluence on their behavior. It may directly determinewhat strategies make sense to generate more revenue.

Policies that influence the competitiveenvironment through regulations or contracting cansignificantly alter the contestability of health care goods

and services. Similarly, information asymmetry can bereduced by policies that increase the availability of goodinformation on health services, enhance health careproviders’ institutional capacity to deal with suchinformation, and improve patients’ understanding abouthealth problems.

Such policies not only address some of theunderlying contestability and measurability problems,but they also shift both the contestability/measurabilitygrid and the boundaries of needed governmentintervention to ensure favorable outcomes (Figure 11).

HighContestibility

LowContestibility

Shifting the Contestability/MeasurabilityGrid and Needed Public Policies

MediumContestibility

Public Financing

ProductionRegulations and Contracting

Information Disclosure

Hig

hM

easu

rabi

lity

Low

Med

ium

Mea

sura

bilit

y

Figure 11

Conversely, in a less competitive environmentwith weak policies and data to overcome informationasymmetry, the grid for services that fall into the upperleft corner (Type I, II, and IV) may contract, with thegrid in the lower right corner (VI, VIII, and IX)expanding.

Market Imperfections in Service Delivery

There are two related problems in the marketstructure of service delivery in most segments of thehealth sector. First, little or no competition mayemerge—reducing the pressures on the provider todeliver “value for money” to maximize profits.Alternatively (or in addition), competition may emerge,but it may be dysfunctional. Both cases are discussed inthis section.

Some health services, especially tertiary andquaternary, exhibit scale economies in production. Thisrelieves incumbent hospitals from pressure from newentrants. Geographic monopoly over certain servicesmay leave buyers very little leverage to negotiate with

18

service providers. There are many examples of strongcollusion among medical doctors that creates a virtualmonopoly, thereby shifting the grid for ambulatorymedical care toward the left—and strengthening theneed for direct provision or other policy intervention.Public monopolies and policies that prevent public fundsfrom being used to contract services from the privatesector have the same negative effect on contestability.

Even for services where monopoly power is not anissue, providers may still capture market share ormaximize profits through various forms of distortionarybehavior (Figure 12).

In a competitive market, firms seek to maximizetheir profits—and by using any method that makes sensein that environment. In a healthy market environment,they will try to capture market share from theircompetitors by better pleasing customers, maximizeprofits by reducing costs through efficiency gains, andexpand their product lines through imitation orinnovation. Wherever possible, however, they will seekto exploit or construct advantages. Where this ispossible, the pressures for efficiency and qualitygenerated by the market may be weak.

Market Forces That InfluenceCompetition

Bargaining Power of Suppliers

BargainingPower of

Customers

Threat of Substitute Products or

Services

Threat ofNew

Entrants

The IndustryJockeying for

PositionAmong Current

Competitors

Figure 12

Such distortionary features of health servicemarkets often enable providers to: (a) counter thebargaining power of suppliers, patients, or purchasers;(b) ward off threats posed by new entrants and imitationproducts; and (c) control a large share of the relevantmarket.

Information asymmetry in the health sectorexacerbates these problems. For example, medicaltreatment is largely a “bundled” good where the seller

(doctor) guides patients’ consumption decisions—whichhospital to go to for surgery, which lab to use fordiagnostic services, and so on. Thus, providers canparley their information advantage into control over arigid and lucrative referral chain. Doctors may “forwardintegrate” into diagnostic labs or pharmacies and steertheir patients toward consumption where they have afinancial stake. Hospitals may “backward integrate” bycreating strong links with doctors, thereby cornering partof the market where they experience little or nocompetitive pressure. Medical professionals arefrequently able to create cartels, limiting competitivepressures that strengthen the influence of patients andpurchasers.

Since patients and payers know less thanproviders about the true value or cost of health services,providers can cream-skim, selecting patients who costless to treat than other patients. Thus, providers canincrease their profits, not by delivering better service tocapture market share or cutting costs but by choosingmore profitable patients.

Most of these market imperfections in servicedelivery can be corrected through appropriateregulations and contracting arrangements.

A few examples will illustrate this point. Equal access tocapital and antitrust legislation, limiting the power ofprofessional cartels, can significantly decrease the entrybarriers for some segments of the health care market,especially for clinical services that fall in the middleband of the contestability/measurability grid. The samewould be true for contracting practices that are open toboth public and private providers and which leave openpossibilities for choosing alternative providers orexercising “exit” strategies.

In other instances, supplier cartels, combinedwith low quality-control standards, shift activities suchas retail sale and distribution of pharmaceuticals andmedical equipment into the lower right corner, eventhough such activities belong in the upper left area ofhigh contestability and measurability.Market Imperfections of Private Health Insurance

Even if private health insurance is contestable,due to severe information asymmetry, such services areoften deliberately crowded out for strategic reasons byrestrictive policies and public financing. This topic isbeyond the scope of discussion in this paper but is

19

discussed in detail elsewhere and briefly in Box 4.50

Box 4. The Intractable Market Imperfectionsof Private Health Insurance

Private voluntary health insurance is one area that is particularlyprone to market imperfections, many of them related to informationasymmetries.

Insurance may succeed in protecting some people againstselected risks, but it usually fails to cover everyone who wants tosubscribe to insurance plans and often excludes individuals who needhealth insurance the most or who are at greatest risk of illness. Thishappens because insurers have a strong incentive to enroll onlyhealthy or low-cost clients (risk selection or cream-skimming).Private insurers also have incentives to exclude costly conditions orto minimize their financial risk by using benefit caps and exclusions.This limits protection against expensive/ catastrophic illnesses.

Because of these factors, individuals who know they are atrisk of illness have a strong incentive to conceal their underlyingmedical condition (adverse selection). Individuals who are—or thinkthey are—healthy will often try to pay as low premiums as possible.This prevents insurers from raising the funds needed to cover theexpenses incurred by sicker or riskier members. Worse, the healthymay even deliberately underinsure themselves, in the hope that freeor highly subsidized care will be available when they become ill(free-riding). When third-party insurers pay, both patients andproviders have less incentive to be concerned about costs, and somemay become careless about maintaining good health. This leads notonly to more care being used (the reason for insurance), but also toless effective care, or care that would not be needed if peoplemaintained good health (moral hazard).

Purchasing Mechanisms

Finally, provider payment systems alsoinfluence goods properties by interacting with three ofthe five key elements of the internal incentive regime ofhealth care organizations: distribution of residual claims,market exposure, and provision for social functions.Service providers, in particular, respond differently toalternative funding and payment mechanisms. Forexample, collective purchasing by a strategic socialhealth insurance fund in Germany sends a different setof signals to providers than regulated competition in theUnited States, consumer-driven demand through out-of-pocket payments in India, medical savings accounts inSingapore, and monopsonistic purchasing in the UnitedKingdom.

Although reforms in governance may endow anorganization with formal claims to residual revenue indifferent categories, the structure of the paymentssystem will directly determine whether this claim hasany real meaning or incentive effect. If, for example,services must be delivered at prices below cost, there

will be no residual to claim. Thus, the relation of coststo the price-setting and capital-charging formula in thepayments system is a critical determinant of theincentives of the model. The crucial factor is whethermarginal cost-saving efforts by the provider cangenerate revenue flows that the provider can keepwithout deterioration in quality or effectiveness.

When reforms in organizations such as hospitalsentail a shift in revenue earning by delivering services“in a market,” what kind of market emerges becomes acrucial issue. Often, government is the largest or onlybuyer. In this case, the process and terms on which thegovernment purchaser engages the provider may welldetermine the degree of pressure on the provider to“deliver the goods.”

To gain maximum benefits from reforms thatexpose the public sector to competition with the privatesector, it is crucial that adequate steps are taken tosecure competitive neutrality.

Two sets of policies must be built into providerpayment systems to achieve competitive neutrality:

• moneterization of social functions such as explicitsubsidies that cover the cost plus a reasonablemargin in delivering services to nonpaying oruninsured patients

• leveling of the playing field through astandardization of the fee structure and cost ofcapital for both the public and private sector.

H. GETTING FROM HERE TO THERE

We have presented a strong argument for acontinued and even an enhanced role for the state inproviding strong sectoral stewardship and securingequitable and sustainable financing for the health sector.But it challenges the principles and nature of publicintervention pursued by many governments, especiallyin the area of the public production of health services.

Many countries today have large inefficientpublic sectors producing goods and services that couldbe bought from nongovernmental providers. Movingfrom one system to another will not easy (Figure 13). Itwill take time and must be accompanied by capacitybuilding in areas such as contracting, regulation, andcoordination of nongovernmental providers.

20

A three-step process can be used to movegradually from one balance to another in the public-private mix in service delivery. First, when there isalready a large private sector, the public sector can beginby recognizing its existence and slowly increase use ofthese resources through better coordination, contracts,and a positive regulatory environment.

Once some learning has taken place incoordinating and contracting with existing providers, thepositive lessons from this experience can be transferredto other priority areas where nongovernmental providersmay not yet be active. Finally, in some cases where thepublic sector is clearly engaged in inefficient activitiessuch as public production of many inputs, these can beconverted through outright privatization andsubsequently bought from the private sector.51

At the same time, the public sector may not beinvolved in areas of strategic importance such assecuring financial protection against the cost of illnessand failing to provide critical sectoral oversight in termsof its stewardship function. Parallel to moving out of thearea of production of goods and services, a strongargument can be made for a more integrated approachand greater public sector involvement in health carefinancing, sectoral coordination, regulation, monitoring,and evaluation.

S

GI PS

F

Unbalanced - Governmentexcessively involved in“Rowing Function” –Generation of Inputs andProvision of Services.

More Balanced -Government activitiesconcentrated on “SteeringFunctions” - Stewardshipand Financing

Appropriate Role of Government

Generationof

Inputs

Provisionof

Services

Steward-ship

GI PS

FinancingPublic

Private

Figure 13

21

ENDNOTES 1 World Health Organization (WHO), World Health Report

2000: Health Systems Performance (Geneva: WHO, 2000).2 D. Black, An Anthology of False Antitheses (London:

Nuffield Provincial Hospital Trust, 1984). Flexibility andrigidity; access and privacy; animal experiment an animalwelfare; medicine and alternative medicine; science andcompassion; acute and chronic sector; hospital orcommunity; treatment or prevention; science or compassion;access and privacy.

3 Papyri are ancient Egyptian clay tablets. J. K. Mason and R.A. McCall Smith, Law and Medical Ethics, 2d ed., (London:Butterworths, 1987), p. 4, quoting A. Castiglioni, A Historyof Medicine, translated and edited by E.B. Krunbhaar, 2d ed.1947).

4 In this famous cuneiform legal code of the first BabylonianDynasty, has 9 of its 282 statutes relate to the services ofhealers. Statutes 215–17 and 221–23 deal with lawsgoverning the fees to be received for certain services;Statutes 218–20 deal with penalties to be inflicted on thehealer in the case of unsatisfactory therapeutic results anddeath. Careleton B. Chapman, Physicians, Law, and Ethics(New York: New York University Press, 1984), pp. 4–5.

5 Control of membership and secrecy, reflected in theHippocratic Oath, was characteristic of all trades. BritishMedical Association (BMA), Handbook of Medical Ethics(London: BMA, 1984), p. 6.

6 Today, the United States, Mexico, and Turkey are threeexceptions in the Organization for Economic Cooperationand Development (OECD) where universal access has notyet been secured. For a review of the introduction ofuniversality in the OECD, see A.S. Preker, The Introductionof Universality in Health Care (London: IIHS, 1989).

7 For a comprehensive discussion see R.G. Evans, StrainedMercy (Toronto: Butterworth, 1984). The classics include:F. Bator, “The Anatomy of Market Failure,” QuarterlyJournal of Economics 72:3 (1958): 351–79; A.B. Atkinson,and J.E. Stiglitz, Lectures on Public Economics(Maidenhead: McGraw-Hill, 1980); and R.A. Musgrave andP.B. Musgrave, Public Finance in Theory and Practice, 4thed. (New York: McGraw-Hill, 1984).

8 For a more comprehensive discussion see Barer, L. Morris,Thomas E. Getzen, and Greg L. Stoddart (eds.), Health,Health Care and Health Economics: Perspectives onDistribution (Chichester, West Sussex, England: John Wiley& Sons, 1998); and E. van Doorslaer, A. Wagstaff, and F.Rutten (eds.), Equity in the Finance and Delivery of HealthCare: An International Perspective, (Oxford: OxfordMedical Publications, 1993). The classical reference is K.W.Arrow, “Uncertainty and the Welfare Economics of MedicalCare,” American Economic Review 53:5 (1963): 940–73.

9 For a comprehensive review, see P. Young, PrivatizationAround the Globe: Lessons from the Reagan Administration(Houston: National Center for Policy Analysis, January

1986); and J.S. Vickers and G.K. Yarrow, Privatization: AnEconomic Analysis (Cambridge, Mass.: MIT Press, 1992).

10 N. Barr (ed.), Labor Markets and Social Policy in Centraland Eastern Europe (Oxford: World Bank/OxfordUniversity Press, 1994); and World Bank, “Investing inPeople and Growth,” 1996 World Development Report:From Plan to Market, ch. 8 (New York: Oxford UniversityPress, 1996) pp. 123–32

11 One of the first proposals for this approach was published byA. Enthoven, “Consumer Choice Health Plan,” NewEngland Journal of Medicine 298:12 (1978): 650–58 and298:13 (1978): 709–20; A. Enthoven, Theory and Practiceof Managed Competition in Health Care Finance (NewYork: North-Holland, 1988).

12 World Health Organization, 1999 World Health Report(Geneva: WHO, 1999); WHO, European Health CareReforms: Analysis of Current Strategies, Series No. 72,(Copenhagen: WHO Regional Office for Europe, 1996);World Bank, 1993 World Development Report: Investing inHealth (New York: Oxford University Press, 1993); WorldBank, Sector Strategy: Health, Nutrition, and Population(Washington: World Bank, 1997); and UNICEF, State of theWorld’s Children (New York: UNICEF, 1999).

13 For a review of the health care problems in the formersocialist states see A.S. Preker and R.G.A. Feachem, MarketMechanisms and the Health Sector in Central and EasternEurope, Technical Paper Series No. 293 (Washington:World Bank, 1996); (translated into Czech, Hungarian,Polish, Romanian, and Russian).

14 Most health economists—even those favoring a morecompetitive marketplace—recognize that government needsto play a significant role in the health sector. For anexcellent recent review on this topic see T. Rice, TheEconomics of Health Reconsidered (Chicago: HealthAdministration Press, 1998). For a more detailed discussionon the theory and empirical evidence of public and privateroles in health care financing, see P. Musgrove, Public andPrivate Roles in Health: Theory and Financing Patterns(Washington: World Bank, 1996); and G. Schieber (ed),Innovations in Health Financing (Washington: World Bank,1996).

15 See D. Osborne and T. Gaebler, Reinventing Government(New York: Plume, 1993).

16 See R. Saltman and J. Figueras (eds.), European HealthCare Reform: Analysis of Current Strategies (Copenhagen:WHO, 1997), pp. 213–14.

17 See J.Q. Wilson, Bureaucracy (New York: Basic Books,1989); J.D. Donahue, The Privatization Decision: PublicEnds, Private Means (New York: Basic Books, 1989); andWorld Bank, World Development Report: The State in aChanging World (New York: Oxford University Press,1997).

18 C.J. Wolf, “A Theory of Non-Market Failure,” Journal ofLaw and Economics 22:1 (1979): 107–39; A. Peacock, “On

22

the Anatomy of Collective Failure,” Public Finance 35:1(1980): 33–43; D.L. Weimer and A.R. Vining, PolicyAnalysis: Concept and Practice (Englewood Cliffs, NJ:Prentice Hall, 1989); A.R. Vining and D.L. Weimer,“Government Supply and Government Production Failure: AFramework Based on Contestability,” Journal of PublicPolicy 10: 1 (1990): 1–22.

19 Much of this discussion is not new but firmly rooted inpolitical and moral philosophy. See K. Arrow, Arrow’sTheorem: The Paradox of Social Choice (New Haven: YaleUniversity Press, 1980.)

20 Although there are no technical limits, few countries likeSwitzerland use plebiscites, and only for major issues ofnational interest. Even then, it is an imperfect instrument dueto low voter turnout.

21 Four interrelated causes have been identified that explainthis observation (A.R. Vining and D.L. Weimer,“Government Supply and Government Production Failure: AFramework Based on Contestability,” Journal of PublicPolicy 10:1 (1990): 15). First, direct democracy usuallyprevents the overseer from knowing the preferences ofsociety. Second, overseers (representatives) typically paymore attention to constituencies and sensational issues thatare most likely to be influential in their reelection. Third,public officials usually have more than one overseer, oftenleading to conflicting and unstable political demands.Finally, since overseers do not fully benefit themselves fromthe effectiveness of their oversight, they often devoteinadequate time and effort to this task, leaving bureaucratswith a large range of independence.

22 This conceptual framework was developed by M. Moore,Creating Public Value (Boston: Harvard University Press,1995).

23 We will discuss this further in a later section under agencyproblems. T.L. Beauchamp, and J.F. Childress, Principles ofBiomedical Ethics, 2d ed. (Oxford: Oxford University Press,1983), p. 4; and E.J. Lemmon, E.J., “Moral Dilemmas,”Philosophical Review 71 (1962): 139–52.

24 See theoretical section for a more complete discussion of theprincipal-agent concepts. S. Bennett, B. McPake,and A.Mills (eds.), Private Health Providers in DevelopingCountries: Serving the Public Interest? (London: ZedBooks, 1997).

25 E.F.Fama and M.C. Jensen, “Agency Problems and ResidualClaims,” Journal of Law and Economics 26:2 (1983): 327–49.

26 D. Kaufmann and R. Ryterman, “Global CorruptionSurvey,” World Bank Working Paper (Washington: WorldBank, 1998).

27 A. Hirschman, Exit, Voice, and Loyalty: Responses toDecline in Firms, Organizations, and States (Boston, Mass.:Harvard University Press, 1970; World Bank, SectorStrategy: Health, Nutrition, and Population.

28 See J. Robinson, “Physician-Hospital Integration and

Economic Theory of the Firm,” Medical Care Research andReview 54:1 (March 1997): 3-24

29 OECD, Regulatory Reform, Privatisation and CompetitionPolicy (Paris: OECD, 1992), p. 17.

30 D.E. Sappington, “Incentives in Principal-AgentRelationships,” Journal of Economic Perspectives 5:2(1991): 45–66.

31 E.F. Fama, “Agency Problems and the Theory of the Firm,”Journal of Political Economy 88:2 (1980) 288–307; M.C.Jensen and W.H. Meckling, “Theory of the Firm:Managerial Behavior, Agency Costs and OwnershipStructure,” Journal of Financial Economics 3 (1976): 305–60.

32 See O. Williamson, “Comparative Economic Organization:The Analysis of Discrete Structural Alternatives,”Administrative Science Quarterly 36 (June 1991): 69–296.

33 Two useful references include: O. Williamson, TheEconomic Institutions of Capitalism: Firms, Markets andRelational Contracting (New York: Free Press, 1985); andO. Williamson, “Transaction Cost Economics,” ch. 3 in R.Schmalensee and R. Willig (eds.), Handbook of IndustrialOrganization (New York: North-Holland, 1989).

34 See P. Milgrom and J. Roberts, “Bargaining Costs, InfluenceCosts, and the Organization of Economic Activity,” in J. Altand K. Shepsle, (eds.), Perspectives on Positive PoliticalEconomy (New York, Cambridge University Press, 1990),pp. 57–89.

35 P. Milgrom and J. Roberts, Economics of Organization andManagement (Englewood Cliffs, NJ: Prentice-Hall, 1992) p. 291-292.

36 Milgrom and Roberts, Economics of Organization andManagement, ch. 9.

37 For example, a person may own a house but not have theright to occupy it if he has leased it out. He may own a carbut not have the right to transfer it freely if he has a loansecured by the car.

38 M. Olson, The Rise and Decline of Nations: EconomicGrowth, Stagflation and Social Rigidities (New Haven: YaleUniversity Press, 1982).

39 N. Girishankar, “Reforming Institutions for ServiceDelivery: A Framework for Development Assistance with anApplication to the HNP Portfolio,” World Bank PolicyResearch Working Paper 2039 (Washington: World Bank,1999).

40 W.J. Baumol, J.C. Panzar, and R.D. Willing, ContestableMarkets and the Theory of Industrial Structure (New York:Harcourt Brace Jovanovich, 1982); and W.J. Baumol,“Towards a Theory of Public Enterprise,” AtlanticEconomic Journal 12:1 (1984): 3–20. For a critique of thistheory, see W.G. Shepherd, “Contestability vs.Competition—Once More,” Land Economics 71:3 (August

23

1995): 299–309; and W.G. Shepherd, “Contestability vs.Competition,” American Economic Review 74:4 (1984):572–87.

41 An asset is specific if it makes a much larger contribution tothe production of a good than its value in alternative uses.See B. Klein, R.G. Crawford, and A.A Alchian, “VerticalIntegration, Appropriable Rents and the CompetitiveContracting Process,” Journal of Law and Economics 21:2(1978): 297–326.

42 Adapted from N. Manning, “Unbundling the State:Autonomous Agencies and Service Delivery,” Draft WorldBank Discussion Paper, 1998.

43 See A.S. Preker and A. Harding (eds.), Innovations inHealth Service Delivery: Corporatization in the HospitalSector (Baltimore: Johns Hopkins University Press, 2000);forthcoming. In other parts of the public sector, suchautonomous and corporatized units are also variouslyreferred to as public or executive agencies, independentpublic organizations (IPOs), quasi-autonomousnongovernmental organizations (QANGOs), and state-owned enterprises (SOEs). There is no standard functionaldistinction between these different organizationalmodalities.

44 C. Ham, “Priority Setting in Health Care: Learning FromInternational Experience,” Health Policy 42:1 (1997): 49–66.

45 Musgrove, P, “Public Spending on Health Care: How AreDifferent Criteria Related?” Health Policy 47 (1999): 207–23.

46 See A. Harding and A.S. Preker (eds.), Private ParticipationToolkit (Washington: World Bank, 2000); forthcoming.

47 R. Herzlinger, Market Driven Health Care: Who Wins, WhoLoses in the Transformation of America’s Largest ServiceIndustry (Reading, Mass.: Perseus Books, 1997).

48 N. Manning, “Unbundling the State: Autonomous Agenciesand Service Delivery”; and Bennett, McPake, and Mills,Private Health Providers in Developing Countries.

49 See Preker and Harding, Private Participation Toolkit.50 See Musgrove, Public and Private Roles in Health.51 See Harding and Preker, Private Participation Toolkit.