Belli, P. (1997).

The comparative advantage of government: a review

(15 p.) Washington, DC: The World Bank (Working Paper, 1834). (026301)

The Comparative Advantage of Government:

A Review
Pedro Belli

This paper presents preliminary and unpolished results of analysis and is circulated to encourage discussion and
comments. Citation and use should take into account its provisional character. The findings, interpretations, and
conclusions expressed in this paper are entirely those of the author and should not be attributed to the World Bank, to
its affiliated organizations, to members of the Board of Executive Directors, or to the countries they represent.

A:\PUBLIC3.DOC
The Comparative Advantage of Government: A Review

The important thing for government is not to do things which

individuals are doing already and to do them a little better or a
little worse: but to do those things which are not done at all.
J. M. Keynes, “The End of Laissez-Faire”
1. Worldwide, the private sector is increasingly providing goods and services that a few
decades ago were deemed to be properly in the domain of the public sector. Several reasons
account for the shrinking role of government in this regard. First, the economic development
paradigm no longer considers it appropriate for governments to act as entrepreneurs. Second, an
inconclusive, albeit growing, body of evidence indicates that the public sector tends to use funds
less efficiently than the private sector when engaged in market-oriented activities.1 Third,
technological changes are making it possible to have competition in markets that traditionally have
been considered natural monopolies. Is there still a justification for public provision of goods and
services, or should the private sector be the sole provider?
2. The economic answer depends partially on whether a country’s welfare is likely to increase
more if the public rather than the private sector is the provider. In turn, this depends on a host of
conditions that vary from country to country and, within a particular country, from year to year:
for example, institutional arrangements; legal, regulatory, and political conditions; and external
circumstances. In addition to economic considerations, there are distributional, political and
strategic considerations. Consequently, there are no hard and fast rules by which to conclude
unmistakably that one or the other sector is the appropriate provider–in the end this decision is
largely a matter of judgment.2 This paper reviews the main arguments for public intervention in the
provision of goods and services and shows how the traditional tools of economic analysis of
projects can shed light on this important question.

General Considerations
3. Why should governments be involved in the provision of any good whatsoever? As far
back as 1776, Adam Smith argued in The Wealth of Nations that in competitive markets, an
individual pursuing private gains would promote the common good:
He intends only his own gain, and he is in this, as in many other cases, led by an invisible hand to
promote an end which was no part of his intention. Nor is it always the worse for society that it
was no part of it. By pursuing his own interest he frequently promotes that of the society more
effectually than when he really intends to promote it.

1 There are no theoretical grounds for supposing that private sector enterprises are more efficient than public
enterprises, nor is there conclusive evidence showing that one is more efficient than the other. Examples of
efficiency and inefficiency can be found in both sectors. Yet even those economists who make strong cases for
government intervention side with the popular notion that public enterprises are less efficient (see, for example,
Stiglitz, 1994, p. 237).

2 In the context of Bank work, the initial justification for public provision ought to stem from an analysis of the
country’s public expenditure program and, should be justified in the Country Assistance Strategy.

A:\PUBLIC3.DOC
 2

In the 1950s Arrow (1951a) and Debreu (1959) formalized Adam Smith’s insight in what are now
known as the two fundamental theorems of welfare economics. The first theorem says that under
certain conditions every competitive equilibrium is Pareto-efficient— that is, in an economy that
reaches a competitive equilibrium, no one can be made better off without making someone else
worse off. The second theorem says that under certain conditions every Pareto-efficient allocation
of resources can be obtained through a decentralized market mechanism. These theorems are
relevant to any discussion of the role of government in resource allocation: they imply that under
the conditions assumed by Arrow and Debreu, no government or central planner, however
omniscient and well-intentioned, can improve on the results obtained by the free market system.
The best of all possible planners might do as well as competitive firms attempting to maximize
their own profits, but they would never do better. This point is especially relevant for project
analysis because under the ideal conditions posited by Arrow and Debreu, a project’s net benefits
from an investor’s perspective would give an exact measure of the net benefits from society’s
perspective. The economic and financial evaluation of a project would yield exactly the same
result.
4. If the real world fulfilled the assumptions of the fundamental theorems of welfare
economics, the market would produce every good in demand and there would be no need for
governments to provide any good or service. Equity considerations, then, would be the only
economic justification for government intervention. However, the real world, is a far cry from the
idealized Arrow-Debreu world. In many cases private markets fail to produce the socially optimal
quantities of goods and services and, in principle, government intervention can enhance welfare.
5. Market failures (departures from the ideal conditions posited by Arrow and Debreu) occur
because (a) competition is imperfect (someone may have monopoly power, for example); (b)
producers may impose a cost or confer a benefit to other producers (or consumers) without paying
for the cost or charging for the benefit (that is, there are externalities); (c) the process produces a
“public good” for which it is impossible or undesirable to levy a charge; (d) markets are incomplete
(they do not extend infinitely far into the future and they do not cover all risks); or (e) information
is incomplete and imperfect. There is an a priori rationale for public sector involvement whenever
the market cannot or will not produce the socially desirable quantity of the good or service. But the
public sector should be the residual provider of goods, intervening only when the market does not
produce the socially optimal quantity of the good or service.
6. The nature of government involvement, however, merits careful consideration. In some
cases it may be appropriate for the government to produce goods (roads, for example); in others,
financing production of the service might be just more advisable (primary education, for example);
in yet others, a subsidy might be the most suitable intervention (subsidizing a forest that sequesters
carbon dioxide, or the access of poor people to safe water, for example). In all cases, the analyst
must ask three fundamental questions: (a) What market failure leads the private sector to produce
more or less than the socially optimal quantity of this good or service? (b) What sort of government
intervention is appropriate to ensure that the optimal quantity is produced? and (c) Is the
recommended government intervention likely to have the desired impact? If there is a strong case
for government intervention, we must assess the costs and benefits of government involvement and
show that the benefits are likely to outweigh the costs. We cannot assume that government
bureaucrats will succeed where markets fail. Government interventions, often poorly designed and
implemented, may create more problems than they solve. The rest of this paper will review some of
the most common market failures and the rationale for public intervention in each case. At the end
of the paper we will also discuss two reasons for government intervention, merit goods and poverty
reduction, that are separate from market failures.

A:\PUBLIC3.DOC
 3

Market Failures

Natural Monopolies
7. Natural monopolies provide one of the oldest justifications for government provision of
goods and services. Adam Smith’s invisible hand works well only in competitive markets. In many
markets competition does not exist, and in others competition is inefficient. Some production
processes enjoy economies of scale; that is, unit costs of production fall as output rises. A common
example is the supply of electricity: in densely populated regions, it is more efficient to supply
electricity through an integrated network than for every household to have its own generator. When
economies of scale are present, large firms produce more cheaply than small firms and tend to
dominate their markets; eventually they may drive smaller firms into bankruptcy and, in extreme
cases, may become monopolies. Industries in which the conditions of demand and supply are such
that production by a single firm minimizes costs are known as a natural monopolies. Unregulated
monopolies of any kind, natural or not, tend to charge too much and produce too little. Whenever
natural monopolies are present, government intervention, at least in principle, can lead to more
production at a lower price.
8. What kind of intervention is appropriate? The first alternative is to do nothing. This
solution might be optimal when the product or service has close substitutes and monopoly power is
weak, that is, when the ability to charge prices that result in excess profits is insignificant. In the
case of cable television, for example, the presence of close substitutes reduces the monopoly power
of cable providers enough to obviate government intervention. Before deciding that some form of
government intervention is called for, we need to assess the welfare losses from the exercise of
monopoly power. For a methodology for estimating the welfare losses from monopoly, see
Harberger’s (1954) seminal article, and the extension by Cowling and Mueller (1978). Ferguson
(1988) provides a summary of several studies on the subject.3
9. The traditional solution in many countries has been to have a public enterprise provide the
good or service. In many countries electricity is publicly provided, and many water companies
around the world are public enterprises. The assumption has been that a public enterprise would
maximize social rather than private welfare. To induce public enterprises to maximize social
welfare is extremely difficult because social welfare is tough to measure and hence so is the
performance of managers of public enterprises. Therefore, we must usually use proxies that at best
are imperfect substitutes. As a result, what managers of public enterprises maximize is not
necessarily welfare, but something else. Peltzman (1971) postulated that managers of public
enterprises maximize political support. His theory predicts that public enterprises will set a price
below the profit-maximizing price, voters will pay lower prices than nonvoters, and public
enterprises will use less price discrimination than private firms. Evidence from developed countries
supports Peltzman’s theory and shows that public enterprises tend to charge lower prices than
regulated private monopolies, practice less price discrimination, and adjust rates less frequently
(Peltzman, 1971, Moore 1970).
10. Another traditional solution has been to have a private, but regulated, enterprise provide
the good or service. In some countries, telephone companies are private, regulated monopolies.
Regulation itself has benefits and costs. The benefits are the reduction in deadweight losses in
efficiency that would exist under monopoly. The costs include the direct costs of regulatory

3 It should be noted that technological advances are making it possible to have competitive markets in areas that in
the past were considered natural monopolies (telecommunications, for example).

A:\PUBLIC3.DOC
 4

agencies, higher production costs because of changed incentives for regulated firms, and
unintended side effects of regulation. For regulation to be effective, the regulatory agency must
induce the firm to provide the good or service at average cost pricing, which in turn requires that it
have cost and demand information. For a discussion of the costs and benefits of regulation, see
Viscusi, Vernon, and Harrington (1966).
11. A solution that is becoming more common is to auction off the franchise to private firms.
The franchise is awarded via competitive bidding to the firm that offers to provide a given quality
of service at the lowest price. In theory, a large number of bidders drives the price down to the
point where the eventual provider earns a normal return. Franchise bidding should thus avoid the
need for regulation while achieving the same result. In practice, franchise bidding has been much
more complex, and it is not at all clear that it has generated socially desirable solutions. Viscusi,
Vernon, and Harrington (1996) provide a good review of experience in the United States.
12. Which is the preferred solution for dealing with natural monopolies— a regulated private
firm, a public enterprise, or franchise bidding? It is difficult to rank the alternatives in order of
preference. The evidence concerning the relative efficiency of regulated privately owned utilities
compared to public utilities is mixed, though the weight of the evidence points to greater efficiency
in regulated private enterprises (Moore, 1970; DiLorenzo and Robinson, 1982). The experience
with franchise bidding in the United States indicates that government quickly turns from mere
auctioneer to regulator. Nevertheless, because franchise bidding provides a greater role for
competitive forces, it is the most promising.

Externalities
13. Externalities provide another traditional argument for government intervention. Sometimes
activities generate benefits and costs that are not reflected in the benefits and costs of the firm. A
forest, for example, may lower the level of carbon dioxide in the world, but the owner of the
forest— who bears the full cost of planting and maintaining the forest— cannot charge for this
benefit. As a result, the forest may be smaller than desirable from the world’s point of view. In
some other cases, a project may use resources for which it does not pay. Consequently it may
produce more than is socially desirable. An irrigation project, for example, may lead to reduced
fish catch downstream. The project’s effect on the downstream fisheries are costs to society that
are not reflected in the project’s accounts. This discrepancy between private and social costs leads
to a larger scale of irrigation than is socially desirable. Externalities are among the principal
justifications given for public involvement in the provision of education services and prevention of
communicable diseases.
14. Government can intervene in various ways to induce firms to produce the socially optimal
quantity of goods whose production process is subject to externalities. Again, if the magnitude of
the externality is insignificant one alternative is to do nothing. Automobiles have been polluting the
air since they were invented, but the problem did not become serious until automobiles became
numerous. Another solution is to regulate. The Clean Air Act in the United States, for example,
sets ambient quality standards irrespective of cost considerations. A third solution is to tax the
producer of negative externalities to discourage their production and to subsidize the producer of
positive externalities to encourage their production. The Global Environmental Facility funds
production of goods and services that reduce global environmental externalities, for example.
15 Conceptually, at least, optimal solutions can be reached through taxes and subsidies.
Figure 1 shows the market for good X. The production of this good is subject to an externality that
increases the social cost of production (SMC) above the private cost (PMC). The marginal benefit

A:\PUBLIC3.DOC
 5

of good X is given by the demand curve. Without government intervention, the market will produce
Q units as compared to the optimal quantity Q* and the optimal price P*. An optimal tax equal to
P* - P would raise the price of X to P* and induce production of Q* units. Instead of a tax, the
government could impose a quota to limit production of X to Q* units. Eventually the market will
drive the price of X up to P*. Government could also intervene by producing good X and limiting
its output to Q*. If the externality were positive, the position of the SMC and PMC would be
reversed and the optimal intervention would be a subsidy.

$
Social MC

Private MC

P*

P
Marginal Benefit (demand)

O Q* Q Production of good X
FIGURE 1. MARKET SOLUTION VS. SOCIAL OPTIMUM WHEN EXTERNALITIES ARE PRESENT

Public Goods
16. The third traditional argument for public provision concerns the nature of the goods and
services themselves. All goods provided by the private sector share one important feature, namely,
that the provider of the good can charge those who wish to consume it and make a profit in the
process. Not all goods, however, share this characteristic. There is a broad category of goods
called public goods for which it is either impossible or undesirable to charge. The private sector
usually shies away from producing public goods; or if it does produce them, it usually charges too
much and produces too little of them. For example, cleaning up the air in Mexico City would be of
great benefit to the city, but no private sector company would do it because it could not charge for
the service.

A:\PUBLIC3.DOC
 6

17. Exclusion difficult or costly. Some public goods are not produced by private markets
because it is impossible to prevent anyone from consuming them, even if they do not want to pay
for them. These public goods are called nonexcludable goods. Consider national defense. If an
army succeeds in defending the national territory against an enemy, every citizen benefits, whether
he/she paid to sustain the army or not. Similarly, spraying an area to rid it of malaria-carrying
mosquitoes benefits every nearby inhabitant, but it is difficult, to charge everyone for the service.
Those who refuse to pay for spraying will get a free ride. If a sufficiently large number refuse to
pay, spraying may not take place. Because of these difficulties, the private sector will not usually
produce nonexcludable public goods (or will produce suboptimal quantities). Public production of
nonexcludable public goods has been considered to enhance public welfare and therefore to be a
proper function of government.
18. In some cases exclusion is possible, but costly. Roads are nonexcludable, but toll roads are
excludable. The costs associated with limited-access roads, however, are considerably higher than
those of normal roads: exclusion comes at a high cost. Whenever a project produces a good for
which the cost of exclusion is high, there is also a strong presumption for public provision.
19. Nonrival goods (exclusion undesirable or inefficient). Private goods also share another
important characteristic, namely that the marginal cost of consumption is high. In the case of
nonrival public goods, however, the marginal cost of consumption is zero or very low. Although
private production of nonrival goods is possible, the private sector will produce suboptimal
quantities. Socially optimal pricing requires that the price of goods or services be equal to the
marginal cost of consumption. If the price is set to equal marginal cost, private provision may be
unprofitable. Once a bridge is built, for example, the marginal cost of letting another car use it is
virtually zero (up to the point of congestion). For an uncongested bridge, optimal pricing would
require a very low toll, too low to recover the initial investment and hence too low to interest the
private sector. If the toll were set high enough to interest the private sector, too few cars would use
the bridge. Likewise, the cost of informing one thousand consumers over the air waves is the same
as the cost of informing two thousand, and the information available to a thousand additional
consumers does not reduce the amount available to others: the marginal cost of consumption is
zero. Whenever the marginal cost of consumption is zero or extremely low, it is undesirable to use
the price system to ration the good, as efficiency requires that goods and services be provided at the
marginal cost of providing them. The private sector, which would charge more than the marginal
cost of provision, would provide suboptimal amounts of these goods. Low marginal cost of
consumption is often used as an argument for public provision of research and extension, utility
services, and public information services (agricultural prices and weather patterns). The argument
for public involvement in the provision of nonrival public goods is strong, but the nature of the
involvement need not be provision of the good, as provision of funding may be optimal in many
cases. For example, the optimal quantity of research and extension services may be achieved with
public funding of private provision.
20. Some goods are both nonexcludable and nonrival. National defense, for example, is both
nonexcludable and nonrival: the cost of protecting a given territory against foreign invaders is the
same, whether the inhabitants of the territory amount to one or two million. Pure public goods are
both nonexcludable and nonrival. Private goods, say shoes, are both excludable and rival. In
between these two extremes, there is a host of goods that may be only partially nonexcludable or
partially nonrival. Information and technology, for example, are nonrival and partially
nonexcludable goods. The provider of pure public goods incurs all the costs but reaps none of the

A:\PUBLIC3.DOC
 7

benefits, the provider of private goods incurs all of the costs and (except for taxes) reaps all of the
benefits.1 As general rule, there is no economic justification for public provision of private goods
in competitive markets.
21. Figure 2 shows the ease of exclusion along the horizontal axis and the cost of exclusion
(marginal cost of consumption) on the vertical axis. Pure public goods appear in the lower left-
hand corner. These goods are nonexcludable and nonrival. Pure private goods appear in the upper
right-hand corner, where the marginal cost of consumption is equal, or nearly equal, to the average
cost, and the cost of exclusion is low. Impure public goods are only partially excludable or
partially nonrival. A bridge is an example of an impure public good. Once built, it is relatively
simple to exclude anyone from using it, but because the marginal cost of an additional automobile
crossing is low (up to the point of congestion), it is not socially desirable to charge a fee that
exceeds the marginal cost. A crowded limited-access highway is an example where exclusion is
easy, but where the marginal social cost of a vehicle entering might be very high.

Pure
High marginal cost of
private
consumption -¡
Publicly provided
private goods
Shoes
• goods

Desirability Crowded
of
Exclusion
• highway

Uncrowded
Public bridge
Health
National
Roads
Lighthouse •
Low marginal cost • Defense
• •
Exclusion difficult Exclusion easy

Ease of Exclusion

FIGURE 2: PURE IMPURE PUBLIC GOODS5

22. Many Bank projects finance public goods. A difference between a project that produces
private goods and one that produces public goods is the distribution of costs and benefits among
various groups in society. In the case of private goods, the benefits (save for taxes) as well as the

4 That is, only if the production of the good is not subject to externalities.

5 After Stiglitz (1988).

A:\PUBLIC3.DOC
 8

costs accrue to the project entity. In the case of public goods, the project entity incurs the costs but
society at large enjoys the benefits.
23. The very nature of public goods implies that there are no markets for them, and hence it is
difficult to value the benefits of producing them. As the Bank’s business moves away from
financing the production of private goods to financing the production of public goods, the
traditional tools of economic analysis need to be extended to evaluate the benefits of goods and
services for which there are no ready markets. In some cases (roads, for example) the tools are well
developed, in other cases (health, for example) they are controversial, and in yet others
(environment) they are rudimentary. Table 1 shows in schematic form a hypothetical distribution of
costs and benefits for a pure public good.
Table 1: Hypothetical Distribution of Costs and Benefits of a Public Good

Project Entity Government Others Society

Benefits 0 0 150 150
Costs (80) 0 (80)
Net benefits (80) 0 150 70
Subsidies 80 (80) 0 0
Total 0 (80) 150 70

24. In this case the provider incurs all of the costs but does not receive a monetary reward for
any of the benefits because it cannot charge for them. The benefits accrue to the country at large,
or “others.” The provider needs to be subsidized to survive and as a result there is a negative fiscal
impact of 80. As the last column shows, the project generates gross benefits of 150 and net benefits
of 70. Note that if the project produced a private good, the economic costs and benefits would be
exactly alike. What would differ would be the distribution of benefits and costs among the various
groups in society. If we looked only at the last column, we would not be able to tell whether the
good is public or private.
25. By exploiting the information embedded in the differences between private and social
prices and private and social flows, the tools of economic analysis of projects enable us to
construct tables showing the distribution of costs and benefits. They thus provide very valuable
information that can guide the decision to place a project in the public or in the private sector. For
an application of these techniques, see Belli, Anderson, Barnum, Dixon, and Tan, 1996.

Asymmetric Information and Incomplete Markets

26. Perfect information, equally shared among all consumers and producers, is a basic
assumption of the two fundamental theorems of welfare economics. Another basic assumption is
the existence of complete markets (a market for every type of good and service, for every type of
risk, extending forever into the future). Neither of these assumptions is ever fulfilled. Information is
always imperfect, and markets seldom provide all goods and services for which the cost of
provision is less than what individuals are willing to pay. When information is imperfect and
markets are incomplete, the actions of individuals have externality-like effects that result in
suboptimal production of goods and services (Greenwald and Stiglitz, 1986). Information-based
market failures differ from the market failures discussed above in two important respects. First, for
the most part, the former or “older” market failures are related to an easily identifiable source, and
second, they can be corrected (at least conceptually) with well-defined government interventions.

A:\PUBLIC3.DOC
 9

Market failures based on imperfect and costly information and incomplete markets, on the other
hand, are pervasive in the economy and difficult (if not impossible) to correct, as nearly all markets
are incomplete and information is always imperfect. Full corrective policy would be impractical as
it would entail taxes and subsidies on virtually all commodities. We now consider some arguments
for public intervention when information is imperfect and markets are incomplete.
27. Ignorant consumers and informed producers. Parties to market transactions rarely share
the same information. Producers usually know more than consumers do about the product they are
selling. Bank managers and bank owners, for example, know more about the financial health of
their institutions than consumers do. Buyers of used cars usually know less about the car than the
owner and may get stuck with a lemon. Patients usually know less about how to treat a disease
than their doctor and will accept the treatment prescribed, even if there is no need for it.
Asymmetric information is pervasive. If information were complete and equally shared, more
transactions would take place as fewer parties would fear “being taken.” Government interventions
that improve information flows can lead to more transactions and hence to increased welfare.
Although in principle taxes and subsidies would lead to optimal allocation of resources and hence
to improved welfare, in practice most interventions aiming at correcting information failures do not
rely on taxes and subsidies, but on the coercive power of government. Thus, in many countries
banks are required to disclose financial information, sellers are required to disclose information
about the goods they are selling to potential buyers, and there are strict disclosure requirements for
publicly traded stocks.
28. Informed consumers and ignorant producers. Asymmetric information also works against
producers. It is often argued, for example, that lack of information induces lenders not to lend to
certain groups, or lend at rates that cover not only costs and risks but also the lack of information.
Hence, they lend less than the social optimum. If information about the borrowers were better, it is
often argued, lenders would be willing to lend at lower rates and hence would attract more clients.
29. The World Bank is an example of an intervention that improves information and capital
markets. By virtue of its charter, the World Bank has access to information about its borrowers
that commercial banks do not and therefore it can assess country risks better than a commercial
bank. Because it can also influence policy, it can diminish country risks and improve the flows of
private capital into developing countries. The Grameen Bank in Bangladesh is an example of an
intervention that completes an incomplete capital market. This rural bank was created to provide
credit and organizational support to individuals who, for lack of material collateral, would be
excluded from the formal credit system. The Grameen Bank tackles the problem of asymmetric
information by lending only to small groups of self-selected individuals who are known to one
another and who monitor and provide the repayment guarantees for each other. It tackles the
enforcement problem by linking the individual’s continued access to credit to the group’s
repayment record. Among other things, its proximity to its clients facilitates the flow of
information, which in turn enables it to supplement an incomplete market and improve its
efficiency.
30. The rationale for public intervention in activities that provide information is strong. Stiglitz
(1988), argues that in many ways, information is a public good. First, it is nonrival, as giving
information to one more individual does not reduce the amount available to others. Second, it is
largely nonexcludable, as the marginal cost of giving information to one more individual is low and
at most equal to the cost of transmitting the information. Efficiency requires that information be
given at the marginal cost of providing it. Because the marginal costs of provision may be close to
zero, the private sector, which charges more than the marginal cost, often provides too little
information. Although the case for public intervention in the provision of information is strong, the

A:\PUBLIC3.DOC
 10

rationale for public provision of information is weaker. Publicly funded tornado warning services,
for example, may be provided over private radio stations; they need not be provided over public
radio stations.
31. Asymmetric information gives rise to adverse selection and moral hazard, two problems
that arise in a variety of contexts, but that are most clearly seen in the case of insurance markets.
With either problem, private markets tend to provide too little of a good or service, and appropriate
government intervention can enhance welfare.
32. Adverse selection. It is human nature to seek protection when we are at risk and to be less
careful when we are not at risk. In any given population, some people are at higher risk than others
with respect to a certain event. Residents of California are at higher risk from earthquake damage
than residents of Virginia. As a result, Californians generally buy earthquake insurance and
Virginians do not. If everyone in the United States purchased earthquake insurance, rates would be
lower. But buyers of insurance are not necessarily randomly selected. They are self-selected: only
those who consider themselves at risk buy insurance. From an insurance company’s viewpoint,
buyers are adversely self-selected. As a result, insurance companies charge higher premiums than
they would if the selection were random. Higher premiums lead to thinner markets. Whenever the
insurance premiums are set so high as to compensate for adverse selection, large numbers of people
who would be willing to pay for the actuarially fair rate to protect themselves from a particular
risk do not buy insurance. Adverse selection has been cited as a justification for public intervention
in, for example, the health sector. As a result of adverse selection and ensuing high insurance
premiums, many people may be left without insurance against catastrophic illness, leading to large
discrepancies between social and private benefits (Hammer, 1996).
33. Moral hazard. It is also human nature to be less careful when we do not have to suffer the
consequences of our own actions. Thus, we tend to speed if there is no fear of being caught. Banks
that are insured by the government against bad loans may not exercise as much care as banks that
incur the full cost of bad loans gone sour. Likewise, people who are insured against theft may be
less careful with their belongings than those who are not insured. The lack of incentive to take care
when we do not have to suffer the consequences of our own actions is called moral hazard. In
markets where moral hazard is present, too little of a certain good is typically offered. For
example, consumers would like to buy more insurance, and insurance companies to sell it— but
only if consumers exercise due care. However, insurance companies, cannot be sure that
consumers will exercise due care, and therefore they sell less insurance than consumers are willing
to buy: they ask consumers to bear some of the risk, that is, to self-insure and to copay. Many
authors think that moral hazard problems are not a good justification for public intervention (see,
for example, Lal, 1994, and Varian, 1987), but rather a limitation of the extent of insurance that
can be provided. But Greenwald and Stiglitz (1986) have shown that, in theory, if moral hazard
problems are present, governments can effect Pareto improvements by taxing and subsidizing
various commodities that induce people to take greater care. For example, by subsidizing the use of
seatbelts, the government can encourage drivers to exercise greater care, and by taxing cigarettes
and alcohol, it can discourage smoking in bed in a state of inebriation and therefore reduce the
incidence of fires. Government provision of information can have effects on incentives akin to those
of taxes and subsidies.
34. Risk aversion. The public sector, as representative of a country’s entire population, can
spread risk over every citizen in the country and is therefore in a unique position as an investor.
For this reason, Arrow and Lind (1970) argued that when governments act as investors, they
should be risk-neutral, that is neither prefer nor avoid risk. Governments, Arrow and Lind argued,
should normally choose projects on the basis of their expected net present value and disregard the

A:\PUBLIC3.DOC
 11

variance around the mean of the net present value. For private investors, who are normally risk-
averse, there is always a trade-off between risk and return, often expressed as a trade-off between
the variance and the mean. If problems of moral hazard did not exist and insurance markets were
complete, private investors would be able to buy insurance against commercial failure and
undertake riskier projects. But investors cannot insure against commercial failure and normally shy
away from excessively risky projects. The absence of an insurance market against commercial
failure and government risk neutrality imply that some risky projects may be attractive to the
public sector but not to the private sector (Arrow and Lind, 1970). If a project is not attractive to
the private sector because it is too risky, public provision may be justified, even if the project
produces a private good.
35. Complementary Markets. In some cases, the production of a good requires the production
of a complementary good— computers and computer programs, for example. Software companies
flourished only after the advent of personal computers. This example of complementary markets
involves only two goods. In some cases, many markets— and large scale coordination— must be
involved. Public intervention in urban renewal programs and rural development have been justified
on the grounds of this market failure. The renewal of a large section of a city or the development of
rural areas requires extensive coordination among many actors, including factories, retailers,
landlords, transport, and so on. Similarly the development of rural areas requires extensive
coordination among various actors. If markets were complete, the coordination would take place
through the price system. Incomplete markets require that someone act as coordinator.
36. Cost of Capital. In a perfect and undistorted capital market, the market rate of interest
would reflect the cost of capital to a country: on the demand side, the market rate of interest would
be equal to the marginal productivity of capital, and on the supply side, it would be equal to the
rate of time preference for consumption. Taxes, however, drive a wedge between the private and
the social opportunity cost of capital. On the demand side, the private after-tax return is lower than
the social return, that is, lower than the marginal productivity of capital in the private sector. On
the supply side, also because of taxes, the marginal return to savers is lower than the social return,
that is, lower than the rate of time preference for consumption. The cost of capital to the public
sector then, viewed as the weighted average of the social marginal productivity of capital in the
private sector and the social rate of time preference for consumption, is usually higher than the
private cost of capital. Under certain circumstances, however, the cost of capital to the private
sector might be higher than the cost of capital to society. For example, the public sector may have
access to low-cost sources of funds, say IDA, while the private sector may not. When the private
sector looks at a project, therefore, it may use a higher discount rate than the public sector and
reject projects with, for example, long gestation periods. For these reasons, some projects that may
be highly beneficial to society may not be attractive to private investors. In these cases, government
intervention, through provision or subsidies, can improve welfare.
37. Size of the Project. The size and strategic nature of the project may be another justification
for public provision. The public sector, as the representative of a country’s entire population, can
command more resources than any single private sector entity and can thus undertake large,
strategic projects that require capital investments that are beyond the financial reach of the private
sector. Sometimes large projects may be attractive to foreign investors, but many countries are
reluctant to allow foreign ownership of strategic resources. Public provision may be justified even
if the project produces a private good, when the nature and size of the project are such that the
domestic private sector would not be able to undertake it.

A:\PUBLIC3.DOC
 12

Other Grounds for Public Intervention

Poverty Reduction
38. Public intervention to reduce poverty may be justified on ethical and political grounds.
Even in the idealized Arrow-Debreu world, Pareto-efficient solutions achieved by the decentralized
market system depend upon the initial allocation of resources among all the actors in society. It
would be entirely possible for a Pareto-efficient solution to be glaringly inequitable, leaving some
with “too much” and others with “too little.” In most countries the distribution of resources is
unacceptable, and in some it is politically unsustainable. A case can be made for public provision
of goods that the poor consume relatively more of than the nonpoor— that is, for goods with low
income elasticity— on grounds of redistribution. Some types of health care may qualify. But low
income elasticity is not the only grounds for government intervention in the provision of goods and
services for the poor: many types of health and education services, that have high income elasticity,
are provided to the poor on grounds of redistribution. Moreover, sometimes leakage is either
technically inescapable or the political price of poverty reduction: to benefit the poor it may be
necessary to benefit some of the nonpoor. Nevertheless, if poverty reduction is given as a rationale
for public provision of any good or service, it is often desirable to target project benefits towards
the poor.

Merit Goods
39. Another argument for government intervention even in the absence of market failures
arises from the belief that individuals may not always act in their own best interest, and the
government must intervene to see to it that they do. Mandatory use of seatbelts in cars and of
helmets for motorbikers and mandatory elementary education are examples of a class of goods
known as merit goods. The paternalistic argument for government intervention is different from the
externalities and information arguments discussed above. Bikers may know the benefits of wearing
helmets, yet may continue to ride without them. Those who advocate government intervention
believe that it is not enough to provide information and force those who “misbehave” to pay for any
externalities through taxation. As Stiglitz (1988) states, “Those who take the paternalistic view
might argue that individuals should not be allowed to smoke, even in the privacy of their own
homes, and even if a tax, which makes the smokers take account of the external costs imposed on
others, is levied. This paternalistic role undoubtedly has been important in a number of areas, such
as government policies toward drugs (marijuana) and liquor (prohibition), as well as compulsory
education.”
40. Using the merit goods argument for justifying government intervention is very delicate and
controversial. Many economists believe that no group has the right to impose its will on another
group. Moreover, they fear that special interest groups will attempt to use the government to
further their own views about how individuals should act or what they should consume.

Summary
41. In theory, market failures are necessary but not sufficient conditions for justifying
government intervention in the production of goods and services. Even in the absence of market
failures, there might be a case for government intervention on grounds of poverty reduction and
merit goods.
42. In every case, a case for government intervention must first identify the particular market
failure that prevents the private sector from producing the socially optimal quantity of the good or

A:\PUBLIC3.DOC
 13

service, second, it must select the intervention that will most improve welfare, and third, it must
show that society will be better off as a result of government involvement; that is, it must assess
the costs and benefits of government involvement and show that the benefits will outweigh the
costs. Table 2 lists the most common rationales giving for public intervention and the types of
interventions most commonly used.
43. It is impossible to judge a priori whether or what type of government intervention is
appropriate to a particular circumstance or even to a class of situations. Such judgments are both
country- and situation-specific and must be made on a case-by-case basis. To be sure, it is easier to
make such judgments in the case of the old market failures than in the case of the new ones. Market
failures rooted in incomplete markets and imperfect information are pervasive: markets are almost
always incomplete and information is always imperfect. This does not mean that there is always a
case for government intervention and that no more analysis is needed. On the contrary, there is a
keener need for analysis. The welfare consequences of the “new market failures” are more difficult
to measure. Therefore, the contribution of government interventions to welfare is likely to be more
difficult to assess and the case for intervention (especially provision) more difficult to make.
44. It is important to keep in mind that government interventions are often poorly designed and
excessively costly. Poorly designed interventions may create market failures of their own. For
example, governments concerned about low private investment in high-risk projects may guarantee
them against risk, but in the process create problems of moral hazard and induce investors to take
no actions to mitigate such risks. Some interventions may turn out to be too costly relative to the
posited benefits. In seeking to provide extension services, for example, governments may incur
costs that are higher than the benefits received by farmers.

Table 2: Rationale for and Examples of Public Interventions

Rationale Examples of Intervention
Natural monopolies Franchise bidding, regulation, provision
Externalities Taxes and subsidies, regulation, provision
Public goods
Exclusion difficult Provision
Exclusion undesirable Subsidies, provision
Information failures Regulation, taxes and subsidies, provision
Incomplete markets Provision, taxes and subsidies, regulation
Equity objective Subsidies, provision
Redistribution Provision, subsidies
Merit goods Regulation, provision

A:\PUBLIC3.DOC
 14

Bibliography

Arrow, K.J. 1951a. Social Choice and Individual Values. New York Wiley.

Arrow, K.J. 1951b. An extension of the basic theorem of classical welfare economics. In
Proceedings of the Second Berkeley Symposium on Mathematical Studies and Probability, ed.
J. Neyman, 507-532. Berkeley: University of California Press.

Arrow, K. J., and Lind, R. C. 1970. “Uncertainty and the Evaluation of Public Investment
Decisions” American Economic Review 60: 364-78.

Belli, P., Anderson, J., Barnum, H., Dixon, J., and Tan, .-P. 1996. Handbook on Economic
Analysis of Investment Operations. Processed.

Coase, R. “The Problem of Social Cost.” Journal of Law and Economic, 60: 1-44.

Cowling, K., and Mueller, D. C. 1978 “The Social Costs of Monopoly Power. Economic Journal
88: 727-48.

Debreu, G. 1959. The Theory of Value. New York: Wiley, New York.

DiLorenzo, T. J., and Robinson, R. 1982. “Managerial Objectives Subject to Political Market
Constraints: Electric Utilities in the U.S.,” Journal of Public Economics 13: 259-76.

Dixon, J. A., Scura, L. F., Carpenter, R. A., and Sherman, P.B, 1994 Economic Analysis of
Environmental Impact. 2d ed. London: Earthscan Publications.

Ferguson, P. 1988. Industrial Economics: Issues and Perspectives. London: Macmillan.

Greenwald, B. C., and Stiglitz, J. E. 1986. “Externalities in Economies with Imperfect Information
and Incomplete Markets.” Quarterly Journal of Economics 101: 229-64.

Galal, A., and Shirley, M. 1995. Bureaucrats in Business: The Economics and Politics of
Government Ownership World Bank Policy Research Report. New York: Oxford
University Press.

Hammer, J. S. 1996. Economic Analysis for Health Projects. Policy Research Working Paper
1611. Washington, D.C.: World Bank.

Harberger, A. C. 1954. “Monopoly and Resource Allocation.” American Economic Review 44:
77-87.

Harberger, A. C., 1968. The discount rate in public investment evaluation (Conference
Proceedings of the Committee on the Economics of Water Resources Development,
Waste Agricultural Economics Research Council, Report no. 17, Denver, CO.

A:\PUBLIC3.DOC
 15

Hufschmidt, M. M., James, D. E., Meister, A. D., Bower, B. T., and Dixon, J. A. 1988.
Environment, Natural Systems, and Development: An Economic Evaluation Guide.
Baltimore: The Johns Hopkins University Press.

Lal, D. 1994. The Role of the Public and Private Sectors in Health Financing. HRO Working
Papers HROWP 33. Washington, D.C.: World Bank.

Moore, T. G. 1970. “The Effectiveness of Regulation of Electric Utility Prices. “ Southern

Economic Journal 36: 365-75.

Peltzman, S. 1971. “Pricing in Public and Private Enterprises: Electric Utilities in the United
Sates,” Journal of Law and Economics 14: 109-47.

Sandmo, A. and Dreze, J. H.1971. “Discount Rates for Public Investment in Closed and Open
Economies.” Economica 38: 395-412

Sjaastad, L. A., and Wisecarver, D. L. 1977. “The Social Cost of Public Finance.” Journal of
Political Economics, 85: 513-547

Smith, Adam. 1937 The Wealth of Nations. New York: Modern Library (originally published in
1776).

Stiglitz, J. E. 1984. Pecuniary and Market Mediated Externalities: Towards a General Theory of
the Welfare Economics of Economies with Imperfect Information and Incomplete
Markets. National Bureau of Economic Research, Working Paper Series No. 1304.
Washington, D.C.

Stiglitz, J. E. 1988. Economics of the Public Sector. New York: W.W. Norton

Stiglitz, J. E. 1994. Whither Socialism? Cambridge, MA: The MIT Press

Stiglitz, J. E. 1996. “Some Lessons from the East Asian Miracle.” World Bank Research
Observer 11: 151-177.

Varian, H.R. 1987. Intermediate Microeconomics, a Modern Approach. 3d ed. New York: W.W.
Norton.

Viscusi, W. K. Vernon, J. M. and Harrington, J. E. 1996. Economics of Regulation and Antitrust.

2d ed. Cambridge, MA:. The MIT Press.

A:\PUBLIC3.DOC