Rethinking the state

The important thing for government isnot to do things which individuals are doing

already, and to do them a little better or a littleworse; but to do those things which at present are

not done at all.

-JOHN MAYNARD KEYNES,

"The End of Laissez-Faire"

The agenda for reform that has emerged in thecourse of this Report calls for governments to in-tervene less in certain areas and more in othersfor the state to let markets work where they can,and to step in promptly and effectively where theycannot. In many countries this calls for a strongerorientation toward the market and a more focusedand efficient public sector role. History suggeststhat this is the surest path to faster growth in pro-ductivity, rising incomes, and sustained economicdevelopment.

Judging by their recent activities, many govern-ments in industrial and developing countries havecome around to this view. But economic policycannot be implemented in laboratories; it has to bemade to work in the real world. Reformers face avariety of political constraints on their actions. Inmany developing countries, one of the obstacles toreform has been its political costs, actual or poten-tial. Political instability and other political consid-erations go a long way toward explaining why, inthe first place, many of these countries adopted, totheir economic disadvantage, the policies they did.And they underline the difficulty many countriesface in changing course swiftly. So it is importantto ask whether sufficiently broad support for thesorts of reforms that have been recommended can

128

be built. It has often been argued, for instance,that democracy and structural adjustment do notmix well. Is this true?

Governments also have other objectives in addi-tion to faster economic growth. Employment gen-eration is a related goal. Most think it right to alterthe distribution of income in helping the poor or inimproving equity. How is this best achieved? Dosuch policies serve the goal of faster economicgrowth, or act as an additional constraint? And inthe narrower economic domain, how is the publicsector's performance to be improved? These ques-tions also are all part of reconsidering the role ofgovernment in development.

The political economy of development

Political instability is a fact of life in many coun-tries. The past forty years have seen scores of ra-cial, tribal, communal, and guerrilla wars. Coupsd'etat have occurred in many of the Latin Ameri-can countries (except in Mexico, Costa Rica, and afew of the Caribbean island nations); in manyNorth African and Middle Eastern countries (forexample, Algeria, Egypt, the Islamic Republic ofIran, Iraq, Lebanon, Libya, the Syrian Arab Re-public, and Turkey); and in many parts of Asia andSub-Saharan Africa. Since 1948, there has been atleast one coup attempt per developing country ev-ery five years (Table 7.1).

And there is more to political stability thanmerely avoiding coups. Separatist movements, re-gional rivalries, ethnic frictions, and other some-times violent social conflicts can plague even themost secure executive. Repressive governmentscan create a semblance of stability even when they

entirely lack popular support, as Eastern Europeshowed until recently. In 1987, roughly half of theworld's governments were not democratic (Figure7.1), whereas about three-fifths of nonindustrialcountries fell into that category.

Social consensus helps governments to establishlegitimate authority to govern. Without this au-thority, even the most basic functions such as taxa-tion and allocation of public spending can becomeproblematic. When it began its modernization inthe second half of the nineteenth century, Japan'sper capita income was the lowest among the coun-tries which are today classified as industrialized.However, the country was already politically well-developed, and this was undoubtedly a great as-set. The government was not democratically elec-ted, but it was perceived to be legitimate by thepopulation; it had a strong administration and abroad tax base. All this helped it to undertake itsmajor reforms after the Meiji Restoration in 1868.

As the industrial countries have discovered inthe course of their history, economic moderniza-tion creates new sources of wealth. This can shakethe coalition on which traditional social order wasestablished. A transition of this sort affects manydeveloping countries today. Fragile social con-sensus, entrenched special interests, and weak ad-ministrative capacity have influenced their choiceof economic policies, and the outcomes.

To a large extent, governments everywhere tendto tailor economic policies to balance conflictinginterests. Political rather than economic considera-tions explain why the governments of manyOECD countries intervene in support of ailing in-dustries or regions. The resurgence of protection-ism among OECD countries in the 1980s, the prob-lems encountered in the current round of GATTnegotiations, and the slow pace at which some in-dustrial countries addressed their macroeconomicimbalances in the 1970s and 1980s highlight how,even in societies with secure institutions, much-needed economic reforms can be blocked.

Constituencies and interventions

In many developing countries, political and eco-nomic instability have put the social consensus un-der strain. Such difficulties are hardly new. Formany years they have tended to subordinate eco-nomic policy to the task of securing the support ofinfluential groups for the government. The crudepolicy instruments which many governments of-ten have as a result of these skewed priorities,combined with their often-weak administrative ca-

Table 7.1 Irregular executive transfers: average occurrence per country, 1948-82

Figure 7.1 Nation-states by type ofgovernment, 1850-1987

Number

160

140

120

100

80

60

40

20

0

1850 1875 1900 1925 1950 1975

o Military rule 0 One-party system

o Parliamentary system 0 Presidential systemo Absolute monarchy

Sources: Vanhanen 1979, 1990.

Note: Number of countries considered is in parentheses. Both successful and unsuccessful executive transfers are included. Irregular successfulexecutive transfers are changes in the office of the national executive from one leader to another, outside conventional legal or customaryprocedures for transferring power. Unsuccessful irregular executive transfers are failed attempts at such irregular transfers. Countries are rankedaccording to their per capita income in 1988.Sources: Taylor and Jodice 1983; data base supplied by the Inter-University Consortium of Political and Social Research. Income groupclassifications are from the World Bank.

129

Income group 1948-52 1953-58 1959-64 1965-70 1971-76 1977-82

Low-income 1.0 (21) 1.1 (24) 1.2(39) 1.4(51) 1.3 (53) 0.9(55)Middle-income 1.6(30) 1.7 (32) 1.4(41) 0.8(47) 0.9 (51) 0.6(55)High-income 0.0 (23) 0.2 (23) 0.1 (24) 0.2(25) 0.1 (28) 0.1 (28)

pacity, have compounded the problemand madeit potentially more damaging. Typically, govern-ments have tended to centralize economic re-sources and decisionmaking. This tendency wasreinforced by the prevailing belief in the 1950s,1960s, and 1970s, among many policymakers anddevelopment economists, and sometimes in exter-nal aid and finance agencies, that developingcountries could not rely upon markets and the pri-vate sector alone to develop their industries.

In the 1950s and 1960s, utilities, oil companies,plantations, and assorted manufacturing indus-tries were nationalized in many developing coun-tries, including Algeria, Brazil, Chile, Egypt, SriLanka, and Tunisia. Governments regarded them-selves as too weak administratively at the time totax and regulate private enterprises at arms'length. The nationalization of the major privateBolivian mining companies in 1952 followed de-cades of attempts by governments to tax the fami-lies that owned the mines. An inability to regulateand supervise the banking system led many gov-ernments to nationalize banks or intervene directlyin the allocation of credit. Interventions in agricul-ture have followed a similar pattern. Farming inSub-Saharan Africa, for example, has been highlytaxed through currency overvaluation, state mar-keting boards paying low procurement prices, andexport taxes.

In many developing countries, it is common tofind tariffs, tax incentives, or special regulationsdesigned to protect special-interest groups. Insome cases, "predatory" states have designedpolicies and programs to transfer resources to verynarrowly defined interest groups, and they haveresorted to coercion when the legitimacy of suchpolicies was questioned. The strength of urban in-terests in Latin American and Africa helps to ex-plain why the industrialization strategies adoptedby many countries in these regions were stronglybiased against farming.

Many governments have assumed the role ofemployer of last resort, partly as a result of con-cerns about the social and political implications ofunemployment. Until recently, university gradu-ates had guaranteed employment in governmentin several Sub-Saharan African countries. In theGambia, the civil service doubled between 1974and 1984. The central governments of Argentinaand Sri Lanka reckon that a fifth of their staffs areredundant; Brazil's government puts the figure athalf.

State-owned enterprises (SOEs) have been usedto create employment (though seldom for the

130

poorest), raise incomes in certain regions, or meetthe demands of powerful groups such as the mili-tary. In the 1970s and early 1980s, the Sri Lankangovernment built several textile factories andsugar refineries in backward areas with high ruralunemployment. In Argentina throughout the1970s, industries run by the military were highlyprotected.

Programs of public expenditure have financedthe underpricing of utilitieswater, electricity,telecommunications, railway, or city transportand supported untargeted food subsidies. Moststaple foods in Egypt and Mexico were subsidizeduntil recently, as was wheat flour in Brazil. Thesesubsidies usually benefited the politically activeurban population at the expense of the agriculturalareas where most of the poor live. Uneconomicpublic investment projects are often politically mo-tivated: a very large power project in Zaire, forinstance, was intended to improve the govern-ment's control over an unruly region. In somesuch cases, corruption in the execution of expendi-tures is also a problem (see below), sometimes in-volving foreign suppliers.

The costs

By the 1980s, persistent difficulties in financing ex-ternal and public sector deficits made the costs ofthese interventions plain to see. When the supplyof external finance dwindled after 1982, demandsfor special treatment outran the economy's capac-ity to deliver.

Buying support at the cost of economic effi-ciency was, in the end, self-defeating. Govern-ments reacted to mounting civil service wage billsby letting nominal wages lag behind inflation; thisgenerated the resentment of public servants andled to low morale and poor service. Together withdiscretionary interventions, their situation fueledcorruption. Then, in some countries, corruptionbrought the government down. The hope of em-ployment in government induced migration fromrural to urban areas, aggravating the problem ofurban unemployment. Underpriced and over-staffed utilities meant poor serviceschronicpower cuts, silent telephones, bad public trans-port. This caused further dissatisfaction. Moregenerally, this highly interventionist approachslowed growth, which in many countries againundermined political stability.

Piecemeal interventions also made it harder toestablish essential public institutions. For instance,direct control of the financial system meant there

was no effort to build capacity in banking supervi-sion. High tariffs and the inflation tax made it lessnecessary to broaden tax bases. The expansion ofagricultural state banks, which was intended tomake credit widely available to farmers, made itless urgent to develop cadastres and clarify owner-ship rightsthat is, to address the underlying rea-sons for the high cost of rural credit. Meanwhile,in many developing countries the agriculturalbanks have failed to deliver, so farmers are asbadly off as before.

Market failure and government failure

As the earlier chapters of this Report made clear,intervention by the public sector is not undesirablein itself. On the contrary, many sorts of interven-tion are essential if economies are to achieve theirfull potential. An abbreviated list of indispensableinterventions would include the maintenance oflaw and order, the provision of public goods, in-vestments in human capital, the construction andrepair of physical infrastructure, and the protec-tion of the environment. In all these areas (andarguably more) markets "fail" and the govern-ment must step in. But the countless cases of un-successful intervention suggest the need for cau-tion. Markets fail, but so do governments. Tojustify intervention it is not enough to know thatthe market is failing; it is also necessary to be con-fident that the government can do better.

Governments are prone to fail, at least in eco-nomic terms, for a variety of reasons. As notedabove, economic goals may not be their highestpriority. A combination of political objectives andconstraints and weak administration may leadgovernments to intervene in ways that are eco-nomically harmful. Also, the consequences of eco-nomic interventions are difficult to predict. For in-stance, in the 1950s many Latin Americancountries protected their industries to (amongother things) reduce their dependence on imports.Later it became evident that they had increasedtheir dependence, because the new urban manu-facturing sector that evolved under protection re-lied heavily on imported inputs and machinery.

Private firms are not always better at makingdecisions or predicting their consequences. Buttests of performance are usually clearer to privatefirms, which enables them to take corrective actionfaster. Furthermore, without the help of the gov-ernment, it is harder for private firms to shift thecosts of their mistakes onto taxpayers.

Another difficulty is that government interven-

tion creates vested interests which make it difficultto change the policy. Not all interventions need tobe reversed: investments in infrastructure, for ex-ample, can generate enough resources to covertheir costs. But protection for manufacturing in theearly stages of industrialization can only succeed,if at all, as long as it is temporary. Once protectionis granted, however, it is exceptionally difficult toremove.

Protection creates rents: owners of some labor,capital, or land obtain higher returns than theywould in the absence of intervention. This drawsnew resources to the protected industry until, atthe margin, the rents disappear. Removing theprotection penalizes not only the owners who firstreceived the rents as a windfall, but also those whocame later, seeking normal returns. Removing thetariff, in this case, can force into loss the firms thatgained least. Thus, even when protection has notbeen created by industrial interests, protection cre-ates industrial interests. These then become a for-midable obstacle to liberalization.

Corruption

Excessive intervention breeds corruption. Again,the problem is by no means confined to govern-ments, or to the developing countries. In somecountries, it has grown to alarming and destruc-tive proportions.

Corruption weakens a government's ability tocarry out its functions efficiently. Bribery, nepo-tism, and venality can cripple administration anddilute equity from the provision of governmentservicesand thus also undermine social cohe-siveness. Corruption was identified as a seriousproblem in ancient China and India, in the Otto-man Empire in the fourteenth century, in Englandin the early 1800s. Every other year a scandal is areminder that it continues in Europe, Japan, andthe United States. Corruption has also contributedto the fate of many governments: it was a majorjustification for the military overthrow of theGhanaian civilian government in 1981 and theNigerian one in 1983; an important theme in the1982 Mexican presidential campaign; a major rea-son for the fall of the government in the Philip-pines in 1986; and a problem the authorities con-sider of the utmost gravity in the USSR.

Corruption manifests itself in a variety of ways.A common one is bribery of customs officials, whothen allow in illegal imports, or legal imports atbelow-legal duties, or expedite clearance pro-cedures. This has been a serious problem in nu-

131

Box 7.1 Fighting corruption

Being a tax official in the Philippines Bureau of InternalRevenue (BIR) in the early 1970s was so lucrative thatjobs and transfers to the bureau were sold. Manila'smost "extensive, expensive, and lavish assortment ofcars" was in the BIR's parking lot. Then-president Fer-dinand Marcos's New Society, announced in 1972,aimed to alleviate poverty and fight corruption, a fightthat intensified in 1975 when 2,000 government offi-cials suspected of improper conduct were fired. In thatsweep, the BIR's commissioner was replaced by JusticeEfren Plana.

Problems

After a few months, Plana identified a number of se-rious problems. Chief among these were practiceswhereby officials would require payments to process atax matter, provide a record, or make a routine clear-ance; accept bribes to lower tax assessments or stop theharassment of taxpayers with no tax obligations; em-bezzle funds; illegally print fiscal labels and stamps;succumb to cash, nepotism, and influence for person-nel decisions, such as transfers and appointments; andbreak down the internal auditing systems (officials incharge of investigating others routinely accepted bribesfrom those being investigated).

The bureau was virtually free of corruption whenPlana left it in 1980 to become deputy minister of Ii-

nance and, shortly thereafter, a justice of the supremecourt.

Solutions

Plana's success was based on six innovations. First,supervision and auditing were improved by using agroup of highly skilled outsiders teamed with irre-proachable incumbent senior officials. Second, admin-istrative systems were introduced to monitor perfor-mance on the basis of objective criteria such as thenumber of tax assessments and taxes collected. Third,about 100 high-level corrupt agents were punished bybeing dismissed or reorganized. Fourth, tax laws weresimplified to make them more efficient and reduce thediscretion left to tax officials. Fifth, control systemswere tightenedtax payments began to take placethrough banks rather than tax agents, and confirma-tion letters were sent to check tax-payers' payments.Sixth, personnel practices were improved. Recruitmentbecame meritocratic, an antinepotism regulation for-bade the appointment of even distant relatives, andpromotions were based on performance. But theseachievementsin a country where corruption re-mained widespreaddid not last. In the early 1980s,nepotism became a problem once again, and tax assess-ments and tax collections dropped significantly.

merous countries: in the United States at the turnof the century, in Singapore in the 1960s, in Indo-nesia in the 1970s, and in Cameroon in the 1980s.Police indulgence of extortion and other crimes inHong Kong led to the creation of an anticorruptionoffice in the 1970s. In the late 1970s, an inquiry inMassachusetts revealed that 76 percent of a sampleof public buildings had at least one "structural"defect that could not have occurred without in-spectors' complacency. Two-thirds of the nameson the civil service roster in 1978 in Zaire werefictitious. These and less malign forms of corrup-tionabsenteeism, moonlighting, or lack of dedi-cationundercut public administration.

Corruption can seldom be reduced unless itslarger underlying causes are addressed. It flour-ishes in situations where domestic and interna-tional competition is suppressed, rules and regula-tions are excessive and discretionary, civil servantsare underpaid, or the organization they serve hasunclear or conflicting objectives. In Cameroon, ob-taining all the authorizations and permits neces-sary to start a new business takes two years even

132

for a well-connected businessman; the law re-quires twenty-four different steps involvingtwenty separate offices. Anticorruption campaignsare periodically undertaken, sometimes with suc-cess (Box 7.1). But often the root causes remain:weak agencies fighting market forces with controlssociety considers excessive, discretionary, orillogical.

Remedies: democracy and institutions?

Authoritarianism often has been seen as a useful,if regrettable, expedient for effective policymakingin the face of political instability. A strongly heldview from the 1950s through the 1970s was thatdevelopment policies took time to bear fruit, andthat this was inconsistent with the politics ofshort-term electoral cycles. Democracies were seenas having a built-in inclination toward populistpolicies (Box 7.2). Benevolent authoritarian reg-imes (led by philosopher-despots) were needed, itwas argued, to push through unpopular reformsand tame an unruly or otherwise ineffective ad-

ministration. Economies managed with varyingdegrees of authoritarianism have made progress atdifferent times in the past, for example, Brazil,Chile, Spain, and some of the East Asian econ-omies. Yet at the same time, some democraciesold ones as in India or new ones as in the Philip-pineshave been unable so far to make rapidprogress.

During the 1980s, however, severe disenchant-ment with authoritarian regimes set in. Now it isbetter understood that such regimes are no lesslikely to yield to the interests of narrow constituen-cies. Few authoritarian regimes, in fact, have beeneconomically enlightened. Some of the East AsianNIEs are the exceptions, not the rule. Dictator-ships have proven disastrous for development inmany economiesin Eastern Europe, Argentina,Central African Republic, Haiti, Myanmar,Nicaragua, Peru, Uganda, and Zaire, to name onlya few.

Democracies, conversely, could make reformmore feasible in several ways. Political checks andbalances, a free press, and open debate on thecosts and benefits of government policy could givea wider public a stake in reform. The need to pro-

duce good results in order to be reelected couldhelp, rather than hinder, economic change: it in-creases governments' incentives to perform welland keeps predatory behavior in check.

Experience allows no hard and fast conclusion.Peru is going through one of the worst economiccrises in its history, mostly as a result of policiesimplemented in the late 1980s by a democraticallyelected government. Bolivia has been unable toimprove its government's administrative capacitydespite almost a decade of democracy. Literacyrates in China in 1950 were similar to those in In-dia, and four decades later they are twice as high.Yet India is one of the oldest and most sophisti-cated democracies in the developing world.

Democratic governments are not necessarilymore adept at managing reform, either. Transi-tional democratic governments, perhaps becausetheir political base is still fluid, appear to be par-ticularly vulnerable (Tables 7.2 and 7.3). Demo-cratic governments have a better record than au-thoritarian governments in countries that are notpolitically polarized; the reverse seems to be truein polarized societies. On the whole, the evidencesuggests that the democratic-authoritarian distinc-

Box 7.2 Populist experiments

The populist experiments in Latin AmericaAllende inChile (1971-73), Peron in Argentina (1946-49), andGarcia in Peru (1985-88)are extreme examples of theinteraction between political and economic processes.Populist policies have emphasized growth and short-run distributional goals, brushing aside the risks of in-flation and excessive deficits and ignoring external con-straints and the responses of firms and households totheir aggressive anti-market policies. Addressing pov-erty and income distribution issues, which populistregimes viewed as the source of social conflict and po-litical instability, could not be done, however, throughunsustainable economic policies.

In a typical populist cycle, the new administrationsets in motion a marked shift in policies. Excess capac-ity and the availability of foreign reserves at first sup-port higher output growth, which in many cases isaccompanied by an increase in real wages. Inflation iskept low with the help of price controls. But bot-tlenecks soon appear as a result of the strong expan-sion in domestic demand; because of dwindling for-eign reserves, these cannot be bypassed by increasingimports. Shortages, accelerating inflation, and declin-ing reserves lead to capital flight and the demonetiza-tion of the economy. The budget deficit worsens as

subsidies increase, and taxes decline in real terms. Inthis unsustainable position, the government is forcedto devalue the currency and cut subsidies. Inflationaccelerates and real wages fall.

The Chilean experience of 1970-73 clearly illustratesthis sequence of events. To achieve rapid growth andimprove the living conditions of low-income groups,the government stepped up public spending. Publicsector wages were increased, adding to the fiscal defi-cit. Agrarian reform was intensified, and the miningand banking sectors as well as parts of the industrialsector were nationalized. The combination of price con-trols and expansionary demand policies fueled re-pressed inflation; the parallel market flourished. For-eign reserves were so low that it was impossible tomeet the surge in demand by increasing imports. By1972, the government was forced to devalue the escudoand adjust public sector prices. It was unable to controlwages, however. Between 1970 and 1973, inflation in-creased from 35 percent to about 600 percent a year,and the fiscal deficit jumped from 2.7 to 24.7 percent ofGDP. GDP growth accelerated to 9 percent in 1971 butturned negative in 1972 and 1973, when output fell by5.6 percent.

133

Table 7.2 The success of economies withdiffering political systems in implementingan IMF adjustment program(percent)

Note: Based on reform episodes in seventeen countries from the1950s through the 1980s.Source: Haggard and Kaufman 1990.

Table 7.3 The success of economies withdiffering political systems in controllingrapid inflation

Percentage of inflation episodeswhich ended in stabilizationIn nonpolarized environments 75 62In polarized environments 29 67

Percentage of adjustment programsthat led to breakdown of systemtwelve months or less afterprogram started 11 14

Note: Based on 114 standby arrangements from 1954 to 1984 signedby nine Latin American countries.Source: Remmer 1986.

tion itself fails to explain adequately whether ornot countries initiate reform, implement it effec-tively, or survive its political fallout.

But as indicated in Chapter 2, there is suggestiveevidence that links features of democratic systemspositively with overall aspects of development andwelfare. A further result emerges from the empiri-cal literature on the relation between economicperformance and political systems: by developinghuman resources and, more particularly, byinvesting in education, countries have been foundto strengthen the basis for open political systems.Some studies suggest that for a given level of in-come, improvements in social indicators are asso-ciated with freedom and liberty. Other studiessuggest that political instability declines not onlyas income rises but also as education improvesalthough further research is necessary to confirmthis finding.

134

Institutions and development

Another approach to the problems of political in-stability, fragile social consensus, and weak gover-nance is to build more effective institutions. This isan extremely broad concept. It encompasses thepublic bodies through which the state dischargesits most fundamental responsibilities: maintaininglaw and order, investing in essential infrastruc-ture, raising taxes to finance such activities, and soon. But the idea goes further. It extends to theconventions that govern the way people deal witheach other: property rights, contracts, and normsof conduct. The discussion of how society's insti-tutions affect economic performance has been oneof the liveliest in the economic literature in the pasttwo decades. Although understanding of these is-sues is far from complete, it is clear that a primarytask of institutional development is to improve al-locative efficiency and reduce transactions coststhe costs of people dealing with people (Box 7.3).

People's values and ideologies affect institu-tions, and these in turn affect the economy. An-alyzing the role in development of such factors asculture, religion, law, and politics has a strong in-tellectual foundation in the work of Hayek, Hegel,Marx, and Weber. Centralized political institutionsbacked by a strong bureaucracy are argued to havestifled entrepreneurship and productivity growthin ancient Chinaeven though technologically thecountry was far ahead of what is now the West. Atthe level of organizations, recent research suggeststhat the superior performance of Japanese manu-facturing results (among other factors) from normsof behavior that promote the flow of informationbetween workers and supervisors; these lowerfirms' internal transactions costs and help themadapt to markets demanding high-quality prod-ucts with short life cycles. Another study hasfound that when workers in the United States geta share of their firm's profits, it seems to have afavorable effect on their productivity.

Often, the institutions of government can affecteconomic performance more directly. Fiscal defi-cits have led to very high inflation in Latin Amer-ica but not in South Asia, where central banks aremore independent. Credit programs for small andmedium-size industries have been much more suc-cessful in Sri Lankawhere they have been imple-mented by a competent and motivated civil servicerelatively independent of political interferencethan in Bangladesh. For the same reason, ruraldevelopment programs have enhanced produc-tivity in some parts of South Asia, but less so in

Democratic AuthoritarianMeasure systems systems

Continuous Continuous TransitionalPercentage of democratic authoritarian democraticadjustment years systems systems systems

In which fiscaldeficits fell 49 50 25

In which expendi-tures as per-centage ofGDP fell 38 46 29

In which creditexpansionslowed 61 62 43

Box 7.3 The contribution of institutional innovations to development

Over the centuries, market-mediated transactions havebeen a major force in institutional development, whichin turn has been a major force in economic develop-ment. As markets have expanded, market participantsspontaneously have defined rights, formulated con-tracts, and evolved norms of behavior with a view toimproving the efficiency of their interactions.

The letter of credit, a contract that emerged in theMiddle Ages in Italy, increased the scope of exchangeand contributed to the expansion of internationaltrade. By better defining creditors' rights in regard to afirm's assets, public liability companiesan innovationin late-eighteenth-century Englandallowed firms totake risks and attract resources to activities that other-wise could not have developed. Since the 1970s, leas-ing contracts have allowed enterprises to reduce therisks associated with large investments in equipment.In Bangladesh, the Grameen Bank found innovativeways to lend to low-income groups while keeping de-faults low. This was achieved by establishing contracts

that made the community, not only the borrower, re-sponsible for payments.

Behavior also adapts to market needs and influencestransactions costs. Stealing and trading have the samelinguistic root in various languages because of the dis-honesty of early traders. Only after markets becomeestablished, transactions become regular, and competi-tion increases do traders have an incentive to establishand maintain their reputation. Traders in industrialmarket economies are more honest not only becausesanctions are administered more efficiently, but alsobecause a good reputation reduces transactions costs.

Norms of behavior not yet adapted to the needs of amodern economy substantially increase transactionscosts. Pilferage is serious in the ports cities of manydeveloping countries partly because stevedores aremore loyal to their families, clans, or tribes than to theorganization employing them. Not pilfering and beinghonest deprives their families of a source of additionalincomebehavior families would consider dishonest.

Africa arid Latin America. State-owned enterpriseshave been efficient in Singapore and Taiwan,China, where they were subject to competitionand their access to the budget was restrictedbutnot in Argentina, Bolivia, and Nigeria.

In many instances, the state has stimulatedgrowth by restructuring institutions: the abolitionof feudal arrangements and the standardization ofcurrency, taxes, weights and measures, and inter-nal tariffs in revolutionary France in the 1790s; pa-tent laws in nineteenth-century Europe and theUnited States; the integration of customs, com-mercial, and civil and commercial law in both Ger-many and Italy in the nineteenth century; themodernization of Meiji Japan in the second half ofthe 1800s, and that of Turkey in the early part ofthis century; Brazil's company-law reforms in theearly 1970s; the creation of stock exchanges in EastAsia and the economic integration of Western Eu-rope after 1945. All of these depended on stateaction. They molded the framework of enterprisein ways which increased entrepreneurial securityand eased the flow of resources and people. Inmost developing countries, strengthening or creat-ing institutions remains a difficult but necessarytask (Box 7.4).

Supporting institutional development requires astate with well-developed administrative struc-tures and agencies responsive to markets' needs.

The political weaknesses of developing countriesare often, however, manifested in the efficiency oftheir bureaucracies. By itself, an efficient bureau-cracy does not guarantee successful development,nor can it substitute for market forces. As indi-cated above, it can even retard development.Nonetheless, an efficient bureaucracy enables gov-ernments to govern. It was key to the survival ofancient civilizations such as Egypt (3000 B.C.) andChinawhere the well-structured bureaucracywhich had existed since at least 200 B.C. was stilloperating less than a hundred years ago. The basicprinciples of bureaucracy were already well under-stood by the ancient Chinese. The civil servants,the mandarins, were recruited by competitive ex-amination. There were systems of promotion, ca-reer patterns, and job security. Serving the statewas a privilege reserved only for those with dem-onstrated talents. Building efficient bureaucracieswas also an essential step in the process of nation-making in Europebut remains a priority in manydeveloping countries.

Nongovernmental organizations

Nongovernmental organizations (NGOs) have be-come an important force in the development pro-cess which, to some extent, has mitigated the costsof developing countries' institutional weaknesses,which often include administrative shortcomings

135

Box 7.4 Setting priorities for institutional development: easier said than done

The priorities for institutional development naturallyvary with a country's history, culture, economic poli-cies, and stage of development. For most of EasternEurope, the priority is to establish the institutions nec-essary for a market economy to function efficiently:property rights, corporate and bankruptcy laws, com-mercial courts, banking legislation, and stock ex-changes. For low-income Africa and Latin America, thepriority is to improve the management of the publicsector, a goal that often requires a simultaneous reduc-tion in the size of the government and a strengtheningof its capacity.

Elsewhere priorities may be less clear-cut. Particularcountries have their own accomplishments and needs:

In South Asia and some parts of Latin America,training and visit programs have had a strong effect onagricultural productivity.

In Sri Lanka, a recent change in civil courts pro-cedures has greatly improved the workings of bank-ruptcy laws and reduced financial intermediationcosts, after several years of complaints from the bank-ing community.

In Brazil, mechanisms are being devised to im-prove the flow of information among universities, re-search institutes, and industrymaking research moreresponsive to industry's needs.

In Malaysia, a recently created government bond-rating system is expected to reduce private firms' fi-nancing costs significantly.

In northern Brazil, Egypt, India, Indonesia, andSri Lanka, the improvement of cadastres and land tit-ling is overdue and could greatly improve the effi-ciency of rural credit markets and reduce the generallyextremely high costs of rural credit.

In many countries, better banking supervision isimportant for successful financial liberalization.

Identifying institutional needs is not easy, however.First, institutions essential in industrial societies mayprove superfluous in developing countries. Stock ex-changes, treasury-bill markets, credit-rating bureaus,land-titling offices, and metrology and standards bu-reaus are expensive to set up, and it is difficult to de-cide whether they are being developed ahead of mar-ket needs. Second, some institutions are unproductivein the presence of systemwide problems. For instance,an underpaid civil service renders most public institu-tions a hindrance rather than a help to markets. Poorlyplanned public spending deprives institutions of cur-rent inputs and reduces their efficiency. Third, thereare no simple indicators of institutional needs and pri-orities. There is scope, however, to develop quantita-tive indicators of the efficiency of public institutions;for example, how long does it take to register a busi-ness, obtain a passport, clear customs, get an importlicense, or pay taxes?

and an inability to carry out efficiently essentialdevelopment tasks, such as providing social ser--vices or protecting the environment. In response,NGOs have grown rapidly in recent years, both innumbers and in the volume of resources they mo-bilize. In 1987, NGOs transferred about $5.5 billionfrom industrial to developing countriesnearly $1billion more than the International DevelopmentAssociation.

Most of NGOs' resources (about 60 percent) areraised by themselves. The rest ($2.2 billion in 1987)are from official aid agencies, which channel fundsthrough NGOs because such organizations aremore effective in bringing about popular participa-tion, in working at the grassroots level, and inoperating in remote areas. NGOs have also beenimportant in sensitizing governments and interna-tional aid and finance agencies to the social andenvironmental aspects of development. In addi-tion, in many countries, they have taken the leadon controversial development issues such as fam-

136

ily planning. Although many developing-countrygovernments are suspicious about some NGOs'self-appointed role as agents of change, govern-ments of countries such as Bolivia, Egypt, India,Jordan, Mexico, the Philippines, Togo, andUganda are seeking ways to encourage more NGOaction.

NGOs vary in coverage and effectiveness. InBangladesh, NGOs specialized in health and fam-ily planning reach only one-sixth of the country's80,000 villages. Many small NGOs' managerial ca-pacity needs to be developed before they can beeffective. For others, little is known about fund-raising costs. In addition, even the most effectiveNGOs cannot fulfill all the gaps left by the com-mercial and public sectors. Aside from their grow-ing numbers and the volume of resources they mo-bilize, the importance of NGOs lies in their abilityto involve communities and grassroots organiza-tions more effectively in the development processand in addressing poverty.

Equity and redistribution

Governments have always been concerned withequity. Income transfers in OECD countries (ex-cluding interest payments but including social se-curity payments) amount to 40 percent of govern-ment expenditure and are as high as 20-30 percentof GDP in Austria, France, Germany, the Nether-lands, and Sweden. A better distribution of in-come may facilitate economic management. Politi-cal scientists have suggested that mechanisms toredistribute income by sharing the benefits ofgrowth more equally have helped some OECDgovernments to diffuse opposition against market-oriented reforms; the short-term victims of changeare cushioned.

An analysis of thirty-two countries (twenty-fivedeveloping and seven OECD countries) showedthat the higher the risk of term of trade shocks thata nation faces in international markets, the morelikely it is to increase trade barriers. It also showedthat the larger its social-insurance programs, theless likely it is that the government will be protec-tionist (Bates, Brock, and Tiefenthaler 1991). Otherrecent research suggests that wage negotiationsthrough nonmarket mechanisms (negotiations be-tween unions, industrialists, and governments)that take equity into account may explain the rela-tively low unemployment in the Nordic countries(Jackman, Pissarides, and Savouri 1990). Someeconomists have also suggested that the relativelyegalitarian distributions of income in Asia allowedcountries there to adjust to the external shocks ofthe 1970s more rapidly than their Latin Americancounterparts.

Despite such evidence, greater equality of in-come is still considered by some to be inimical togrowth. Increasing the capital stock, it is argued,requires high saving rates; this in turn implies adistribution of income that is tilted toward the(high-saving) rich. The Republic of Korea's tax re-form of 1973 largely excluded capital income (inter-ests, dividends, capital gains, and other returns onassets) from the tax base. The conventional wis-dom among industrial countries as well as policy-makers in developing countries has been thatthings ought to be done "one at a time": first,economic growth; second, social equity; third,civil and political liberties.

In fact, there is no evidence that saving is pos-itively related to income inequality or that incomeinequality leads to higher growth. If anything, itseems that inequality is associated with slowergrowth (Figure 7.2). The notion of a trade-off be-

Figure 7.2 Income inequality and the growthof GDP in selected economies, 1965-89

GDP growth per capita (percent)

8

7

6

5

4

3

2

0

-2

Rep. of Korea- % SingaporeIhiwan,

S Hong Kong- IndonesiaChina S Thailand

Yjigoslavia

Hungar7Malaysia

-Egyp

- \ Tunisia S BrazilS Zortugai,/ S

ISSyna S- Sri LankaIsrael5 Turkey

____-Colombia Bcuador

Spain MexicoS SIndia S Kenya S Panama

Costa RicaUruguayS- Bangladesh

Philippines

S %.__ S Cote d'IvoireTrinidad and

- Chile" Tobago S PeruArgentina

S Venezuela

I I

0 5 10 15 20 25 30

Income inequalitya

35 40 45

a. The ratio of the income shares of the richest 20 percent andpoorest 20 percent of the population. Data on income distributionare from surveys conducted mainly in the late 1960s and early 1970s.Sources: World Bank data; Berg and Sachs 1988.

tween growth and equity, which helped to en-trench antigrowth policies in socialist economiesand antiequity policies in conservative ones, hasbeen further discredited by the many economiesthat consistently outperform the rest on bothcounts: Costa Rica, Indonesia, Japan, Korea, Ma-laysia (Box 7.5), and the Scandinavian economies.

Greater equality is not achieved through incometransfersexcept in the case of safety nets for vul-nerable, small, and well-targeted groups of thepopulation. World Development Report 1990 showedthat the pattern of development has strong distri-butional implications. Industrial protection anddiscriminatory taxes on farming help to explainwhy income inequality is more severe in LatinAmerica than in Asia. The bulk of developingcountries' revenues generally consist of indirecttaxes, which are generally less progressive thanincome taxes. Subsidies for capital (in the form oftax incentives, subsidized credit, or currency over-valuation) invariably lead to more capital-intensivemodes of production, and thus worsendistribution.

Another lesson is that public expenditure canhave powerful redistributive effects. Various

137

Box 7.5 The politics of inclusion: Malaysia and Sri Lanka

Similar starts

Both Malaysia and Sri Lanka were British coloniesuntil 1963 and 1948, respectively. Both had well-devel-oped, export-oriented, tree-crop plantations in the1960srubber and palm oil in Malaysia, rubber and teain Sri Lanka. Both had sophisticated bureaucracies.Both had advanced democratic political institutions.And both had relatively well-educated populations,with 90 percent primary school enrollment. Both coun-tries also had problems caused by the presence ofhighly differentiated ethnic groups, with a majoritythat was economically underprivileged but politicallydominant. In Malaysia, Bumiputras (Malays and otherSons of the soil) accounted for 55 percent of the popula-tion, Chinese 35 percent, and Indians 10 percent. In SriLanka, Sinhalese accounted for 72 percent of the popu-lation, Tamils 18 percent, and other groups 10 percent.Both countries adopted discriminatory policies specifi-cally to improve the lot of the majority groups (legis-latively in Malaysia and de facto in Sri Lanka).

Both countries used public enterprises not only inthe plantation sector, but also in other areas such asairlines, cement, banks, and manufacturing. They sup-ported rice farmers through subsidized fertilizer,credit, and irrigation. They gave preferred access topublic employment and public procurement to the ma-jority ethnic group. And they emphasized educationand health services for allbut biased higher educationin favor of the majority.

Different results

In the early 1960s, Malaysia's per capita income (at$320) was twice Sri Lanka's. Three decades later, Ma-laysia's per capita income is five times Sri Lanka's.Malaysia has also contained conflict among its ethnicgroups without serious violence. By contrast, since1983, ethnic and regional conflicts in Sri Lanka have

-w a

claimed tens of thousands of lives. The cost of de-stroyed infrastructure and forgone income from dis-rupted economic activities has been estimated to beclose to two-thirds of GDP. In addition to its superiorgrowth, Malaysia has reduced the incidence of povertyfrom 50 percent in 1970 to 10 percent todayand re-duced the inequalities between and within ethnicgroups.

Reasons for the difference

Unlike Sri Lanka (until 1977, when the economy wasliberalized) Malaysia's authorities adjusted the anti-growth elements of their policiessuch as foreign in-vestment and industrial licensing ruleswhen growthrates faltered. Trade policies were kept open, withmoderate tariff levels (although, in selected importantcases, highly effective protection was retained). Privateenterprises did not need permits to expand productionor invest. Nor were they harassed by currency con-trols, extensive quantitative trade restrictions, or thethreat of nationalization without compensation. Minor-ity businesses that were discriminated against in do-mestic markets were thus not shut out from businessopportunities abroad. They could use their income tobuy abroad goods and services (such as education) thatwere being denied them at home.

Sri Lanka's heavy regulatory framework before 1977gave ample opportunities for discretion and discrimi-nation. Economic controls ended up becoming controlson individualsdespite Sri Lanka's democratic tradi-tions. Travel was restricted because of exchange con-trols, and simple business transactions (such as obtain-ing a permit to invest, to import, or to expandproduction) often ended up being highly politicized.The perception was that government could influence,and was influencing, the distribution of assets amongethnic groups.

studies have found that education is the most im-portant single variable influencing income in-equality. Investments in education, health, andnutritionif well-designed and -implementedcan improve distribution and at the same time pro-mote development in other ways. The reform pro-grams of the 1980s and 1990s thus have empha-sized more and more the need to protect socialprograms during fiscal adjustment.

When markets work well, greater equity oftencomes naturally. For instance, labor markets arefragmented in many countries. People with similarattributes are unable to obtain similar rewards oremployment: sex, ethnicity, location, and indus-trial occupation consistently appear as determi-

138

nants of wages, regardless of productivity. Help-ing women to participate in labor markets hasbeen an important reason for the improving distri-bution of income in Malaysia and Indonesia. Gov-ernment spending on improving infrastructureand the delivery of social services has traditionallybeen the main mechanism to integrate markets,and this remains of major importance. A variety ofother public programs that can reduce inequalitieswhile improving allocative efficiency and spurringgrowthfor example, programs designed to im-prove access to infrastructure, credit, and land.

Land reform often seems to have raised the in-comes of the poor. China, Japan, and Korea are allregarded as outstanding examples of economies

that have succeeded at land reform. The evidenceon its effect on agricultural efficiency is moremixed, however. It is hard to separate the effects ofland redistribution from the effects of the comple-mentary investments and institutions oriented to-ward increasing agricultural productivity that havetypically accompanied land reform. There does ap-pear to be evidence, however, that the social stabil-ity resulting from land reform has contributed tofaster growth.

For all these reasons, efforts to improve equitycan sit comfortably within reform programs aimedat promoting growth. It is clear, however, thatmarket-distorting and overzealous redistributioncan quickly pose overwhelming financial prob-lems. For example, the cost of food subsidies inBrazil in the late 1970s, and more recently inEgypt, ballooned as international food prices wentup. Subsidies to protect declining industries haveto rise continuously to achieve the same effect, be-cause maintaining preference requires that dy-namism elsewhere in the economy must be offset.In Europe, for instance, maintaining farm incomesrelative to other incomes has become more andmore expensive because of faster growth in othersectors.

Also, crude transfers through market-distortinginterventions almost always end up worsening thedistribution of income rather than improving it.Fertilizer subsidies in Bangladesh, Brazil, Ecuador,Egypt, India, and Pakistan accrue mainly either tofertilizer manufacturers or to better-off farmers.The large subsidy on wheat in the 1970s in Brazilreduced the demand for Leans grown by smallfarmers. Production of beans declined. Farmerssold their land and migrated to the cities, wherethey increased the demand for subsidized wheat.Rich commercial farmers bought the migrants'land at distress prices.

Reforming the public sector

In about the fourteenth century, an Arabic treatiseargued: "Commercial activity on the part of theruler is harmful to his subjects and ruinous to thetax revenue . . . crowds out competitors; dictatesprices for materials and products which could leadto the financial ruin of many businesses. When theruler's attacks on property are extensive and gen-eral, affecting all means of making a livelihood, theslackening of business activity too becomes gen-eral" (Ibn Khaldun 1981). One of the most strikinglegacies of the 1980s is the rediscovery of theseancient truths. Many governments are reconsider-

ing their involvement in the economy, reviewingtheir spending priorities, and undertaking fewercommercial activities. For this reappraisal to suc-ceed, the administrative capacity of the state willneed to be improvedand governments will haveto cope with opposition from the vested interestscreated by decades of excessive intervention.

Rationalizing public expenditure

Government expenditure accounts for slightlymore than 20 percent of GDP in low-income coun-tries, and close to 30 percent in middle-incomeones. These ratios are much lower than in indus-trial countries today, but much higher than in in-dustrial countries at a comparable stage of devel-opment (Tables 7.4 and 7.5). The evidencesuggests that many of the developing countries'public spending programs provide very lowreturns.

PuBLIc INVESTMENT. The quality of public in-vestment significantly depends on the quality ofthe economic climate (see Chapter 4). But somedeveloping countries are experiencing economicdifficulties because the projects themselves, veryoften financed with the support of external agen-cies, were ill-advised. A loss-making silver-smelt-ing plant in Bolivia, a value-subtracting shoe fac-

Table 7.4 Percentage share of governmentexpenditure in GNP or GDP, industrialcountries, 1880-1985

GNP.GDP.

Source: World Bank, various years.

Table 7.5 Percentage share of governmentexpenditure and consumption in GNP orGDP, industrial and developing countries,1972 and 1986

Year France Germany Japan SwedenUnited

KingdomUnitedStates

1880a 15 10 11 6 10 81929a 19 31 19 8 24 101960b 35 32 18 31 32 281985b 52 47 33 65 48 37

Expenditure Consumptionb

Economy group 1972 1986 1972 1986

Low-income 19 23 12 13

Lower-middle-income 15 27 11 14

Upper-middle-income 25 27 12 14

Industrial market 28 40 14 19

a. GNP.b. GDP.Source: World Bank, various years.

139

tory in Tanzania, and irrigation systems with lowrates of return in Sri Lanka are only a few of count-less possible examples. The costs can be consider-able. In Zaire, the hydropower and transmission-line project that was mentioned earlier in thischapter cost almost $3 billion in 1990 pricesabouta third of the country's external debt. The projecthas never operated at more than 30 percent of itscapacity, and it is now in the midst of extensiverehabilitation, although it began operating only in1982. This is an extreme case, but unproductiveprojects on a less spectacular scale are all toocommon.

WAGES AND THE CIVIL SERVICE. Wage bills are a

large part of government expenditure in mostcountries. Before the reform programs, the wagebill absorbed more than 60 percent of current reve-nues in the Central African Republicand morethan 40 percent in the Gambia. The tendency tooverstaff and underpay that in the last few dec-ades has prevailed in many developing countriesmeans that much of this spending is wasted. Theproblem of poor motivation has often been com-pounded by ill-defined career structures, and bypoliticized recruitment and senior appointments.In some countries the institutional structures andsystems originally established to staff and operatethe civil service have collapsed. In Uganda, a civilservice census turned up not only numerous non-existent workers but also entire nonexistentschools.

As a result, in Latin America, South Asia, andAfrica, civil service reform has become a high pri-ority for many governments. Civil service reformprograms generally have three components. Thefirst is a retrenchment effort to downsize the civilservices to more manageable numbers of em-ployees. The second is pay and grading restructur-ing to increase incentives, reduce moonlightingand corruption, and provide a better frameworkfor career development. The third is institutionalrebuilding to create the control structures and op-erating procedures needed to manage a modernand efficient civil service.

Most civil service reform programs have movedon all fronts simultaneously. The more successfulAfrican programs have cut back on the numbers ofpublic employees (Ghana, the Gambia, andGuinea). But their progress has been limited toimproved pay structures and some reform of insti-tutional structures. No African country's ongoingreform program has completely rebuilt its civil ser-vice structures. The Ghanaian program, in opera-

140

tion since 1985 and probably the farthest-reaching,has yet to establish an effective system to controlrecruitment.

None of the programs now under way appearsto embrace a serious examination of governmentfunctions to determine which can be privatized,delegated to the local community, or stopped alto-gether. Given the need for smaller, more efficientpublic sectors and a more dynamic private sector,future civil service reform efforts would definitelybenefit from tackling such larger issues.

SUBSIDIES AND TRANSFERS. Expenditure on sub-sidies and transfers accounts for about 3 percent ofGDP on average for a large sample of countries. Itis difficult to generalize about them because theyare one of the most heterogeneous categories ofspending. Moreover, reporting systems are weakin most countries. The real cost of subsidies andtransfers could easily be twice what is on thebooks. Subsidies often result from government in-terventions in prices, and they may apply to allmanner of goods and services: wheat in Egypt andthe Soviet Union; bus travel in Sri Lanka; fertilizerin Bangladesh, Ecuador, and India; and so on.Transfers are usually made to state-owned enter-prises. They become necessary either because theSOEs are inefficient, or because price controls andother restrictions force them to operate at a loss.These transfers, however, are generally insuffi-cient to meet the enterprises' needs for capital in-vestment; as a result, standards of service havedeteriorated dramatically in some countries. Thetelephone system and railways in Argentina andthe bus service in Egypt, for instance, have suf-fered from far too little investment.

MILITARY SPENDING. The world spends $1,000billion on the military every year. In the late 1980s,military expenditure totaled $860 billion a year inhigh-income countries and $170 billion a year indeveloping ones. Of this $170 billion, $38 billionwas spent on imports of arms, mostly from indus-trial countries.

If global military expenditure were reduced, theworld would undoubtedly be a better place. But isthis realistic? Humanity is no stranger to wars andconflictstwentieth-century humanity least of all(Box 7.6). The recent war in the Gulf region; ensu-ing conflicts there; continued violence in Af-ghanistan, Angola, Central America, and Indoc-hina; civil wars in Ethiopia, Mozambique,Somalia, and Sudan; and the snail's pace of super-power disarmamentall make it only too clear

how difficult progress toward lasting peace willbe.

Unsurprisingly, military spending is higher inthe developing countries that face external or in-

ternal threats. Military spending is more than 10percent of GDP in several countries. After the eth-nic conflict that erupted in 1983, Sri Lanka's mili-tary expenditure has increased from less than 1

Box 7.6 War and development

The two world wars involved unprecedented numbersof nations and resulted in unprecedented loss of life.But regional wars and civil upheavals since 1945 havealso claimed lives and have devastated individualcountries, many of them in the developing world (Boxtable 7.6). The commonly reported estimates of 450,000deaths in the Iran-Iraq conflict equal about 1 percent ofthose countries' combined populations at the start ofthe conflict in 1979. The 2 million losses in the Ethio-pian Civil War constitute more than 7 percent of thecountry's 1974 population.

Battlefield death tolls underestimate war's impact.War makes heavy claims on the most productiveworkers. In World War I, only 4.5 percent of Ger-many's fatalities were more than forty years old; 63percent were between ages twenty and thirty. In addi-tion, soldiers are not the only ones to die. Civilianssuccumb directly fighting and from war-related famineand disease; military mobilization results in lower birthrates. Totaled this way, the loss of life during the period1914-21 (which includes the Soviet civil war) may ex-ceed 60 million. Only about 8 million of these weremobilized men.

The costs of fighting involve more than the cost ofbullets, uniforms, and equipment. The 1969 SoccerWar between Honduras and El Salvador lasted just 100hours. About 2,000 people died. But 100,000 peoplebecame refugees. The fighting destroyed half of El Sal-

Box table 7.6 Deaths during wars, 1900-89

vador's oil refining and storage facilities and paralyzedthe Central American Common Market. Military ex-penditure and forgone output during the first fiveyears of the Iran-Iraq conflict cost more than $400 bil-lion. The cost by the end of the war in 1988 was muchhigher. Economic disruption is similarly severe in civilwars. The conflict in northern Ethiopia's Eritrea terri-tory has cut the labor force; bombs and mines havecaused farmers to avoid some land, thus effectivelytaking it out of production-40 percent of land wasestimated to have been left idle in 1987, which partlyexplains the food shortfalls in the region.

Inevitably, war retards development. The combinedcosts of replacing lost equipment, providing healthcare for the wounded, and enduring lower produc-tivity are paid long after armistice. In the Nigerian CivilWar of 1967-70, the government sought to finance thewar without triggering high inflation or causing deteri-oration in the balance of payments. It restricted bankcredit to restrain internal demand; it raised taxes, cutcapital investment, and sharply decreased nondeferiseexpenditures, including those for general administra-tion, social and community welfare, and economic ser-vices. But because of the high cost of importingweapons and of forgone exports, these policies couldnot prevent the deterioration of Nigeria's balance ofpayments position.

Note: All but 11,000 of the war deaths following 1949 took place in developing countries. All numbers are estimates and are subject to significanterror. Domestic conflicts are not always clearly definable as civil wars, and thus coverage of these statistics varies among studies. A variety ofcivil disruptions are excluded. For example, estimated deaths during purges and collectivization in the Soviet Union during the 1930s, whichrange from 5 to 20 million people, are not included. Figures for deaths resulting from other such events after World War II are also excludedbecause of poor data. Rough estimates for these range up to 15 million. Also, some wars are counted as civil even when foreign interventionoccurred. Deaths during the Korean War, deaths during the Viet Nam War for the 1965-75 period, and the war in Afghanistan for the 1978-89period are included under international wars.a. Totals include total estimated deaths. When breakdowns are unavailable, deaths are omitted from civilian and military subcategories. Totalsmay also differ as a result of rounding. Deaths were prorated when reporting periods spanned more than one decade.Sources: Sivard 1988, 1989.

141

Period

Number of warsDeaths during

international wars (thousands)Deaths during

civil wars (thousands)

Total deathsas percentage

of worldpopulationCivil International Civilian Military Total' Civilian Military Total'

1900-09 10 6 230 12 243 25 139 166 0.021910-19 15 9 7,045 13,470 20,556 1,140 139 1,327 1.131920-29 11 8 21 42 109 39 111 371 0.021930-39 11 8 933 838 1,770 646 1,109 1,796 0.171940-49 13 7 20,176 19,110 39,285 1,007 5 2,182 1.701950-59 20 5 1,073 1,926 3,031 1,571 253 1,879 0.171960-69 12 9 622 605 1,256 1,827 1,222 3,301 0.131970-79 18 7 639 606 1,246 3,543 1,236 4,957 0.161980-89 29 6 702 931 1,733 1,899 179 2,081 0.081900-89 141 63 31,440 37,539 69,229 11,697 4,393 18,059

Table 7.6 Public expenditure on the military compared with that on social sectors, 1986(percentage of GNP)

Militaryexpenditure 1-1.9 2-4.9

percent of GDP to about 5 percent. Many poorcountries expend more on the military than on so-cial sectors (Table 7.6). During the past three de-cades, military and civilian regimes appear to havespent roughly the same share of their GDP ontheir armed forces.

In high-income countries, military spending hasbeen increasing at the same rate as GDP. In devel-oping countries, military expenditure has been de-cliningfrom 6-7 percent of GDP in the late 1970sto about 5 percent in the second half of the 1980s,This was mainly the result of a drastic reduction ofmilitary spending in the Middle East (especially inSyria and Egypt) and in Latin America (after thefiscal crisis of the 1980s). But 5 percent of GDP is

142

Expenditure on health and education

BarbadosCyprus

AlgeriaCentral

African Rep.Côte d'IvoireFijiJamaicaMalta

BulgariaChileCongoCzechoslovakiaGabonGerman Dem. Rep.HungaryItalyKenyaLesothoLiberiaMadagascar

BahrainCubaEgyptEthiopiaGreeceHondurasKorea, Rep. ofKuwait

Iran, I.R. ofIsraelJordanOman

5-9.9

Gambia

Papua NewGuinea

SwazilandTrinidad and

TobagoVenezuela

MalawiMauritaniaPolandSenegalSomaliaSouth AfricaSpainTanzaniaThailandTogoYugoslaviaZambia

MalaysiaMoroccoSingaporeTunisiaUnited StatesYemen, Arab Rep.

SyriaUSSRYemen, P.D.R.

lOand above

Costa RicaLuxembourg

AustriaFinlandIrelandJapanSwitzerland

AustraliaBelgiumBotswanaCanadaDenmarkFranceGermany, Fed. Rep.NetherlandsNew ZealandNorwayPanamaPortugalSweden

United KingdomZimbabwe

GuyanaLibyaNicaraguaSaudi Arabia

Note: The ranges given in this table are illustrative of the differences in expenditures in the different categories; they are not necessarily indicativeof precise differences across countries because of some differences in the definition of the categories. The estimates of social expenditure do notcover those by local bodies.Source: Sivard 1989.

still an enormous sum; in many countries it wouldbe more than enough to double governmentspending on infrastructure or on health andeducation.

Governments need to take every possible stepto reduce military expenditure. Costa Rica is anoutstanding example of a government which de-cided to reduce military spending and focus itsefforts on the provision of health and educationan approach which improved equity and achieveda degree of political stability unusual in the devel-oping world. Costa Rica's poor soils and sparsenatural resources meant, however, that the coun-try had few enemies; its experience may not beeasy to replicate.

5-9.9 Chad Sri LankaChina SudanPakistan United ArabPeru Emirates

10 and above AngolaIraq

Less than 1 Brazil MexicoGhana Niger

1-1.9 Nigeria Argentina HaitiParaguay Bangladesh Nepal

Cameroon PhilippinesColombia RomaniaDominican Rep. RwandaEcuador Sierra LeoneGuatemala

2-4.9 Uganda Burundi IndiaZaire Benin Indonesia

Bolivia MaliBurkina Faso MyanmarEl Salvador TurkeyGuinea Uruguay

Many countries have to deal with bigger internaland external threats than those facing Costa Rica;even so, these threats hardly justify the sums be-ing spent today on armed might. Aid and financeagencies are entitled to ask whether it makes senseto help governments whose first priority is not todevelop but to add to their military strength.

Privatizing and reforming state-owned enterprises

In the 1980s and 1990s so far, transferring state-owned enterprises to the private sector has beenan important government objective both in OECDcountries such as New Zealand and the UnitedKingdom, and in developing countries such as Ar-gentina, Brazil, Chile, Ghana, the Republic of Ko-rea, Malaysia, Mexico, Nigeria, Togo, and Turkey.In most of these countries, privatization has meantmuch more than merely transferring assets to theprivate sector. It has been part of a broader exer-cise aimed at stabilizing and liberalizing the econ-omy on several frontsregulation, prices, trade,the financial sector. Governments have con-sciously set out to redefine the economic role ofthe state. As part of this shift, they have curtailedthe SOEs' privileged access to the budget or creditsystem, tariff or nontariff protection for their prod-ucts, and regulatory protection from private com-petitors. They have signaled a new determinationnot to pursue narrow distributional goals at theexpense of efficiency.

Many countries have taken the view that unlessprivatization were part of such a broad program ofreform, it would be an empty gesture. It wouldmerely transfer the control of rents from the publicto the private sector. There have been variants onthis general approach. In China, deregulation hasbeen accompanied by new institutional arrange-ments that allow the government to retain owner-ship while improving the enterprises' efficiency.In Argentina, and in Mexico to a lesser extent, pri-vatization with heavy foreign participation hasbeen used to reduce external debt and increaseinvestment in basic infrastructure, such as energyand telecommunications. In Argentina and Brazil,revenues from privatization are also expected tocontribute significantly to balancing the fiscaldeficit.

PROBLEMS OF IMPLEMENTATION. Privatizationhas proven to be an arduous exercise, however.Thin markets for domestic capital, adverse eco-nomic conditions, and the resistance of tradeunions and civil servants have slowed the process

virtually everywhere. Except for relatively ad-vanced economies such as Argentina, Brazil, andMexico, the infrastructure of privatizationlaw-yers, accountants, merchant bankers, and entre-preneursis largely missing in most developingcountries. The need to build up this infrastructureis particularly acute in Eastern Europe, where it isdifficult even to find qualified individuals to serveas directors of companies. Departments compe-tent to handle privatization must be created, thenadequately staffed and funded: a challenge in itselfduring times of financial stringency.

Legal issues also complicate privatization. InMexico, constitutional amendments had to be pas-sed in 1983 before privatization could go forward.In Turkey, sales were canceled when the courtsdeemed them illegal. In socialist economies, lawsmust be passed defining property rights, legaliz-ing private ownership, establishing guidelines forarticles of incorporation, and protecting minorityshareholder interests; all these are necessary if thelegality of a private purchase of a company is to beestablished. Similarly, the legality of the sellermust be established. Contrary to popular belief,governments in socialist countries did not holdclear titles to firms. In some instances, assets werenationalized shortly after World War II, but thepromised compensation, which would have legal-ized the nationalizations, was never made.Czechoslovakia and the former German Demo-cratic Republic established new laws granting pre-vious owners priority rights to compensation orthe return of their original property. Uncertaintyover whether there is a prior claim on the firm'sassets has made many potential investors wary ofprivatization.

Previous attempts at decentralization in socialisteconomies had given workers in many enterprisesrights that traditionally belong to shareholders inWestern countries. Workers' councils in Polandhave the right to decide on mergers and enterprisedissolution, asset sales, and appointment of chiefexecutive officers; in Yugoslavia, workers' rightsare codified even more extensively. In addition,state-owned assets are one of the few positive leg-acies of years of communist rule; the people insiston a fair distribution of this wealth as partial com-pensation for the suffering of the past. As a result,there is strong resistance to passing it into thehands of the old communist nomenklaturathemanagerial class, linked through party ties, whoran the economy. Yet this group is among the rich-est, and it has the best information on the realworth of enterprises and the business connections

143

to make firms run. These is a fear that an openmarket sale of assets will restore the nomenklaturato its earlier dominance.

The experience of the former German Demo-cratic Republic suggests that, even under favorablefinancial, legal, and technical conditions, sellingenterprises will take some time. The 9,000 stateenterprises to be privatized in eastern Germanycorrespond to about 30,000-40,000 firms in a mar-ket economy. Even though state enterprise salesby Germany's privatization agency (Treuhan-danstalt) have now reached 300 per month (com-pared with 25 per month during ten years in theBritish privatization program), it will take years tocomplete the process. Uncertainty about the valueof the companies was initially so great that pur-chase prices were to be settled later by arbitration.Although asset and enterprise values are nowclearer, price contingency clauses are being in-cluded in most sales contracts to enforce commit-ments by buyers.

Overall economic conditions, political consid-erations, and technical aspects of the process alsocomplicate privatization. In Chile, some of thefirms privatized during the period 1974-78 wererenationalized within a few years to salvage themfrom the bankruptcies that followed the deep eco-nomic crisis. In the mid-1980s in Nepal, a privatiz-ation was reversed because of opposition to thetransfer of the enterprise to a minority ethnicgroup. In Bangladesh, unresolved questions ofshare pricing and debt overhang led privatizedfirms to neglect investment in maintenance andnew capacity. Instead, the firms concentrated ongenerating immediate cash flowand much of theefficiency gains expected from privatization evapo-rated. In the former German Democratic Republicand Hungary, the first heads of the privatizationagencies resigned within a year; in Argentina, al-leged corruption in some privatizations led to acabinet reshuffle. Even where privatization hasencountered fewer setbacks, achievements haveoften been modest. In Mexico, for instance, two-thirds of SOEs have been privatizedbut thesesales account for less than 20 percent of the SOEs'total assets.

A REVOLUTION NONETHELESS. The recent changein government thinking on privatization has been,despite such difficulties, extraordinary. Much ofwhat has been achieved would have been unthink-able ten years ago. In Argentina, the governmenthas privatized two television stations, andawarded sales contracts for the telephone corn-

144

pany, the national airline, some components of thenational petroleum company, and the main dis-tributor of electricity. More privatizations are ex-pected in the near future. In Chile, privatizationshave reversed the emergency nationalizations ofthe preceding years; sectors traditionally domi-nated by the government, such as steel, petro-leum, and telecommunications, may be privatizedin the near future.

In Côte d'Ivoire, the private sector, already in-volved in water supply, is also going into powergeneration; in Togo, textile firms have been sold toforeign investors. More privatizations are expectedsoon in Brazil, Peru, Sri Lanka, and Turkey. Liq-uidations of nonviable SOEs are going ahead inseveral African countries.

The first phase of drafting and implementingnew legal statutes has been largely completed inCzechoslovakia, Hungary, Poland, and Yugo-slavia. Czechoslovakia and Poland seem commit-ted to a quicker pace of privatization and to thecreation of a broader shareholding base. Hungaryhas opted to go more slowly; it is creating joint-stock companies whose shares are deposited witha state holding company until the enterprises canbe valued and sold through public offerings. It ex-pects to privatize about two hundred companies inthis way during 1991.

In Poland most shops, gas stations, and trucksare already owner-operated, and a significant por-tion of housing is now private. Typically, the as-sets have been leased, not bought outright. Auc-tions of small assets, coordinated by the centralauthorities, have already started in Czecho-slovakia. For the larger firms, programs of "free"distribution of shares are being planned to acceler-ate privatization. The Polish plan calls for the con-version of several hundred large firms into joint-stock companies, with most shares allotted toworkers, pension funds, banks, and other finan-cial intermediaries (acting as trust funds for thepopulation at large); the remaining shares will besold to private investors. Similar arrangementsmay be worked out in Czechoslovakia.

LESSONS. Privatization is necessary and highlydesirable, even though difficult and time-consum-ing. It is not to be undertaken as end in itself, butas a means to an end: to use resources more effi-ciently. Removing price distortions and controls asquickly as possible is essential for that purpose.Unless prices are true indicators of costs and con-sumer demand, the true profitability of an enter-prise can not be known, so its asset cannot be

properly valued. Selling the enterprise at an ap-propriate price may be impossible, and meanwhilemanagers will be unable to make informed deci-sions on investment and production. Letting theprice system work as it should means removingdistortions such as price controls, distorted trans-fer prices between enterprises, subsidized loans,and preferential access to the budget and creditsystem; it also means getting macroeconomic p01-icy right, and that includes avoiding an overvaluedexchange rate.

It would be difficult to privatize all SOEs in thenear future even if governments wanted to. Mean-while efforts to raise productivity cannot wait.Governments need not hesitate to liquidate inher-ently unprofitable enterprises, and the remainingSOEs can be managed much more effectively. Softbudget constraints, interference in managementand recruitment, and restrictions on competition(either in product or factor markets) need to bereduced or eliminated. Efficient state enterprises

are to be found in many economies: for instance,Brazil, Ethiopia, France, Italy, Korea, Malaysia,and Singapore. These show that SOEs can be runas efficient commercial concerns responsive toconsumers. In many developing countries, im-proving the performance of SOEs is as urgent asprivatization in its own right.

The challenge of reform

The challenge for governments is to implement re-forms in the face of sometimes trenchant politicalopposition. Structural reforms may hurt powerfulinterests. Streamlining the civil service threatensurban workers with unemployment, especially inAfrica, where the public sector often employs upto half the salaried work force. Groups that canorganize against reform have to be reckoned with;the beneficiaries are often dispersed and unor-ganized, making it difficult for the government tocount on their support.

Box 7.7 From a centrally planned to a market economy

Transforming a centrally planned economy into a mar-ket economy requires complex and unprecedented re-forms. There is no experience to guide transitions ofthe current magnitude. And most countries in transi-tion are simultaneously creating a new political order.There is relatively little disagreement that the transi-tions have to be made, but there is much controversyabout the theory, timing, scope, speed, and sequenc-ing of reforms.

Three sets of issues arise. One concerns the eco-nomic implications of a policy sequence: will one kindof reform achieve its objectives while other economicdistortions remain? Another question is political: willmounting opposition derail reforms scheduled near theend of the sequence? Finally, there is technical fea-sibility. New legal, accounting, and financial systemswill require greater technical expertise and longer ges-tation periods than reforms that include only pricederegulation.

One school of reform proposals puts change in own-ership at the head of the sequence before or alongsidechanges that address macroeconomic stability and mar-kets. The rationale is partly political. With early priva-tization, there is less risk that the economy will remainstate-controlled and greater pressure for complemen-tary market-oriented reforms. But another school ofthought begins with macroeconomic and market-build-ing reforms: it leaves privatizationat least for largestate enterprisesto a second stage. (Under both pro-

posals some agricultural, retail, and residential assetswould be privatized early.) The rationale is that privateownership requires financial institutions, experience,and expertise that do not yet exist in the transitionaleconomies. Without this infrastructure, rapid privatiz-ation could lead to widespread corruption and eco-nomic and political chaos. Within each school there arefurther differences on the proper order for addressingparticular distortions.

No single reform sequence will fit all the transitionaleconomies. Reform histories vary; unlike others, Hun-gary has had more than two decades of experiencewith decentralized economic decisions. Macro-economic conditions range from great instability (theSoviet Union) to relative stability (Czechoslovakia). Pri-vate sector activity has been relatively higher in pre-dominantly agricultural countries such as China andViet Nam but negligible in more industrialized nations.

A preferred sequencing (Box figure 7.7) would in-clude early steps to stabilize the macroeconomy andderegulate domestic- and external-sector prices to giveclear, accurate signals for economic activity and for thevaluation of enterprises. These steps would be accom-panied and followed by intense efforts to rationalizeenterprises, improve economic decisionmaking, re-form trade policy, and build managerial skills and astrong financial sector. Privatization of large state en-terprises would become the next priority. Protection

145

146

Box 7.7 (continued)

would be cut and the economy would be opened toforeign competition on a firm, preannounced sched-ulefirst in goods and later in capital markets. Institu-tion building would be a basic theme from the start andat all levels: the legal contractual system, the structureof ownership, and the roles of key organizations in theeconomy would require reform and restructuring.

Large-scale privatization would not be at the head ofthe sequence, but to address the risk in delaying it,there would be early legal commitments (the distribu-tion of shares) that would guarantee private ownershipwithin a reasonable time. The program would be fast,in that each type of reform would move at the maxi-mum rate consistent with developing institutional ca-

Box figure 7.7 The phasing of reform

Note: Darker shading indicates intensive action. QRs, quantitative restrictions.

pacities. Indeed, a three-to-five-year time span seemsoptimistic in light of the progress achieved so far in thetransitional economies.

Reforms will surely involve painful adjustments. In-flation and unemployment will worsen as price con-trols are removed and the real economic losses of someactivities are revealed. Political opposition may mountwith these developments and with the rise in incomeinequality that comes after radical change in the incen-tive structure. But progress in exports and the availabil-ity of consumer goods could soon follow. And, giventhe relatively strong human resource endowments inEastern Europe, prospects for growth could beexcellent.

Area of reform

0

Year of reform

1 2 3 4 5 6 7 8 9 10

I I I I I I

Macroeconomy Maintain stability

MarketsGoods and services

Prices

Trade

Distribution

Labor market

Financial market

es Liberalize prices of some necessities (including housing)

Remove QRs Adjust tariffs to modest level

Privatize; demonopolize Develop

Deregulate hiringand firing Liberalize wage bargaining

Restructure and develop Liberalize and privatize

Ownership structure

Small enterprises

Large enterprises

Foreign investment

Develop and privatize

Evaluate Restructure and privatize

Revise regulations

Government

Legal framework

Institutionalframework

Social safety net

r'ti Extend reforms to other areas

Reform legal and regulatory institutions and fiscal administration

gencies Institutionalize

If output responds quickly to a program of re-forms, support for the program will increase andthe changes can be consolidated. Rapid growth inexports bolstered reforms in Indonesia, Korea,and Turkey. Strong export growth can also help toprevent policy reversals caused by balance of pay-ments problems and dwindling foreign exchangereserves. Reforms that improve the investment cli-mate are more likely to be sustained because newinvestors will add to the forces in support. Ex-panding output and investment would enlarge thetax base, raise tax revenues, and reduce the budgetdeficit. All this argues for reforms that are boldenough to elicit a prompt supply-side response.Timid programs are unlikely to win converts to thecause of reform. Many reformers in Eastern Eu-rope have taken these lessons to heart (Box 7.7).

Despite the political difficulties, many govern-ments have shown great ingenuity in implement-ing controversial reforms. For instance, the gov-ernments of Bolivia, Ghana, Korea, and Mexicostrived to convince the public of the costs of inac-tion and to explain the thinking behind their re-forms. International diplomacy can lend credibilityto the cause. Agreements with the EEC helpedGreece, Israel, Portugal, and Spain adopt tradereforms in the 1960s; accession to the GAUhelped Mexico in 1986. Some governments havenegotiated social pacts to distribute the burden ofadjustment equitably between labor and business,as in Mexico in the 1980s and Israel in 1986. InChile in the early 1980s and Sri Lanka in themid-1980s, the success of trade liberalization grad-ually changed the orientation of the manufac-turers' association from import substitution toexport promotion. In general, governmentscommitted to addressing their societies' problemshave rarely lost power because of this deter-mination.

Of course, there are limits to persuasion. Often,it seems, reform can be born only of a full-blowneconomic or political crisis. Examples range fromMeiji Japan to contemporary Argentina, Ghana,Peru, and Poland. Sometimes, even in the face ofeconomic collapse, reform is blocked by the gov-ernment's key supporters. In such cases, externallenders and aid agencies face an extremely uncom-fortable fact: despite the country's need of externalresources, such support may do more harm thangood by helping to keep the anti-reform adminis-tration in power.

In countries that are not paralyzed by politicalforces, and where reform can go forward, the taskfor external aid and finance agencies is to promoteit. They can do this by avoiding support for un-productive activities or for new projects thatwould be implemented under severely distortedconditions. In many countries, external agenciesmust help to strengthen the public institutionswithout which development assistance is likelyto be ineffectual. Reform sometimes imposesheavy costs on those least able to bear them: thepoor. Well-designed safety nets (such as the emer-gency adjustment funds introduced in Bolivia andGhana in the 1980s) can help the most vulnerableand, in doing so, broaden the constituency fordevelopment.

Reform will remain a formidable task, requiringpolitical courage and economic vision. Combiningthe many different elements described here and inother chapters is, in itself, enormously difficult.The appropriate combination of factors will varyfrom country to country, according to circum-stances. And even when reform is well-designed,governments are sure to encounter unforeseen set-backs, some of them entirely beyond their control.Development is indeed a challengebut, as his-tory tells us, one that can be met.

147