Public Finance for Developing Countries
From Planning Paradigm To Market Paradigm
Opinions on the appropriate role of government policy, including
budget policy, in developing countries have to be based on some vision of what the
basic mechanisms of the development process are. The specification of such a vision is
crucial also for the choice of analytical techniques when studying developing countries.
This is why this chapter puts budget policy, or public finance, into the broad perspective
of the development process. This also makes it natural to see budget policy as a
complement and/or substitute for other types of policy.
The dominant vision of the development process during the 1940s, 1950s,
and early 1960s was that the market failures in less developed countries were so huge that
the market mechanism could not be much relied on in such countries. This view was
certainly characteristic of economists such as Gunnar Myrdal, Ragnar Nurkse, Raul
Prebisch, Paul Rosenstein-Rodan, and Hans Singer, even though their emphases on specific
aspects of asserted market failures differed strongly.
Various forms of "structuralism" were also popular. Developing
countries were asserted to be characterized by pronounced "structural
inflexibility" in the allocation of resources; in other words, low (or even zero)
elasticities and long time-lags with respect to the economic incentives of the supply and
demand for goods, services, and factors of production,
I am grateful for useful comments on an
earlier draft from Jorgen Appel gren, Jagdish Bhagwati, Arne Bigsten, Lawrence Krause,
Anne O. Krueger. Deepak Lal, Sven-Olof Lodin, Mats Lundahl, and Amartya Sen. Karl Gustav
Hansson and Reza Firuzabadi have assisted with statistical compulation.
. 103 .
and, indeed, of productive effort and entrepreneurship in general,
were thought to exist.  Also based on such structuralist views
of the economies of developing countries was the idea of binding saving or
balanceof-payments constraints on economic growth owing to asserted weaknesses in the
response of saving and investment to changes in income and interest rates, as well as of
exports, imports, and long-term capital movements to changes in exchange rates, the terms
of trade, and the rates of return on assets. A special variant was the "two gap"
theory of savings and balance-of-payments constraints, asserting strict limits on the
raising of taxes by the national government and difficulties in turning domestic saving
into capital formation via the exchange of traditional export products for capital goods
on international markets owing to inelastic world demand for the former (Chenery 1965,
From views like these which were far from uniform followed both a strong
distrust of the price mechanism and, as a mirror image, considerable enthusiasm for
government regulation (such as licensing systems), as well as economywide central planning
of inputs, outputs, exports, imports, and investment activity. Moreover, as the
manufacturing sector, in contrast to agriculture, was often asserted to be characterized
by constant or even increasing returns to scale, a high propensity to save, and rapid
technological progress, government-enforced industrialization at the expense of
agriculture and handicraft production was usually strongly advocated. Without drastic
government actions in these fields, developing countries were asserted to be doomed to
"underdevelopment equilibrium traps" or "vicious circles of poverty."
Also popular was the notion that the entire manufacturing sector could be treated as an
"infant industry," though this notion is difficult to distinguish from general
political and ideological preferences for industrialization as the essence of
economic and social "modernization."
The recommended, and indeed often also the actual, role of public
finance during the first decades after World War II should be seen in this
perspective: (1) as helping provide an industrial base by way of heavy public investment
in both physical infrastructure and manufacturing, often in the form of large-scale
government projects; (2) as squeezing private consumption by increasing the aggregate
saving ratio by way of
. 104 .
taxes, in particular on traditional exports and on the large
agricultural population (though in reality agriculture was perhaps squeezed more by
overvalued exchange rates and regulated output prices than by explicit land or
agricultural taxes); and (3) as directing the allocation of economic resources in general
by means of government enterprises, subsidies, government credits, taxes, and tariffs
(import-substitution policy) as a complement to "command," which was to be
implemented by physical regulations of various types.
These ambitions made the government budget a main tool of aggregate and
disaggregate "national economic planning" for the mobilization and allocation of
resources a point emphasized, for instance, by A. Waterston (1965). While attempts to
direct the economy by way of taxes, tariffs, and subsidies must have been based on the
idea that private agents do react to economic incentives though unguided markets
were asserted to give the "wrong" incentives the recommendations for reliance on
public enterprises and on command of private firms by way of physical regulations were
more consistent with a structuralist view of the world (in other words, weak responses of
private agents to economic incentives).
It may be argued that the policy recommendation to raise the aggregate
saving and investment ratio was the most valuable feature of the predominant development
paradigm during the first few decades after World War II, and that the drastic increase in
saving and investment ratios to some 20 25 percent of GNP in most developing countries was
the most important achievement of actual development policies. Developing countries have
usually been much less successful in increasing economic efficiency, and hence in speeding
up the rate of productivity growth. For instance, while output in manufacturing in
developing market economies increased by 5.1 percent per year during 1960â83,
the accompanying increase in labor input was as high as 3.5 percent, which implies that
the increase in labor productivity was only about 1.6 percent (calculations of time trends
based on data in United Nations 1983).
By itself, a strong increase in labor input is, of course, favorable
from the point of view of mitigating unemployment, or "underemployment," but
output growth, and hence labor-productivity growth, has certainly been much weaker than
would have been possible in reasonably efficiently functioning economic systems. Indeed,
even if very laborintensive methods of production are used, labor productivity should be
able to grow by several percentage points per year "simply" by the intro-
. 105 .
duction of better technologies and organization, as illustrated by the
experience of developed countries, and indeed also of some developing countries. Many
developing countries have simply been getting too small an increase in labor productivity
from their investment. We also notice extremely high marginal capital/output ratios in
many developing countries, such as in Africa and Latin America.
It has in fact become obvious that vastly different rates of growth of
GNP, and perhaps in particular of consumption, can be achieved with about the same rates
of capital accumulation depending inter alia on the allocative efficiency of investment
and production. For instance, reference is often made to the success of a number of
countries in Pacific Asia (Taiwan, South Korea, Singapore, Hong Kong) that have relied to
a considerable extent on economic incentives and an outward-looking development strategy
though active and competent governments in these countries have certainly also stimulated
growth and economic efficiency by way of institutional reforms, redistributions of assets
and infrastructural investment. To some extent, these countries have been engaged in
growth forecasting ("indicative planning"), and, in varying degrees, operate
state-owned firms, but it is difficult to have a definite opinion about the role of these
features, which they share with many other, less successful nations. What has clearly
differentiated these countries from many others is that governments have tried to support
rather than restrict the activities and initiatives of the private sector.
It is for these various reasons natural that the emphasis in analytical
discussions of economic policy in developing countries has gradually shifted to issues
relating to the allocation, and not just the total volume, of investment, and indeed to the
allocative efficiency in general of production activities. This is probably an
important explanation for the increased respect for decentralized decision making by way
of markets, price signals, and economic incentives, and, as a mirror image, an increased
skepticism about the usefulness of direct government regulation and central planning.
. 106 .
More generally, it has become increasingly understood during the past
few decades that, contrary to previous assertions, both the aggregate and the
"fine" (disaggregate) micro structure of the economies of developing countries
respond quite strongly to economic reward, including profitability prospects and relative
prices and wages if governments allow such a response. This is
perhaps what we should expect, as poor people have no less reason to respond to the
opportunities for improving their economic situation than do more affluent people, perhaps
rather the reverse.
Moreover, after having brought about large infrastructural investments,
and in some cases a considerable mobilization of resources, it is natural that problems of
economic efficiency, and hence resource allocation, become more interesting. It has also
become clearer over time that the potentialities of import substitution in manufacturing
are rapidly exhausted in most countries owing to the smallness of domestic markets. It
would also seem that the attractiveness of the Soviet planning model subsided when it
became more widely understood that this was more of a "mobilization model" than
a prescription for economic efficiency indeed, that the model stimulated
When relying on markets, governments must, of course, help ensure that
the market signals are "right," in the sense that prices reflect opportunity
costs and preferences. However, it is also crucial that the price signals
"work," in the sense that various institutional "filters" in society
do not distort, or even "abort," the information and incentive content of market
signals. Thus there is a potentially important role for the government both as regards
improvement of the information and incentive structure and helping strengthen (and
perhaps, at an early stage of economic development, even helping create) market-oriented
institutions. In broad terms, the main contribution of public finance to the
economic development of market-oriented developing countries would then probably be (1) to
provide infrastructure and public goods; (2) to help relative prices reflect opportunity
costs and preferences, which is
. 107 .
often more a question of avoiding and removing distortions previously
introduced by the government itself than of fighting "spontaneous" market
distortions; (3) to contribute to the redistribution of income and wealth (according to
certain values concerning equity) by methods that are as market-conforming as possible;
and (4) to contribute to the development of market-oriented institutions that respond
satisfactorily to market signals, for instance, in the fields of finance, trade, labor
markets, consulting, and the transfer of technology. However, it is also important to
emphasize the crucial role of (5) macroeconomic stabilization policy, as failures in that
field have often been extremely damaging to attempts to liberalize the economies of
If governments start to rely more on markets for the supply of ordinary
goods, the public sector can increasingly concentrate on those activities, mentioned
above, that only the government can pursue, or where government at least has a comparative
advantage relative to private agents. The administrative resources that are then released
in the public sector can instead be used to improve the quality of the public sector's
remaining functions, to the extent that such resources are not simply transferred to the
private sector. This is an important consequence of the shift of a developing country to a
more market-oriented system, as the majority of these countries are characterized by a
shortage of competent civil servants. Administrative "overload." which has
recently been much discussed in developed countries, is an even more characteristic
feature of most developing countries a problem that could be mitigated by a shift to a
more market-oriented system. Indeed, managers of firms would then also be able to devote
more time to "ordinary" business, rather than bargaining with government
officials, or "rent seeking" (Krueger 1974), whereas today the latter often
yields higher returns than do attempts to improve efficiency within firms. A removal of
regulations could change that.
More generally, in market-oriented economies, the role of government
planning and public finance is largely to "plan" the physical and psychological environment
of private agents rather than to plan what those agents are supposed to do.
Method of Analysis
On the basis of the dominating view in the early post-World War II
period that private agents in developing countries react (if at all!) completely
differently to economic incentives than do agents in developed
. 108 .
countries, and that markets in developing countries will not be able
to function properly in the foreseeable future, there followed a profound skepticism about
traditional methods of economic analysis. Developing countries were often asserted to be
"different kinds" of economies, requiring both new tools of analysis and
different behavior assumptions; indeed "development economics" was often
asserted to represent a new and separate branch of economic analysis . By contrast,
this chapter is based on the assumption that there are great enough similarities between
behavior patterns and economic mechanisms in general in developed and developing countries
to make standard economic analysis relevant for the latter type of country as well.
It is indeed quite easy to illustrate the relevance of standard economic
analysis for developing countries. For instance, in a similar way as in developed
countries, "overvalued" currencies, by keeping down profitability, tend to
reduce output and employment growth in the tradable sectors, which often also results in
(increased) budget deficits owing to the negative consequences for the tax base. Moreover, it is well established by now, not only for developed but
also for developing countries, that high real wage rates, in particular when combined with
low real interest rates, tend to favor capital-intensive methods of production. Regulated
wage rates tend to accentuate unemployment for certain types of labor (for instance where
minimum wage rates are binding), while for other types of labor more or less permanent
vacancies tend to prevail. Pegged interest rates create credit shortages, with
"arbitrary" credit rationing and an inefficient allocation of credit and capital
as a result. High tariffs and large subsidies to specific sectors tend to expand these
sectors at the expense of others, in particular sectors for which government regulations
have kept down prices, such as agriculture. Rent control hits house building, creates
housing shortages, and results in a deterioration of the housing stock as well as a
reduction in labor mobility.
We also note that, as in developed countries, regulations breed new
regulations, as politicians and public administrators try to fight the unintended, and,
for them, often unexpected, side effects of previous regulations. And, probably even more
than in developed countries, regulation is "the mother of corruption," as
corruption presupposes that politicians and administrators have "something to
sell" such as licenses, tax concessions, tariffs, or subsidies. Indeed, there is most
. 109 .
also a "reverse causation": corrupt politicians and public
administrators have a strong self-interest in promoting regulations.
Moreover, the possibilities of substitution between labor and capital,
and indeed between inputs in general, have come to be regarded as much more promising than
they were earlier thought to be a development that in economic theory is symbolized by the
replacement of the rigid Harrod-Domar growth equation with more flexible aggregate growth
models Ã¡ la Solow and (in more detailed and quantified form) Denison. Such
possibilities of substituting labor for capital in developing countries have proved to be
particularly promising in multiple-shift operations and in ancillary services, such as
maintenance, and handling of material and other transport services, such as packing
(Morawetz 1976; White 1978).
My conclusion is that it is quite appropriate to regard
"development economics" as an application of the standard tools of economic
analysis (whether at micro-or macroeconomic levels) to long-term growth and development
issues for all types of countries, in the same way as trade theory is the
application of micro-and general equilibrium theories to issues of international trade
regardless of what types of countries are studied. In particular, the traditional
microeconomic theories regarding prices, markets, and incentives are probably no less
useful for the analysis of developing countries than for that of developed countries. This
also means that the entire arsenal of tools and insights from applied fields of economic
analysis such as industrial organization, money and banking, labor economics, and, as
illustrated by this chapter, public finance, can be put into operation in analyses of
developing countries and not just of developed ones.
Of course, it is important to take into account various institutional
peculiarities in the analyses, though by treating institutions not as insurmountable
obstacles to development, which was typical for early postwar development theories, but
rather as the "filters" through which incentives, as well as commands,
necessarily have to pass. However, that cannot be done by assuming some kind of
"standard" developing country institutional setup. Owing to the wide variations
in institutional conditions among developing countries heuristically speaking, with
stronger variations than among developed countries the institutional conditions have to be
specified separately for each country.
The main contribution of the earlier evolution of economic analysis for
less developed countries as a specific field of economics is then mainly that it has
increased awareness of the importance in economic
. 110 .
analysis (of both developed and developing countries) of watching out
for various institutional peculiarities and the changes in these over time.
However, when discovering "institutional peculiarities," which
sometimes makes markets look unfamiliar, it is important to realize that what in the light
of traditional models may (superficially) appear as a "market distortion" may in
fact be a simple reflection of costs that are not apparent when examining only
conventionally defined production costs. It is, for instance, well known that price
differences between apparently similar goods, services, or factors may reflect differences
in risk, information costs, or "interlocking" markets (Stiglitz 1985)
Concern for the interaction between incentives and institutional
conditions is important also in the field of politics and public administration, and not
just in markets. Indeed, if it is agreed that differences in government policies are
responsible for much of the variation in economic performance among nations, it must be a
research topic of the uppermost priority to try to establish which institutional
circumstances are conducive to various types of policies. More specifically, policy
recommendations that do not rely on a realistic assessment of the functioning of the
political systems and of the administrative capabilities of the countries concerned often
do more harm than good. In particular, policy advice that is based on the assumption that
governments and public administrators behave like well-informed, competent, and highly
"benevolent guardians of the public good," maximizing some asserted social
welfare function, are bound to lead to disappointing results. Indeed, it may be argued
that a policy advisor should base his advice on a hypothesis about the effects of
his advice on the actual policies (Lindbeck 1973).
Against this background it is important to include in the analysis not
only traditional concepts of static efficiency and Pareto optimality, but also broader
ideas about the functioning of the economic and political system, such as J. M. Clark's
vision of competition as a dynamic process ("workable competition"); Joseph
Schumpeter's idea of competition as "creative destruction" (when new kinds of
competition, technologies, and products threaten existing ones); Friedrich von Hayek's
view of competition as a decentralized search for ways of using existing knowledge more
efficiently; Harvey Leibenstein's vision of X-efficiency, reflecting other types of
economic efficiency than traditional allocative efficiency; and Herbert Simon's theory of
"satisfying behavior." "bounded rationality," and endogenous changes
in the aspiration levels of agents.
For instance, rather than taking the production function as given, it is
important to analyze the process by which the production function is chosen or developed,
not only via research and investment in physical
. 111 .
and human capital, but also through organizational modification,
innovation, and entrepreneurship. And when analyzing economic policy it is necessary to
regard political and administrative decision making as an endogenous process with its own
patterns of behavior, as suggested by the Public Choice School (Buchanan and Tullock 1962)
and others. The latter point also emphasizes the importance of studying constitutional
rules and political culture, including not only the distinction between democracies and
authoritarian regimes, but also the rules of election, the degree of political
centralization, the character of party competition, and the use of referenda (Lindbeck
1985). Moreover. Amartya Sen (1981) has argued that the existence of a free press and an
active political opposition has helped prevent crop failures from resulting in starvation
for various population groups.
In other words, analyses of the development process and appropriate
government policies, including budget policies, have to include much broader, though often
less rigorous, aspects than those on which formalized general equilibrium theory and
optimization analysis arc built. Institutional conditions and institutional change, the
political and administrative processes, and the environment for human creativity and
entrepreneurship are factors that seem to have played too small a role in economic
analysis of both developed and developing countries.
This methodological point also implies that it is important that the
genuine complexity and diversity of the development process are taken into consideration,
and that evaluations of the performance of both markets and governments are made against
much less ambitious benchmarks than perfect competition, perfect information, Pareto
efficiency, and the maximization of social welfare functions. That we shall do below.
If a country starts to rely more heavily on markets, economic
incentives, and decentralized private initiatives, the "classical" roles of
government spending for the allocation of resources come to the forefront: (1) infrastructural
investment in physical and human capital, including the supply of goods that are
produced with particularly high fixed costs relative to the variable costs, such as
harbors, bridges, and some other transportation systems; and (2) the supply of public
(collective) goods including the legal system, education, basic research, and
environmental protection. Moreover, as poor people in developing countries cannot be much
helped by tax reductions, government expenditure policies become crucial also in (3) redistributional
. 112 .
Infrastructural expenditures hardly need advertisement today. However,
as the low productivity level in developing countries derives to a considerable extent
from the lack of human capital, it may nevertheless be worth advertising infrastructural
investment in the form of the accumulation of human resources in a wide sense of the
term including not only education, but also health, sanitation, and, in many countries,
food supply to particularly poor sections of the population
As it has been increasingly recognized that the social rates of return
in most developing countries are higher for primary education and vocational training than
for most forms of higher theoretical education (Psachropoulous 1981), a change in the
composition of public spending on education in favor of the former seems to make sense for
many developing countries. It is also important that public policies in the field of
education and training recognize the need to build up competence that is relevant at the microeconomic
level in society (for example, and in particular, within individual firms). This would
probably be greatly facilitated if firms, rather than government institutions, were
largely in charge of vocational training programs, so as to make them practically
applicable and strongly market-oriented the latter being particularly important in
economies that follow a market-oriented strategy of economic development. However, as
firms have suboptimal incentives to provide general training, owing to the mobility of
labor between firms, it is natural to recommend combinations of general schooling provided
by public institutions and specific training within firms, though the latter, too, could
be at least partly tax-financed. Modified to fit the specific conditions in the countries
concerned, the apprenticeship system in West Germany may be a model worth following.
It is also conceivable that the import and domestic dispersion of
technology could be stimulated by government initiatives, even if the actual import is
certainly best done by individual firms. For instance, government initiatives to help
establish private import and service firms in the field of technology may be worth
pursuing for a while in the least developed countries.
In many fields of technology it would certainly be useful if research
could take place in the countries of the Third World themselves though in many cases
in cooperation with institutions and individuals in developed countries. The rationale is
to increase the probability that the research will be relevant for the developing
countries. Such activities could also perhaps, at least for a while, be stimulated by
government initiatives. Obvious examples of such fields are aspects of agriculture
. 113 .
and health care, such as tropical agriculture, soil analysis,
integrated development in arid areas, aquaculture, and tropical medicine. As the results
of research and development in these fields are characteristically public goods owing to
the smallness of the firms and the externalities in the ecological systems, government
spending programs have a particularly important role to play here. For instance, the
experiences of agricultural extension services, financed by the government, often seem to
have been rather good. Indeed, the atomistic structure of the agricultural sector makes it
important that governments lake initiatives to help establish such services in precisely
Beside basic general skills and technological competence, one of the
most important bottlenecks in the area of human resources in developing countries is managerial
skills in the private as well as in the public sector. The role of managerial
skill is not only an issue of the competence of managers at the top of organizations.
Particular attention must probably be focused on middle-level management, supervisory
staff, and people providing specialist services at the middle level for example,
procurement and inventory management; production management; control research; marketing;
advertising; tool room service; raw material and product testing; machine assembly;
equipment maintenance; staff recruitment; and project development Government initiatives
to stimulate training in such fields are clearly a promising type of government investment
in developing countries.
Whereas the various types of training mentioned above are relevant for
all types of developing countries, the buildup of entrepreneurial capabilities is
particularly important in liberalizing developing countries. Experience from many
countries over long historical periods illustrates the enormous role of entrepreneurship
in the development process, not least in small and medium-sized firms and newly started
firms. Because of the difficulties involved in formalizing and quantifying the role and
importance of entrepreneurship, it easily "disappears," not only in economic
theory, but also in domestic development plans and political and administrative
discussion, which has often concentrated on existing firms and large firms. However, all
developing countries that opt for industrialization and modernization along
market-oriented lines are strongly urged to facilitate the emergence of vital
entrepreneurship not only by way of formal education in business, but also, and in
particular, by allowing and stimulating entrepreneurial initiative. Indeed, in
addition to land, capital, and labor, it is reasonable to regard entrepreneurship as a
fourth factor of production of crucial importance for economic
. 114 .
development a factor the government may promote by stimulating the
buildup of facilities for the training of entrepreneurs and by deregulation, which tends
to release entrepreneurial skills.
So much for the "creation" part of Schumpeter's vision of
"creative destruction." It is, of course, also important that politicians allow
the "destructive" part of the process to operate by avoiding subsidizing
contracting or even "dying" sectors, firms, and production processes, even
though the "demand" for such protection is certainly one of the most powerful
forces in the political process in most countries. An important "constitutional"
issue is then what types of political decision-making rules are conducive to preventing
more subsidization than the electorate, and a majority of politicians, on reflection
The stimulation of entrepreneurship is also an issue of the political
and social attitudes in society toward entrepreneurial activities. Indeed, the attitudes
in society toward entrepreneurship may be regarded as an important collective good in a
private enterprise economy.
In the context of government expenditure policies, it is often also
important to look over the rules, incentives, and practices of the management of publicly
owned firms and agencies, which play an important role in many developing countries.
Indeed, if governments of developing countries are serious about the shift to a more
market-oriented system, it is important that this shift incorporate stale-owned
enterprises too, in the sense that these are either handed over to the private sector or
given the same incentives and freedom of action as private business and that they
also feel the same need as private firms, by way of competitive pressure, to react to
market signals. The withdrawal of "automatic" government financing of losses of
government-owned firms is crucial from that point of view.
This chapter is not designed to deal with issues related to specific
production sectors. However, it is nowadays well understood that the modernization and
industrialization process is easily damaged, or may even fail completely, if agriculture
is neglected, which may also have serious consequences for the distribution of income.
"Negative" illustrations of this thesis are provided by several South American
and especially African countries, while positive examples are found in several countries
in Pacific Asia countries that have also been strikingly successful in manufacturing.
Widespread experiences suggest, for instance, that government marketing
boards in agriculture tend to destroy incentives. Such boards usu-
. 115 .
ally do not have either the competence or the incentives to create a
level and structure of producers' prices that is even remotely rational for economic
efficiency. Moreover, they tend to direct the rents from agriculture to inefficient and
highly protected manufacturing firms by way of import-substitution policies. Thus, as a
way of improving the production incentives for farmers, it is, as a rule, useful to
abolish government marketing boards in agriculture. This would also contribute to the
allocation of the rents and profits of agricultural production via market mechanisms
rather than by bureaucratic fiat and government subsidies.
Favorable production opportunities in agriculture are important also for
achieving a geographically dispersed income growth in society among both regions and
population groups. Obvious policies to promote this, besides a reasonably conducive price
policy toward agriculture, are a decentralized public infrastructure and regionally
dispersed public services in rural areas in general. Of course, this is something that is
important regardless of whether the economy is liberalized or not. However, as
decentralization of decision making may in some cases accentuate regional income
differences, at least for a while, there is a case for conscious government attempts to
speed up the process by which higher income in the national economy as a whole raises
incomes in regions that lag behind.
This raises the general problem of poverty and income distribution.
Considering the enormous misery among a minority at the bottom of the income distribution,
humanitarian values should certainly make us emphasize redistributive actions in favor of
the very poorest members of society, in the same way as redistribution policy in the
presently developed countries started with "poor laws" even before the
Industrial Revolution. On the expenditure side of the budget, such programs could very
well to a considerable extent consist of what has been called a "basic needs
strategy" for the purpose of providing basic food, shelter, water, sanitation,
health, and education to the very poor, which is perhaps to some extent most effectively
done by way of transfers in kind. Indeed, experience suggests that it is possible for a
country to achieve substantial improvements in a number of "social indicators"
at relatively low levels of per capita income (Balassa 1983), examples being life
expectancy, infant mortality, and child death rates.
Thus there is evidence of the usefulness of "direct" attempts
to satisfy some "basic needs" in nutrition and health, even at low levels of per
capita income. A "basic needs approach" is therefore quite compatible with
. 116 .
a growth-oriented approach based on economic incentives, provided that
neither is pushed to the extreme. Indeed, a great number of studies indicate that
increased spending on health, nutrition, and education up to certain levels may give
considerable boosts to productivity growth (Balassa 1983).
So far, it has been typical for developing countries to provide income
support via indirect subsidies of commodities rather than via direct transfers. One reason
has certainly been administrative feasibility. However, another reason has simply been
that strong pressure groups in the urban centers have "demanded" low food and
housing prices, which has not only reduced production incentives in agriculture and
housing but has often also assisted people with income levels considerably above those of
the rural poor.
Sooner or later, when the "modern sector" starts to dominate
the national economy, we would expect the same types of demands for social security
systems for a broad section of the population as in developed countries. This is, again,
not something which is confined to marketoriented systems. But market-oriented systems are
certainly often blamed for creating inequalities and insecurities. Thus, in order to
maintain their legitimacy among broad sections of the population, such systems need not
only to pursue redistributions to the very poor, but also to provide social security
systems for the population as a whole, though powerful urban pressure groups may benefit
the most from such systems.
Indeed, while social security spending is only about 1.2 percent of GNP
in low-income developing countries (with per capita income below $400 in 1983 prices), the
figure is as high as 6.6 percent in uppermiddle income countries (per capita incomes above
$1,600); see figure 3.1. (Other categories of public spending do not differ much between
developing countries with different per capita incomes. )
Eventually, of course, public spending programs may raise the same type
of welfare state incentive problems that are intensively discussed today in most developed
countries. Indeed, there may be a risk of the creation of "premature welfare
states," in the sense that the economic
. 117 .
foundation for an elaborate system of social security, transfer
payments, and redistributions of income among large population groups does not yet exist
(Uruguay is often mentioned as an example). "Early" attempts to create elaborate
social security systems may not only create severe incentive problems in the private
sector; they may also strain the taxation system so much that governments will not be able
to provide what they have a comparative advantage in supplying: collective goods and
infrastructure facilities. These points lead directly to the problem of taxation.
The level and structure of taxes, subsidies, and tariffs differ so
much between various developing countries that generalizations about their taxation
problems are difficult. However, the general level of taxation in this group of countries
is still usually much lower than that in developed countries typically 10â25
percent of GDP, as compared to about 30â55 percent in Western Europe, 33
percent in the United States, and 25 percent in Japan. Among eighty-two developing
countries for which information is available, the average tax share of GNP seems to have
been about 18 percent in the early 1980s (Tanzi 1986), with taxes on goods and services
and foreign trade playing a much more important part than in developed countries. (Figure
3.2 provides statistics for "central government" revenues; figures for
"general government" are probably, on average, one or two percentage points
higher.) Thus disincentive problems for private agents, owing to a generally high
level of tax rates, cannot possibly be a serious problem in most developing countries.
However, these figures underestimate the "tax bite" for those sectors that
actually pay the bulk of the taxes. (For instance, if we assume, as an extreme case, that
agriculture pays no taxes at all, the average tax rate for the rest of the economy would
be about 24 percent as compared to 18 percent for the entire economy.)
One dilemma, though, is that a given level of taxes (as a percentage of
GDP) may "hurt" people more in a poor than in a rich country by depressing an
already very low level of private consumption. However, this is not really a problem of
economic disincentives for private agents, but rather an issue concerning the ability, or
disability, of the political process to generate a reasonable allocation and distribution
of resources between private consumption, public consumption, and investment. However, it
is also an issue of the efficiency of the allocation of investment.
. 118 .
Central Government Expenditure for Low, Lower-Middle, Upper-Middle, and Industrial Economics
. 119 .
Central Government Central Revenue for Low, Lower-Middle, Upper-Middle, and Industrial Economics
. 120 .
In a country where this allocation is more efficient than in other
countries, it is not necessary to squeeze private consumption so much by way of taxes.
It is not clear whether liberalization of the economic system requires,
or can be expected to result in, a higher or lower total level of taxation. Reductions in
the level of tariffs no doubt raise the need for new tax sources. Moreover, if increased
reliance on markets results in a more uneven distribution of income, and indeed if it is
already believed that this is the case, the political process may generate strong
pressure for more redistributions via public budgets, and hence a need for tax increases.
However, there are also factors that reduce the need for taxes in connection with
economic liberalization, since some expenditure items would tend to disappear, or at least
fall. In particular, a reduction in the need for subsidies to enterprises private as well
as government would be an important part of a shift to a more market-oriented
economic system, provided profitability is kept up sufficiently, which requires that the
real exchange rate is not overvalued. The need for taxes would also recede to some extent
if public authorities shifted to a greater reliance on users' fees for various types of
public services, a reform for which there are well-known allocative and efficiency
arguments. A reduction of the public bureaucracy and increased competition for public
service agencies from private agents would further reduce the need for taxes partly
because competition would most likely increase efficiency in the public sector, and partly
because the size of the public service sector would be smaller owing to private
Another feature of economic liberalization is that increased incentives
for private saving would reduce the need for public saving, though the recent fall in
public saving in some developing countries is a reason for not advising the authorities to
make substantial tax reductions on this ground.
However, even though the level of taxes in most developing countries
cannot be regarded as excessive, or "harmful," the structure of taxes,
subsidies, and tariffs is certainly already a serious problem. It is hardly necessary to
say that high and strongly selective tariffs and taxes on foreign trade export discrimination
policies often result in suboptimal foreign trade, and hence in an underutilization of the
gains from the international division of labor. Indeed, it is today quite well established
that such policies, often designed to promote "import substitution," when
pursued for long periods (such as several decades), have been highly detrimental to most
developing countries that have pursued them.
. 121 .
It is also quite clear that highly selective commodity taxes, as in
fact implemented, have often created inefficiencies in the allocation of resources in the
private sector, on both the consumption and production side, without always providing
advantages in terms of the distribution of income. And in cases where selective taxes and
subsidies have actually improved the distribution of income (on the basis of certain
values), the same improvements could perhaps in many cases have been achieved by a
structure of taxes and subsidies with smaller efficiency costs.
Severe distortions may also come out of some other taxes. The provision
of accelerated depreciation for investment and subsidies to some investment, in
combination with payroll taxes on labor, may contribute to raising the wage/rental ratio
though recent increases in interest rates may have reversed this feature in several
countries. Moreover, some developing countries do have quite high statutory marginal
income tax rates for high-income groups, and sometimes also for middle-high income earners
such as households with two or three times the national average. For instance, according
to a study by Sicat and Virmani (1986) referring to the situation in the early eighties of
married couples with one earner and a "standard" number of children (three),
about half out of fifty developing countries studied had marginal income tax rates above
30 percent for incomes three times the average, and about a third of the countries had
marginal income tax rates of at least 40 percent for that (relative) income bracket. (The
mean marginal rate was 34 percent for that income bracket, as compared to 19 percent for
the middleincome bracket.) If separate statistics had been available for the formal sector
in urban areas, we would certainly have found much higher figures. Of course, legal
avoidance and illegal evasion mean that many taxpayers do not in fact pay those rates,
even remotely. However, the statutory marginal rates are an indicator of the incentives to
avoid and evade, which is part of the process by which marginal tax rates distort the
allocation of resources and human effort.
Of course, an overhaul of the tax system makes perfectly good sense even
without a shift to a more market-oriented system. However, a reform of the price and
incentive system is clearly more significant for a pronounced market-oriented system than
for a highly regulated and centrally planned system, as the gains of shifting to a more
marketoriented system cannot be fully achieved without reforming the price and incentive
system in conformity with efficiency criteria. Indeed, this truth has been well
illustrated in recent decades by the attempts of socialist countries in Eastern Europe to
rely more heavily on markets and
. 122 .
economic incentives. Moreover, if capital movements too are
liberalized, it is important to adjust capital taxation to levels that make domestic and
foreign wealth owners and firms willing to invest enough in the country in question,
rather than abroad and to allow remittances of earnings, as well as providing guarantees
of property rights in general. Stability of the domestic currency is, of course, also
crucial in this context. Indeed, the problem of "capital flight" in developing
countries is to a large extent a "confidence problem" concerning property rights
and the value of the domestic money.
What, then, are the most important specific changes to be considered in
the tax system when an economy shifts from regulations and central command to increased
reliance on markets? Some economists may want to base their policy advice on sophisticated
calculations of optimal tariffs, taxes, and subsidies. There is no doubt that the
literature on optimum taxation has helped us understand the general problems of taxes and
subsidies in cases where compromises have to be made between the requirements of tax
revenues, on one hand, and the losses of economic efficiency owing to the "excess
burden" of positive marginal tax rates on the other. The literature on optimum income
taxation has, for example, given precision to the old idea that marginal tax rates should
be higher, the smaller the elasticity of effort with respect to rewards is, ceteris
paribus. And the literature on optimum commodity taxation has formalized old views
among economists about how to make a compromise between the allocative efficiency of
consumption and concern for the distribution of income. While in the interest of economic
efficiency tax rates should be relatively high on goods and services for which the demand
and supply elasticities are small, the rates should, for distributional reasons, be high
also on goods and services that play a relatively important part in the consumption
pattern of groups of households that are supposed to be discriminated against in
redistribution policy; these groups are often, of course, high income earners. Taxes
should, ceteris paribus, also be high on goods and services that for the consumers
closely complement untaxed, or indeed untaxable, goods and services, such as leisure.
Quite complex formulas have in fact been derived to strike a balance between these
different, often conflicting, aspects, using a Social Welfare Function as the criterion
for the trade-off (Atkinson and Stiglitz 1980; Stern 1984).
However, there are strong objections to the strategy of using such
calculations as a basis far actual policy advice in developing as well as in
. 123 .
(1) First, formulas of optimum taxation catch only one, or possibly a
few, types of mechanisms for adjustment by the individual agents, such as a shift between
leisure and work, and/or between the consumption of different commodities. In reality
there are, of course, myriads of other adjustment mechanisms for taxes, such as variations
in the amount of do-it-yourself work and the intensity of work; adjustment of the level of
saving and the composition of portfolios; changes in the level and type of investment in
physical and human capital; changes of profession, workplace, or location of residence;
emigration across national borders; the use of more time to search for tax loopholes, or
even for cheating with taxes. Formulas that would simultaneously reflect all such major
adjustment mechanisms, or most of them, are quite simply beyond the range of useful
(2) Second, even to derive an optimum tax formula that takes into
account just one single type of adjustment mechanism, such as the choice between
leisure and income, or between different consumer goods on the demand side, it is in fact
necessary to rely on extremely special assumptions, such as identical preferences of all
individuals, and special forms of the production function, such as Cobb-Douglas functions.
(3) Third, all optimum tax formulas, even rather modest ones, require
statistical information that is not very reliable. Not only do we need an
"arbitrarily" chosen Social Welfare Function, but also information about
individual capabilities and preferences specific enough to quantify the sensitivity of the
response to contemplated tax rate changes of the various types of adjustments that are
supposed to be covered in the study. On most of this we shall never get sufficiently
reliable information. This is serious, as the tax rates that are derived from optimum lax
formulas are very sensitive to alternative specifications of the various functions and the
Indeed, if all the necessary information on individual capabilities and
preferences that is required for empirical application of the theory of optimum taxation
were available, we would not be too far from the type of knowledge that is necessary to
design lump-sum taxes and transfers, and hence avoid the economic distortions that are the
reason for choosing an optimum tax approach in the first place! More specifically, in
order to design optimum tax systems, we would need information about both
. 124 .
individual abilities and preferences, whereas it is simultaneously the
difficulty in obtaining information about matters like these that is the basic reason why
we are not able to use lump-sum taxes, and hence why there is a case for a second-best
solution by way of "optimum taxation."
(4) Applications of the theory of optimum taxation are also confronted
with a severe aggregation problem. More specifically, the tax rate that is assigned to a
specific product in the context of an optimum tax formula depends crucially on the degree
and type of aggregation of commodities. For instance, if furniture is put into the group
of durable consumer goods, it gets one tax rate, but if it is put into some other group of
goods, it would get a different rate or even a zero rate or a subsidy. In other words, the
tax rate of an individual good will be highly arbitrary depending on which other goods and
services it is lumped together with. This means that optimum taxation will to a large
extent be "arbitrary taxation." It is in reality also difficult to group the
goods in such a way that only consumers' goods are included in the tax base. Many inputs
in the production process will in fact also be taxed, which means that distortions of the
allocation of resources will arise from the production side as well, without these
distortions being considered in the calculation of the optimum tax structure. Attempts to
differentiate the tax rates of one and the same type of good when used for direct
consumption and when used as an input in the production process are bound to raise severe
problems of administration and evasion.
Several of these difficulties with optimum taxation are, of course, well
known by the adherents of optimum taxation. But it seems to me that they have not taken
these problems seriously enough when ruling out "traditional" principles of
uniformity of commodity taxes and tariffs, as well as "comprehensive" income
taxation with similar tax rates for all sources of income (for instance, the so-called
(5) However, there is an even more fundamental objection to using
optimum taxation as a basis for policy advice. There is no reason to assume that
politicians and public administrators would follow advice that is based on calculations of
optimum taxation by the help of some (assumed) Social Welfare Function. Politicians have
their own targets and ambitions, which may not bear much relation to the ideas that lie
behind calculations by economists of opti-
. 125 .
mum tax or tariff structures. Indeed, this is exactly the background
for various attempts in recent years to endogenize the behavior of politicians and public
administrators, as well as for suggestions tying the behavior of politicians to various
types of "rules." More generally, there is no reason to assume that tax rates
that are the outcome of political processes with conflicts and compromises between various
political parties and interest group organizations would be much correlated to the tax
structure that some economists may derive from optimization calculations.
This point about political behavior is important regardless of whether
the main ambition of politicians is to satisfy some strong interest group, to increase the
probability of being elected, or to adhere to some personal or ideological idiosyncrasy.
Politicians and public administrators, with the help of their economic advisors, can
always present some reasonable-sounding argument for their particular choice of a
differentiated structure of taxes and subsidies for instance, by referring to aspects that
have not been considered in the calculations of the optimum tax specialist, or by
exploiting the wide choice of "reasonable" elasticities of the demand and supply
responses to taxes. The fact that calculated "optimum rates" for individual
goods depend crucially on how goods are aggregated also opens the possibility of various
interest groups arguing about the "proper" way of aggregating goods in official
If optimum tax theory then is not the most appropriate basis for tax
policy advice in developing countries (or in developed countries for that matter), what
types of considerations should be used instead? The general answer, in my judgment, is
that it is better to rely on approaches that are less ambitious and less demanding
concerning knowledge about private behavior patterns, statistical information, and
administrative competence, but instead more ambitious with respect to basic insights about
the functioning of the political process. In other words, it is important to focus on more
"pedestrian," practical, and "commonsense" aspects than those
emphasized in the optimum tax literature. For instance, the following types of tax reforms
are worth considering in a great number of developing countries:
(1) As several developing countries, in particular those in Latin
America, often have high and highly variable rates of inflation, an
. 126 .
adjustment of tax assessment and tax collection to inflation is often
crucial. Though inflation functions as an implicit "inflationtax" on the stock
of money and government bonds (with less than fully adjusted interest to inflation),
inflation often also implies that the government loses "explicit" tax revenues
in real terms owing to the timelags in tax collection (the so-called OliveiraTanzi
effect). Inflation often also erodes the tax base because of the deductibility of nominal
interest rates, which in an inflationary situation means that part of the amortization of
the debt, in real terms, can be deducted from the tax base. If the latter two effects
dominate the former, inflation will generate (or accentuate) higher budget deficits (in
real terms), which then often feed back into even higher inflation rates. Obvious reforms
to solve these problems are to shorten the time lags in tax collection; make inflation
adjustments to the tax rates (or tax brackets); and redefine the tax base in order to make
a distinction in real terms between (deductible) interest payments and (nondeductible)
(2) As the tax base is often very narrow in developing countries, the
tax rates are often relatively high for certain sectors and groups of taxpayers. For
instance, in many developing countries today income taxation is, in fact, mainly a tax on
public servants and the employees of large corporations rather than on capital owners or
on employers and employees in agriculture or in the "informal" urban sectors.
The tax system is also plagued with other types of "asymmetries" with respect to
the effective tax rates for different agents and sources of income, and these asymmetries
are often accentuated by inflation. These asymmetries could, of course, be mitigated by
moving in the direction of a Haig-Simons type "comprehensive" income tax, with
uniformity between different sources of incomes, different assets, and different types of
income earners. Basically that would mean a broadening of the base and a lowering of the
rates as has recently been tried in some developed countries. The tax system would then
most likely be improved in terms both of efficiency and equity (and perhaps even
equality), whereas attempts to differentiate based on optimum tax formulas are likely to
be exploited by various interest groups. In other words, uniformity, as the basic rule of
taxation, is probably less vulnerable to manipulation by powerful interest groups than is
the principle of "differentiation" according to optimum tax principles.
. 127 .
(3) As the small degree of household saving that exists in most
developing countries largely takes place among the very top income earners, attempts to
redistribute income from these groups on a large scale are likely to conflict with
ambitions to stimulate private saving and to increase the supply of risk capital. This is
a reason for being careful about heavy increases in income and wealth taxes for
upper-middle and high income groups. Moreover, as corporate saving plays an important part
in aggregate saving in some developing countries, and could play an even more important
part in the future, there is a strong case also against raising taxes drastically on
Superficially, it may be argued that tax increases that reduce private
saving are not really a problem, as it does not matter if saving is done by private agents
or public authorities. However, this argument is seriously flawed for two reasons. First,
it neglects imperfections in the political process in the sense that tax increases
originally designed to increase public saving often, in fact, release increased public
consumption or subsidies of private consumption. Second, public saving is not a
perfect substitute for private saving in market-oriented economies, as private saving
contributes to a decentralization of decisions regarding investment, the entry of new
firms, and innovation. To keep down private consumption is not the only purpose of saving;
another important role, which it plays in market-oriented economies, is to allow and
stimulate the emergence of a system of decentralized decision making, and hence to
help channel resources to alternative types of assets in an efficient way.
Indeed, it may be argued that one of the most important prerequisites
for a successful shift to a more market-oriented economic system is just this, namely the
stimulation of private saving a conclusion that follows from informal commonsense
views on the functioning of markets (such as those of Hayek and Schumpeter), rather than
from formalized general equilibrium theories.
In fact, the history of economic development in the West during the past
few hundred years illustrates vividly the importance of private saving and private supply
of capital for the entry and growth of new firms, for "entrepreneurship," and
hence also for innovation. In other words, it is difficult to preserve an important role
for the entrepreneur if the private capitalist is "destroyed" partly because
these are often the same persons. Thus, while both centrally regulated economies and
market-oriented economies have to be careful not to destroy incentives to work, it is also
important in market-oriented economies to watch out
. 128 .
for disincentive effects on private saving and entrepreneurship. This
means, of course, that certain "sacrifices" have to be made in the ambition to
redistribute incomes from high- to low-income groups though less so in the case of
redistributions from economically "passive" groups, such as
"traditional" wealth holders who keep their assets in idle land and various
types of collectors' items.
What, then, are the conclusions for indirect (commodity) taxation? My
basic assertion is that if we opt for a reasonably nondistorting structure of taxes,
tariffs, and subsidies, it is advisable to choose, at least as a starting point, a uniform
structure (same tax rate on every commodity and the same tariff rate on every importable),
rather than attempting to find some optimum tax structure. The rationale is simply to
avoid a situation where something even further away from an optimum structure will in fact
be implemented by way of party competition and the influence of strong interest groups. It
is difficult for voters to judge if a highly differentiated tax structure reflects an
attempt to implement "optimum tax rates" or if it is simply designed to assist
politically powerful pressure groups. It is probably easier for voters to judge on this
matter if the norm is a proportional structure of indirect taxes rather than some asserted
Moreover, it is likely that highly selective taxes and subsidies, like
direct regulations, breed both corruption and "rent seeking" via political
lobbying (Myrdal 1968). This means that while a liberalization of the economic system in
developing countries is most likely a necessary requirement for a drastic removal
of rent seeking and corruption, this outcome could partly be jeopardized if highly
selective taxes and subsidies are introduced as suggested by proponents of optimum tax
An obvious objection to this reasoning is that it may, in fact, be
difficult to induce politicians to follow a norm about uniform indirect taxes and tariffs.
However, it is well known that politicians sometimes may find it in their own interest to
"straightjacket" themselves by accepting various types of norms obvious examples
being international GATT rules on tariffs. Indeed, commitment by way of binding rules has
been discussed frequently in game theory in recent years, emphasizing that this may be a
method of preventing various groups in society from exploiting the difficulties
experienced by government, without such rules, in committing itself in advance to a policy
it would like to pursue. A braver strategy would be to use a uniform tax structure as the
basis for the tax system, but to allow some additional selective taxes on goods for which
there are really strong and uncontroversial reasons to believe
. 129 .
(1) that the supply and demand elasticities are very low, and (2) that
the goods are consumed proportionally much less among citizens with low incomes than among
high-income groups (assuming that an improvement in incomes of low-income groups relative
to high-income groups is desired). Of course, there is then an equivalent case for
deviating from the basically proportional tax structure by low tax rates (or even
selective indirect subsidies) on goods and services that there is overwhelming reason to
assume have the opposite characteristics.
Such a modification of a strategy of uniform indirect taxes would be a
modest attempt to accomodate some of the basic ideas of the optimum tax theory without
using that theory as the basic foundation for tax policy recommendations. It differs from
the idea of optimum taxation in the sense that (1) attempts to adjust tax rates to
differences in demand and supply elasticities would be the exception rather than the rule,
and (2) considerations of the functioning of the political process would be paramount.
However, in a short- and medium-term perspective, the most important
aspect of tax reform in developing countries is probably to improve tax collection and
tax compliance. Not only is the administrative capacity of the tax authorities often
weak, but in addition, firms are often small and difficult to control, and loyalty to the
national state is often rather limited in some developing countries owing to historical
experience. Greater uniformity and less differentiation in the treatment of different
taxpayers, products, and sources of incomes may often facilitate both tax administration
and tax compliance. Moreover, it may be advisable to use a sales tax on wholesale trade
rather than a comprehensive value-added tax or a sales tax in retailing. Reductions in the
element of "arbitrary discretion" by local tax collectors may also help to
increase voluntary tax compliance, as would implementation of the earlier suggestion about
a broader base and lower rates.
To summarize my general points on tax policies: policy that relies on
(1) sophisticated analytical techniques, combined with (2) extreme oversimplifications of
the functioning of the economic system, (3) enormous requirements of sophisticated
statistical information, and (4) total neglect of the functioning of the political and
administrative system is likely to create more distortions than simple rules of thumb
using uniform and broadly based taxes and tariffs possibly modified by selective taxes or
subsidies where the case for such a modification is particularly strong. Thus, there is a
strong case for the "traditional" recommendation in public finance of a
"comprehensive" income tax with uniform
. 130 .
rates between different sources of income assets and types of income
earners, and a similar case for uniformity of the tax rates for commodity taxation as the
basic starting point, though exceptions may be made where strong cases can be put forward.
Growth, Distribution, and Poverty
Allocative efficiency and economic growth should not, of course, be
regarded as "ends in themselves," but rather as means of raising the material
well-being of "the ordinary man," and in particular of the poorest fraction of
the population. This observation raises the classical question of the relation between the
distribution of income, on the one hand, and allocative efficiency and economic growth on
the other, hence highlighting the celebrated conflict between equity and efficiency
"the Big Trade-off" in Arthur Okun's terminology (Okun 1974). However, it is
also obvious that equity and efficiency considerations are in many cases complementary
rather than conflicting objectives. Indeed, it is important to try to find strategies and
instruments of growth and redistribution policy for which such complementarities exist.
The most solid empirical observation on the relation between growth and
distribution is perhaps that the fruits of economic growth, at least after a while, tend
to become dispersed enough to result in an increase in the standard of living of both the
great majority of the population and of the bulk of low-income groups (Kuznets 1955,
1963). Thus "immiserizing growth" (Bhagwati 1985) seems to be quite unlikely in
a long-term perspective.
However, there is also rather strong evidence from time-series analysis
that the relative overall distribution of income (as measured by, for instance, the
Gini co-efficient) often becomes more uneven during the very first decades of economic
growth, with a reversal of this tendency later on. This is, of course, the empirical
background to the celebrated "Kuznets' Law" about the inverted U shape of the
relation between per capita income and the inequality of the overall distribution of
income. Indeed, for a given point of time, cross-country studies, too, support the
hypothesis of such an inverted U relation (Ahluwalia 1976 and references in Bigsten 1986).
The usual theoretical explanation for this asserted empirical regularity
is that economic growth to begin with tends to be concentrated in the initially very small
"modern" sector of the economy where per capita income is higher and more
uneven, and often also tends to rise more rap-
. 131 .
idly than in the traditional (usually rural) sectors. When, later on
during the growth process, the modern sector becomes a larger share of the total, and the
intrasector distribution of skills tends to become more even, the overall distribution of
national income, too, tends to become more even possibly also more even than in the
"initial" situation before the "take-off" of modern economic growth
A similar pattern seems to hold if we look at the relative position of
low-income groups (such as the very lowest income deciles) rather than at measures of the
overall distribution of income: though low-income groups often seem to gain absolutely
also during the very first few decades of economic growth, they usually seem to lose
ground relative to other groups during that phase of economic development.
Thus, though there is hardly any reason to be pessimistic about the
possibilities of raising the standard of living of low-income groups by way of economic
growth, both concern for the relative positions of people during early phases of economic
growth and an eagerness to help the very poorest certainly make a case for deliberate
policy actions to help the fruits of economic growth spread to the poorest fraction of the
population. Thus Kuznets' Law should be regarded as a "tendency" of given
economic and social policies, rather than some "iron law of distribution" that
cannot be repealed by appropriate economic and social reforms.
When discussing such reforms, it may be useful to distinguish between
four (closely related) aspects: (1) attempts to redistribute the ownership of human,
financial, and physical assets in favor of low-and low-middle-income groups; (2) removals
of institutional obstacles keeping these groups from participating in the process of
income growth; (3) redistributional considerations when designing general economic
policies; and (4) fiscal policy actions specifically designed to improve the living
standards of people in the above mentioned groups.
(1) It is true (and practically tautological) that a relatively even initial
distribution of human and physical capital, in particular of land, helps to spread the
fruits of growth widely. However, it would seem that successful land reforms have to
fulfill two requirements: they should opt for private ownership, largely in the form of
family farms, and they have to be once-and-for-all actions, so
. 132 .
that the owners of land can be sure about their property rights.
Otherwise a serious conflict between equity and efficiency considerations easily arises in
agriculture. We would also expect educational reforms during a process of growth to even
out the overall distribution of income, in particular if there is a strong emphasis on
literacy and vocational training for broad population groups.
(2) Obvious examples of methods to remove institutional obstacles
to a dispersed distribution of the fruits of economic growth is the stimulation of the
buildup of credit market institutions that reach low-and low-middle-income groups, in
particular in rural areas. In the case of farm families characterized by ample
availability of labor and scarcity of land, it is also important to remove legal and
institutional obstacles for additional land tenure (such as the leasing of land). It is
also obvious that the possibilities of poor farmers participating in the process of income
growth may be drastically improved by government infrastructural investment, both to
increase the productivity on the farms (for example, by irrigation systems) and by
creating a favorable infrastructure for nonfarm activities in the countryside, where
family labor may get additional earnings.
(3) The distributional consequences of alternative general policy
strategies is a more difficult and controversial matter. However, both casual observations
and systematic research indicate that a shift to an outward-oriented growth strategy tends
to favor not only overall economic efficiency, but also employment and redistribution of
income to unskilled labor (Krueger 1978; Bhagwati 1978). One of the explanations is that
import protection in manufacturing turns the terms of trade against agriculture, where the
majority of poor people are to be found in most developing countries. Another explanation
is that such policies tend to increase competition and hence reduce monopoly profits.
However, even more fundamentally, an outward-looking growth strategy in labor-abundant
countries favors labor-intensive sectors and labor-intensive production techniques simply
because free trade tends to allocate factors of production in conformity with comparative
advantages, which tends to turn the composition of national income in favor of labor
income, in particular for low-skilled workers. Poverty will then tend to be reduced, as
there is overwhelming evidence that increased employment is of utmost importance of the
removal of poverty (Fields 1984). Indeed, as
. 133 .
pointed out by Ranis (1978, 407), "The only sure method of
achieving a sustained improvement in equity lies in hastening the advance of
commercialization, i.e. the end of the labor surplus conditions."
The importance of high demand for labor for an even distribution of
income, and for the mitigation of poverty, also creates a strong case for fighting various
regulations in the labor market that keep low-skilled labor unemployed. Obvious examples
are minimum-wage legislation and wage policies by unions with similar characteristics,
though such legislation and policies do help some low- and middle-income employees if they
do not lose their jobs by way of such actions. Here then is another illustration that
efficiency and equality do not always conflict.
However, when stimulating employment, it is, from the point of view of
distributional considerations, important to avoid using methods, such as aggressive demand
expansion, that generate high and fluctuating inflation, which easily hurts low-income
groups. This is particularly important as inequalities generated by inflation are usually
"nonfunctional," in the sense that they do not contribute to efficient economic
incentives (probably the reverse), in contrast to inequalities that are caused by
differences in productive effort. As high-inflation economies tend to get large
distortions of relative prices, including the real exchange rate, here is another example
where allocative and distributional aspects are complementary.
The upshot of all this is that those who, for efficiency reasons, are in
favor of outward-oriented development strategies have certainly no reason to be shy about
their position from the point of view of the distribution of income rather the opposite.
Recent experiences in countries like South Korea and Taiwan illustrate this point. There
countries have demonstrated the possibility of reconciling efficiency and distributional
considerations both by choosing an outward-looking development strategy and
by making early redistributions of the ownership of land and human capital. Thus, even if
examples are easy to find where specific tools of egalitarian policies, such as high
marginal tax rates, do harm economic growth, there is no presumption of an unavoidable
negative reverse causation, according to which increased economic growth would necessarily
be unfavorable for income equality even in a short- and medium-term perspective.
. 134 .
The situation is more complex for countries with abundant natural
resources rather than labor, as a growth process based on comparative advantage in this
case tends to result in high rents. It is then often important to achieve institutional
conditions, including a well-functioning capital market, dynamic entrepreneurship, and tax
and expenditure policies that help these rents to flow to other sectors where the country
has a comparative advantage. rather than using these rents for subsidies to
(4) Of course, various arguments for institutional reforms in favor of
low-income groups, redistribution of assets, and an outward-looking development strategy
do not rule out the possibility of more direct policy actions specifically designed to
improve the consumption of poor people. As pointed out in the section on expenditure
policies, one obvious example is the provision of water, sanitation, health care, and
food. It is. however, important to warn countries against choosing methods of
redistribution that harm growth, as after a while economic growth usually tends to be
accompanied by an equalization of the overall distribution of income.
Problems of Transition and Macroeconomic Instability
So far our discussion has been confined to various aspects of the
allocation of resources and the distribution of income. However, when analyzing public
finance, or economic policy in general for liberalizing developing countries, there is a
strong case for emphasizing stabilization policy aspects as well. For instance,
greater reliance on foreign trade, which is an expected consequence of more
market-oriented economic policies, would be expected to accentuate the size of
"imported" disturbances from world markets, though at the same time domestically
generated disturbances will have smaller effects on the domestic economy, as part of the
effects "leaks out" through the balance of payments.
However, it is conceivable that a market-oriented and outward-looking
country will be so much more flexible than a highly regulated one (with the emphasis on
import substitution) that the effects on the domestic economy of foreign shocks
will not be greater in the former than in the latter type of country. Indeed, Bela Balassa
(1984) asserts that a number of highly "outward-oriented" economies, in
particular in Pacific
. 135 .
Asia, though being exposed to greater international shocks than most
other developing countries, have recently been able to "absorb" the shocks
better than more regulated, "inward-looking" developing countries. However, it
is equally striking that attempts to liberalize the economies of some developing
countries, in particular in South America, have backfired because of their inability to
deal with problems of macroeconomic instability during the period of transition to a more
market-oriented economic system. These important issues will be dealt with below.
One obvious problem during a period of transition to a more
market-oriented economic system is that the redistribution of income and wealth, which is
induced by the accompanying shifts in relative prices, profitabilities, and the
composition of demand, will be resisted by various interest groups in society. New market
uncertainties will often also be created during the transition period. One reason is that
when an economy is originally in a situation characterized by pronounced disequilibria,
which is a characteristic feature of a regulated economy, it becomes difficult to predict
what will happen to various relative prices, demands, and supplies when the economy is
liberalized. It is therefore important that there be confidence among private agents that
the shift of system is permanent, as otherwise economic agents may not be willing
to fully adjust their activities, in particular their investment, to the new information
and the new incentives that are transmitted via the market system. An important
explanation as to why the economic liberalization in West Germany in 1947â50
and in Pacific Asia in the 1960s and 1970s was so successful is probably just that there
was considerable confidence that the shift to new rules of the game was permanent. More
efficient market-oriented institutions in product and factor markets would also help
smooth the transition, obvious examples being highly flexible credit institutes and labor
exchange agents. It would also be useful if the government could adjust the public
infrastructure rapidly, it possible even in advance of the liberalization of product and
Another obvious transitional problem is the emergence of severe risks of
large increases in frictional and structural unemployment, as the requirements of
reallocations of labor may outstrip the ability of the labor market to achieve this
smoothly. Thus, it may be hazardous to liberalize product markets without at the same time
removing important obstacles to the flexibility of the factor markets. From that point of
view it would be wise if, for instance, minimum-wage regulations and interest-rate
ceilings were removed before, or simultaneously with. a liberalization of
. 136 .
the product market. Recent studies indicate, however, that severe
unemployment problems have not been caused by the liberalization attempts during the
seventies except in a few countries that pursued stabilization policies that severely
damaged the employment situation (Michaeli 1986).
Indeed, experience suggests that the most important transition problem
concerns macroeconomic instabilty the difficulties of avoiding fluctuations in capacity
utilization, inflation, and the balance of payments. For instance, inflation and balance
of payments problems will most likely be accentuated when price controls and other types
of regulations are removed. Indeed, this is exactly what has happened when socialist
countries have experimented with a freer system of price and wage formation. The obvious,
and generally accepted, conclusion is that a liberalization of the economic system has to
be combined with increased concern for the management of stabilization policy with fiscal,
monetary, and exchange-rate policies as the main tools.
However, many years of experience in developed countries certainly show
that even rather well designed macropolicies may not be enough to prevent high, and
perhaps even rising, inflation. A main reason for this is that the mechanisms of wage
formation are characterized by a pronounced inflationary bias in most, perhaps all, of the
world's market economies. Some adjustment of these mechanisms is therefore worth
considering when developing countries plan to liberalize their economic systems (in the
same way as such modifications are today discussed in various developed countries). This
argument may carry particular weight in countries with fairly strong labor unions, as in
some Latin American countries.
Obvious candidates for such reforms are: removals of price index clauses
in wage contracts, and the introduction of new contract forms with bonus systems that tie
wage increases to productivity increases or profits. When developing countries introduce
unemployment insurance systems, a strong case can also be made for letting unions and
firms in each separate sector bear the bulk of the insurance costs, rather than shifting
these costs onto the taxpayers. The idea is, of course, to internalize the negative
consequences on the employment level of aggressive wage increases. Clearly, government
budget policy has a direct responsibility for the rate of wage increases in the public
sector, which in some developing countries has an important bearing on wages in the entire
labor market. However, in some developing countries, where labor unions are weak, general
macroeconomic policy with a concentration on getting the budget surplus, the monetary
aggregates, and the ex-
. 137 .
change rate right may be not only a necessary but also a sufficient
condition for reasonable macroeconomic stability if international disturbances are not too
In countries with large fluctuations in the terms of trade (owing, for
instance, to heavy exports of raw materials), there is also a strong case for pursuing
monetary and fiscal policies that avoid large swings in the real exchange rate for
nontraditional exports to prevent a "Dutch disease" in that sector in connection
to export booms. The "sterilization" of extreme export earnings (for example, by
taxation) or of their consequences for domestic financial markets (by monetary policy
actions) to reduce instability in the real exchange rate is a crucial aspect of a
successful stabilization policy for such countries.
There are also a number of specific transition problems connected to the
liberalization of foreign transactions trade as well as capital movements. For instance,
as a tariff reduction initially reduces the prices of import goods
("importables") relative to both export goods ("exportables") and
nontraded goods ("nontradables"), both domestic absorption of importables and
the production of nontradables are stimulated. In other words, there is an appreciation in
the "real exchange rate." In combination with the reduced competitiveness of
domestic production of importables, we would, in short-term perspective, expect temporary
unemployment problems in that sector, and most likely also an increased current account
deficit. This is, of course, the rationale for the traditional policy advice to devalue
the currency in connection with a general tariff reduction implying a slowing down (or
reversal) of the appreciation of the real exchange rate. An additional complication is
that the inflationary effects of the devaluation may dominate the deflationary effects of
the tariff reduction, which is, of course, a reason to suggest that the devaluation be
combined with restrictive demand management policies.
As direct regulations regarding imports are often more disruptive to the
efficiency of the economy than tariffs, it is worthwhile starting a process of import
liberalization not only with an "evening out" of the structure of tariffs
between different groups of commodities but also with liberalizations of direct import
controls. Attempts to reduce the average level of tariffs could then be delayed
somewhat if rapid devaluations prove to be difficult.
With sufficiently restrictive management of domestic aggregate demand,
the combination of tariff reduction and devaluation could then, in principle, bring about
the desired shift of resources not only from less to more efficient parts within the
tradable sector, but also from the
. 138 .
nontradable to the tradable sector, with an expansion of foreign trade
without a deterioration of the current account and without increased inflation.
However, if nominal wage increases (owing, for instance, to explicit or
implicit indexation of wage rates) follow price increases closely, the result may be both
stagflation and (with fixed exchange rates) a gradual disappearance of the gains in
competitiveness that the devaluation was designed to provide. What would then remain of
the entire operation would be a more efficient level and structure of tariffs, and hence a
more efficient allocation of resources, though at the cost of more inflation and possibly
also a deterioration in the current account of the balance of payments. Both the higher
rate of inflation and the increase in the current account deficit then easily release
political forces that result either in the reintroduction of a protectionist stance or in
strongly restrictive unemployment-creating policy actions. All this highlights the
importance of an appropriate combination of trade liberalization and macroeronomic policy
during a transition period.
Of course, an increased current account deficit could, in principle, be
financed by capital inflows, which are facilitated by the liberalization of capital
movements, in particular if domestic interest rates are initially higher than the rates on
world markets. However, such a development requires both high confidence in the permanence
of the liberalization of capital movements and a sufficiently high real rate of return on
domestically held assets, implying an appropriate and stable real exchange rate. As such
conditions cannot always be taken for granted, a country that is about to liberalize
foreign exchange should build up its reserve and/or lines of credit in advance, for
instance by way of foreign borrowing by the government.
It would seem that the unsuccessful liberalization attempts in some
countries in South America in particular, in the so-called Southern Cone countries, Chile,
Argentina, and Uruguay in the late seventies and early eighties were closely connected
with serious deficiencies just in short- and medium-term macropolicies. These policies
were simply not "consistent" enough with the liberalization attempts (Khan and
Zahler 1984). An overexpansion of domestic aggregate demand, often via a large budget
deficit and a rapid increase of the money stock (partly to finance the budget deficit,
partly owing to large capital inflows), during the process of liberalization contributed
both to inflation and to rising current account deficits, which then resulted in severe
liquidity and confidence problems for the national economy.
. 139 .
In these countries expansion of domestic aggregate demand pulled up
prices in the nontradable sectors and wages in the entire economy, which with fixed
exchange rates severely harmed production and employment in the tradable sectors. This
also illustrates the dangers of trying to fight inflation with a fixed exchange rate if
domestic aggregate demand policies are not kept under control in particular, if wages are
indexed to the domestic price level for consumer goods (with a strong component of
nontradables). The experiences of Argentina and Chile during the late seventies illustrate
Thus, the most important aspect of the transition problem is probably to
avoid combining a liberalization of trade and capital movements with (1) an expansion of
domestic aggregate demand by way of big budget deficits and a large expansion of the
monetary aggregates, and (2) unstable real exchange rates, in particular heavy
appreciations, for instance, owing to rapid wage increases with fixed (or lagging)
exchange rates. Such policies have, of course, particularly serious consequences in an
international environment characterized by drastically increased interest rates, a
deterioration of the terms of trade, and a cyclical weakening of foreign markets, as in
the early eighties.
However, it is also difficult to introduce trade liberalization in a
situation in which the government tries to bring down inflation by restrictive demand
policies, which will accentuate the unavoidable disturbances in the labor market resulting
from the trade liberalization programs. This dilemma has induced some observers to suggest
that a trade liberalization program has to be postponed until inflation has been brought
down substantially (Bruno and Sachs 1985). It is also obvious that the order of trade
liberalization is important. For instance, the experience of Argentina illustrates the
danger of liberalizing imports before exports, and the experience of Chile illustrates the
risk of liberalizing capital movements at an early stage of the liberalization process if
effective action is not undertaken to prevent inflationary effects from capital inflows.
Finally, it is also important to avoid destabilizing macropolicies after
a period of transition to a more market-oriented economic system, as both allocation and
growth policies are likely to fail if the macroeconomic policy is not pursued with
reasonable skill. An open economy, in particular a small one with large foreign trade,
will be disrupted if aggregate demand is much too high, or much too low for that matter,
and if the most important relative price of all the real exchange rate is excessive.
Experience shows that this simple point cannot be stressed too much.
Ahluwalia, Montek T. 1976. "Income Distributions and
Development: Some Stylized Facts." American Economic Review 66 (May):
Atkinson, Anthony B., and Joseph E. Stiglitz. 1980. Lectures
on Public Economics, chaps. 12â14. New York: McGraw-Hill.
Balassa, Bela. 1983. "Public Finance and Social Policy
Explanations of Trends and Development: The Case of Developing Countries." In Public
Finance and Social Policy, 41â58. Proceedings of the 39th Congress of the
International Institute of Public Finance, Budapest.
. 1984. "Adjustment Policies in Developing Countries: A
Reassessment." World Development 12 (September): 955â72.
Bauer, P. T. 1971. Dissent on Development. London:
Weidenfeld & Nicholson.
Bhagwati, J. N. 1966. The Economics of Underdeveloped
Countries. New York: McGraw-Hill
. 1968. The Theory and Practice of Commercial Policy:
Departure from Unified Exchange Rates. Essays in International Finance no. 8.
International Finance Section, Department of Economics, Princeton University.
. 1978. Foreign Trade Regimes and Economic Development:
Anatomy and Consequences of Exchange Control Regimes. Cambridge, Mass.: Ballinger.
. 1985. "Growth and Poverty." Lecture delivered at
Michigan State University, East Lansing, April 4.
Bigsten, A. 1986. "Poverty, Inequality and
Development." University of Gothenburg. Mimeographed.
Bruno, Michael, and Jeffrey D. Sachs. 1985. Economics of
Worldwide Stagflation. Cambridge, Mass.: Harvard University Press.
Buchanan, James M., and Gordon Tullock. 1962. The Calculus of
Consent: Logical Foundations of Constitutional Democracy. Ann Arbor: University of
Chenery, Hollis B. 1975. "The Structuralist Approach to
Economic Development." American Economic Review 65 (May): 310â16.
. 1979. Structural Change and Development Policy. London:
Oxford University Press.
Fields, G. S. 1979. "Decomposing LDC Inequality." Oxford
Economic Papers 31 (November): 437â59.
. 1984. "Employment, Income Distribution and Economic Growth
in Seven Small Economies." Economic Journal 94 (March): 74 83.
Haberler, G. 1959. International Trade and Economic
Development Cairo: National Bank of Egypt.
International Labor Organization. 1976. Employment, Growth and
Basic Needs. Geneva: ILO.
International Monetary Fund. 1985. Government Finance
Statistics Yearbook. Washington, D.C.: International Monetary Fund.
Khan, M. S., and R. Zahler. 1984. Trade and Financial
Liberalization in the Context of External Shocks and Inconsistent Domestic Policies.
IMF Document DM 84/44. Washington, D.C.: International Monetary Fund.
. 141 .
Krueger, Anne O. 1974. "The Political Economy of the Rent-Seeking
Society." American Economic Review 64 (June): 291â303.
, ed. 1983. Trade and Employment in Developing Countries.
Vol. 3, Synthesis and Conclusions. Chicago: University of Chicago Press.
Kuznets, S. 1955. "Economic Growth and Income
Inequality." American Economic Review 45 (March): 1 28.
. 1963. "Quantitative Aspects of the Economic Growth of
Nations: III: Distribution of Income by Size." Economic Development and Cultural
Change 11 (January): 1â80
Lal, Deepak. 1983. The Poverty of Development Economics.
Hobart Paperback no. 16. Washington, D.C.: Institute of Economic Affairs.
Lindbeck, Assar. 1973. "Endogenous Politicians and the
Theory of Economic Policy." Seminar Paper no. 35. Stockholm: Institute for
International Economic Studies.
. 1985. "What Is Wrong with the West European
Economics?" World Economy (June): 153â70.
Little, I. 1982. Economic Development: Theory, Policy and
International Relations. New York: Basic Books.
Little, I., T. Scitovsky and M. Scott. 1970. Industry and
Trade in Some Developing Countries: A Comparative Study. London: Oxford University
Michaeli, M. 1986. "The Timing and Sequencing of a Trade
Liberalization Policy." World Bank papers prepared for conference in Lisbon, June
Morawetz, D. 1976. "Elasticities of Substitution in
Industry: What Do We Learn from Econometric Estimates?" World Development 4
Myint, H. 1967. "Economic Theory and Development
Policy." Economica 34 (May): 117â30.
Myrdal, Gunnar. 1968. Asian Drama: An Inquiry into the Poverty
of Nations. New York: Twentieth Century Fund. Reprint, Penguin Books.
Okun, A. M. 1974. Equality and Efficiency: The Big Trade-Off.
Washington, D.C.: Brookings Institution.
Psacharopoulos, G. 1981. "Returns to Education: An Updated
International Expansion." Comparative Education 17:321â41.
Ranis, Gustav. 1978. "Equity with Growth in Taiwan: How
'Special' Is the Special Case?" World Development 6 (March):
Schultz, Theodore W. 1964. Transforming Traditional
Agriculture. New Haven: Yale University Press.
Sen, Amartya. 1981. Poverty and Famines: An Essay on
Entitlement and Deprivation. New York: Oxford University Press.
Sicat, G. P, and A. Virmani. 1986. "Personal Income Taxes in
Developing Countries." World Bank. Mimeographed.
Stern, N. H. 1984. "Optimum Taxation and Tax Policy" IMF
Staff Papers 31 (June): 339â78.
Stiglitz, Joseph E. 1985. Economics of Information and the
Theory of Economic Development. NBER Working Paper no. 1. Cambridge, Mass.: National
Bureau of Economic Research.
. 142 .
Tanzi, V. 1986. "Quantitative Characteristics of the Tax System
in Developing Countries." In D. Newberg and N. Stern, eds., Modern Tax Theory for
Developing Countries. Washington, D.C.: World Bank.
United Nations. 1983. Industrial Yearbook. New York: UN.
Viner, J. 1952. International Trade and Economic Development.
Lectures delivered at National University of Brazil. Glencoe, Ill.: Free Press.
Waterston, A. 1965. Development Planning: Lessons of
Experience. Baltimore: Johns Hopkins University Press for the World Bank.
White, L. J. 1978, "The Evidence on Appropriate Factor
Proportions for Manufacturing in Less Developed Countries: A Survey." Economic
Development and Cultural Change 27 (October): 27â59.
World Bank. 1985. World Development Report, 1985
Washington, D.C.: World Bank.
. 1986. World Development Report, 1986. Washington, D.C;.: