I N T R O D U C T I O N TO PART 1

T.N. SRINIVASAN
Yale University

Part 1 of this Handbook is, appropriately enough, concerned with the concepts of
and approaches to the study of the economics of development. Development, as
contrasted with mere growth of the economy, is, according to Schumpeter, "a
distinct phenomenon, entirely foreign to what may be observed in the circular
flow or in the tendency towards equilibrium. It is spontaneous and discontinuous
change in the channels of flow, which forever alters and displaces the equilibrium
previously existing" [Schumpeter (1961)]. While Schumpeter wrote in 1911 and,
therefore, did not have contemporary developing countries in mind, economists
writing after the Second World War viewed the developing economies as being
caught in an equilibrium variously described as a "vicious circle of poverty",
"low-level equilibrium trap", and so on. The problem of development was
described as taking the economy out of this equilibrium and setting it on a path
of self-sustaining growth.
Unlike Schumpeter, who emphasized the role of innovations and the concom-
itant private anticipation of entrepreneurial profit as major factors explaining
development, most development economists writing in the late 1940s and 1950s
visualized a dominant role for the state in initiating and sustaining the develop-
ment process. A corollary view was the advocacy of comprehensive national
development planning under state auspices. This view, however, came to be
tempered considerably by the sobering experience of the decades following the
Second World War with disappointing, if not entirely counterproductive, state
interventions in the economy in one developing country after another.
In the dominant neoclassical paradigm of the postwar era, Adam Smith's
parable of the invisible hand was rigorously restated in terms of the two
fundamental theorems of welfare economics. First, under a set of mild restric-
tions on production technology and individual preferences, the equilibrium of a
laissez-faire market economy is a Pareto optimum. Pareto optimality obtains if
any departure from it cannot make everyone better off, that is, benefit for some
can come only at the expense or loss in welfare of others. The fact that a
laissez-faire equilibrium is Pareto optimal does not of course imply that it is fair
in some well-defined sense. More simply put, the distribution of income or
Handbook of Development Economics, Volume I, Edited by H. Chenery and T.N. Srinivasan
© Elsevier Science Publishers B. V., 1988
4 T.N. Srinivasan

welfare associated with a laissez-faire Pareto optimum could be highly unequal.

Second, gix;en a set of rather restrictive assumptions, including the absence of
technological externalities and increasing returns to scale in production of goods,
and convexity of consumer preferences, any Pareto optimum is also a competitive
equilibrium PrOvided income (or initial endowments) of individuals can be
redistributed through lump-sum transfers or other non-distortionary means.
Thus, both efficiency of resource allocation and distributional equity (in the sense
of Pareto optimality) can be achieved in a competitive equilibrium, albeit one
with redistribution of incomes or endowments, rather than one that is laissez-faire.
The "static" or "certainty" version of the theorems stated above was extended
by Arrow and Debreu to cover resource allocations in time, space, and in
uncertain states of the world. The essential feature of this extension is that a
complete set of markets for the purchase or sale of goods contingent on each
possible combination of time, space, and the (uncertain) state of nature is
assumed to exist. In both the original and extended versions there was no room
for strategic behavior and no costs of transactions by assumption.
Early development economists based their analytical case for state intervention
in the developing economies on the exceptions to the validity of the static version
of the second theorem. They argued that, in early stages of development,
externahties are pervasive and scale economies are significant, particularly in
sectors such as transport and communications which provide the infrastructure
for the functioning of an economy. Capital markets are likely to be segmented
and imperfect; and the financial system primitive and not performing intermedia-
tion to any significant extent. The entrepreneurial class is likely to be miniscule, if
not altogether absent or alien and unassimilated. But even if entrepreneurs were
believed to exist in adequate numbers, the markets for risk pooling and sharing
were thought likely to be absent, so that the scope for entrepreneurial activity
would be severely circumscribed. Under such circumstances, it was argued,
market failure was the rule rather than an exception. The state would then have
to perform the role of the Schumpeterian entrepreneur as well as intervene in
existing markets, through appropriate taxes and subsidies, to ensure that exter-
nalities were appropriately reflected in private calculations. Recent theoretical
work on development has taken the exceptions to the extended version of the
second theorem as its point of departure, in emphasizing the absence of a
complete set of contingent markets and asymmetries in information and strategic
behavior. This research investigates the origins and consequences of incomplete
markets, imperfect information, transactions costs, and imperfect competition.
One important consequence is that the allocation of resources through existing
markets may not be Pareto optimal, and once again a role for government
intervention can emerge.
In contrast to the neoclassical approach to development which is ahistoric and
based on virtual time, there is the historical, real-time Marxian approach that is
introduction to Part l 5

perhaps closer to the spirit of Schumpeter's analysis. In this approach, the forces
of production, represented by available technology at any point in historical time,
and the existing relations of production, represented by the institutions governing
ownership and access to the means of production, determine an "equilibrium".
Exogenous technical progress, i.e. shifts in the forces of production, disturbs such
an equilibrium inducing a realignment of relations of production that replaces
those of the pre-existing equilibrium which are no longer functional. In this linear
view of history, capitalistic production relations supplanted feudal relations in
the historical context because changes in technology (particularly after the
industrial revolution) made the feudal relation obsolete. In turn, the recurrent
crises that Marx so confidently predicted for mature capitalism were expected to
lead to socialism and, eventually, to communism. As will be noted below, the
recent rational-choice school of Marxism (perhaps it should be called neo-classi-
cal Marxism) does not view the world in the rigidly deterministic way described
above and allows some scope of action for individuals.
In the structuralist view of development, rigidities of various kinds preclude
the rapid and quantitatively significant response of the economy to changing
incentives as conveyed by market signals. Thus, for example, the supply response
to improved terms of trade for agriculture, the dominant sector in the early stages
of development, was said tb be negligible, at least in the short to medium run.
The prospects of acquiring growing amounts of f~reign exchange deemed essen-
tial for obtaining or financing vital imports through exports of traditional
products were considered poor and so on. Even if market failures in the
neoclassical sense were absent, development will be hampered unless strategies
(e.g. import substituting industrialization) that alleviate the structural rigidities
are adopted. Dualism between the traditional and rural agriculture and urban
and modern industries is another example of a structural rigidity. However,
neither the origin of the rigidities nor their persistence is adequately explained in
behavioral terms by structuralists.
All three approaches to development - neoclassical, Marxian, and structuralist -
are represented, at least minimally, in this Handbook. Part 1 elaborates some of
the more prominent features of each approach and compares their explanations
of some of the stylized facts of development. It is fair to say that most
development economists, and certainly the authors of this Handbook, would not
wish to be tied down exclusively to one approach. They would rather apply that
approach which is most suited to the analysis of each problem. Besides, the
divergence in the explanations of stylized facts as put forth by different schools
appears to be narrowing.
In much of neoclassical analysis, inter-personal comparison of welfare is
avoided so that for comparing situations in which the welfare of different
individuals may be affected differently only a partial ranking by appealing to
Pareto optimality is used. That is, unless in one situation every individual is
6 T.N. Srinioasan

better off (or worse off) compared to another, two situations are not ranked.
Sen's philosophical analysis of the concepts and indicators of development in
Chapter 1 makes it plain that a major difficulty with the concept of development
arises from value-heterogeneity, i.e. that people differ in their views as to what
things are valuable to promote, and value-endogeneity, i.e. in fact that the
process of change involved in development alters the valuations of people. Yet,
Sen strongly argues that it is often possible and desirable to separate the
relatively uncontroversial judgements from the controversial ones related to
value-heterogeneity and value-endogeneity. In his view much of development
policy analysis involves valuation problems that are not excessively problematic.
Arthur Lewis (Chapter 2) traces the historical roots of development thought in
the writings of economists of the pre-Adam Smith era and finds them to be
surprisingly modem in identifying constraints on growth imposed by agricultural
growth, foreign exchange, and saving. Rudiments of the theory of bank credit,
human capital, and public finance are also found in these early writings.
Bardhan (Chapter 3) compares three alternative approaches, which he terms
"neo-classical", "Marxist", and "structuralist-institutionalist" , to five broadly
defined problems: theory of household behavior, resource allocation, income
distribution and growth, trade and development, and economic policy and the
state. His discussion of Marxist approaches is heavily weighted in favor of a
so-called "rational-choice" Marxism as contrasted with traditional Marxism, the
former being closer to neoclassical economics in positing that individuals have
room for choice and are not completely constrained by technology and history.
He argues that although superficially the three approaches may appear very
different there is a significant measure of methodological similarity among the
more sophisticated versions of each. Institutions and modes of behavior such as
share-cropping and inteflinkages between markets that were once condemned by
Marxists as exploitative and by neoclassicists as inefficient, may turn out to be
rational responses to individuals to problems of risk, moral hazard, and/or costs
of monitoring work-effort. While Marxists tended to emphasize the role of
ideology as legitimizing exploitation or false consciousness as masking it, and
while neoclassicists ignored it altogether, recent work on political economy views
ideology as a resource-saving way of keeping free riders in check in the provision
of public goods of all kinds.
Another area where Bardhan argues that modern Marxian and neoclassical
analysis are converging is in explaining the existence of unemployment in
equilibrium. While Marxists suggest that the threat of unemployment disciplines
workers and enables the capitalists to extract labor power from them, neoclassi-
cists point out that compared to the alternatives of monitoring workers closely or
making them put up bonds for good behavior, the threat of unemployment may
be an instrument for reducing shirking by workers, thus extracting effort from
them at a reasonable cost. Another area of convergence of views can be discerned
Introduction to Part I 7

in the waning enthusiasm for centralized bureaucratic planning for economic

development and a grudging recognition of the efficiency of the market mecha-
nism on the part of some Marxists and structuralists and the recognition by
neoclassicists of a role for government for reasons other than the standard one of
market failure.
Labor markets are likely to be segmented and imperfect in early stages of
development when more than two-thirds of the labor force is employed in
agriculture. Development, almost by definition, involves a transfer of labor from
agriculture to manufacturing and services. Following in the footsteps of Arthur
Lewis, development economists have lavished analytical attention on this transfer
and the factors that influence it. Ranis (Chapter 4) elaborates the influential
dual-economy model of development in which an essentially market-oriented and
capitalistic urban manufacturing sector expands by drawing labor from an
agricultural sector based on traditional family farms. Until a unified labor market
emerges, the real wage in urban manufacturing is related to the agricultural wage
by such factors as rural-urban terms of trade, the cost associated with moving
from rural to urban areas, and any wedge introduced by public policy interven-
tion. The rate of capital accumulation in urban manufacturing determines the
pace of withdrawal of labor from agriculture in the Ranis model.
In no real world economy, developed or otherwise, does a complete set of
contingent commodity markets exist. The non-existence of some markets as well
as the imperfect functioning of others are characteristic of the state of underde-
velopment. Stiglitz (Chapter 5) analyzes the causes and consequences of these
phenomena. According to him the two important consequences are: first, equi-
libria based on transactions in existing markets need not be Pareto optimal, and
second, contractual arrangements other than "arm's length" market transactions
between anonymous parties may be widespread. The absent markets most often
cited are those related to insurance against risks and those related to financial
intermediation, broadly construed. One of the major factors contributing to the
non-existence of markets is the problem of asymmetric information leading to
moral hazard, adverse selection, etc. Stiglitz is able to provide a rationale for
share-cropping (often with a 50-50 split of output between the landlord and
tenant whatever be the crop grown), for inter-linking of share-cropping with
credit and marketing, and sharing of costs with sharing of output, etc. An
important but troublesome implication of his analysis is that although laissez-faire
equilibria in such contexts are not Pareto optimal, and thereby a role for
government intervention exists, there is no presumption that Pareto-superior
equilibria can always be achieved through government intervention.
Stiglitz is also concerned with efficiency wage theories. Efficiency problems
arise from a number of sources including: incentives for shirking due to imperfect
monitoring and supervision of workers, self-selection considerations in a context
of asymmetric information about worker productivity, and the possibility that
8 T.N. Srinioasan

workers' productivity depends on their nutritional status, the latter being posi-
tively related to their consumption out of their wages.
Taylor and Arida (Chapter 6) study the macroeconomics of growth by identify-
ing Schumpeter's circular flow equilibrium with long-run steady state growth.
Five versions of the steady state are elaborated, three of which are constrained
from the supply side and two from the demand side. The three supply-con-
strained models include Solow and Swan's neoclassical models of growth. In
these models the exogenous steady growth of labor in efficiency units determines
the growth rate of the system. In contrast, in another supplied-constrained model
the growth rate of one sector's output is exogenously specified to which rate the
growth of output of other sectors converge, inter-sectoral equilibrium being
maintained through relative price changes. In the demand-determined models of
Taylor and Arida, output and capacity utilization respond to changes in demand
with no presumption of full employment or full capacity utilization. The con-
straint on growth in demand arises because the pattern of income distribution
needed to sustain a full capacity eqtiilibrium differs from that associated with
existing production relations. Implicit in this story is a whole host of structural
rigidities, and inflexibilities that either prevent prices from adjusting or resources
from being reallocated in response to price changes. Taylor and Arida explore the
implications of rigidities of technology and in behavior of agents.

Reference

Schumpeter, J. (1961) The theory of economic development. New York: Oxford University Press.