Lance Taylor - Structuralist Macroecnomics - Applicable Models For The Third World-Basic Books Inc (1983)

APPLICABLE MODELS FOR
THE THIRD WORLD
Lance Taylor
ISBN O-MbS-DflEB^-M
Structuralist
Macro¬
economics
APPLICABLE MODELS FOR
THE THIRD WORLD
Lance Taylor
Economists in developing countries have long

realized that their options for policy and indeed
for economic change are severely limited by spe¬
cific and durable structural features of their econ¬
omies. A macroeconomic theory that does not
incorporate these features into its models is of lit¬
tle use. Yet only recently have serious attempts
been made to formalize and test structuralist hy¬
potheses about development.
This path-breaking book by one of the country’s

leading young economists presents for the first
time a comprehensive and coherent macroeco¬
nomics for developing countries. It develops sim¬
ple, useful, and tested models for different groups
of developing countries—mineral exporters,
countries with an important food sector, econo¬
mies with endemic inflation, and the like. While
placing special emphasis on problems of eco¬
nomic stabilization (and in particular on those
policies, such as devaluations and interest manip¬
ulations, that affect the balance of payments),
(continued on back flap)
□3A3
Structuralist Macroeconomics
Structuralist
Macroeconomics
Applicable Modelsfor
the Third World
Lance Taylor
Basic Books, Inc., Publishers New York

A version of chapter 10 appeared in the
Journal of International Economics (December 1981).
Library of Congress Cataloging in Publication Data
Taylor, Lance.
Structuralist macroeconomics.
Bibliography: p. 220
Includes index.
1. Underdeveloped areas—Mathematical models.
2. Macroeconomics. I. Title.
HC59.7.T372 1983 339’.0724 82-72408
ISBN 0-465-08239-4
Copyright © 1983 by Basic Books, Inc.

Printed in the United States of America
Designed by Vincent Torre
10 987654321
Contents
Acknowledgments
1. The Structuralist Perspective

1.1 Economic Structure 5
1.2 Analytical Tools 7
1.3 Readers’ Guide 10
2. Adjustment Mechanisms—The Real Side

2.1 An Industrial Economy 13
2.2 Comparative Statics 21
2.3 A Neoclassical Digression 28
2.4 Profit Squeeze and Growth Dynamics in the Medium
Run 32
3. Adjustment Mechanisms—Two-Sector Models

3.1 An Economy with an Important Food Sector 38
3.2 Comparative Statics for the Economy with an
Important Food Sector 43
3.3 Intermediate Inputs in a Mineral-Exporting
Economy 48
3.4 Balance of Trade Complications 54
VI Contents
4. Short-Run Adjustment in Practice 57

4.1 A Social Accounting Matrix for India, 1980-81 58
4.2 A Macro Model for India 63
4.3 Parameterization 71
4.4 The Jacobian 73
4.5 Numerical Results 76
4.6 Conclusions 84
5. Money and Other Assets in the Short Run 86

5.1 The Real Side of the Economy 87
5.2 Money and Other Assets 91
5.3 The Model without Gold—Short Run 95
5.4 The Macro System with Gold 98
5.5 Money in the India Model 103
6. Money, Inflation, and Growth 109

6.1 Short-Run Adjustment under Rational
Expectations 110
6.2 Wage Dynamics in the Medium Run 114
6.3 Inflation and Growth in the Medium Run (with no
Gold) 116
6.4 Gold and Other Matters 121
Appendix 123
7. Trade Balance Complications 127

7.1 Internal and External Balance 128
7.2 The Two Gaps 130
7.3 Exchange Rate and Output Adjustments 135
7.4 Interest Costs and Export Performance 140
7.5 Exports and Growth 143
8. Foreign Assets and the Balance of Payments 145

8.1 Asset Markets and the Capital Account 146
8.2 Dynamics of Exchange Adjustment 152
8.3 Extensions and Reality Checks 157
8.4 Real and Financial Programming 157
Contents Vll
9. Two-Sector Models of Inflation, Distribution, and

Growth ' 161
9.1 Inflation from Lagging Food Supply 162
9.2 Growth in an Economy with an Important Food
Sector 166
9.3 Effects of Income Redistribution in an Industrial
Economy 170
9.4 Income Distribution and Patterns of Growth 173
10. Trade Patterns and Southern Growth 111

10.1 Growth in the North 179
10.2 Growth in the South 182
10.3 How the Regions Interact in the Short Run 185
10.4 Steady State 186
10.5 Final Observations 189
11. Policy Lessons 190

11.1 Orthodox Stabilization Ploys 191
11.2 Monetary Contraction 192
11.3 Devaluation—Real Effects 193
11.4 Intersectoral Complications 195
11.5 Mongering Financial Reform 196
11.6 Crawling Pegs and External Financial
Reform 198
11.7 Income Redistribution and the Export Push 200
11.8 Summing-Up on Stabilization 201
11.9 Toward the Long Run 202
11.10 Terms of Trade 203
11.11 Income Distribution and Patterns of Growth 204
11.12 Inflation, Class Struggle, and Growth 205
11.13 Poor Countries in the World System 207
Notes 209
References 220
Index 227
*
Acknowledgments
MOST scholarly books are pastiches of other people’s works. This one is
no exception, and draws heavily on studies by people from many different
parts of the world (though not from different time periods since most of the
research on formal structuralist models has taken place during the past
three or four years). I am particularly grateful to the following people:
Edmar Bacha, Francisco Lopes, Eduardo Modiano, Andre Lara-Resende,
and Persio Arida at the Pontifical Catholic University (PUC) in Rio de
Janeiro; Alejandro Foxley, Jose Pablo Arellano, and Rene Cortazar at
CIEPLAN in Santiago; Alain Ize and Nora Lustig at El Colegio de Mexico
in Mexico City; J0m Rattsp at the University of Trondheim; Hiren Sarkar
at the National Council of Applied Economic Research in New Delhi; Jack
Duloy, Sweder van Wijnbergen, Alan Gelb, Wafik Grais, Pâdeep Mitra,
and Desmond McCarthy at the World Bank; Youssef Boutros-Ghali at the
International Monetary Fund; Frank Lysy at Johns Hopkins; Jere Behrman
and Ed Buffie at the University of Pennsylvania; Graciela Chichilnisky at
Columbia; Carlos Diaz-Alejandro at Yale; Bill Gibson at the University of
Massachusetts in Amherst; Steve Marglin at Harvard; Eliana Cardoso at
Boston University; Amitava Dutt, Paul Krugman, Rudi Dombusch, Emma
Rothschild, and Dick Eckaus at MIT. Cohorts of students at PUC, MIT,
El Colegio de Mexico, and in seminars at various other places suffered
through successive elaborations of the models here; they paid back their
tormentor for his “minor slips” in turn. My wife, children, and pets pro¬
vided the appropriate quantum of distractions from the algebra. Linda
Dorfman did a super job of typing. My thanks to them all.
Structuralist Macroeconomics
1
The Structuralist
Perspective
A
m JL N ECONOMY has structure if its institutions and the behavior of
its members make some patterns of resource allocation and evolution sub¬
stantially more likely than others. Economic analysis is structuralist when
it takes these factors as the foundation stones for its theories.
According to this definition, North Atlantic or neoclassical professionals
do not practice structuralist economics. Their standard approach to theory
is to postulate a set of interlocking maximization problems by a number of
“agents” and ask about the characteristics of the solutions. Institutions are
conspicuously lacking in this calculus, as is recognition that men, women,
and children are political and social as well as economic animals. Noneco¬
nomic, or even nonmaximized, forces affecting economic actions are ruled
out of discussion. Moreover, allowable economic actions are curiously cir¬
cumscribed. For example, markets are almost always postulated to be price¬
clearing when it is patently obvious that many functioning markets are
cleared by quantity adjustments or queues.
In applied economics, this theoretical cabala breaks down, as it must. But
the impact of the theory on the practitioner remains strong. If markets are
routinely supposed to clear by price, then getting the prices right or, more
grandly, assuring a Pareto efficient allocation, becomes an obsession. If
4 Structuralist Macroeconomics
neoclassical growth theory teaches that increased saving automatically

transubstantiates itself into higher investment and faster growth, then poli¬
cies that stimulate saving become ends unto themselves.
Until recently, the narrowness of these ways was more apparent in the
Third World than the First. The neoclassical synthesis held sway in ad¬
vanced economies from the 1950s until the mid-1970s, in part because they
behaved neoclassically. Such fortune never befell the poor countries of this
world, and their economists always talked about structure. So did econo¬
mists from the richer countries until after the Second World War—John
Maynard Keynes is a classic example. Because their epistemology was
founded on physics, mathematics, and derivation of economic theories from
maximizing “first principles,” the framers of the neoclassical synthesis
devoted scant thought to issues of structure, though some could do it very
well. They passed their bad habit along to their intellectual daughters and
sons.
The analysis in this book is an attempt to rectify this omission in the
context of developing countries. It amounts to a combination of abstract
descriptions of the structures that Third World scholars have always em¬
phasized with analytical models largely developed by Keynes, Michal Ka-
lecki, and their followers of the Cambridge School. Like most blends of two
lines of thought, the product will partly displease both sets of progenitors.
But—as is clear from the long list of names in the Acknowledgment at the
front of the book—it strikes an increasing number of people as a useful way
to think about macroeconomics for the majority of countries in the world.
The remainder of this chapter is devoted to three tasks. First, in section
1.1 there is a recital of the structural characteristics of developing countries
that are emphasized in the analytical models that follow. No one country
has all or even many of the characteristics, but all poor countries have some
of them. There is no such thing as a single model of underdevelopment, but
certain aspects reappear in different combinations. Not all existing combi¬
nations are discussed in this book, but enough models are developed to show
how they can be.
In section 1.2 there is a discussion of the main analytical tools applied,
the tricks in the models. There is a strong tendency in orthodox economics
to make a fetish of technique. In part this book responds to fashion—it
purports to show that structuralist thought can be made rigorous. This
attempt risks banality. True insight into how the economy functions easily
degenerates into boring algebra. Our goal in section 1.2 is to give a preview
of how the algebra goes about its tasks.
Finally, in section 1.3 there is a compilation of the topics covered in
The Structuralist Perspective 5
chapters 2 through 10. Most of what one learns from the analytical models
that is of practical or policy use is summarized in chapter 11. The reader
mainly interested in inquiring if formalized structuralist macroeconomics
is of any earthly use is encouraged to jump directly there.
1.1 Economic Structure
The approach to models taken here begins by asking what variables

adjust to assure that overall macro balance is reached—that saving is equal
to the value of investment, in the usual shorthand. Three adjustment mech¬
anisms naturally present themselves:
1. Output meets demand, made up of autonomous elements like investment and

output-sensitive components like private consumption.
2. Supply is fixed, so demand must adjust to it. One means is through price changes
that limit consumption (in the simplest model) to total output minus investment.
3. Some component of demand varies freely to bring overall balance—competitive
imports or government spending are two possibilities.
Consider an economy in which commodity production requires inputs of

labor, imported intermediate goods, and a stock of durable means of pro¬
duction called “capital.” One aspect of structure emphasized here is that
the wage is determined institutionally, in effect by class conflict in ways
described more fully below. A second key point is the economy’s critical
need for imported inputs. Since the energy crisis, these have entered North
Atlantic theory also, but their importance never eluded economists in the
Third World.
In this production setup, which of the three adjustment mechanisms just
mentioned can be assumed to apply? The answer is that “it depends.” At
times and for some purposes the macroeconomy can be assumed to be at
“full employment” in the sense that capacity is fully utilized and output
fixed; at other times and for other purposes the opposite assumption makes
sense. Moreover, some sectors may have fixed outputs, and adjusting prices,
whereas others have supply responding to demand. Different structural
models follow from different sets of assumptions, as pointed out repeatedly
in the following chapters.
Subject to these sorts of hypotheses about commodity market clearing,
changes in the functional income distribution are a fundamental part of the
economy’s approach to macro balance. Classes are distinguished by sources
of income and saving and consumption behavior. Some gain and others lose
when the macro situation changes; the losers may try to regain their eroded
position by forcing further changes in macro variables over time. One
example is workers who find their real incomes decline when prices rise to
bring macro balance under conditions of fixed output in some sectors. They
press for higher wages, and in so doing find that their claims for real income
conflict with those of capitalists, the government, foreigners, or other
groups in the system. From this process of conflicting claims, inflation,
perhaps linked with more severe social disruption, is an inevitable result.
Most structuralist theories of inflation (certainly the ones in this book)
proceed along these lines. They can be made subtle when it is recognized
that the relative bargaining strength of capitalists and workers changes with
the situation of the economy—workers may be in a position to obtain higher
real wage gains when employment is high or inflation accelerates. Such
assumptions underlie the profit-squeeze theories of cyclical adjustment that
appear in the medium-run models presented below.
Foreign trade enters this interplay in several ways. The open economy
must adjust so that two balances are satisfied:
1. The trade deficit (foreign saving) must equal investment minus national saving.
2. The trade deficit must also equal the sum of imports (capital, intermediate, and
consumer) and net interest payments on debt, minus exports and other net
current foreign exchange receipts such as emigrant remittances.
Different models can be constructed around the two variables that adjust
to assure these relationships hold. The growth rate, capacity utilization, and
the trade deficit are possible accommodating variables among macro quan¬
tities, while the exchange rate and the price level (implicitly, the income
distribution) play the same role in the price system. Some pairs of adjusting
variables are more plausible candidates in the context of development eco¬
nomics than others. The choice defines the structure one chooses to impose
for analyzing a particular policy program. Selection requires a judgment
from outside the model about the key forces that make the economy oper¬
ate. Returning to a point made earlier, postulating maximizing behavior on
the part of all agents suppresses the need to judge.
Money and financial variables affect the real side of the economy along
with trade. Investment responds to the interest rate as well as to the profit
rate (or the income distribution) and capacity utilization. But interest pay¬
ments also enter firms’ prime costs if they are required to pay in advance
for labor and intermediate inputs. Facing poorly articulated transport and
commercialization networks and financial systems with a low degree of
intermediation, firms in poor countries may find these working-capital ex¬

penses to be a large fraction of their cash flow. One important implication
is that by driving up interest rates, contractionary monetary policy is likely
to push up prices and reduce activity levels from the side of costs.
Interest rate cost-push is left out of standard models, which inevitably
show monetary contraction to be anti-inflationary. Similarly, the fact that
exchange-rate devaluation raises costs through intermediate imports does
not enter the usual formulations. The implications are that devaluation does
not raise prices and stimulates output, when in fact the opposites are true,
in the short run, in many developing countries. Models have to include these
linkages if they are to be at all relevant to policy formulation in the Third
World.
To summarize, structuralist models incorporate important technical and
behavioral relationships of the type mentioned above. They focus attention
on how the income distribution and output levels vary to satisfy the major
macro balance equations in the short run. Longer-run developments depend
on the reactions of gainers and losers to these initial adjustments. Finally,
judgment is required as to the key forces shaping economic change and the
directions they move. An economic world substantially richer than that of
orthodox theory is constructed in an attempt to capture the diversity of
underdevelopment.
1.2 Analytical Tools
Perhaps surprisingly, in light of the way economic analysis has developed

since the 1930s, simple mathematics more than suffices to address the issues
considered here. The reason is that separate models are constructed to
describe specific economic structures, with no attempt at generalization (in
the mathematicians’ sense) for its own sake. Eschewing maximization,
many agents, general nonlinear or topological descriptions of choice and
production sets, and all the other paraphernalia of modem economics
means that one can concentrate on essentials—specifically, how do eco¬
nomic groups respond broadly to price and output changes at the macro
level?
This question already focuses attention on the two main macro balance
equations mentioned above: investment equals foreign plus national saving,
and foreign saving equals imports plus net interest, minus exports and net
remittances. These two balances hold “identically” in the national accounts
data upon which all applied models must be based, but they also describe
situations of virtual disequilibrium. Suppose investment momentarily ex¬

ceeds saving. What can adjust to restore macro balance? Will the adjust¬
ment be stable in the sense that once it has occurred, equilibrium will
persist? Most analysis in this book is set up in terms of these questions, and
for that reason standard local stability analysis of the differential equations
describing adjustment toward equilibrium is employed. But the manipula¬
tions around the Jacobian matrixes for stability should not shroud the
simplicity of the issues involved. Macro equilibrium in which aggregate
prices, quantities, and the income distribution all satisfy the “identities” can
come about in various ways. We want to explore these paths.
In terms of details, a few formulations are repeatedly employed. They
include the following:
1. The price level is often assumed to be determined by a fixed producers’ markup

over prime cost, including labor, import, and working-capital finance costs. The
markup specification broadly fits the data and is far simpler to manipulate than
the obvious alternative of neoclassical cost functions.
2. Saving propensities are assumed to differ by class—higher for profit than wage
recipients. Once the difference is recognized, one retains little besides detail if
the workers’ saving propensity is assumed to be positive. Hence, it is set to zero.
3. Even short-run models are often set up in terms of variables normalized by
division by the value of capital stock. This trick throws emphasis on growth and
profit rates as opposed to levels of investment and payments to factors; one can
argue that the former type of variable provides a more intuitive feel for how the
system works. Also, setting up initial models in normalized form allows an easy
translation to analysis of growth later in the book.
4. In several cases overall macro balance is decomposed sectorally— the stability
stories are set up around mechanisms through which excess demand in each
sector adjusts toward toward zero in the very short run. As mentioned above,
one sector may be assumed to have an adjusting price and the other an adjusting
output level. This simple “flexprice-fixprice” model1 describes a surprising
range of economic structures and is employed repeatedly.
5. The two-sector formulation calls attention to differences in consumption pat¬
terns, in particular to the fact that some industries face demands that are
inelastic with respect to income and price. More or less complicated complete
systems of consumer demand equations can be used to formalize this manifesta¬
tion of Engel’s Law. Very simple equations using constant marginal propensities
to consume and one or two intercept terms are used here.
6. Financial variables are incorporated on lines set out by Keynes and elaborated
by James Tobin. Specifically, a set of assets is postulated, and accommodating
variables adjust until rentiers desire to hold the assets in the quantities and at
the prices that exist. In the simplest case, for example, firms borrow at a variable
interest rate to finance working capital outlays. Rentiers can either lend to firms
or put deposits in banks. The interest rate changes until the excess demand for
loans (or, given the balance sheets of the banks, the excess supply of deposits)
is equal to zero. Differential equations are deployed to formalize conditions

under which this adjustment process stably converges. (Sometimes it does not
—a good reason to worry about working-capital cost-push.)
7. Investment demand is determined by firms and plays a key role in both short-
run adjustment and longer term growth. The growth rate of capital stock
(investment divided by the existing stock, ignoring depreciation) is assumed to
depend positively on the profit rate (firms’ cash flow divided by the value of
capital) or the rate of capacity utilization. This specification reflects Cambridge
theory. When the interest rate is brought into the models, capital stock growth
depends on the difference between profit and interest rates. This is very Keyne¬
sian or Tobinesque. In two-sector models, investment-demand functions may of
course differ between industries, with corresponding influence on long-run eco¬
nomic structure.
8. Imports are split several ways. There are, at times, three types of “noncompeti¬
tive” imports.2 Capital-goods purchases abroad are proportional to total invest¬
ment demand. Intermediate imports depend on levels of output. Consumption
imports, when considered, come from demand functions of the type discussed
above. Finally, “competitive” imports are sometimes entered as an adjusting
variable to close the gap between demand and supply of a commodity produced
within the economy. Under this last set of circumstances, the “Law of One
Price,” which states that internal prices are determined from the world market,
may or may not be assumed to hold.
9. Exports may be set exogenously, or assumed to be determined by the difference
between the internal and world price of the commodity exported (in a simple
linear equation), or vary to equilibrate commodity demand and supply in the
same fashion as competitive imports.
10. Long-run issues are investigated by setting up transitions between steady states
in which all variables grow at constant rates. For example, in one-sector models
with money, the steady-state conditions are two. First, the rate of money expan¬
sion should equal the rate of output (or capital stock) growth plus the rate of
inflation. Second, the real wage should be constant, or wage inflation should
equal price inflation. These conditions are equivalent to stating that the ratios
Money Stock/Value of Capital and Price/Wage stay constant. Differential equa¬
tions are set up in terms of these “state variables,” and their stability properties
are analyzed. Of course, the equations incorporate hypotheses regarding bar¬
gaining or class conflict over wage-adjustment rules, the role of the interest rate
in shaping working-capital cost-push and investment demand, and similar mat¬
ters. In two-sector models, as another example, the relevant state variable is the
ratio of capital stocks in the two industries. To hold the ratio constant, both
sectors must grow at the same rate in the steady state.
11. Finally, directions of causality among variables differ from model to model—
that is what the analysis is all about. A reader carrying this thought at the back
of his mind will not be confused by the transitions.
1.3 Readers’ Guide
Topics covered in the chapters unwind more or less in the order just set
out. Chapter 2 takes up quantity, price, and competitive import adjustment
toward macro balance in a one-sector model. Comparative static exercises
center around currency devaluation (which can prove “stagflationary,”
with price increases and output contraction at the same time) and income
redistribution. A digression shows that neoclassical formulations give essen¬
tially the same results as in the simpler models used here, except when
macro causality is set up to let saving determine investment. Then the
substitutability assumptions at the base of neoclassical models have an
important role to play. Finally, a profit-squeeze business-cycle theory is
proposed as a first extension of the analysis toward the medium run.
Chapter 3 is devoted to a pair of fixprice/flexprice models, with price¬
clearing agricultural and infrastructure sectors in order. Industry is the
quantity-clearing sector in both models. The effects of big investment
pushes, food consumption subsidies, and other matters are discussed in the
first agriculture/industry system. The second model focuses on the absorp¬
tion problems of a mineral exporting economy. Both theoretical specifica¬
tions underlie chapter 4, where an empirical computable general equilib¬
rium model for India (with two price-clearing and three quantity-clearing
sectors) is used to illustrate the vagaries of short-run adjustment in practice.
Financial markets are introduced in chapter 5. The public can hold three
assets: money in the form of bank deposits; loans to firms to finance working
capital; and a collection of nonproductive assets called “gold.” A reduction
in money supply increases the interest rate on loans to firms and thus the
cost of working capital. A sharp enough increase can drive up the price level
from the side of cost-push—monetary contraction can be inflationary in the
short run. Similarly, attempts to draw deposits toward the banking system
can lead to a contraction in rentiers’ desired volume of productive loans to
firms instead of holdings of gold. The outcome is once again stagflation.
Orthodox financial policy can easily prove counterproductive. This lesson
is extended to the long run in chapter 6, where it is shown across different
steady states that slower money growth can lead to faster inflation, a lower
growth rate of output, and a more unequal income distribution. The skele¬
ton key to these unusual results is again interest rate cost-push resulting
from restrictive monetary policy.
Chapter 7 addresses open economy problems of stabilization and growth,
using the two balance equations described above. These are first interpreted
in terms of the development economists’ familiar two-gap model (due fun-
damentally to Hollis Chenery) and then put through adjustment paces. The
chapter also discusses the roles of foreign interest payments and an export
push in affecting internal and external balance.
One of the specifications in chapter 7 is short run, with the trade deficit
and the level of economic activity responding to investment demand (or the
capital stock growth rate), the exchange rate, and other variables. This
description of the real side of the economy is combined with financial
markets in chapter 8 to inquire about the effects of revising the rate at which
the nominal exchange rate is depreciated in a crawling peg. Recent or¬
thodoxy recommends that the rate of crawl be slowed as an anti-inflationary
device. This policy is shown to have strongly destabilizing tendencies in the
long run. The chapter closes with a review of how the gap equations in
chapter 7 plus the monetary system’s accounting identity (money supply
= bank credit to the private sector -f bank credit to the government +
foreign reserves) are applied in practical economic work by agencies such
as the World Bank and the International Monetary Fund.
Chapter 9 takes up longer-term patterns of growth and income distribu¬
tion in two two-sector models. The first extends the agriculture/industry
model of chapter 3 to discuss food-price inflation and its real effects in
steady state. The second model shows how policies aimed at income redis¬
tribution may accelerate growth in an economy with “wage” and “luxury”
good industries.
Chapter 10 turns to international issues in a long-term context, where the
“South” is assumed to export a price-clearing commodity for which the
income elasticity of demand in the “North” is less than one. The North
exports investment goods and luxuries to the South. It is shown that growth
in the South is largely dependent on what happens in the North in these
circumstances, and that a productivity increase in its export industry will
generate adverse shifts in its terms of trade, available saving, and rate of
growth. This last finding is a classic structuralist result.
Finally, as mentioned above, chapter 11 summarizes policy conclusions.
Adjustment
Mechanisms—
The Real Side
M
AVA ACROECONOMICS begins with the notion that the value of
saving generated by all participants in the economy must by one means or
another come into equality with the value of investment in the short run.
The investment decisions are typically made by public authorities, the
managers of firms, and the families that choose to construct new housing.
In a capitalist system, saving comes from financial institutions (insurance
companies, pension funds, and so forth), retained earnings of firms (over
which managers exercise partial control), the public sector’s budget surplus,
households, and foreign institutions that lend money to accumulate “our”
country’s financial obligations or cover its deficit on foreign trade. These
accounting truisms are repeated here to emphasize one point: the groups
making saving and investment decisions only partly overlap; hence, there
must be mechanisms that affect their behavior to bring their totals of saving
and investment into accord. In any functioning economy, of course, adjust¬
ment occurs in many ways at the same time. The art of macroeconomic
modeling is to pick out the dominant changes for more careful, at times
Adjustment Mechanisms—The Real Side 13
mathematical, scrutiny. By the same token, fundamental debates in macro¬

economics almost always are about which mechanisms are central to the
adjustment process.
Although economic history and anthropology show that other modes are
possible, saving and investment shifts in modern economies are mediated
through money or its close substitutes. The implication is that banks and
the financial system play central macroeconomic roles. However, the banks’
actions are often (though not always) subordinate—they are overwhelmed
by more aggressive actors, like Bertram in All's Well that Ends Well. For
that reason, this and the next two chapters deal with adjustments in prices
and flows of output that stem from the economy’s nonmonetary or real side.
The focus here is on countries at middle and low levels of gross domestic
product per head. The institutional and class alignments assumed to charac¬
terize these economies will emerge during the course of the discussion. With
some modification, the models set out here also apply to northern, industri¬
alized nations; indeed, many of the basic ideas are due to progressive
economists from that corner of the world. But for the most part the analysis
here is directed toward the southern hemisphere.
We begin the discussion in section 2.1 with a simple one-sector model in
which three types of short-run macro adjustments are possible—changes in
capacity utilization, shifts in the income distribution through “forced sav¬
ing,” and movements in the level of available saving stemming from changes
in the foreign trade deficit. Contrasts between these adjustment mechanisms
are illustrated by comparative static exercises in section 2.2. The formula¬
tion of the model is not neoclassical in the sense that it works with fixed
production coefficients and markup pricing. However, the results in prac¬
tice differ little from more orthodox formulations. This is the topic of
section 2.3. How the economy functions under full employment of both
labor and capital or capacity is also discussed there. Finally, an initial
analysis of growth and distributional issues in the medium run appears in
section 2.4.
2.1 An Industrial Economy
Through manufacturing and related activities account for far less than
one-half of total value added and an even smaller share of employment in
most developing countries, they play a central macroeconomic role. The
industrial sector often grows faster than the rest of the system, and is more
responsive to government policy tools. Moreover, it is often characterized

by excess capacity, and for that reason its macro adjustments take place on
the quantity side. By contrast, manufacturing prices are likely to be fixed
in the “short run” (one or a few quarters) by relatively stable markups over
variable cost. The existence of markups can be rationalized in various ways,
but presence of spare capacity and oligopolistic structures among firms
(buttressed in most developing countries by ample tariff protection from
foreign competition) is probably the best single explanation.1 For present
purposes, we simply postulate a markup rate for such firms (fixed or vary¬
ing, according to whether capacity is fully utilized) and work out the
implications of this pricing rule.
On the cost side, a firm’s variable inputs will typically comprise labor,
intermediate goods in process, and raw materials. For the economy as a
whole, standard input-output calculations show that these input flows can
be reduced to labor services and imported intermediates. One firm’s sale is
often another’s intermediate purchase, and such transactions can be netted
out in macro accounting. To keep the algebra as simple as possible, assume
that this consolidation has already taken place.2
The costs per unit of labor and intermediates will be taken as fixed in
nominal terms in the short run. For imports, this assumption amounts to
saying that they can be obtained on the world market at a fixed price in
terms of foreign exchange, and that the nominal exchange rate is stable. For
labor, a fixed money wage is assumed—workers are hired up to the levels
of firms’ demand at this wage. Between periods (or “moments” of time) the
money wage will adjust according to past inflation rates, labor market
conditions, and the relative bargaining strengths of workers and capitalists.
The details of what happens when the markup, as opposed to the wage rate,
shifts are given in section 2.4.
With fixed nominal input costs and markup pricing in force, the pricing
rule for the industrial sector will take the form:
P = (1 + t)(w6 + eP$a0) (2.1)
where P is the sector’s output (or producer) price; r is the markup rate,
w is the nominal wage rate; b is a labor-output coefficient (the inverse of
average labor productivity); e is the exchange rate or price of the local
currency in terms of foreign currency (for example, ten “locals” trade for
one dollar, and e — 10); PJ? is the foreign currency price of imported
intermediates; and a0 is the intermediate import-output coefficient. Another
decomposition of price, useful later on, is:
P = [wb + T(wb + eP$a0)] + eP%a 0 (2.2)
Here, the term in brackets is value-added per unit output, in turn separated
into wage and nonwage components wb and r(wb + eP$a0) respectively.
Value-added plus intermediate imports per unit output (ePJa0) sum to the
price level P.
Assume that the value of physical capital (plant and equipment) owned
by industrial firms is PK. A rigorous justification would be that there is just
one type of industrial product that can be used as an intermediate, a final
consumption good, and fixed capital: the amount K serving as capital is
valued at the current producer price P. Nobody seriously believes such
stories but many economists still find it convenient to work with profit rates
of the form:
PX — wbX — eP$aoX
PK
In this ratio, the numerator is total industrial sales PX (when the output
flow is X), less wage and intermediate import costs. Dividing this measure
of total profits by the value of capital stock gives the profit rate r. After
substitution from (2.1), r can be rewritten as
t X T
r —-= -u (2.3)
1 + t K 1 + r
where u is the output-capital ratio. So long as there is excess capacity in

the industrial sector, u is free to vary and can serve as a measure of capacity
utilization. Equation (2.3) shows that the profit rate is an increasing func¬
tion of the output-capital ratio. For constant u, r also increases with the
markup rate r, as one might expect.
Now assume for the moment that the industrial sector is the whole
economy (or that the rest of the economy behaves just like industry). The
purpose of this drastic simplification is to set out the major macro adjust¬
ment mechanisms in simplest form. To that end, note that excess demand
(demand less supply) for output will equal zero when the condition
C + I + E - X = 0 (2.4)
is satisfied. As usual, C in (2.4) stands for household consumption, / for

investment (stock changes plus capital formation), and E for net exports
(total exports less imports of commodities similar to or “competitive” with

domestic products). Note that government purchases of goods and services
are not included in the accounting—leaving them out amounts to ignoring
fiscal policy in favor of an emphasis on private sector actions in this chapter.
The next step is to reduce (2.4) to a statement of saving-investment
equilibrium for the economy. To do so, let the value of total consumption
be given by:
PC = wbX + (1 - sr)rPK (2.5)
That is, wage earners consume all their income, while a share, s„ of profit
income is not consumed, or saved. This formulation leaves out whatever
saving workers may undertake; however, much of their accumulation ulti¬
mately goes to finance housing and purchase of consumer durables and can
be left out of a model focused on industrial investment demand. Saving from
profits will be undertaken both by firms (retained earnings) and those
presumably well-off households that receive significant amounts of property
income. Depending on accounting conventions, in many countries surpluses
of public enterprises would also contribute to the profit-based saving
flow.3
Substitution of (2.5) into (2.4) and manipulation based on the cost decom¬
position (2.2) give the equation:
PI - (ePZaoX - PE) - srrPK = 0
This expression can be interpreted as an investment-saving balance. When

investment PI less the trade deficit (eP*a0X — PE) less saving from profits
srrPK is equal to zero, then the economy is in macro equilibrium. An
equivalent statement from (2.4) is that excess demand for commodities is
zero. In this chapter, excess demands for both labor and financial assets are
left out of the picture; hence, (2.4) or the investment-saving balance is
equivalent to Walras’s law.
In the rest of this chapter, a notation that expresses the investment-saving
balance in growth-rate terms will be used to pave the way for analysis of
growth models later in the book. Define three new symbols as <p — eP$a0/
(wb + eP%a 0), or the share of intermediate imports in variable cost; g =
I/K, or the growth rate of capital stock; and € = E/K, or the ratio of
exports to capital stock. Then dividing the above investment-saving balance
by PK allows it to be rewritten as:
g + f — (T-,<f> + sr)r = 0 (2.6)

In words, saving per unit of capital in the economy is the sum of saving from
profits srr and the nonconsumed flow of income resulting from intermediate
imports (f>r/r. Total saving is exhausted by capital stock growth g and net
exports of home-produced goods as a fraction of capital stock, e.
To close out a description of the macroeconomy, we have to write down
demand functions for investment and net exports. Export demand e can
most usefully be held constant for the moment. Investment demand can be
expressed as:
6 — z0 I Z |' I ~
, 1 + T
— Zq + Z1 -(- -Z 2 (2.7)
The theory underlying the equation in the first line is that there is some base
level z0 of public and private investment that is unaffected by current
economic conditions—in the traditional phrase, z0 reflects the “animal
spirits” of private corporate managers and public enterprise functionaries.
Investment will rise with both the profit rate r and the index of capacity
utilization u as indicators of future profitability. The variable u resembles
an accelerator term in investment demand functions; econometric evidence
suggests that the accelerator coefficient z2 is likely to exceed the profit rate
coefficient z, in (2.7). The second line in this formula uses (2.3) to reduce
investment behavior to a function of the profit rate and the markup r.
Equation (2.6) is drawn as the “Saving supply” curve in figure 2.1, under
the assumption of an exogenously fixed level of net exports. Even if invest¬
ment demand g is zero, domestic saving is required to finance the corre¬
sponding trade surplus. The intercept of the saving supply curve on the r-
axis represents the profit rate that will generate the needed saving flow.
The “investment demand” schedules in the diagram correspond to (2.7),
with two different levels of the animal spirits parameter z0. In the middle
quadrant, capacity utilization and the profit rate are related positively
according to (2.3) until a capacity limit u is reached. Thereafter, capacity
use stays constant while profit rate increases are accompanied in the bottom
quadrant by a falling real wage to or w/P:
= w_= » = * - 4> (2 8)
P (1 + T)(wb + eP%a q) (1 + r)b
The markup rate is constant so long as capacity utilization is below its

maximum value u. The real wage takes on corresponding value to. When
Macro Equilibrium with Capacity Adjustment (Point I) and Forced Saving (Point
II)
the economy reaches full capacity, further increases in the profit rate can
only come from a higher markup rate and a falling real wage. In formal
terms, transition to full capacity imposes an additional restriction on the
macro system. It can only be met by an additional dependent variable, and
the obvious candidate is the markup rate. As described in detail below, a
rise in r drives up the price level P, which in turn limits demand to the
available capacity level UK.
Now consider the investment-saving equilibrium points in figure 2.1. At
point I, the growth rate is g(, the profit rate is r,, and there is excess
capacity since u i lies below u. Any small increase in investment (or export)
demand could be met by increased capacity use at a constant real wage co.
Macro adjustments in the vicinity of equilibrium I follow the classical

Keynesian rules.4 From (2.6) and (2.7) the mechanism can be described in
formal terms as an adjustment in the profit rate of the form:
^7
at
=/[£ + €- (t ‘<f> + O'*]
— f\z 0 + € + (2.9)
(Zl + iT^) “ (7 + *')
where/(0) = 0 and the first derivative of /is positive. The profit rate (and
capacity utilization) will be constant when commodity excess demand is
zero and saving equals investment, but will rise when excess demand is
positive. The adjustment is stable if the derivative of / with respect to r is
negative, so that a rising r will drive excess demand back toward zero. It
is easy to verify that this will be the case when the slope of the investment
function is less than that of the saving function, as shown in figure 2.1.
At point II, the saving-investment equilibrium occurs at full capacity u.
As argued above, the markup rate and real wage are endogenous in this
situation. Any increase in animal spirits or exogenous investment z0 will
create excess demand for output that can only be reduced by rising price
level P and markup rate r. From (2.8), there will be a falling real wage a).
Since the propensity to consume from wage income exceeds that from
profits, total consumption will be reduced and excess demand driven back
to zero. This is a forced saving adjustment, of the type emphasized by
Kalecki and Kaldor.5 Note that its efficacy requires a constant money wage
w in (2.8), or at least wage increases that lag prices. Forced saving is
impossible when (1) saving propensities of workers and profit recipients (or
of all groups) are the same, or (2) wages (or all nominal payment flows) are
instantaneously and fully indexed to price increases. In an economy with
a plausible set of medium-term contract obligations, behavioral parameters
and apparatus for class repression, these conditions will not be satisfied and
forced saving is a viable adjustment mechanism.6
To give a formal version of the adjustment, note from (2.3) that at full
capacity the profit and markup rates are linked by the equation (1 -f t)/t
= u/r. Substitution into (2.9) then gives the profit rate response to excess
demand as:
— = f[z0 + € + (z2 - (f>)u + (z, + <f> - sr)r]

dt
For a plausibly high saving rate, df/dr < 0, and the forced saving adjust¬
ment will be stable.
So far, the variable e, representing net exports as a share of capital stock,
has been held constant. Moreover, little would be changed if net exports
adjusted according to a rule such as e = e(eP%/P), where P* is the foreign
currency price of “our” exports. This equation simply states that exports
are an increasing function of the ratio between the price at which they can
be sold and the domestic producer price. Depending on the circumstances,
export price responsiveness may or may not be macroeconomically impor¬
tant, but it does not modify the adjustment mechanism underlying figure
2.1 in any essential way.
Adjustment can be affected, however, if net exports are treated as com¬
pletely endogenous. Assume that the economy is at full capacity and, more¬
over, that workers can muster sufficient power to stabilize the real wage at
level ci>2 in figure 2.1. At full capacity, the profit rate will then be fixed at
r 2 and saving supply (per unit of capital) at g2. How could a further increase
in investment demand be accommodated? Figure 2.2 illustrates one means
—an upward shift in the saving function mediated by falling net exports.
FIGURE 2.2
Macro Adjustment by Reduced Net Exports at Full Capacity Utilization and a
Fixed Real Wage
As € declines for a given profit rate r, (2.6) shows that more saving is made
available to finance investment demand. The saving supply curve shifts up
while the country’s foreign trade surplus declines (or its deficit increases).
“Foreign saving” provides the means for the economy to move its equilib¬
rium from II to III while the real wage stays fixed.
Evidently, other saving flows could be made endogenous to satisfy a
similar function. Changes in the public sector surplus by continuously
adjusting tax or subsidy rates are examples. As we will see in subsequent
chapters, such adjustments almost always have monetary repercussions that
may not be easy to control. But along with variations in capacity use and
forced saving, endogenous shifts in the levels of foreign or public sector
saving are an important adjustment device.
2.2 Comparative Statics
The best way to illustrate the economic significance of these mechanisms

is through exercises in comparative statics. Two main variants are presented
here—a reduction in the markup rate r and currency devaluation or an
increase in the exchange rate e.
Cutting the markup rate amounts to an income redistribution from profit
recipients to wage earners. In a relatively closed economy, lower markups
could be achieved by profit taxes coupled with income transfers or with
price controls or more forceful state intervention in the affairs of industrial
firms. In an open economy, reduction of tariff barriers and other forms of
trade liberalization could do the trick. If these moves are made when there
is excess capacity, there will be three major effects. First, there will be an
increase in aggregate demand as income flows are redirected toward rela¬
tively low-saving workers. At a given level of capacity utilization u, the real
wage will rise and the profit rate will fall when r is reduced, as shown
respectively by (2.8) and (2.3). The curve relating u and r will rotate
clockwise, as shown in the middle quadrant of figure 2.3. The outcome will
be higher output from increased demand at the initial profit rate rx.
The second effect is on investment demand. When the coefficient z2
exceeds zero in (2.7), investors will respond to increasing capacity use by
raising their rate of capital accumulation. The investment demand function
will rotate counterclockwise in the upper quadrant of figure 2.3, and the
result, when the saving function doesn’t shift, would be higher rates of
growth, profits, and capacity utilization.
FIGURE 2.3
Effects of a Reduction in the Markup Rate r
The third effect—a shift in the savings function—works in the opposite

direction. From (2.6) saving supply at a constant r increases when r de¬
clines; in effect, imports amount to a larger share of redirected income flows.
The resulting leftward rotation of the saving function is contractionary,
reducing aggregate demand. The joint movement of investment and saving
functions has an ambiguous impact on both r and g, as can be seen from
the following algebraic expressions for their changes:
dr r
(<t> ~ z2)
dT At7
and
~T~ = + z2(4> - sr)]

ar At2
where A = (t~'4> + 5,) — [z, + r_1(l -f r)z2] is positive from the overall
stability conditions.
These equations show that the growth rate will increase with a decline
in r so long as the saving rate sr exceeds the import cost share <f>. The
conclusion is that profit and growth rate increases are more likely conse¬
quences of a reduction in the markup rate when the economy is not highly
dependent on intermediate imports. Otherwise, the profit rate could easily
fall, as illustrated in figure 2.3. Note also that capacity utilization increases,
though this change could go in either direction as well. In the diagram, a
reduction in markups gives rise to faster growth, more capacity utilization,
and a higher real wage (bottom quadrant). At the same time, the profit rate
falls. Resistance on the part of capitalists toward such a policy would be
likely.
This model is often interpreted the other way round, with a high markup
presented as the root cause of economic stagnation—low capacity utiliza¬
tion and slow growth. This line of thought has been frequently repeated in
India, as in the following quotation from Deepak Nayyar:7
Clearly, a large proportion of the demand for industrial products originates from
a narrow segment of the population. However, manufactured goods sold to the
relatively few rich can use up only so much, and no more of the capacity in the
intermediate and capital goods sector. Only a broad-based demand for mass con¬
sumption goods can lead to a full utilization of capacity . .. but that in turn requires
incomes for the poor. Thus an unequal income distribution operating through
demand functions might well restrict the prospects of sustained industrial growth.
The simple model used here does not capture Nayyar’s intersectoral nu¬
ances, but does reflect his emphasis on lagging aggregate demand. Indeed,
with a low enough intermediate import share, even capitalists might benefit
from the sort of income distribution he implicitly favors. Such a condition
may hold in India. Also, as shown in chapter 9, effects of the type discussed
here carry over to a multisectoral macro model.
To pursue this theme further, one might ask about the consequences of
income redistribution under full capacity use. Since r is an endogenous
variable when there is full capacity, income shifts via forced reduction in
markups are not feasible. To explore other devices we can use the price
decomposition:
rP
P = wb H-b eP *a o (2.10)
u
which follows from (2.2) using the substitution rPK = r(wb + eP*a0)X
for total profits. Equation (2.10) can in turn be solved as:
wb + ePâ o
r = u
P
where, under full capacity, u will take the value u.

The trade-offs in this equation are between real wage w/P (or w) and
profit rate r. Suppose that it is possible to raise w/P, for example, by money
wage increases under strict price controls. Then r will fall, as in response
to an increase in w above the value g>2 in figure 2.2. As r declines from
r2, the first effect would be macro disequilibrium. The excess of investment
over saving (excess commodity demand) might be met by some endogenous
saving source—for example, the balance of payments deficit or fiscal sur¬
plus. After the resulting upward shift in the saving supply schedule, the final
effects would be slower growth, a higher real wage, and some dampening
of aggregate demand from the fiscal or trade balance change. If demand
were not damped, prices would tend to rise and the result would be reestab¬
lishment of macro equilibrium at point II after the wage increases had fully
been passed along into the price level. Unless accompanied by policies to
contract demand, attempts at income redistribution under full capacity can
easily set off an outburst of inflation, resulting in no redistribution at all.
The trick when attempting to shift income toward low-saving classes is to
recognize when capacity bottlenecks are reached and the appropriable sur¬
plus from excess capacity has run out. Although other circumstances were
adverse, the experience of historically slow-growing Chile in the early 1970s
suggests that the feasible once-for-all increase in GDP might be about 8 to
10 percent.8
Income redistributions also play a substantial role in devaluation, the
other comparative static shift explored here. An increase in the exchange
rate e has several macro effects. It will initially lead to a rise in <f>, the share
of imported intermediates in variable cost. From (2.8) the real wage will fall
when there is excess capacity and the markup rate is fixed. From (2.6) there
will be more saving generated for a given profit rate r. As shown in figure
2.4 the saving supply schedule will rotate to the left. Finally, net exports
may respond to a higher exchange rate, raising the initial surplus on trade
of competitive goods, and shifting the intercept of the saving schedule to
the right.
FIGURE 2.4
Effects of a Devaluation (an Increase in the Exchange Rate e)
The macroeconomic result of these changes is ambiguous. Figure 2.4

shows a case in which export response is relatively weak and the share of
imported intermediates in variable cost is high. The leftward rotation of the
saving schedule outweighs its rightward shift, leading to slower growth and
reduced capacity use. Or, to put it another way, the initial increase in the
local currency trade deficit from higher import costs exceeds its reduction
from more exports, leading to increased foreign saving and lower aggregate
demand. Put either way, the resulting output contraction is also accom¬
panied by a price increase from (2.1) and a lower real wage. As illustrated
on the “Trade deficit” axis, the devaluation does lead to balance of pay¬
ments improvement, at least in terms of foreign prices. However, the deficit
in domestic prices (the foreign price deficit multiplied by the higher ex¬
change rate e) may get larger. As discussed in chapter 8, the higher deficit
can lead to contraction in the money supply, which will strengthen the
effects being discussed here.
Empirical evidence suggests that the output contraction and adverse
income redistribution from devaluation illustrated in figure 2.4 may be
fairly common; however, much depends on the export response. To get a
formal expression, let 17 be the elasticity of exports with respect to their
relative profitability, viz:
dE (eP\/P)
d(eP*E/P) E
Using this elasticity, one can show that, leaving aside changes in invest¬
ment (z, = z2 = 0), the response of output to devaluation is given by:
al (1 + t)(1 — 4>) ( Tj E _ <f> |

de X srT -I- <f> (eP*E/P) X 1 + rj
From this expression it is easy to see that output expansion requires the
export elasticity to satisfy the condition:
eP*E (eP*aoX/PX)
17 >- (2.11)
P E/X
In an application, the price ratio eP*/P should be fairly close to one

(depending on the extent of export taxes or subsidies). In chapter 7 it is
argued that exports will often exceed intermediate imports. Most develop¬
ing countries import substantial quantities of capital goods as well as inter¬
mediates; indeed, the stylized fact is that financial capital inflow or the trade
deficit is less than the value of imports of physical capital goods. Under such
circumstances, exports exceed intermediate imports to generate sufficient
foreign exchange to pay for the balance of capital goods imports.
The implication from condition (2.11) is that the lower bound on the
elasticity for devaluation not to be contractionary is likely to be less than
one. After a few quarters for sales to build up, rj might be expected to take
a value around unity. Hence, devaluation may be initially contractionary,
and then lead to economic expansion after the export response takes hold.
However, riding out the contractionary phase may not be easy. The ex¬
pected tenure of finance ministers who devalue is often very short.9
The effects of devaluation at full capacity can vary, depending on the
specific assumptions about the responses of different classes of income
recipients that one chooses to make. When u is fixed at u and the markup
rate is endogenous, the saving supply and investment demand functions
(2.6) and (2.7) can respectively be written as:
V = g + {e(eP*E/P) - (eP$a0/P)u j (2.12)
and
g = (z0 + z2u) -f zxr (2.13)
Suppose that the domestic price level P is held constant when e is raised
so that the nominal devaluation is also “real.” Then the relative cost of
imported intermediates (eP*a0/P) will go up while the return to exports
(eP*E/P) will rise. The change in the bracketed term on the right side of
(2.12) can, in principle, take either sign, but a “successful” devaluation
would improve the trade balance and lead it to increase. It is easy to see
from (2.12) and (2.13) that the outcome would be a higher capital stock
growth rate g and profit rate r (so long as the basic stability condition
sr > z, is satisfied).
What about real wages? The cost decomposition (2.10) can be restated
as:
Evidently, if both r and the relative cost of intermediate imports eP*a0/P

rise, the real wage must fall. And the fall in the real wage is the mechanism
by which consumption demand can decline so that exports and investment
can rise when output is held constant by full capacity utilization. Under
both fixed and variable capacity, devaluation basically works by cutting real
wages. Once again, its political liabilities stand out strikingly.
2.3 A Neoclassical Digression
The model sketched so far has scant possibility for substitution of labor
and capital, capital and imports, and other such trade-offs that neoclassical
economists adore. The omission stems in part from a desire to simplify—
fixed coefficients and markup pricing permit easy algebra and diagrams that
emphasize directions of causation underlying macro adjustment paths. At
the same time, the basic neoclassical hypotheses of perfect competition,
facile substitution of production inputs, and all the rest lack credibility in
industrial economies, let alone the Third World. Finally, a neoclassical
formulation adds little to what has already been said, except under strong
additional hypotheses. Demonstration of this last point is the purpose of
this section.
To begin, note that the standard story about substitution in production
can be closely mimicked with the apparatus already developed. The key
neoclassical hypothesis is that the rate of profit should rise with the output-
capital ratio. This relationship is already built into equation (2.3) for a fixed
markup rate r. Making r a nonlinearly increasing function of output (there
is conflicting econometric evidence on this idea10) would not change the
fundamentals of macro adjustment as previously analyzed. However, there
would be a blending of the two first adjustment mechanisms described
above. That is, there would be forced saving from the increasing markup
even before high levels of capacity utilization were reached, whereas if the
function r = t(X) were made steep near “full capacity,” the kink in the
capacity utilization function of figure 2.1 and the abrupt switch between
Keynesian and forced saving adjustment regimes would not show up. For
applied models, smoothing out adjustment processes in this way is often
convenient—it helps prevent the numbers coming out of the computer from
taking disconcerting jumps. One strength of neoclassical formulations
comes from the fact that they have a lot of smoothness built in.
To illustrate smooth adjustment in more detail, it is simplest to work with
a model in which there are no intermediate imports, so that only labor and
capital enter as production inputs. Suppose further that there is an econo¬
mywide constant elasticity of substitution (CES) production function of the
form":
X = ) BtL(<r_,)/<r + B*A'(<r-,Vo'}a/<<r“,)
where L is the level of employment, cr the elasticity of substitution, and

B/ and B* distribution parameters in the CES function.
If producers minimize costs subject to this production technology, then

the labor-output and capital-output ratios they choose will be given by:
(2.14)
and
K
(2.15)
~X
That is, the intensity of use of each factor is a decreasing function of its real
cost to the producer.
To fit these relationships into a macro model, assume in the neoclassical
spirit that there are no differences in saving parameters between recipients
of wage and profit incomes. Then the value of consumption is given by:
PC = (1 — s)PX
where s is the economywide saving ratio. Insertion of this condition into

the commodity excess demand function (2.4) and use of (2.15) then give an
investment-saving balance of the form:
g + a - s = 0 (2.16)
This equation should be compared to (2.6). The two relationships clearly

have the same general form, though when <r is not unity (the production
function is not Cobb-Douglas), (2.16) is nonlinear in the profit rate r.
There is one more apparent difference between the neoclassical formula¬
tion and the one used before, in that the CES function permits capital-labor
substitution. Observe that the CES production function can be rewritten as:
<r/(cr— 1)
L
(2.17)
~K
This relates the labor-capital ratio uniquely to the output-capital ratio, or

u. The relationship is not a straight line, as we have been assuming so far.
Rather, it shows decreasing returns to labor in generating output as in the
southwest quadrant of figure 2.5.
Capital Stock
Growth Rateg
Saving
Supply
Investment
Demand
Employment-Capital
Ratio L/K Profit Rate r
Capacity
Utilization
Output-Capital
Ratio u
FIGURE 2.5
Macro Equilibrium When There Is a CES Production Function (cr < 1) and
Producer Cost Minimization
Adding decreasing returns to the analysis suggests that wage-cutting

ought to be an effective way to raise employment in the neoclassical model.
However, such optimism is not well founded when the price level is free to
vary. The reasoning is as follows:
1. The profit rate r is set in the macro equilibrium condition (2.1t>) when exports
and the growth rate are predetermined variables.
2. The output-capital ratio follows from (2.15) and the labor-capital ratio from
(2.17).
3. Solving (2.14) and (2.15) for the real wage w/P gives:
w L
P ~ ~K r
so that w/P is determined macroeconomically. Any attempt to reduce w

will just pull P down along with it, and the real wage and resource alloca¬
tion will be left unchanged. Unless the price level is set from outside (by
fiat as the numeraire, or else by foreign trade), a nominal wage reduction
will not reduce employment despite ample possibility for labor-capital
substitution.12
The equilibrium model is illustrated in figure 2.5. Equation (2.16)

becomes the nonlinear saving function in the northwest quadrant. Immedi¬
ately below, the output-capital ratio is an increasing function of the profit
rate, while in the southwest quadrant employment rises along with output.
The saving supply and capacity utilization curves are drawn under the
empirically based assumption that the elasticity of substitution cr lies below
unity. The implication is that successively higher levels of the profit rate are
associated with decreasing increments in the output-capital ratio. This hy¬
pothesis approximates the shape of the kinked capacity utilization schedule
in figure 2.1
An increase in investment demand would lead to higher levels of the
profit rate and output-capital ratio in figure 2.5. The labor-capital ratio
would also rise, and from (2.17) the real wage w/P would have to fall. The
adjustment to increased investment is a mixture of forced saving and
Keynesian responses.
One final observation is that the economy might be supposed to function
with full employment of both capital and labor. In that case, the labor/capi¬
tal ratio would take on a predetermined value. From figure 2.5 one can see
that the output-capital ratio, the profit rate, and [from equation (2.17)] the
real wage would all be determined. What about saving and investment?
With exports either fixed or depending strictly on price relationships, equa¬
tion (2.16) is a condition for macroeconomic equilibrium that must be
respected. Hence, investment and the capital stock growth rate would have
to adjust to the level of g following from the full employment assumption
and the economy’s technological and savings parameters.
Along with output changes, forced saving, and endogenous shifts in the
saving function, an investment function that always moves to meet saving
supply is a fourth macro adjustment mechanism that one might consider.
The problem is that it is not obvious what is changing to permit the
investment function to shift (as opposed to the trade deficit or government
budget surplus when there is an endogenous saving function). In fact, it is
probably simpler to dispense with the investment function altogether, and
assert that the capital stock growth rate results only from the forces of
productivity and thrift. This sort of macro adjustment might properly be
labeled neoclassical since it is based on the notion that the wage, profit rate,
and labor- and capital-output ratios can always vary in such a way as to
permit full employment of both inputs. That investment has to be endoge¬
nous is a natural corollary of such a system.
To illustrate the implications of this interpretation of neoclassicism, the
comparative statics experiment of increasing the saving rate is useful. In a
version of figure 2.5 with the investment function suppressed, the outcome
would be a counterclockwise rotation of the saving supply and a higher

capital stock growth rate g. Neither the profit rate nor the ratios describing
labor and capital use would change. The contrast with previous formula¬
tions is striking since in those a higher saving rate is associated with slower
growth, a lower profit rate, and either a lower level of output or a higher
real wage.
Which set of responses is more realistic is a question of the institutions
and historical circumstances in which an economy happens to be. Presuma¬
bly, neoclassical economists tend to think that there really is full employ¬
ment, while Keynesians and structuralists do not. The main thing to recog¬
nize is that a neoclassical formulation of production relationships is not
fundamentally different from a story involving excess capacity and markups
unless full employment is also assumed. But that amounts to a statement
about the grand structure of the macro system and not its specific de¬
tails.13
2.4 Profit Squeeze and Growth Dynamics in the Medium Run
As a final exercise, we can use the nonneoclassical model of sections 2.1

and 2.2 to explore growth and distribution over time as a prelude to the
models of chapters 6, 8, and 9. Suppose that the markup rate t is fixed at
any moment, determining the level of capacity utilization u as in figure 2.1.
If we treat u as the adjusting variable, then its value follows from the
condition that excess demand equal zero.
(2.18)
where the first term in parentheses represents investment demand (assumed

to depend only on the profit rate r — tu /(\ + t) (because of the coefficient
z,) and the second term is saving supply. The stability condition is that
saving responds more strongly to an increase in capacity utilization than
investment, or (4>/r) -(- sr > z,.
Dynamics in the medium run will center around changes in the markup
rate. Specifically, we adopt a chronology that appears in American business
cycles, as follows:
Phase of Profit Capacity Profit

Cycle Rate Utilization Share
Upswing Rises Rises Rises

Peak Falk Stable Falk
Downswing Falls Falk Falk
Radical economists (especially) interpret these cyclical variations in terms

of a theory of profit squeeze. At the top of the cycle, employment and labor’s
bargaining strength rise. Workers can force a reduction in the share of
profits in output, or a decline in the ratio r/(l -f r). The resulting fall in
markups also reduces the profit rate, leading to a contraction in investment
demand. Aggregate demand as a whole declines in a multiplier process, and
the cycle turns down. Over time, these effects may cumulate to lead to a
secular decline in profits and the rate of growth, unless the central govern¬
ment intervenes to prop up demand.14 Profit-squeeze or class-conflict theo¬
ries of distribution and growth of this type state that the effect of capacity
utilization on the markup is opposite to that in the neoclassical model in
the last section. The evidence for advanced economies (at least) seems to
run against the neoclassical hypothesis.15
To work out a formal model, we need to know the comparative statics
of responses of the profit rate r, the capacity utilization rate u, and the
capital stock growth rate g to changes in r. From (2.18) an increase in
r will raise investment demand by the amount zxu/( 1 + r)2. It will
cut savings generated directly and indirectly from intermediate imports by
— 4>u/(l + r)2, and raise savings from profits by sru/( 1 + r)2. Assume
investment minus saving rises, or z, > sr — <p. Then capacity utilization
will go up with the markup rate. The solutions for the other variables are:
=
r A
and
g = (r '4> + sr)^-
A
where
A = t-1<{> + sr — h > 0
By direct calculation, dr/dr > 0 and dg/dr > 0.

The profit-squeeze story depends on the notion that r falls when employ¬
ment and labor militancy are high. The labor-capital ratio L/K is equal to
bu. Let the labor supply consistent with high employment be L—in the
Marxist idiom the reserve army of the unemployed would shrink toward a
low level as L rose toward L. Let X = L/L and k — K/L. Then by easy
substitution:
Assume that the markup rate varies over time according to the rule
dr/dt = a( 1 - X) = a(l - buk) (2.19)
so that r falls when X grows and full employment is approached.16

If the fully employed labor force grow at the rate g*, then the growth
rate of k is given by:
dk/dt = (g - g*)k (2.20)
since (lettering a “hat” over a variable denote its growth rate)
k = (dk/dt)/k = K - t
and
K = (dK/dt)/K = g
In growth models such as the system (2.19) and (2.20), it is simplest

analytically to consider dynamics around a steady-state solution like the one
defined by X = 1 and g = g*. That is, there is full employment of labor
and the capital stock is growing at the same rate as the labor force at the
steady state.
Around such an ideal growth trajectory, the variables adjust according
to the Jacobian of (2.19) and (2.20):
— abk(du/dr) —abu
, (2.21)
k(dg/dr) 0
where the entry in the southeast vanishes since g = g* at equilibrium.

Following tradition for applied economists, in this book we will consider

only conditions assuring local stability of differential equations for adjust¬
ment around an equilibrium solution where time derivatives of “state”
variables such as r and k are equal to zero. For a two-dimensional system
like (2.19) and (2.20), the stability conditions are that the trace of the
Jacobian matrix must be negative and the determinant positive.17
The determinant of the matrix in (2.21) is positive, while the trace will
be negative when du/dr > 0. The conclusion is that when the level of
capacity utilization responds positively to the markup rate, then distribu¬
tional changes induced by a profit squeeze will lead the economy toward
steady growth. As indicated by the spiral path sketched in figure 2.6,
cyclical convergence toward the equilibrium could be expected.
Dynamics of Adjustment to Steady-State Growth in a Profit-Squeeze Model
The key linkages underlying the stable adjustment are the positive associ¬
ation of capacity utilization with the markup rate due to
1. a strong investment response to a rise in r, and

2. a saving decrease at the bottom of the cycle due to a lower relative domestic
valuation of intermediate imports as r starts to rise.
Forced dissaving from imports at low values of r together with the

investment response cause an initial expansion that is finally throttled by

the profit squeeze. Stability of the process evidently requires that there exist
a source of saving separate from that by capitalists. Fluctuations in the real
value of this saving source as well as a Keynesian adjustment underlie the
cyclical dynamics of figure 2.6. This adjustment process is complex, but it
does broadly fit the data. It would fit even better if it took into account the
role of inflation in prolonging forced saving processes over time. Such topics
are addressed in the version of the present model for an inflationary econ¬
omy, which appears in chapter 6.
A final question regards secular changes in r as the economy shifts from
the vicinity of one steady state to another. From the condition g = g*, the
equilibrium value of r is:
<t>(g* ~ z0)
T —
Z\g* - s,ig* - z0)
A positive value of t requires labor force growth to exceed autonomous

investment demand (g* > z0) and a relatively high investment response
parameter z ].
Comparisons across steady states follow from observing that in the short
run dg/dr > 0, dg/dz, > 0, and so forth.
Holding g equal to g* and then taking the total differentials shows that
the markup rate will fall secularly as labor force growth declines, invest¬
ment responsiveness to the profit rate rises, or the saving rate falls. Any or
all of these shifts could contribute to a long-run profit squeeze.
Adjustment
Mechanisms—
Two-Sector Models
O NCE any model of the economy is extended from one to several

sectors, changes in relative prices of the different outputs can play a key role
in the short-run adjustment process. The specific details, however, vary
from system to system. In this chapter, the price mechanism is illustrated
in two models corresponding to economic structures that appear widely in
the Third World. The first describes an economy in which food production
accounts for a large share of economic activity and consumer demand.
There is a nonagricultural sector similar to the industrial economy de¬
scribed in the last chapter, with a fixed markup and capacity adjustment.
The food, or agricultural, sector, by contrast, has fixed (or price-inelastic)
supply in the short run, while its price varies to clear the market. A model
is presented in section 3.1, with comparative statics in section 3.2.1
The other kind of economy described in this chapter is one in which there
are large mineral exports. The mineral sector provides resources in foreign
exchange, and the problem is how to transform dollars into effective non-
mineral production within the country. There is assumed to be a sector

producing nontraded goods with fixed capacity in the short run (think of
the construction sector, transport and energy, and other infrastructure
services). The price of such “universal intermediates” varies to clear the
market. One effect of a price increase will be to raise production costs in
the rest of the economy, assumed to have excess capacity and markup
pricing (over wage, national intermediate, and imported intermediate
costs). Products from the “manufacturing” sector are assumed to be used
for export and consumer demand.
The details of the model are set out in section 3.3, where it is shown that
expansion of mineral production can lead to a reduction in output and
exports of the manufacturing sector (as well as increases in prices and a fall
in the real wage). Hence, reduced internal and external diversification of the
economy may go hand in hand. These results are extended to describe
balance of payments problems in section 3.4, where the merits of devalua¬
tion and manipulation of the mining sector to improve the balance of trade
are assessed.
3.1 An Economy with an Important Food Sector
Two sectors are assumed to be macroeconomically important. We will

call them the agriculture (or the A-sector) and nonagriculture (or the
N-sector) sectors for short. The N-sector resembles industry and urban
services in most market economies by operating with chronic excess capac¬
ity and a fixed markup. If intermediate imports as an item of cost are
ignored for simplicity, then the N-sector price Pn is given by:
Pn = (1 + r)wb„ (3.1)
where r is the markup rate, w is the money wage, and b„ is the labor-output
ratio.
The agricultural sector, by contrast, is resource-limited: supply does not
respond to price or other incentives in the short run.2 For purposes of
formal modeling, land can be assimilated with capital so that sustained food
output increases can only come from investment activities such as land¬
clearing, mechanization, and works for irrigation. Available employment
opportunities are limited by capital in the A-sector in terms of full-time
equivalent jobs. For macroeconomic purposes here, we do not go into
Adjustment Mechanisms — Two-Sector Models 39
details about how full-time jobs may be split among numerous family
members and employed workers. Also, no distinction is drawn between
food sector capitalists and employees. All income flows go to “peasants,”
who are assumed to share common saving and consumption functions.3
In formal terms, agricultural output is determined by available capital
stock:
Xa = aKa (3.2)
where a is the output-capital ratio. Variations in a can reflect good or bad

weather and other supply shocks in food production.
Total income in agriculture is PaXa, where Pa is the market-clearing food
price. Nonagricultural workers are assumed not to save, while the saving
propensities for agriculturalists and nonagricultural profit recipients are
sa and sn, respectively. Then the value of total consumption demand D is
given by:
D = (1 - sa)PaXA + wnbnXn + (1 - sn)TwnbnXn

= 7 aPaXa + J nP „X „ (3.3)
where ya — 1 — sa andy„ = [1 + (1 — s„)t]/( 1 + r) are the propensities

to consume from agricultural and nonagricultural income flows, respec¬
tively. Note that dy„/dr = —s„/( 1 + t)2 < 0. Not surprisingly, an
increase in the markup rate reduces the propensity to consume from the
N-sector income.
Consumption spending is split between the two sectors. Probably the
best-established stylized fact in economics is Engel’s law, which states that
food consumption is inelastic with respect to income—a 10-percent rise in
real income will lead to substantially less than a 10-percent increase in real
food consumption. The simplest way to capture Engel’s law in a formal
specification is to write sectoral consumption functions as:
PaCa = aD + QPa (3.4)
and
PnCn = (1 - a)D - &Pa (3.5)
where Ca and C„ stand for consumption of food and nonfood, respectively.

It is easy to verify that food demand will be income-inelastic (and nonfood
demand income-elastic) when 0 > 0. These equations have the great

advantage of being linear in income and prices. They are a simplification
of the econometrician’s well-known linear expenditure system (described
more fully in chapter 4).
Excess demand functions for the two commodities can be written as:
EDa = Ca + E — Xa (3.6)
and
EDn = C„ + Ia + /„ - Xn = C„ + / - Xn (3.7)
where E stands for net exports of food, and 1 is investment demand (the
sum of sectoral investment demands /„ and Ia, on the assumption that
capital goods are produced only in the N-sector).
The conditions for macro equilibrium in the economy are EDa =
EDn = 0. One can multiply (3.6) by Pa and (3.7) by Pn, sum the equations,
and simplify to get the investment-saving balance:
PnI + PaE — saPaXa — sn(rw„bnXn) = 0 (3.8)
As in the last chapter, investment and net exports are financed in equilib¬
rium by domestic saving, the sources this time being agricultural and profit
incomes.
Rather than express all variables as growth or profit rates, it is simplest
in the present model to work out adjustment processes in terms of “level”
variables of the form presented so far. The price Pa rises when there is
excess demand in the food market, while X„ balances the N-sector. The
formal specification depends on the excess demand functions, which can be
written in detail as:
[(1 - a)yaXa - Q]Pa

EDn = - [1 - (1 - a)yn]X„ + + / (3.9)
Pn
and
EDa = aynPnXn _ (i aya)Xa -f 0 + E (3.10)
The specific adjustment rules are:

Adjustment Mechanisms—Two-Sector Models 41
dXn
f[EDn(Xn,Pa)]
dt /
and
—~ = g[EDa{Xn,Paj\
dt
where the functions / and g satisfy /(0) = g(0) = 0 and have positive first
derivatives. These derivatives only describe speed of adjustment (which is
presumably fast since we are dealing with a short-run model). Stability itself
depends on the properties of the Jacobian matrix of the excess demand
functions, which can be written as:
(1 - a)yaXa - 0
-[1 - (1 - a)yn]
ay nP„ ay„P„X„
Stability analysis follows the usual procedures.4 To check the trace condi¬
tion, note that the bracketed term in the northwest is one minus the propen¬
sity to consume nonagricultural products from nonagricultural income,
which will be a fraction. Similarly, the term in the southeast (aside from the
minus sign) is proportional to the propensity to consume food from nona¬
gricultural income, also a fraction. Both terms are clearly negative, and so
is the trace (their sum). The determinant of the Jacobian is:
ay nP„ (1 - a)yaXa - 0
A = r-[l - 0 - a)7.] ~ (3.11)
For A to be positive, the condition:
(1 - a)yaXa - 0 p
Pn[\ - (1 - a)yn\ °
must be satisfied. But note that when excess demand for the nonagricultural
product is zero, then from (3.9) Xn is given by:
Equilibrium in the Two-Sector Model When Demand for Food Is Relatively

Income-Inelastic (Upper Diagram) or Elastic (Lower Diagram)
Note: The shifts in the Agricultural market lines represent increased net exports or lower production
in the agricultural sector.
So long as investment demand / is positive, (3.12) shows that the above

condition for a positive determinant will be satisfied. The conclusion is that
the adjustment process based on X„ and Pa is locally stable.
Equation (3.12) is one relationship between the two endogenous variables
in the model. Either the investment-saving balance (3.8) or the food excess
demand function (3.10) could be used to finish solving the system. For
present purposes, it is clearer to use (3.10). With EDa set to zero, X„ can
be expressed as:
[(1 - aya)Xa - 0 - E]Pa

x, (3.13)
OLJnPn
To see how the system works, begin with the excess demand function for
the N-sector, equation (3.9). Demands are from investment / and nona-
gricultural and agricultural incomes. An increase in the agricultural price
reduces real incomes in the N-sector while it increases those of agricultural¬
ists. Thus, a rise in Pa causes demand for X„ from N-sector incomes to fall
(from both income and substitution effects) and demand from A-sector
incomes to rise. The former effect is likely to dominate when Engel effects
are strong, leading to a relatively large real income loss in the N-sector as
Pa goes up. In the algebra, a large value of 0 is what matters since this
parameter represents the portion of food demand that is insensitive to
changes in incomes and prices. In formal terms, 0 > (1 — a)yaXa is the
condition for X„ to respond negatively to Pa, as can be seen directly from
(3.12). The two possible cases are illustrated by the “Nonagricultural mar¬
ket” schedules in figure 3.1, with the upper diagram showing what happens
when Engel effects are strong.
The analysis of food market equilibrium is simpler. As (3.13) shows,
X„ will be an increasing function of Pa so long as food output Xa exceeds
the sum of marginal food demands from farmers (ayaXa), the constant
demand level 0, and net exports E. Equation (3.13) is graphed as “Agricul¬
tural market” schedule in figure 3.1.
3.2 Comparative Statics for the Economy with an Important Food

Sector
As usual, changes in policy or exogenous variables can be viewed as

shifting one or both of the curves in figure 3.1. The diagrams then show how
the agricultural price and nonagricultural output adjust to reestablish the
disturbed equilibrium. Several examples are presented in this section. They
reveal that the interaction of the two sectors creates enough potential
surprises to keep the macro policy team alert.
First, figure 3.1 itself demonstrates what happens when there is an in¬
crease in net food exports. At a given level of X„, excess food demand in
(3.10) rises in response to the increased sales abroad. The only way it can
be driven back to zero is via reduced domestic demand induced by higher
prices, so Pa must rise. The outcome is a clockwise rotation of the agricul¬
tural market schedule. Hence the domestic food price goes up in response
to increased export demand. If Engel effects are strong, the price increase
will choke off demand for N-sector products and Xn will decline (upper
diagram). Otherwise, both Pa and Xn will rise in response to the higher net
exports.
This result can be used to illustrate institutional responses in several
countries. When Engel effects are strong, nonfarmers would benefit in two
ways from lower net food exports (or higher imports)—the food price would
be lower and there would be a higher level of N-sector economic activity.
Under such circumstances, a political coalition in favor of gaining easy
access to imports could form. Examples might be the groups favoring repeal
of the Com Laws in England in the last century, or nationalistic supporters
of low-cost Japanese acquisition of rice from Taiwan and elsewhere before
World War II. By contrast, export promotion might well be favored politi¬
cally when Engel effects are not important at the macroeconomic level, and
the lower diagram of figure 3.1 applies. Exports would stimulate both farm
prices and industrial activity in this case, perhaps characteristic of food
exporters such as the United States in recent years.
These examples are implicitly based on the notion that the volume of
agricultural trade is controlled by the government, perhaps through a state
trading corporation. In fact, regulation of both internal and external com¬
merce in food products is the rule in most countries because of the ex¬
tremely low demand and supply elasticities that characterize the market in
the short run. Low elasticities naturally give rise to large fluctuations in
food prices and quantities offered for sale, which governments do their best
to avoid by regulatory maneuvers.5
If, improbably, free trade in food products were allowed, then the domes¬
tic price Pa would in principle be determined from the world market. Food
traders could make a profit by importing for domestic resale until such time
as the internal price came down to the world level. In practice, such conver¬
gence to the “Law of One Price” may be glacial, but the assumption is often
maintained in theoretical discussions.6
If Pa as determined by the world market increases from an initial equilib¬
rium position in figure 3.1, then adjustment would occur by a clockwise
rotation of the agricultural market schedule as net food exports go up. It
is easy to see from the excess demand function (3.10) that the immediate
export response would be given by:
- aya) - 0 - E]
Note that this response will be weak insofar as net surplus production
beyond fixed food demands (the term in brackets) is small. Indeed, one
could imagine cases in which this surplus would be negative, for example
when the propensity to consume food from food producer’s incomes (aya)
is high. The particular model developed here does not have a solution with
positive values of Xn and Pa under these circumstances, and other models
are unstable. Nonetheless, a perverse or very weak export response to world
price increases has often been pointed out in the literature as a serious policy
problem confronting such food exporters as Argentina or Thailand.7
Turning to other experiments, note that increased investment demand
will shift up the intercept of the nonagricultural market schedule in figure
3.1, driving up both X„ and Pa. A push toward more rapid growth should
then be accompanied by a shift in the terms of trade toward agriculture.
This possibility is of considerable historical interest, for example, in the
rapid growth spurt experienced by the Soviet Union in the first five-year
plan between 1928 and 1932. During that period, the share of investment
in national income rose from 14.8 percent to 44.1 percent, and national
income itself rose by 60 percent. By the end of the period, the investment
share in national income was more than twice as big as that of agricul¬
ture.
The traditional interpretation is that much of the increment in savings
necessary to support the extra investment came from the agricultural sector.
However, the above comparison of the income shares of investment and
agricultural value-added shows that the role of agriculture as a savings
source cannot have been great. Rather, the terms of trade shifted strongly
in favor of food producers (from 100 in 1928 to 164 in 1930 and 130 in
1932), and real urban wages dropped by about 50 percent. The resulting
forced savings on the part of workers supported much of the investment
effort. At the same time, Stalin’s dekulakization campaign generated an
ample supply of displaced agricultural labor for employment in urban
industries—in many ways capacity limitations on industrial output could
be broken by (inefficient) use of raw labor.
This interpretation of Soviet experience is broadly consistent with the
model at hand, where an investment push is bound to raise the agricultural
terms of trade.8 Other historical conjunctures would be interesting to ex¬
plore.
A more contemporary policy issue regards consumer food subsidies,
which have been installed in a number of poor countries. They satisfy
political necessity to pacify urban workers’ pressures to maintain their
living standards, but also are often criticized by international agencies as
being “distortions” leading to improper resource allocation. It is reasonable
to ask about the macro implications of modifying subsidy policies.
Suppose the consumer price of food is (1 — z)Pa, where z is the subsidy

rate. From appropriately extended versions of (3.12) and (3.13), one can
show that an increase in z will shift both schedules in the diagrams of figure
3.1 upward. The food price Pa will assuredly rise, and in fact one can show
that the consumer price (1 — z)Pa will also go up. The economics is simple.
The extra subsidy gives an increment in purchasing power to consumers,
but food supply is fixed. Hence, the consumer price must rise more than
proportionately to the subsidy to limit demand. Incremental food imports
would permit the consumer price to fall and real food intake to rise, but of
course could lead to balance of payments problems. One practical reason
for objections to food subsidies by regulatory agencies such as the Interna¬
tional Monetary Fund becomes clear.
Another observation is that decreases in food subsidies may possibly
reduce manufacturing output. In the upper diagram of figure 3.1, a down¬
ward shift in both schedules can lead Xn to fall. With strong Engel effects
present, the loss in purchasing power induced by the fall in the farmgate
price Pa can actually cut nonfood demand. By contrast, an increase in
subsidies will raise the level of economic activity. Evidently, food subsidies
can have important macro repercussions. Their installation or modification
should be subject to serious prior consideration.
Next consider a fall in food production. In figure 3.1, the agricultural
market schedule will rotate clockwise (just as from increased net exports),
while the nonagricultural schedule will move the same way. In the diagrams
of figure 3.1 the results are ambiguous; however, total differentiation of
(3.12) and (3.13) shows that X„ must fall and Pa rise in response to this
supply shock.
To illustrate the economics, figure 3.2 presents a four-quadrant diagram
of the model. From (3.9) total demand for N-sector output can be split into
the sum of investment and demand from the A-sector, or / -t- (Pa/Pn)
[(1 — a)yaXa — 0], and also demand resulting from N-sector income itself.
The first component is shown (with strong Engel effects) in the northwest
quadrant of figure 3.2; the other component is the straight-line function of
X„ beginning at point B. For the agricultural price PJt nonagricultural
output X„ is determined by the intersection of total demand for N-sector
output with supply along the 45°-line.
The lower quadrants show how the model can be solved using the invest¬
ment-saving balance (3.8). An amount PnI — s„TO)nb„X„ of saving is
required in equilibrium from agriculture, after nonagricultural saving sup¬
ply is subtracted from investment demand. This declining function of
X„ appears in the southeast quandrant. Finally, saving supply from agricul-
A Four-Quadrant Diagram for the Economy with an Important Food Sector
Note: A reduction in agricultural supply reduces both demand for nonagricultural output and potential
saving from the sector, leading Pa to rise and X„ to fall.
ture is (saxa — E)Pa, shown in the southwest quadrant. For overall equilib¬
rium, both excess demand for the nonagricultural commodity and econo¬
mywide excess demand (investment minus saving) must be driven to zero
at Pa and Xn. In the diagram, adjustment occurs via upward or downward
shifts of the X„ demand function in the northeast quadrant.
Now consider a food supply shortfall, or a fall in Xa. For a given Pa, both
farmers’ demand for X„ and their supply of saving will decline. In the
diagram, the schedules in both the northwest and southwest quadrants
rotate toward the Pa axis, and the outcome is a higher Pa and lower Xn.
The sort of supply shock shown in figure 3.2 can easily set off a burst of
inflation if money wages respond to the rising agricultural price and drive
up N-sector costs in turn. A sequence of such shocks could keep inflation
going if, for example, population growth runs ahead of the growth rate of
agricultural supply. This sort of inflationary process is discussed formally
in chapter 9.
Two final exogenous shifts of interest are increases in the urban nominal
wage w and the markup rate r. Observe from (3.12) and (3.13) that the
whole model is linearly homogeneous in Pa and Pn—the output side would

be unaffected by any proportional increases in the two prices. An increase
in w is automatically passed along into Pn via the markup without affecting
other variables or parameters. Because of homogeneity, the only response
of the system would be an equiproportionate increase in Pa. A money wage
increase is purely inflationary and has no effects on resource allocation at
all. Of course, this result would not carry over if intermediate imports were
introduced as elements of N-sector cost, as in the model of the last chapter.
A money wage increase would then make imports relatively cheaper, reduce
potential foreign saving, and lead to both higher Xn and Pa.
An increase in the markup rate both drives up P„ and reduces the
nonagricultural consumption propensity yn. As with a wage increase, the
first effect is purely inflationary in the present model. However, the decrease
in y„ will reduce demand for food. The implication is that Pa will rise less
than proportionately to r, and might even fall. N-sector output unambigu¬
ously declines in response to lower aggregate demand. As shown in the
growth version of this model in section 9.2, these results regarding price
(and inflation) effects of changing the markup carry over into the long
run.9
3.3 Intermediate Inputs in a Mineral-Exporting Economy
The sort of model used for agriculture and industry in the previous
sections can be extended to deal with many other economic structures in
which prices vary to clear some markets and output movements clear
others. In this section, we sketch a model in which intermediate inputs play
a crucial role. The economy in question exports a fixed quantity of a
commodity, such as a mineral, which requires nationally produced nonim¬
portable intermediates like transport services, energy, and other infrastruc¬
ture. These “universal” inputs are assumed to be in fixed supply, and feed
either mineral production or another “manufacturing” sector with markup
pricing. Mining itself typically generates only modest wage flows that lead
to demand for manufactured goods since export profits come in the form
of foreign exchange. If mineral output expands it will pull intermediates
from use in manufacturing on the one hand, and generate demand for
manufacturing output on the other. In addition, there can be cost-push
price increases in manufacturing if the intermediate goods price rises to
clear that market. The final outcomes for manufacturing prices, exports (if
they are price-responsive) and outputs is unclear. And the situation

becomes even more complicated if foreign exchange is in relatively short
supply (the case for most mineral exporters except those with major petro¬
leum pools). With a binding restriction on the amount of foreign exchange
the country can spend, short-run variations in the exchange rate may be
required to stimulate manufacturing exports and keep imports in line.
Feedbacks from the exchange rate to the rest of the system then have to be
taken into account.10
To set this up formally, let mineral output (equals volume of exports) be
Xm, which is fixed in the short run. The volume of output of universal
intermediates, Xu, is also fixed, but output of manufactures (or, really, the
rest of the economy) can vary. The level of production in this N-sector is
Xn, and the price Pn is given by:
Pn = (1 + T)[wbn Puaun + eP%a0n] (3.14)
In this equation, the markup (at rate r) is taken on wage cost per unit output
wbn, cost of the universal intermediate Puaun, and imported intermediate
costs eP$a0n. The input-output coefficients bn, au„, and a0„ are assumed
constant in the short run.
On the demand side, assume that Engel elasticities are unity. To save
notation, all profits will be assumed to be saved and all wages spent on
consumption. Then the consumer demand functions for N-sector outputs
will be
PnC„ = (1 - a)w[bnxn + buXu + bmXm]
and
PUCU — Q-w\bnXn + buXu + bmXm]
where it is assumed that the same wage w is paid in all three sectors.11 The
coefficient a will presumably be positive but small, as much less than
one-half of consumer spending goes toward energy, transport, and other
products from the U-sector. In these equations, the contribution of the
mineral sector to consumption demand is likely to be low since only wage
income is consumed. (The share of mining wages in the total across the
economy might be as low as 5 percent.12) In practice, some share of the
export receipts from mining will go to the national government, and will
be spent. This important potential source of demand is left out in the present
formulation, which concentrates on short-run responses.
Excess demand balances for the nonmineral sectors are:
Cn 4~ E -f- G — Xn — 0
and
Cu + ounXn + aumXm 4- 0/ — Xu — 0
In the second equation, demands for the universal intermediate come

from consumption, intermediate sales (aunXn 4- aumXm), and investment.
The key assumption regarding the latter is that a fraction, 0, of the demand
level / is produced nationally by the U-sector. In practice, 0/ represents
the construction component of capital formation, and the implicit hypothe¬
sis is that output of construction services is limited by skilled labor and
equipment shortages. Such an assumption makes sense in some countries
but not in others; it is mostly adopted to generate an interesting story here.
On the side of manufacturing, assume that volume of sales abroad E is
given by:
E = [€o + €,( eP*/Pn)]Xn = eXn (3.15)
An essentially noncompetitive market in exports is assumed. The price of

national products sold abroad is Pn/e, or domestic price divided by the
exchange rate. The price in the international market of products against
which “our” manufactured exports compete is P*. As P* rises relatively
to Pn/e, a larger share of domestic output goes abroad. The elasticity of this
share with respect to eP*/Pn can easily be shown to exceed one when
€0 < 0, and be less than one when e0 > 0. With any magnitude of the
elasticity, this export response will of course take time to build up, but such
lags are ignored for simplicity.
Putting all these equations together gives excess demand functions for the
manufacturing and intermediate sectors as:
(1 - a)w
[bnXn + buXu -f bmXm\
(1 + T)[wbn + Puau„ + eP*a0n]
eP*
+ 6 + €o + *i
(1 4- r)[wbn 4- Puaun +
- X, = 0 (3.16)
and
— \b„Xn + buXu + bmXm] + ou„Xn

* u
+ aumXm 4- 0/ — Xu = 0 (3.17)
The adjusting variables are X„ for (3.16) and Pu for (3.17).13 The equa¬
tions are written out in all their glory to illustrate the fact that inclusion
of intermediate inputs in a model substantially complicates its accounting.
Compared to (3.9) and (3.10) for the agriculture/industry model, these
excess demand functions both have more entries and (more importantly) are
nonlinear. Because the intermediate price Pu enters into the determination
of the N-sector price P„, terms in products or quotients of Pu and manufac¬
turing output Xn are unavoidable in (3.16) and (3.17).
Even without writing the Jacobian, it is straightforward to see that (3.16)
gives a negatively sloped relationship between Pu and X„. Excess demand
for manufacturers falls when Xn rises (more supply) and also when Pu goes
up (less consumer and export demand from intermediate cost-push into a
higher Pu). Hence, the two adjusting variables must trade off inversely in
the (Pu, X„) plane. By contrast, an increase in X„ generates more consumer
demand for the U-sector in (3.17), which must be limited by an increase in
Pu. A positively sloped curve results.
The two schedules are shown in figure 3.3. The shifts are due to an
FIGURE 3.3
Response of the Mineral-Exporting Economy to an Increase in Mineral Output
Note: The intermediate price rises and manufacturing output can shift either way.
increase in mineral output. More mining activity generates greater interme¬

diate demand, so the “Intermediates” schedule must shift to the right as
Pu rises to ration the fixed supply. At the same time, the mineral wage bill
goes up, stimulating demand for manufacturers, and that schedule shifts up.
The outcome is a rise in Pu and an ambiguous response in Xn.
To track down the ambiguity, we can begin with an algebraic expression
that determines the sign of the N-sector output response:
awL{\ — a)wbn
awbm + aumPu
[(1 - a)wL + t,eP*nXn}(\ + r)aun (3.18)
in which L = bnXn + buXu + bmXm, or total employment.

The first positive term represents the consumer demand generated by
more mining employment. The negative entries show how an increase in
P„ from intermediate cost-push reduces consumption and export demand
for manufacturing output. Depending on which effects are stronger, the
output response can go either way.
Figure 3.4 illustrates the comparative statics for two special cases. In the
upper diagram, aun is set to zero, so that manufacturing does not require
intermediates. With unitary income elasticities (no Engel effects), Xn is
independent of Pu in (3.16) and rises along with miners’ consumption
demand from Xm. By contrast, a is set to zero in the lower diagram, so that
there is no consumer demand for U-sector products. N-sector output is
determined in (3.17), and falls as its available intermediates inputs are
reduced by an increase in Xm. The relative importance of these two effects
depends finally on the relative magnitudes of a and aun.
The ambiguous response of Xn to an increase in mineral production
illustrates a classic problem of single-export economies. When output in the
export sector goes up, it may choke off or at best weakly stimulate nonex¬
port production. Indeed, if manufacturing exports are reduced by interme¬
diate cost-push, declines in both “external diversification" (of the export
basket) and “internal diversification” (of the production structure) go
together.14 More mining drives down both exports and output levels in the
rest of the economy. The loss in production flexibility and the noneconomic
retardation caused by lagging domestic production of everything but a
mineral can both be very great.
Other comparative static responses depend on the linkages just discussed.
A higher level of investment demand puts pressure on the intermediate
Manufacturing
Effects of an Increase in Mineral Output When the Manufacturing Sector Does

Not Require Intermediate Inputs (Upper Diagram) and When There Is No
Consumer Demand for Products of the Intermediate Sector (Lower Diagram)
goods market, so that Pu rises and Xn falls. An increase in intermediate

output Xu has the opposite effect, with corresponding responses in the
endogenous variables. Higher government spending G pulls up both Xn and
Pu. Finally, a rise in the wage rate w drives up Pu by creating consumer
demand. The effect on Xn is ambiguous but positive if additional consump¬
tion demand by workers for N-sector products out weights the reduction in
exports resulting from a higher price level (Pn/e) of national products
abroad.
Finally, note that the real wage defined in terms of the workers’ consump¬
tion basket is:
w w
(Pny~a(Pu)a (^«)aKl + r)[wbn + eP*a0n + Puaun]Y~a
so that the real wage falls when Pu goes up. On the saving-investment side
of the model, the corresponding increase in real profits underlies a forced
saving macro adjustment process. Setting out the details is left to the reader.
The derivations are tedious, but follow directly along the lines of the models
already described.
3.4 Balance of Trade Complications
Though there are striking exceptions (mostly in the Persian Gulf), min¬
eral-exporting countries by and large suffer from balance of payments pres¬
sure. The historical sequence is that after an initial phase of building up
foreign exchange holdings when mineral wealth arrives, pressures to in¬
crease imports rise and the corresponding coefficients shift up. The govern¬
ment learns that it can finance more spending on N-sector goods by holding
foreign exchange in the form of central bank reserves and creating national
money against that base.15 The outcomes in terms of the model sketched
in last section are relatively high levels of G, (1 — 0), and a0„. The balance
of payments deficit in terms of dollars becomes:
B = P*0a0nXn + nO - 0)/ - PIXm

- [€o + €y{eP*/PH)\X„(PK/e) - F (3.19)
where F is an endogenous level of capital inflows net of interest and amorti¬

zation costs. The deficit B may be near zero or positive (indicating loss of
central bank reserves) even though F is high because a mineral-exporting
country often has good credit in Eurocurrency and similar financial markets
abroad. The other new terms in the equation are P* or the cost of imported
investment goods in the world market, and P* or the world price of mineral
exports.
Suppose that the trade deficit suddenly gets worse, from drying up of
foreign loans, an increase in investment demand, or some such similar
shock. Reserve outflow can accelerate for a time, but finally some sort of
adjustment must be made. Devaluation is one standard policy choice.
The effects of an increase in the exchange rate e on the trade deficit
basically run through intermediate imports and manufactured exports since
the mineral export price P* is fixed in terms of dollars. The longhand
expression can be derived by totally differentiating (3.19) and plugging in
expressions for dXJde and dPu/de from the model of last section. The
result is lengthy but enlightening. It is:
€o —(1 + T)aun
e
(3.20)
where A > 0 is the Jacobian determinant of the (stable) system of two

excess demand functions, and y = G/Xn.
To disentangle the knots, begin with the last expression in (3.20). This
term comes from the expression multiplied by (Pn/e) in (3.19), and shows
that when e0 > 0 devaluation makes the trade deficit worse. The explana¬
tion is that an increase in e lowers the world price of national exports
(Pn/e) immediately. However, export volume presumably rises in response
to the price improvement. If a time lag is involved, the trade surplus may
first grow smaller and then gradually increase. Graphed against time, the
surplus looks like a fish hook, or the letter J. Hence we have a “j-curve”
dynamic export response.16 For present purposes it is simplest to ignore this
process, though it may take well over a year to work itself out. The main
point is that the overall payments response must include the world price
increase induced by a rise in e.
The bracketed term at the beginning of (3.20) shows how N-sector de¬
mand responds to devaluation. When the export response term €) is small,
Xn will decline when e goes up—this is the contractionary response to
devaluation illustrated in figure 2.4. By way of variety, we assume an
expansion in output when e rises, or in other words when the export
response is strong.
However, to get an improvement in the balance of payments {dB/de < 0),
not all hurdles have been crossed. One helpful effect (at least when the
exchange rate elasticity of exports is less than unity, or e0 > 0) comes from
the second term in the curly brackets, showing how Pn rises from intermedi¬
ate cost push, partly offsetting the decline in competitiveness resulting from
the initial increase in e. But finally, we also require e(Pn/e) — P%a0n > 0,
or the export share in N-sector output (measured in dollar prices) exceeds
that share of intermediate inputs. If this condition is satisfied by a wide
enough margin, then the j-curve will turn up and devaluation will improve
the balance of payments. The strength of the response clearly depends on
many parameters in the system.
The conclusion is that devaluation, if expansionary, will help the mineral
exporter’s trade deficit, with both internal and external diversification. One
policy cost would be a reduction in the real wage, from an increase in

Pu and e in (3.18). This shift, plus possibly weak responses of manufactured
exports to exchange rate changes in the medium run, makes a devaluation
strategy possibly infeasible and certainly difficult to implement over
time.17
The alternative is always to manipulate the level of mineral exports in
response to a payments crunch. How feasible such a policy is depends very
much on the characteristics of the commodity involved. Changing oil pro¬
duction volume or even capacity may be fairly simple; for copper or tin the
required modifications in output flows could be fairly hard.
There is also a question of which way to go. The obvious response is to
increase production from the mines and suffer the loss of economic flexibil¬
ity entailed in lower internal and external diversification (as discussed
above). But in some cases it is possible that reduced mineral production
would lower Pn and raise Xn enough to increase net exports. Contemplation
of the model suggests such an outcome is more likely when export responses
to price are strong and the share of nonmineral exports in the total is already
high. Such a situation might apply for a country that has already pursued
the devaluation strategy for quite some time. But until then, devaluation
and (more plausibly) export incentives may be an unavoidable policy
choice. The details in the incentive case are interesting, but are left to the
reader to work out.
4
Short-Run Adjustment
in Practice
TA HE previous chapters described four mechanisms through which

aggregate saving supply is brought into equality with the value of invest¬
ment demand. These are variation in the level of output; forced saving
caused by changes in prices relative to a nominally fixed magnitude such
as the wage or exchange rate; variation in the value of sales or purchases
of a commodity (for example, endogenous determination of exports or
competitive imports through the condition that excess demand for a com¬
modity must equal zero); and determination of the level of investment by
available saving. Now we will illustrate the roles that the first three mech¬
anisms play in a practical model. We leave neoclassical determination of
investment by saving for others to pursue.
The model described here is for the Indian economy in the national
accounts year 1980-81 (beginning April 1st). It generally follows the theo¬
retical structure of the agriculture/industry model set out in sections 3.1
and 3.2, but with five production sectors (two with flexible prices and three
with varying outputs) instead of just two. Similar systems have been con¬
structed for several countries besides India, so the discussion really refers
to a class of models in active use for macroeconomic analysis in the early
1980s rather than an isolated example.1
The data base for the India model (and all the others) is arranged in the
form of a social accounting matrix, or SAM. The Indian SAM is described
in section 4.1—it amounts to a blown-up version of the accounting relation¬
ships developed above. Section 4.2 sets out the equations for the model,
which are closely linked to the structure of the SAM. A brief discussion of
how values are assigned to the parameters appears in section 4.3. Section
4.4 describes the numerically calculated Jacobian of the Indian model’s
excess demand equations. This leads to the analysis in section 4.5 of a
number of comparative static exercises with the model, as parameters or
exogenous variables are perturbed to shift it away from its base solution
(which is described by the SAM). The emphasis is on the role that different
adjustment mechanisms play in this applied model. Conclusions are sum¬
marized in section 4.6.
4.1 A Social Accounting Matrix for India, 1980-81
A social accounting matrix is a tabular representation of the accounting

identities stating that incomes must equal outlays for all sectors of the
economy. Typically, deliveries of commodities or incomes are put in the
rows of the matrix, and receipts in the columns. Corresponding row and
column totals must be equal to make the double-entry income/outlay book¬
keeping balance out. During the 1970s, such matrices were constructed for
a large number of countries—they seem to be generally accepted as a
compact device for presenting input-output and national income ac¬
counts.2
The structure of the Indian matrix is illustrated schematically in figure
4.1. The principal blocks are labeled in the diagram. It makes sense to
describe them briefly before going on to the numerical SAM.
In the northwest corner is the interindustry matrix, which gives the
values of intermediate commodity flows among the five sectors distin¬
guished in the economy. Because the matrix identities are valid only in value
(and not “real” or quantity) terms, these flows are in 1980—1981 Indian
prices (in units of crores, or 10 million rupees). The five sectors are food-
producing crop agriculture (cereals, pulses, and vegetables), other agricul¬
ture, manufacturing and construction, infrastructure and energy-producing
sectors (transport, mining and quarrying, electricity, gas, and water), and
services. In the model described below, the two agriculture sectors are
assumed to be price-clearing, whereas the others reach demand-supply
balance through changes in output.
Short-Run Adjustment in Practice 59
FIGURE 4.1 Row Totals

Block Diagram of the Indian Social Accounting Matrix
Immediately below the input-output flow table are blocks giving sources
of household and government revenues, sources of foreign commodities (or
uses of foreign exchange for imports), and taxes and subsidies. All the
accounting in these rows is in current Indian price terms except the import
row, where purchases are valued in c.i.f. border prices (multiplied by the
rupee/dollar exchange rate). Differences between prices of imports at the
border and the prices at which they are sold internally are captured in
entries in the tax and subsidy rows of the SAM.
The first five column sums give the cost structures of the producing
sectors, exhausted by intermediate purchases, payments to the government
and households, import costs, and indirect taxes less subsidies. By the
standard SAM accounting rule, these column sums must equal the corre¬
sponding row sums, which represent total sales by the sectors.
To the right of the input-output matrix are blocks for uses of resources.
The household sector purchases outputs of the sectors, saves, receives subsi¬
dies, and pays direct taxes. The government also buys commodities on its
current account (its capital expenditures are in the investment column), and
saves and pays subsidies. In the investment columns appear outlays for
gross fixed capital formation and stock changes—their total is of course
balanced by the sum of the entries in the saving block. Exports (positive)
and competitive imports (negative) are valued in border price terms. The
difference between border and internal prices of exports is made up in the
export subsidy column. Exports less competitive imports comprise sources
of foreign exchange; the other major sources are remittances from emigrant
workers that go into household income, and foreign saving, which comes
into the economy on foreign capital account to balance the current account
deficit.3
Table 4.1 is the numerical SAM for our Indian model.4 Some of its details
are worth pointing out. Note first that income flows to households are split
three ways. Agricultural households in row 6c get income from the first two
sectors, while income generated in nonagriculture goes to wage-earning and
nonwage-earning households. The nonagricultural households get govern¬
ment wage, transfer, and interest payments (column 7) and emigrants’
remittances (column 9a). Note also that government revenues in row 7
include income from public sector production activities outside of agricul¬
ture.
Household uses of income appear in column 6. Only totals are shown in
the matrix, though in the model described below distinct consumption
patterns are postulated for agricultural and nonagricultural households.
Note the entry with negative value —545 in row 11. This represents subsi¬
dies to food consumption; along with producers’ subsidies signed with a
minus in columns 1-3, it is offset by the government’s total subsidy outlay
of 1327 in column 7.
Besides its production activities, the government’s revenue sources in row
7 include indirect and direct taxes (columns 11 and 12). The sources of these
taxes appear in rows 12 and 13. All direct taxes are assumed to be paid by
households (2941 in column 6), whereas the indirect tax burden falls on
production and sales in the nonagricultural sectors. Government uses of
funds appear in column 7.
Transactions with the rest of the world are recorded in row 9 for imports
and columns 9a and 9b for exports and competitive imports. The negative
entries for the latter in the commodity balance rows 1 through 5 are offset
by a positive total of 3782 in row 9, so that the column 9b sum is zero. Most
competitive imports come into the manufacturing sector. Besides exports
and remittances, the other source of foreign exchange include 1800 as the
current account deficit or foreign saving in row 8. Household saving is
22637 and government current saving is 2980; their total of 27417 is equal
to investment demand, or the sum of the column totals 8a and 8b.
04 v© vo m T*
O Tf —< oo
a 11 oo
CM
m
v©
04
O m
m
S *->
*—< o-
m m
«—ii
On
O'
vO
ON
m o4 — vo 3 04 04 -
m 04
& *
QK
■'T
ON
o
^ £
je K
8.1
%
I I I
II
I4
04
O' O'
Ss 04
3
!
A Social Accounting Matrix for India, 1980-81
"IS ac
tog
rS -3
o s;
c „
TABLE 4.1
?5 g ON Tf OO
8
» ? 5 On
Tt-
mm
3 O' —*
04
O
ON
*1!
<3
O' —
m oo
04 oo 3
s<s,
Os 04 m oo
O oo
ir\ m
3
On
04
rj-
OO
On
—■ On
All values are in crores (ten million rupees) at current prices.
to
3I SI & Tt O OO «—• vO
^ o s:4 on
On
m
OO 04
^g^ o
to «B
m m 04 04
S''S.'S
2 S' c O O vo
04 vo
O oo
04 m
m oo ON ^ ^ O
cI 3 04
■5S 3
rf m *—'
■*t
04
■<t
04
VO
O'
04
.£ C
60^53 O' ON
0^3
•O O Q O' o
O m
•§ e 3
r>j ^0
On
OO
^
—
—'
«—'
O
O-
04
•“1 ''T
^^ a
<u
00 6
3 5 c o
C
3 •5 G q c "§ + X
'g o .2C OG «u «oo *73
5 XI 3 x o. 2 o
oo
CO 1 ££> <23 jg op -a 00 co 3 <u
£ , x
X
c
S E S c a E5 3
.2 § v
60
2
g O T CO « O B
x> P-
I § e ° g- s 3 UJ
^ e
£ O *5 2 *5 e
<U
<u Z
(/)
S Z c < .S vg £
£ O .£ o W Cl, w 5 H o
H
O
Z
—■ 04 m & OO O' O «—< 04 m
With its double-entry bookkeeping, the SAM provides a data base around
which it is easy to construct a general equilibrium model. The SAM in fact
represents the base solution of the model since, as we will see in section 4.3,
the parameters of its equations are calculated to make their numerical
solution agree with the SAM. Of course, one has to first produce a SAM
to be able to follow this procedure.
Fairly complete recipes for constructing social accounting matrixes are
available in the literature, so there is no need to go into great detail
here.5 The usual approach is to begin with an input-output table, the
national accounts, and information from a household survey on patterns of
consumer spending. In general outline, the following steps will produce the
matrix:
1. Sectoral levels of value-added from the input-output and national accounts data
must be made consistent, typically by adjustment of both.
2. Next comes disaggregation of the value-added totals into income flows to house¬
hold and government sectors to be distinguished in the analysis. For example,
sectoral value-added in the Indian SAM had to be split among the entries in rows
6a, 6b, 6c, and 7 of table 4.1, following national accounts and other information
about the functional distribution of income.
3. The national accounts will give a control total for household saving. Savings
levels from the different income recipient classes (H of them) have to be estimated
to satisfy this total, typically on the basis of cross-sectionally estimated saving
parameters.
4. After taxes and saving are subtracted from household incomes, one is left with
H totals for values of household consumption by income recipient class. If there
are N sectors in the economy, the input-output table gives N totals for consump¬
tion of the different sectoral outputs. One thus has the row and column sums
(perhaps after correction for consumer taxes and subsidies) for the N X H matrix
of values of household consumption by type of product and income recipient
group. Subject to these control totals, the matrix has to be filled in. One typically
begins with existing estimates of household expenditure patterns, then modifies
them until the control totals are satisfied. The computations can either be done
informally or with a matrix-balancing computer algorithm such as the procedure
called RAS.6
5. Input-output and foreign trade information can be utilized to fill in the rest of
the sectoral final demand columns and the import row. Government fiscal data
usually suffice for the tax, subsidy, and government expenditure accounts. The
final check on filling in the matrix is that saving supply should equal investment
demand. After enough informal reconciliations between inconsistent data sources
are made, overall balance can usually be worked out.
4.2 A Macro Model for India
With the SAM constructed, the next step is to set out equations for the
model around it. The Indian model is based on the notion that the two
agricultural sectors are price-clearing, while output levels vary to satisfy
demand in the rest of the economy. Much literature in India points to an
“infrastructure bottleneck,” which suggests that sector four might best be
treated as price-clearing subject to fixed supply (like the U-sector in the
model of sections 3.3 and 3.4). However, numerical results show little
demand variation for sector four, so it was deemed simpler to treat it as
quantity-clearing to avoid complications in the model solution algorithm
described below.
The model’s equations appear in table 4.2 and the variables and parame¬
ters are described in table 4.3. The equations are set out in blocks, to be
described in succession.
Block I contains input-output balances, which set demand (to the left)
equal to supply for the five sectors. The entries correspond to the first five
rows of the SAM, with the convention that base-period prices for sectoral
outputs are all set to one. Note in equation (4.1) an entry (GPX — FA ,
— FN i) that reflects government food market interventions described
below. Also, in (4.3) national production of intermediate inputs to sectors
one and two is pegged at the level Xf. This quantity represents an exoge¬
nously specified level of fertilizer output. The balance of fertilizer demand
(«31*, + anX2 — X/) is assumed to be met by imports.
Block II contains an equation making accumulation of private food
stocks an increasing function of the market food price. The equation is
included in an attempt to capture the macroeconomic effect of food stock
speculation, which is often pointed out as a possible source of inflation in
India. Equation (4.6) gives a positive feedback from a rise in the food price
to increased stock accumulation and hence to a still higher price. The
practical importance of this potentially destabilizing demand response is
discussed below.
Block III has equations defining agricultural and wage incomes Ya and
YValue-added in the two agricultural sectors is calculated in (4.7) to give
Ya. Note that farmers pay a regulated price Pf for their fertilizer inputs.
The government also gives them a guaranteed “procurement price” P* for
a fixed quantity GP \ of food sales (the procurement). Production above the
procurement level (Xx — GP ,) is sold at the market price P,. The govern¬
ment’s accounts for this complex set of food market interventions are set
out below.
TABLE 4.2
Equations for the India Model
I. Demand-Supply Balances by Sector
2 a\jXj + {GP, - FA, - FN,) + C, + AS, + G, + Ex - Mx = Xx (4.1)

y=1
5
2 ayX} + ^2 + SS2 G2 -\- Ei — M1 = X1 (4.2)

t-1
5
Xf 4- 2 aJ/^0 + C3 + '£'3+^3 + SS y + G 3 4-^3 — 4/3 = .T3 (4.3)

>-3
5
2 a+jXj 4- C4 4- AS14 + C4 + £4 - A/4 = T4 (4.4)
y-'
5
2 BsjXj + GS5 4" C5 4- AS} 4- Gs + £j — M y — X y (4.5)

y-1
//. Cereal Stock Speculation
A 5, = ASK/V/’ir* (4.6)
III. Generation of Nonprofit Income Flows
Y„ = WA) + W, - GA) - (2 «/>A + <*3.^/ + 2 «iiW.

1=1 1=4
2 5
+ 2 — ^ fl/2^/ — anPf — S i)X2 (4.7)
/=1 1=4
5
CPI = 2 a,/*, (4.8)

/=1
h-, = H-* 4- MC/Y - CPT). / = 3, 4, 5 and g (4.9H4.12)
Kh, = Wjb,Xi 4- iv?Lg -f“ ^ 4" 7^ (4.13)

/=3
/K Price Equations for Nonagricultural Sectors
T/ = kj^Xt/KtYi, i — 3 and 5 (4.14H4.15)
^3 - hs/’i + anPi 4- at}Pt 4- a})Py

1 - (1 + rj)(l 4- Tj)flj3(l - sub3)
+ ^ yb y 4- OoS^JhO — ^ 03)] (4.16)
„ (1 4- r4Xl 4- f4)
l3* — . ... . [ .IL ai*Pi + ai*Ps + **'4^4 + aMeP04(1 - '04)] (4.17)
1 - (1 4- t4)(1 4- '4)044 /-i
„ (1 4- T,X1 + ty) *
P> = . ... , Iz a»pi + ‘*'3*5 + a0}eP$,(l - /os)] (4.18)
1 - (1 + Ts)(l + t,)a„ ,-1
V. Equations for Variable Costs in Nonagricultural Sectors
^3=2 i + ^3^3 + “ *(») “ (4.19)

/-I
5
B/ = 2 a<A + + Oo//,J;(1 - t0j),j = 4 and 5 (4.20)—(4.21)
/-I
VI. Generation of Income from Nonagricultural Markups
5
Y, = 2 - GR,) - (1 - 4>)DEP + GINT (4.22)
1=3
VII Consumption From Agricultural and Nonagricultural Incomes
Da = (1 - M (4.23)
£>„=(!- + (1 - OO - t,)Y, - eP*cC0 (4.24)
VIII. Sectoral Consumption Functions
-
Da = 2 ©W (4.25)
1=1
D„ m X ©7^, (4.26)
/=1
Ca, = ©“ — F/4, + (mUPiXD. + P,FAl - Da) (4.27)
C°, = 0? + (.mVPtXD. + P\FA, - D.), / = 2.5 (4.28H4.31)
C? = ©7 - FNX + (mf/i>,)tZ>„ + (P, - P,)FN, - D„] (4.32)
Cn, = ©7 + (m"/Pi)[D„ + (/>, - P,)/W, - 3,], i=2-5 (4.33M4.36)
C, = C| + FA, + C? + FN x (4.37)
C, = C° + C},i = 2, .... 5 (4.38H4.41)
IX. Model Closure
X, - X, == 0, / = 1 and 2 (4.42H4.43)
X. Governmental Balances
Tmd = i /,(1 + r (4.44)

i=3
5
= r,n + rind + 2 (4.45)

/=3
G£XP = P.Tvl, + (/>, - P,)/W, + (Pi - P*)GPi + T’sGSj + 2 Pfii

i= 1
+ (P i — cP*/)Xf + (eP*/ — Pf)(a j\X i + 033X2) + wgL * + 7*

5
+ 4>DEP + GINT + sub,a33X3 + X (^ - eP*)E,

i= 1
5
+ 2 a OiePtit 0iXj (4.46)

1=1
X7. 7>o*/e Deficit

5 5
DEE = 2 W + eP&Co + 2 ePtfl0,X/ + ePJ/Zo

/=1 /=1
5
4- ep*f(ai\X\ + (*32X2 — X/) — 2 eP*iP*i ^ (4.47)

1=1
XII. Saving-Investment Balance
SaYa + swYw + 5,(1 - on + (C/?£K - G£X/>) + DEF
= P3/3 + ePoVo + 2 S, + />>(Gi>, - FAi - /W,) (4.48)

/=1
TABLE 4.3
Symbols for the India Model
I. Sectors
1. Food agriculture
2. Other agriculture
3. Manufacturing industries
4. Intermediate production (energy and transportation)
5. Services
II. Endogenous Variables
X, — sectoral output levels, / = 1 , .... 5

Pi — sectoral price levels, / = i, . . . , 5
C, — sectoral levels of consumption demand, i = i , . . . , 5
C? — sectoral consumption levels from agricultural income, / = 1 , . . . , 5
C" — sectoral consumption levels from nonagricultural income, / = 1 .5
— sectoral wage rates, / = 3,4,5
wt — government wage rate
B, — variable costs per unit of output, i = 3,4,5
T3 — markup rate in sector 3
t5 — markup rate in sector 5
Ya — income in agriculture
Yw — wage income
y2 — markup income
Da — consumption spending from agricultural income
D„ — consumption spending from nonagricultural income
Da — floor-level consumption from agricultural income
D„ — floor-level consumption from nonagricultural income
AS, — change in private stocks of sector 1 products
CPI — consumer price index for urban consumption
y’ind — indirect taxes
GREV — government revenue
GEXP — government spending
DEE — trade deficit
III. Exogenous Variables
GP\ — procurement of cereals at prespecified price P*

GS> — government demand for services on development budget
G, — government demand for commodities, / = 1 , . . . . 5
A5* — base-year change in stocks in sector 1
AS/ — change in sectoral stocks, i = 2 , .... 5
E, — sectoral exports, / = 1.5
M/ — competitive imports, / = 1 ..... 5
x{ — domestic production of fertilizer
U — investment demand for sector 3 products
10 — investment demand for imports
n — border price of I0
pr — procurement price
p; — subsidized price of fertilizer from sector 3
e — exchange rate
R — remittances
Lg — government employment
TR — transfer payments
P* — border prices of intermediate imports, / = 1.5
t0i — subsidy rates on intermediate imports, f = 1 , . . . , 5
sub 3 — subsidy rate for manufacturing industries
t4 — markup rate in sector 4
ti — indirect tax rates, /' = 3,4,5
DEP — total depreciation
<#> — share of depreciation for government capital
GRi — government profits from enterprises, i = 3,4,5
tz — tax rate on profit income
FA | — food distribution under rural food-for-work program
FN i — food distribution in ration shops, “issue”
Pi — issue price
C0 — consumption of imported goods
P*c — border price of consumption imports
P*f — border price of fertilizer imports
P* — border price of sectoral exports, / = 1.5
A", — capital stock in sector i, = 3 and 5
GINT — government interest payments
Xj — fixed output levels, / = 1,2
IV Parameters
ay — sectoral input-output coefficients, i, j = 1 , . . . , 5

a0/ — input coefficients for intermediate imports, / = 1 , . . . , 5
— marginal propensity to consume from agricultural income, / = 1 , 5
m" — marginal propensity to consume from nonagricultural income, /' = 1 , 5
©; — floor-level consumption from agricultural income, / = 1 , . . . , 5
0" — floor-level consumption from nonagricultural income, / = 1 , 5
bj — sectoral labor-output ratios, / = 3,4,5
sa — savings ratio for agricultural income
sw — savings ratio for wage income
sz — savings ratio for profit income
a, — weights in the consumer price index, i = 1,5
kTj — constants in the supply-response functions, i = 3 and 5
<Tj — elasticity in the supply-response function, / = 3 and 5
km — wage-indexation coefficient
crs — elasticity of stock speculation
V. Initial Values
w° — beginning-of-period wage level, / = 3,4,5 and g

CPF — beginning-of-period consumer price index
P] — beginning-of-period price level of sector 1
Wages in the nonagricultural commodity-producing sectors and for gov¬

ernment employees are assumed to be partially indexed to the cost of living
within the model’s solution period of one year. Equations (4.8)-(4.12) give
the details of how wages respond to changes in the consumer price index
CPI. Wage income Yw in (4.13) comes from production activities, govern¬
ment employment, emigrant remittances (eR), and transfer payments from
the government (TR ).
Block IV gives price equations for the nonagricultural sectors, with prices
determined by a markup over variable costs. In manufacturing and services
(sectors three and five) the markup rates r3 and r5 are assumed in equations
(4.14) and (4.15) to rise with levels of output, again to capture a potentially
inflationary phenomenon emphasized by Indian economists. Prices them¬
selves are given in the simultaneous system (4.16)—(4.18), which includes
terms for sales taxes at rate /, for sector i, as well as subsidies granted
manufacturing production (sub)) and intermediate imports (at rates t0l).
With the prices computed, it is an easy matter to give final expressions
for variable costs per unit of output in block V. The level of income from
profits then follows in block VI, as the sum of markups over variable cost
less prespecified government revenues GR, from public enterprise produc¬
tion. Depreciation on capital stock is subtracted from markup incomes and
government interest payments are added to give the final value of Yr
In block VII appear equations for levels of consumer spending from
agricultural and nonagricultural incomes (Da and Dn, respectively). Agri¬
cultural saving at rate sa is subtracted from income Ya to give D„ in (4.23).
Along with saving parameters, income taxes at rate t, on profit income
Yz also enter the determination of D„ in (4.24). A term for a prespecified
(small or zero) level of noncompetitive consumption imports is also sub¬
tracted.
Block VIII specifies sectoral consumption levels according to the widely
used complete set of demand equations known as the linear expenditure
system, or LES.7 There are two sets of parameters for the LES—base levels
of commodity consumption (0“ and ©?, / = 1.5), which are assumed
to be independent of income and prices, and marginal propensities to con¬
sume (ml and m", i = 1, . . . , 5) from incomes above the levels Z)„ and
A t “
D„ required to pay for the base-level purchases. Both price and income
responsiveness of consumption are summarized by these parameters.
In the Indian context, further complications come from the facts that
agricultural families receive food-for-work cereal donations at zero price in
quantity FA ,, while nonagricultural families can get an "issue” of grain
(total quantity FN,) at a subsidized price P, in the largely urban ration
shops. Working through the utility maximization that underlies the LES
when these food subsidies are included in budget constraints gives the
demand equations that appear as (4.27)-(4.36). Adding the subsidy quanti¬
ties gives sectoral consumption levels in (4.37)-(4.41). Finally, note back in
equation (4.1) that the net change in the government’s food stocks is given
by its procurement GP, less food-for-work distribution FA , and the ration
issue FNX.
So far, we have described 41 equations in 43 variables (that is, all the
endogenous variables in section II of table 4.3 except T'nd, GREV, GEXP,
and DEF, which are described below). The two equations of block IX that
specify that demand less the fixed supply for each agricultural commodity
must equal zero serve to close the system. Thinking in terms of excess
demand functions shows how it can be solved. An algorithm contains the
following steps:
1. Guess sectoral output levels X\.X5 and a trial pair of prices P ^ and P2 for
the agricultural sectors.
2. Calculate nonagricultural markup rates and prices from the equations in block
IV (incorporating the feedback from price levels to wage rates in equations
(4.8)-(4.12) when it is assumed to apply).
3. Using the trial values of prices and outputs, calculate wages and wage and
agricultural incomes in block III, variable costs per unit output in block V, and
profit income in block VI.
4. Calculate sectoral levels of consumer demand from blocks VII and VIII, and the
foodstock change from block II.
5. Add up totals of sectoral demand, using the expressions to the left of the equality
signs in equations (4.1)-(4.5). If any sectoral demand level differs from its corre¬
sponding trial value by more than a prespecified tolerance, set the trial value equal
to the demand level and go back to step 2. Continue until demand levels are very
close to trial output values for all sectors.8
6. Check if output levels X i and X2 are very close to the prespecified agricultural
supply levels Xi and X2. If not, modify the prices P, and P2 and return to step 1.
Continue until excess demands for the agricultural products are effectively equal
to zero.9
When this solution procedure is complete, all sectors are in demand-

supply equilibrium and as a consequence saving must equal investment.
This derived relationship appears as equation (4.48) of table 4.2. To get to
this expression for macro balance, government and foreign trade accounts
are defined in (4.44)-(4.47). These equations summarize the fiscal and
balance of payments effects of all the tax and subsidy programs described
above.
For the government, (4.45) gives its current revenue GREV, made up of
direct tax receipts, indirect taxes [calculated in (4.44)], and revenues from
enterprise. The first two terms in equation (4.46) for current expenditure
GEXP are costs of food-for-work and ration shop food subsidies; the cost
differential associated with grain procurement, (P, — P*) GPU follows.
The next terms represent purchases of sector outputs from the development
budget (P GS ) and current outlay accounts; then come two terms captur¬
5 5
ing the effects of the fertilizer subsidy. The following five terms, respec¬
tively, represent wage payments to functionaries, transfers, depreciation on
government capital stock, government interest payments, and the subsidy
to manufacturing industry.
The next term is the sum of export subsidies in the five sectors and merits
some commentary. The law of one price is not imposed in the model; as a
consequence domestic and border prices of commodity exports may differ.
The price differential creates an income gain or loss to be absorbed by
someone in the system. In practice, both firms and the government will
share the benefit or burden, but in the Indian institutional context where
the government actively subsidizes exports, it is appropriate in a simplified
model to give it the full role. The penultimate term in (4.46) shows that the
government pays a variable export subsidy to firms, so that they can sell at
the same price both internally and externally. This outlay, of course, repre¬
sents government dissaving, and plays a role in the model’s overall adjust¬
ment to equilibrium. (A corresponding entry, which should appear for
variable tariffs on competitive imports, is omitted since its magnitude is
likely to be small.) The final term shows that the government also subsidizes
intermediate imports, at fixed sectoral rates t0r
Equation (4.47) defines the trade deficit in rupee prices. Because taxes or
subsidies on competitive imports are not included in the specification, they
enter the definition of the deficit DEF at internal prices (first term to right
of the equality sign). The other terms are at border prices multiplied by the
exchange rate e, and represent items of foreign commerce already de¬
scribed.
The saving-investment balance (4.48) is of standard form. Saving to the
left of the equals sign comes from households, the government current
account, and the trade deficit. Investment on the right includes capital
formation and stock changes. The other item is the government’s net stock
acquisition GP\ — FA 1 — FN,, valued at the market price Px.
As in all the models of this book, the saving-investment balance is not
an independent restriction; for that reason it is equation number 48 in a
system of 47 endogenous variables. However, it provides a convenient
numerical check on the computer solution of a model like the one here. In
solving the India model, tolerance limits were set to make saving equal to
investment to five or six significant figures. This degree of precision assures
that the algorithm to solve the model will generate a balanced SAM; it is
well beyond the accuracy of the data at hand.
4.3 Parameterization
The main reason for setting up a model based on a social accounting

matrix is that it can be used to address questions of income distribution and
intersectoral interaction that are aggregated away in one- or two-sector
specifications. A corresponding drawback is that time series information
can be used to estimate only a few relationships in the model, since the
national accounts do not appear in the requisite detail. Indeed, as described
above, even the construction of a SAM is usually based on a process of
imposing accounting identities on disparate and inconsistent data sources.
A social accounting matrix will scarcely ever be a correct description of the
macro economy, but at least it is a consistent one.10
If one takes consistency and accounting identities as central to the exer¬
cise, then it is clear that model equations must be parameterized in such a
way as to satisfy the SAM. The main parameters are listed in part IV of
table 4.3. Observations on their calculations follow:
1. The input-output coefficients ay and a0J follow directly by division of intersec¬

toral flows by gross outputs of the sectors. The main complication in figuring out
these “physical” coefficients is to make sure that taxes and subsidies are taken
out of the valuation of commodity flows in their numerators and denominators,
but getting the details straight essentially requires only tenacity and respect for
the accounting identities.
2. Parameters of the LES are not independent. Like all demand systems based on
additive utility functions, it has only N + 1 independent parameters for an
N-sector demand structure. Recipes are available for estimating the complete
parameter set based on (for example) income elasticities and minimal information
regarding price response.11
3. Saving rates must be computed to satisfy the overall saving-investment balance
of the SAM. In practice, both saving rates and levels for the different income
recipient classes are calculated together, incorporating whatever econometric
information happens to be available about saving behavior.
4. Labor-output ratios bj come from the wage bill and the wage rate by sector. In
a dynamic model, one would have to worry about rates at which the bt decline
from productivity change. Note also that in realistic country data, wage and profit
rates will always differ across sectors; an a priori assumption that they should be
equalized by competition does not fit the facts.
5. The other parameters in the table were estimated more or less ad hoc for the India
model, again subject to sensitivity analysis and the results of formal statistical
estimation.
6. The initial values of the variables of the model either can be read directly from
the SAM or else follow from simple manipulations (like dividing sectoral gross
output levels by total variable costs to get markup rates). Again, accounting
consistency permits straightforward computation of the long list of symbols in
table 4.3.
Key parameters from these manipulations appear in table 4.4. Saving

propensities from the various income flows differ but are all in the range
from 0.12 to 0.23. Income elasticities for consumer demand are plausible,
and there is substantial price-responsiveness in the consumption functions.
Note the value of 0.3081 for mau the marginal propensity to buy cereal
products from total consumer spending generated from agricultural in-
TABLE 4.4
Main Parameters of the Model
Uses of Agricultural Income

s„ = 0.12
Income Own Price
Sector m* 0* Elas. Elas.
1 0.3081 1.36314 0.6162 -0.5213

2 0.13239 0.18677 1.1654 -0.6380
3 0.35388 0.14801 1.6497 -0.8869
4 0.01743 0.07376 0.6354 -0.3296
5 0.1882 0.19846 1.3027 -0.7170
Uses of Nonagricultural Income

sw = 0.1482 s, = 0.2223 t, = 010936
Income Own Price
Sector m? ©? Elas. Elas.
1 0.09665 0.69737 0.7548 -0.4409

2 0.14771 0.25756 1.2207 -0.6731
3 0.33929 0.5521 1.2528 -0.7572
4 0.02103 0.09038 0.7816 -0.4074
5 0.39532 1.20443 0.9488 -0.6850
Sectoral Markup and Indirect Tax Rates

Markup Indirect
Sector Rate Tax Rate
3 0.1466 0.1243
4 0.3066 0.10157
5 0.50158 0.06987
comes. The parameter shows that about 30 percent of agriculturalists’ extra

spending goes for consumption of products of sector one. This economically
reasonable assumption strongly influences behavior of the model, as dis¬
cussed in the following section.
4.4 The Jacobian
The basic equations are (4.1)—(4.5), which can easily be rewritten as

excess demand functions for the five sectors. When excess demands are not
zero, equilibrium is reestablished via price increases in the first two sectors
and output increases in the other three. For the variables Pu P2, Xif X4,
and X$, three adjustment matrices are shown in table 4.5: the Jacobian for
the five-equation excess demand system, a matrix of elasticities of excess
TABLE 4.5
Response Matrices for the India Model
Jacobian Matrix for Excess Demand Functions
Pi Pi x3 X4 x5
Pi -0.561 0.500 0.026 0.039 0.053

Pi 0.091 -0.481 0.111 0.059 0.075
X3 0.217 0.190 -0.658 0.254 0.296
x< 0.010 0.020 0.037 -0.796 0.032
X, 0.012 0.128 0.264 0.242 -0.662
Column
total -0.231 0.357 -0.220 -0.202 -0.206
Elasticities of Excess Demand Function

with Respect to Investment
-0.242 0.216 0.107 0.017 0.149

0.039 -0.207 0.458 0.026 0.212
0.094 0.082 -2.716 0.112 0.835
0.004 0.009 0.153 -0.351 0.090
0.005 0.055 1.090 0.107 -1.867
Inverse of the Jacobian Matrix (x — 1)
3.133 4.317 1.534 1.419 1.497

1.267 4.309 1.347 1.193 1.252
2.010 4.052 3.282 2.114 2.193
0.212 0.457 0.273 1.467 0.261
1.179 2.693 1.697 1.636 2.750
demand with respect for goods from sector three (industry and construc¬
tion), and the inverse of the Jacobian.12
An increase in any one of the five adjusting variables will reduce excess
demand in its corresponding sector—that is, the meaning of the negative
entries along the main diagonal of the Jacobian. Note, however, that there
are strong feedbacks from some sectors to others. The most important
example is the coefficient in row one, column two. Its value of 0.5 comes
directly from the parameter value of 0.3081 for the marginal propensity to
consume m\ pointed out in connection with table 4.4 above. The linkages
go as follows:
1. An increase in P2 raises agricultural income proportionately to sector two output

of 2.6546 (the entry in row 14, column 2 of table 4.1 with quantities measured
in ten thousand crores, or one hundred billion rupees in 1980-81).
2. The direct effect on sector one consumer demand is given by 2.6546 (1 - j()ra“
= 2.6446(0.88)(0.3081) = 0.717.
3. Subtracting intermediate input costs (a 12 = 0.0174 in the input-output matrix),
substitution effects in consumption and other minor feedbacks gives the coeffi¬
cient of 0.5 appearing in table 4.5. Despite these emendations, there remains a
strong income effect from increases in the noncereal price P 2 on demand for
products of sector one. Through this linkage, an increase in the price of a major,
relatively stable production flow has macroeconomic repercussions throughout
the system.
Along similar lines, there are strong demands from all other sectors for
sector three, and from sectors three and four for sector five. Perhaps surpris¬
ingly, not much additional demand for infrastructure services from sector
four is generated by price or quantity increments in the other sectors. The
Jacobian replicates the finding mentioned earlier that output in the infra¬
structure sector is relatively independent of activity in the rest of the econ¬
omy.
In algebraic terms one can write the response of the endogenous variables
to a change in investment demand for sector three products as follows:
1
_1
j\\ jn ' 0 ‘ ’o'
jn dp1 0 0
dX 3 + dh — 0
dXA 0 0
_1
3
y 51 0 0
1
where they,* are the elements of the Jacobian. From this formulation it is
evident that the column sums of the Jacobian (multiplied by minus one)
provide a differential form of the saving-investment balance in the model,
that is
~ On + ju + • • • 4- ji\)dPi — (y,2 + j22 + • • • -F jn)dP2

— • • • ~ O’is + 7 25 + • • • + jss)dX5 = dli,
The numerical matrix in table 4.5 shows that a unit increase in P, generates
0.231 units of saving, and so on.13 All saving responses are positive except
for P2 ■ The increase in demand for food and other products that a rise in
this price generates actually reduces saving overall. Such an effect could
potentially make the excess demand functions unstable, but it is not strong
enough to do so in the present model.14
The elasticity matrix shows how prices and quantities respond in relative
terms to an increase in investment demand. Since the rise in investment first
affects sector three, its own elasticity of 2.716 is the largest in absolute terms
in the matrix, as might be expected. The elasticities for the prices are in
general less than for the quantities, basically because consumers’ price
elasticities of demand for food are fairly small, as shown in table 4.4. The
consequence is that wide swings in prices as equilibrium shifts may be
expected. This observation is verified by the inverse of the Jacobian (with
all entries multiplied by minus one) in table 4.5. It shows that all the
endogenous variables will respond positively to an increase in any sector’s
demand. The elements in the first two rows giving price responses are large,
especially in comparison to the initial normalization of both prices at unity.
The direct and indirect elasticities of the endogenous variables to an in¬
crease in sector three investment demand (calculated from the third column
of the inverse Jacobian matrix) are: Pu 3.557; P2, 3.123; X3, 0.819; X4, 0.62;
X5, 0.602. The fivefold difference in magnitude between price and quantity
elasticities comes from the economics of the system. The price-clearing
agricultural sector plays a major role in the Indian economy, and is charac¬
terized by inelastic supply and low elasticities of demand. Hypersensitive
prices are the structural outcome.
4.5 Numerical Results
In this section, trial variations in parameters and data for the India model
are described, with emphasis on its adjustment mechanisms. We first take
up how the system responds to changes in investment demand, coupled with
foodstock speculation, partial wage indexation, rising markup rates, and
variations in the level of agricultural supply. Next comes an analysis of the
possible effects of devaluation, followed by a discussion of how the govern¬
ment might adjust its food procurement policy to compensate for output
fluctuations in sector one. Results from the model, including a demand-
supply balance for money, are given in section 5.5.
Table 4.6 shows how the system responds to an increase in investment
demand for products of sector three from 2.3186 to 2.41 hundred billion
rupees (about 4 percent). The rise in aggregate demand is reflected in
increases of both agricultural prices and nonagricultural levels of output,
as shown in the second column of panel A of the table. The arc elasticities
of response are close to those calculated from the Jacobian in table 4.5, and
the output multipiers are^3, 3.335; X4, 0.281; X5, 1.743. These values are
consistent with the usual Keynesian “leakages” from the income-demand
linkage through saving and intermediate imports, coupled with the feed¬
back of increases in agricultural prices to demand for nonagricultural pro¬
ducts.
Panel B of table 4.6 shows how saving flows adjust to meet higher
investment demand. The value of investment rises from 2.7417 to 2.928
hundred billion rupees. This increase reflects both the greater quantity of
capital goods demanded and the rise in the sector three price (due to
cost-push of intermediate inputs purchased from agriculture) from 1.0 to
1.0318. As shares of investment we have the following changes in sources
of saving:
Saving Sources Initial Final
Agricultural incomes 0.196 0.213

Wage incomes 0.234 0.224
Profits 0.194 0.192
Trade deficit 0.066 0.072
Government saving 0.049 0.055
Depreciation 0.261 0.244
There is a forced saving in the sense that real wages in nonagricultural

sectors decline as prices go up, causing the proportion of saving from wages
to the total to fall. The real value of nominal depreciation allowances goes
down for the same reason. The share of saving coming from nonagricultural
TABLE 4.6
Effects of a Four-Percent Increase in Investment Demand for Sector Three
Products
A. Prices and Quantities

Investment Increase Coupled With
Only Foodstock Rising
Base Investment Price Wage Markup
Run Increase Responsiveness Indexation Rates
Pi 1 1.1484 1.1917 1.2182 1.1988

Pi 1 1.1325 1.1527 1.1916 1.1748
Pi 1 1.0318 1.0372 1.0597 1.0666
p* 1 1.0098 1.0015 1.0336 1.0229
P5 1 1.0138 1.0164 1.0388 1.0435
9.5712 9.8760 9.9012 9.9299 9.8833
*4 1.0216 1.0473 1.0501 1.0530 1.0505
*5 6.5387 6.6980 6.7128 6.7350 6.7052
B. Savings and Investment

Agr. sav. 0.5373 0.6237 0.6437 0.6626 0.6511

Wage sav. 0.6414 0.6546 0.6558 0.6702 0.6552
Prof. sav. 0.5324 0.5635 0.5675 0.5850 0.6094
Trade def. 0.1800 0.2108 0.2142 0.2246 0.2243
Gov. sav. 0.1357 0.1606 0.1607 0.1516 0.1720
Deprec. 0.7148 0.7148 0.7148 0.7148 0.7148
Total 2.7417 2.9280 2.9567 3.0087 3.0268
C. Income Shares
Agr. income 0.3895 0.4171 0.4239 0.4248 0.4198

Wage income 0.3765 0.3545 0.3497 0.3479 0.3420
Profit income 0.2340 0.2384 0.2264 0.2273 0.2382
Total 1.0000 1.0000 1.0000 1.0000 1.0000
profits stays about constant, while the output increases in the nonagricul-
tural sectors draw in more intermediate imports and drive up the trade
deficit’s share.
Both price and quantity changes enter into the adjustment. In distribu¬
tional terms, the outcome in panel C of table 4.6 is an increase in agricultur-
alists’ share of income, and a fall of that of wage earners. In relative terms,
profit incomes hold roughly stable.
These patterns change when additional price and quantity reactions are
introduced into the system, as shown in the other columns of the table. The
columns headed “Foodstock price responsiveness” show what happens
when the speculation parameter crs in equation (4.6) is changed from zero
to one. There are sharper price increases, and some output response in the
nonagricultural sectors. More forced saving occurs at the expense of work¬
ers, and the income distribution shifts further toward agriculture.
In practice, partial indexation of urban wages in India to price increases
within the year appears to be the rule. The columns headed “Wage indexa¬
tion” show what happens when 20 percent of the increase in the consumer
price index as calculated in (4.8) is passed along to nominal wages in
(4.9)-(4.12). Not surprisingly, the effect is inflationary. Moreover, workers
lose from the indexation, as their income share falls from 0.3545 when
investment increases without indexation, to 0.3479. The reason is that food
supplies are assumed fixed, so that an initial increase in demand resulting
from higher money wages has to be cut back by still sharper food price
increases. The outcome is an additional shift of the income distribution
toward agriculture and a lower real wage. Profit recipients gain from bigger
markups on higher nominal variable costs.
The columns headed “Rising markup rates” summarize the effects of
raising the parameters <x3 and <x5 from zero to 2.0 and 1.5, respectively, in
equations (4.14) and (4.15). With these changes in specification, markup
rates rise along with the output-capital ratios in sectors three and five,
essentially reflecting demand-pressure on prices.15 Once again, price in¬
creases go beyond those induced by a simple increase in investment demand.
In panel C of table 4.6, it can be seen that the distributional effect favors
agriculturalists and harms wage earners; the profit share as usual holds
steady.
Each of the three shifts in the specification—foodstock speculation, wage
indexing, and markups rising in response to demand—is mildly inflationary
taken by itself, adding 5 percentage points or so to the average 14 percent
rise in agricultural prices caused by higher investment demand. However,
when the three effects are put together, the feedbacks between them are
extreme.
The agricultural prices rise by 64 and 49 percent, as shown in the “Infla¬
tionary effects” column in panel A of table 4.7. There are also substantial
increases in sectoral outputs, and the income distribution swings 7 percent¬
age points toward agriculture.
TABLE 4.7
Effects of an Investment Increase Linked with All
Inflationary Mechanisms and Either Increased Imports of
Sector Two Commodities or Supply Response in Sectors
One and Two

Agric.
Base Inflationary Increased Supply
Run Effects Imports Response
P\ 1 1.6396 1.1853 1.0998

Pi 1 1.4856 1.1342 1.0850
Pi 1 1.2133 1.0689 1.0511
P* 1 1.1184 1.0377 1.0270
X3 9.5712 10.1302 9.8628 9.8551
X* 1.0216 1.0798 1.0473 1.0466
X5 6.5387 6.8687 6.6906 6.6955

Agric.
Agr. sav. 0.5373 0.8826 0.6353 0.6006

Wage sav. 0.6414 0.7083 0.6660 0.6617
Prof. sav. 0.5324 0.7456 0.6123 0.6026
Trade def. 0.1800 0.2954 0.2464 0.2234
Gov. sav. 0.1357 0.1596 0.1688 0.1710
Deprec. 0.7148 0.7148 0.7148 0.7148
Total: 2.7417 3.5063 3.0435 2.9742
C. Income Shares
Agric.
Agr. income 0.3395 0.4626 0.4110 0.4000

Wage income 0.3765 0.3006 0.3489 0.3563
Profit income 0.2340 0.2368 0.2401 0.2432
Total: 1.0000 1.0000 1.0000 1.0000
These shifts are too large to be credible, but do reflect the model’s (and
to a degree the Indian economy’s) sensitivity to the agricultural sectors.
However, they are easily moderated by supply changes. In table 4.7 two
options are presented. First, “Increased imports” columns show the effects
of raising competitive imports into sector two from 0.0035 to 0.0235 hun¬
dred billion rupees (about 0.75 percent of initial supply). Second, the “Agri-
cultural supply response” columns show what happens when output during
the year in each of the agricultural sectors rises in the response to higher
prices. The supply elasticities were set at values 0.1 and 0.2 in sectors one
and two, respectively, on the notion that there is likely to be somewhat less
market sensitivity for the staple commodities produced by cereal growers.
For short-run elasticities these values are perhaps on the high side, but were
chosen to reflect as much stabilization from supply response as might
reasonably be expected.
In both cases, the price increases are greatly moderated, as one could
already anticipate from the large entries in the northwest comer of the
inverse Jacobian matrix of table 4.5.16 Both import manipulation and farm¬
ers’ price-responsiveness (as well as other policies such as changes in pub¬
licly held foodgrain stocks) have very important macroeconomic effects in
the model. This observation also applies to the real economy of India.
Turning to another topic, we examine the effects of a 15-percent devalua¬
tion of the rupee in table 4.8. Three simulations are given, with elasticities
of export volumes from all sectors with respect to the exchange rate as¬
sumed to be zero, 0.5, and 2.0, respectively.
When there is no export response, the model follows the contractionary
devaluation scenario of chapter 2. Devaluation drives up the costs of inter¬
mediate imports, which are directly transmitted to higher prices of nona-
gricultural products. Real wages fall, pulling down aggregate demand. The
outcome is reduced production in the quantity-adjusting sectors, and lower
agricultural prices. There is an overall inflationary push, despite the reduc¬
tions in F, and P2.
Increased nominal values of exports and remittances and the reduced
volume of intermediate imports dominate higher import costs, so that the
trade deficit in rupee terms declines (panel C).17 At the same time, the
nominal value of investment demand rises along with nonagricultural
prices. The required saving comes mainly from the government (panel B)
as its nominal export subsidies fall. The net increase in saving from govern¬
ment and the trade deficit is (0.2247 — 0.1357) + (0.1539 — 0.1800) =
0.0629. As discussed in section 5.5, where money is brought into the present
model, the outcome is a reduction in money creation by 0.0629 as well. At
a time when devaluation is driving up nonagricultural prices, money supply
drops with a contractionary effect on output and growth. Finally, note that
devaluation shifts the income distribution slightly away from agriculture
(panel D) as the sector’s prices fall.
When exports respond to the exchange rate, the contractionary effects
vanish. Already an exchange rate elasticity of 0.5 induces enough export
TABLE 4.8
Effects of a Devaluation by Fifteen Percent
Export Export
Base No Export Response Response
Run Response Elast = 0.5 Elast = 2.0
P\ 1 0.9994 1.1245 1.2052

Pi 1 0.9994 1.1014 1.1593
P3 1 1.0189 1.0629 1.0922
P4 1 1.0061 1.0303 1.0465
PS 1 1.0052 1.0394 1.0632
9.5712 9.5283 9.6598 9.7748
*4 1.0262 1.0190 1.0350 1.0520
*5 6.5749 6.5246 6.6034 6.6658

Export Export
Agr. sav. 0.5373 0.5364 0.6060 0.6485

Wage sav. 0.6414 0.6460 0.6603 0.6709
Prof. sav. 0.5325 0.5344 0.5780 0.6120
Trade def. 0.1800 0.1539 0.1516 0.1392
Gov. sav. 0.1357 0.2247 0.2300 0.2431
Deprec. 0.7148 0.7148 0.7148 0.7148
Total: 2.7417 2.8102 2.9407 3.0285
Dom. inv. 2.3186 2.3625 2.4644 2.5325

Imp. inv. 0.1211 0.1393 0.1393 0.1393
Pr. stocks 0.4048 0.4113 0.4399 0.4596
Gov. stocks -0.1028 -0.1028 -0.1028 -0.1028
Total: 2.7417 2.8102 2.9408 3.0285
C. Trade Balance
Export Export
Compet. imp. 0.3782 0.3853 0.4021 0.4133

Imp. inv. 0.1211 0.1393 0.1393 0.1393
Imp. fertilizer 0.0588 0.0672 0.0677 0.0677
Imp. intermed. 0.5719 0.6547 0.6638 0.6717
Total: 1.1300 1.2464 1.2728 1.2919
Total exports 0.7100 0.8165 0.8452 0.8767

Remittances 0.2400 0.2760 0.2760 0.2760
Trade def. 0.1800 0.1539 0.1516 0.1392
D. Income Shares
Export Export
Agr. income 0.3895 0.3878 0.4065 0.4150

Wage income 0.3765 0.3781 0.3586 0.3476
Profit income 0.2340 0.2341 0.2350 0.2374
Total: 1.0000 1.0000 1.0000 1.0000

demand to produce expansion in the nonagricultural sectors. However,

there is also demand pressure on agriculture, which as usual must be met
by rising prices. With an export elasticity of 2.0, this inflationary effect is
quite strong. The rise of the agricultural income share by 2.5 percentage
points is part and parcel of the food price inflation.
Effects of changes in the level of foodgrain supply from sector one are
summarized in table 4.9. Note the price decreases that accompany higher
cereal production. As discussed in sections 3.1 and 3.2, a fall in the food-
TABLE 4.9
Food Procurement Policy

Output Output
Base 3.6 % Up + Procur. 6.6% Up 4- Procur.
Run Sector I Increase Sector 1 Increase
P\ 1 0.8309 0.99727 0.7303 0.9912

Pi 1 0.9769 1.0846 0.9654 1.1539
Pi 1 0.9821 1.0298 0.9721 1.0541
P4 1 0.9864 1.0138 0.9786 1.0249
Pi 1 0.9844 1.0225 0.9756 1.0407
X3 9.5712 9.5468 9.6726 9.5316 9.7524
Xt 1.0216 1.0201 1.0343 1.0193 1.0445
X5 6.5387 6.5446 6.6249 6.5494 6.6943

Output Output
Base 3.6% Up 4- Procur. 6.6% Up 4- Procur.
Run Sector 1 Increase Sector 1 Increase
Agr. sav. 0.5373 0.4821 0.5690 0.4494 0.5941

Wage sav. 0.6414 0.6347 0.6503 0.6308 0.6575
Prof. sav. 0.5324 0.5209 0.5666 0.5145 0.5948
Trade def. 0.1800 0.1828 0.2084 0.1873 0.2320
Gov. sav. 0.1357 0.1432 0.1439 0.1471 0.1511
Deprec. 0.7148 0.7148 0.7148 0.7148 0.7148
Total 2.7417 2.6785 2.8529 2.6439 2.9442
Total inv. 2.7417 2.6785 2.8529 2.6439 2.9442
C. Income Shares
Output Output
Base 3.6% Up 4- Procur. 6.6% Up 4- Procur.
Run Sector I Increase Sector 1 Increase
Agr. income 0.3895 0.3675 0.3954 0.3533 0.3995

Wage income 0.3765 0.3918 0.3659 0.4015 0.3577
Profit income 0.2340 0.2407 0.2387 0.2452 0.2428
Total: 1.0000 1.0000 1.0000 1.0000 1.0000

grain price resulting from higher output can either stimulate or cut back
on demand for nonagricultural products. Farmers’ income drops as their
output price declines sharply in the face of increased supply, while the real
purchasing power of incomes generated in nonagricultural sectors goes up.
Demand for nonagricultural products will rise or fall, depending on how
these opposing income changes balance out.
In the India model, falling cereal prices slightly retard demand for sector
three products (industry) and stimulate it for sector five products (services).
At the same time, the income distribution swings against agriculture and
in favor of wages and profits. All prices fall as foodgrain output rises, due
to consumer demand and input-output linkages in the system.
The policy variable that the Indian government uses to stabilize the
economy against food output shifts is procurement. The government has to
honor the procurement price, which, with its initial value of one, lies above
the market price in both cases of output increase shown in table 4.9. The
“procurement increase” columns show what happens when the government
buys enough grain to hold the market price steady at its initial (normalized)
value of one. There is overall expansion, and government saving increases
in nominal terms despite its outlays to buy grain.
Varying procurement to stabilize prices makes GP\ almost an endoge¬
nous policy variable. When procurement is treated in this fashion, it is
natural to ask about trade-offs between the amount of government pur¬
chases required to hold market food prices constant and the level of produc¬
tion. Figure 4.2 depicts the iso-price relationship between GP] and output
Xt. The graph is essentially linear, and shows that a procurement increase
of about three billion rupees’ worth of grain will counteract a production
increase of ten billion. Procurement appears to be an effective price stabili¬
zation instrument in India.18
FIGURE 4.2
Levels of Grain Production and Procurement That Maintain the Market Price at
Its Initial Level
4.6 Conclusions
The foregoing exercise shows in a near-practical context how price and

quantity adjustments enter a macro model constructed along lines sug¬
gested in the previous two chapters. In most cases, the reactions of the
adjusting variables to shifts in parameters or endogenous variables have
“correct” signs, but some may be too large in magnitude to make sense.
Since the model is aggregated, it leaves out enormous detail regarding the
Indian economy—demand-supply adjustment mechanisms for specific agri¬
cultural crops, regional variation, time lags achieving sectoral balance—a
host of complications.
To point out the major features of adjustment, it makes sense to leave
these complexities out. In applied policy analysis, it would be better to
incorporate them insofar as possible, or change the specification to smooth
them away. One obvious trick is to make agricultural supply dependent on
the current period’s price level, as in table 4.7. Equally, saving rates could
be made dependent on income levels, or real investment might be assumed
to fall if prices go up (on the hypothesis that investment outlays are planned
in nominal, not real, terms).19 Some of these changes would make the
specification more neoclassical, as discussed in section 2.3; others would
depend on policy judgment. But all of them amount to ways of sweeping
innumerable microeconomic feedbacks to the macro system into a tidy pile;
they add little to our understanding of how the system operates in the first
place.
5
Money and Other

Assets in the Short Run
T HE DISCUSSION now shifts toward the finances of the economy,

under simplifying assumptions about the real side. All intersectoral and
trade balance considerations are dropped, to put emphasis on how the
general price level may respond to financial perturbations in a short run
during which the money wage, price expectations, and the capital stock and
capacity utilization are all fixed. In chapter 6, these restrictions are lifted
in a full monetary growth model.
The chief innovation in the discussion is inclusion of interest payments
on working capital in variable cost. This specification reflects the fact that
with poor articulation of both transport and financial systems in developing
countries, entrepreneurs typically have to make large advances for working
capital in the form of labor and intermediate goods; interest on these ad¬
vances looms large in their statements of profit and loss. A reduction in the
money supply drives up interest rates and can thus increase the price level
from the side of cost. At the same time, higher interest rates will reduce
investment demand. To see how these two effects can lead to stagflation—
simultaneously rising prices and falling levels of economic activity—we
begin the discussion with a simple short-run model in section 5.1. Since the
emphasis is on inflationary processes, full capacity use and a variable price
level are assumed, as in chapter 2.
Money and Other Assets in the Short Run 87
In section 5.2, the financial side of the economy is described. Households

hold three kinds of wealth: bank deposits, loans to firms, and a group of
nonproductive assets called “gold.” Excess demand functions for loans and
gold are derived, with the interest rate and gold price as market-clearing
variables. Demand and supply of bank deposits come into balance by Wal¬
ras’s law, although all three asset-market excess-demand functions may be
shifted by a change in an exogenously specified deposit interest rate.
Section 5.3 describes short-run equilibrium in a two-asset model without
gold. Depending on whether cost-push from working capital or demand
reduction from less investment is more important, a rise in the interest rate
can either drive up or reduce the price level. Comparative statics depend
on which effect dominates. When working-capital costs are more important,
it is shown that monetary contraction can lead to stagflation in the short
run. Similar problems arise in section 5.4, with all three assets. An increase
in the deposit rate can lead to stagflation unless it strongly shifts asset
preferences away from gold, as opposed to productive loans from the house¬
hold sector to firms.
The final section of the chapter is an empirical illustration of how stagfla¬
tion can be induced by monetary contraction in a version of the India model
of chapter 4 extended to include a money market. Monetary repercussions
of currency devaluation are also assessed.1
5.1 The Real Side of the Economy
Since the emphasis in this chapter is on price instead of output changes,

it is simplest to assume full utilization of capacity. Excess demand will
adjust toward zero by forced saving in the short run, as described in chapter
2. The markup rate is an endogenous macro variable, while output is
determined by available fixed capital,
X = aK (5.1)
where X is output, K is capital stock, and a is the output-capital ratio. An

alternative, somewhat richer model would result if (5.1) were replaced by
a rule such as r = r (X/aK). As discussed in section 2.3, an increase in
aggregate demand would then be associated with forced saving from the
rising markup rate r as well as output expansion. Without algebraic detail,
how this more “neoclassical” model would behave is pointed out from time
to time in the following discussion.
An extension beyond chapter 2 is consideration of demand for working

capital by firms. The assumption is that entrepreneurs must have money on
hand to pay in advance for the services of current inputs into the production
process. In a poor, inflationary economy, suppliers of inputs are too finan¬
cially constrained to allow their payments to wait.2
There has been substantial debate in the literature as to whether working-
capital costs should be measured by the nominal interest rate i or a “real”
rate defined as / — it, where tt is the expected rate of price inflation. The
choice really hinges in what one considers to be the “short run.” Assume,
for example, that production takes place during a period of T days, during
which employment stays fixed at L. Workers have to be paid in advance,
so working-capital expenses are given by borrowing costs at time t, when
production is initiated, less the labor content of sales at time t + T,
Cost = L[w,( 1 + zT) — w(+r]
in which i is the daily interest rate, and w, and w,+T are wages at the
beginning and end of the period. If the money wage stays constant (w, + r
= wt), then evidently Cost = Lw,iT. Alternately, if the wage is perfectly
indexed to expected inflation and the nominal interest rate is equal to some
real rate j plus inflation (z = j + tt), then we have:
Cost = L[w,{ 1 + (/' + 7t)T) — w,(l + ttT)\ = LwjT
so the real interest rate would be the appropriate concept of cost.

In reality, of course, time profiles of input use vary widely across produc¬
tion processes—ships and houses tie up streams of raw materials and inter¬
mediates over extended periods, while crude oil feedstocks for refineries
may be in tankers or storage for a few weeks or even days. In the formal
model discussed here, continuous flow input of labor is assumed, while
excess macro demand is driven to zero during a period (implicitly of days
or weeks) during which the money wage does not change. On these assump¬
tions, working-capital costs should be measured by the nominal inter¬
est rate, so that the payment flows that exhaust the value of output are
given by:
PX = (1 + i)wL + rPK (5.2)
The income decomposition underlying this equation is threefold—work¬

ers get a wage bill wL, firms receive a return rPK on their fixed capital
(r is the profit rate, as usual), and banks get interest payments iwL on their
working capital advances for employment of labor.3 As just discussed, the

labor input flow is proportional to the flow of output,
L = bx = baK (5.3)
Substitution of (5.3) into (5.2) gives an expression for the profit rate as:
(1 -f 0 wb
r = a (5.4)
~~P
Evidently, r rises with the price level but falls with both the interest rate
and the money wage. Inverting (5.4) gives:
aw( 1 + i)b
(5.5)
(a - r)
so that the maximum profit rate is the output-capital ratio a, reached when
the real wage falls to zero.
Equilibrium excess demand for output must be zero:
C + I + G - X = 0
where C is household consumption, / is investment, and G is government

consumption demand. As usual, we assume that there is no saving from
wage income, and that the share of total profits rPK that is saved is a
constant, s. This parameter captures saving from both firms (retained earn¬
ings) and personal incomes (dividends). The same sort of decomposition
should apply to saving from interest income iwL, except that retained
earnings of banks and especially moneylenders may be small, while interest
recipients such as the traditional “widows and orphans” may not save much
either. To save a parenthesis or two under these circumstances, it is simplest
to assume that interest income is entirely consumed.4 The value of private
consumption thus becomes:
PC = (1 + i)wL + (1 - s)rPK
Substitution of this expression into the excess demand equation and ma¬
nipulations such as those in chapter 2 give the investment-saving balance as:
g -f ya — sr = 0 (5.6)
where g — I/K is the growth rate of capital stock (and output) and y =
G/X is the share of government demand in output.
Together with (5.4), equation (5.6) determines the equilibrium price level
P. To get a final expression for P, a theory of investment must be added
to the demand side. As before, returns to investment are measured by the
profit rate r. A new feature is the cost of tying up resources in physical
capital measured by the alternative uses to which they could be put. Since
capital formation is a long-term process, the real interest rate / — 7r is
the appropriate cost concept. The simplest formulation makes the growth
rate of capital depend on the difference between profit and real interest
rates, or5
g = go + h[r - (/ - 7r)] (5.7)
Solving (5.4) (5.6) and (5.7) together for P gives:
p _ _ (5 - h)aw( 1 + i)b _
(5.8)
[s - (h + y)]a - g0 + h(i - n)
Forced saving is built into (5.8) since an upward shift in autonomous

investment g0 drives up the price level relative to the money wage w. The
consequent distributional shift toward profits generates the saving counter¬
part of the higher investment demand. An increase in / has an ambiguous
effect on P. A rising interest rate may push up the price level by increasing
working-capital costs or reduce it by cutting back on investment demand.
To check the conditions under which one or the other effect dominates, one
can differentiate (5.8). The sign of dP/di depends on the quantity:
[s — (h + y)]o - g0 - h( 1 + tr)
For given values of the other parameters, this expression shows that work¬
ing-capital cost-push makes dP/di positive when the expected inflation rate
7T is relatively low. With more inflation, forced (dis) saving from reduced
investment demand may come into play strongly enough to make the
derivative negative. For future reference, note that simple manipulation
makes the condition (s — h)abw > Ph a requirement for working-capital
cost-push to dominate.6
Before going on to discuss money and finance, we can derive two addi¬
tional results from the saving-investment balance. First, P is a function of
i and tt from (5.8), and the profit rate r follows from (5.5). From (5.7),
investment demand thus depends only on i — it. Going through the algebra
shows that the growth rate is given by the expression:
sg o + hya — sh(i — n)
g = ---—^-- (5.9)
s
s — h
So long as s > h (the effect on saving of profit rate increases exceeds the
effect on investment—the standard stability condition), then an interest rate
increase unambiguously reduces the rate of growth.
The second result is simply a closed form expression for commodity
excess demand, which is:
(s - h)a{ 1 + i)wb
---h(t - tt)
+ go — [s — (y + h)]a = 0 (5.10)
Excess demand on the left side of the equality is a falling function of P. This
effect represents the forced saving adjustment mechanism. Discussion of
how excess demand responds to interest rate changes can safely be post¬
poned until after details of the financial market are set out.
5.2 Money and Other Assets
The ultimate holders of wealth in a country are its citizens. It is custom¬

ary in macro models to treat them as being homogeneous in their choices
regarding assets in which to store their patrimony, though, of course, rich
and poor people behave entirely differently in this regard. However, the rich
hold the preponderance of assets in capitalist economies. Data are hard to
come by, but in England and the United States even the top 1 percent of
wealth-holders own respectively 30 and 20 percent of financial assets.7
Corresponding figures cannot be much lower (and may well be higher) in
the Third World. Given such concentration in the control of wealth, assum¬
ing uniform behavior on the part of its owners may be an acceptable first
approximation.
A second question is how to describe wealth. In any functioning econ¬
omy, there are numerous formal financial instruments, informal debt and
credit relationships in both urban and rural areas, productive economic
assets such as physical capital, livestock and land, and unproductive ones
like antique weapons and gold that are held for sentiment or speculative
thrill. The usual models are based on the assumption that all assets must
be held by someone, with their returns adjusting accordingly. The “adding-
up” restriction that the value of assets should equal wealth puts strong
restrictions on portfolio choice, as we will see later on in this chapter.
Hence, selection of assets to be included in a formal model is not trivial,
even if one residual, such as the “bond” of traditional macro theory, is
included to absorb all the slack.
For the present, we will distinguish three sorts of assets held by persons.
The first is money, conventionally held in the form of deposits with banks.
The second takes the form of loans to firms, which in turn borrow from both
banks and the public to finance fixed and working capital needs. (Firms are
assumed not to sell equity, for the simple reason that stock markets are at
best a decoration in most poor lands). Finally, the public is assumed to hold
a congeries of assets such as currency, precious objects, and land and real
estate, which we call “gold.” This gold has its price fixed by simple market
clearing and has no value in use. This asset specification is set up to empha¬
size the differences between wealth held with banks, or the formal financial
system, and that held (perhaps with some sacrifice of liquidity) beyond the
bankers’ and tax collectors’ view. Of our three assets, deposits are fully
“monetized” or “intermediated” and gold is not. Loans to firms are some¬
where in the middle. Informal credit markets are widespread in most coun¬
tries, are often competitive and agile, and provide an area in which to hold
resources besides a mattress or a hole in the ground or the banks. They also
play an important macroeconomic role, as we will see shortly.
The balance sheets for the assumed financial system appear in table
5.1.8 At the bottom, the public holds its wealth in the form of bank deposits
Dp, loans to firms Lp, and gold. The fixed gold stock is Z. its price is
Pz and the value of gold holdings is P,Z. At the top of the table, the central
bank holds government fiscal debt F as its only asset. The corresponding
liability is commercial bank reserves H, which in the table’s simplified
accounting scheme are equivalent to the money base, or high-powered
money.
The commercial banks have as assets their reserves, plus loans or net
credit to the private sector. For simplicity, all credit is assumed to go to
firms (in volume Lh). Bank liabilities are deposits held by the public and
firms (respectively DP and Df). Firms receive loans both from banks and
the public, while their assets are their bank deposits and the value of their
fixed capital PK. After cancellation of all items entering on both sides of
the accounts, wealth W turns out to be equal to H + PK + P,Z, or the
sum of the money base and the values of capital and gold. Consolidation
of the bank’s account shows that H + Lb = Dp + Dfi or outstanding
credit to the government and firms just equals the money supply. These
TABLE 5.1
Financial Balance Sheets
Assets Liabilities
Central Bank
Government debt F Bank reserves H
Commercial Banks
Bank reserves H Deposits from firms Df

Loans to firms Lb Deposits from the public DP
Firms
Physical capital PK Loans from banks Lb

Loans for working
capital (= deposits) Df Loans from the public
Public
Deposits with banks DP

Loans to firms LP Wealth W
Gold PzZ
Wealth identity: W = H + PK + PzZ

Banking balance sheet: H + L b = Dp + Dj
wealth and banking system “identities” of course hold when everyone’s

portfolio adjustment has arrived at an equilibrium.
What about working capital? Firms require advances in the amount wL
— wabK to pay the wage bill. If workers’ own money holdings are small,
which seems likely, they will respend their advances on consumption. In
turn, firms will use these receipts to buy intermediate inputs and pay their
own workers. In practice, then, working capital will mostly be held by
enterprises and not the public. The simple hypothesis is that firm deposits
are equal to working capital, or:
Df = wbaK — H -\- Lb — DP
where the second equality follows from the consolidated balance sheet of
the banking system.
The next assumption is that banks don’t hold excess reserves and that
the banking authorities enforce a reserve requirement of the form H =
lx(Df + Dp), with ju, < 1. (We could alternately—and realistically—assume
that the authorities restrict bank credit Lb to a given level and get essentially
the same results.) Finally, assume that the public’s demand for deposits
takes the form DP — i|/(/', id, tr, ttz)W, where W is total wealth and the
arguments of the function v|> are discussed in more detail below. In effect,
the public is assumed to keep a fraction ij/ = DP/W of its wealth in the
form of bank deposits, providing liabilities for the banking system to trans¬
form into loans to firms. From the equation for Df just above, excess
demand equilibrium for these loans becomes:
wbaK — (5.11)
The interest rate / is supposed to vary rapidly to drive excess demand to

zero in the market for loans. This assumption marks a dramatic simplifica¬
tion of reality, but it seems best to adopt it for analytical purposes
here.9
To close out the description of the asset market, only one additional
excess demand function must be specified (with the wealth identity taking
care of market clearing for the third asset.) Note first that real returns to
the three assets are as follows:
Deposits DP\ id — 7r
Loans to firms LP: i — tt
Gold Z: 7TZ — 7T
A controlled interest rate id is assumed to be paid on bank deposits—the

real deposit rate is then id — tt. Firms pay the rate / to both banks and
the public for their loans, if market segmentation between lenders (or a
controlled bank lending rate) is assumed away for simplicity. The real
return on such loans is i — tt. Finally, the only pecuniary reason to hold
gold is in anticipation of capital gains. If the expected rate of growth of the
price of gold is 7r*, then real expected capital gains are tt. — 7r.
In the short run, gold is assumed to have a market-clearing price. If the

public wants to hold a fraction £ of its wealth as gold, then the short run
condition for the excess demand to equal zero is:
£(/', id, tt, 7r,)—-Z = 0 (5.12)

* T
The gold price P. varies to clear this market.

To summarize, the expected inflation rates tt and tt. are given at any
moment, so that the interest rate and gold price adjust to bring demands
in line with supply. The excess demand balances (5.11) and (5.12) are
brought into equilibrium in markets for loans and gold. When excess de¬
mands for these two assets are zero, then excess demand for money (by
Walras’s Law) will be as well.
5.3 The Model Without Gold—Short Run
The sensible way to figure out how a model with three assets works is to
attack it by stages. In this section, we follow this strategy by leaving gold
out. The relevant equations are commodity excess demand (5.10) and loan
market excess demand (5.11), dropping the terms tt2, Pz and Z. For a fixed
nominal wage w these equations constitute an equilibrium system for the
commodity price level P and nominal interest rate /. The growth rate g
follows as a dependent variable from (5.9). For the moment the hypothesis
of static expectations is maintained, so that expected inflation tt is a prede¬
termined variable.
The Jacobian for (5.10) and (5.11) takes the form:
—a (s — h)awb
--h
P P
tyK i|f,W
where:
a = [s - (y + h)]a - go + h(i - n) (5.13)
This long expression represents the excess of saving over investment when
the profit rate takes its maximum value a, and in practice will be positive.
Along with the condition i|/, < 0 (demand for deposits declines when the
interest rate on loans rises), a > 0 assures that the trace of the Jacobian
will be negative.
The determinant is:
A = — ATi|>[(s — h)awb — hP]} (5.14)

To check the stability condition A > 0, observe that the slopes of the excess
demand functions for commodities and loans respectively are:
dP ($ — h)awb — hP
di p a
(S h)awb- {[5 - (h + y)]a - h( 1 + 77)1 (5.15)

a
and
dP _ vj/,TF
(5.16)
~di\.i|)K
Since \|/, < 0, the loan market equilibrium derivative is positive. The
detailed explanation depends on the peculiarities of the asset structure set
out in table 5.1, but would go through in its essentials under other specifica¬
tions. In the present model, the immediate impact of an increase in the price
level is to raise the value of firms’ physical capital PK. With an agile equity
market, this capital gain would soon pass through into an increase in
price/earnings ratios on common stock. When equities are not widely
traded (the assumption underlying table 5.1, and the truth in most poor
countries), the assumption has to be that sooner or later firms will put their
windfall gains to use by demanding more loans. Bank lending is limited by
the reserve requirement; hence, public lending LP must rise. The requisite
resources are drawn into the loan market by an increase in the interest rate
and a reduction in bank deposits as determined by vjb-
In macro equilibrium, comparison of (5.15) and (5.16) with (5.14) shows
that the determinant A will be positive when the slope of the loan market
schedule from (5.16) exceeds the commodity market slope (5.15). In con¬
nection with (5.8) it was already noted that this latter relationship between
P and i can have either a positive or negative slope, depending on whether
cost pressure or investment demand reduction from a rising interest rate is
stronger. Macro instability occurs when working-capital cost-push is force¬
ful enough (or expected inflation it low enough) to make the curve repre¬
senting commodity market equilibrium steeper than the curve for the mar¬
ket for loans.
Figure 5.1 illustrates determination of short-run equilibrium, when the
price level rises in response to higher interest costs. Adjustment dynamics
are shown by the small arrows. In the upper quadrant, output is in excess
demand below the commodity market curve and loans are in excess demand
to the left of the loan market schedule.
FIGURE 5.1
Determination of Short-Run Equilibrium When the Only Assets Held by the
Public Are Bank Deposits and Loans to Firms
Note: There is assumed to be a strong effect of the interest rate on the price level via working-capital
costs.
The main comparative static result is that the loan market curve will shift
to the right either when there is monetary contraction or when the money
demand function x)/ shifts autonomously upward.10 The outcome will be a
rise in the interest rate i and declines in investment demand and the growth
rate g. In a more complete model with variable capacity utilization, the
story would be similar but slightly more complex. A rising interest rate
would reduce investment demand through (5.9). At the same time, it would
be reasonable to postulate an aggregate supply function of the form
X = X[P/{ 1 + /»]
At the initial price level P, an increased i would lead to cost pressures that
would induce producers to reduce their output X. If supply went down
more than demand, the resulting aggregate excess demand would lead to
a rising price along the lines of figure 5.1. Thus, monetary contraction
would not only slow growth and drive up prices, but also reduce output and
employment in the short run. The possibility that stagflation may result
from economic stabilization policies that restrict the money supply has been
heavily emphasized in macro-policy literature in developing countries in
recent years.11
A second comparative static exercise of interest has to do with changes
in the expected inflation rate n. In equation (5.8) an increase in n will
reduce the real interest rate and stimulate investment demand, shifting up
the commodity market market curve in figure 5.1. For a given price level,
the loan market curve has to shift to the right, with a higher value of i
countering the increase in 7r to maintain equilibrium in (5.11). The outcome
is a higher level of P—an increase in inflationary expectations drives up
aggregate demand and the current output price. A similar linkage will turn
out to be important in the long-run stability analysis under rational (not
static) expectations about inflation that appear in chapter 6. But before
going on to ask how well short-run results carry over into longer periods,
we turn to the next section for a description of the macro system when there
are assets besides money and loans to firms.
5.4 The Macro System with Gold
One standard recommendation of orthodox experts from institutions like

the World Bank or International Monetary Fund when they visit poor
countries is that the local banking authorities should raise deposit rates.
Such a move is supposed to increase saving as well as lead to an influx of
deposits into commercial banks. Under full employment of capacity, the
rise in saving will lead to faster growth. Higher bank deposits in principle
expand . . the real size of the banking system and hence the net flow of
real bank credit to finance investment.”12 More bank lending capacity
should stimulate long-term growth, as well as help satisfy working capital
needs and thus reduce inflationary pressure at a given level of aggregate
demand.
It is interesting to give these ideas about financial liberalization a run for
their money in a full capacity model like the one in this chapter, where the
positive effect of a higher saving rate on the growth rate is a foregone
conclusion. (Things would, of course, go differently under excess capacity,
as discussed at length in chapter 2.) The main conclusion is that an increase
in the deposit rate may not lead to more overall credit to firms with
associated benefits. The key question is whether the new deposits are drawn
from existing productive loans from the public to firms or from other assets
such as “gold.” In the former case, the deposit rate strategy could easily
backfire, as we demonstrate shortly. This possibility perhaps explains why
generously offered advice from flying experts about interest rate policy in
the Third World has been almost universally declined.13
The model with gold is based on the three excess demand functions
(5.10)—(5.12). For the record, one can write out the comparative statics of
a change in the deposit rate in the following matrix equation form:
a B
0 dP 0 0
~P ~P
r\fK v)fiW x\fZ di + w^d did = 0 (5.17)

_1
£K £,W (1 - g)z
*t3
^3
0
Pz _ N
1
I
The symbol a has been defined in (5.13), B stands for (s — h)awb — Ph,
and the subscripts / and d respectively denote partial derivatives with
respect to the interest rate on loans to firms and the deposit rate.
The standard conditions for stability of a three-dimensional system like
(5.17) are that the trace of the Jacobian should be negative, the sum of the
three second-order principal minors (or 2 X 2 determinants linking each
possible pair of excess demand functions) should be positive, and the third-
order determinant should be negative. Both trace and second-order condi¬
tions follow easily when i|/, < 0 and £, <0, or an increase in the interest
rate on loans pulls wealth holding away from both bank deposits and gold.
After some manipulation, the third-order determinant can be written as:
r = [aW(x|/,(1 - 0 + v|/0) + BxJ/AT] (5.18)

pjp
It is easy to see that T will be negative so long as B is not too large. A

positive value of B occurs when a rise in the interest rate leads to a price
increase due to cost pressure instead of a price reduction due to reduced
investment demand. As in the model without gold, the stability condition
is that the working-capital cost-push cannot be “too strong.”
The derivative of the interest rate on loans with respect to the deposit
rate is:
di (XZtV _ y\ \
[(1 - 0v|fd + iJ/0]
t y i /r ,
(5.19)
did TP7P
di / o [7 « > 3+ ©
Note that the partial derivative is positive since an increase in the

deposit rate id presumably leads people to hold more deposits. By contrast,
is negative and the sign of the derivative in (5.19) is unclear. A negative
sign for di/did occurs when the absolute values of both £ and £d are
relatively large—the share of wealth held in the form of gold (£) and the
willingness of the public to shift out of gold toward bank deposits (£d) both
have to be high. Under these circumstances, raising the deposit rate will pull
money into banks for relending to firms, drive down the loan rate 4 and
lead to faster growth from the investment demand function (5.9). An added
benefit will be a reduction in the price level when it responds positively to
the interest rate from working-capital costs.
The economics behind these pleasant events can be illustrated further by
manipulating the identities that link the different assets. It is straightfor¬
ward to show that di/did < 0 when the following condition holds:
\\>d (H/n) - wbaK

Zd LP -f Dp
This condition is more difficult to satisfy when the absolute value of £d is

small, as just noted. It is also more likely to be violated when the commer¬
cial bank reserve requirement is large. Thus as ju. approaches zero (and
banks generate pure “inside” money by credit expansion), any ratio of
deposit rate derivatives on the left side will satisfy the inequality. On the
other hand, when fx is one (100 percent reserves and pure "outside” money)
the condition becomes:
H — wbaK
U < H + PK
in which the right side is certain to be less than one. Under these cir¬
cumstances, financial liberalization will not bring good results, at least un¬
der the standard assumption that financial assets are gross substitutes, or
H'rf >
The difference between the two extremes hinges on the degree of financial
intermediation that the commercial banks can legally provide. The informal
market operates with no reserve requirement, channeling resources toward
firms with great efficiency. By contrast, a fraction of each new deposit in
the commercial banks goes to reserves and cannot be loaned out. Thus,
unless banks largely draw hoarded assets into deposits when id goes up, the
overall effect of reform can be stagflationary.14
These effects can be illustrated in a diagram if we first solve for the price
of gold by eliminating the wealth variable (tV = H + PK + PzZ) from
(5.11) and (5.12):
^ (H/fx) — wbak
(5.20)
vj7[ Z
Clearly, Pz falls as Z rises with unit elasticity. But the sign of its response
to interest rate changes is ambiguous since both and i|/( are negative. If
|£,-/£;| > then the derivative dPz/di will be negative. Again, the
assumption is that the public is loath to hold gold when returns to alterna¬
tive assets rise.
Plugging PzZ from (5.20) back into (5.11) allows a solution of the system
in the variables i and P alone. For the case of a positive response of the price
level to the interest rate due to working capital cost, the comparative statics
are shown in figure 5.2.
FIGURE 5.2
Effects of an Increase in the Deposit Rate When Increased Deposits Come Mainly
from Gold Holdings and a Rise in the Interest Rate Drives Up the Price Level by
Working-Capital Cost
The deposit rate increase pulls resources into the banking system, having
the same impact as an increase in the money supply. The loan market
schedule shifts to the left. At the same time, if the saving rate goes up in
response to a rise in id, there will be less demand pressure at a given interest
rate and the commodity market schedule shifts down. The outcome will be
lower prices and interest rates, coupled with faster growth. The financial
reform package works well, with the only untoward consequence being a
possible rise in the price of gold in response to the interest rate changes.
Gold speculation resulting from strong expectations about a rise in its price
could be destabilizing in the medium run, as discussed further on.
If the loan market curve in figure 5.2 were shifted to the right instead
of to the left, it would correspond to a situation in which the deposit rate
increase mostly pulls asset demand from loans to firms. If, in addition, the
saving rate response to id is weak (a small downward shift in the commodity
market schedule), the outcome would be a higher price level and slower
growth. In this case, as in the monetary contraction discussed in connection
with figure 5.1, orthodoxy gives rise to stagflation in the short run.
Finally, figure 5.3 illustrates monetary contraction, or a deposit rate
FIGURE 5.3
Effects of an Increase in the Deposit Rate When Increased Deposits Come Mainly
from the Public's Loans to Firms and a Rise in the Interest Rate Reduces the
Price Level by Decreasing Investment Demand
increase with stable gold demand in the case when the price level responds
negatively to the interest rate. The result is more traditional in that growth
slows and prices fall.
These outcomes suggest that simple pronouncements about financial pol¬
icy don’t make much sense, even in poor countries where financial markets
are not notably complex.
Moreover, it is easy to spin out disquieting medium-run asset market
scenarios set off" by short-run events such as depicted in figure 5.2. For
example, an increase in the deposit leads the interest rate i and output price
P to fall and the gold price Pz to rise. Suppose that expected price changes
7r and ttz follow in the same direction. Then the returns to holding deposits
and gold (respectively id — tt and itz — ir) will rise relative to the return
to loans (/ — 7r). The corresponding attempts to shift portfolios toward
money and gold could lead to an accelerating gold price increase and shifts
away from loans and (perhaps later) money. The final outcome could be
rising interest rates and finally economic collapse as working capital finance
dried up. Such unstable expectational bubbles are not easily ruled out when
there are three assets on the scene.15
Even leaving aside unstable dynamic stories, there is quite a lot else that
can happen in the economy as the results of these last two sections are
extended toward the medium run. But the lesson remains that if assets held
as gold, land, or other hoards are not easily relinquished when bank deposit
rates go up, then aggressive interest rate policy will do little good for the
economy in the short to medium run.
5.5 Money in the India Model
Some of the effects just discussed can be illustrated in practice by adding

a demand-supply balance for money to the model for India from chapter
4. Both for simplicity and because of lack of data we eschew an extra asset
in the system such as gold.
Tables 5.2 and 5.3 set out changes and additions to the model’s original
equations from table 4.2. Three major modifications are included:
1. Working-capital cost-push is assumed to be present in the industrial sector (num¬

ber three). Thus the cost equation (4.16) is modified to include the interest rate
i in nominal costs. Total working capital demanded by the sector is kwcB 3^3,
where X3 is the output level, B 3 is variable cost per unit output, and kwc is a
stock-flow conversion factor that relates the sector’s stock of outstanding credit
(estimated in practice from flow-of-funds data) to its flow of outlays for current
production costs. The same factor kwc enters, multiplying the interest rate in the
new version of (4.16).
TABLE 5.2
Equations to Put Money in the India Model
I. Working-Capital Markup in Sector 3 [New Version of (4.16)]
(1 + U)0 + TjXl + hî)

P. =
1 - (1 + U)0 + r3)(l + kwci)a33(1 - sub]) /=i
aoP 1 + 2 aoP1
+ wjb) + aêP 03(1 — / 03)] (4.16)
II. Investment Demand
h = r> + h(i - «°)_(4.49)
III. The Money Market
GBANK = GEXP + GINV + />,(G/>, - FA, - FN,) + NA - GREV

- GBFOR - GBDOM - OTD (4.50)
DRES = GBFOR + CBFOR - DEF - OTA (4.51)
CURR, = CURR 0 + CURR (4.52)
GBANK, = GBANK a + GBANK (4.53)
DRES, = DRES0 + DRES (4.54)
CBANK, = k„B}X} (4.55)
CBANK = CBANK, - CBANK0 (4.56)
RLIAB, — RLIABq + RLIAB (4.57)
BLIAB, = BLIABq + BLIAB (4.58)
MDEM = -!-(Yw + Y. + + X GR,) (4.59)

V0 + V|/ /_3
7E Demand-Supply Balance for Money
MDEM - (GBANK, + CURR, + DRES,

+ CBANK, - BLIAB, - RLIAB,) = 0 (4.60)
2. Investment demand is assumed to respond to the nominal interest rate in (4.49).

The coefficient h was assigned a value of 0.7, in line with econometric and Indian
economists’ intuitive assessments of investment demand responsiveness in the
economy.
3. The demand-supply balance for money is tied back to variables generated on the
real side of the system, with the interest rate acting as the corresponding adjusting
variable.
The details of calculating money supply appear in equations (4.50)

through (4.58). It is assumed that changes in flows such as government
borrowing from the banking system or the balance of payments surplus
affect the money stock within the model’s solution period of one year. This
treatment differs from the formal models of this and following chapters,
where flows affect the time derivatives of stocks at “the present moment”
and only cumulate into stock changes over time. The distinction arises
because the formal models work with differential equations for mathemati¬
cal tractability, whereas an applied model based on annual data has to take
into account such banalities as the fact that the government spreads its
borrowing from the banking system over the national accounts year instead
of waiting to do it all at 11:59 p.m. on March 31.
The components of the money supply are shown in the term in paren¬
theses in (4.60). The stock of bank credit outstanding to the government is
GBANKX’, its change during the year is GBANK from (4.50). The govern¬
ment’s borrowing from the banking system rises with its current spending
GEXP, investment GINV, foodstock accumulation PfGPx — FA , — /TV,),
and items that are not otherwise accounted (NA) for. Its sources of funds
are current revenue GREV and loans from nonbank sources both home and
abroad. For historical reasons the government of India (not the banks) has
small currency liabilities to the public, accounted as CURR , at the end of
the period. Foreign assets of the banking system (4.60) are DRES \—their
change DRES is defined in (4.51). DRES is the difference between govern¬
ment and commercial borrowing from abroad (GBFOR and CBFOR, re¬
spectively) and the current deficit DEE and other transactions.
The third main component of the money supply is credit to the private
sector CBANKX, which is related to flow demand for working capital in
(4.55). The items BLIAB x and RLIAB, are nonmonetary liabilities of the
commercial banks and the Reserve Bank of India. They can be varied to
reflect monetary policy changes, as in the simulations that follow.
The entire equation (4.60) sets demand minus supply of money (or excess
demand) equal to zero as an equilibrium condition. Money demand itself
is MDEM, and is given as a function of the interest rate and total value-
added in the economy in (4.59). The particular functional form assumed
states that money velocity (or value-added divided by the money stock) is
an increasing function of the interest rate; people hold less money as / goes
up. The parameter values are v0 = 1.5215 and v, = 5.0, in line with
empirical estimates for India.
Table 5.4 shows results from two simulations with this extended version
of the model. In the first, a monetary contraction is induced by reducing
TABLE 5.3
New Symbols for Putting Money in the India Model
I. Endogenous variables
h — investment demand for sector 3 products (exogenous in chapter 4 model)

GBANK — within-period net government borrowing from the banking system
GBANK, — end-of-period government debt to the banking system
DRES — change in foreign asset holdings of the banking system
DRES, — end-of-period foreign asset holdings of the banking system
CBANK — within-period change in net bank credit to the commercial sector
CBANK, — end-of-period net bank outstanding credit to the commercial sector
CURR, — end-of-period government of India currency liabilities to the public
RLIAB, — end-of-period nonmonetary liabilities of Reserve Bank of India
BLIABt — end-of-period nonmonetary liabilities of commercial banks
MDEM — demand for money
i — nominal interest rate
II. Exogenous variables
n — autonomous investment demand for sector 3 products

GINV — value of government investment spending
GBFOR — government borrowing from abroad (rupees)
GBDOM — government borrowing from nonbank financial sector and the public
CBFOR — commercial sector borrowing from abroad (rupees)
CURR — within-period change in government currency liabilities to the public
BL1AB — within-period change of nonmonetary liabilities of commercial banks
RLIAB — within-period change of nonmonetary liabilities, Reserve Bank of India
NA — government current expenditures, not accounted otherwise
OTU — other government domestic transactions, from "statutary liquidity ratio”
OTA — other foreign transactions
III. Parameters
v0, V| — money demand parameters

h — interest rate parameter for investment demand
kwc — stock-flow coefficient between working capital demand and sector 3 output
level
IV. Initial values
CURR o initial government currency liabilities to the public

GBANK o initial government debt to the banking system
DRESo initial foreign assets of banking system
RLIAB o initial nonmonetary liabilities. Reserve Bank of India
BLIAB0 initial nonmonetary liabilities, commercial banks
CBANK 0 initial stock of bank credit to the commercial sector
io beginning-of-period nominal interest rate
TABLE 5.4
Effects of Contractionary Monetary Policy and Devaluation in the India Model
Incorporating Money

Base Mon. Base Mon.
Run Policy Deval. Run Policy Deval.
1 1.0665 1.0003 *3 9.5712 9.5621 9.5253

Pi 1 1.0488 0.9997 *4 1.0262 1.0254 1.0191
Pi 1 1.0446 1.0226 *5 6.5749 6.5703 6.5276
P4 1 1.0198 1.0075 h 2.3186 2.2974 2.3168
P5 1 1.0244 1.0069 i 0.1200 0.1503 0.1229

Base Mon. Base Mon.
Run Policy Deval. Run Policy Deval.
Agr. sav. 0.5373 0.5723 0.5370 Dom. inv. 2.3186 2.3998 2.3692
Wage sav. 0.6414 0.6478 0.6463 Imp. inv. 0.1211 0.1211 0.1393
Prof sav. 0.5325 0.5773 0.5388 Pr. stocks 0.4048 0.4272 0.4127
Trade def. 0.1800 0.1962 0.1554 Gov. stocks -0.1028 -0.1028 -0.1028
Gov. sav. 0.1357 0.1368 0.2260
Deprec. 0.7148 0.7148 0.7148
Total: 2.7417 2.8421 2.8183 Total: 2.7417 2.8421 2.8183
C. Income Shares
Base Mon.
Run Policy Deval.
Agr. income 0.3895 0.3956 0.3881

Wage income 0.3765 0.3626 0.3774
Profit income 0.2340 0.2418 0.2345
Total 1.0000 1.0000 1.0000
nonmonetary liabilities of the Reserve Bank of India to give a drop in money

supply of 4 percent. The simulation is done with wage-indexation, foodstock
speculation, and rising markups all incorporated in the model, along the
lines of table 4.7.
The impact effect of the monetary contraction is to increase the interest
rate, which leads in turn to cost-push in sector three and a decline in
investment demand. With our assumptions regarding parameters, the out¬
comes include price increases in all sectors, output decreases in the nona-
gricultural sectors, and a shift in the income distribution from wages toward
agriculture and profits. It seems reasonable to characterize the monetary
contraction as stagflationary.
Somewhat similar results come from a 15-percent devaluation when
exports are assumed not to respond to a higher exchange rate. As noted in

connection with the discussion of table 4.8, devaluation in the model with¬
out money leads to a decline in the total of bank credit expansion induced
by government borrowing and the balance of payments—as a consequence
the money supply drops. The interest rate rises, and as shown in table 5.4
the result is price pressure and a fall in investment demand. Hence, devalua¬
tion has contractionary effects through monetary as well as real channels.
Under other parameterizations of the model, these results reverse them¬
selves—they should not be taken as predictions about what will happen in
India. Rather, they are presented here to illustrate the fact that stagflation
from devaluation or contractionary monetary policy is by no means ex¬
cluded as an outcome of a plausible general equilibrium model of an actual
economy. Under certain circumstances, stagflation occurs in the real world
as well as the model.
6
Money, Inflation,
and Growth
T HE ANALYSIS in chapter 5 was strictly for the short run. Now it

is time to see how well it extends to a longer period of time. The “medium
run” is considered here for an economy in steady state, with all real varia¬
bles growing at a constant rate, and the growth rates of output and the price
level summing to the rate of growth of money supply. One question asked
regards the effect of changing the growth rate of money on other variables,
across distinct steady states. A second question is whether or not the
adjustment process between steady states will be stable. The answers are
that slowing the growth of the money supply from an initial steady state
can in certain circumstances lead to faster inflation, slower output growth,
and a lower real wage at a new steady state. Moreover, the transition
between trajectories can be stable under quite orthodox assumptions about
the dynamics of the system. The results suggest that monetary contraction
can be stagflationary in the medium run in real economies. The finding is
novel, but perhaps not so surprising in light of the important role played
by working-capital cost-push in the models of chapter 5.
Causality in this chapter’s model runs along Marxian lines, from conflict
between workers and capitalists about the rate of increase of the money
wage in a generally inflationary situation. Rules for indexing the wage to
past inflation are assumed to be in place. However, if inflation accelerates,

the real wage will fall due to lags in the indexing process. Workers will force
an acceleration in the growth rate of the money wage to try to regain their
real income, and as a consequence there will be a profit squeeze. How this
process works in detail is described in section 6.2.
With the real wage (or, for analytical convenience, the price-wage ratio)
given by such class interaction, the macro system in the short run must
adjust to predetermined levels of the price-wage ratio and the money stock
(or, again for convenience, the ratio of money to the value of capital). A
change in the price-wage ratio will affect available saving by shifting the
income distribution, and investment must in some way respond. We adopt
A
the fashionable rational expectations notion that the rate of inflation P

A
varies to alter the real interest rate i — P and drives investment to equality
with saving. As in most Marxian models, investment is finally determined
by saving, but with the particularly un-Marxian rationale, expectations
twist. How this saving-investment adjustment is linked with financial mar¬
kets is described in section 6.1.
With wage and inflation dynamics set up in the first two sections, the
story about transitions between steady states is told in section 6.3, when
money and loans by the public to firms are the only assets in the system.
The possible role of “gold” and other extensions are briefly sketched in
section 6.4, while an appendix contains the mathematical underpinnings of
some of the analysis that is done graphically in the body of this chap¬
ter.1
6.1 Short-Run Adjustment under Rational Expectations
One of the standard tricks in working with growth models is to invent

“state variables” that fit into tractable differential equations so that their
trajectories can be studied over time. The model from the last chapter can
be restated usefully if we define two new variables as R = P/w, or the
price-wage ratio, and V = H/PK, or the ratio of the money supply to the
value of the capital stock. The depth or degree of intermediation of the
financial system is loosely indicated by V—a higher value means more
intermediation.
In terms of R and V, the excess demand functions (5.10) and (5.11) for
commodities and money can be restated as:
(.s — h)( 1 + i)ab

- h(i — P) + g0 — [s — (y + h)]a = 0 (6.1)
R
Money, Inflation, and Growth 111
and
ba + 4/(7 - P)(V + 1) - j V = 0 (6.2)
where the actual inflation rate P replaces the expected rate tt in the invest¬
ment demand term h(i — P) and the money demand function v)/(i — -Pi-
Adjustment stories in perfect foresight models are often farfetched, per¬
haps justly so in light of the incredibility of the hypothesis itself. In the
present model, R and V are given at any point of time from dynamic
processes to be described below. Hence we need two variables to adjust
“within the moment” to assure that (6.1) and (6.2) are satisfied.
Look first at the market for commodities. If the price-wage ratio R rises
from an initial equilibrium level, the macroeconomic outcome is an incipi¬
ent excess of saving over investment (given that capitalists’ saving rates
exceed those of workers). In an agile financial market, one might expect the
real rate of interest to fall, to stimulate investment demand, and to restore
equilibrium. This bit of entelechy can occur through either a decrease in the
nominal interest rate / or an increase in the inflation rate P. Assume the
latter, that inflation accelerates when there is initial excess saving supply.
Then if the interest rate goes up when there is excess demand for loans in
(6.2), we have a locally stable short-run adjustment process. Financiers are
assumed to sense potential imbalance in commodity markets and adjust
their inflation expectations and thus the cost of borrowing rapidly enough
to prevent disequilibrium from actually coming about.2
This short-run market balance can be perturbed by shifts in the state
A
variables. To see the details, we ask how P and / respond to maintain

equilibrium in (6.1) and (6.2). For the financial market, the answer is easy.
Money demand depends on the real interest rate i — P. To maintain
balance, any change in i has to be matched by an equal change in P—the
two variables trade off along a 45-degree line.
In the commodity market, the story is a bit more complex. For a
given value of the interest rate, an increase in the price-wage ratio R
generates extra saving by moving the income distribution in favor of profits.
As discussed above, financiers catch this shift in the wind and increase their
A
expected inflation rate P in anticipation of an investment boom to come.

Now turn to interest rate reactions. In the last chapter it was shown that
at a given inflation rate, the nominal interest rate can either rise with the
price-wage ratio when working capital costs are important, or fall to stimu¬
late investment even further when R goes up. The outcome is an ambiguous
slope of the (i,P) equilibrium relationship in the commodity market. With
dominant working-capital effects, the slope is negative—both P and / rise

with R and must trade off inversely to hold R constant. When investment
A
demand effects are more important, the equilibrium (i,P) locus will have
a positive slope.
Comparative statics of the dominant working-capital case are shown in
figure 6.1. In the upper diagram an increase in V, the ratio of money to
FIGURE 6.1
Comparative Statics of an Increase in the Money-Capital Ratio V (Upper
Diagram) and Price- Wage Ratio R (Lower Diagram) When an Interest Rate
Increase Drives Up the Price Level through Higher Working-Capital Costs in the
Short Run
capital stock, requires a lower real interest rate to maintain money market
balance; for a given P, i must decrease. After commodity market reactions
are taken into account, a rise in V is associated with a lower interest rate
and a higher rate of inflation.
In the lower diagram of figure 6.1, we see that an increase in the price-
wage ratio R is associated with a higher interest rate at a given inflation rate
in the commodity market (as just discussed). In the loan market, an increase
in R raises the value of the term (R /fx)V = H/jjlwK or the degree of
monetization measured in wage units. Once again the real interest rate must
fall, so the loan market curve shifts to the left. A higher price-wage ratio
is associated with faster inflation, but an ambiguous shift in the rate of
interest.
Figure 6.2 shows what happens when investment demand responses dom-
Comparative Statics of an Increase in the Money-Capital Ratio V (Upper

Diagram) and the Price- Wage Ratio R (Lower Diagram) When an Interest Rate
Increase Reduces the Price Level by Cutting Back on Investment Demand
inate the interest rate effect. In the upper diagram, an increase in the
money-capital ratio reduces both the interest and inflation rates. The lower
A
diagram shows an ambiguous response of both P and i to a higher price-

A
wage ratio. However, P will rise if investment demand is insensitive to

changes in the real interest rate and the commodity market curve is rela¬
tively flat.
6.2 Wage Dynamics in the Medium Run
To incorporate these comparative static responses into a growth model,

the next task is to bring in wage and price changes. Assume that money
wages are fully indexed to the rate of price inflation. However, even with
full indexing, lags in the institutional adjustment process can lead to a loss
in workers’ purchasing power when the rate of price increase goes up. The
time path of the real wage consistent with indexing rules will subsequently
be taken as a standard around which actual wage inflation varies, but for
the moment we concentrate on formulation of the standard itself.3
A
Suppose there is ongoing price inflation at the rate P, where as usual the
A
“hat” denotes a logarithimic time derivative: P = (dP/di)/P. Wages,

however, are indexed only at discrete points in time, for instance, every year
or every quarter. If X stands for the indexing interval, then at times t,
t + X, and so forth, workers get a 100-percent adjustment in nominal wages
to compensate for immediately past inflation, that is, wl+x/w, — epx. At
time / + X, the real wage will satisfy a condition such as w,+x/P,+x =
w*, where the target real wage to* is assumed to be constant over time.
Figure 6.3 illustrates how the real wage changes between readjustments.
FIGURE 6.3
Variations in the Real Wage over Time When There Are Periodic Money Wage
Adjustments with an Ongoing Price Inflation at Rate P and the Target Wage Is
0)*
There is a cycle between the values o>* and a)*e~^x, which is of course
repetitive when the inflation rate stays constant.
Over time, the average real wage to is given by:
assuming P stays constant.

A
If we let PX = p, then from PHospitaPs rule one can easily show that
co approaches co* as p approaches zero, that is, the average wage becomes
the target either when inflation falls to zero or when there is instantaneous
indexing. Similarly, one has:
a co co*
ap p2
But since ep = l-t-p + p2/2 + ..., this derivative is negative for p >
0. The real wage declines when either the inflation rate is higher or the
indexing period gets longer. If the period is one year, the following numbers
illustrate the impact of increases in the inflation rate on the real wage (and,
for later reference, its inverse or the price-wage ratio):
Annual
Inflation Real Price- Wage
Rate Wage Ratio
0.0 1.0 1.0

0.10 0.952 1.051
0.25 0.885 1.130
0.50 0.787 1.271
0.75 0.704 1.422
1.00 0.632 1.582
With a fixed-period indexing scheme, an acceleration in inflation can cut

back sharply on workers’ real purchasing power.
To extend (6.3) to the macro system, one has to recognize that contracts
are not all renegotiated or indexed on fixed calendar dates—there is likely
to be some spread of wage bargaining over the year. If the time span between
each group’s readjustment is always X, then it is easy to see intuitively (and
can be verified formally) that (6.3) remains valid when readjustments occur
uniformly over time. An example would be a yearly cycle in which Vn of
workers get readjustments at the end of January, Vn at the end of February,
and so on. In countries where 100-percent wage indexing has been applied,
such uniformity is not observed.4 However, we will assume that strictly
uniform wage adjustment exists to make analytical discussion of the dynam¬
ics of inflation easier.
6.3 Inflation and Growth in the Medium Run (with no Gold)
Equation (6.3) is a relationship between the average real wage and infla¬
tion that would hold at all times if the economy’s wage indexing rules were
followed exactly. It summarizes institutional arrangements that give equi¬
librium on the supply side of the labor market (with “supply” referring not
to the number of people seeking jobs, but rather to the real wage consistent
A #
with indexing rules when the inflation rate is P). By inverting the equation,
the supply side value of the price-wage ratio R s can be written as:
P\
Rs = (6.4)
g>*(1 _ e-K)
The derivative of Rs with respect to both the inflation rate P and the
indexing period X will be positive. The second derivative is also positive, or
A
the impact of an increase in P on the equilibrium price-wage ratio strength¬

ens as the inflation rate is higher.
Since it represents a bargained wage consistent with inflation in the
medium run, R, plays a role in the following analysis similar to the money
wage of the short-run model—it provides a fixed magnitude against which
the inflation rate can vary. At any time the observed ratio R can differ from
Rr The assumption is that the money wage will grow faster than inflation
when the actual real wage is high relative to the institutionally warranted
level from (6.4). In other words, a tight labor market leads to faster than
normal money wage increases because the indexing rules get bent by wage
drift, extra bonuses, and so on. This hypothesis amounts to a restatement
in a model with inflation of the profit-squeeze cyclical adjustment outlined
in section 2.4. Here, as in that model, stability will depend critically on shifts
in the functional income distribution over time.
In formal terms, the money wage increases according to the equation:
When the labor market is tight, R lies below Rs, and w > P. Restated in
terms of R, this out-of-equilibrium adjustment rule becomes:
R = P — w = <f>(R — Rs) (6.5)
Note from (6.4) that Rs is a function of P, while from (6.1) and (6.2) P is
determined by R and V Hence, (6.5) is a shorthand representation of how
R evolves over time as a function of itself and V We have one of the two
differential equations for the state variables that we require.
The other equation can easily be written as:
A A A A
V — H — P — K (6.6)
For the moment, assume that the growth rate of the money base (H) is
exogenous. Equations (6.1) and (6.2) give P and i, while (5.9) gives K (or
g) after the current inflation rate is substituted for the expected rate tt. In
effect, V becomes a function of R and V. Given equations (6.5) and (6.6),
the next step is to consider their stability as represented by the Jacobian
matrix:
'dV/dV a V/dR"
dR/dV dR/dR
The sign pattern under appropriate assumptions takes the form:
— or + -
so that stability occurs if the term in the southwest corner is not too strongly
negative. The rationale goes as follows:
1. In the southeast of the Jacobian, begin by observing from figures 6.1 and 6.2 that
a reduction in the price-wage ratio R pulls down the inflation rate in the short
run, unless investment demand is highly sensitive to real interest rate changes.
But then from (6.4) the price-wage ratio consistent with the wage indexing goes
up. Workers’ nominal wage demands moderate since the difference between the
current real wage and the target declines (as also would the level of employment
in a more realistic model with variable capacity utilization). If money wage
growth slows sufficiently, R = P — w will rise. Hence, the derivative dR/dR
takes a negative sign. From (6.5) it is clear that this result depends on a large
A A
positive derivative of Rs with respect to P in (6.4), and of P with respect to

R in (6.1) and (6.2). We assume that the derivatives are large enough for stability
—an initial profit squeeze is followed by enough deflation to make the squeeze
disappear. We have a restatement under inflation of the real-side dynamics of
section 2.4, and the key adjustment mechanism in the present model.
2. The derivative dR/dV in the southwest corner of the Jacobian can take either
sign. As argued in connection with figure 6.1, an increase in V lowers the interest
rate and drives up the inflation rate in the short run when working capital
dominates interest rate responses. The higher P raises the institutionally accept¬
able price-wage ratio Rs or lowers the indexed real wage. Workers’ wage demands
are animated as the indexing rules start to penalize them and money wage
increases accelerate. As a consequence, the growth rate of the price-wage ratio
slows, or dR/dV < 0. The contrasting case is the reduction in the inflation rate
resulting from a higher level of V in the upper diagram of figure 6.2 (referring
to the case where investment demand effects dominate the interest rate response).
The process just described is reversed, and dR/dV > 0. The switch in signs of
the partial derivative gives either a positive or negative slope to the locus of values
of R and V, along which R — 0. There is an exact parallel with the ambiguous
slope of the commodity market relationship between P and / that was discussed
in chapter 5, as is illustrated in figures 6.4 and 6.5.
3. The signs of the derivatives a V/d V and a V/dR depend on the effects of shifts
in the state variables on inflation and the rate of capital growth, as can be seen
from equation (6.6). If working-capital cost-push is the dominant effect of interest
rate increases, figure 6.1 shows that P rises along with both V and R. When
investment demand effects dominate as in figure 6.2, P may decline, but this
response is likely to be weak. All the diagrams show that the real interest rate
i — P declines when either V or R rises; from (5.9) the capital stock grows faster.
The conclusion is that the growth rate V is likely to slow when either state
variable gets bigger, that is, V and R trade off inversely to hold V = 0.
The overall stability condition is that the slope of the locus of points along
which V = 0 should be less (more negative) than the locus along which
R = 0 in a diagram with V on the horizontal axis and R on the vertical.
An increase in the rate of growth of the money supply (H) shifts the V
— 0 locus to the right—either the inflation or growth rate has to rise to meet
the steady growth condition H = P + K, and either shift is associated with
increased monetization V for a given price-wage ratio R.
If stability is assumed, figure 6.4 illustrates the dynamics in the dominant
working-capital case. The adjustment involves a gradual increase in V and
a decline in R from the model’s initial equilibrium at A. Since at the new
steady state at B the labor market will be in equilibrium, (6.4) shows that
the inflation rate P will be lower. Also, since H = P + K the growth rate
will be higher. Compared across steady states, faster money growth leads
to lower inflation, more rapid capital stock growth, and a higher real wage
when interest rate increases drive up working-capital costs. The faster
FIGURE 6.4
Adjustment between Steady States under Rational Expectations in an Economy
Where an Interest Rate Increase Drives Up the Price- Wage Ratio from Working
Capital Cost-Push in the Short Run
Note: Faster money growth gives lower inflation, an increased real wage, and faster output growth.
Adjustment between Steady States in an Economy Where an Interest Rate

Increase Reduces the Price-Wage Ratio by Cutting Investment Demand
Note: Faster money growth gives higher inflation, a reduced real wage, and faster output growth.
monetary expansion permits more rapid growth and less inflationary cost
pressure in the long run.
Details of the transition between steady states depend on the specific
dynamic assumptions of the model. In the case at hand, when money
growth is jumped up from an initial steady state, V or H/PK will start to
rise while R is momentarily stationary. The implication from figure 6.1 is
that inflation will speed up. Thereafter, R will begin to fall (as it would with
monetary expansion in the static expectations story of figure 5.1) and the
inflation rate will follow. The adjustment pattern for “half-cycle” conver¬
gence as assumed in figure 6.4 will be initial acceleration and then decelera¬
tion of P. There could be further oscillations with the adjustment trajectory
spiraling in toward point B, but it is convenient to assume them away.
Adjustment in the alternative commodity market case in which an inter¬
est rate reduction drives up the price-wage ratio by increasing aggregate
demand is shown in figure 6.5. As usual, the results are more traditional:
an increase in the growth rate of money base leads to higher steady-state
inflation by driving up the price-wage ratio R. The money-capital ratio
A
V rises as well, so that the real interest rate i — P will fall. From (5.9)
investment demand will be stimulated, so that the growth rate goes up in
this case as well. There is not a sufficient fall in the interest rate to allow
real wages to rise. Forced saving supports faster growth in the long run.
These results are fairly straightforward extensions of the short-run analy¬
sis of section 5.2, to a situation where plausible price and wage dynamics
apply. With volatile rational expectations built into the specification, the
model’s stability properties are bound to be tenuous, as indeed they are. If
there is stability, it comes from two linked relations. First, a rise in the
price-wage ratio R (or a fall in the real wage) is linked to a faster inflation
A
P through aggregate demand. Second, rising inflation pushes up the target

price-wage ratio R s (or reduces the target real wage) and thus slows the
increase in R from (6.5). If these links are strong enough, stability can be
observed; otherwise not. They could be strengthened by more wage lag, for
example, less than complete indexation in (6.3) by reduction of the target
real wage, lengthening of the adjustment period, or postponement of adjust¬
ment for last period’s inflation until well into this period’s calendar run. The
fact that an institutional wage squeeze helps bring macro stability has not
escaped the notice of monetarists. Their stabilization programs ostensibly
(or ostentatiously) concentrate on reduced H, but incomplete indexation of
wages often also enters in. A dose of civil repression may be added to assure
that workers stay in line.
A second observation is that so far H has been treated as exogenous. This
hypothesis makes sense under the circumstances in which the central bank
has both monetary and nonmonetary assets in its portfolio and can sell them
back and forth (typically in open market or foreign exchange operations)
with other agents in the system. Strictly speaking, such maneuvers are ruled
out here by the balance sheets of table 5.1. Indeed, since there are no taxes
in the model, monetary expansion is determined by the needs of government
finance: dH/dt = PG. Plugging this expression in to the rest of the equa¬
tions shows that the growth rate of money base is given by:
H = 1 (6.7)
ju, v|(R + ba
A A
Evidently, H depends on both i and P through the money demand function

v|/ and the markup R. The effects are ambiguous, but working through the
algebra suggests that the negatively sloped equilibrium locus for V in figures
6.4 and 6.5 will remain. Money growth can be stimulated when (6.7) holds
either by raising the government expenditure share y or cutting the reserve
requirement ju.
6.4 Gold and Other Matters
Bringing gold into the system means that another asset market and state
variables have to be considered. Define the state variable as U = PzZ/
PK, or the value of gold holdings relative to the capital stock. The short-run
equilibrium conditions incorporating gold become:
ba + vj//?(i' - P, id - P,PZ - P)(V + U + 1) - 0 (6.8)
for the loan market, and
Z(i - P, id - P, Pz - P)(V + U+\)— U = 0 (6.9)
for gold. Stability analysis could be done as before, by inverting (6.8) and
(6.9) for P and Pz and then substituting the resulting expressions into
differential equations for the state variables U and V. We forego the gory
details and make only brief comments on the nature of steady-state equi¬
libria.
First, the growth of money base must now be written as:
£ = (l - £) - ^ yaR
n v|//? + (1 — i)ba
Note that a rise in the deposit rate id increases v|/ and reduces £, so that its
effect on money growth is ambiguous. However, enthusiasts for financial
liberalization presumably suppose that the effect of the change in v|/ would
be stronger, so that money growth falls. In the case where aggregate de¬
mand effects from interest rate changes dominate, figure 6.5 shows that
inflation would drop off and the real wage would rise, but the growth rate
would decline. A strong saving response to higher deposit rates could
salvage the growth rate, although empirical evidence suggests little to hope
for on this score. The story in figure 6.4 (dominant working-capital effects)
is worse, for there the inflation rate would rise and real wage decline in the
long run.
The conclusion is that unless coupled with expansionary monetary policy
from some other source (perhaps a reduction in reserve requirements),
financial liberalization will do little to benefit economic performance in the
medium run. Moreover, it may not even lead to an increase in financial
depth, as measured by the ratio H/PK. Simple manipulation of (6.8) and
(6.9) shows that this variable can easily move in either direction across
steady states.
A final question might be raised regarding the two aggregate demand
cases so scrupulously distinguished in the analysis: Which one really mat¬
ters in a given economy over a given time span? Regarding timing, invest¬
ment demand responses to interest rate changes may take longer to build
up than working-capital costs. The implication is that the effects shown
in figure 6.5 (faster money expansion increases both inflation and growth)
will hold in the longer run, although one could scarcely be sure. In any
case, which effect dominates is an empirical matter that remains to be
explored.
The evidence from a few econometric studies seems to show that the
working-capital channel is the one that matters in the short run in develop¬
ing economies, while traditionally Keynesian responses are more important
in the longer run.5 Proponents of deposit rate increases might draw support
from these conclusions since their nostrum works better just under such
circumstances. Nonetheless, the situation remains complex—too complex
for any simple policy prescription along financial liberalization (or other)
lines to be worth serious thought.
Appendix
A more formal stability analysis for the model of this chapter is straight¬
forward, but tedious to work out. A sketch begins with solving the commod¬
ity market demand-supply balance (6.1) for the interest rate,
{[* - (7 + h)]a -g0 - hP]R - (s - h)ab

l = -——-- (o. lu)
(s — h)ab — hR
Note that this equation takes the form:
_ A \R - B
B - A2R
which will generate a positive value for i so long as both R and B are
positive. The sign of the derivative depends on A \ — A 2 or [5 — (y + h)]
A
a — go — h(\ -4- P), the criterion already discussed in connection with

equation (5.8). For given values of the parameters, di/dR > 0 when the
inflation rate is relatively low, from working-capital cost-push. For higher
inflation rates, investment demand effects dominate and di/dR < 0. The
expression (s — h)ab — hR is positive when working-capital costs domi¬
nate and negative when investment demand is more important in determin¬
ing price response. This distinction will be important in the stability analysis
below.
A
A further step in solving (6.1) and (6.2) for i and P is to rewrite the latter
equation as:
R V/y, — ba
i — P — A (6.11)
R(V + 1)
where A is the inverse function of ij/ and has a negative first derivative. It
A
is easy to show that the real interest rate i — P must fall in the loan
market when either the price-wage ratio R or the degree of monetization
V goes up. Let A* = AV[R(V + l)]2 < 0, where A' is the derivative
of A. Then for reference the expressions for the partial derivatives of
(6.11) are:
(6.12)
and
— = A *ba(V + 1) (6.13)
dR
Solving (6.10) and (6.11) for i and P is straightforward. The results are:
. = ft* ~ (7 + h)]° - go + hA}R _ j Uv

(5 - h)ab
and
P _ ([*-(? + hâ ~ ~ A[(s - h)ab - hR] _ ^

(5 - h)ab K ^
When both commodity and loan market equilibrium are taken into ac¬
count as in these equations, di/dR still has an ambiguous sign. The positive
effect comes from the bracketed term multiplying R in (6.14). Equation
(5.13) shows that the bracket just equals a > 0. The negative effect comes
from the derivative of A with respect to R, as previously discussed. The
corresponding partial derivative makes di/dV clearly negative.
Recall that the term [(5 — h)ab — hR] in (6.15) is positive when an
increase in the interest rate drives up prices, and negative otherwise. Under
A
rational expectations, dP/d V exceeds zero under dominant working-capital

effects and is less than zero when the interest effect on investment demand
# , A
is more important. The same effects go into dP/dR, for which the formal
expression is:
bP_
[a - 8A*ba(V + 1)] (6.16)
Jr
where:
8 = (s — h)ab — hR
= (S-~-h)ab {[5 - (h + y)]a -g0 - h(\ + P)] (6.17)

a
Since 8 can be negative (when under high inflation investment demand

effects dominate the interest rate response), dP/dR can in principle take
either sign. However, a positive derivative will soon show up as a partial

sufficient condition for stability of R and V in economic growth.
We are finally ready to deal with dynamics. The state variables change
according to the rules:
V = (.* +,U£
\ss —
- h )
'.V/ji — ba
[ 1R(V + 1)
- i(V,R)
sg o + hya
+ H (6.18)
s — h
and
R = <M* - (6.19)
Equations (6.18) and (6.19) are unpleasantly nonlinear in V and R. For

that reason, we will be content with stability analysis around a steady-state
growth path where both variables are constant. One implication is that the
value of capital stock grows at the same rate as money base, or P + K
= H. Secondly, there will be equilibrium in the labor market at steady state,
or R = R v.
At an initial steady state, the Jacobian of (6.18) and (6.19) is proportional
to the matrix:
{shab + 8) • A*Rba(V + 1) — a
(f> 1 - -^f[a - 8A*Rba(V + 1)]

l dP J.
When 8 > 0 and working capital cost-push is the dominant interest rate
A
effect, the northwest term will be negative. Since aRs/dP > 0, the south¬
west term will be negative when a — 8A*Rba{V + 1) is strongly positive.
A
But from (6.16) this condition just means that dP/dR should be large. The
trace condition for stability will be satisfied for 8 > 0 under these condi¬
tions. Higher inflation or stronger investment demand would make 8 < 0,
but by continuity of the functions stability should still be observed for values
of 8 not far below zero.
The determinant of the Jacobian is proportional to the quantity:
A*R ^1 — + 8
Since A* < 0, a positive determinant is more likely when 8 < 0 and

investment demand effects dominate the macro response to interest rate
changes. When 8 > 0 the condition:
dRs
A
a < 0
dP
is required for stability. After a bit of manipulation, the relevant partial

derivative can be written as:
dRs
A
dP 1
Since a is a fraction, dRs/dP will have to be substantially greater than one

to make the determinant positive. If the zero-inflation value of R , is set to
unity by choice of units, then the numerical illustrations of section 6.2
suggest that with inflation, Rs will take a value between one and two. A bit
of additional computation reveals that the derivative will be large only when
the indexing period X is longer than one unit of time in the model. A
question of choice of time units is clearly involved, but with monthly
inflation rates of, for example, 10 percent, relatively infrequent annual
indexing might be required to assure macro stability when working-capital
cost-push is the main channel for interest rate effects on the system. The
situation is, of course, better when investment demand effects dominate.
7
Trade Balance
Complications
a UITE A LOT has already been said about the commodity trade
entry in the balance of payments. Recall the contractionary devaluation tale
of chapter 2, the travails of agriculture and mineral exporters in chapter 3,
and the empirical wrinkles of chapter 4.
Perhaps surprisingly, trade poses still more riddles, especially about di¬
rections of causality in the macro economy. There are two key equations
for what orthodox theorists call internal and external balance, but about a
half dozen potentially endogenous variables enter into them. Hence, infor¬
mation beyond the equations must be used to specify chains of economic
causes and the corresponding adjustment processes. There is no reason to
expect the model to be the same for all countries or all times.
Section 7.1 sets up internal and external balance equations, or the saving
gap and trade gap as they are known in the development literature. Section
7.2 explains the disequilibrium two-gap planning model interpretation of
the balances, as well as two equilibrium adjustment processes for the rate
of growth. The first has investment determined from saving, while competi¬
tive imports vary to close the balance of trade. The second, more realistic
for application in poor countries, has growth fixed by available capital goods
imports, while the income distribution changes to equilibrate saving supply
with investment demand. Two additional pairs of adjusting variables are

discussed in section 7.3. The exchange rate and capacity utilization as
endogenous variables give a model resembling the mineral exporter of
chapter 3, while endogenous capacity utilization and trade deficit comprise
a system to be used in chapter 8 for analysis of macro stabilization policy.
Detailed though they are, these trade models say nothing about the
capital account of the balance of payments. This omission begins to be
remedied in section 7.4, where interest-bearing capital inflows to a sovereign
borrowing country are analyzed in a two-gap framework. The chapter
closes in section 7.5 with brief observations on the role of exports in trade
models and their possible effects on economic growth.
7.1 Internal and External Balance
To keep the discussion within bounds, we will work with a one-sector

economy, leaving out the interesting interindustry complications of chapter
3. The balance between supply and demand of a nationally produced good
can be written as:
PX = PC 4- PQI + [e0 + el(eP*/P)]PX (7.1)
As in the mineral exporting country of section 3.3, only a fraction 0 of total

investment demand / is produced at home. The remainder, (1 — 0)7, must
be imported. Roughly speaking, the home country is assumed to produce
construction and unsophisticated equipment, while more elaborate capital
goods come from abroad. Secondly, (7.1) incorporates the chapter 3 specifi¬
cation for the share of national production that is exported: E/X = e0 +
€i (eP*/P). As usual, e stands for the exchange rate and P* for the world
price of goods similar to home exports. The export share will have an
elasticity greater than one with respect to profitability eP*/P when €0 is
negative. Also, the world price of “our” exports is P/e, which is essentially
determined by internal factors. For this reason, the models of this chapter
can best be interpreted as referring to a semi-industrialized country. A pure
primary product exporter would have its export price determined from
abroad.
To (7.1) we add a hypothesis of markup pricing,
/> = (!+ T)(wz + ePta 0) (7.2)

Trade Balance Complications 129
where /**() is the world price of intermediate imports and a0 is the corre¬
sponding input-output coefficient. The money wage is w and the labor-
output ratio is z. All wage income is consumed, and a share 5 of profits is
saved, so that the value of consumption is given by:
PC = wzX + (1 — s)t(wz + eP*a0)X (7.3)
Note that there is consumption only of the nationally produced good.

Imports of consumer goods could be considered, but except for one point
raised in the next section they just complicate the presentation. For that
reason—and also because most poor countries ban almost all direct con¬
sumption imports by quantitative restriction—they are left out.
After all this preparation, it is simple to set down the value in domestic
currency of the deficit on commodity trade as:
eB = eP$a0X + eP* (1 - 0)/ - [e0 + e,(eP*/P)]PX (7.4)
where B is the deficit in dollars.1 Insertion of (7.2), (7.3), and (7.4) into (7.1)
gives a condition for zero excess demand of the nationally produced good:
[P0 + eP*( 1 — 0)]/ — st(wz + eP$a0)X — eB = 0 (7.5)
Unsurprisingly, this equation states that in equilibrium, the value of invest¬

ment is equal to national saving (from profits) plus the deficit on current
account.2 Excess demand for foreign exchange (times the exchange rate)
can be expressed as:
[ePtao - exeP* - e0P]X + eP* (1 - 0)7 - eB = 0 (7.6)
Equations (7.5) and (7.6) represent internal and external balance of the
economy, respectively. Two variables must adjust to assure the balances
hold in equilibrium. Various stable marriages are discussed in the following
two sections. In preparation, we follow our usual strategy of dividing (7.5)
and (7.6) by the value of capital stock, PK, to restate them as:
ST
[0 + (1 - ®)qPf]g-« - qb = 0 (7.7)
1 + T
and
[q(P$a0 - e,P*) - e0]u + qP*( 1 - Q)g - qb = 0 (7.8)

where g = I/K is the rate of growth of capital stock and u = X/K

measures capacity utilization. On the trade side, the current account deficit
relative to capital stock is b = B/K, and q = e/P is the ratio of the
nominal exchange rate to the domestic output price. We follow the conven¬
tion of calling q the “real exchange rate,” though at times in the literature
this label is applied to P/e, or q's inverse.
There are five possible variables to adjust in these two equations, viz g,
u, and b on the quantity side and r and q (or P and e) in terms of prices.
The mechanisms can be described more easily if we adopt the “export
optimist” assumption that €0 < 0. or the export share of national output
has a greater than unitary elasticity with respect to the real exchange rate.
Also note that the quantity:
h = q[( 1 - ©)PTg - b] (7.9)
is almost always positive for poor countries: net capital inflow (or the trade
deficit) is less than the value of capital goods imports.3 From (7.8) the
condition A > 0 implies that [q(P*a0 — e,P*) — e0]w < 0- The interpreta¬
tion is that exports exceed intermediate imports, or that the country runs
a surplus on current trade to pay for the portion of its investment imports
not financed from abroad. This sort of trade surplus strongly influences
external adjustment in a semi-industrial country, as will be discussed in
detail shortly.
7.2 The Two Gaps
One useful way of looking at (7.7) and (7.8) is in disequilibrium terms.

To that end, assume provisionally that capacity utilization u, the real
exchange rate q, and the markup r are fixed. What then are the trade-off
relationships between the trade deficit and growth? The relevant derivatives
are
dg _ _ <7
db Int 0 + ?(1 - Q)P*
from the internal balance condition (7.7) and
g
qP*(\ - 0)
from (7.8).
The internal balance slope is less than the one for external balance. A unit
increase in b permits g to go up by more along (7.8) than (7.7) because only
a portion of the commodities required for capital formation is supplied from
foreign trade. This distinction is illustrated by figure 7.1.4
Capital Stock
Growth Rater? External Balance
Internal Balance
or Saving Gap
Ratio of Trade Deficit

to Capital Stock b
FIGURE 7.1
External Balance (Trade Gap) and Internal Balance (Saving Gap) Trade-Offs
between the Trade Deficit and Capital Stock Growth
In a planning exercise, the diagram shows that an increase in the trade

deficit will generate more growth if (7.8) is the constraining relationship
between b and g than if (7.7) applies. Moreover, “something” must adjust
to assure equality in the nonbinding constraint. As will be discussed shortly,
if the external balance relationship or trade gap binds, then either the
markup rate or capacity utilization could vary to assure the equals sign in
(7.7) . If the saving gap binds, an extra import term could bring balance in
(7.8) .
Many planning and development programming models have been based
on this setup since it was proposed in the early 1960s—the simplest variant
is outlined in section 8.4.5 Usually, the models are set up to calculate
feasible levels of investment for given capital inflows. The diagnosis of the
“binding gap” required for such an assessment comes informally from what
is happening in the macro economy. Domestic saving may constrain the
system if central bank reserves or consumption imports are on the rise,
while foreign exchange may be in short supply if capacity utilization is low
or investment falling off.
These quick looks at the system replace specification of adjustment mech¬

anisms in the two-gap approach. This absence of rules about what closes
the “gap between the gaps” by driving left-hand sides of both (7.7) and (7.8)
to zero has often been criticized. To illustrate the conclusions to which the
criticisms lead, two formal processes are set out here.
First suppose that the saving gap binds, so that capital stock growth
g increases when there is “excess supply” of saving. This is the same as
writing the left side of (7.7) with the sign reversed:
— = Fg ((1 + t)~'stu + qb - [Q + (1 - Q)qP*]g\ (7.10)

dt
Also allow competitive imports of nationally produced goods to enter the

economy in volume M. These imports (net of local commodity exports) will
be the variable that adjusts to bring external balance; hence, we ignore the
export supply term (qt\P* + e0)u for the moment. Let m — M/K and
assume that m rises when there is an excess of foreign exchange:
dm
Fm[qb - qP* (1 - 0)g - qP$a0u - m\ (7.11)
~dt
The nominal exchange rate is assumed constant and since there is free
foreign trade the law of one price applies: P — eP*. The markup rate is
determined from (7.2), and the economy is supposed to operate at full
capacity so the u is fixed. (Flexible wages and free capital-labor substitution
may be postulated to back this hypothesis up.) Finally, the allowable trade
deficit b is set by capital market conditions from abroad.
Under all these assumptions, it is easy to see that (7.10) and (7.11) define
a stably adjusting system in g and m. The economics is simple. A unit
increase in b, the trade deficit relative to the capital stock, generates enough
saving to raise the growth rate by q/[Q + (1 — Q)qP* ], from (7.10). But
from (7.11) the extra growth uses only P* (1 — 0) in imports of capital
goods, and the rest of the increment in foreign resources can go to competi¬
tive imports m. What happens when b is reduced and m tends to fall below
zero is left unexplained.
This story clearly focuses on the saving gap—conditions of productivity
and thrift within a country and capital donors’ largesse abroad provide the
saving that determines g. Then competitive imports adjust to keep trade in
balance. The two-gap dilemma is dissolved by free trade. Not surprisingly,
neoclassical trade theorists are the principal exponents of the foregoing sort
of model.6
The main objections center on institutions—few countries permit free

imports, full employment of resources is rarely observed, and the law of one
price is more often than not in abeyance. Moreover, trade flows need not
adjust with much rapidity. If m in (7.11) is negative (or the country is a
net exporter of competitive commodities—the likely case), then in fact it
may be very difficult to raise exports in the short run. The country may well
be quantity-constrained in world markets, and run straight into two-gap
problems.7 Under such circumstances, the natural approach is to take the
trade gap as ultimately binding on the growth rate, and ask what adjustment
mechanisms act to bring national saving in line. The obvious possibilities
are forced saving and output changes. We illustrate the former in a model
where the markup rate r and the growth rate g are endogenous variables.
The differential equations for adjustment are most easily expressed after
multiplying (7.7) and (7.8) by 1 + r. With response functions FT( ) and
Fg( ) they are:
= 7\{(1 + — ®)Pf - b] ~ stu} (7.12)

at
and
— = Fg{qb - qP*( 1 - 0)g

at
- [q(P0*flo - «iP*) - (1 + t)€0]«) (7-13)
where q = e/(wz + eP*a0), or the nominal exchange rate divided by

variable cost. Along with capacity utilization u and the trade deficit b, q
is assumed exogenous. The adjustment story is that the trade deficit deter¬
mines g from the external balance equation (7.13), while the income distri¬
bution shifts through changes in r to satisfy internal balance in (7.12). An
equilibrium can easily be seen to be locally stable when h = q[{\ — 0)
Pfg — b] > 0 and e0 < 0.
Figure 7.2 illustrates the model; the small arrows indicate the stable
dynamics. The two-gap interpretation hinges on two observations. First,
when the saving gap binds, a higher growth rate requires a higher markup
r to the left of the kink at point A. Second, when the trade gap binds to
the right of A, the import content of higher investment provides enough
extra saving to drive r down. The dominance of the trade gap in the overall
specification stems from the fact that an increase in the deficit b gener¬
ates q/q( 1 - 0)P? additional growth from (7.13). The higher growth rate
requires (1 + r)0 + q( 1 — Q)P* in saving from (7.12). The required extra
Markup Rate
Note: The dashed lines show the effect of an increase in the intermediate import price, or an "oil shock."
saving exceeds that provided by a higher deficit alone, and r must rise. In
an alternative specification with capacity utilization as the adjusting vari¬
able, a larger deficit would permit both higher output and faster growth.
As a final exercise, consider what happens under an “oil shock” or
increase in the intermediate import price P* in the two specifications. When
growth is saving-determined in (7.10) and (7.11), it is easy to see that a rise
in P* will cut back on both g and competitive imports m. Figure 7.2 shows
what happens in the forced saving model. The higher import bill represents
additional savings so that for a given r, growth will be slowed from the trade
gap, or that schedule shifts leftward. The outcome is a lower markup and
(from arguments similar to those above) slower growth. In an output¬
adjusting model, capacity utilization would decline from lower aggregate
demand, as in the model of contractionary devaluation of chapter 2. A
lower markup is how the system adapts to fixed capacity utilization.
Regrettably, taxonomy runs rampant in this discussion of the two-gap
model; it often does when foreign trade is the subject at hand. The differ¬
ences in response are substantial enough to suggest that one should be quite
clear about causal linkages in the system before jumping into policy advice.
From a structuralist perspective, a model on which income distribution or
level of output bears the brunt of adjustment looks decidedly more relevant
than one in which freely adjusting foreign trade smooths all external shocks
away. But as we see in the following section, there are further possibilities.
Balance equations alone cannot define what is happening in an economy;

information outside the models has to say what is pushing the algebra
around.
7.3 Exchange Rate and Output Adjustments
Why can’t the exchange rate vary to smooth the two-gap kink away? If
not constrained by the asset market (see chapter 8), in principle it can bear
some part of macro adjustment. But even in commodity markets there are
limits—no plausible exchange appreciation is going to turn Kuwait into a
net importer. Moreover, the macro correlates of exchange rate adjustment
may not be pleasing. To see why, consider a model in which capacity
utilization rises in response to excess commodity demand in (7.7) and the
exchange rate goes up when there is excess demand for dollars (a positive
left-hand side) in (7.8). The growth rate, the markup, and the trade deficit
are assumed to be fixed. This specification is not greatly different from that
of the mineral-exporter model of chapter 3, and addresses the same issues
discussed there.
The details of this model closure show up in the total differentials of
equations (7.7) and (7.8):
ST h
du —q(\ + r)a0h
1 + T <1
+
h
3
o
dq qa0u[ 1 - €0(1 + t)]

u
*
'0 + q( 1 - Q)P* ' g "o"
+ dg - db = (7.14)
q( 1 - &)P* q_ 0
With e0 < 0 (elastic exports, again) the trace of the Jacobian matrix
multiplying du and dq will be safely negative. The same condition assures
a positive determinant A:
hz srue0
A =-
ug (1 + r)q
Adjustment of the Output-Capital Ratio or Capacity Utilization to Reach Internal

Balance and the Real Exchange Rate to Reach External Balance
Note: The dashed lines show the effect of greater foreign exchange availability, as from foreign transfers
or mineral exports.
The adjustment process is illustrated in figure 7.3. The output-capital ratio

u responds stably to maintain internal balance and c0 < 0 guarantees that
the exchange rate does the same for foreign trade.8
An interesting comparative static exercise is an increase in the trade
deficit b. More foreign aid or bigger mineral exports from a purely enclave
sector could generate this pleasant macroeconomic shock. In the trade
balance at a given capacity level, (7.14) shows that an increase in b reduces
excess demand for dollars and consequently the real exchange rate q. Since
a high trade deficit represents additional saving, capacity utilization u will
also fall from reduced aggregate demand with a given q. As shown in figure
7.3, the external balance schedule shifts leftward, while internal balance
moves down.
Clearly, output or capacity utilization drops due to a standard disincen¬
tive effect from additional foreign exchange. From (7.14), (7.7), and (7.9),
the sign of the change in the real exchange rate depends on:
So long as there is some capital goods production at home, the extra

resources brought in by an increase in the trade deficit exceed domestic use,
and appreciation of the real exchange rate is an inevitable result.9
The extra foreign exchange turns out to be not so pleasant after all.
National production declines in the short run, and in a dynamic model the
lower real exchange rate would give scant incentive to investment demand.
But then from (7.14) a lower growth rate would further reduce q, as both
external and internal diversification declined. The long-term analog of our
short-term formal results is a cumulative process in which freely available
foreign resources reduce the real exchange rate, thereby cutting back invest¬
ment and leading to internal stagnation in the long run. Gifts from the gods
of politically determined foreign aid or bounteous exports are not without
their sting.10
Even if these long-run difficulties don’t arise, further problems can come
from the asset market. In particular, if capital flows are not tightly con¬
trolled, they may influence the exchange rate, internal price level, or the
interest rate via the capital account of the balance of payments. These
responses depend greatly on institutions and modes of capital market inter¬
vention. Examples are given in chapter 8. In preparation for the analy¬
sis presented there, one more current account model is set out, in which
the trade deficit instead of the real exchange rate varies to assure exter¬
nal balance. Also, the asset market may play a role in determining the
price level. To capture this possibility, the markup expression r/( 1 + t)
for profits per unit output in (7.7) is replaced by (1 — coz — qP*a0)

where co = w/P is the real wage. The internal balance equation is thus
rewritten as:
[0 + (1 - ®)qP*]g - *[1 - wz — qP$a0]u - gb = 0 (7.15)
The adjustment process is an increase in u when the left side of (7.15)

exceeds zero, while b goes up in response to excess demand for foreign
resources in (7.8). Foreign asset holders are ignored, so the higher trade
deficit is financed by a reduction in national holdings of foreigners’ liabilities
—either central bank reserves or currency and other tradable wealth held
in private hoards.
The total differentials of the balance equations resemble (7.14). The roles
of dq and db are reversed, and dropping the markup term slightly alters
the price responses. For the record, the expression is:
h + ®g
-q du
u
—sP$a0u -|- 0 + qPf( 1 - 0)

q
dq - (7.16)
e0u
-•*
—\
©
"0
1
y
q
The determinant of the Jacobian is Qqg/u, safely positive.

For comparative statics, note first that (7.8) and (7.15) can be directly
rewritten as equations in the ratios (u/g) and (b/g). An increase in g will
raise u and b in the same proportion, or capacity utilization and the trade
deficit are unit elastic functions of the growth rate. The same result of course
can be verified from the heavier algebra of (7.16).
To check the effect of a shift in the real exchange rate, recall the assump¬
tion made in connection with (7.9) that h > 0, or the country has a surplus
of exports over intermediate imports. The term [q(P$a0 — e,P*) — 60]«
in (7.8) must then be negative. Suppose further that e0 < 0, or exports are
more than unit elastic to exchange rate changes. The implication is that
q(t\P* — P*aQ) > 0, or an increase in q drives up the difference between
exports and imports. This healthy trade response raises capacity utilization
since from (7.16):
du u
— = — - P*a0u) + sP%a0u]
dq 0g
In effect, the assumption that €0 < 0 rules out contractionary devaluation

of the type discussed in chapter 2. A positive value of c0—or a less than unit
elasticity of exports to the exchange rate—would permit (e,/** — P*a0) to
be negative, and perhaps du/dq would be less than zero as well. Determined
to be sanguine about trade in the balance of payments analysis of chapter
8, we will there postulate contractionary devaluation away.
Responses are illustrated in figure 7.4. From (7.16), the slope du/db of
FIGURE 7.4
Effects of an Increase in the Real Exchange Rate q on Capacity Utilization and
the Trade Deficit When Devaluation Is Expansionary (Upper Diagram) or
Contractionary (Lower Diagram)
the internal balance or saving gap equation (7.15) will be less negative than
that of the external balance or trade gap (7.8). In (7.15) an increase in
q raises capacity utilization u at a given trade deficit b when h > 0. Under
this condition, along with e0 < 0, a higher q reduces the deficit for a given
level of u in (7.8). The outcome is shown in the upper diagram of figure 7.4
—capacity utilization goes up and the deficit falls. For e0 > 0, b can rise
along with q, as shown in the lower diagram. Devaluation under such
circumstances may worsen both the trade deficit and the level of economic
activity.
If we banish the lower diagram by assumption, then this last model is
quite standard—the exchange rate does the usual things and higher capital
stock growth stimulates aggregate demand. Note, however, that from (7.15)
the profit rate (1 — coz — qP*a0)u initially falls when q goes up. Hence,
capital stock growth might be expected to fall with devaluation or speed up

when the exchange rate appreciates. How such an investment response
complicates stabilization policy in an open economy is discussed informally
in the following chapter.
7.4 Interest Costs and Export Performance
The foregoing models leave out an important contemporary item in the

current account of developing countries—the interest cost on outstanding
debt.11 To see how debt and interest can influence trade behavior of develop¬
ing countries, we can extend the accounting underlying (7.7) and (7.8) by
letting F be the outstanding value of “our” country’s foreign debt, and
f = F/K. The definition of the current deficit in (7.4) has to be written as:
eB — ePtaoX + eP* (1 — 0)7 — e0 -f e, PX + ieF (7.17)
where i is the (real) interest rate on F If reserve changes are zero, then in
the overall balance of payments the current account B will have to be offset
by increases in F, or capital inflows:
—B -f — = —B + gfF = 0 (7.18)
where gf is the growth rate of F. A higher current account deficit reduces

foreign reserves, while a capital inflow (or an increase in foreign nonmone¬
tary debt) adds to them. Finally, national saving decisions determine the
level of consumption, so (7.3) has to be rewritten as:
PC = wzX + (1 — ^)[t(h,z + eP$a0)X — ieF] (7.19)
Equations (7.17)-(7.19) can be combined with (7.1) and (7.2) to give new
versions of the internal and external balance conditions (7.5) and (7.6).
Assume that / is constant so that gf = g, where g is the growth rate of
capital stock. Division of these new balance equations by PK (the value of
capital stock) then permits them to be stated as:
[0 + qP*{\ - 0)]g - s(r - iqf) - qfg = 0 (7.20)

and
-eu + qP*( 1 - 0)g + iqf — qfg = 0 (7.21)
where r = tu/(\ + r) is the national profit rate and e = q(e^P* —

P*ao) + Co is the share of exports net of intermediate imports in total
output. The internal balance equation (7.20) says that investment demand
is met by national saving plus net capital inflows (per unit of capital stock)
qfg. The external balance (7.21) shows that interest payments iqf plus
investment imports qP* (1 — 0)g are balanced by capital inflows and net
exports eu.
To illustrate these relationships, we can return to the planning approach
of section 7.2, treating exports as a variable the country tries to control to
determine its growth rate, given a level of the debt/capital ratio f. The
export ratio can be manipulated by exchange rate policy, as well as tariffs,
quotas, and nonmarket incentives. We do not go into the details of these
interventions here. The fixed “debt-to-equity” ratio / is assumed to arise
naturally from credit-rationing in international capital markets—lenders do
not allow any nation’s obligations to grow without bound.12
Equations (7.20) and (7.21) are illustrated in figure 7.5. In (7.20) for
Effects of an Increase in the Debt/Capital Ratio f on the Capital Stock Growth

Rate g and and Net Export Share e
internal balance, the growth rate is independent of the export share, but its
derivative with respect to / is given by:
dg_ qig - si)_

(7.22)
df _ 0 + qP* (1 - 0) - qf
So long as si (the loss of national saving induced by net interest payments

abroad) is small, an increase in capital inflows permits faster growth insofar
as investment-saving balance is concerned.
In (7.21) for external balance, a higher export share € permits higher
capital goods imports and faster growth. The effect of more capital inflow
is:
dg_ eu — iqP* i 1 — 0) qig - 0

(7.23)
df. [qP*i\ — 0) — qf]1 q[P?( 1 ~ ©) ~f\
The term in the numerator after the first equality is the difference between
the ratio of net exports to capital stock and the interest costs on capital
goods imports. This expression represents the recurring foreign exchange
surplus that can be generated by putting foreign debt to use for capital
formation. Countries without good export prospects could find this surplus
negative; the implication is that they would be unlikely to have access to
capital markets in the first place. For present purposes we assume the
numerator positive so that an increase in debt permits faster growth along
the trade gap, but bear in mind that some economies may not be so for¬
tunate.13 The expression after the second equality shows that having a
growth rate less than the real interest rate on one’s loans may be a sign
of this sort of international impoverishment (under our standard assump¬
tion that financial capital inflows do not pay for physical capital imports:
PH\ -&)>/).
Suppose that the country passes this means test and is eligible for interna¬
tional loans. How does their presence affect the relative slopes of the trade
and saving gaps? Recall from the discussion of figure 7.1 that the trade gap
is usually assumed to be steeper, because the denominator in the term after
the second equality in (7.23) is less than the denominator in (7.22). Note
now, however, that if the interest rate i is high relative to the growth rate
g, the usual condition can reverse and the derivative in (7.23) can be less
than that in (7.22). The reason is that interest charges add to the current
account deficit one-for-one, while they only reduce national saving by the
fraction s. On a long-term basis, an increase in a country’s creditworthiness
as demonstrated by a rise in / may add more to its national saving than to
its available foreign exchange. The effect on the economy would be contrac¬
tionary—in a model with output adjusting to national excess demand, the
outcome would be a lower level of capacity utilization, for example. While

interest costs are probably not large enough in most practical cases to give
such results, the possibility is worth bearing in mind.
Refer to figure 7.5. The growth rate is independent of net exports in
(7.20), so this equation is represented by the horizontal line. There is a
positive relationship between exports and growth in (7.21), as shown by the
“External balance” schedule. A higher debt ratio / permits faster growth
in both equations; both schedules shift up. How do net exports respond?
The usual argument was already made in section 7.1 for the model with
competitive imports. Extra foreign resources add more to growth through
generating saving than through augmenting capital goods import capacity.
There is spare foreign buying power, and net imports can rise or net exports
fall. But once again, this result can reverse if international real interest rates
are high. More capital flows may require a country to step up its export
effort to meet payments on its foreign debt! Figure 7.5 shows this possibility
through the ambiguous shift in e as / goes up. In a diagram like figure 7.1,
the induced rise in exports from higher foreign borrowing will occur pre¬
cisely in the “perverse” case where the saving gap has a steeper schedule
than the trade gap.
7.5 Exports and Growth
There is a final observation on the role of exports in all the models of this
chapter: there is no very clear linkage between the net export share and
other variables of interest, such as the rate of growth. A country at neoclas¬
sical full employment would demonstrate a negative relationship between
exports as a fraction of potential output and the capital stock growth rate.
Physical resources for higher investment would have to be diverted from
net sales abroad, and the foreign deficit would have to increase to provide
the saving counterpart of increased investment demand. If there is spare
capacity, on the other hand, higher exports can stimulate growth, as in the
Hobson-Luxemburg theory of imperialism discussed in chapter 10. And, of
course, there may be shifts between regimes from time to time.
All this stands in sharp contrast to recent statistical exercises in which
GNP growth is supposed to be monotonically related to the (average or
marginal) export share or to the growth rate of exports.14 First note that
a positive relationship between GNP growth and the growth of exports
would simply reflect a constant export share. The resulting regression coeffi-
cient is irrelevant to the question of how changes in the share might be

associated with growth performance. Secondly, the models of this chapter
strongly suggest a nonlinear relationship, shifting over time, between the
two variables. Simple linear correlations fall well short of addressing the
complex interactions between a country’s balance of payments position and
its rate of economic growth.
8
Foreign Assets and the

Balance of Payments
D ETAILED though they are, last chapter’s models say nothing

about how private capital movements may affect the balance of payments.
For developing countries, portfolio shifts between domestic and foreign
assets can have important effects on the real exchange rate, and through it
on inflation and output growth. A model based on recent stabilization
attempts in several countries is developed in this chapter to address these
issues. The specific policy question is the impact of a reduction in the rate
at which the nominal exchange rate is depreciating, or what happens when
a crawling peg is slowed. A three-asset model like the one in chapters 5 and
6 is developed to show that the “normal” financial market response to the
slower crawl would be an upward shift in the schedule of bank lending to
firms. A higher price level (or a lower real exchange rate) or a lower interest
rate and faster growth would be the expected results, as shown in section
8.1. However, the opposite responses could also occur, for example, when
a slower crawl leads wealth-holders to prefer nonmonetized assets (the
“gold” of chapter 5) for purchase of which they can conveniently obtain
bank loans. How these short-run stories work in steady state is sketched in
section 8.2, and section 8.3 reviews in informal fashion several extensions
to the basic model.
An additional point to note is that the crawling peg models really depend
on the two-gap (or internal and external balance) specification of chapter
7, plus the accounting identity for the banking system that says that the
money supply must equal the sum of banks’ foreign assets and their domes¬
tic credit to the government and private sector. The two accounting gaps,
or balances, and the money identity also underlie other approaches to
macro analysis in the Third World. The practice of agencies such as the
World Bank and International Monetary Fund is discussed in section 8.4.
8.1 Asset Markets and the Capital Account
Few developing countries have a floating exchange rate. Several, how¬

ever, revise their nominal rates regularly (every few weeks or months) in
a “crawling peg.” This sort of policy was first used in Latin America in the
1960s. Its original purpose was to hold the real exchange rate steady in the
face of domestic inflation to maintain incentives for private sector exporters.
A second objective was to avoid the massive capital flows that a fixed rate
system stimulates when national prices keep going up. If the exchange rate
is held constant, exports might be expected to stagnate and imports stream
in. Sooner or later a maxidevaluation to restore the current account has to
occur. But before then, anyone with access to financial capital will send it
abroad, via legal channels, if they exist, and illegal ones like overinvoicing
of imports, underinvoicing of exports, and exaggerated interest payments
and royalties. Once the maxi occurs, capital comes back to reap its windfall
increase in local purchasing power. The sequence of minidevaluations that
makes up a crawling peg can substantially smooth these speculative ebbs
and flows.1
More recently, a third objective of a crawling peg has been enunciated
in the context of monetarist stabilization programs. The notion is that the
government should pre-announce the crawl at a rate lower than current
inflation. Inflationary expectations are supposed to be reduced, causing
actual price increases to decelerate after a time.
The policy of trying to wave down inflation with the exchange rate has
usually been applied in connection with the lifting of controls over capital
flows. The results were unexpected—capital came streaming in and led to
rapid accumulation of central bank reserves. Subsequent credit expansion
kicked off additional inflation, while the real exchange rate went down,
taking net exports with it. In other contexts the inflationary push was not
Foreign Assets and the Balance of Payments 147
so overwhelming, but interest rates rose very high, leading to investment

stagnation and slow growth.2
Both sorts of responses can be traced to asset-holders’ reactions to the
slower crawl. The return to retaining foreign assets is /* -f ire — it, where
i* is the foreign interest rate (which we set to zero for convenience), ne is
the expected rate of devaluation, and 7r is the expected internal inflation.
A slower crawl reduces TTe, leading those with wealth to try to shift their
portfolios toward other assets. How the shift occurs and how much it is
multiplied by credit creation in the banking system determine the macro-
economic result. To begin to see the details, look at the balance sheets in
table 8.1.3
Foreign exchange is assumed to be held by two parties—the central bank,
which has dollar reserves in the quantity R (national value eR), and private
individuals, who have quantity eJ.4 Let e(R + J) = eV Under a fixed
exchange rate, in the short run V is constant. If the public wishes to hold
fewer dollars, the central bank must absorb them at the given rate in
TABLE 8.1
Balance Sheets for the Open Economy Model
Assets Liabilities
Central Bank
Government fiscal debt F

Bank reserves H
Foreign reserves eR
Commercial Banks
Bank reserves H
Loans to firms L Deposits from the public D
Loans to public (1 - B )PzZ
Firms
Loans from banks L
Physical capital PK
Net worth N
Public
Deposits with banks D

Foreign exchange eJ Wealth W
Net gold B PzZ
Banking balance sheet: eR + [F + L 4- (1 — B)PzZ~\ = D

Foreign assets: e(R + J) = eV
Wealth with public: W = D 4- eJ + BPzZ
Total wealth: F + eV + PK + PzZ = W + N
exchange for domestic deposits. The total of bank and private holdings (or
V) can only change over time through a current account deficit or surplus.
As in table 5.1, the central bank also holds government debt F, which
sums with eR to give high-powered money H. Along with H, the assets of
the commercial banks are loans to firms L, and loans to the public (7 — B)
PzZ. Regarding the latter, the hypothesis is that, as in chapter 5, the public
holds a nontradable asset, gold (or perhaps more realistically, “land”).
Under existing financial customs, when people acquire Z, they finance a
fraction (1 — B) of the purchase with bank loans. Think of mortgages.
Hence, bank assets include outstanding loans of this type, (1 — B)PzZ. The
public holds the balance, BPzZ. Individuals also have deposits D, a banking
system liability.
The consolidated balance sheet of the central and commercial banks takes
the form:
eR + F + [L + (1 — B )PzZ] = D (8.1)
In words, money supply D is the sum of foreign assets of the banking system
(eR), bank credit to the government (F) and bank credit to the private
sector [L + (1 — B)FzZ]. The definitional decomposition of money supply
into the credit items of the left-hand side of (8.1) is the basis of the “financial
programming” exercises of the International Monetary Fund, as discussed
in section 8.4.
The treatment of firms’ assets differs somewhat from that in chapters 5
and 6. Working capital and associated cost-push are left out for simplicity.
Also, in the absence of an active equity market, firms are assumed to own
themselves. In practice, in poor countries a very restricted part of the
population plus the government will control most enterprises using large
amounts of physical capital, so this assumption is not so far-fetched as it
seems. Its implication is that PK—the value of capital stock—is the total
of firm assets, while their liabilities are loans from banks (L) and net worth
(N). The bank credit finances a fraction a of investment spending; hence,
in the notation of chapter 7, loan demand Ld is:
Ld = a [OP + (1 - 0)eP*]7 (8.2)
The public’s wealth is W, the sum of bank deposits D, foreign exchange

eJ, and gold BP2Z. Total primary wealth in the system includes physical
capital PK, total foreign assets e(R + 7), fiscal debt F, and gold PzZ. It
is easy to see that these sum to W + N. The public’s wealth W will shortly
be shown to be an endogenous variable with a changing value. Firms’ net

worth N is the residual that picks up the slack.
Public demand for assets is described by the equations:
D = v|/(/ — TT, TTe — 7T, TTZ — 7r) W (8.3)

eJ = X(/ — 7T, ITe — TT, TTZ —7t)W (8.4)
and
BPzZ = £(/ — TT, TTe - TT, 7Tz —7t)W (8.5)
where i|/ + X -f £ = 1. The real return to holding deposits is the interest

rate / (very probably controlled by the authorities) less expected inflation
77. Holdings of foreign exchange and gold respond respectively to net ex¬
pected capital gains tt€ — tt and ttz — tt.
To demonstrate the endogeneity of the public’s wealth, we can substitute
equations (8.3)—(8.5) into the definition:
W = D + eJ + B PzZ
If we follow the chapter 5 assumption that the authorities impose a reserve

ratio ju. on the commercial banking system so that H = ju,D (0 < jul < 1),
then wealth turns out to be:
W = --- (8.6)
ja(l - 0 + X(1 - &
Wealth is higher, the higher is the share of assets held in gold (|) and the
lower in foreign exchange (X). The impact of a rise in the reserve require¬
ment ju is to reduce wealth by cutting banks’ credit creation. If gold is
ignored (£ = 0) then (8.6) becomes:
eV + F
W = (8.7)
i - iK1 - p)
which shows that W rises as more deposits are held (higher i|/) and as the
reserve requirement is lower.
Using (8.6) and a bit more manipulation, one can derive an expression
for loan supply to firms as:
= B[1 - Ml - I) - Ml - rt] - l(ey + -

Bind - © + X(1 - JLC)]
= <#>(/ — 7r, 7re — 7r, 7rz — 7r)(eF + F) (8.8)
Available loans rise along with the total primary financial assets (eV +
F), but otherwise depend in a complex way on the workings of the financial
system, as illustrated by the expression after the first equality. For future
reference, all the interactions are boiled down into a single credit multiplier
<f> in the second expression.
When there is no gold, (8.8) simplifies dramatically to the equation:
<Ki - m)
V (eV + F)
1 - ij>(l - p)
The arguments of i|/ are i — tt and ne — 7r, with positive and negative
partial derivatives, respectively. Now suppose that the government pro¬
mises to slow the crawling peg, and is believed. Then ire will fall and i|/ will
rise as wealthholders shift toward domestic assets by trading in foreign
currency at the bank. There is credit multiplication, and the loan supply to
firms goes up. Our first conclusion is that in a very simple financial model,
slowing the crawl with credibility will lead to domestic credit expansion.
When gold enters the picture, the wealthy have another option. After
they convert their foreign holdings to national deposits, they can turn
around and take out loans to finance gold purchases. If the initial deposits
are relatively small (or are restricted by regulations on capital inflow) and
the private loan demands large, then loans to firms will be cut back. Con¬
traction of economically productive credit follows in wake of the slower
crawl.
To see the details, assume that di, = —yd\, with 0 < y < 1. The new
parameter y is the fraction of their reduction in foreign exchange holdings
that members of the public direct toward gold. Its magnitude will depend
on both asset-holders’ behavioral patterns and the institutional restrictions
that the authorities place on the exchange market (for example, minimum
maturity requirements on foreign loans, interest rate premia, and so forth).
Differentiation of (8.8) shows that loan supply to firms will contract when
the following condition is satisfied:
[p. + A(1 - p)]y + £(1 - p)
PT + 0 - p)
As y approaches one, this expression simplifies to:
B < + (A + £)
so that an easily accessible mortgage market (B < 1) and high reserve

requirements (ju, near one) go together with high gold or land demand (y
near one) to squeeze loans to firms.
The easiest way to check the macroeconomic consequences of these
financial adjustments is to set excess demand for productive loans [the
difference between (8.2) and (8.8)] to zero, after dividing by PK. The re¬
sult is:
a[0 -I- (1 — &)qP* ]g

aV 4- (F/P)
— <f)(/ — 7T, TTe — 7r, 7Tz — 7t)-- = 0 (8.9)
K
where <fi is defined in (8.8). The foregoing arguments can be summarized

by the observation that d<f>/dne can have either sign.
The financial market model just sketched must be coupled with a descrip¬
tion of the real side of the economy to say something about macro policy.
We will use the second model of section 7.3 for this purpose. Its balance
equations are (7.8) for the trade gap and (7.15) for the saving gap. Compara¬
tive static responses are summarized by equation (7.16) and figure 7.4, with
the real exchange rate q and the capital stock growth rate g as shifting
variables, and capacity utilization u and the current account deficit b
endogenous. A credit market equation like (8.9) can be used to determine
either g or q. A change in q = e/P must take place via a change in P since
at any moment the level of the nominal exchange rate e will be fixed, though
of course e rises in a predetermined fashion over time according to the
crawling peg. Also, real side macro adjustment occurs solely through
changing capacity utilization, the real wage having been assumed fixed in
(7.15). The underlying labor market story is that the money wage is instan¬
taneously indexed to the price level.
There are two obvious possibilities for the adjusting variable in (8.9). One
is the interest rate, which could rise in response to loan excess demand. The
effects would be to increase bank deposits and thus loan supply through
credit multiplication and to reduce investment demand g. Both responses
would give a stable adjustment.
A second, more monetarist approach is to assume that the price level
P is determined by money market conditions. This line of analysis presup-
poses that prices are determined neither from markup rules as in chapter
2 nor from trade arbitrage between domestic and competitively imported
goods, which would enforce the Law of One Price. Under either hypothesis
the price would be fixed in the short run, and changes in money or credit
supply would spill over into the balance of payments, as in the International
Monetary Fund financial programming models discussed in section 8.4. The
IMF in fact follows what is usually called the monetarist approach to the
balance of payments, and our determination of the price level here by
internal monetary conditions would be considered heretical (but perhaps
acceptable to monetarists if current account foreign transactions are as¬
sumed constricted by protection). Be that as it may, we are after a structur¬
alist story under circumstances where money creation may be inflationary
in any case.
With the price allowed to vary, it is reasonable to assume that it will
increase when there is an excess supply of loans in (8.9), in a case of financial
credit chasing producer’s goods. With a pegged nominal exchange rate at
any point in time, this adjustment amounts to postulating that the real
exchange q — e/P will rise with loan excess demand. The process will be
locally stable when 0 > 0, or some share of investment demand is satisfied
with goods produced at home.5
Suppose that loan supply (f> goes up with the slower crawl. Then the
alternative adjustment mechanisms would lead to either lower interest rates
and faster growth, or rapid price increases and exchange appreciation
(lower q) after the policy shift. Alternatively, if loan supply goes down, the
outcome could be high interest rates and investment stagnation, but a drop
in inflation and real devaluation (with a consequent push to exports). De¬
pending on interest rate restrictions, market dynamics, and other considera¬
tions, a combination of interest rate and price response could of course
happen. By and large, the symptoms of credit expansion were observed
during stabilization attempts in Argentina, Uruguay, Israel, and Spain in
the late 1970s, but the reverse was true for Chile.
8.2 Dynamics of Exchange Adjustment
To tell a fully dynamic story about changes in foreign exchange holdings

and the inflation rate over time, we have to settle on one of the two
short-term adjustments just discussed—interest rate or price level variation.
The analytics is easier and exchange rate dynamics more interesting with
the latter, so assume that the price level (or real exchange rate) varies to
assure loan market equilibrium in (8.9), while the interest rate is held
constant by bank regulation. The capital stock growth rate g is supposed
fixed for the moment, from singularly stable investment demand. Finally,
the national component of the money base is ignored. In table 8.1 F is set
to zero and the central bank’s only assets are foreign reserves eR. Similarly,
the term (F/P) disappears in (8.9).6
The second model closure in section 7.3 now comes into play since it
shows how b (the trade deficit divided by the capital stock) is determined
by g and the real exchange rate q. This information is of interest since total
foreign assets held by the central bank or the public can only change from
the trade account:
It was noted in connection with (7.16) that b is proportional to g in the

model at hand. Set b — rjg and substitute in the above equation to get:
^-= - VgK (8-10)

dt
where
„*a _ 0x _0[(Po*Qo - e,P*)<7 - c0]_

V " q[s(l - a>b - qP%a0) + q{Pta0 - €,/>*) - €0]
This expression is dreadful enough to leave out of the following discussion,

but do recall our foreign trade optimists’ assumption that real devaluation
helps the current account, or dr)/dq < 0. This hypothesis is important since
from the loan market balance (8.9), an increase in V lowers q. But then
7] will rise, and from (8.10) dV/dt < 0. Foreign assets demonstrate a stable
adjustment around an equilibrium point.
The next step, routine in the circumstances, is to assume that the econ¬
omy is initially in a steady state, V = K = g. Putting this condition
together with (8.9) and (8.10) gives an expression in the real exchange rate,
—q<f>r) = a[0 + P?(l — ©)#]g (811)
which determines a q consistent with steady state. The minus sign on the
left side implies that 17 < 0. Since the model omits foreign aid, emigrant
remittances, and similar transfers from abroad, the only way the home
country can maintain a growing money supply is to run a trade surplus.
Now add two more pretty incredible assumptions. First, expected infla¬
tion rates in the prices of commodities and gold (v and tt2) are static.
Second, all agents really believe that the authorities will stick with their
announced exchange rate crawl, so that ne = e is given.
If e is reduced under these circumstances, the “normal” response in (8.9)
is that (f> should rise. Then the real exchange rate has to appreciate from
(or drop below) its steady-state value q to maintain equilibrium in the loan
market. Since 017/3q < 0, the trade surplus will diminish, and V/K will
fall over time. The real exchange rate will gradually rise back toward q,
implying that price inflation P will be slower than the crawl rate e. The
dynamics are shown in figure 8.1. The “loan market” schedule in the
FIGURE 8.1
Adjustment of Foreign Asset Holdings to a More Slowly Crawling Peg under
Static Expectations
Note: The supply of loans to firms rises when the crawl is slowed. When the policy shift is (credibly)
announced, the real exchange rate jumps down from A to B, and than gradually rises from B to C as
domestic inflation declines below the rate of crawl.
diagram corresponds to (8.9). The initial steady state is at point A, with the
exchange rate q determined from the “ V = g" line corresponding to (8.11).
A decline in the crawl rate e increases loan supply <$>qV/K. At the initially
given level of V/K, the real exchange rate must shift down to maintain loan
market equilibrium—q jumps from A to B as the loan market schedule
adjusts. The trade deficit worsens, and q gradually rises back to q at point
C, with a lower steady-state foreign exchange/capital ratio V/K.
To summarize, if everybody believes in it, a slower crawl will first kick

off a price increase and real exchange appreciation, followed by a period of
concomitant trade deficits, declining inflation, and persistent real devalua¬
tion, until a steady state is attained. In Argentine practice, at least, the first
stage occurred, but citizens ceased to believe in the second-stage adjustment
before it reached consummation. The Chilean mirror image—real devalua¬
tion followed by steady appreciation (mellowed for profit recipients by wage
repression and unemployment)—may work better in practice, though it
presupposes export optimism and (as shown shortly) unstable exchange
dynamics in the long run.
Much the same occurs under other assumptions about expectation forma¬
tion. For example, under fashionable rational expectations, ne = e, tr =
A A
P, and 7tz = Pz. If we ignore details of the gold market to keep the problem
tractable, then we can look at dynamics in (8.9) with the arguments of tf>
restated as / — e + q, q, and Pz — e + q. These substitutions transform
(8.9) into an implicit differential equation for q as a function of q, with e
as a shift variable. In the case when a slower crawl boosts loan supply, the
first partial derivative of <j> will be positive and the second, negative (with
absolute magnitude smaller than the first, from an assumption that assets
are gross substitutes). An increase in q will drive up <f>, and from (8.9) will
reduce q. Inverting this relationship makes q a declining function of q, or
(8.9) defines a locally stable adjustment process of adjustment for the real
exchange rate.
The dynamics of adjustment for (V/K) and q are shown in figure 8.2.
FIGURE 8.2
Rational Expectations
Note: The supply of loans to firms rises when the peg is slowed.
There could be oscillations, but if half-cycle convergence is assumed (as

shown) there will be a spell of real exchange appreciation after the crawl
is slowed, followed by depreciation back to the steady-state value q at which
q — 0. The country as a whole will shed foreign assets along the way,
though central bank reserves may rise as individuals shift out of foreign
exchange.
In the case where a slower crawl contracts loan supply, (8.9) makes q an
increasing function of q. The outcome is saddlepoint dynamic instability,
as shown in figure 8.3. All trajectories share rational expectations, but only
FIGURE 8.3
Rational Expectations When the Supply of Loans to Firms Rises
Note: The adjustment process demonstrates a saddlepoint instability. To reach the trajectory converging
stably to C, the real exchange rate must jump from A to B immediately when the crawl is slowed.
the one linking points B and C converges to a steady state. The other paths
diverge in a speculative bubble of some sort. If, after the crawl is slowed,
the government, speculators, or the gods assure that the real rate jumps
from A to B, then there will follow steady appreciation, trade surpluses, and
convergence to point C in the long run. Otherwise, the clock just unwinds.
8.3 Extensions and Reality Checks
Like the analysis of chapter 6, figure 8.3 suggests that one should take
long-run arguments involving formation of expectations with a grain of salt.
Real people are always capable of changing their minds and mucking up
the model makers’ conceits. Even when they aren’t supposed to be following
acrobatic saddlepaths to steady state as in figure 8.3, sensible folk revise
expectations in response to both economic and noneconomic information.
Their changing opinions can substantially alter the policy regime. For
example, politically powerful groups, such as Argentine exporters, found
scant profits during the course of a stately real devaluation along the lines
of figures 8.1 and 8.2. Use of an overvalued peso in an attempt to control
inflation led inevitably to a squeeze on local industry. Capitalists revolted,
there was a run on the peso, and a maxidevaluation was the final result.
Monetarist rationalization of a slow crawling peg was faulty—it buckled
under current account duress.
Besides these exchange rate effects, monetarist stabilization is prone to
other ills. Working capital cost-push along the lines of chapters 5 and 6
could be added to the present analysis, with predictable but relevant re¬
sults.7 A more interesting twist can be based on the observation that in both
Argentina and Uruguay a slower crawl was initially followed by an invest¬
ment boom and a rapid approach to full employment. Thereafter, the
economy went in a cyclical trough and recession accompanied the export¬
ers’ blues.8
The initial investment burst follows naturally from (7.15), where initial
appreciation of the real exchange rate can be seen to lead to an increase in
the profit rate (1 — o)z — qP*a 0). Higher investment demand would be
a natural consequence. Thereafter, profit squeeze along the lines of section
2.4 could set in, acting together with lagging exports to exacerbate a reces¬
sion. The reader is encouraged to work out the details.
8.4 Real and Financial Programming
The final issue to be discussed in connection with the balance of payments

is how the real hardhats go about their work of maintenance and repair. We
refer, of course, to the operatives of the World Bank and International
Monetary Fund, whose preferred tools are, respectively, the saving and
trade gap equations (7.7) and (7.8), and the money supply decomposition
(8.1). Note that these equations are virtually accounting identities. Working
with the simplest possible models is the professionals’ stock in trade.
Most policy-related work with the balance or gap equations is based on
using them independently to get alternative predictions of the trade deficit
for a given rate of growth (or the other way round). The parameters in
these computations are usually derived by a combination of guesswork
and simple regression equations between such variables as intermediate
imports and GDP. The regressions may be run over time spans of a very
few years because those are the statistical degrees of freedom that the data
permit.
If the two gaps or trade deficits come out about the same in the projec¬
tions, then at least that bit of consistency is something to rely on. It can be
tested by alternative specifications of input coefficients, saving rates (includ¬
ing that of the government, omitted for simplicity from the models here),
capacity utilization, and the rate of capital stock growth. Evidently, (7.7)
and (7.8) will yield determinate solutions for b and g. The question is
whether these projections are consistent with available finance for the trade
deficit, aspirations about growth and income distribution, anticipated
changes in the real exchange rate, world prices, and the parameters. The
equations are useful devices for organizing country experts’ thoughts
around these issues. The widely used RIMSIM, or Revised Minimum Stan¬
dard Model of the World Bank (unfortunately undocumented in the aca¬
demic literature), amounts to (7.7) and (7.8) programmed out in enormous
national and capital payment (interest and amortization of loans) account¬
ing detail. Its simulations often seem to call for cutting imports (and capac¬
ity utilization) or maintaining profits. But such results may not be surpris¬
ing when coming from an institution that makes shibboleths of a virile
balance of trade and fast growth.
The Fund and the central bank governors it monitors go through similar
exercises on the basis of (8.1), but with conclusions even more predictable
than those of planning ministers and the World Bank. The IMF model can
be interpreted as an applied version of monetarist balance of payments
theory (though chronologically it came well before the academic explosion
of the 1970s) and as such is documented in the journals.9
Fund-style financial programming adds the familiar equation of ex¬
change,
DV = PX (8.12)
to the accounting identity (8.1), slightly revised as:

eR + F + LP = D (8.13)
/
in which LP is total bank credit to the private sector. The other new symbol
is V (redefined from last section), the velocity or turnover rate of money:
_ Value of transactions PX
Money stock D
In planning financial developments, an output projection (even the two

gaps) will give X, and notions about cost-push or the Law of One Price give
a number for P. Velocity is not supposed to change much from historical
values; hence, one has a value for V “Required” money supply follows
directly from (8.12).
Now plug this value for D in the right side of (8.13). An improvement
in the trade balance—the Fund’s real concern—will show up as an increase
in eR. How to accommodate this reserve increase within the projection for
money demand? If credit to the private sector is maintained, then evidently
bank credit to the government has to fall. If, for political reasons, it can’t
decline enough, then the public has to tighten its belt as well. Finance for
working capital and investment is sacrificed to the projections for money
demand and trade surplus.
A letter of intent from a country to the Fund (but more or less dictated
by the latter) will contain specific limits on increases in F and Lp. Just to
make sure that the trade balance gets better, a devaluation may be thrown
in. And wage repression will help shift the income distribution toward
capitalists, thus generating the saving that under the model’s implicit as¬
sumption of stable capacity use will turn automatically into increased in¬
vestment and growth.
As in all short descriptions, there is an element of caricature in what has
just been said. The perceptive members of the Fund (and Bank) staff con¬
tribute more to policy than economic mystification based on accounting
tricks. Nonetheless, the average mission falls well short of what might be
desired. All the topics discussed in this book—contractionary devaluation,
working-capital cost-push, unexpected asset market responses to exchange
rate manipulation, interindustry transmission channels, and terms-of-trade
changes—lead one to believe that trying to stabilize or regulate an economy
by routine application of accounting identities is going to hurt more than
it helps.
At least two problems arise with overly simplistic models. First, the
formal analysis here suggests that orthodox policies all act the same way,
and their simultaneous application amounts to overkill. A contractionary

devaluation on top of stagflation from working-capital cost-push does not
allow much room for keeping up investment, operating near full capacity,
or reaching an equitable income distribution. Second, there are enough
variables in the accounting equations to ensure that developments will never
turn out as planned. The emphasis on different possible adjustment mech¬
anisms in this chapter and elsewhere precisely underlines the uncertainty
that policy makers always face.
Fund efforts at stabilization have not been remarkably successful—a
point recognized even from within.10 Less dogmatic and more elastic appli¬
cation of their accounting tools by staff members might help improve their
performance in years to come.
9
Two-Sector Models
of Inflation, Distribution,
and Growth
T HE MODELS of the last four chapters went into great detail regard¬
ing finance and the balance of payments but said nothing about how link¬
ages between sectors may affect the pattern of economic growth. In this
chapter, the emphasis switches to intersectoral complications, under the
hypothesis that money and trade adjust to whatever happens in the real
domestic economy. Postulating “passive money” under inflation has a long
tradition in structuralist macroeconomics, and is adopted here to keep the
discussion within bounds.1 A foreign sector could be added to the chapter’s
two models without much difficulty, but again to keep things relatively
simple this extension is left for the interested reader to pursue.
The first two sections are devoted to interactions between a food-produc¬
ing sector and the rest of the economy, along the lines of chapter 3. In
section 9.1, a model is set up to describe inflation resulting from conflicting
claims to product between workers and profit recipients under conditions
of lagging food supply. (The assumption of passive or accommodating
monetary policy clearly underlies this approach.) Thereafter, section 9.2

contains an analysis of how food and nonfood sectors may expand together
in the long run, as usual interpreted as a constant growth steady state.
The second model has two sectors with markup pricing and excess capac¬
ity, as opposed to the agriculture/nonagriculture dual system in which food
price adjusts to clear its market. Demands for “wage” and “luxury” goods
differ between classes, and the key question is how policies aimed at chang¬
ing the income distribution would affect patterns of intersectoral resource
allocation and growth. The short-run model appears in section 9.3, and the
growth story follows in section 9.4, where it is shown that the growth rate
can move either up or down after redistribution.
9.1 Inflation from Lagging Food Supply
As in chapter 3, suppose that there are two sectors in the economy,

agriculture and nonagriculture. We are interested in how inflationary pro¬
cesses can be set off from lagging food supply. To that end, the price Pa for
the agricultural sector is assumed to rise when there is excess demand:
Pa = Ba[P.{Xi - Xa)) (9.1)
where Xda is the current level of demand for food, Xa is supply, and B„ is
a constant describing speed of adjustment. Equation (9.1) says that the food
price responds to the value of excess demand Pa{Xda — Xa)—little would
change in the analysis if it were assumed to respond to demand volume
Xda — Xa instead. Finally, food supply Xa is assumed fixed, as in chap¬
ter 3.
The price Pn in the nonagriculture sector is determined by a markup
Pn = (1 + T)wbn (9.2)
where r is the markup rate, w the money wage, and bn the labor-output
ratio. There is spare capacity, and output Xn rises to meet excess demand:
Xn = Bn[Pn(Xdn - X„)] (9.3)
Again, the demand level Xd follows from aggregate income and prices.
Two-Sector Models of Inflation, Distribution, and Growth 163
And, as usual, the difference between investment and saving (/ — S) is

equal to the sum of excess demands in the economy:
I - S = PflXda - xa) + PflXi - X„) (9.4)
The key factor in inflation dynamics is the response of wage-earners to

a fall in real income caused by a rise in food prices. For rigor, their real
purchasing power should be defined in terms of the true cost of living index
associated with the utility function underlying their demand equations for
food and nonfood.2 For present purposes such sophistication is pretentious;
we simply assume that workers react to a price index of P of the form:
P = PaaPnx~a (9.5)
The weighting parameter a is usually taken as the food share in the con¬
sumption budget in some base year.
To make use of (9.5), add a definition of the agriculture/nonagriculture
terms of trade cf>:
<f> = ~ (9-6)
*n
Then if the real wage is defined as &> = w/P, it is easy to show that
a) = [cf>a(l + T)bn]~l (9.7)
In words, the real wage is a decreasing function of the terms of trade cf>, the
markup rate r, and the labor-output ratio (inverse of productivity) b„.
The next step is to build this trade-off between the real wage and terms
of trade into a model of inflation when claims to food output exceed the
available supply. The approach here differs somewhat from the profit-
squeeze dynamics of chapters 2 and 6, which amount to a fairly complex
(though realistic) theory of how worker-capitalist conflict can underlie a
cyclical approach to steady-state growth under inflationary conditions.
Here the emphasis is on inflation resulting from workers’ attempts to main¬
tain their real consumption standards when food is a major component of
their expenditure basket. The appropriate hypothesis is that if the real wage
falls below some target level a>*, then workers push for increases in the only
price over which they have any influence, the money wage. But from (9.7)
the real wage target co*, can be used to define a target terms of trade, <f>*.
Hence, the equation for changes in the money wage becomes:
to = Pn = Bw(4> - <f>*) (9.8)
where price inflation Pn follows from (9.2). This equation asserts that wage
and price inflation are more severe when the terms of trade shift in favor
of agriculture. For many developing countries, (9.8) appears to be an accu¬
rate rule.3 Putting it together with the rest of the model leads to an expres-
A A A
sion for the change in the terms of trade, (f> = Pa — Pn:
$ = B.[(/ - S) - Pn(Xi - X„)\ - B„(<f> - 4>*) (9.9)
The inflationary process can easily be illustrated as in figure 9.1. The

“/ — S" schedule shows levels of output X„ and terms of trade <J> consistent
with macro equilibrium in which I — S = 0 in (9.4). At a given level of
<p, investment exceeds saving below this line (under the standard stability
condition that an increase in Xn stimulates saving supply more strongly
A
than investment demand). Similarly, the = 0” locus shows equilibrium

in the nonagricultural market. From the discussion in chapter 3, this curve
can slope either way and demonstrates a low income elasticity for food
demand in figure 9.1. As the arrows show, output X„ falls when supply
FIGURE 9.1
Determination of the Equilibrium Nonagricultural Output Level and Terms of
Trade in a Model Where Inflationary Pressure Arises from a Low Real Wage
exceeds demand above the market equilibrium line. Finally, the vertical line
at <j>* indicates the level of terms of trade at which there is no wage inflation.
The next step is to characterize the locus “(f> = 0,” along which the terms
of trade stay constant. To begin, suppose the economy is at point A, where
<f> = <f>* and Xdn = X„. However, investment exceeds saving, so that from
(9.9), the terms of trade will rise (as illustrated by the arrow). At B, by
contrast, cf> < 0 since investment equals saving and the nonagricultural
market is in equilibrium, but (f> > <f>*.
With opposite directions of motion for <f> established at A and B, it is easy
to imagine moving along the line between them to some intermediate point
Q at which <f> = 0. Here the wage push from high terms of trade is just offset
by excess demand for food and <p stays constant.
Much the same analysis can be applied along the “/ — S” locus to get
another stationary point for <f>. At C, cf> is rising since there is excess food
demand (which follows from (9.4), where I — S = 0 and there is excess
nonagricultural supply). Hence, there will be a point R between B and C
at which <f> is constant. Connecting Q and R is the curve “<j> = 0” for stable
terms of trade. To its left <J> rises and to its right cf> falls.4
Given this adjustment dynamics, it is clear that there is a stable equilib¬
rium at Q, with the nonagricultural commodity market in balance and
unchanging terms of trade. However, as argued above, there will be excess
food demand, and from (9.1) the food price Pa will rise at a constant rate.
Given the constant terms of trade, w and Pn will increase at the same speed
as well.
The model thus generates persistent inflation. Of course, if workers were
willing to accept a lower target wage, the rate of price increase would be
reduced. In figure 9.1 <f>* would rise and Q would shift to the right, lowering
excess demand. For typical levels of living in poor countries, such meekness
on the part of labor is improbable, unless imposed. Without wage repres¬
sion, a more promising policy is to increase food supply. The impact would
be to shift the “/ — S” locus down and B to the left. Point Q would also
shift left, lowering the rate of wage inflation. Extra food could come from
imports or (better) more efficient agriculture. Land reform has often been
proposed to this end.5
Finally, note that the present model at best applies to the short run since
it generates persistent macro disequilibrium. In any functioning economy,
the observed market responses would be reductions in food stocks and
queues. The latter might persist, but stock reduction, imports, and similar
adjustments guarantee that the national accounts will never register an item
for “excess demand.” If only for this trite statistical reason, the model
would be difficult to fit to data. In practice one would have to assume that
the food price responds much more rapidly to excess demand than does the
money wage to shortfalls of the real wage below its target. In symbols, Bfl
would be much larger than B* and the stable inflation point Q would be
close to the macro equilibrium B in figure 9.1. Having gone this far, it is
simpler to assume that the food price varies to clear the market in the very
short run (say a period of days), while wage inflation persists until the real
income shortfall is closed, as in (9.8). With a homogeneous model like the
one developed in chapter 3, prices would also increase at the same rate as
the wage. Long-run policies that may affect rates of both price and output
increase are discussed in the following section.
9.2 Growth in an Economy with an Important Food Sector
To set out the details of growth in a food-producing economy, a precise

description of short-run excess demand patterns in required. Since it is
ready at hand, we use the model of sections 3.1 and 3.2 here. The two
relevant equations are the investment-saving balance (3.8) and the market
equilibrium condition for food (3.10). The nonagricultural output level
X„ and the food price Pa are supposed to vary to assure both macro and
food market equilibrium in the short run.
To study economic expansion, it is useful to rewrite these equations in
terms of growth and profit rates. Let gn = IJKn and ga = Ia/Ka be
growth rates of capital stocks in the two sectors (where the investment levels
/„ and Ia sum to / in (3.8)). Profit rates in the N-sector and A-sector,
respectively, are:
and
PaXa Pa
a
In the definition of ra the assumption that labor and capital incomes are not
separated in agriculture is maintained, so that the profit rate can be deter¬
mined on the basis of the total value of output PaXa. With a fixed output-
capital ratio a in the food sector, ra turns out to be proportional to the terms
of trade <f>. As we will see shortly, high agriculture profits and inflation go
hand in hand.6
With this much notation set out, the investment-saving and food market
balances can be written as:
\go + g» - Xra[sa - (e/a)] - snrn = 0 (9.10)
and
1 + T ®/Ka
-o.y„r„ + \ra (1 - aya) = 0 (9.11)
T a
Besides the N-sector markup rate t, the parameters carried over to these
equations from chapter 3 include a and 0 from the demand functions (3.4)
and (3.5), and the consumption propensities ya and y„ defined in connec¬
tion with (3.3). Two new symbols are X = Ka/K„ and € — E/Ka. The first
measures the relative size of the two sectors, while e indicates the impor¬
tance of agricultural exports relative to the sector’s capital stock.
Equations (9.10) and (9.11) almost describe a tractable growth model, but
two simplifications are still required. First note that if 0 is constant, then
the ratio Q/Ka will steadily decline as the economy expands. Since 0 > 0
when food demand is income inelastic, a decreasing ratio of 0 to Ka means
that Engel effects become less important as growth proceeds. On the other
hand, a steady state can never be reached so long as &/Ka is changing. For
that reason, it is simplest to set 0 to zero and assume that demands for both
goods have unit income elasticities. This is standard practice in growth
models. A more realistic alternative (which the reader might want to pur¬
sue) would be to assume that 0 grows at a constant rate to represent
population pressure.
The second simplification is to set the two sectoral saving rates equal:
sa = s„ = s. This assumption is mostly adopted to shorten lengthy alge¬
braic expressions and could be relaxed.
We first describe how the economy behaves in steady growth, and then
briefly discuss stability. At the steady state, set ga = gn — g. Then (9.10)
and (9.11) can be solved to get the two sectoral profit rates as:
1 + X a[l + (1 — s)r]_
r, g (9.12)
X A
and
- 0 + X)r[. - aC - ,) - («/«)], (9.13)
where
A = ,s(t + a) — (e/a)[a( 1 + (1 — j)t) + sr] (9.14)
Note that ra declines and r„ rises as X (or Ka/Kn) goes up. A shift in
capital stock toward agriculture lowers the profit rate there and raises
returns in the nonagriculture sector. We will shortly specify an investment
function that responds to a difference between the two profit rates strongly
enough to drive them to equality. When ra = r„ — r at steady state, the
corresponding value of X is given by:
q[l + (1 - *)t]
(9.15)
t[1 + a(l - s) — (e/a)]
Contemplation (or differentiation) of (9.15) shows that the steady-state

capital stock in agriculture will be relatively large when a is high (a large
consumer budget share for food), e/a is high (large agricultural exports),
s is low and r is high. Both supply and demand factors enter into the latter
two results. Finally, X is independent of the steady-state growth rate g,
though from (9.12) and (9.13) it is clear that profit rates increase along with
growth.
The equilibrium is shown in diagrammatic form in figure 9.2. Its stability
follows if we postulate that the agricultural growth rate is fixed at g from
technical considerations—food supply is the limiting restriction on expan¬
sion of the economy. Under such circumstances, a plausible investment
function for nonagriculture would take the form:
gn = g + z(r„ - ra) (9.16)
N-sector entrepreneurs are assumed to aim for balanced expansion of their

capital stock at rate g to maintain a normal level of capacity utilization.
However, extra high profits rn (relative to agricultural profits ra) will stimu¬
late investment demand.
The growth rate of the capital stock ratio is X = ga — gn = g —
g„. Substitution from (9.16) gives the equation:
^ = *(ra - r„) (9.17)

FIGURE 9.2
Determination of Steady-State Values of the Profit Rate r and Capital Stock
Ratio X for a Given Food Sector Growth Rate g
Note: There will be wage inflation from low real purchasing power of workers when r lies above r*.
to describe medium-run adjustment around the steady-state growth path.

From figure 9.2, when X exceeds the steady-state value X, then ra < rn and
A —
X < 0. Thus X represents a stable equilibrium. At an initial steady state,

an increase in the agricultural budget share a, for example, will raise ra and
lower r„ from (9.12) and (9.13). From (9.16) investment demand will shift
away from the N-sector until a new steady state with a higher value of X
is reached.
Now consider inflation. From (9.8) money wages will increase so long as
the current terms of trade <f> lie above a critical value <f>*. But since r0 is
proportional to <f>, there will also be inflation when the steady-state profit
rate 7 lies above some value r*. Figure 9.2 depicts such a situation. To ask
how the inflation problem might be resolved, it is useful to look at an
explicit solution for r, as follows:
r[l - (e/a)] + a_
r -:-S (9.18)
Note that when e/a = 0 (no food exports), then r — g/s. More generally,
(9.18) is a standard formula for forced saving models, showing here that the
profit rate must rise to generate additional saving when investment demand
to support faster agricultural growth goes up. This result carries directly
over to steady-state growth from the short-run models of chapter 2.
The lesson of (9.18) is that an increase in the saving rate s will permit
a lower long-run profit rate and less inflation when the economy is limited
by growth (and investment requirements) in one sector. Public sector in¬
vestment financed by taxes on profits could do the trick. More rapid agricul¬
tural expansion, on the other hand, would worsen inflation by putting
pressure on profits and the terms of trade. The profit rate would fall with
faster food supply growth only if it came from an increase in agriculture’s
output-capital ratio when exports were positive. Under the circumstances,
the export effort would be less costly in terms of domestic resources and
inflation would drop off.
A final policy to be considered is reduction of the markup rate r, for
instance, from industrial price controls. From (9.7), any target real wage
would be consistent with higher terms of trade, hence r* in figure 9.2 would
shift up. This change might or might not be offset by other macro impacts
of a change in t. In section 3.2, it was observed that the agricultural price
Pa can respond either way to a higher markup in the short run. Differentia¬
tion of (9.18) shows that the same conclusion carries over to the steady-state
profit rate. With high enough saving rates, a fall in r will reduce r. If
agriculturalists and profit recipients save a lot, then a reduction in the
markup that the latter receive will permit a lower profit rate and less
inflation in a resource-limited long run. A plausible justification for urban
price controls could be developed along these lines.
9.3 Effects of Income Redistribution in an Industrial Economy
Results of the same type as those just developed carry over into other
models with one fixed-price and another flexible-price sector. For example,
one can extend the analysis of internal and external diversification of the
mineral exporting economy of chapter 3 in fairly straightforward fashion
to the long run. This story is not pursued here to save space. Rather, the
discussion in section 2.2 of the possible expansionary effects of income
redistribution is extended to an economy with two sectors, respectively
producing wage (subscript w) and profit recipients’ (subscript p) goods.7
A system in which food or intermediate input shortages are not restrictive
is assumed; hence, there is excess capacity and markup pricing in the two
sectors according to the rules:
/»* = (1 + r„)wbw (9.19)
/
and
Pp = (1 + Tp)wbp (9.20)
For simplicity, workers in both sectors are assumed to receive the same
wage w. However, markup rates and labor-output ratios can differ. As
discussed below, employment and distributional impacts of policy changes
depend crucially on intersectoral variation in these parameters.
All wage income goes to consumption, while a proportion s of profits
(from both sectors) is saved. Workers devote a fraction aw of their con¬
sumption spending to the wage good and (1 — aw) to the more luxurious
commodity; profit recipients follow the same pattern but with a parameter
ap. The names of the sectors already imply that aw > ap. Finally, the
P-sector is the only one to produce capital goods. The W-sector can be
imagined as making food products, clothing, and so forth, while the P-
sector provides cars, consumer and producer durables, and services catering
to high-income tastes. (In a poor country, the latter would range from
domestic servants to good restaurants to high tech health care supply.)
To set out balance equations, let X now stand for the ratio of W-sector
to P-sector capital (X = Kw/Kp) and define sectoral profit rates as:
(9.21)
and
(9.22)
1 + tw Pp Kw
The price ratio Pw/Pp in (9.22) is required to put profit rates on a common
base. As discussed in chapter 4, one can always measure quantities of goods
in such units as to set base year prices to one. We assume therefore that
PJPp = 1.
With this notation, one can immediately set down the saving-investment
balance as
Xgw + gP - s(rw\ + rp) = 0 (9.23)
where gw and gp are growth rates of capital stocks in the two sectors. The
condition that excess demand for the wage good is equal to zero takes the
form:
(1 /Tw)[aw + a/l — i)r, — (1 + rw)KX
+ (1 /tp)[clw + a,( 1 - s)Tp]rp + (aw - ap)(T/PpKp) = 0 (9.24)
The last term introduces a tax/transfer policy variable that can be used to
redistribute income. A nominal amount T is taken from profit recipients’
consumption spending (as by a consumption tax) and transferred to wage-
earners. The equation shows that as T rises in relation to the P-sector
capital stock, then demand for wage goods will go up.
In the very short run, sectoral capital stocks are fixed and X is predeter¬
mined. If, for simplicity, we assume that investment demand levels gw and
gp are constant, then (9.23) and (9.24) can be solved for rw and rp. Sectoral
outputs will, of course, be proportional to the profit rates, from (9.21) and
(9.22). After the algebraic dust settles, the solutions turn out to be:
£*X gp
rP =
A - £ (9.25)
1 + A
and
1 g»X + gp
= + { (9.26)
X(1 + A)
where
(1 + tw) — [a„ a,(l — s)t„]

A = (9.27)
aw -I- a„(l - 5)r„
and
k = (a* - ap)(T/PpKp) (9.28)
The coefficient A amounts to a profit rate multiplier. For plausible values

of the parameters, it will have a value close to (perhaps a fraction greater
than) one. Equations (9.25) and (9.26) show that profit rates and output
levels in both sectors respond to increased investment demand. When there
is an income transfer from profit recipients to wage earners (an increase in
T and £), rp will fall and rw will rise. If L is total employment, one can write
out an expression for its level as:
1 -
L — bwKw r,p
If bw and Kw are small in comparison to bp and Kp, it is clear that employ¬

ment may fall as a consequence of an income transfer toward wage-earners.
During the 1970s many input-output calculations of possible impacts of
redistribution on output patterns, employment, foreign exchange use, and
so forth, were made using blown-up versions of the present model (with tens
of sectors, classes of income recipients and so on). The results usually gave
changes on the order of a percent or so in sectoral outputs in response, for
example, to an income redistribution involving 5 or 10 percent of GNP.
Employment sometimes fell in the experiments, basically as a consequence
of a high service component in the consumption basket of the well-to-
do.8
In retrospect, these “small” changes in the production structure after
redistribution should not be too surprising. After all, they depend on differ¬
ences in income elasticities and budget shares between classes, such as the
(aw — ap) term in (9.28). For empirically estimated parameters, these
differentials are usually quite small.
A more interesting question, which the empirical studies did not address,
is how initial distributional realignments might cumulate into growth pat¬
terns in the longer run. An attempt at an answer requires specification of
investment functions, as in the following section.
9.4 Income Distribution and Patterns of Growth
To take up the longer term impacts of redistribution, it makes sense to

work with fairly general investment functions such as those specified in
chapter 2. Following equation (2.7), investment demand by capitalists in the
wage goods sector can be written as:
(9.29)
Here, z0w is a base rate of growth for which entrepreneurs aim. They
respond to higher profits with a coefficient zlw and to capacity utilization
according to z2w. As noted with chapter 2, the capacity utilization effect
amounts to an accelerator term in investment demand. Econometrics sug¬
gests that z2w > zlw.
Capitalists making luxury goods follow basically the same pattern; hence,
their investment demand is:
1 + TB
SP — zop + •ip 4- z T-P ' rP = z0p + <f>prp (9.30)
As it turns out, most of the action in the model depends on the difference
between and <f>p, which empirically likely sectoral distinctions in
markup rates (or shares of profits in value-added) will make nonzero even
if all investment demand parameters are the same. To save space, we thus
set zoh. = z0p — Zo in what follows.
It is easy to solve (9.29) and (9.30) for profit rates in terms of growth
rates. Substitution into (9.25) and (9.26) then gives expressions for growth
rates only:
<f>w /£»X + gp
- Zo — (9.31)
(1 + A)X V ~
and
&P ( g^ 4 gp ^ _
gP ~ Zo (9.32)
1 + A \ ~
Consider a steady state in which gw = gp = g, and the transfer term

£; is equal to zero initially. Then it is easy to see from (9.31) and (9.32) that
the value of the steady-state capital stock ratio is:
<K
X = (9.33)
hA
Routine differentiation shows that X will rise when aw or ar goes up, or

s falls. These are the responses one might expect. The effects on X due to
changes in the markup rates are ambiguous since the investment parameters
and A move in opposite directions.
Now suppose that ^ rises from zero as an income transfer is attempted.
Even at steady state, (9.31) and (9.32) form a nonlinear system in g, A, and
£. Resorting to total differentiation shows that A will rise in response to an
increase in £. Across steady states, a transfer from profit to wage incomes
will shift the structure of production toward wage goods. This is a direct
analog of the short-term result in the last section.
Regarding growth, the situation is more complex. One can show that the
derivative of the steady-state growth rate is:
Clearly, dg/dB, can have either sign. From (9.33) it will be negative when
the existing steady-state capital stock ratio Kw/Kp is less than one. A
relatively small W-sector (in these terms) will occur if A is larger than one
and the investment response of P-sector captitalists to higher profits is as
strong as the response of their colleagues in the W-sector. The implication
is that an income transfer toward wage recipients will make the growth rate
fall. Or, the other way round, policies aimed at worsening the income
distribution might stimulate long-run growth.
Evidently, a serious empirical question is involved in these results. It was
already noted in chapter 2 that Indian economists, in particular, have
argued that stagnation can result from an unequal income distribution. By
contrast, one can find Latin Americans asserting that deteriorating equity
goes hand in hand with faster growth.9 Cumulative processes of the type
illustrated here clearly underlie these judgments, and it would be interesting
to explore empirically whether they are valid or not. Unfortunately, this
task would not be easy since specification and estimation of investment
functions at a sectoral level is very hard to do.
The final matter to be discussed is stability of the steady state. Again
setting £ to zero for simplicity, one can solve (9.31) and (9.32) for the
growth rates. Since A — gw — gp, substitution shows that the growth of
the captial stock ratio (around an initial steady state) is
[(1/X) + 1] - „ Vo + x>
(1 + A )s
where
(9.34)
is the determinant of the equation system. For stability, the derivative of

A _
X with respect to X must be negative. Such will be the case when T is

positive. From (9.34) this condition requires that s—the saving response to
profit rate increase—exceed a weighted average of the investment responses
(f>„ and <f)p. As usual in models without capacity limits, long-run macro
stability occurs when savers respond more strongly than investors to profit
rate shifts.
10
Trade Patterns
and Southern Growth
TA HERE IS a long tradition of analysis in international trade theory

that is based on the notion that countries are dissimilar—in particular, that
developing economies in the “South” are shackled by trade patterns im¬
posed by colonialism and their own lack of technical mastery when they
were integrated into the world system in the nineteenth century. There are
recent exceptions to this generalization—the large petroleum exporters and
the so-called newly industrializing countries, or NICs. But by and large
most underdeveloped countries remain in the position of being hewers of
raw materials and drawers of oil for the industrialized world.1 In this
chapter, a simple formal model of such dependence in trade is presented.
The key assumptions go as follows:
The industrialized countries of the “North” grow in Keynesian fashion,
with excess capacity that can be reduced by brisker export sales to the
South. Northern investment demand and saving supply both respond to the
local rate of profit. When saving equals investment in macro equilibrium,
the North’s growth rate, profit rate, and output level are all determined. By
contrast, investment and growth follow from available saving in the South,
where output is constrained by supply conditions as opposed to demand in
the North. Surplus labor fixes the South’s real wage.
Consumption patterns differ in the two regions, but in the North the
Engel elasticity for the South’s export is less than one. The South is also
dependent on the North for supply of its capital goods; that is, technology
for production of commodities viable under the existing trade regime re¬
quires capital inputs traceable only to the North. The amount of Northern
financial capital transferred to the South (that region’s trade deficit) is
assumed to be determined by political factors, and is fixed in the short run.
(Alternately, capital flows responding positively to interest rate differences
between the regions and negatively to their own volume, or “lender’s risk,”
could be postulated without changing the model’s basic results.)
Finally, there is a third region in the system selling an imported interme¬
diate input to the North at a nominal price of its own choosing. Part of the
proceeds from sales of the intermediate flow back to the North as export
demand. A rise in this price is a stylized version of an oil shock.
All these assumptions go together to give conclusions that are very
strong, viz
1. There are not enough degrees of freedom in the international system to allow the
South to choose its own growth rate or terms of trade. Macro equilibrium is
determined by saving and investment functions in the North, plus its determina¬
tion of the magnitude of capital flows.
2. A productivity increase, or an oil shock by itself, will reduce aggregate demand
in the North, leading to slower growth in both regions and a fall in the South’s
terms of trade. In the medium run, the nominal wage rate in the North might
be expected to rise following either event, to restore aggregate demand. Any
inflationary impact would be beneficial to the South.
3. Faster capital flows from North to South stimulate the latter’s growth rate and
terms of trade; however, the quantitative response may be small. Capital flows
stimulate Northern capacity utilization and output growth. These responses fol¬
low from determination of output by aggregate demand in the North (the Hob-
son-Luxemburg theory of imperialism) and the role of surplus labor and capital
inflows in shaping saving supply in the South.
4. Productivity increases in the South will reduce its terms of trade and growth rate
so long as the North’s Engel elasticity for consumption of Southern exports is less
than one. Such impoverishing effects of technical change are often invoked in
structuralist discussions of the terms of trade, as several authors have pointed
out.2
5. In steady state, the Southern growth rate remains determined by the Northern
rate, and the response coefficient is equal to one. A perturbation to steady state,
such as an oil shock or a Northern productivity increase, will reduce the coeffi¬
cient, so that Southern growth will lag. The South carries a major share of
worldwide economic adjustment costs.
The reasons why these conclusions arise are singled out in the formal
discussion, which begins in section 10.1 with a description of growth in the
Trade Patterns and Southern Growth 179
North. Section 10.2 discusses growth in the South, and section 10.3 takes
up interactions of the two regions in terms of the system’s equations for
excess demands. Section 10.4 is devoted to comparative dynamics, and
section 10.5 concludes with a few additional observations and sketches
directions for further research.
10.1 Growth in the North
The strategy will be to set up a model in which the North grows along
lines sketched for the industrial economy in chapter 2, while surplus labor
characterizes the South. In addition, the North purchases an intermediate
imported input from a third party, “OPEC”. The three countries are linked
through trade and transfers and have their own behaviorally determined
demand and supply patterns.
The natural way to describe the North is as a Keynesian economy in
which firms are on their labor demand curve and labor can be hired as
necessary at the current nominal wage (with workers drawing on knowledge
that real wages have increased with productivity over time.) There is excess
capacity and markup pricing according to the rule:
P* = (1 + T)(wnbn + vm) (10.1)
In this equation, the North’s price Pn is determined by a markup over

variable input costs. The markup rate is r, and the inputs are labor and the
imported intermediate—the wage is wn, the labor-output ratio is b„, the
cost of the intermediate is v, and its input-output coefficient is m.
Income flows in the North come from wages and profits. If all wages are
consumed and a share s„ of markup income is saved, then consumption
demand from Northern incomes is:
[wnbn + (1 - s„)T(wnb„ + vm)]Xn = yX„
where Xn is the Northern output level. Total spending yXn is devoted to

purchases of Northern and Southern commodities. One key structuralist
hypothesis is that the South’s output has a low income, or Engel elasticity,
in Northern markets. Following the model of chapter 3, the simplest way
to capture this assumption formally is in the following expenditure equa¬
tions:
Spending on North’s commodity = (1 — B)yX„ — QPS

Spending on South’s commodity = ByXn + OPs
Here Ps is the price of the South’s output. So long as the constant 0 is

positive, the Engel elasticity for the South’s commodity is less than one. The
marginal propensity to spend on this commodity from total consumption
in the North is B (and the propensity from output X„ is By). It is easy to
verify that the sum of expenditures for the two goods is equal to yX„.
Now we want to describe macroeconomic equilibrium in the North. Let
the values of net exports to the South and OPEC be PnZ and (1 — s0)
vmXn, respectively. How real net exports to the South (Z) are determined
is taken up later; OPEC is simply assumed to save a fraction s0 of its income
vmXn and redirect the balance to consumption imports of Northern goods.
If the North’s investment level is /„, then its savings-investment balance
takes the form:
(Pn - y - vm)Xn + [vmXn - PnZ - (1 - s0)vmX„] = />„/„ (10.2)
The first term on the left of (10.2) is saving generated by Northern produc¬
tion, and the second term is the trade deficit. The two sources of saving sum
to the value of investment, PnIn.
This expression can be simplified if we define rn, the Northern profit rate
on capital, as
P„Xn - (w„b„ + vm)X„

PnKn
The interpretation of this string of equalities is the same as in chapter 2.

Total profits are sales less costs, or the numerator in the expression after
the first equality sign. As the volume of output Xn increases, then so do total
profits and the rate of profit rn on existing capital. The last equality shows
that r„ is proportional to the output-capital ratio u.
Along with r„ and u, we define three other new symbols. The export to
capital stock ratio Z/K„ is z; the share of intermediate imports in Northern
variable cost vm /(w„b„ 4- vm) is <f>; and the growth rate of Northern capital
stock In/K„ is g„. In this notation, the savings-investment balance (10.2)
can be rewritten as:
(5„ + T-'<KK = Z + gn (10.4)
Saving supply in the North is proportional to the profit rate r„, with a
coefficient determined by the North’s saving rate sn and net savings gener¬
ated by imports from OPEC. The saving is used to finance growth of capital
stock at rate gn and also net exports to the South (measured relative to
capital stock).
To determine macro equilibrium in the North, investment demand must
balance with saving supply. If investment rises with the profit rate, then the
demand function can take the form:
gn = go + hr„ (10.5)
Solving (10.4) and (10.5) together determines the growth and profit rates
g„ and rn. The resulting equilibrium will be stable under Keynesian output
adjustment in the North if investment responds less strongly to the profit
rate than saving, or s„ + r -1 <f)S0 — h > 0. A graphic presentation appears
in the right-hand quadrants of figure 10.1. The northeast quadrant shows
FIGURE 10.1
Determination of Macroeconomic Equilibrium under Trade between the North
and South
Note: The shifting schedules depict the response of the system to an “oil shock.
savings-investment equilibrium, and then the output-capital ratio u follows

along the “Capacity utilization” schedule from the profit rate in the south¬
east quadrant.
Comparative statics are straightforward. We present three exercises here
in preparation for later discussion. First, an increase in the price of the
imported intermediate will raise <fi, its share in variable cost. As imports
become more costly to the North, potential saving goes up. From (10.4) the
change takes the form of a rotation to the left of the savings schedule. The
diagram shows that the Northern response to something like an oil shock
will be slower growth, a lower profit rate, and less capacity utilization.
Second, <f> will also rise when there is a reduction in the labor-output ratio
bn, that is, a productivity increase. Once again, there will be economic
contraction. The cause this time is loss of jobs due to the productivity shift.
In practice, one would expect increases in both productivity and the import
price to be offset after a lag by a rise in the money wage w„. The wage
increase following a productivity gain is noninflationary and represents the
normal way in which a capitalist economy responds to technical change.
The increase after an oil shock of course adds to inflation but also stimulates
aggregate demand. These dynamic responses are not our main concern here,
but they illustrate the fact that the North possesses mechanisms to deal with
deflationary shocks. The same is far less true of the South, as we will see.
Finally, consider an increase in the North’s export surplus with the
South, z. From (10.4) an upward shift in z slides the intercept of the saving
schedule on the r„-axis in figure 10.1 to the right. The outcome is faster
growth, higher profits, and more capacity use. The Northern economy is
stimulated by higher net exports, as stressed by John Hobson and Rosa
Luxemburg long ago.3
10.2 Growth in the South
With its large subsistence sector, the South will have a labor supply
function elastic to the real wage. Moreover, labor supply conditions will
determine wages in the medium run, precisely because there is no history
of rising real income to induce workers to be somewhat tolerant of employ¬
ment lapses from time to time. To keep to simple formulations, we will
assume that the South’s real wage is constant in terms of its own product
—there is an infinitely elastic labor supply.
To explore the implications of surplus labor, we must specify conditions
of production in more detail. There are two key assumptions. First, capital
scarcity is the rule in the South, so that there is full capacity utilization (as
opposed to determination of output from the demand side in the North).
That is, Southern output Xs is proportional to the capital stock, Ks:
Xs = aKs (10.6)
The second assumption is that capital goods are imported. This polar case
in effect amounts to saying that all capital takes the form of “machines” that
are made only in the North. Our main results remain valid if part of capital
(“construction”) is Southern-made, but the algebra becomes cumbersome
enough to be left out.
With a fixed real wage and capital valued at the Northern price Pn, the
Southern profit rate rs can be written as:
_ PSXS - PswsbsXs _ Ps( 1 - w5bs)a

PnKs ~ Pn ‘ }
Here, w>5 is the South’s wage and bs is its labor-output ratio.

Since real wages are fixed in terms of the South’s product, it is natural
to assume that all consumption demand from labor income is devoted to
the home-produced good. Further suppose that profit recipients consume
only imports. In our scheme of things, Southern capitalists (or landlords)
form a “comprador” class living off local surplus extraction but focusing
their demand pattern abroad. Equating the values of demand and supply
for Southern output gives the equation:
(ByXn + QPS) + PswsbsXs = PSXS (10.8)
The term in parentheses on the left of (10.8) is Northern demand for the
South’s exports, and the other term shows consumption from labor income.
These two sources of demand exhaust the value of Southern output, PSXS.
Equation (10.8) determines the South’s price level (and its terms of trade
Ps/P„) from Northern aggregate demand, Xn. From (10.7) the South’s
profit rate becomes:
(1 - wsbs)a ByXn
(1 - wsbs)Xs - 0 Pn
(1 - wsbs)Xs Byu ^
(10.9)
(1 - wsbs)Xs - 0 P,
where X = Kn/Ks is the ratio of capital stocks in the two regions.

Equation (10.9) is graphed as the “Export demand” schedule in the
southwest quadrant of figure 10.1. The rightward rotation corresponds to
an oil shock or a Northern productivity increase not offset by a wage
change. The algebra becomes clear if we make the substitution:
_ wnbn + (1 - sn)T(wnbn 4 vm) (1 - <f>) + (1 - 5„)r

Pn (1 -f r)(wnbn + vm) 1 + r
into (10.9). A higher value of <f> (the share of imports in Northern variable
costs) reduces the slope of the schedule relating rs to u. An increase in the
Northern wage rate wn following a rise in productivity or the price of oil
will reduce cf> and benefit the South. However, such a response in the North
is always in danger of being cut off by anti-inflationary moves.
Perhaps surprisingly, an increase in Southern productivity also cuts back
on the Southern profit level (and export terms of trade) for a given level of
Northern capacity utilization u. This effect shows up algebraically in a
positive derivative of the expressions for rs in (10.9), with respect to the
labor-output ratio bp when 0 is greater than zero. Inelastic Northern
demand for the South’s export leads to a fall in its profit rate when bs
declines due to a productivity gain. Adverse effects of technical progress on
the South’s terms of trade (and, as we shall see, growth) are a staple
component of structuralist doctrine. They come out clearly from the de¬
mand conditions facing Southern exports in the model here.4
The final task is to determine the growth rate in the South. Sources of
saving are from income of capitalists and the North’s trade surplus PnZ,
which serves as a source of foreign saving for the rest of the world. We
assume that only a fraction t of this capital inflow is in fact invested, the
rest feeding into consumption imports from the North.5 On these hypothe¬
ses, the South’s saving-investment balance becomes:
ss(l - wsb,)P,Xs 4 tPnZ = P„IS
where Is is the South's real investment level. Dividing this expression by

P„KS and substituting gives the relationship:
gs = ssrs 4 tzX (10.10)
where gs is the growth rate of Southern capital stock (or Is/K,). This
equation appears as the "Saving supply” curve in northwest quadrant of
figure 10.1. It shows that Southern growth increases with the profit rate
rs and the Northern trade surplus z (though the quantitative importance
of the latter is attenuated by the consumption leakage parameter t).
10.3 How the Regions Interact in the Short Run
To put the model together, observe that in equilibrium it reduces to three

excess demand functions:
North Commodity Market
o - B)y^ _ ®Ps 1 (i - -OP - h’sbs)psxs
+ (1 - t)Z + --s«)vmX" + /„ + Is _ xn = 0 (10.11)
South Commodity Market
(ByX„
+ + wsbsXs — Xs = 0 (10.12)
V Ps
South Saving Supply
ss(\ - w5bs)PsXs + tP„Z - PnIs =0 (10.13)
The market for the North’s commodity is quantity-cleared; hence, the

equilibrating variable is output Xn. The Southern commodity is sold in a
price-clearing market, so that Ps (or the terms of trade Ps/Pn) adjusts.
Finally, if the North’s trade surplus (or the South’s deficit) PnZ is fixed,
then Southern investment Is must be the endogenous variable in (10.13).
Alternately, an independent investment function in the South would require
the region’s trade deficit to be endogenous. Essentially on political grounds,
we rule this possibility out. The North determines its export surplus domes¬
tically (savings directed abroad cannot be used for public or private invest¬
ment at home), and Southern investment accommodates to bring macro
equilibrium about. An important aspect of dependence is the inability of a
country to determine its own investment demand.
In equations (10.11)—(10.13), equilibrium is easy to find. Equation (10.12)
makes Ps a function of X„, so that from (10.13) Is depends on Xn as well.
Substitution into (10.11) gives a solution. If the North’s macro equilibrium

is stable, then overall stability is assured when (1 — wsbs)Xs — 0 > 0, that
is, the South generates a surplus of product beyond its own worker’s de¬
mand and the “floor” level of Northern imports, 0.6
These equations also give rise to figure 10.1. The flow of causality in the
model is clearly illustrated by the diagram. Macro equilibrium is deter¬
mined by the North’s saving and investment functions in the northeast
quadrant. Northern output and capacity utilization follow in the southeast
and determine real demand for the South’s export. This demand determines
the South’s terms of trade and profit rate (the southwest quadrant), and its
growth rate (the northwest quadrant). Growth in the South is ultimately
determined only by what happens in the North and OPEC.
The main comparative static results that follow from figure 10.1 have
already been derived, and can be summarized as follows:
1. When not offset by wage increases, Northern productivity growth or an oil shock
will reduce growth and profit rates in the two regions, and the North’s output-
capital ratio will fall.
2. An increase in Northern net exports to the South raises output, growth, and the
profit rate in the North. It also increases the South’s terms of trade, profit rate,
and growth rate, but the quantitative responses will depend on the North’s
marginal propensity to consume Southern goods (B) and the share of capital
inflows to the South that is actually saved (t). Both parameters may be small.
3. A productivity increase in the South will reduce the region’s terms of trade, profit
rate, and growth.
Other exercises could be pursued, but these suffice to illustrate the short-
run workings of the model. To close off the analysis, we turn to growth
analysis in steady state.
10.4 Steady State
To illustrate how North and South grow together, it is simplest to impose

balanced trade, or z =0. Also, note from (10.9) that unless the North’s
floor level of consumption (0) increases over time, it becomes quantitatively
unimportant in the determination of the profit rate rs as Xs goes up. A time
trend in 0 would probably be observed in the data (as is usually the case
when the linear expenditure system is estimated econometrically), but for
present purposes it is simplest to set 0 to zero, so that the Engel elasticity
of demand for the South’s commodity is equal to one. Under these assump-
tions we can solve (10.4) and (10.5) to get growth and profit rates in the
North as:
(s„ -h T 'fojgo
(10.14)
sn 4- T~'(f)s0 — h
and
_go_
rn (10.15)
s„ + r~'(f)s0 — h
From (10.9) and (10.3), the South’s profit rate rs is determined by r„ in the
expression:
rs = BX[r '(1 — <f>) + (1 — sn)]rn (10.16)
where it may be recalled that X is the ratio of capital stocks, K„/Ks. Finally,
note that when there are no capital flows the growth rate in the South is
given by:
gs = ssrs = 55BX[t ‘(1 — <f>) + (1 - sn)]rn (10.17)
The capital stock ratio X is the variable that adjusts to generate steady
states. It satisfies the simple differential equation:
dX W x
— = Mgn - gs)
dt
of the same general type as discussed in chapter 9. A steady state will be

observed when the growth rates in North and South are equal, so that
dX/dt = 0. Such a growth path is locally stable since g„ is independent of
X, whereas g* is an increasing function of X from (10.17). Substitution shows
that the standard stability condition is satisfied—dX/dt is a decreasing
function of X itself. For the geometry, look at figure 10.1, which is drawn
with g„ > gs. From this inequality, X will increase, causing the South’s
schedule for export demand (southwest quadrant) and growth rate (north¬
west quadrant) to rotate clockwise. The Southern growth rate will increase
until gs = g„ and the adjustment converges.
Equating gn to gs determines the steady-state value of X as
s„ + T~'(f>S0
X* = (10.18)
5sB[t-'(1 - <f>) + (1 - -O]
With this expression in hand, it is easy to characterize the steady state. First
insert (10.18) into (10.16) to get the South’s steady-state profit rate:
S„ + T-'(j>S0
rc = ---r.
The implication is that profit rates are not equalized. Rather, if the North¬
ern saving rate sn exceeds the Southern rate ss, then the South’s profit rate
will be higher than the North’s (even ignoring the saving generated from
intermediate imports). This finding is generally in line with the stylized
facts.
Second, from (10.4) and (10.17), the two growth rates are related as:
* - - ♦) + (■ - O l
Sn + T-'<f)S0
When X takes its steady state value X*, this expression reduces to gs =
g„. But suppose that the steady state is perturbed by an oil shock or a
Northern productivity gain. The import cost ratio <p will rise, and from
figure 10.1 or equation (10.14), the Northern growth rategn will fall. Since
the coefficient in (10.19) tying gs to g„ declines as well, Southern grow'th
is doubly hurt in the short run. In the longer run, (10.18) shows that the
steady-state capital stock ratio X* must rise in response to the lag in South¬
ern growth. The final outcome is a new steady state at a lower growth rate,
but with the South having carried the brunt of the adjustment cost.
The conclusion is that growth in the North serves as a locomotive for the
South—equation (10.19) assures that fact. However, the locomotive loses
traction whenever Northern productivity or the real cost of intermediate
imports goes up. Any time either event occurs, the coefficient relating
gs to gn becomes less than one. Even in terms of growth rates, conjunctural
perturbations help assure that parity between the two regions is indefinitely
postponed.
10.5 Final Observations
The foregoing results are quite striking, and should serve as a useful
purgative for anyone who is sanguine about benefits to the South from the
international economic system as presently arranged. Moreover, the analy¬
sis can be extended along several lines.
Can the framework here be broadened to deal with other issues in the
international order? One possibility is to consider dualities within the dual
regions of the present model. In the North, for example, a number of writers
have pointed to an emerging dichotomy between human service and envi¬
ronmentally oriented sectors with high income elasticities of demand on the
one hand, and the more traditional industrial sectors on the other. Within
the South, this same sort of industry is dual to the rural/agricultural sector
in ways that are illustrated in chapters 3 and 9. Addition of another com¬
modity to each country (nontraded services in the North, industry in the
South) would permit discussion within something like the present model of
such issues as to how sectoral polarities within each country are (or are not)
mutually reinforced through the system of international trade. Exploration
of such dualities has long been the stock in trade of dependency theorists;
there is no reason why mainline economists shouldn’t learn something from
it as well.
Finally, both the present model and dependency theory reach strong
conclusions about what is in store for the South. It is caught in a web of
economic restrictions from which the exits are difficult and few. Realign¬
ment of export patterns to more “modern” products, reduction of the need
for imported capital goods, elimination of surplus labor—these changes are
as difficult as they are essential to a more equable alignment of international
growth. The North will not willingly yield the upper hand. Internally
oriented policy and great self-reliance might enable the South to seize it in
the fullness of time.
11
Policy Lessons
TJHL WO MAJOR THEMES emerge from the model-building exercises

in the previous chapters. The first is of a long-run nature and reflects the
diversity of countries that make up the Third World. Each economy con¬
fronts a special set of structural constraints on its possible evolution, and
these restrictions vary greatly from place to place and time to time. Re-
source-rich countries face different problems than do resource-poor ones;
equal and unequal income distributions give rise to different patterns of
demand and possibilities for growth; the relative importance of the agricul¬
tural sector strongly influences inflationary pressures and formation of
aggregate demand. No single set of intuitions or equations can capture all
this variety, but country-specific models taking into account the salient
features of the economy at hand can say something about its medium- and
long-term prospects. The details appear in the rest of the book, but the
lessons for policy are spelled out in nonmathematical language in sections
11.9 through 11.13.
The second major theme is that standard short-run stabilization tools can
easily have undesirable effects in the Third World, again because there are
important structural features of poor economies that the usual models omit.
This point is argued in sections 11.1 through 11.8, where first the orthodox
stabilization package is described and then its weaknesses pointed out. Some
suggestions are offered about more reasonable approaches to stabilization,
but panaceas do not exist. There is ample room for innovation in the
short-term macro policy field.
Policy Lessons 191
11.1 Orthodox Stabilization Ploys
Imagine a country facing balance of payments difficulties, accelerating

inflation, and blatant mismanagement in key sectors of its economy. The
latter problems may include huge purchases of military equipment by gen¬
erals and admirals who happen to be running the political show, insatiable
greed and speculative binges on the part of their military and civilian
counterparts in charge of public enterprises, or the destabilizing impact of
rapidly growing demands for money wage increases on the part of workers
pushing for immediate and massive income redistribution. Compound all
this with sudden reluctance on the part of foreign commercial bankers to
increase their lending and possible political pressures from the superpowers.
What is the economic policy team supposed to do?
In most capitalist or mixed economies (and increasingly even in socialist
ones) there is a standard set of remedies that tends to be applied, often
through the intermediation of the International Monetary Fund. One can
distinguish at least seven distinct policies that go into orthodox packages.
The first three in the following list are almost always recommended, and
the other four show up frequently as well:
1. Monetary contraction, usually due to the imposition of separate ceilings on credit

from the banking system to the private sector and the central government.
2. Devaluation, either explicit or in disguised fashion through cuts in quotas, higher
prior deposits for imports and subsidies for exports, and general tightening of
trade restrictions. In-country economists may prefer the latter package of indirect
measures, while the IMF typically presses for a devaluation up front.
3. Abolition or reduction of government intervention in the price system, for exam¬
ple, consumer food subsidies.
4. Internal financial reform and liberalization, in particular, attempts to raise inter¬
est rates.
5. External liberalization in the form of reduction of barriers to trade and capital
flows. Again, there may be strong differences of opinion as to whether these will
be useful policies, but the consensus among the staffs of the Fund and the World
Bank and many academic consultants is that liberalization is a good thing.
6. With or without discrete devaluation, there may be advice to slow the rate at
which the local currency is devalued over time in a crawling peg. The slower
crawl is supposed to dampen expectations and be an anti-inflationary force.
7. Freezing of wage demands is often recommended to cut inflationary pressure and
perhaps shift the income distribution toward high-saving profit recipients and the
upper middle class.
How is a specific package fashioned from these elements likely to work?

The whole will certainly differ from the sum of the parts, a point taken up
in section 11.8. But first we have to look at the components, specifically

monetary contraction in section 11.2, exchange rate changes in section 11.3,
effects of intervention in food and other markets in section 11.4, interest rate
reform in section 11.5, foreign capital markets and the crawling peg in
section 11.6, and attempts at income redistribution in section 11.7.
11.2 Monetary Contraction
There are two justifications for monetary contraction, which get muddled
in minds of practitioners of applied orthodoxy who jet out to the Third
World. The first is based on the hoary equation of exchange from the
quantity theory of money, which states that the value (price times quantity)
of transactions in the economy is proportional to the money supply. If the
quantity of transactions or output stays constant, then a smaller money
stock ought to be associated with a lower price level. On this rationale,
cutting the money supply or at least its growth is a key anti-inflationary
move.
Open economy monetarism retains the hypotheses of the equation of
exchange and stable output (or, effectively, full employment) but adds the
notion that the price level is determined in the short to medium run either
from competition of imported goods or from labor and intermediate import
costs, that is, the levels of the wage and exchange rate. The IMF couples
these assumptions with the balance sheet identity for the banking system
—any rise in the money supply (the banks’ main liability item) must be
balanced by increases in loans to the government or private firms, or else
by higher foreign exchange reserves (banks’ loans and foreign exchange
holdings are their principal assets). Monetary restriction takes the form of
ceilings on credit to the government and private sector; the main benefits
are supposed to come in the form of higher foreign reserves, not lower
prices.
To apply the open economy monetarist model in detail, first make projec¬
tions of output and the price level; then you have money demand. Put limits
on government and private borrowing from the banking system. From the
banks’ balance sheets, the change in foreign reserves “must” be determined
as a residual item. Or, in other words, if domestic borrowing from the
banking system is restricted, then the balance of payments will improve.
This reasoning persuades many, and is certainly neat—causality runs
crisply from monetary restraint to lower prices and better external bal¬
ance. 1 Doubts can be raised, but necessarily center on inelegant details such
as the likely effects of monetary contraction in practice.
Policy Lessons 193
One objection—put forth as often by conservative businessmen as by

progressive economists—is that tight money drives up interest rates on
loans to firms for working capital and investment, and thus increases costs.
The normal business response would be to cut back on activity and attempt
to pass increased costs through to higher prices. Even if aggregate demand
falls under monetary constraint, aggregate supply may fall by more, so that
excess demand for commodities (demand minus supply) goes up. Further
inflationary pressure results.
The implication, argued in chapter 5, is that monetary restraint may be
stagflationary in the short run, increasing prices and causing output con¬
traction if interest rate cost-push is strong enough. The first monetarist
model—less money means lower prices—becomes insecure. The second
model—less money improves the foreign balance—goes through because
output contraction reduces intermediate imports. Can trade improvement
be made less burdensome? In the orthodox script, devaluation now plays
a central role.
11.3 Devaluation—Real Effects
There is little room in the monetarist model for devaluation. The ex¬
change rate may enter into import costs and the price level, but otherwise
is singularly absent from the accounting. To give devaluation a macro role,
two additional relationships have to be considered: internal balance or the
saving gap, and external balance or the trade gap. Details appear through¬
out this volume, especially in chapter 7.
Internal balance says that investment demand must be met by saving
from the government, private citizens, and foreigners. Foreign saving is
equivalent to a current account deficit, with the corresponding capital
account surplus representing an increase in foreigners’ saving in the form
of claims on “our” national wealth. External balance says that the current
account deficit is equal to imports of intermediate, capital, and consumer
goods plus interest payments less exports and net foreign remittances.2
These balances represent two restrictions on the macro system—there have
to be two corresponding endogenous variables. Candidates on the side of
real magnitudes include investment demand (or output growth), the level
of economic activity as a determinant of intermediate imports and national
saving, and the current account deficit itself. Possible endogenous prices are
the exchange rate and the aggregate price level (and implicitly the distribu¬
tion of income if the money wage is fixed). Different models follow from
different hypotheses regarding which variables are exogenous and which are
endogenous. In analyzing short-run stabilization, it is reasonable to take the

exchange rate as a policy variable and investment as fixed. The level of
output, the current account deficit, and the internal price level become
endogenous if the latter is determined by producers’ markups over interme¬
diate import and wage costs.3
Devaluation or an increase in the local currency to dollar exchange rate
has several effects. Prices will rise within the economy as intermediate input
costs are driven up, even if the money wage stays fixed. The real wage will
fall, and the income distribution may shift from low-saving workers toward
high-saving capitalists or markup income recipients.
A second increase in potential saving is from the trade deficit if (as usual)
devaluation is the last resort when there is an unfavorable foreign balance.
The rise in local currency prices of traded goods gives an income gain to
exporters with one hand but extracts a loss from importers with the other.
The loss exceeds the gain when there is a trade deficit, and real demand for
national products falls. Equivalently, foreign saving goes up.
The outcome from higher potential saving is bound to be economic
contraction coupled with a worsening income distribution unless export
volume responds smartly to the higher exchange rate. There is trade balance
improvement through reduced intermediate imports, but it comes at high
social cost.
Despite these ill effects, orthodox stabilizers see devaluation unambigu¬
ously as a useful thing. The reason is that the rise in domestic value of
exports is supposed to stimulate their production and sale abroad and offset
the contraction (though not the price increase nor the real wage decline).
Is this a likely event? Results differ from model to model, but in the simplest
cases of chapters 2 and 7 it can be shown that devaluation will be expansion¬
ary when the elasticity of export volume with respect to the exchange rate
exceeds the ratio of exports to intermediate imports. Note from the external
balance relationship that intermediate imports + capital goods imports =
exports + deficit if minor items are omitted from the accounting. The
stylized fact is that the trade deficit or financial capital flow toward a
developing country is likely to be less than its capital goods imports.4
Hence, exports must exceed intermediate imports, and the export elasticity
that guarantees expansionary devaluation can be less than one.
An export elasticity near one with respect to the exchange rate is reason¬
able in most countries, but only in a “medium” run of more than a few
quarters. Before then the export response will be weak, and devaluation
stagflationary. If monetary restraint is also stagflationary, there is ample
possibility of excessive contraction, or what has been called “overkill” by
the standard package. We return to this point in section 11.8.
Policy Lessons 195
11.4 Intersectoral Complications
The effects of any macro policy are complicated by the fact that in
many developing countries distributional and price interactions between
sectors are complex. The paradigm case is an agricultural (or food) sector
for which the market is cleared by a varying price coupled with a non-
agricultural sector in which output responds to demand. We can begin to
understand this system by looking at how income effects influence con¬
sumer decisions.
In the market for food, an increase in nonagricultural output and income
will raise demand. If supply is fixed or not very price elastic, the incipient
demand increase will have to be cut back by a higher price. Since demand
for food consumption is not elastic, the increase may be sharp.
In the market for nonagricultural products, things are harder to figure
out. A rise in the agricultural price, on the one hand, will increase farmers’
incomes and their demand for industrial products. By hypothesis, for a
generally oligopolistic industrial/service sector, the demand increase will be
met by increased supply.
On the other hand, an increase in food prices will reduce real income
elsewhere in the economy (bread or rice becomes more expensive and your
overall purchasing power falls), leading to less buying of nonagricultural
products by nonagriculturalists. Hence, demand for nonfood products can
either rise or fall with the food price. Consider the two possibilities in an
economy where policy is suddenly redirected toward an increase in agricul¬
tural exports.5
Farmers’ income would rise with the policy change; it cuts internal
supply and drives up agricultural prices. If demand for nonagricultural
products is dominated by rural income, there would be an export-induced
boom. By contrast, if the initial real income loss in the nonagricultural
sector is the dominant effect, demand for industrial products would fall and
recession would result. In the first case, the food price increase is likely to
be large since demand is price inelastic; in the second case, the real nona¬
gricultural output loss could be serious. Agricultural exporters are perched
on a knife-edge. Devaluation or other stimuli to exports do not ease their
balancing act.
The same line of reasoning applies to cutting food subsidies, another
orthodox policy move. Demand for food products will fall, leading to a
decline in either imports or the internal producer’s price (or both). There
may be a spillover into demand for nonagricultural goods, leading to overall
output contraction. Once again, a foreign balance improvement is obtained
at the cost of reduced economic activity and a price-induced shift in the
income distribution against the poor, who devote a large share of their
spending to food.
These two examples plus others presented in section 3.2 show that the
terms of trade between the agricultural and nonagricultural sectors are in
principle highly sensitive to attempted policy changes. Given that food
products make up a large share of total production and consumption activ¬
ity in poor countries, the consequent shifts in income distribution and
resource allocation could be expected to be large. In practice, violent re¬
sponses are not frequently observed. The reason is that governments almost
always intervene in food markets, separating and stabilizing consumer,
producer, and import prices to try to maintain or at most marginally alter
the status quo among the interest groups involved.6
This policy stance responds to the structure of a poor economy in which
price formation and output responses vary significantly across sectors. This
sort of problem is less serious in a more homogeneous developed country.
Simple attempts to alter the terms of trade by shifting food subsidy rates
or enacting agricultural export promotion schemes are bound to have gen¬
eral equilibrium effects that will call forth reaction on the part of groups
that get hurt. In some circumstances, such policy changes might be sensible;
they are just extremely difficult to apply.7
11.5 Mongering Financial Reform
Monetary contraction, devaluation, and tinkering with the commodity

price system represent the first line of orthodox stabilization policy. Now
we turn to more subtle moves, for example, a recent thrust toward financial
reform. This policy change is discussed in detail in chapters 5 and 6, but
the main conclusions are summarized here.
In practice, the reform package for internal finance is constructed around
an attempt to raise interest rates. Rates in “official” money markets (espe¬
cially on bank deposits) are often controlled in developing countries as a
natural counterpart to central bank regulation of the banking system
through credit restrictions as opposed to rediscounting and open-market
operations. These latter, market-based interventions are infeasible when the
local financial system is not highly articulated, as is the case in much of the
Third World.
The fixed interest rates are often negative in real terms, that is, the
nominal rate less the observed rate of inflation is less than zero. Negative
Policy Lessons 197
real rates are common in advanced economies, but have been pointed to as
a “distortion” ripe for liquidation in the Third World. The current rate of
inflation provides a convenient benchmark against which to measure what
the nominal interest rate “should” be.
What are higher interest rates supposed to achieve? One benefit, long
term at best, would be increased use of the banking system and greater
financial intermediation. More financial depth would permit open-market
operations and other flexible regulation policies and undoubtedly is a goal
to be pursued, but is not relevant in a stabilization context. What is relevant
is the notion that interest rate increases will raise saving rates and available
deposits in the banking system. If, as in neoclassical growth theory, the
extra saving gets translated automatically into higher investment, the econ¬
omy can be boosted painlessly onto a faster growth path.
It is argued in chapter 2 that the world probably does not operate neoclas-
sically—an ex ante increase in saving is more likely to lead to economic
contraction than a sudden investment spurt. Nonetheless, let us assume for
discussion that saving does rise with increases in bank interest rates, but that
investment demand is interest-sensitive as well. The question becomes:
sensitive to what interest rates? In many countries, the relevant rates would
appear to be the ones ruling in informal financial markets. Except for the
largest, both industrial and agricultural firms have recourse to informal
lenders instead of the banking system for their finance. Curb and village
credit markets are often efficient and well tailored to the needs of small
borrowers. Not surprisingly, the interest they charge is also quite high.
Now assume that bank deposit rates are increased, and in fact more
deposits come in. What is their origin? One possibility is a reduction in
informal lending, and a second is a drawing-down of hoards of gold, idle
balances, and sales of real estate and similar goods. If the latter occurs, there
will be a transfer from productively useless to useful assets. Bank lending
can increase, investment demand be stimulated, and working-capital costs
fall—all in all a favorable outcome.
The other story is more grim. If higher deposit rates induce moneylenders
or their backers to pull resources out of informal markets and put them in
banks, there can easily be an overall credit contraction because reserve
requirements or credit ceilings in official institutions are bound to be more
strict than those along the curb. The outcome could be working-capital
cost-push and a decline in investment demand and the level of economic
activity—stagflation once again. Under these circumstances, the increase in
ex ante saving would only make things worse.
Which outcome is more likely can only be judged by informed observers
in a specific country; broad generalizations are beside the point. The impor¬
tant conclusion is that the existing financial structure can frustrate simplis¬
tic attempts at reform by “getting the interest rate right.” A second observa¬
tion is that even if the policy is successful it will work only by getting the
financial system tail to wag the capital formation dog. Directly stimulating
investment and technical progress would be a straighter route to sustained
economic growth.
11.6 Crawling Pegs and External Financial Reform
Besides revising interest rates, recent orthodoxy has focused its attention
on the market in foreign assets and the role of the exchange rate in influenc¬
ing portfolio choice. To set the stage, consider an economy where the
authorities have been steadily devaluing the nominal exchange rate in a
“crawling peg” at a rate more or less equal to the difference between the
domestic and foreign rates of inflation.
Such policies were first introduced in Latin America in the 1960s. As
discussed in chapter 8, their rationale was to keep the real return to export
activities stable, and to preclude speculation against a maxidevaluation such
as occurs in a fixed-rate exchange regime.8
More recently, a third justification for a crawling peg has been proposed.
If the crawl is slower than the actual rate of inflation, it can dampen
expectations and gradually direct the economy toward an equilibrium with
stabler prices.
When crawling pegs were slowed in practice during the 1970s, the shift
came as part of a package involving reduction of restrictions on both foreign
trade and international capital flows. The results were unexpected.
First, foreign-held capital came streaming toward the domestic banking
system, which at any time had to trade national currency for dollars at the
ruling exchange rate. Slowing the speed at which the local exchange rate
was raised in periodic minidevaluations reduced the short-term return to
holding assets abroad, which can be written as foreign interest rate +
expected depreciation of national currency — expected foreign inflation
rate. So long as asset-holders continued to believe that the slower crawl
would be followed, they had strong incentives to repatriate wealth held in
dollars.
Second, the capital inflow led to substantial credit expansion by the
banking system as new deposits were lent out. The jump in money and
Policy Lessons 199
credit in the system led to a burst of inflation and instant appreciation or

a fall in value of the real exchange rate (defined as the nominal rate divided
by the national price level).
Third, there was an investment boom, easily explicable by euphoria in the
capitalist system and the increase in the profit rate entailed by lower real
import costs.
Now come the difficulties. After the initial exchange appreciation, there
was supposed to be a period during which the real exchange rate gradually
rose (or depreciated) toward a long-term level consistent with equilibrium
in the balance of payments. The rate of change of the real rate is the speed
at which the nominal rate is rising minus price inflation. If the real rate is
to rise, inflation must run slower than the crawling peg. The whole system
was expected to proceed to a low-inflation, steady-state equilibrium.
Of course, this intricate minuet never was properly danced. The initial
investment boom was followed by wage pressure, a profit squeeze, and a
cyclical downturn in standard market economy fashion.9 More seriously,
the lagging real exchange rate held down exports and led to unrest on the
part of businessmen in the export sector. Not willing to wait for slow, real
depreciation and return to external balance, they put their money abroad
in speculation against the nominal exchange rate. Meanwhile, export stag¬
nation worsened the internal recession. All signs pointed toward a max¬
idevaluation with consequent inflationary burst, output shock, and balance
of trade improvement, and that finally occurred. An elegantly contrived
long-term stabilization scheme ran afoul of the short-run facts of life.
What one learns from these failures is that it is perilous to trust in things
working without hitches in the medium to long run. Trying to wave down
inflation with a stable exchange rate presupposes that people with access to
wealth will behave the way you want them to. Such an assumption is
dangerous, doubly so when financial liberalization removes most obstacles
to capital flows. Too much was attempted with the slower crawling pegs
because the models underlying the policy were too simple and their ar¬
chitects paid scant heed to the historical failure of such policies before. The
fiasco of stabilization based on a slower crawl strongly suggests that the
model results be taken with a grain of salt, especially in the long run.
11.7 Income Redistribution and the Export Push
Usually, income distribution against labor and the poor is implicit in

stabilization attempts. Wages may be frozen, food subsidies cut, and exports
stimulated, all with benefit to landlords and rentiers.
Justifications for these policies are diverse: to increase saving and “there¬
fore” investment and growth (the neoclassical growth model again); to
contract “excessive” aggregate demand; to impose greater discipline on the
system, and so on.10
Progressive economists have not been much more coherent in their judg¬
ments about possible effects of redistribution. One strand—ranging from
the Franco-Italian Sismondi early last century to the Russian narod-
niki to modern Indian scholars—argue that an income distribution to¬
ward the poor will stimulate aggregate demand, lead to an investment
increase through accelerator effects, and raise the economy to higher
growth. Alternatively, the Latin American structuralist school in the 1950s
and 1960s explicitly stated that income concentration might be necessary
to stimulate demand for “luxuries” such as consumer durables, which
would be the basis for economic expansion after an “easy” phase of import-
substituting industrialization had passed.11
In the context of stabilization policy, these long-run models suggest that
the income concentration associated with most stabilization efforts may
dampen aggregate demand overall and at best stimulate only certain sectors
of the economy. The stimulation may or may not feed back into higher
investment and more economic animation after the stabilization episode is
over.
One particular issue that comes up repeatedly in both short- and medium-
run contexts is the possible effect of an export push. A number of factors
have to be considered:
First, is the economy at full employment of available capacity? If so, an
export effort will divert resources from either investment or consumption.
If growth is maintained, there would have to be income concentration to
lower the aggregate propensity to consume and permit greater sales abroad.
If the economy is not at full capacity, then exports add to aggregate demand
and permit higher output and growth. The discussion of devaluation in
section 11.3 presupposes excess capacity, but recall from there that the
beneficial output effects may take some time to appear.
Second, intersectoral redistribution may be important. Do saving and/or
consumption propensities differ across income flows generated by export
and nonexport sectors? One example in which demand-composition effects
Policy Lessons 201
matter is in mineral-exporting economies. The export revenues usually

accrue to the government, which preferentially demands nontraded goods.
The outcome is a higher home goods price, a lower real exchange rate, and
a lack of stimulation for production of nonmineral tradable goods at
home.12 Other less extreme models can also be produced, suggesting that
an export drive sets off general equilibrium effects that are not easy to trace.
Like attempts to get one or another price right, efforts at stimulating partic¬
ular sectors or activities can easily go awry.13
Finally, the models in section 7.4 show that high interest payments can
skew relationships between internal and external balance sufficiently to
make export expansion contractionary when there is less than full capacity
and can also create other paradoxes. In part these results are curiosa, but
approach reality in countries that borrowed heavily during the 1970s.
11.8 Summing-Up on Stabilization
Table 11.1 presents in summary form the judgments one can reach about
the effects of orthodox stabilization policies taken one by one. The results
are decidedly a mixed bag. Moreover, it is easy to show with simulation
models, such as the one in chapter 4, that often unfavorable effects are
TABLE 11.1
Likely Effects of Orthodox Stabilization Policies
Improve Improve
Maintain Balance of Reduce Income
Output Payments Inflation Distribution
Monetary contraction
a. Working capital important - + — -
b. Otherwise — + + +
Devaluation — initially + — —
-I- later
Commodity price reform
(especially food) — + + -
Higher interest rates
a. Draw resources from idle
balances + ? + ?
b. Otherwise — ? — ?
Slower crawling peg -)- initially -f initially — ?
— later — later
Export push (at excess capacity) + + — —
strengthened when two policies are applied simultaneously. For example,

devaluation coupled with monetary contraction can, under certain circum¬
stances, lead to extreme output loss, higher inflation, and income concentra¬
tion in the short run.
A second point is that the policy effects can differ, depending on condi¬
tions under which they are applied. If working-capital costs are important,
monetary contraction can drive up prices; if it pulls resources from produc¬
tive loans to the banking system, an interest rate increase will be counter¬
productive. But in other situations both policies could work well. Orthodox
programs implicitly postulate favorable circumstances. Less adventurous
packages can be based on the recognition that things are not necessarily
going to be all right.
All these complications exist in real economies and are not put forward
to argue that stabilization is not important or not necessary in some cases.
Rather the structuralist view emphasizes complexity and the need for wis¬
dom and receptiveness to how the economy at hand seems to work. A
structuralist stabilization package would no doubt include many of the
policies listed in table 11.1. Not many others are known. But it would not
apply them all in the usual directions, and would also incorporate distribu¬
tional considerations and nonmarket interventions explicitly. The basic
question at hand is how best to avoid the ill effects of standard policy tools
and to forge new ones. Structuralists have much work to do in these
directions.
11.9 Toward the Long Run
Many of the factors that condition long-run change in developing coun¬

tries have already been touched upon. Nonetheless, their secular influence
merits restatement and emphasis. The main topics discussed in previous
chapters are the following:
1. The terms of trade between agriculture and industry or (not the same thing)
traded and nontraded goods are less free to vary in some structural situations than
others, with consequent effects on sectoral growth and income distribution.
2. The income distribution itself may affect patterns of commodity demand, and
through induced sectoral investment the long-term growth prospects of the econ¬
omy.
3. The rate of inflation, rate of growth, and income distribution all interact in
complex fashion under different monetary growth rules and fiscal policy regimes.
Policy Lessons 203
4. Poor countries as a group may find their possibilities for autonomous develop¬
ment circumscribed by their dependent position in the international system.
These four issues are taken up in the following sections.
11.10 Terms of Trade
A fair policy question to ask is how long-term developments may be

altered by adjustments in key relative prices in the economy, for example,
between agriculture and industry or traded and nontraded goods. The
models in this book suggest that prospects for such changes are not always
favorable.
The traded/nontraded price ratio, for example, is a measure of the real
exchange rate. If it is low, there will be little stimulation for production of
tradable goods. Consider a mineral-exporting country as a case in point. In
the long run, world financial market conditions place an upper bound on
the ratio of its feasible trade deficit to GDP.14 To find the country’s import
capacity, its mineral export total should be added to the deficit since, with
a typically low cost of extraction and weak linkages to the rest of the
economy, the mineral sector essentially is a conduit for foreign exchange
transfers from the rest of the world to the home country.
Few economies resist the temptation to use up the transfer, through
imports, internal inflation, and capacity adjustment. Absorption of a large
volume of tradable goods requires their prices to be relatively low. The
productive structure is skewed toward nontradables, and both agriculture
and industry lag. Nonmineral exports stagnate under the low real exchange
rate, and both internal and external diversification of the economy decline.
The lack of diversification is rooted in the transfer; agricultural mono¬
exporters do not face the same problems, for example, because much of their
export revenue takes the form of generation of domestic value-added. If a
broadly based economy is taken as a policy goal, then one has to ask what
tools a mineral exporter can use to achieve it.
A frontal assault on the low real exchange rate through nominal devalua¬
tion seems unlikely to succeed. Traded goods prices will initially rise, but
the pressures to import remain. The usual outcome is that wages and
nontraded prices go up pari passu, and the real exchange rate is not altered.
More subtle strategies are piecemeal and dispersed: an export subsidy in
one sector, making another sector effectively nontraded through protection,
trying to cut back on imports by curtailing mineral production, or putting

the proceeds in long-term assets abroad. Whether such ploys can be effective
is open to question-—it is easy to point to mineral-exporting countries
around the world that have never effectively diversified. Their real terms of
trade between traded and nontraded goods remain low because of the
structure of their situation.
Somewhat similar issues arise between agriculture and industry, where
policy and custom separate internal prices in the food sector from those in
the world market.15 The agriculture/nonagriculture terms of trade are de¬
termined by internal supply-demand balance. A particular problem is that
a lagging agricultural sector can drive up the terms of trade and thereby
severely constrict demand for industrial products. Demand-creating poli¬
cies will not be useful in such a situation since much of their effect will be
dissipated in rising food prices.16
Here, the structural bottlenecks can only be broken by secular increases
in agricultural supply, or demand shifts away from food. New production
techniques may help on one front; urbanization on the other. But once
again, existing technology and demand patterns strongly limit variation in
the intersectoral terms of trade and resource allocation.
11.11 Income Distribution and Patterns of Growth
Additional terms-of-trade issues could be mentioned; for example, the

discussion in chapter 3 of the risk of high food prices, industrial stagnation,
and adjustment instability in agricultural mono-exporters. But it makes
sense to shift focus and inquire as to long-term prospects in economies
where supply restrictions are not of overriding importance. In particular,
how do the income distribution and patterns of demand and growth inter¬
act?
As already mentioned in section 11.7, underconsumptionist theories have
a long tradition on the left. The theme is that shifting the income distribu¬
tion toward workers and the poor will raise demand in the short run and
stimulate growth over a longer period. A more recent variation is that
demand-composition effects may preclude this happy result.
Theory does not take one very far in narrowing down the possibilities.
An investment-driven model is presented in chapter 9 to show that things
can go either way. An income redistribution toward labor could raise the
growth rate if the wage-goods sector is initially large enough and if its
Policy Lessons 205
entrepreneurs respond to increased demand with an investment push. More

supply-oriented models cited in this chapter give similarly ambiguous re¬
sults.
These exercises suggest that the feasibility of effective redistribution cou¬
pled with growth depends crucially on the institutions and detailed charac¬
teristics of the economy. It is very difficult to say whether (and under what
circumstances) progressive redistributive policies will work.17 Here—as in
suggesting truly innovative stabilization measures—we have an area where
structuralism is weak. The lacunae are more of fact than theory, but that
does not mean they should not be filled.
11.12 Inflation, Class Struggle, and Growth
Inflation as a macroeconomic phenomenon is poorly understood—one

need only contemplate the profusion of theories to recognize the point.
Structuralist models are not complete but at least provide a consistent
framework for understanding the phenomenon. Three mechanisms are im¬
portant. First, in many economic systems conflicting claims between classes
to the available product get translated into steadily rising prices. The classic
example is the attempt by workers to maintain living standards when there
is lagging agricultural supply. Food prices go up, and they bid up money
wages in consequence. The higher wages are passed along into higher
output prices, and a cumulative inflationary process (described in chapter
9) is set up. Were workers willing to accept a lower standard of living, the
inflation would not occur. But the same can be said of capitalists, the people
who mark up rising money wages into higher prices. Inflation is a sign that
a capitalist economy is trying to sweep class conflict over distribution of
product under the rug.
The second mechanism hinges on the observation that class power to
enforce claims depends on the economic situation—workers may well be
more powerful when inflation is fast or employment high. The result in the
medium run is the business cycle, and in the longer run a fair stability in
the functional income distribution. The cyclical picture is as follows: an
investment boom in the first instance shifts the income distribution toward
capital through forced saving. But as employment rises or inflation speeds
up, labor gains strength and forces an increase in the real wage. Profit rates
fall, investment stagnates, and output growth declines. Wage claims then
erode, the distribution shifts towards profits, and the stage is set for another
upswing.
Third—and more specific to the models of chapter 6—conflicting claims

are worked out subject to given institutional rules. One common case is
lagged wage adjustment to inflation, perhaps a wage increase in response
to past inflation rates every second quarter or every year. Under this type
of wage indexation, an acceleration in inflation will reduce the real wage
averaged over the indexing period. In inflationary economy practice, rules
are bent according to the macro situation, as in the profit-squeeze mecha¬
nism just described. The main point is that the rules have forced saving
through a wage lag built into them.
To explore the implications of these inflation responses, we have to bring
in the financial side of the economy. Two possibilities present themselves,
as described in section 11.2. Short-run monetary expansion, for example,
may reduce prices by moderating interest rate cost-push, or it may not. In
the longer run, these outcomes can be assessed when there is an increased
growth rate of money supply. To do theory tractably, we start from a
postulated initial steady state (described more fully in chapter 6). Across
steady states, can increasing the money supply growth rate reduce the rate
of inflation?
Surprisingly, the answer to this question is yes: an initial acceleration in
money supply growth will lead, under rapidly adjusting (or “rational”)
expectations, to an initially faster rate of inflation. Given lagged adjustment
of the money wage, the real wage consistent with indexing goes down.
Workers respond by being more aggressive in wage demands, and ulti¬
mately the actual real wage (or wage-price ratio) begins to rise. Saving
supply is squeezed and investment slows. The result is reduced demand
pressure on commodity markets and incipient decline in the rate of infla¬
tion. The ratio of money supply to value of output rises and, as a conse¬
quence, interest rates fall as well. Working-capital cost-push is reduced,
making room for higher real wages and a lower steady-state inflation rate
consistent with the indexing rules.
For this process to occur stably, investors have to be threatened enough
by increased wage demands to cut back aggregate demand and reduce
inflationary pressure; a profit-squeeze theory is involved in this sense. Sec¬
ond, the interest rate decline in response to a higher money/output ratio
must be strong enough to permit a higher steady-state real wage. If not, then
the second phase of the adjustment won’t go through and faster money
growth will lead to more rapid inflation, as usually postulated.
This scenario (described fully in chapter 6) shows clear long-run “non¬
neutrality” in the economy. Different money supply growth rates are
uniquely linked to different rates of price inflation and output growth—the
Policy Lessons 207
signs of the linkages depending on the strength of working-capital cost-

push. The mechanisms are structural and plausible in a Third World insti¬
tutional setting. Orthodox models that leave them out may be a poor basis
for formulation of long-term monetary and growth policy.
11.13 Poor Countries in the World System
One final topic is the position of poor countries as a group in the interna¬
tional system. Aggregating all the Third World together is perilous—pre¬
cisely the point of the previous discussion is that there is great structural
diversity in underdevelopment. Nonetheless, for the grand view one has to
work with grandly aggregated models. One such is presented in chapter 10.
Following structuralist tradition, it tries to spell out the essential asymme¬
tries in the international roles of rich and poor countries (or “North” and
“South”), underlining the barriers that hinder the latter group’s overall
growth. The key assumptions go as follow:
1. The North grows in Keynesian fashion, with excess capacity that can be reduced
by brisker export sales to the South. When saving equals investment in macro
equilibrium, the North’s growth rate, profit rate, and output level are all deter¬
mined. By contrast, investment and growth follow from available saving in the
South, where output is constrained by supply conditions as opposed to demand
in the North. Surplus labor fixes the South’s real wage.
2. Consumption patterns differ in the two regions, but in the North the Engel
elasticity for the South’s export is less than one. The South is also dependent on
the North for supply of its capital goods. As discussed in section 11.3, the amount
of Northern financial capital transferred to the South (that region’s trade deficit)
is likely to be less than the value of capital goods imports. The actual magnitude
is assumed to be determined largely by political forces, and is fixed in the short
run.
3. Finally, there is a third region in the system selling an imported intermediate
input to the North at a nominal price of its own choosing. Part of the proceeds
from sales of the intermediate flow back to the North as export demand. A rise
in this price is of course a stylized version of an oil shock.
All these assumptions go together to give conclusions that are very

strong:
1. There are not enough degrees of freedom in the international system to allow the
South to choose its own growth rate or terms of trade. Macro equilibrium is
determined by saving and investment functions in the North, plus its determina¬
tion of the magnitude of capital flows.
2. A productivity increase or an oil shock by itself will reduce aggregate demand

in the North, leading to slower growth in both regions and a fall in the South’s
terms of trade. In the medium run, the nominal wage rate in the North might
be expected to rise following either event, to restore aggregate demand. Any
inflationary impact would be beneficial to the South.
3. Faster capital flows from North to South stimulate the latter’s growth rate and
terms of trade; however, the quantitative response may be small. Capital flows
stimulate Northern capacity utilization and output growth. These responses fol¬
low from determination of output by aggregate demand in the North, and the role
of surplus labor and capital inflows in shaping saving supply in the South.
4. Productivity increases in the South will reduce its terms of trade and growth rate
so long as the North’s Engel elasticity for consumption of Southern exports is less
than one. Such immiserizing effects of technical change have often been invoked
in structuralist discussions of the terms of trade.
5. In steady state, the Southern growth rate remains determined by the Northern
rate, and the response coefficient is equal to one. A perturbation to steady state
such as an oil shock or a Northern productivity increase will reduce the coeffi¬
cient, so that Southern growth will lag. The South carries a major share of
worldwide economic adjustment costs.
The reasons why these conclusions arise are singled out in the formal
discussion of chapter 10, but their implications for the South seem clear.
Poor countries are caught in a web of economic restrictions from which the
exits are difficult and few. Realignment of export patterns to more ‘‘mod-
ern” products, reduction of the need for imported capital goods, elimination
of surplus labor—these changes are as difficult as they are essential to a
more equable alignment of international growth. Autonomous development
of the South may well require an inward-looking strategy delinked from
Northern pressures to the maximum possible extent. South-South trade and
relaxation of NIC restrictions on imports from their poorer neighbors
would help greatly toward these ends. But in the final analysis, each country
has to go it alone and make its own way out of the structural cage it is now
within. The failure of the New International Economic Order negotiations
of the 1970s showed how tightly Southern countries are bound. The suc¬
cesses will come only when they break their constraints.
Notes
Chapter 1
1. The terminology is due to Hicks (1974). For additional references, see chapter 3.
2. Commodities imported in “non-competitive” or “complementary” fashion cannot be
produced domestically within a reasonable time period (for example, a year) or without
unreasonable changes in existing techniques. “Competitive” imports are made up of
commodities that are (or can easily be) produced at home.
Chapter 2
1. Good references on the microeconomic rationale for markup pricing are Sylos-Labini
(1979) and Kalecki (1971a). A useful discussion is given by Lara-Resende (1979), while
Ros (1980) provides econometric justification in the case of Mexico.
2. More detail on input-output accounting is given in chapter 4 and Taylor (1979). When
intermediate input prices can vary, further complications are added to the adjustment
process, as illustrated in chapter 3.
3. Public sector enterprises fit uneasily into standard accounting categories but generate a
substantial cash flow. They are handled explicitly in an illuminating study of Brazil by
Wemeck (1980).
4. Or at least output variation is the main adjustment mechanism in Keynes’s General
Theory (1936). Previously, he had concentrated on forced saving type adjustments in the
Treatise on Money (1930), as discussed below.
5. See Kaldor (1955). Forced saving is the main adjustment process in Keynes’s Treatise on
Money (1930), where its continual replenishment of profit flows is likened to the biblical
parable of the widow’s cruse. Lopes (1977) gives a nice synthesis of the arguments
involved.
6. Note that in practice the government may serve as another important source of forced
saving. If tax receipts rise along with the price level P but government spending is fixed
by the budgetary process in nominal terms, then government saving will increase. This tax
effect is present in countries like the United States, where income is withheld during the
210 Notes
year, but may work the other way in many developing countries where taxes on incomes
earned in one year may not be collected until the middle of the next.
7. This quote is from a political economy analysis in the Indian newspaper, Economic and
Political Weekly (Special Number, August 1978). The model in the text follows Dutt
(1982).
8. Bitar (1979) is a standard economists’ reference on the Allende period in Chile. His and
other evidence shows that GDP rose by about 8 percent in 1971, Allende’s first full year
in the presidency. The major stimuli were income redistribution and higher aggregate
demand. In 1972, limits on both foreign exchange and capacity were reached, and inflation
(fed by full and unlagged wage indexation) accelerated rapidly. Needless to say, substantia]
internal and external political disruption abetted the inflationary process.
9. The notion that devaluation may lead to simultaneous price increases and economic
contraction has appeared for many years in the literature. Early versions in English are
Hirschman (1949) and Diaz-Alejandro (1963), and there are papers in Spanish and Por¬
tuguese as well. Cooper (1971a, b) provided both theoretical and empirical support for the
idea, and discussed the fate of ministers who devalue. The formal model here is similar
to that of Krugman and Taylor (1978).
10. Sylos-Labini (1979) argues that there are at most very weak linkages between markup rates
and capacity utilization in industrial economies; Ros (1980) finds a negative relationship
for Mexico.
11. For details on the CES function collected in one place, see Taylor (1979), Appendix D.
12. Another way of putting this result is that the present model is linearly homogeneous in
P and tv, so the price level will change by the same amount as the wage in percentage
terms, and w/P must stay constant. Adding an additional input such as intermediate
imports would permit the real wage to change when the money wage was modified (even
when the price level is not pegged), but the effect on resource allocation would probably
be weak. For algebra on this point see Taylor and Lysy (1979), who paraphrase the elegant
argument in Keynes’s General Theory (1936), chapter 19.
13. The whole argument in this chapter draws heavily on the literature about how different
schools of economists close their models in different ways, consistent (usually) with their
ideological stance. Sen (1963) is perhaps the first published paper on the topic, while
Marglin (1983) gives it an illuminating and thorough treatment. Note also in passing that
neoclassical full employment can be assured when there is an independent investment
function if both saving and investment are sufficiently responsive to the rate of interest
(introduced as an adjusting variable in the system, distinct from the rate of profit). Since
interest rates presuppose the existence of money and other assets, their analysis is post¬
poned to later chapters. In practice, private saving is not strongly affected by the interest
rate in poor countries, and the classical “loanable funds" adjustment is a weak foundation
upon which to construct a full employment general equilibrium. For evidence, see Giovan-
nini (1982), who severely criticizes regression results by Fry (1980) that purport to show
interest-responsiveness of saving in several Asian economies.
14. The radical story has its roots in the theory of the political business cycle put forth by
Kalecki (1971b), though he argued that capitalists might induce a recession to discipline
labor when there is full employment rather than have their own profits squeezed by
workers’ militancy. Current versions involving a profit squeeze are presented for the
United States by Boddy and Crotty (1975) and for Britain in a longer time frame by Glyn
and Sutcliffe (1972).
15. A complete, useful review of the American evidence is given by Weisskopf (1979).
16. Note the possibility that L > L. at least for some period of time—actual employment can
exceed the “normal" full employment level.
17. References on the stability of differential equations include Hirsch and Smale (1974) at
a fairly rigorous level, Gandolfo (1980), who stresses economic applications, and many
other texts. In this book, only local stability analysis based on the Routh-Hurwitz theorem
is employed.
Notes 211
Chapter 3
1. Sections 9.1 and 9.2 in chapter 9 extend the model further to deal with questions of
inflation and growth.
2. “Short run” in the context of the present model means a period of a year or less, during
which farmers are not likely to switch their patterns of input use markedly in response
to price or other signals. In the longer run, no one denies that farmers revise input and
yield patterns under institutional circumstances that reduce risk and guarantee some
security of tenure. Even so, total aggregate output from agriculture will depend on availa¬
bility of capital, land, and technical change. In a relatively land-scarce economy, a produc¬
tion function specification like (3.2) makes sense.
3. For a version of the model with a class structure in agriculture, see Taylor (1982b).
4. For references on stability of differential equations, see note 17 of chapter 2.
5. For more on the instability of food prices, see the model for India in the following chapter.
6. Reasons why the Law of One Price is frequently violated are reviewed by Kravis and
Lipsey (1977) and Isard (1977).
7. Chichilnisky (1981) gives one such model. In the present version, a full linear expenditure
system (with another constant term for nonagricultural demand corresponding to 0 for
agriculture) would permit existence of the unstable case in which the agricultural market
schedule of figure 3.1 would have a negative slope and cut the nonagricultural market
schedule from above. A rise in food exports would lead to a lower agricultural price in
this case.
8. See Ellman (1975) for discussion of the Soviet first plan experience.
9. Models of the terms of trade between food and nonfood sectors along the lines developed
here show up in various contexts throughout development economics. The behavior of the
large general equilibrium model for South Korea proposed by Adelman and Robinson
(1977) can be understood in terms of the simple model presented here. More interesting
political economy appears in Mitra (1977) and in empirical models developed by deJanvry
and Sadoulet (1983). Okun (1981) gives a lucid presentation of a developed country version
of the model.
10. There is a large structuralist literature about mineral-exporting economies from both Latin
American and Carribbean authors. See, for example, Girvan (1973), Brewster (1972) and
Sunkel (1969). The present formal treatment closely follows Boutros-Ghali (1981).
11. There is usually a wage differential in favor of the mineral sector (perhaps created by the
export companies to maintain a labor queue), which we ignore. For evidence regarding
Jamaica on this matter, see Brewster (1968) and, more generally, Nankani (1979). The
present model ignores the interesting policy question of what happens when the wage
differential is reduced.
12. Or at least 5 percent was the mineral sector wage share in Chile in 1977, according to the
input-output table.
13. It is worth noting that in many countries the production of U-sector products is likely to
be in the hands of public enterprises, which often engage in transactions that are not
rationed by price. If, however, the export sector is given priority when intermediates are
in short supply, then most of the results that follow will still go through.
14. These useful terms are from Boutros-Ghali (1981).
15. A good historical review of such developments in non-Persian Gulf oil exporters is given
by Gelb (1981).
16. Magee (1973) gives a good empirical description of the j-curve.
17. See Cooper (1971a) for statistics regarding the number of finance ministers who fell not
long after effecting a devaluation.
212 Notes
Chapter 4
1. The work reported here draws heavily on joint research with J0rn Ratts0 and Hiren Sarkar
on contructing a macroeconomic model for India. A directly parallel effort for Mexico is
being undertaken by the author with Bill Gibson and Nora Lustig. Models closely related
to the present one are described in McCarthy and Taylor (1980) and Taylor (1979), while
more distant cousins are scrutinized by Dervis, deMelo, and Robinson (1982).
2. Social accounting matrices derive ultimately from the work of Richard Stone at Cam¬
bridge University; see Stone (1966). Expositions of the usefulness of SAMs in developing
countries are given by Pyatt and Thorbecke (1976), Pyatt and Roe (1977), and Taylor
(1979).
3. Note in passing that the same sector can have both competitive imports and exports—this
happenstance is due to aggregation of trade in many commodities over time and space.
4. The SAM in table 4.1 was constructed by a team at the National Council of Applied
Economic Research in New Dehli; see Sarkar and Subba Rao (1981).
5. For applied detail in constructing social accounting matrices, see Pyatt and Roe (1977)
or Eckaus, McCarthy and Mohie-eldin (1981).
6. For discussion of the RAS procedure (due like social accounting matrices to Richard
Stone), see Bacharach (1970).
7. The original presentation of the linear expenditure system is by Stone (1954). A useful
monograph is by Lluch, Powell, and Williams (1977).
8. The output iteration just described amounts to a Gauss-Seidel procedure for solving a
closed Leontief model in which prices are fixed. Convergence will be rapid so long as
“leakages” of income flows to saving (instead of consumer demand) are nonnegligible.
9. The reduced form excess demand equations for sectors one and two with the correspond¬
ing prices as endogenous variables are highly nonlinear; hence a fairly sophisticated
algorithm must be used to find their solution. The solutions described here were obtained
with an algorithm due to Powell (1970). It is no longer state-of-the-art, but has proved
effective in several computable equilibrium models besides the one described here.
10. Taylor (1979) discusses the strengths and weaknesses of using SAMs in applied macro
analysis in developing countries.
11. Taylor (1979), Appendix B, discusses “guesstimation" of LES parameters in applied
models.
12. The Jacobian in table 4.5 is “exact.” An algebraic representation of the Jacobian was
obtained with the symbol-manipulation computer package MACSYMA at MIT; the
numerical version followed from plugging in the base solution parameters and values of
the variables from the SAM. The table 4.5 matrix does not include effects of wage
indexation, rising markup rates, and foodstock speculation as discussed below.
13. In the numerical solutions, initial prices are set to one, while quantities are scaled around
unity by figuring magnitude in units of one hundred billion rupees (instead of crores, or
ten million, as in table 4.1). In practice, computable general equilibrium models seem to
be easier to solve when all variables are normalized around one (or some other convenient
order of magnitude).
14. A quick check shows that all 2 X 2, 3 X 3, and so forth, principal minors of the Jacobian
have alternating signs—these determinants basically show whether or not groups of sec¬
tors taken in isolation show stable adjustments. Somewhat more complicated calculations
show that all eigenvalues of the matrix are negative, as required for stability. Their values
are —0.996, —0.834, —0.701, —0.574, and —0.052. The small magnitude of the last
eigenvalue reflects potentially destabilizing feedbacks between sectors one and two, as
described in the text.
15. The prices of the infrastructure services aggregated into sector four are fixed by the
government; hence, the markup rate is held constant.
Notes 213
16. The Jacobian in table 4.5 underestimates price responsiveness because the inflationary
feedbacks discussed here are not incorporated into its underlying equations. The price
responses in the expanded model are substantially larger than the values shown in table
4.5.
17. As discussed in chapter 2, the local currency trade deficit can move either way after
devaluation. It improves in the India model but deteriorates sharply in the companion
model for Mexico described by Gibson, Lustig, and Taylor (1982).
18. For completeness figure 4.2 shows how procurement would have to fall to compensate for
production decreases (points to the left of the base run.) Numerical results are not
reported, however, since changes in the procurement are not a likely policy under such
circumstances.
19. The extent to which the model needs smoothing depends on circumstances. In the Mexico
model described by Gibson, Lustig, and Taylor (1982), for example, agricultural price
fluctuations are much less brusque than those summarized here. The difference in behavior
probably reflects the reduced role of the agricultural sector in the Mexican economy as
compared to India’s.
Chapter 5
1. The short-run model of stagflationary effects of monetary restriction presented here is a

restatement and simplification of Taylor (1981a). The inspiration of that paper owes much
to Bruno (1979) and especially van Wijnbergen (1983b). Three-asset models in which
changes in the rate paid on bank deposits can reduce overall credit availability have been
worked out by several authors, including Buffie (1983), van Wijnbergen (1983a), and
Krugman (1981). Friedman (1978) presents similar notions in the context of the American
economy.
2. Emphasis on working-capital costs goes back far in structuralist tradition. A good empiri¬
cal study is by Morley (1971) for Brazil; see also Lara-Resende (1979) on the same country.
The current emphasis stems from Cavallo (1977) and the formalization by Bruno (1979)
and van Wijnbergen (1983b).
3. In practice, of course, some fraction of working capital will be financed by firms from
current income flow—perusal of flow-of-funds data suggests that the self-financed portion
may exceed one-half. Both the level and possible interest rate responsiveness of the
self-finance ratio are ignored in the following discussion. The main point is that its value
is well below one.
4. What may happen when the saving propensity from interest income exceeds zero is
discussed in note 6 below.
5. This investment function is a variant on the “q” theory of Tobin (1961).
6. If /< 1 is the share of working-capital cost financed by banks and the informal credit
market and s, > 0 is the saving propensity from interest income, then the numerator of
(5.8) should be rewritten as [(s — h)( 1 + i) — fs,i]wab. Its derivative will be positive
when s — fsj — h >0. The results of the text will be unchanged if / and s, are small
enough fractions. In practice, they probably are.
7. See Atkinson (1975) and Thurow (1975) for the relevant data.
8. These particular balance sheets follow van Wijnbergen (1983b). Buffie (1983) uses another
financial structure with similar results.
9. In practice, of course, financial markets are segmented with many interest rates and also
quantity-clearing from queues, moneylenders’judgments, bankers’ committees, and minis¬
ters’ telephone calls. Using one interest rate is at best a pallid miniature of all this
complexity. Regrettably, most of us can’t paint like the Younger Brueghel.
214 Notes
10. The monetary contraction could take the form of a lower value of H in (5.11). Strictly
speaking, there would have to be fiscal policy changes to permit H to vary over time, but
we ignore this complication here. Alternately, the reserve ratio p could be increased or
(in any functioning economy) myriad other instruments could be deployed to induce
monetary tightness.
11. See the references in note 2.
12. The quote is from Kapur (1976), which is more or less the approved formalization of the
institutional and historical arguments presented by McKinnon (1973) and Shaw (1973).
The general equilibrium asset market model used here of course follows in the tradition
of Tobin (1975). Recent specific applications to poor countries include Krugman (1981),
van Wijnbergen (1983a), and Buffie (1983), all of whom note that changes in the deposit
interest rate can move total private sector credit either way.
13. Argentina in the late 1970s was perhaps the nearest thing to complete credit market
liberalization. The results were disastrous but complicated by volatile capital flows in and
out of the country. See chapter 8 for details.
14. To their credit, McKinnon, Shaw, and their disciples would also push for lower reserve
requirements and loosening of credit restrictions to increase bank intermediation and the
effectiveness of interest rate reform. But these policies would just look like monetary
expansion in figure 5.1. Why bother with the deposit rate incantations?
15. Kindleberger (1978) gives a fascinating history of financial panics from destabilizing
speculation.
Chapter 6
1. The growth model presented here has a fairly complicated history. A first version of the
medium-run forced saving story of section 6.3 appeared in a 1976 first draft of Taylor
(1979), following the institutional analysis of Georgescu-Roegan (1970). Cardoso (1979)
added a monetary adjustment mechanism. The institutionally based wage adjustment is
based on Bacha and Lopes (1983), with mathematical assistance from Sweder van Wijn¬
bergen. Bacha and Lopes also extend the analysis to include the real balance or inflation
tax effects that are part of orthodox monetary growth models. The possibility that faster
money growth may slow inflation in steady states is pointed out by Sargent and Wallace
(1981). Their mechanism is financial market crowding due to government demands for
loans to finance the interest burden of its debt. Working-capital costs give rise to the
“perverse” long-run effect of slower money growth in the present model. This channel has
apparently not been noted before. Finally, for comparison of the wage dynamics adopted
here with more orthodox Marxian model closures, Marglin (1983) is a useful source.
2. The need for such extreme assumptions about investors' perspicacity stems from the way
P behaves in (6.1) and (6.2). Note that partial derivatives of both expressions with respect
to P are positive, ruling out simple excess demand based adjustment hypotheses.
3. The loss of real purchasing power of fixed money wages in the face of inflation is an old
Latin American notion—diagrams similar to figure 6.3 have been around the continent
for decades. The present formalization follows Bacha and Lopes (1983).
4. Chilean labor contracts in the 1960s had their renegotiations spread over a quarter or two
following legislation on government wage readjustments that could be passed anytime
during the first few months of a new year. For details, see Cortazar (1983).
5. Econometric evidence for South Korea and Argentina comes from van Wijnbergen (1982)
and Cavallo (1977), respectively.
Notes 215
Chapter 7
1. Factor income flows are omitted for the moment from (7.4)—repatriated profits from
multinationals or interest on sovereign debt going out and emigrants’ remittances or
intergovernmental transfers coming in. These items are introduced in section 7.4.
2. A wrinkle on (7.5) is that all of the current account deficit (or capital inflows less reserve
changes) may not be saved. Griffin (1970) and Weisskopf (1972) in influential papers both
pointed out that B in (7.5) should be multiplied by a saving share sb < 1. The foreign
exchange inflows not saved would have to be consumed, adding a term (1 - sb)eB to the
right-hand side of (7.3).
3. Evidence that h > 0 can be pieced together from such sources as the World Bank’s World
Development Report (1981).
4. When not all “capital inflows” are saved, as suggested by Griffin (1970) and Weisskopf
(1972), then the saving parameter sb < 1 will multiply q in the numerator of the expres¬
sion for the internal balance slope. The distinction between the two balances becomes even
stronger.
5. Early two-gap formulations are due to Chenery and Bruno (1962) and McKinnon (1964).
The latter paper is the source for figure 7.1. The basic notion of “external strangulation”
from lack of foreign exchange is due to the Latin structuralists; an early exposition in
English is in chapter 9 of Hirschman (1958).
6. The neoclassical case is argued by Bhagwati (1966) and Findlay (1971). It is further
discussed in connection with figure 7.5.
7. This interpretation of the two-gap hypothesis follows Bacha (1983).
8. The condition e0 < 0 resembles Marshall-Lerner restrictions on elasticities in orthodox
balance of payments theory. Dornbusch (1981) gives the canonical explanation.
9. We follow the usual confusing terminology about the exchange rate. Depreciation or
devaluation refers to a rise in q, and appreciation to a fall. When q (or e) rises, then our
country’s real (or nominal) purchasing power abroad goes down; hence, the decline,
depreciation, or devaluation of the local money’s force.
10. Tricks for formally setting up such cumulative processes as mineral abundance giving rise
to slow growth are presented in chapter 9. The obverse story is that a price shock to an
oil importer is likely to reduce u but raise q. The resulting stimulus to domestic activity
may accelerate growth in the long run.
11. In nominal terms, interest payments are a large component of the current account of
countries that borrowed heavily in world capital markets during the 1970s, such as
Mexico, South Korea, and Brazil. Real interest payments are substantially less (the extent
depending on the price deflator subtracted from nominal interest rates to make them
“real”), but still can amount to more than 10 or 20 percent of exports. In what follows,
we implicitly work with a real interest rate /. For more on debt and interest problems,
see Bacha and Diaz-Alejandro (1982).
12. For theoretical discussion of how credit rationing naturally arises in international capital
markets, see Eaton and Gersovitz (1981) and Sachs (1982).
13. Condition (7.23) follows Bacha (1983). At a full capacity level of the output-capital ratio
u, the quantity iP*(l — &)/u becomes a floor level on the export share e. If its value
of € falls below the floor, a country is not likely to get access to foreign commercial loans.
14. For examples, see Michaely (1977) and Balassa (1978).
216 Notes
Chapter 8
1. The name crawling peg was proposed by Williamson (1965). Bacha (1979) gives an
insightful review of the practical effects of a crawl.
2. Diaz-Alejandro (1981) gives an influential survey of the historical and institutional under¬
pinnings of responses in Latin economies to a slower crawl. Reviews of the Chilean and
Argentinian cases are given by Ffrench-Davis and Arrellano (1982) and Frenkel (1982),
respectively. The importance of economists who study real economies is emphasized by
the large number of theoretical papers that these studies stimulated, for example, Buffie
(1981), Calvo (1981), Krugman (1980), Dombusch (1981), van Wijnbergen (1983c), and
the present treatment, among many others. A related issue is switching by nationals
between local currency and readily available dollars. This is an important policy issue in
countries such as Egypt or Mexico; see Boutros-Ghali (1980) and Macedo (1983), or Solis
and Rizzo (1982).
3. The details are a bit different, but the general approach in the following model comes from
Krugman (1981).
4. As pointed out by Dombusch, et al. (1982), currency outstanding from the United States
Federal Reserve Bank is on the order of $500 per American. Even allowing for organized
crime, a large proportion of these dollars must be held by private individuals outside the
United States.
5. An excess of credit is nearly equivalent to an excess supply of money—if the latter drives
up the price level, then so should the former as well. Note from table 8.1 that money supply
Ds = H/p while (ignoring loans for gold for simplicity) loan supply Ls = (1 — p)
H/p. The simplest goods market clearing equation is PI = sPX, and loan demand
Ld = aPI (ignoring imported capital goods). Then Ls — Ld = (1 — p)H/p — PsX
is excess supply of loans. But this expression is proportional to a typical equation for the
excess supply of money of the form D1 2 — kPX where k = as/(l — p) can be interpreted
as an institutionally determined money demand parameter. The money and credit markets
are almost mirror images of each other, especially in an economy with little financial
depth, and one can model price pressures in one as well as the other.
6. This assumption precludes discussion of important questions regarding coordination of
monetary and exchange rate policy; see Krugman (1981) and Dombusch (1980) for
various rules.
7. Van Wijnbergen (1983c) has an interesting model along these lines.
8. Buffie (1981) reviews this cyclical history in Argentina and Uruguay, and presents a model
incorporating traded and nontraded goods and a Phillips curve.
9. The model by Polak (1957) enshrined the IMF approach—this is undoubtedly one of the
most practically important macroeconomic papers written since the Second World War.
The Fund’s nitty-gritty is further displayed by Robichek (1975).
10. Reichman (1978) and Reichman and Stilson (1978) back up the assertions made here in
official IMF publications. Buira (1983) presents a thoughtful synthesis of Fund experience.
Chapter 9
1. The label “passive money” and an interesting formalization appear in Olivera (1970).
2. See Deaton and Muellbauer (1980) for a good discussion of true cost-of-living indexes and
other demand theory concepts.
Notes 217
3. See Ahluwalia (1979) for evidence on India, for example.

4. This elegant argument follows Cardoso (1981).
5. Land reform as a device to improve agricultural productivity and reduce inflation was at
the key of the Latin America structuralist world view. The classic paper is Sunkel (1960).
An early (and neat) formalization appears in Garcia d’Acuna (1964).
6. The same general observation applies if a worker/capitalist (or landlord) class structure
exists in the agricultural sector. For details, see Taylor (1982b).
7. For the mineral sector long-term story, see Boutros-Ghali (1981). Growth models involv¬
ing distributional considerations along the lines of sections 9.3 and 9.4 are presented by
Taylor and Bacha (1976) and deJanvry and Sadoulet (1983).
8. The sort of model developed here underlies the guarded optimism about equity improve¬
ments presented in such books as Redistribution with Growth by Chenery, et al. (1974).
The input-output studies mentioned in the text are reviewed in part by Morawetz (1974).
9. Lustig (1980) and (1982) reviews this literature and runs it through empirical tests using
Mexican data. In a framework similar to the one here, deJanvry and Sadoulet (1983) study
inequality using data from several countries.
Chapter 10
1. The model presented here is used in one form or another by Prebisch (1959), Nurkse
(1959), and Lewis (1980). Its political economy is discussed by Taylor (1982a). The
formalization here follows Taylor (1981b). For a more neoclassical presentation see Find¬
lay (1981).
2. Bhagwati’s (1958) model of impoverishing growth (really exogenous technical change) has
been very influential, as evidenced by its transformation of the strange word “immis-
eration” into economic jargon. Bacha (1978) states a similar result in the context of the
debate about unequal exchange, further reviewed by Mainwaring (1980) and Gibson
(1980).
3. The references are Hobson (1902) and Luxemburg (1921). As pointed out by Hirschman
(1976), the philosopher Hegel was an (unlikely?) precursor.
4. More precisely, the terms of trade always shift against the South when its productivity
rises. So much is clear from the excess demand function (10.12), where a decrease in
bs cuts Southern workers’ total demand for their product, and leads its price Ps to fall.
However, the profit rate only falls when Northern demand is inelastic and the fall in price
is more than proportional to the productivity gain. Symmetrically to a productivity gain,
a decline in the Southern wage rate ws also reduces the region’s terms of trade, profit rate,
and growth. Though the mechanisms differ, this is the key result of the Sraffa-type unequal
exchange models of Mainwaring (1980) and Gibson (1980).
5. This hypothesis on saving follows Griffin (1970) and Weisskopf (1972). See also chap¬
ter 7.
6. If the model is extended to include endogenous capital flows, then it will be stable when
the North’s export surplus z adjusts according to a rule such as z = z0 + a(rs — r„).
In this equation, the parameter a measures the response of the level of capital flows to
a profit rate differential, while z0 represents foreign investors’ animal spirits. Adding such
a response function to equations (10.11)—(10.13) complicates the model but does not alter
its essential results. This more market-oriented variant of the political story in the text still
rules out independent Southern investment demand.
218 Notes
Chapter 11
1. For equations describing IMF “financial programming” as just sketched, see section 8.4
here and (in another variant) section 9.2 in Taylor (1979).
2. In most of the following discussion, foreign interest payments and remittances are ignored.
For detail on how the former may upset normal macro relationships, see section 7.4.
3. Chapter 2 gives a rationale for markup pricing rules. Note also that we do not permit freely
varying competitive imports or exports to close the “gap between the gaps” as in tradi¬
tional trade theory. For a justification, see section 7.2.
4. In a representative economy, investment might be 20 percent of GDP, made up in turn
of 60 percent nationally produced goods (construction and simple machinery) and 40
percent imported goods (sophisticated machinery). The capital goods import share in
GDP of 8 percent will exceed the feasible trade deficit for most countries.
5. The policy shift could take the form of devaluation or more sector-specific changes such
as export subsidies or acreage controls. Examples might be Bangladesh emphasizing jute
production as opposed to rice, or the Central American republics pushing bananas and
beef instead of corn and beans.
6. One key reason for these interventions is to avoid inflation. If the real cost of food rises
for groups such as urban workers, they can easily bid up money wage demands, which
in turn can be passed along to further price increases and an inflationary spiral. Models
are given in sections 9.1 and 9.2.
7. In some countries—Egypt is an example—nominal consumer prices of basic food items
have been held constant for years. Consequently, the real food price has dropped steadily
as the general price level has gone up. Such a secular decline in an important price sooner
or later has to be undone, but raising food prices suddenly is political dynamite. One
should have sidestepped the problem by slow and gradual price revisions, as in the
crawling peg exchange devaluation schemes discussed herein.
8. If the exchange rate is fixed when domestic inflation is substantially higher than that of
trading partners, imports steadily become cheaper and exporting less worthwhile. Pres¬
sures toward devaluation grow, and asset-holders begin to put wealth abroad through legal
channels or by overinvoicing imports, underinvoicing exports, and similar illegal devices
if necessary. The capital flight finally forces a maxidevaluation, with the shocks to the
economy discussed in section 11.3. Only the wealthy benefit, by bringing home their assets
and enjoying their capital gain.
9. A profit squeeze and consequent cutback in investment at the height of a boom is a
moderately convincing theory of the business cycle in a capitalist economy. For details,
see section 2.4 and chapter 6.
10. One imaginative—and nameless—structuralist economist fancies that the IMF team ar¬
rives in a country on mission with briefcases full of whips and leather dungarees. Mel
Brooks movies to the contrary, such is not the case. But it is at times difficult to avoid
reading self-righteousness or class bias into orthodox policy moves.
11. For references and models, see chapters 2 and 9. Note that distinct consumption propensi¬
ties between rich and poor for different goods, diverse investment and saving functions
and, more generally, demand composition differentials underlie the discrepant theories
mentioned in the text. It is hard to sort out the theoretically dominant forces underlying
different income redistribution models, let alone their empirical relevance.
12. For details, see sections 3.3 and 3.4.
13. For example, in the numerical model for India in chapter 4, devaluation and an export
push are strongly inflationary because they indirectly and directly generate demand for
products of the supply-limited agricultural sectors.
14. The implicit assumption is that the world capital market effectively imposes credit limits
Notes 219
on borrowing countries—there is nonprice rationing of Eurocurrency loans. The limits for

mineral exporters are relatively generous, but they do exist.
15. See section 11.4 and chapter 3 for reasons why governments usually separate internal and
external food prices. For present purposes, we can assume the nonagricultural price to be
determined from the world market or (more likely) from a markup rule.
16. The classic Latin American structuralist theory of inflation is based on conflicting claims
to product in a situation of lagging agricultural supply. See chapter 9 and section 11.12.
17. One cannot help adding, in the context of Britain and America in the early 1980s, that
regressive redistribution has scarcely proved its merits either.
References
Adelman, Irma, and Robinson, Sherman. Income Distribution Policies in Developing Countries:
A Case Study of Korea. Stanford, CA: Stanford University Press, 1977.
Ahluwalia, Isher J. “An Analysis of Price and Output Behavior in the Indian Economy:
1951-73.” Journal of Development Economics 6 (1979): 363-90.
Atkinson, A.B. The Economics of Inequality. New York and London: Oxford University Press,
1975.
Bacha, Edmar L. “An Interpretation of Unequal Exchange from Prebisch-Singer to Emman¬
uel.” Journal of Development Economics 5 (1978): 319-30.
Bacha, Edmar L. “Notes of the Brazilian Experience with Minidevaluations, 1968-76.” Jour¬
nal of Development Economics 6 (1979): 463-81.
Bacha, Edmar L. “Growth with Limited Supplies of Foreign Exchange: A Reappraisal of the
Two-Gap Model.” Mimeo. Rio de Janeiro: Pontifical Catholic University, 1983.
Bacha, Edmar L., and Diaz-Alejandro, Carlos F. "International Financial Intermediation: A
Long and Tropical View." Princeton University: Princeton Essays on International Finance
No. 147 (1982).
Bacha, Edmar L., and Lopes, Francisco L. “Inflation, Growth and Wage Policy: A Brazilian
Perspective,” Journal of Development Economics 12 (1983).
Bacharach, Michael. Biproportional Matrices and Input-Output Change. New York and Lon¬
don: Cambridge University Press, 1970.
Balassa, Bela. “Exports and Economic Growth: Further Evidence.” Journal of Development
Economics 5 (1978): 181-90.
Bhagwati, Jagdish N. "Immiserizing Growth: A Geometrical Note." Review of Economic
Studies 25 (1958): 201-05.
Bhagwati, Jagdish N. “The Nature of Balance of Payments Difficulties in Developing Coun¬
tries.” Measures for Trade Expansion in Developing Countries, Japan Economic Research
Center, Center Paper No. 5, (1966).
Bitar, Sergio. Transicion. Socialismo y Democracia: La Experiencia Chilena. Mexico City:
Siglo XXI, 1979.
Boddy, Raford, and Crotty, James. “Class Conflict and MacroPolicy: The Political Business
Cycle.” Review of Radical Political Economics 1 (1975): 1-19.
Boutros-Ghali, Youssef R. “Foreign Exchange, Black Markets and Currency Substitution:
The Case of Egypt." Mimeo. Department of Economics, Massachusetts Institute of Tech¬
nology, 1980.
Boutros-Ghali, Youssef R. “Single Export Systems and the Dependent Economy Model."
“Essays on Structuralism and Development.” Unpublished Ph D. dissertation, Massachu¬
setts Institute of Technology, 1981.
References 221
Brewster, Havelock. “Wage, Price, and Productivity Relations in Jamaica.” Social and Eco¬
nomic Studies 17 (1968): 107-32.
Brewster, Havelock. “The Growth of Employment under Export Biased Underdevelopment.”
Social and Economic Studies 21 (1972): 153-70.
Bruno, Michael. “Stabilization and Stagflation in a Semi-Industrialized Economy.” In Rudiger
Dornbusch and Jacob A. Frenkel (eds.) International Economic Policy: Theory and Evi¬
dence, Baltimore and Washington: Johns Hopkins University Press, 1979.
Buffie, Edward. “Price-Output Dynamics, Capital Inflows and Real Appreciation.” Mimeo.
Department of Economics, Yale University, 1981.
Buffie, Edward. “Financial Repression, the New Structuralists and Stabilization Policy in
Semi-Industrialized Economies.” Journal of Development Economics, in press, 1983.
Buira, Ariel. “IMF Financial Programs and Conditionality.” Journal of Development Eco¬
nomics, in press, 1983.
Calvo, Guillermo A. “Trying to Stabilize: Reflections on Argentina.” Paper presented to a
Conference on Financial Policies and the World Capital Market: The Problem of Latin
American Countries, Mexico City, 1981.
Cardoso, Eliana A. “Inflation, Growth and the Real Exchange Rate: Essays on Economic
History in Brazil.” Unpublished Ph.D. dissertation, Massachusetts Institute of Technology,
1979.
Cardoso, Eliana A. “Food Supply and Inflation.” Journal of Development Economics 8 (1981):
269-84.
Cavallo, Domingo F. “Stagflationary Effects of Monetarist Stabilization Policies.” Unpub¬
lished Ph.D. dissertation, Harvard University, 1977.
Chenery, Hollis B., et al. (eds.) Redistribution with Growth. New York and London: Oxford
University Press, 1974.
Chenery, Hollis B. and Bruno, Michael. “Development Alternatives in an Open Economy: The
Case of Israel.” Economic Journal 72 (1962): 79-103.
Chichilnisky, Graciela. “Terms of Trade and Domestic Distribution: Export-Led Growth with
Abundant Labor.” Journal of Development Economics 8 (1981): 163-92.
Cooper, Richard N. “An Assessment of Currency Devaluation in Developing Countries.” In
Gustav Ranis (ed.) Government and Economic Development. New Haven: Yale University
Press, 1971a.
Cooper, Richard N. “Devaluation and Aggregate Demand in Aid-Receiving Countries.” In
Jagdish N. Bhagwati et al., Trade, Balance of Payments and Growth: Papers in International
Economics in Honor of Charles P. Kindleberger. Amsterdam: North Holland, 1971b.
Cortazar, Rene. “Employment, Wages and Income Distribution in Chile.” Unpublished Ph.D.
dissertation, Massachusetts Institute of Technology, 1983.
Deaton, Angus, and Muellbauer, John. Economics and Consumer Behavior. London and New
York: Cambridge University Press, 1980.
deJanvry, Alain, and Sadoulet, Elisabeth. “Social Articulation as a Condition for Equitable
Growth.” Journal of Development Economics, in press, 1983.
Dervis, Kemal, de Melo, Jaime, and Robinson, Sherman. General Equilibrium Models for
Development Policy. New York and London: Cambridge University Press, 1982.
Diaz-Alejandro, Carlos F. “A Note on the Impact of Devaluation and Distributive Effect.”
Journal of Political Economy 71 (1963): 577-80.
Diaz-Alejandro, Carlos F. “Southern Cone Stabilization Plans.” In William R. Cline and
Sidney Weintraub (eds.), Economic Stabilization in Developing Countries. Washington,
D.C.: The Brookings Institution, 1981.
Dornbusch, Rudiger. Open Economy Macroeconomics. New York: Basic Books, 1981.
Dornbusch, Rudiger et al. “The Black Market for Dollars in Brazil.” Quarterly Journal of
Economics, in press, 1982.
Dutt, Amitava K. “Stagnation, Income Distribution and Monopoly Power.” Unpublished
Ph.D. dissertation, Massachusetts Institute of Technology, 1982.
222 References
Eaton, Jonathan, and Gersovitz, Mark. “Debt with Potential Repudiation: Theoretical and
Empirical Analysis.” Review of Economic Studies 48 (1981): 298-309.
Eckaus, Richard S., McCarthy, F. Desmond, and Mohie-eldin, Amr. “A Social Accounting
Matrix for Egypt, 1976.” Journal of Development Economics 9 (1981): 183-203.
Ellman, Michael. “Did the Agricultural Surplus Provide the Resources for the Increase
in Investment in USSR during the first Five Year Plan?” Economic Journal 85 (1975):
844-64.
Ffrench-Davis, Ricardo and Arellano, Jose Pablo. “Financial Liberalization and Foreign
Debt: The Chilean Experience, 1973-80.” Mimeo. Santiago Chile: CIEPLAN, 1982.
Findlay, Ronald. “The Foreign Exchange Gap and Growth in Developing Economies.” In
Jagdish N. Bhagwati, et al. Trade, Balance of Payments and Growth: Papers in International
Economics in Honor of Charles P. Kindleberger, Amsterdam: North-Holland, 1971.
Findlay, Ronald. “The Fundamental Determinants of the Terms of Trade.” In Sven Grassman
and Erik Lundberg (eds.) The World Economic Order: Past and Prospects. New York: Saint
Martin’s, 1981.
Frenkel, Roberto. “Financial Liberalization and Capital Flows: The Case of Argentina."
Mimeo. Santiago, Chile: CIEPLAN, 1982.
Friedman, Benjamin M. “Crowding-Out and Crowding-In: The Economic Consequences of
Financing Government Deficits.” Brookings Papers on Economic Activity, No. 3, 593-641,
1978.
Fry, Maxwell J. “Saving, Investment, Growth and the Cost of Financial Repression.” World
Development 8 (1980): 317-27.
Gandolfo, Giancarlo. Economic Dynamics: Methods and Models (2nd revised edition). Am¬
sterdam: North-Holland, 1980.
Garcia d’Acuna, Eduardo. “Inflation in Chile: A Quantitative Analysis.” Unpublished Ph.D.
dissertation, Massachusetts Institute of Technology, 1964.
Gelb, Alan H. “Capital-Importing Oil Exporters: Adjustment Issues and Policy Choices."
World Bank Staff Working Paper No. 475, 1981.
Georgescu-Roegen, Nicholas. “Structural Inflation-Lock and Balanced Growth." Economies
et Societes 4 (1970): 557-605.
Gibson, Bill. “Unequal Exchange: Theoretical Issues and Empirical Findings." Review of
Radical Political Economics 12 (1980): 15-35.
Gibson, Bill, Lustig, Nora, and Taylor, Lance. “Impactos Distributives de las Politicas del
Sistema Alimentario Mexicano en un Marco de Equilibrio General." Mimeo. El Colegio
de Mexico, Mexico City, 1982.
Giovannini, Alberto. “The Interest Elasticity of Saving in Developing Countries: The Existing
Evidence.” Mimeo. Department of Economics, Massachusetts Institute of Technology,
1982.
Girvan, Norman. “Multinational Corporations and Dependent Underdevelopment in Mineral
Exporting Economies.” Social and Economic Studies 19 (1970): 490-526.
Glyn, Andrew, and Sutcliffe, Bob. Capitalism in Crisis. New York: Panthenon, 1972.
Griffin, Keith B. “Foreign Capital, Domestic Savings and Economic Development.” Bulletin
of the Oxford University Institute of Economics and Statistics 32 (1970): 99-112.
Hicks, John R. The Crisis in Keynesian Economics. New York: Basic Books, 1974.
Hirsch, Morris W. and Smale, Stephen. Differential Equations, Dynamic Systems and Linear
Algebra. New York: Academic Press, 1974.
Hirschman, Albert O. “Devaluation and the Trade Balance: A Note." Review of Economics
and Statistics 31 (1949): 50-53.
Hirschman, Albert O. The Strategy of Economic Development. New Haven: Yale University
Press, 1958.
Hirschman, Albert O. "On Hegel, Imperialism, and Structural Stagnation." Journal of Devel¬
opment Economics 3 (1976): 1-8.
Hobson, John A. Imperialism: A Study. London: J. Nisbet, 1902.
References 223
Isard, Peter. “How Far Can We Push the ‘Law of One Price’?” American Economic Re¬
view 67 (1977): 942-48.
Kaldor, Nicholas. “Alternative Theories of Distribution.” Review of Economic Studies 23
(1955): 83-100.
Kalecki, Michal. “Costs and Prices” and “Distribution of National Income.” In Selected
Essays on the Dynamics of the Capitalist Economy. New York and London: Cambridge
University Press, 1971a.
Kalecki, Michal. “Political Aspects of Full Employment.” In Selected Essays on the Dynamics
of the Capitalist Economy. New, York and London: Cambridge University Press, 1971b.
Kapur, Basant K. “Alternative Stabilization Policies for Less-developed Economies.” Journal
of Political Economy 84 (1976): 777-95.
Keynes, John Maynard. A Treatise on Money. London: Macmillan, 1930.
Keynes, John Maynard. The General Theory of Employment, Interest and Money. London:
Macmillan, 1936.
Kindleberger, Charles P. Manias, Panics and Crashes: A History of Financial Crises. New
York: Basic Books, 1978.
Kravis, Irving B., and Lipsey, Robert E. “Export Prices and the Transmission of Inflation.”
American Economic Review (Papers and Proceedings) 67 (1977): 155-63.
Krugman, Paul. (1981) “The Capital Inflows Problem in Less Developed Countries.” Mimeo.
Department of Economics, Massachusetts Institute of Technology, 1981.
Krugman, Paul, and Taylor, Lance. “Contractionary Effects of Devaluation.” Journal of
International Economics 8 (1978): 445-56.
Lara-Resende, Andre. Inflation, Growth and Oligopolistic Pricing in a Semi-Industrialized
Economy: The Case of Brazil. Unpublished Ph.D. dissertation, Massachusetts Institute of
Technology, 1979.
Lewis, W. Arthur. “The Slowing Down of the Engine of Growth.” American Economic
Review 70 (1980): 555-64.
Lluch, Constantino, Powell, Alan A., and Williams, Ross A. Patterns in Household Demand
and Saving. New York and London: Oxford University Press, 1977.
Lopes, Francisco L. “Lucro, Juros e Moeda: Um Ensaio em Dinamica Keynesiana.” Revista
de Estudos Economicas 7 (1977): 221-51.
Lustig, Nora. “Underconsumption in Latin American Economic Thought: Some Considera¬
tions.” Review of Radical Political Economics 12 (1980): 35-43.
Lustig, Nora. “Characteristics of Modern Economic Growth: Empirical Testing of Some Latin
American Structuralist Hypotheses.” Journal of Development Economics, in press, 1982.
Luxemburg, Rosa. Die Akkumulation des Kapitals, Leipzig: Frankes Verlag, 1921.
Macedo, Jorge Braga. “Currency Diversification and Export Competitiveness: A Model of the
‘Dutch Disease’ in Egypt.” Journal of Development Economics, in press, 1983.
Magee, Stephen P. “Currency Contracts, Pass-Through and Devaluation.” Brookings Papers
on Economic Activity 1 (1973): 303-23.
Mainwaring, L. “International Trade and the Transfer of Labor Value.” Journal of Develop¬
ment Studies 17 (1980): 22-31.
Marglin, Stephen A. Growth, Distribution and Prices: Neoclassical, Neo-Marxian and Neo-
Keynesian Aporoaches to the Study of Capitalism. Cambridge, Mass.: Harvard University
Press, 1983.
McCarthy, F. Desmond, and Taylor, Lance. “Macro Food Policy Planning: A General Equi¬
librium Model for Pakistan.” Review of Economics and Statistics 62 (1980): 107-21.
McKinnon, Ronald I. “Foreign Exchange Constraints in Economic Development and Efficient
Aid Allocation.” Economic Journal 74 (1964): 388-409.
McKinnon, Ronald I. Money and Capital in Economic Development. Washington, D.C.: The
Brookings Institution, 1973.
Michaely, Michael. “Exports and Growth: An Empirical Investigation.” Journal of Develop¬
ment Economics 4 (1977): 49-54.
224 References
Mitra, Ashok. Terms of Trade and Class Relations. London: Frank Cass, 1977.
Morawetz, David. “Employment Implications of Industrialization in Developing Countries.”
Economic Journal 84 (1974): 411-52.
Morley, Samuel. “Inflation and Stagnation in Brazil.” Economic Devlopment and Cultural
Change 19 (1971): 184-203.
Nankani, Gobind. “Development Problems of Mineral Exporting Countries.” Washington
D.C.: World Bank Staff Working Paper No. 354, 1979.
Nayyar, Deepak. “Industrial Development in India: Some Reflections on Growth and Stagna¬
tion.” Economic and Political Weekly. Special Number, Augusto 1978.
Nurske, Ragnar. Patterns of Trade and Development. Stockholm: Almquist and Wicksell,
1959.
Okun, Arthur M. Prices and Quantities: A Macroeconomic Analysis. Washington D.C.: The
Brookings Institution, 1981.
Polak, J.J. “Monetary Analysis of Income Formation and Payments Problems.” International
Monetary Fund Staff Papers 6 (1957): 1-50.
Powell, M.J.D. “A Hybrid Method for Nonlinear Equations” and “A FORTRAN Subroutine
for Solving Systems of Nonlinear Equations.” In Philip Rabinowotz (ed.) Numerical Meth¬
ods for Nonlinear Algebraic Equations. London: Gordon and Breach, 1970.
Prebisch, Raul. “Commercial Policy in the Underdeveloped Countries.” American Economic
Review 49 (1959): 251-73.
Pyatt, F. Graham and Roe, Alan. Social Accounting for Development with Special Reference
to Sri Lanka. New York and London: Cambridge University Press, 1977.
Pyatt, F. Graham and Thorbecke, Erik. Planning Techniques for a Better Future. Geneva:
International Labor Office, 1976.
Reichman, Thomas M. “The Fund’s Conditional Assistance and the Problems of Adjustment,
1973-75.” Finance and Development 15 (no. 4) (1978): 38—41.
Reichman, Thomas M. and Stilson, Richard T. “Experience with Problems of Balance of
Payments Adjustment: Stand-by Arrangements in the Higher Credit Tranches, 1963—
1972.” International Monetary Fund Staff Papers 25 (1978): 293-309.
Robichek, E. Walter. “Financial Programming Exercises of the International Monetary Fund
in Latin America.” Mimeo. IMF Institute, International Monetary Fund, 1975.
Ros, Jaime. “Pricing in the Mexican Manufacturing Sector.” Cambridge Journal of Eco¬
nomics 4 (1980): 211-31.
Sachs, Jeffrey. “LDC Debt in the 1980’s: Risk and Reform." Mimeo. Department of Econom¬
ics, Harvard University, 1982.
Sargent, Thomas J. and Wallace, Neil. “Some Unpleasant Monetarist Arithmetic." Federal
Reserve Bank of Minneapolis Quarterly Review (Fall, 1981) 1-17.
Sarkar, Hiren, and Subba Rao, S.V. “Social Accounting Matrix for India for 1980-81." New
Delhi: National Council of Applied Economic Research, 1981.
Sen, Amartya K. “Neo-Classical and Neo-Keynesian Theories of Distribution.” Economic
Record 39 (1963): 54-64.
Shaw, Edward S. Financial Deepening in Economic Development. New York and London:
Oxford University Press, 1973.
Solis, Leopoldo, and Rizzo, Socrates. “Oil Surplus and Financial Openness: The Case of
Mexico." Mimeo. Santiago, Chile: CIEPLAN, 1982.
Stone, Richard. "Linear Expenditure Systems and Demand Analysis: An Application to the
Pattern of British Demand.” Economic Journal 64 (1954): 511-27.
Stone, Richard. “Multiple Classifications in Social Accounting” and “British Economic Bal¬
ances in 1970: A Trial Run on Rocket.” In Mathematics in the Social Sciences and Other
Essays. London: Chapman and Hall, 1966.
Sunkel, Osvaldo. "Inflation in Chile: An Unorthodox Approach." International Economic
Papers. No. 10 (1960): 107-31.
References 225
Sunkel, Osvaldo. “National Development Policy and External Dependence in Latin America.”
Journal of Development Studies 6 (1969): 23-48.
Sylos-Labini, Paolo. “Industrial Pricing in the United Kingdom.” Cambridge Journal of
Economics 3 (1979): 153-63.
Taylor, Lance. Macro Models for Developing Countries. New York: McGraw-Hill, 1979.
Taylor, Lance. “IS/LM in the Tropics: Diagrammatics of the New Strucuturalist Macro
Critique.” In William R. Cline and Sidney Weintraub (eds.) Economic Stabilization in
Developing Countries. Washington D.C.: The Brookings Institution, 1981a.
Taylor, Lance. “South-North Trade and Southern Growth: Bleak Prospects from a Structural¬
ist Point of View.” Journal of International Economics 11 (1981b): 589-602.
Taylor, Lance. “Back to Basics: Theory for the Rhetoric in the North-South Round.” World
Development 10 (1982a): 327-35.
Taylor, Lance. “Food Price Inflation, Terms of Trade, and Growth.” In Mark Gersovitz, et
al. (eds.) The Theory and Experience of Economic Development: Essays in Honor of Sir W.
Arthur Lewis. London: Allen and Unwin, 1982b.
Taylor, Lance, and Bacha, Edmar L. “The Unequalizing Spiral: A First Growth Model for
Belindia.” Quarterly Journal of Economics 90 (1976): 187-219.
Taylor, Lance, and Lysy, Frank J. “Vanishing Income Redistributions: Keynesian Clues about
Model Surprises in the Short Run.” Journal of Development Economics 6 (1979): 11-29.
Thurow, Lester C. Generating Inequality: Mechanisms of Distribution in the U.S. Econ¬
omy. New York: Basic Books, 1975.
Tobin, James. “Money, Capital and Other Stores of Value.” American Economic Review
(Papers and Proceedings) 51 (1961): 26-37.
Tobin, James. “A General Equilibrium Approach to Monetary Theory.” Journal of Money,
Credit and Banking 1 (1969): 15-29.
van Wijnbergen, Sweder. “Stagflationary Effects of Monetary Stabilization Policies: A Quan¬
titative Analysis of South Korea.” Journal of Development Economics 10 (1982): 133-70.
van Wijnbergen, Sweder. “Interest Rate Management in LDC’s.” Journal of Monetary
Economics, in press, 1983a.
van Wijnbergen, Sweder. “Credit Policy, Inflation and Growth in a Financially Repressed
Economy.” Journal of Development Economics, in press, 1983b.
van Wijnbergen, Sweder. “Capital Flows, the Real Exchange Rate and Preannounced Crawl¬
ing Peg Policies.” International Economic Review, in press, 1983c.
Weisskopf, Thomas E. “The Impact of Foreign Capital Inflow on Domestic Saving in Under¬
developed Countries.” Journal of International Economics 2 (1972): 25-38.
Weisskopf, Thomas E. “Marxian Crisis Theory and the Rate of Profit in the Postwar U.S.
Economy.” Cambridge Journal of Economics 3 (1979): 341-78.
Werneck, Rogerio L. “Rapid Growth, Distributional Equity and the Size of the Public Sector:
Trade-Offs Facing Brazilian Economic Policy in the 1980’s.” Unpublished Ph.D. disserta¬
tion, Harvard University, 1980.
Williamson, John. (1965) “The Crawling Peg.” Princeton University: Essays in International
Finance No. 50, 1965.
World Bank. World Development Report: 1981. Washington, D.C., 1981.
Index
Adjustment mechanisms: compar¬ 167-70; in short run, 60, 61, 64-

ative statics, 21-27; *n industrial 66, 68, 72, 78, 84; in social ac¬
economy, 13-36; neoclassical ap¬ counting matrix, 60, 61, 64-66,
proach to, 28-32; profit squeeze 68, 72, 78, 84; wages in, 48,
and growth dynamics, 32-36; 162-66
short-run adjustment in prac¬ Amortization costs and balance of
tice, 57-58; two-sector models payments, 54-55
of, 37-56
Agricultural sector, 37-48; con¬ Balance of payments, 26; and
sumer food subsidies in, 45; de¬ amortization costs, 54-55; and
mand in, 38-44, 46, 162-66; and asset markets and capital ac¬
exports-imports, 42, 43, 46; and count, 146-52; and capacity uti¬
free trade, 44; growth in, 166-70; lization, 151; and deposits, 147-
income in, 39, 40, 42-45; and in¬ 50, 158, 159; and exchange rate,
flation, 11, 47, 162-66, 169, 170; 1^45—57; and extensions and real¬
and intersectoral complications, ity checks, 157; and foreign as¬
195-96, 202; investment in, 167; sets, 145-60; and gold, 147-55;
labor in, 45, 162-65; output and and government, 147-50, 159;
production in, 38-47, 162-67, and growth, 145; and inflation,
170; price in, 11, 38-48, 162-66; 146, 147, 149, 152, 154, 155; and
profit in, 37, 38, 47-48, 162, 163, interest, 54-55, 145, 153; and in¬
166-70; savings in, 39, 45, 47, vestment, 153; and loans, 147,
228 Index
Balance of payments (continued) Currency devaluation; see Mone¬

148, 150, 151, 154-59; in mineral¬ tary devaluation
exporting economy, 54-56; and
monetary devaluation, 54-56, Deflation in medium run, 118; see
146; and price, 152-54; and also Inflation
profit, 145, 152; and real and Demand: in agricultural sector,
financial programming, 157-60; 38-44, 46, 162-66; and imports,
and savings, 151,157; and wages, 5; for investment, 21, 25; for
159; see also Trade balance loans and interest rates, 8-9; in
Bank reserves, 92-96 medium run, 120, 123; in North,
178; for output, 5, 15-16, 19; and
Cambridge school of economics, price, 24, 40; and profit squeeze,
4, 9 32, 33; in short run, 64, 73, 95,
Capacity utilization, 13, 20, 21, 31; 99. 106;
and balance of payments, 151; in¬ Deposits: and balance of pay¬
dustrial economy, 15, 17-21, 23, ments, 147-50, 158, 159; in short
28, 30-33, 35, 174; and invest¬ run, 95, 98-102;
ment, 6; in North, 178, 181; and Devaluation; see Monetary deval¬
profits, 17, 22-24, 32, 33; in short uation
run, 97; in South, 181; and trade Diversification of exports, 52
balance, 130, 133, 135, 136, 138,
.
139 143
Capital stock and growth, 8, 9, 11,
Employment: full level of, 5, 31-32;
and profit squeeze, 34; and
17, 25, 28-32, 34 wage-cutting, 30
Chenery, Hollis, 11 Energy crisis, 5
Chilean mirror image, 155 Engel effects, 8, 39, 43, 44. 46, 49,
Class struggle, 5, 9; and policy, 52, 167, 178, 186, 207, 208
205-7; Exchange rates, monetary, 11, 21,
Cobb-Douglas, 29 24-26, 57, 107-8; and balance of
Competition, perfect state of, 28 payments, 145-57; policy on,
Constant elasticity of substitution 194, 199, 2oi, 203; in short run,
(CES), 28-30 70, 80, 81; in social accounting
Consumer price index (CPI), 67, matrix, 70; and trade balance,
68 54-55. I29> 130, 132, 135-40
Consumption: imports, 9; in indus¬ Exports-imports, 9, 20-21; and ag¬
trial economy, 15,16; inelasticity ricultural sector, 42, 43, 46; and
of, 8; in short run, 64-67, 89; in consumption, 9; and demand, 5;
social accounting matrix, 64-67 diversification of, 52; elasticity
Corn Laws, English, 44 of, 26; and income, 23; in indus¬
Crawling pegs, 198-99, 201 trial economy, 17, 20-21, 24-26,
Index 229
28; noncompetitive, 9; and Imperialism, Hobson-Luxemburg

North and South, 180-82, 184; theory of, 178, 182
policy on, 200-201; in produc¬ Imports; see Exports-imports
ers’ markup, 8; productivity in¬ Income: in agricultural sector, 39,
crease in, 11; in short run, 59-61, 40, 42-45; distribution of and
66, 79-81; in trade balance, 6, 7 macro balance, 5-6; in indus¬
trial economy, 16, 21, 22, 27,
Finance costs in producers’ mark¬ 170-76; and inflation, 24; and
ups, 8 monetary devaluation, 25-26;
Flexprice-fixprice models, 8, 10 policy on, 200-202, 204-5; and
Foreign aid, 137, 198 price, 6; and profit, 21, 22; and
Foreign assets and balance of pay¬ short run, 60, 61, 64-66, 68, 76,
ments, 145-60 81, 82, 88, 107, no; in social ac¬
Foreign debt and trade balance, counting matrix, 60, 61, 64-66,
140 68, 76, 81, 82; see also Profit;
Foreign savings: components of in Wages
macro balance, 6, 7; and invest¬ India, social accounting matrix of,
ment, 6, 7 57-73. 87; money in, 103-8; nu¬
Free trade, 44 merical result on, 76-83
Industrial economy: capacity utili¬
Generalization, mathematical, 7 zation in, 15,17-21, 23, 28-33,35.
Gold: as asset, 92; and balance of 174; comparative statics on, 21-
payments, 147-55; medium 27; consumption in, 15, 16; ex¬
run, 121-22; public assets, 10; in cess capacity in, 14; and exports
short run, 92-95, 98-103 and imports, 17, 20-21, 24-26,
Government; see Policy 28; growth in, 23-26, 30-36,171,
Government debt, 92-94 x73—76; income in, 16, 21, 22, 27,
Growth, 6; in agricultural sector, 170-76; inflation in, 16-20, 22,
166-70; and balance of pay¬ 24, 32-36; investment in, 16-20,
ments, 145; in capital stock, 8, 9, 22, 24, 32-36, 174-76; labor in,
11,17, 25, 28-32, 34; in industrial 14, 28-36, 171; monetary de¬
economy, 23-26, 30-36,171,173— valuation in, 21, 24-27; neoclas¬
76; in medium run, 116-21, 122; sical approach to, 28-32; output
in North, 179-82,186,187; policy and input in, 15, 22, 25, 30, 31,
on, 202, 204-7; 'n short run, 90, 33, 171, 173; and policy, 14, 16;
91, 95, 101, no; in South, 182-88; price in, 14-15, 23-27, 29-31,171,
and trade balance, 129-44 172; profits in, 14-36, 170-74;
savings in, 16-23, 25, 3X~33. *74-
Hobson-Luxemburg theory of im¬ 76; taxes in, 21; wages in, 16-31,
perialism, 178, 182 170-72
230 Index
Inflation, 63, 80; in agricultural profits, 6; and savings, 6, 7, 16-

sector, 11, 47, 162-66, 169, 170; 19; in short run, 86, 87, 89, 97,
and balance of payments, 146, 98, 104, 106, 107, no; in South,
147, 149, 152, 154, 155; and con¬ 181, 184; and trade balance, 6,
flicting claims, 6; and income, 128, 142
24; in industrial economy, 16-
20, 22, 24, 32-36; and interest, 7; Jacobian matrixes, 8, 34, 41, 55, 58,
in medium run, 114-26; and 73-75, 80, 95, 99, 117, 118, 125,
money supply, 9, 10; policy on, 135. 138
191,192,198-202, 205-7; *n short
run, 88, 90, 95, 98, no—13; and Kaldor, Nicholas, 19
wages, 14, 78 Kalecki, Michal, 4, 19
Input: in industrial economy, 15, Keynesian economics, 8, 9, 19, 28,
22, 25, 30, 31, 33, 173; in North, 31, 32, 36, 76, 122, 177, 179, 207
177-82, 186, 187; in South, 180,
183-85; and trade balance, 129, Labor: in agricultural sector, 45,
r33, 135-40; see also Output 162-65; in industrial economy,
Interest, 8-9; and balance of pay¬ 14, 28-36, 171; labor-capital
ments, 54-55,145,153; and infla¬ ratio, 29-31, 34; in medium run,
tion, 7; and investment, 6-7; and 116; in mineral-exporting econ¬
loans, 8-9; and money supply, omy, 50; in North, 179; and
10; and trade balance, 6, 140-43; short run, 88, 89; in South, 182,
in medium run, 123-25; policy 183; and trade balance, 129; see
on, 191, 197, 201; in short run, also Wages
86-88, 90, 91, 97, 98, 102-6, 108, Land as asset, 92
110-13; and trade balance, 140-43 Law of One Price, 9, 44, 152, 159
International Monetary Fund, 11, L’Hospital’s rule, 115
46, 98, 146, 148,152,157,158, 191, Linear expenditure system (LES),
192 65, 68-69, 71
Investment, 9; in agricultural sec¬ Loans, 8-10, 192; and balance of
tor, 44, 167; and balance of pay¬ payments, 147,148, 150,151, 154—
ments, 153; and capacity utiliza¬ 59; and interest, 8-9; in medium
tion, 6; demand, 21, 25; in run, 121; in short run, 92-97,102,
industrial economy, 16-20, 22, 104, 105, 112, 113
24, 32-36, 174-76; and foreign Long run, policy in, 202-3
savings, 6, 7; and interest, 6-7;
investment-saving balance, 16, Manufacturing sector in social ac¬
18-19; and macro equilibrium, counting matrix, 64-66
16; in medium run, 126; in Markets, price clearance of, 3
North, 181; and policy, 207; and Markup; see Profit
Index 231
Marxism: conflict in, 109-10; and Narodniki, 200

unemployed reserve, 34 Nayyar, Deepak, 23
Medium run: appropriate condi¬ Neoclassical economic theory, 4, 8,
tions for, 109; deflation in, 118; 28-36, 57, 87, 132, 197
demand in, 120,123; gold in, 121- New International Economic
22; growth in, 116-22; inflation Order, 208
in, 114-26; interest in, 123-25; in¬ Newly industrializing countries
vestment in, 126; labor in, 116; (NICs), 177
loans in, 121; money in, 118-20; North: capacity utilization in, 178,
price in, 114-21; profit in, 116-21; 181; demand in, 178; and exports
rational expectations in, 118-20; and imports, 180-82,184; growth
and stagflation, 109; wages in, in, 179-82,186,187; input-output
114-22, 178 in, 177-82, 186, 187; investment
Mineral-exporting economy, 48- in, 181; labor in, 179; and pol¬
54; balance of payments in, 54- icy, 207-8; price in, 179-81;
56; consumer demand functions productivity in, 178, 179, 188;
in, 49-50; and government, 53, profit in, 177,179-81,186-87; sav-
55, 201, 204; labor in, 50; non¬ ings in, 180, 181, 187, 188; and
export production of, 52-53; short run, 185-86; steady state
output of, 48-56; policy in, 53, in, 186-88; wages in, 178,179,182,
55, 201, 204; price in, 49-56; pro¬ 186
duction levels in, 49; profit in,
49; wages in, 49, 53, 54 Oligopoly, 14
Monetarism, 157, 192, 193 OPEC, 179-81, 186
Monetary contraction, 191-93, 201 Output: in agricultural sector, 38-
Monetary devaluation, 7,10, 21, 80, 47, 162-67, 170; and demand, 5,
81, 107-8, 155; and balance of 15-16, 19; fixation of, 5-6; and
payments, 54-56, 146; and in¬ industrial economy, 15, 22, 25,
come, 25-26; in industrial econ¬ 30, 31, 33, 171, 173; of mineral¬
omy, 21, 24-27; policy on, 191- exporting economy, 48-56; and
94, 201; and savings, 24-25; and money expansion rate, 9; in
trade balance, 26-27; and North, 177-82, 186, 187; output-
wages, 27 capital ratio, 15, 28-31; in short
Money: in medium run, 118-20; in run, 87-89, 97; in South, 180,
short run, 86-108,112,113; in so¬ 183-85; and trade balance, 129,
cial accounting matrix of India, 133, 135-4°; see also Input
103-8
Money supply, 98; and inflation, 9, Pareto efficient allocation, 3
10; and interest, 10; and trade Policy, 14, 16, 170, 190-208; and
deficit, 26 balance of payments, 147-50,
232 Index
Policy (continued) 183—85; in trade balance, 132,134,

159; and class struggle, 205-7; 137, 138, 140, 142
crawling pegs, 198-99, 201; on Profit, 8; in agricultural sector, 37,
exchange rate, 194,199, 201, 203; 38, 47-48, 162, 163, 166-70; and
on exports and imports, 200- balance of payments, 148, 152;
201; financial reform, 196-99; on and growth of capital stock, 9; in
growth, 202, 204-7; on income, industrial economy, 14-36, 170-
200-202, 204-5; in industrial 74; and investment, 6; in me¬
economy, 14,16; on inflation, 191, dium run, 116-21; in mineral-
192, 198-202, 205-7; on interest, exporting economy, 49; in
191, 197, 201; and investment, North, 177-79, 186-87; policy
207; on long run, 202-3; in min¬ on, 199, 205-7; in short run, 64-
eral-exporting economy, 53, 55, 66, 68, 72, 77, 79, 89-91, 99; in
201, 204; on monetary contrac¬ social accounting matrix, 64-66,
tion, 191-93, 201; on monetary 72, 77, 79, 82; in South, 181, 184,
devaluation, 191-94, 201; on 187, 188; squeeze on, 10, 32-36,
monetary exchange rate, 194, 116, 118, 199, 206; and trade bal¬
199, 201, 203; and North, 207-8; ance, 128-29, 133-38, 141
on price, 191, 193, 194; on profit, Public assets, types of, 10
199, 205-7; on savings, 194, 197,
207; and short run, 89, 104-6, Real exchange rate, 130
202; and South, 207-8; stabiliza¬ Reserve Bank of India, 105, 106
tion actions, 191-92, 201-2; on RIMSIM, 158
stagflation, 194, 197; on terms of
trade, 203-4; on trade balance, Savings: in agricultural sector, 39,
191; on wages, 191, 194, 199, 45, 47, 167-70; and balance of
203-5, 207> 2°8 payments, 151,157; forced, 13,17-
Precious objects as assets, 92 21, 57, 90; and growth of capital
Price: in agricultural sector, 11, 38- stock, 17; in industrial economy,
48, 162-66; and balance of pay¬ 16-23, 25> 31-33. 174-76; and in¬
ments, 152-54; and demand, 24, vestment, 6, 7,16-19; and macro
40; and income, 6; in industrial balance, 6; and monetary de¬
economy, 14-15, 23-27, 29-31, valuation, 24-25; in North, 180,
171, 172; in medium run, 114-21; 181, 187, 188; policy on, 194, 197,
in mineral-exporting economy, 207; schedule of and evaluation,
49-56; in North, 179-81; policy 24-25; in short run, 60, 61, 64-
on, 191, 193, 194; in short run, 65, 67, 68, 77, 79, 81, 82, 89,102,
63-69. 73-75. 77. 79-82, 89-91, 107, no; in social accounting ma¬
94-99,101,103-5, Ho-13; and so¬ trix, 60, 61, 64-65, 67, 68, 77,
cial accounting matrix, 63-69, 79, 81, 82; and socioeconomic
73-75. 77. 79-82; in South, 180, class, 8; in South, 181, 184-85,
Index 233
187, 188; and trade balance, 131- in, 65; conclusions on, 84-85; as
35* 139* 141. 142 data base, 62; demand-supply
Short run, 57-113; in agricultural balances in, 64, 73; imports-
sector, 60, 61, 64-66, 68, 72, 78, exports in, 59-61, 66, 79-81; in¬
84; capacity utilization in, 97; come in, 60, 61, 64-66, 68, 76, 81,
commodities in, 102, 112, 113; 82; infrastructure and energy in,
consumption in, 89; demand in, 61; input-output in, 59-69, 73-
64* 73* 95* 99* 106; deposits in, 78, 82; manufacturing sector in,
95, 98-102; exchange rate in, 70, 64-66; numerical results on, 75-
80, 81; exports-imports in, 59-61, 83; price in, 63-69, 73-75, 77,
66, 79-81; gold in, 92-95,98-103; 79-82; profit in, 64-66, 68, 72,
growth in, 90, 91, 95,101, no; and 77, 79, 82; public sector in, bo-
income, 60, 61, 64, 65, 68, 76, 81, 72, 76, 79, 81-84; savings in, 60,
82, 88, 107, no; in industrial 61, 64-65, 67, 68, 77, 79, 81; sec¬
economy, 16-23, 25, 31-33, 174- toral consumption functions in,
76; and inflation, 88, 90, 95, 98, 65; subsidy in, 59-61, 64-66, 69,
no—13; interest in, 86-88, 90, 91, 70; taxes in, 59-62, 64-68, 72;
97, 98,102-6, 108, no—13; invest¬ and utility maximization, 65, 69;
ment in, 86, 87, 89, 97, 98, 104, wages in, 64-68, 76-79
106, 107, no; and labor, 88, 89; Socioeconomic class: conflict, 5, 9,
loans in, 92-97,102,104,105,112, 205-7; and savings, 8
113; money in, 86-108, 112, 113; South; capacity utilization in, 181;
neoclassical approach to, 87; exports-imports in, 180-82, 184;
and North, 185-86; output in, growth in, 182-88; input-output
87-89, 97; policy on, 89, 104-6, in, 180, 183-85; investment in,
202; price in, 63-69, 73-75, 77, 181, 184; labor in, 182, 183; and
79-82, 89-91, 94-99* 101, 103-5, policy, 207-8; price in, 180, 183—
no—13; profit in, 64-66, 68, 72, 85; profit in, 181, 184, 187, 188;
77, 79, 82, 89-91, 99; and ratio¬ savings in, 181, 184-85, 187, 188;
nal expectations, 110-13; savings and short run, 185-86; steady
in, 60, 61, 64-65, 67, 68, 77, 79, state in, 186-88; and trade pat¬
81, 82, 89, 102, 107, ho; and terns, 177-79, 182-89; wages in,
South, 185-86; stagflation in, 86, 177, 182-84
87, 97-98,100,107,108; taxes in, Soviet Union, five-year plan of, 45
59-62, 64-68, 72; wages in, 64- Stagflation, 10; and medium run,
68, 76-82, 88-90, 107, 110-13 109; policy on, 194, 197; in short
Sismondi, 200 run, 86, 87, 97-98, 100, 107, 108
Social accounting matrix (SAM), Stalin, 45
58-83; agricultural sector in, 60, Structuralism, 3—11; analytical
61, 64-66, 68, 72, 78, 84; cereal tools for, 7-9; and economic
stock speculation in, 64; closure structure, 5-7; and neoclassical
234 Index
Structuralism (continued) Urbanization, 204

economic theory, 4; and re¬ Utility maximization and social
source allocation, 3 accounting matrix, 65, 69
Taxes: in industrial economy, 21; Wages: adjustments in, 14, 19, 21,
in short run, 59-62, 64-68, 72 57; in agricultural sector, 48,
Technology, 204 162-66; and balance of pay¬
Tobin, James, 8, 9 ments, 159; indexation of, 77, 78,
Trade balance, 127-44; and capac¬ 114, 115; in industrial economy,
ity utilization, 130, 133, 135, 136, 16-31, 170-72; and inflation, 14,
138, 139, 143; and exchange rate, 78; institutional determination
54-55, 129, 130, 132, 135-40; and of, 5; in medium run, 114-22,178;
foreign debt, 140; and growth, in mineral-exporting economy,
129-44; and input-output, 129, 49. 53» 54; and monetary de¬
133, 135-40; and interest, 6, 140- valuation, 27; in North, 178,179,
43; internal and external, 128- 182, 186; policy in, 191, 194, 199,
30, 131, 136; and investment, 6, 203-5, 207, 208; and short run,
128, 142; and labor, 129; and 64-68, 76-82, 88-90, 107, no-
monetary devaluation, 26-27; 13; in South, 177, 182-84; and
policy on, 191; price in, 132, 134, trade balance, 129, 137; see also
i35> 137. 138, 140-42; and profit, Labor
128-29, I33-38. 141; savings in, Walras’s law, 16, 87, 95
131-35. 139. 141. 142; two-gap ap¬ Wealth: forms of, 91-94; identity
proach to, 130-35; and wages, of. 93-95
129,137; see also Balance of pay¬ World Bank, 11, 98, 146, 157, 158,
ments 191
Two-gap model, 10, 127, 130-35
*
*
(continued from front flap)
Structuralist Macroeconomics also considers in¬
come distribution and attempts to alter the struc¬
tural composition of sectoral outputs.
Throughout, the aim is to present simple and

empirically grounded models that are genuinely
useful and fit real problems, the kind that econ¬
omists in developing countries must grapple with
daily. Lucidly written and brilliantly argued,
Structuralist Macroeconomics is that rarity, an
extraordinarily teachable text (for graduate stu¬
dents in economic development) that at the same
time makes an important contribution both to
macroeconomic theory and to development eco¬
nomics.
LANCE TAYLOR is Professor of Nutrition and

Economics at the Massachusetts Institute of
Technology. He is the author of Macro Models
for Developing Countries (1979) and of numer¬
ous articles in scholarly journals.
JACKET DESIGN BY VINCENT TORRE

ISBN O-MbS-Ofl^-M

Lance Taylor - Structuralist Macroecnomics - Applicable Models For The Third World-Basic Books Inc (1983)

Uploaded by

Copyright:

Available Formats

Lance Taylor - Structuralist Macroecnomics - Applicable Models For The Third World-Basic Books Inc (1983)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lance Taylor - Structuralist Macroecnomics - Applicable Models For The Third World-Basic Books Inc (1983)

Uploaded by

Copyright:

Available Formats

APPLICABLE MODELS FOR

THE THIRD WORLD

Economists in developing countries have long

This path-breaking book by one of the country’s

Basic Books, Inc., Publishers New York

Library of Congress Cataloging in Publication Data

Copyright © 1983 by Basic Books, Inc.

1. The Structuralist Perspective

2. Adjustment Mechanisms—The Real Side

3. Adjustment Mechanisms—Two-Sector Models

4. Short-Run Adjustment in Practice 57

5. Money and Other Assets in the Short Run 86

6. Money, Inflation, and Growth 109

7. Trade Balance Complications 127

8. Foreign Assets and the Balance of Payments 145

9. Two-Sector Models of Inflation, Distribution, and

10. Trade Patterns and Southern Growth 111

11. Policy Lessons 190

neoclassical growth theory teaches that increased saving automatically

1.1 Economic Structure

The approach to models taken here begins by asking what variables

1. Output meets demand, made up of autonomous elements like investment and

Consider an economy in which commodity production requires inputs of

intermediation, firms in poor countries may find these working-capital ex¬

1.2 Analytical Tools

Perhaps surprisingly, in light of the way economic analysis has developed

situations of virtual disequilibrium. Suppose investment momentarily ex¬

1. The price level is often assumed to be determined by a fixed producers’ markup

is equal to zero. Differential equations are deployed to formalize conditions

1.3 Readers’ Guide

mathematical, scrutiny. By the same token, fundamental debates in macro¬

2.1 An Industrial Economy

responsive to government policy tools. Moreover, it is often characterized

P = (1 + t)(w6 + eP$a0) (2.1)

P = [wb + T(wb + eP$a0)] + eP%a 0 (2.2)

where u is the output-capital ratio. So long as there is excess capacity in

is satisfied. As usual, C in (2.4) stands for household consumption, / for

(total exports less imports of commodities similar to or “competitive” with

PC = wbX + (1 - sr)rPK (2.5)

PI - (ePZaoX - PE) - srrPK = 0

This expression can be interpreted as an investment-saving balance. When

g + f — (T-,<f> + sr)r = 0 (2.6)

The markup rate is constant so long as capacity utilization is below its

Macro adjustments in the vicinity of equilibrium I follow the classical

— = f[z0 + € + (z2 - (f>)u + (z, + <f> - sr)r]

2.2 Comparative Statics

The best way to illustrate the economic significance of these mechanisms

The third effect—a shift in the savings function—works in the opposite

~T~ = + z2(4> - sr)]

where, under full capacity, u will take the value u.

The macroeconomic result of these changes is ambiguous. Figure 2.4

al (1 + t)(1 — 4>) ( Tj E _ <f> |

In an application, the price ratio eP*/P should be fairly close to one

V = g + {e(eP*E/P) - (eP$a0/P)u j (2.12)

g = (z0 + z2u) -f zxr (2.13)

Evidently, if both r and the relative cost of intermediate imports eP*a0/P

2.3 A Neoclassical Digression

where L is the level of employment, cr the elasticity of substitution, and

If producers minimize costs subject to this production technology, then