EXPORT iNSTABiLITY. PROTECTION-
NDUCED Gag-FREE .DéSTORTSONS.
AND ECGNQE‘. i3 GROW“! iN" LESS
{LEVELOPED wUNIREES
Dissertation for the Degree of Ph. D.

WOMEN? SKATE UMVERSITY

Ai‘éYEGBUNAM WILUM‘I 033

1974

l

3" LIBRARY ,-'
'; Michigan State '

University

 
 

'Ifi.

  

This is to certify that the

thesis entitled

EXPORT INSTABILITY, PROTECTION-INDUCED
COST-PRICE DISTORTIONS, AND ECONOMIC
GROWTH IN LESS DEVELOPED COUNTRIES

presented by

Anyegbunam William Obi

has been accepted towards fulfillment
of the requirements for

Ph . D. degree in Economics

" J7 U/ nun/L

Major professor

Dategél/7 V

0-7639

 

 

 

ABSTRACT
EXPORT INSTABILITY, PROTECTION-INDUCED

COST-PRICE DISTORTIONS, AND ECONOMIC
GROWTH IN LESS DEVELOPED COUNTRIES

By

Anyegbunam William Obi

During the past few decades, much concern has been
expressed that the relatively large short-term fluctuations
in the export earnings of currently less developed countries
may constitute a major obstacle to their attempts to achieve
an accelerated rate of growth. Implicit in some of the
arguments used to justify such concern is a theory of the
development process in which the rate of growth of the
importing capacity of exports is the critical determinant
of the rate of economic growth, and yet, because of certain
rigidities in the structure of trade and production, maxi-
mum export earnings persistently tend to fall short of
minimum import requirements. In such a framework, indus-
trialization - through import-substitution - is seen as the
only viable means of minimizing the adverse consequences of
trade instability and relieving the foreign exchange
constraint on growth.

In the years following the Second World War, import-
substitution became a widely adopted strategy among the

deve10ping countries for the achievement of rapid growth

(5 Anyegbunam William Obi

and industrialization. This strategy entailed the encour-
agement of domestic manufacturing behind a protective wall
of tariffs, quotas, exchange and other quantitative controls
on trade. But in the past few years, considerable dis—
enchantment with import-substitution has been expressed.

A fundamental criticism of "forced" import-substitution
implicit in the conventional comparative-cost doctrine is
that it inevitably entails a cost-price structure which is
non-Optimal. The present study has examined empirically
the propositions that (a) export instability, and (b) cost-
price distortions engendered by protection in futherance of
import-substituting industrialization have been detrimental
to the growth of the developing countries.

Within an analytical framework suggested by the foreign
exchange gap theory of trade and growth, multiple regression
analysis was used to relate the rate of economic growth to
the degree of export instability and to the extent of two
separate aspects of protection-induced cost-price distor-
tions, viz - the overvaluation of the exchange rate, and
the change over time in the excess of domestic over world
price valuation of domestic manufacturing valued-added.

The statistical tests (a) confirmed the findings of previous
studies of a lack of association between the degree of ex-
port instability and the rate of economic growth among the
less deve10ped countries taken as a group; and (b) indicated

that the degree of cost-price distortions associated with

Anyegbunam William Obi
protection is not systematically related to the rate of
economic growth. The latter finding is consistent with
the view that observed cost-price distortions in developing
countries are more the random results of attempts to force
the pace of economic growth in the face of various struc-
tural disequilibria than the systematic reflections of the

extent of deviation from neo-classical optimal policy.

EXPORT INSTABILITY, PROTECTION-INDUCED
COST-PRICE DISTORTIONS, AND ECONOMIC

GROWTH IN LESS DEVELOPED COUNTRIES

By

Anyegbunam William Obi

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Economics

1974

ACKNOWLEDGEMENTS

I would like to thank Professors M.E. Kreinin
(Chairman), J. Hunter, and M. Weinrobe of my Doctoral
Committee for their advice and encouragement at various
stages in the preparation of this thesis. My deep grati-
tude also goes to my colleague, Dr. S.E. Reynolds, at the
University of Utah for many stimulating discussions.

All errors of omission and commission are of course

mine.

11

TABLE OF CONTENTS

Chapter
INTRODUCTION . . . . . . . . . . . . . . .
THE EXPORT INSTABILITY PROBLEM . . . . . .

Introduction . . . . . . . . . . . . . . .
The Causes of Export Instability . . . . .
The Effects of Export Instability. . . . .

THE MEASUREMENT OF EXPORT INSTABILITY. . .

Introduction . . . . . . . . . . . . . . .
Definition of Export Instability . . . . .
The Export Instability Index . . . . . . .
The Choice of an Index . . . . . . . . . .

ECONOMIC GROWTH AND THE EFFECT OF POLICIES
FOR IMPORT-SUBSTITUTION. . . . . . . . . .

Introduction . . . . . . . . . . . . . . .
The Issue Defined. . . . . . . . . . . . .
The Pattern and Effect of Import-

Substitution in LDC's. . . . . . . . . . .
Measuring the Level of Protection-Induced
Distortions. . . . . . . . . . . . . . . .

THE ANALYTICAL FRAMEWORK . . . . . . . . .

IntrOduction O O O O O O O O O O O O O O O
The Analytical Framework . . . . . . . . .

STATISTICAL ANALYSIS OF THE RELATIONSHIP

BETWEEN ECONOMIC GROWTH, EXPORT INSTABILITY,

.24
.24
.24
.25
.28
.31

.31
.32

.36
.42
.52

.52
.56

AND PROTECTION-INDUCED COST-PRICE DISTORTIONS 69

IntrOduction O O O 0 O O O O O O O O O O O
The Data 0 O O O O O O O O O O O O O O O O
The Regression Results . . . . . . . . . .

CONCLUSIONS. 0 O O O O O O O O O O O O O 0

iii

.69
.71
.76

.94

STATISTICAL APPENDIX . . . . . . . . . . . . 99
Sources. . . . . . . . . . . . . . . . . . . 99
Input Variables. . . . . . . . . . . . . . . 99
Definition of Terms. . . . . . . . . . . . .100

BIBLIOGRAPHY O O I O O O O O O O O O O O O .113

iv

LIST OF TABLES

No.

Measures of the Effect
Protection . . . . . .

O.L.S. Regressions . .
The Total Sample . . .
Input Data . . . . . .
Input Data . . . . . .

Input Data . . . . . .

of

Page

. 90
.101
.104
.107

.110

 

 

 

C»

x

INTRODUCTION

That international trade can play a propulsive role in
the economic expansion of developing countries has been
recognized by economists since the early classical times.
In more recent times, however, some economists, questioning
the applicability to currently less developed countries
(LDC's) of some of the conclusions of the classical,
comparative—cost, free trade theory, have expressed
pessimism regarding the prospects for export-led growth in
these countries; they have maintained that the present
structure of world trade does not offer much opportunity
for most of these countries to expand through strong re-
liance on foreign trade. A principal cause of concern,
which has been a continuing topic of international
discussion as well as the subject of post-World War II
conferences, is the short-term instability in the export
earnings of developing countries, which are, by and large,
exporters of primary products.

Paradoxically, the theory of trade implicit in the
arguments of some of these "trade pessimists" does not
suggest that trade is unimportant for the growth of
developing countries, but rather to the contrary. While
they do not necessarily reject the self-evident logic of

the welfare gains from trade flowing from the conventional

CC

a:

 

 

 

 

 

 

comparative-cost doctrine, these economists argue that
analysis of the long-run growth effect of trade in LDC's
should more meaningfully focus on certain rigidities in

the structure of trade and production found in these
countries. First, it has been argued that by virtue of its
low level of technological development, full utilization of
domestic resources in the typical growing LDC requires a
certain minimum of inputs which can only be imported. This
"import minimum" concept is of course not a peculiarity of
LDC's. Some industrially advanced countries sometimes also
need a minimum of imports (usually raw materials) which
cannot be produced domestically. But in addition to an
"import minimum," an "export maximum" is also said to

exist for LDC's to a degree not faced by developed

countries. The "export maximum,‘ for both primary products
and manufactures, comes about as a result of the particular
pattern of demand by the advanced countries for the exports
of the developing countries and the particular pattern of
resource endowments in LDC's. The implications for growth
in a situation where the "export maximum" persistently
tends to fall short of the "import minimum" have been
systematically formulated in terms of the now well—known
"two-gap" model associated with Professor Hollis B.
Chenery. A slightly different formulation of the same

idea by Professor Dudley Seers brings out clearly how the

rate of growth of the importing capacity of exports is, in

the long-run, the critical determinant of the rate of

 

 

 

 

 

an“.

 

output growth. In such a situation then, trade, far from
being an unimportant vehicle of growth, becomes a potential
"superengine" of growth. It is in the context of this type
of implied theory of trade that the possible consequences
of short-run instability in exports have been viewed.

Reasoning from Professor Seers' model, the conclusion
is inescapable that if the actual rate of export growth
falls short of the minimum required for any desired rate of
rapid growth without inflation, then import-substitution
must be commenced to reduce imports. It is generally known
that in the years following the Second World War, import-
substitution became a widely adopted strategy among LDC's
for the achievement of rapid growth and industrialization.
This strategy has entailed the encouragement of domestic
manufacturing behind a protective wall of tariffs, quotas,
exchange rates and other quantitative controls on trade.
But in the past few years, considerable disenchantment with
import-substitution appears to have set in. Some economists
have suggested that the cost~price distortions engendered
by protection in furtherance of import-substituting
industrialization may in fact have retarded the rate of
growth of developing countries during the period in
question.

The purpose of the present study is to examine
empirically the claims that (a) export-instability and
(b) protection-induced cost-price distortions, have been

detrimental to the growth of developing countries. The

t1

:1.

t}

 

 

 

?«

c\

tests are carried out within an analytical framework
developed from the theory of trade and growth implicit in
the theoretical arguments relating to the effects of export
instability and to the process of import substitution.

It is to be noted at the outset that the literature
on the export instability problem centers around two quite
distinct propositions which should not be confused:

(a) that because LDC's export mainly primary products whose
prices are subject to large and sometimes violent short-
term fluctuations, and also because of the high degree of
concentration of the exports of each country in a small
number of commodities and export markets, these countries
experience a relatively high degree of instability in their
export proceeds; and (b) that instability in export proceeds
is detrimental to prospects for long-run growth. The first
relates to the causes of export instability while the

second concerns the effect of export instability.

Much of the empirical work on the subject of export
instability has been aimed at establishing that most LDC's
in fact experience a high degree of export instability and
at determining the causes of instability. Concurrent with
these studies have been extensive efforts to devise a
variety of schemes for the international stabilization of
the export earnings of LDC's. Now, any effort to stabi-
lize export earnings must weigh the costs of such stabi-
lization against the possible benefits flowing therefrom.

To what extent, if at all, has export instability been

.r.

s

A:

9L

 

 

 

-—-——-‘--—

damaging to long-run growth in LDC's in general? Without
some answers to this question, it is difficult to reach
meaningful policy judgments regarding both the necessity
and sc0pe of any international stabilization policy. Only
very recently have a few attempts been made to investigate
empirically the relationship between export instability and
economic growth. These studies found no evidence of a
negative relationship. Unfortunately, there were a number
of serious statistical and methodological shortcomings in
these studies. The present study attempts to improve upon
previous work in this area by correcting for those short-
comings.

Chapter One presents a resume of the theoretical argu-
ments relating to the causes and effects of export
instability and reviews the empirical literature on the
subject. In Chapter Two we evaluate alternative index
measures of export instability and select the most
appropriate for this study. In Chapter Three we discuss

the process of import—substitution in developing countries,

outline the theoretical case against a strategy of import-

substituting industrialization, and develop two indices
measuring two different aspects of protection-induced cost-
price distortions. Chapter Four develops the theoretical
framework which is used in the statistical analysis of

Chapter Five. A summary and conclusions follow at the end.

 

 

r1“

I'D

;_J.

{u

 

 

CHAPTER ONE

THE EXPORT INSTABILITY PROBLEM

Introduction

 

Although appearing to have been at an unprecedently
high level during the past two decades, concern with the
problem of international trade instability is in fact not
new. Various publications of the League of Nations attest
to considerable preoccupation with the problem during the
period between the two World Wars.1 But the focus of
attention has shifted between then and now: whereas in the
earlier period interest centered on the impact of insta-
bility in raw materials trade on the post-World War I
reconstruction efforts of the industrial countries, concern
today is with the effect of export instability on the
growth prospects of LDC's. The more recent concern appears
to be an integral part of the wider, post-World War II
explosion of interest in the problems of economic develop-
ment in general, coincident with the emergence on the world
scene of an unprecedented number of new nation states in
the Third World.

The continuing interest in the problem of export
instability is reflected in the steady expansion of the

already vast literature on the subject, much of which

Hal“

1“

 

 

deals with various proposals for stabilizing the earnings

2 The empirical part of

of primary exporting countries.
this literature, though considerable, is actually the work
of a few individuals and organizations, most notably,

agencies of the United Nations,3 Coppock,4 Massell,5
McBean,6 and Michaely.7 The purpose of this chapter is to
outline the theoretical arguments regarding the causes and

effects of export instability and to review the empirical

literature on the subject.

The Causes of Export Instability

 

The claim that LDC's experience a relatively high
degree of short—run instability in their exports arises
from the composition as well as the concentration of those
exports. LDC's, by and large, are exporters of primary
products, which have low supply and demand price elas-
ticities as compared to manufactured products. For
primary commodities such as foodstuffs, low short-run
price elasticity of demand may arise from the fact that
they are necessities; have no readily available substi-
tutes; or, as in the case of commodities such as tea,
cocoa, coffee, and rice, are such that taste and custom are
a more important determinant of demand than relative
prices. In the case of raw materials, whose demand is
"derived" from that of the final product, low elasticity
may stem from the fact that raw materials constitute a

small part of total value so that even large changes in

 

 

 

 

 

"'9

m

'11

P
5

 

 

their prices hardly affect the prices of the final
products.

On the supply side, low price elasticity stems largely
from the fixed nature of output levels for both annual
crops and perennials in the short-run. Given low supply
and demand elasticities, autonomous shifts in supply or
demand produce disproportionately large fluctuations in
price. Excepting the infinitely rare cases where changes
in volume exactly offset the effect of price changes, price
fluctuations resulting from these shifts in supply and
demand inevitably translate into fluctuations in proceeds.

The effect on export proceeds of given variations in
supply depends on the elasticity of foreign demand. The
farther away this elasticity is from unity, the greater the
effect. The elasticity of foreign demand facing a country's
export commodity depends partly on the elasticity of the
total demand for the commodity, but more importantly on the
country's share of the world export of the commodity.8 A
country accounting for only a minute prOportion of the
world trade in a given commodity faces a virtually
infinitely elastic demand curve; on the other hand,
countries such as Ghana (cocoa), Brazil (coffee), and
Chile (capper) producing large proportions of the world
output of the given commodities are faced with demand
curves approximating those of world demand. In general,
the smaller a country's share of the total market, the

more elastic the relevant demand curve will tend to be.

 

r—O

r)

(I)

t1

(D

 

 

 

It follows then, excepting the rare cases where a country
is faced with a demand curve with elasticity less than
unity, that demand elasticity is farther away from unity,
the smaller the country's share in the world market. In
consequence, variations in supply can be expected to pro-
duce greater fluctuations in the export proceeds from a
given commodity the smaller the country's share in the
world market for that commodity.

With given elasticities, the effect of fluctuations in
foreign demand on export proceeds depends on the proportion
of the export commodity consumed domestically. The larger
this proportion, the greater the relative change in
proceeds for a given shift in foreign demand. This follows
because the larger the share of domestic consumption of
exportables, the greater will be the relative effect on
the value of exports of a given absolute change in value
(and thus of a given shift in demand).9 Fluctuations in
domestic demand will also affect export proceeds. Shifts
in the domestic demand curve, unless fortuitously offset
by corresponding shifts in the supply curve, will tend to
produce shifts in the net export supply curve. Hence
variations in domestic demand will tend to produce effects
on export proceeds similar to those of variations in supply.

From the foregoing, it follows that given structural
factors such as the proportion of the export commodity

consumed domestically and the country's share in the world

I

an!

a:

VI.

fly.

 

 

 

Cu

at

10

market for the commodity, short-run fluctuations in export
proceeds become a function of short-term shifts in supply
and demand.

For most LDC's, short—term instability in export
supply is attributable mainly to (l) crop-yield fluctua-
tions due to unpredictable weather conditions such as
droughts and floods, and other natural disasters such as
pest and plant disease epidemics; and (2) cyclical output
variations due to the cob-web effect.

Sources of instability on the demand side include
(1) cyclical fluctuations in income and output in the
develOped countries; (2) technological innovations which
alter the proportions of raw material inputs in manufactured
products; (3) random events such as the Korean War and the
Suez crisis; (4) sporadic changes in government policies
such as those relating to strategic stockpiling of raw
materials, surplus disposal programs, and trade and
exchange restrictions; and (5) speculative activity in
anticipation of future price and volume changes. (This
factor can, of course, be stabilizing as well as destabi-
lizing of exports.) The impact of income and output
fluctuations on the demand for primary products varies for
different types of commodities. Foodstuffs, which usually
show low income elasticities in high-income countries,
would tend to be less affected than industrial raw mate-

rials with higher income and output elasticities.

11

The instability of total export proceeds from two or
more commodities depends on the instability of individual
items and the degree of similarity between the commodities.
Commodities affected by dissimilar forces will tend to
fluctuate independently, thus increasing the likelihood
that such fluctuations will offset each other. Hence, the
more diversified the export base, the more stable exports
will tend to be. Export diversification can arise not only
in terms of the nature of the export commodities, but also
in terms of the geographic destination of exports. The
less concentrated a country's export markets, particularly
in the sense that markets are imperfect substitutes for one
another, the greater the likelihood that fluctuations
originating in different markets will cancel out each
other.

Empirical studies, particularly by agencies of the
United Nations, have abundantly demonstrated that most
LDC's experience short-term instability in their exports.10
With regard to the causes of export instability, Coppock,
Massell (1964), and McBean, using cross—section analysis,
found export instability to be positively correlated with
the degree of commodity concentration of exports and with
the proportion of exports derived from primary products,
and negatively correlated with per capita income and with
geographic concentration of exports.11 Michaely found a
non-significant relationship between fluctuations in

export prices and the degree of commodity concentration.12

 

 

 

L

'1'!

(1!

(I)

n’ (1'

n»

(I)

 

”A

r).

 

12

The explanatory variables used in these studies were
defined in a variety of ways, thus making it difficult to
obtain a precise comparison of the different sets of results.

In his more recent study, Massell (1970) employed a
highly comprehensive multiple regression model incor-
porating as explanatory variables the degree of commodity
concentration and of geographic concentration of exports,
the ratio of food items in exports, the ratio of raw
materials in exports, the prOportion of exports consumed
domestically, and the value of merchandise exports used as
a proxy for the export market share coefficient. These
variables are suggested directly by the theoretical argu-
ments presented above. In addition be included (a) per
capita income as "a rough measure of the level of economic
sophistication of a country, which in turn is an impor-
tant determinant of the types of goods produced." He
argued that a more SOphisticated country "tends to produce--
and export--products that are skill - and research-inten-
sive. . . A second point is that a high per capita income
is usually associated with greater flexibility. A more
flexible country is better able to shift resources among
products in response to (or in anticipation of) changes in
demand conditions, and thereby to reduce the impact of
sudden changes in demand;" and (b) a dummy variable dis-
tinguishing between developed and less developed countries.
This variable was used to test whether "there is a

difference in instability between DC's and LDC's over and

 

 

Q1

_1_4_)

r)

(I)

(I)

(TD

(3

ID
1'

’ 1

’1!

 

13

above what can be explained by the other variables." A per
capita income level of $600 was used as a convenient
dividing line between developed and less developed countries.

Using a total sample of 55 countries (19 developed and
36 less deve10ped), the author found a significant positive
coefficient for commodity concentration of exports and a
significant negative coefficient for the food item ratio
and for the value of merchandise exports (used as a proxy
for the export share coefficient). These findings, accord-
ing to the author, "are consistent with the view that
shifts in foreign demand have been a major determinant of
export instability."13 The geographic concentration ratio
showed a highly non-significant regression coefficient,
suggesting that the existence of trading blocs is an
insignificant factor in the presence of other determinants
of inter-country variations in the degree of export
instability. Although the mean instability index for the
LDC group was 50 per cent higher than that for the developed
country group, neither per capita income nor the dummy
variable distinguishing between deve10ped and less developed
countries was statistically significant. Apparently, the
level of development is not a significant factor when
other variables are held constant. The remaining explan-
atory variables also showed non-significant coefficients.

In their study which was not an attempt to "determine
the best possible explanatory equation for inter-country

predictions of the incidence of export instability," Erb

l4

and Schiavo-Campo focused on the relationship between
export instability and overall economic structure, where
structural variables were defined as "those character-
istics which are invariant except in the very long run."14
They noted that most international comparisons of export
instability have contrasted the "developed countries" to

the "less developed countries,‘ i.e., the group of
countries with a high per capita income to the group of
countries with low per capita income. But they argued that

"there are . . . no reasons to suppose that the ratio of

total income to population should, in and by itself, be

 

systematically related to the instability of export earn-
ings." While per capita income "is an aggregate indicator
which is often useful for research and policy purposes
whenever the variables summarized by it are not quanti-
fiable or do not have to be individually considered," if
the level of economic development is thought to bear some
relation to instability of export earnings, "it is necessary
to try and go beyond the simple statistical ratio of income
to population."15
Massell's findings (1970) discussed earlier are indeed
consistent with the notion of a lack of association between
per capita income and export instability. But the fact
remains that LDC's generally exhibit higher levels of
instability than deve10ped countries. To explain this,

Erb and Schiavo-Campo argued that it "is reasonable to

suppose that export instability is related to overall

15

economic structure" and that structural economic variables,
in turn, "partly depend on the absolute economic size of a
country, and particularly those characteristics of a less
deve10ped economy which are thought to have an influence on
the stability of exports." The distinct differences in the
degree of export instability observed between deve10ped and
less developed countries "may be related to the fact that a
world-wide comparison of 'relatively rich' and 'relatively
poor' is bound to be also a comparison between 'large' and
'small,' and in turn a comparison between countries with
different structural characteristics across the board."16
While recognizing that there is some overlapping between
the two groups of countries, the authors cited data showing
that the bulk of less developed countries have considerably
smaller economic size (measured by the level of absolute
income) than most developed countries.

It is well-known that the influence of market size
limitations on the composition of production and on growth
possibilities has been a principal focus of many studies of
economic development. Erb and Schiavo-Campo argued that
market size limitations are most probably directly
connected to the composition and destination of exports as
well to the composition of total production: "If the
domestic economy is heavily weighted towards agricultural
production; if diversification is prevented by the need for
productive units of minimum scale; if the absolute size of

the export sector is small; it is reasonable to presume the

l6

existence of those characteristics which are traditionally

argued to be at the basis of export instability--and mainly,

the geographic and commodity concentration of exports."17

Noting that empirical work has in fact shown a relationship
between country size (as measured by absolute GDP) and the
commodity concentration of exports, the authors then wrote:

On empirical as well as theoretical grounds
it thus appears that the main structural argu-
ments pointing to greater instability of exports
of less developed countries have as their
underlying cause the smaller absolute size of
less developed economies, with the resulting
characteristics of lower overall flexibility to
adjust to short-run external disturbances,
aggregate supply and demand obstacles to pro—
duction and export diversification, speciali-
zation in demand--and income--ine1astic export
commodities owing both to the small market and
to the lack of the variety of natural resources
necessary to begin production in more stable
sectors.

Absolute economic size can thus be taken as
an acceptable proxy variable for the structural
economic differences among countries, and one
would expect it to be inversely related to the
instability of export earnings.18

With a total sample of forty-four LDC's, Erb and
Schiavo-Campo performed contingency tests of the associa-

tion between size (measured by GDP) and instability of

merchandise exports and found a significant inverse relation-

ship for the period 1954-66. They did not find a similar
significant association when instability of total export

earnings rather than of just merchandise exports was used.
Total export earnings are of course the relevant variable

for most research and policy considerations.

17

These studies of the causes of export instability are
clearly important for both economic literature and inter-
national policy regarding ways and means of reducing
instability. But a decision to reduce export instability
must weigh the possible gains from reduced instability
against the costs of reducing instability. To what extent,
if at all, is export instability detrimental to long-run
growth in LDC's in general? Without an answer to this
question, it is difficult to reach meaningful policy
judgments as to both the necessity and scOpe of any schemes

for the international stabilization of the export earnings

of developing countries.

The Effects of Export Instability

 

A variety of arguments have been advanced by several
authors regarding the effects of export instability. These
arguments suggest that the adverse effects of export
instability are transmitted to an LDC economy at both the
micro- and macro-levels . 19

At the micro-level, the uncertainties associated with
excessive fluctuations in export earnings reduce the
attractiveness of devoting resources to production for
exports; the result is a retardation of the growth of the
export sector. An extension of the same argument is that
export fluctuations cause hedging behavior on the part of

farmers which induces them to devote resources to sub-

sistence farm production; this may lead to a perpetuation

18

of low productivity in agriculture.

The arguments relating to the macro-impact of export
proceeds fluctuations seem to be based on the theory that
most LDC's, in the absence of foreign aid, must rely for
their economic growth either directly on the savings-
investment process generated in the export sector, or
indirectly on the imports of development goods purchased

20

with exports. First, it is argued that due to frequent

fluctuations in export proceeds (and hence in the capacity

to import),21

an economy may be prevented from obtaining
the critical minimum level of imports necessary for optimal
growth. Secondly, uncertainties resulting from fluctuations
in the capacity to import may tend to reduce the planning
horizon and hence the effectiveness of develOpment policy.
Finally, it is argued that the national income of LDC's

is highly sensitive to the adverse effects of export
instability because these countries do not have at their
disposal effective counter-cyclical monetary and fiscal
policy instruments. An implication of this statement is
that export fluctuations, given the importance of export
income in LDC's, are expected to cause parallel fluctua-
tions in public revenues and in total domestic demand. In
the face of such cyclical fluctuations, governments may,
because it may be both politically and administratively
easier to handle deflationary then to handle inflationary

pressures, react only to the former; the result is that

short-term export fluctuations lead to a long-term

19

inflationary process.
The above arguments are based entirely on what McBean
has called "a combination of a priori reasoning and casual

empiricism."22

Few studies have been undertaken to date to
determine the empirical relationship between export in-
stability and economic growth. It was McBean who first
undertook an extensive study of the growth effects of export

instability.23

While not denying that short-term fluctua-
tions in exports could pose serious dangers for some
individual countries, he was led by his statistical find-
ings to challenge the validity of the hypothesis that
export instability is generally harmful to the growth of
LDC's. Like COppock in an earlier study, McBean found no
simple correlation between the size of Gross National
Product and export instability.24 But more importantly,
in a more comprehensive analysis, he found a non-signi-
ficant positive coefficient for the export instability
variable in a multiple regression analysis between the
growth rate of Gross Domestic Product (GDP) on the one
hand, and export instability, rate of growth of import
capacity, size of the export sector, and change in the
level of foreign exchange reserves, on the other, for a
sample of twenty-three LDC's for the period 1950-57.25
There are, however, a number of short-comings in
McBean's analysis. First, for lack of data on real GDP,

McBean used for his income growth rate the rate of growth

of GDP at current prices. Given the non-uniform rates of

20

inflation in different countries, serious biases were thus
introduced into the estimated growth rates. Secondly,
McBean used a measure of export instability-~an index

based on moving averages--which embodies serious limi-
tations (to be discussed in Chapter Two, below). In a very
recent paper, Kenen and Voivoidas have updated McBean's

26 Using a different export instability index

study.
(similar to the one to be used in the present study),
larger samples (30 and 50 countries), and longer time
segments (1950-66 and 1956-67), they, in effect, rep-
licated McBean's multiple regression analysis. The
authors, like McBean, failed to find any significant nega-
tive relationship between export instability and economic
growth.27

Kenen and Voivoidas used for their income growth rate
the rate of growth of GDP in current prices, "corrected for
price effects by deflation with a consumer price index."
This is clearly not a satisfactory procedure for obtaining
estimates of growth of real GDP. But a more important
weakness, shared with McBean's analysis, was the absence of
an explicit theoretical model explaining growth and thereby
justifying the particular assortment of explanatory
variables chosen. The present study attempts to improve
upon previous work by (1) developing an explicit model upon

which the statistical analysis is based; and (2) using more

comparable national income accounts data.

FOOTNOTES--CHAPTER ONE

See various issues of the League's World Economic
Survey and Review of World Trade of the period.

 

 

For the most recent of these proposals, see, United
Nations, International Compensation for Fluctuations
in Commodity Trade, New York, 1961; Measures for
International Economic Stability, New York, 1959;
International Bank for Reconstruction and Development,
Supplementary Financial Measures, TD/B/43, 1965;
Report of the United Nations Conference on Trade and
Develppment on its Second Session, TD/L/37, 1968.

 

 

 

 

 

 

 

United Nations, Instability in Export Markets of
Under-Developed Countries, New York, 1952.

 

 

Joseph D. Coppock, International Economic Instability,
New York: McGraw-Hill, 1962.

 

Benton F. Massell, "Export Concentration and Fluctua-
tions in Export Earnings: A Cross Section Analysis,"
American Economic Review, March 1964; "Export
Instability and Economic Structure," American Economic
Review, September 1970.

 

 

Allistair McBean, Export Instability and Economic
Development, Cambridge: Harvard University Press,

1966.

 

 

Michael Michaely, Concentration in International Trade,
Amsterdam: North-Holland Co., 1962.

 

For formulas relating to this and similar points, see
M.E. Kreinin, International Economics, Harcourt,
Brace, Jovanovich, Inc., 1971, pp. 353-55.

 

Cf. Massell (1970), loc. cit., p. 622. Massell notes
that although" . . . the values of the demand and
supply elasticities influence the effect of a shift

in demand, this influence is likely to be small.
Regardless of the elasticities, the price and quantity
effects reinforce each other."

21

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

22

See in particular, U.N. (1952), pp. cit.; International
Monetary Fund, "Fund Policies and Procedures in
Relation to the Compensatory Financing of Commodity
Fluctuations," I.M.F., Staff Papers, Nov. 1960.

 

 

Coppock, op. cit., Chs. 5 and 6, Massell (1964), loc.
cit.; and McBean, g3. cit., Ch. 2.

 

 

Michaely, op, cit., Ch. 5.

 

Massell (1970), 12;. cit., p. 627.

Guy Erb and Salvatore Schiavo-Campo, "Export
Instability, Level of Development, and Economic Size,"
Oxford University Institute of Economics and
Statistics Bulletin, Nov. 1969.

 

 

Ibid., p. 270.

Ibid., p. 271.

 

Ibid., p. 272.

 

Ibid., p. 272.

 

Most of these arguments are to be found in Ragnar
Nuskse, "Trade Fluctuations and Buffer Policies of
Low-Income Countries," Kyklos, Fasc. 2, 1958; and
Symposium II: "Stabilization and Development of
Primary Producing Countries," Kyklos,Fasc. 3, 1959.
McBean, gp. £i£., Ch. 1, presents a comprehensive
survey of the arguments.

This theory underlies the "two—gap" model of growth
discussed in Chapter Four, below.

Export earnings divided by an import price index give
a measure of the capacity to import based on exports
(i.e., the real purchasing power of exports). Cf.
G.S. Dorrance, "The Income Terms of Trade," Review of
Economic Studies, 1948/49. When total export earnings
are adjusted to include net foreign capital inflow
less net payments abroad to foreign owned factors of
production, and then deflated by an import price
index, a measure of the total capacity to import is
obtained. This gives the maximum level of imports
sustainable without compensatory financing.

 

 

McBean, 9p, cit., p. 31.

 

Ibid.

COppock, op. cit., Ch. 5; McBean, op. cit., Ch. 4.

 

 

25.

26.

27.

23

McBean, pp. cit., Ch. 4.

Peter B. Kenen and Constantine Voivoidas, "Export

Instability and Economic Growth," Kyklos, Fasc. 4,
1972.

Ibid.

 

CHAPTER TWO

THE MEASUREMENT OF EXPORT INSTABILITY

Introduction

 

This chapter offers a conceptual definition of the
term "export instability" as it is employed in this study;
reviews alternative index measures of export instability;
and selects the index considered most appropriate for the

subsequent analysis.

Definition of Export Instability

 

Variations in economic time—series are usually
attributable to one or more of the following: (1) secular
forces, which determined the long—run trend of the series,
(2) cyclical factors, which produce regularly recurring
changes over relatively short periods, (3) seasonal factors,
accounting for predictable changes within the year, and
(4) random forces, which are by definition non-recurring
and unpredictable.1 This study is concerned mainly with
the effect of short-term (cyclical and randomz) variations
in exports as distinct from the effect of the trend in
exports; as such, export instability is here defined as
(year-to-year) fluctuations of the export variable about a

general trend.

24

25

The Export Instability Index

 

Various indices of instability have been used by
different authors in the analysis of export instability.
Different indices possess different prOperties and the
purpose at hand is usually the sole criterion for deciding
which properties are the most desirable.

The simplest type of index is an unweighted arithmetic
mean of the yearly variations in exports. Such as index,

used by Michaely,3 can be written symbolically as

 

 

11 = 1 E IXt - Xt-ll
N - l Xt-l
where X = exports
N = number of years in the series
t = a given year

aslight variation of this index, used by the United

4 measures change relative to the terminal high

Nations,
point instead of the starting point, as is conventionally
done in computing percentage changes. The reason for the
different approach is to ensure that upswings and down-
swings are treated equally. Thus, whereas conventionally,
a change from 100 to 150 and back to 100 would be computed
as a 502 increase and a 33-1/32 decrease, the United

Nations method gives a 33-1/32 change in both cases. The

United Nations index can be expressed as

26

.-. 1' 'xt - xt'll
I2 N - 1 z

 

 

max(Xt, Xt_1)
where max(Xt, Xt_1) means the larger of Xt and xt-l'

An important property of both indices is that they are
highly dependent on the trend in X; they show a higher
degree of instability the greater the trend, for the same
variation around the trend. Since our interest lies
strictly in isolating the effect of variations about the
trend from the effect of the trend itself, these indices
are clearly unsuitable for our purposes. One notes in
passing that these indices have been used primarily in
studies of fluctuations in export prices, a variable in
which the trend was not considered of great importance.

A number of indices have been deve10ped which correct
for the influence of the trend factor. Naturally, these
indices vary according to the type of trend assumed as well
as the method used to correct for trend. One category of
these indices gives a measure of the average annual rate of
change, after correcting for the trend. Massell (1964)5
used the following index:

13 = i Z IUt+1 ’ “tl

N max(Xt, Xt_1)

where “t = Xt - Xt

 

and X = a + bt, i.e., the linear, least squares (trend)
value of X.

An exponential, rather than a linear, least squares value of

X can also be used to correct for trend in I3; in other

27

A

words, X can represent the log-linear, rather than the
linear, least squares value of X. A simplified form of
this procedure, using average logarithmic value of X rather

than the log-linear least squares value of X to correct

 

 

for trend was adopted by Coppock:6
I4 = anti log V108
where Vlog = 1 2 (log xt+1 - m)2
- l X
t
and m = 1 log xt+1
N - 1 X

The Coppock index embodies a serious limitation in that,

inasmuch as

 

m = 2 log t+1 1 log N
= N-l X1

it greatly depends on the extreme values of X.

A second category of trend-corrected indices gives a
measure of the average deviation of X from a fitted trend
line. Massell (1964) used a second index which he referred
to as the "normalized standard error," since it was the
standard error of estimate (the square root of the

unexplained variance) in a least squares regression of X

 

against t, divided by the mean of the observation:7
1 = 1 2 (ut)2
5 __ ‘________
x N
1 N
where X = __ 2 Xt, for t = l, 2, . . ., N.

28

Massell (1970) used the same index, with an exponential

rather than a linear trend fitted to the observations on

x.8

Still a third type of instability index corrects for

9 and the

trend by means of a moving average. Both McBean
International Monetary Fund10 used such an index with a

five-year moving average:

2
v
N - 4 Xt
l

5

I6 differs from the previous two categories of trend-

I
1 Ix -XI
16 t t

 

II
M

V
where Xt Xt+1, for t = l, 2, . . ., N.

corrected indices in that it makes no assumption as to the
specific shape of the trend. However, it has a number of
serious limitations. First, when the time-series is not
very long, the loss of observations at both ends of the
series could be a serious handicap. Secondly, there is an
element of arbitrariness as to the number of years to be
used in the moving average. Finally, as theoretical studies
of time-series have shown, the use of a moving average in
the elimination of trend will often produce artificial

oscillations not present in the original series.11

The Choice of an Index

 

As already pointed out, 11 and 12 are undesirable
because it is necessary for us to eliminate the influence of
the trend factor from our measure of export instability; we

require, in other words, a trend-corrected index. In view

29

of the criticisms raised against I4 and I6, we are left
with a choice between I3 and IS' Theoretically, both
measures appear to be equally satisfactory; empirically,
Massell (1964) found that both measures give equivalent
results in terms of country rankings. We have therefore
decided on 15 for the following non-fundamental reasons.
First, it is easier to compute. Secondly, inasmuch as
15 is a measure of average deviations from a trend line,
whereas 13 is a measure of the average annual rate of
change, corrected for trend, the former is more in con-
formity with our a priori definition of export instability.
As pointed out earlier, both linear and exponential
trends can be used to compute instability based on I5.
When we compute this index, we shall decide on which type
of trend to use on the basis of which gives the better fit

to the data for the period under study.

10.

11.

FOOTNOTES--CHAPTER TWO

See Frederick C. Mills Statistical Methods, New York:
Hold and Co., revised edition, 1938, pp. 228-231.

 

The use of annual data obviates the need to consider
seasonal variations.

Michaely, 2p. cit., p. 68.

United Nations, Export Instability in Export Markets
of Underdeveloped Countries, 2p. cit., p. 77.

 

 

Massell (1964), pp. cit., p. 50.

Coppock, 3p. cit., p. 24.

 

Kenen and Voivoidas, loc. cit., also used an index
which is a form of a normalized standard error
measure:

 

 

I:

xH C

where U is the standard error of the regression of
Xt on t and Xt_1.
This measure could be defined as

V

I = u

5
where U is the standard error of the following least
squares regression:

log(X) = a + bt
McBean, 2p. cit., p. 119.

International Monetary Fund, Staff Papers, "Fund
Policies and Procedures in Relation to Compensatory
Financing of Commodity Fluctuations," VIII, No. 1
(November 1960).

 

This is the so-called "Slutzsky effect." See E.
Malinvaud, Statistical Methods of Econometrics,
Amsterdam: North-Holland Publishing Co., Revised
edition, 1970, p. 428.

 

30

CHAPTER THREE

ECONOMIC GROWTH AND THE EFFECT OF

POLICIES FOR IMPORT-SUBSTITUTION

Introduction

 

It is generally known that in the years following the
Second World War, import-substitution (I-S) became a widely
adOpted strategy among the developing countries for the
achievement of rapid growth and industrialization. This
strategy was invariably pursued by the encouragement of
domestic manufacturing behind a protective wall of tariffs,
quotas, exchange, and other quantitative controls on trade.
In very recent years, however, considerable disenchantment
with I-S has been expressed. It has been suggested that
because of the market distortions produced by the devices
used to foster it, I-S may in fact have retarded economic
growth in many countries. Because concern about possible
adverse effects of export instability has been suggested
as one of the major reasons for the adoption of indus-
trialization programs in many LDC's, we consider it a
logical extension of our inquiry to examine empirically the
evidence relating to this proposition regarding the effect
of I-S. The purpose of the present chapter is to define
the hypothesis to be tested, outline the theoretical argu-

ments with respect to the relationship between economic
31

32

growth and policies1 for I-S, and develop indices to be

used in the test.

The Issue Defined

 

Albert Hirschman, in his analysis of the political
economy of the I-S process, notes that there have
historically been four possible motive forces behind import-
substituting industrialization: (l) wars and depressions,
which cause a sudden deprivation of foreign exchange and
imports; (2) balance-of—payments difficulties; (3) official
development policy; and (4) the growth of the domestic
market (as a result of export growth).2 But whatever the
original impulse for it, Hirschman has stressed the differ-
ent character and effect of I-S depending on whether it
occurs when foreign exchange is plentiful or scarce. In
the first case, I-S occurs as a result of export-propelled
expansion of the domestic market in areas of comparative
advantage or high productivity: "As incomes and markets
expand in such a country and some thresholds at which
domestic production becomes profitable are crossed,
industries come into being without the need for external
shocks or government intervention."3 This is not neces-
sarily true where I-S results from official policy in a
situation of foreign exchange constraint. In this case,
investment flows into industries on the basis of the degree

of protection or subsidization, not on the basis of

productivity.

33

The question of whether or not I-S is detrimental to
growth centers around the second type of ("forced") I-S
process. Is forced I-S necessarily a departure from optimal

"yes" according to the

growth policy? The answer would be
conventional (neo-classical) trade theory and policy. The
argument is essentially that given the existing inter-
national and inter-industry structure of comparative costs,
policies fostering I-S inevitably lead to a cost-price
structure detrimental to growth. The answer, however, is

in the negative if one accepts the foreign exchange gap
model (to be deve10ped in the next chapter) which is the
analytical framework adopted for the present study.
According to this model (or at least certain extensions of
it), the rate of growth of the importing capacity of exports
is in the long-run the critical determinant of the rate of
output growth. If the actual rate of export growth falls
short of the minimum rate required for any desired rapid
rate of output growth without inflation, then I-S must be
commenced to reduce imports.

It is to be noted that the question of the long-run
growth effect of policies fostering I-S is not to be con-
fused with the issue of the relative merits of an import-
substitution oriented development strategy as against an
export-promotion biased strategy. Under the foreign
exchange gap model, the only viable alternative to "forced"

I—S in a situation of foreign exchange constraint (and

34

assuming insufficient foreign aid) is "forced" export
promotion aimed at diversifying export activity into new,
non-traditional areas. Both types of policies are, from
the standpoint of conventional theory, merely counterparts
to each other since, in either case, resources will flow
into industries, not on the basis of comparative advantage,
but on the basis of the level of incentives provided by
official policy.

In a recent paper reporting on case studies of the
experience of ten deve10ping countries with exchange con-
trol and quantitative restrictions, Bhagwati and Krueger
found that ". . .countries which have had export-oriented
develOpment strategies appear, by and large, to have inter-
vened virtually as much and as 'chaotically' on the side
of promoting new exports as other countries have on the size
Of import-substitution."4 But the authors also noted that
". . .the economic cost of incentives distorted toward
export promotion appears to have been less than the cost
of those distorted toward import-substitution, and the
growth performance of the countries oriented toward export
promotion appears to have been more satisfactory than that
of the import-substitution oriented countries."5 Bhagwati
and Krueger suggested four possible theoretical explanations
for this finding. First, the costs of excess export
promotion are more visible to policymakers than are those

of I-S. This is because export promotion measures usually

35

involve subsidies and other incentives costly to the
government budget; hence there is a built-in restraint
against oversheltering.

Secondly, an export biased strategy often entails
relatively greater use of indirect interventions such as
taxes and subsidies, rather than direct interventions such
as price and distribution controls. The latter are likely
to be more costly to the economy. Thirdly, no matter how
much sheltering they receive in the domestic market,
exporting firms, unlike import-substitution firms, cannot
be completely protected from foreign price and quality
competition; hence they have a greater incentive to improve
productivity. Finally, where important economies to large
scale production exist, exporting firms are more likely to
exploit these by expanding the existing line of production;
by contrast, sheltered import-competing firms are more
likely to attempt to increase their monopoly positions by
diversifying into new lines of production.

The empirical validity of each of these arguments can,
in the final analysis, be determined only through further
case studies of individual countries. But the fact remains
that (a) from the standpoint of conventional theory, both
forced I-S and forced export promotion equally represent
deviations from optimal policy; and (b) both types of
policy are carried on simultaneously, in varying relative

degrees, in many developing countries. (Pakistan, where a

36

very high rate of protection of import-competing industries
is coupled with an equally high level of sheltering for
manufactured exports is only the most notable example.6)
Thus while, in a partial equilibrium framework, it may be
possible to isolate and compare the welfare cost of in-
centives distorted toward export promotion against that of
incentives distorted toward I-S, no theoretically meaningful
distinction, in a dynamic general equilibrium context, can
be made between the two types of distortions in terms of
their net effect on long-run growth. What may be stated as
a testable hypothesis is that cost-price distortions
associated with policies for I-S are detrimental to long-
run growth. To test this hypothesis, we shall develop in
the latter part of this chapter two indices measuring two
aspects of such distortions. In the next section we discuss
the essential features of the I-S process and outline the
theoretical case against the strategy of import-substituting

industrialization.

The Pattern and Effects of Import-Substitution in LDC's

 

Policies fostering I-S appear to have been adopted
originally in many countries in response to balance-of—
payments difficulties but were soon perceived to be
effective instruments for furthering a strategy of general
domestic industrialization.7 Certain well-known models of
the development process provide persuasive theoretical

justification for a program of general industrialization.

37

In these models, the expansion of the manufacturing sector
relative to the traditional agricultural sector raises the
share of profits in income and helps raise the saving rate.8
However, an important segment of development literature
suggests that adverse movements in international market
conditions for primary exports bear the major responsibility
for both the balance-of-payments difficulties and the
initial decision to embark upon industrialization in many
LDC's.9 In the first place, lagging exports, such as
commenced for many developing countries following the post-
Korean War shift in the world demand for primary exports,
directly create balance—of-payments difficulties. But at
the same time, this same lag in export growth (according to
models of structural inflationlo), as well as short-term
fluctuations in exports (through the process described in
the last section of Chapter One) also lead indirectly to
balance—of—payments difficulties by generating domestic
inflationary pressures. Given a system of fixed exchange
rates, domestic inflation leads to a deterioration in the
Payments balance by raisingcmsts in the export sector while
making importation more attractive. It is then in an effort
to escape this two-pronged assault that countries attempt,
through industrialization, to diversify their exports and

reduce their dependence on imports for the supply of manu—

factures.

38

A deliberate policy of industrialization in general
involves the expansion of the manufacturing sector at the
expense of agriculture which pays the cost of this expan-
sion in the form of lower product prices and higher input
prices. Import-substituting industrialization entails
essentially the protection of domestic manufacturing against
imports. This is carried out mainly by the imposition of
import tariffs, exchange controls and quantitative
restrictions on imports. Average nominal tariffs of around
100% on manufactures have been observed for such countries
as Brazil, Chile, and Pakistan.11 These nominal tariff
rates often given rise to even higher effective tariff
rates.12 The effective rate of tariff protection measures
the percentage excess of value-added in the production
process, resulting from the imposition of tariffs, over
free-trade value-added. The effective rate will exceed
the nominal rate if the tariffs on the final products
exceed those on the inputs.

By increasing the domestic price of imports, these
tariffs induced an increase in the domestic production of
importables and a reduction in the consumption of imports.
This substitution of domestically produced importables for
imports should lead to an improvement in the trade balance
unless the economy operates at such a level of efficiency
that production gives rise to negative value-added at

world prices.13 Such a situation has been found to exist

39

where, for instance, (a) due to, say, transport costs
or the monopolistic position of the seller, the foreign
purchase price of parts and components exceeds that of the
domestically assembled product; (b) the structure of
protection encourages the substitution of materials with
higher world market prices for those with lower prices; or
(c) the resource endowment of the country is not suitable
for the production of a particular commodity whose manu-
facturing is made profitable only by protection.14
The establishment of domestic manufacturing usually
starts with consumer goods and is gradually extended to
intermediate and capital goods as the economy develops.
This is because the cost disadvantage of domestic production
relative to imports is usually assumed to be less for
consumer goods than for the other types of goods which
require relatively more sophisticated techniques. The
process of I-S, far from reducing import requirements, may
often increase them. Seers gives a graphic description as
fOllows:
The immediate effect of accelerated indus-

trialization is to cause imports of capital

goods to rise more quickly than before. It is

true that, for any given industry, this extra

foreign exchange requirement soon ceases, and

one gets the benefit of import-substitution.

But in the first place, the saving of imports

is not as big as it seems because there will

be new demands for imported materials,

intermediate products, and spare parts, and

these imports will now rise more quickly.

Moreover, individual acts of import-

substitution must be seen as part of a
continuing process. Each year a start is made

40

on new industries, involving fresh require-
ments of imported capital goods; and progress
in substitution means entering into types of
production which are more and more capital-
intensive. Furthermore, as industrialization
proceeds, demands mount quickly for electric
power, internal transport, communications,
water supply, etc., which again raises the
requirements for imported-capital goods and
which causes imports of fuel to increase more
rapidly than [income] (in the case of countries
without their own petroleum).

It is true that in the end, by the time a
country is making its own equipment, import—
substitution will bring some reduction in the
prOpensity to import, but this may not be
achieved for two or three decades.

The pressure on the economy's capacity to import re-
sulting from this process is reinforced by the effect of
lagging export growth, fluctuations in exports, as well
as reduction in export earnings due to tariffs which may
have been imposed on inputs used in export industries. All
this leads to a progressive worsening of the balance-of—
payments. In the face of such difficulties, many countries
delay exchange rate adjustment too long, and also seem to
devalue too little. The result is a persistent over-
valuation of the exchange rate. Because it reduces the
domestic price equivalent of the proceeds received by
exporters, an overvalued exchange rate amounts in effect
to a uniform ad valorem tax on exports. To the extent that
incentives for I-S are not equally available for expanding
exPorts, and to the extent that the exchange rate is over-

Vflltmed, resources are induced to shift from the export to

the idnport-competing industries. Thus the I-8 process gives

41

rise to mutually reinforcing pressures on the capacity to
import. Sooner or later, policymakers realize that they
must try to increase export earnings in order to relieve
the increasingly intolerable foreign exchange constraint;
so they adopt a variety of export promotion measures
including subsidies to selected exports, rebates to excise
taxes, and "export bonus" schemes. The latter device
involves an arrangement whereby the exporter surrenders
only part of his foreign exchange earnings to the foreign
exchange authorities; the rest is retained for sale in the
free market or used directly to purchase imports. This
system of import tariffs and export subsidies amounts in
effect to a de facto selective devaluation of the currency
and partially offsets the impact of overvaluation.

The protection of domestic manufacturing against
imports may adversely affect long-run growth in a number
0f ways. First, there are the static or allocative costs
which result when the distortion of relative prices due to
protection interferes with inter-industry specialization
according to comparative advantage between primary pro-
duction and manufacturing as well as within the manufac-
turing sector itself. In principle, this static effect of
Protection is not necessarily inconsistent with Optimal
growth policy if these costs are more than made up for in

the .increased productivity of the protected industries over
time- This is the so-called "infant industry" argument for

prOtecztion.16 However, some authors have argued that in

42

deve10ping countries engaged in I-S, both the government
and producers view protection, not as a temporary measure
to be lifted when the industry matures, but as something to
be continued indefinitely. This expectation on the part of
firms that they will continue to be protected against
foreign and domestic competition leads to a second set of
(dynamic) costs in the form of opportunities foregone to
improve technology and fully exploit available economies of
scale. The latter problem progressively worsens as pro-
duction moves from the relatively easy stage of substi-
tution of consumer goods to the relatively more capital-
intensive intermediate goods and capital goods industries
which require increasingly large markets to achieve low
levels of unit costs. Given the limited size of the domestic
market and the lack of effective competition, domestic
manufacturing takes place at increasingly high unit costs.
Finally, protection adversely affects the capacity to
import through the discrimination against exports resulting
from the overvaluation of the exchange rate and tariffs on

imported inputs used in exports.

Measuringthe Level of Protection-Induced Distortions

 

Protection permits domestic industries to operate with
a value-added higher than that under free trade, thereby
creating incentives for productive resources to move into
protected industries. Devices used for such protection

include import tariffs, quotas, and other quantitative

43

restrictions on imports. The effective rate of protection
is the percentage excess of value-added in the production
process, obtainable by virtue of the imposition of tariffs
and other protective devices, over world market value-added.
In principle, other things being equal, the higher the
average effective rate of protection of manufacturing, the
higher the inducement toward import-substituting indus-
trialization; hence, for the purposes of our analysis,
intercountry variations in the average effective rate of
protection may be taken to represent intercountry variations
in the amount of incentives provided for I-S.

Because the computation of effective rates of pro-
tection requires types of information not readily available
for developing countries (e.g., input-output coefficients,
direct estimates of the difference between domestic and
world market prices), estimates are available only for a
few LDC's.17 In the absence of data on effective rates of
protection, proxies must be sought in terms of the type of
data available for developing countries.

A fundamental conclusion derivable from neo—classical
trade theory is that, in the absence of monopoly and of
restrictions on the free movement of goods or prices, the
effect of foreign trade in a country is to produce a conver-
gence between domestic and international cost-price ratios

for traded goods.18

The essense of import-substituting
industrialization is the protection of domestic manu-

facturing against imports through tariffs, exchange

44

controls, etc. As noted earlier, the process of I—S
generates pressures, to relieve which the government is
forced to intervene even more in the economy. The
cumulative result of this process is a progressive diver-
gence between domestic and international cost-price ratios.
This process may be expressed more precisely in terms of
implicit or effective exchange rates. An implicit exchange
rate is the ratio between the domestic wholesale price of a
commodity in local currency and the foreign price of the
same item, at the port of entry or exit, in some inter-
national currency. Just as one speaks of the average
effective rate of protection for manufacturing, one can
also speak of the implicit exchange rate for the manufac-
turing sector. The latter is some appropriately weighted
average of the implicit exchange rates for the individual
manufactured goods produced and/or sold domestically.19
The effect of protection in furtherance of import-
substituting industrialization is to increase progressively
the implicit exchange rate for manufactured goods. The
extent of this increase over time is a measure of the
degree of distortion over time in domestic versus inter-
national cost-price ratios.

As in the case of effective rates of protection, the
Computation of implicit exchange rates involves severe data
prolDlems. Information on the actual prices of individual
commodities in domestic and foreign currencies is not

86t1e213a11y available for developing countries. AS a proxy

45

measure of the change over time in the implicit exchange
rate for manufactured goods, we have developed an index

as follows:

Let Pd = domestic industrial wholesale price index in
Period 1.
Pf = world trade price index of manufactured goods
in Period 1.
V = index number of domestic manufacturing pro-

duction (value-added) in Period 1.
PA, PE, V. 8 Pd, Pf, V respectively in Period 2.
n = number of years in each period.
t = a given year.
All index numbers are given with reference to the same
base year.
The change between Period 1 and Period 2 in the excess of
domestic over world price valuation of domestic value-added
in manufacturing is given by
1 n n

A . 2 (v' . pgxpg>t - i 2 (v . Pd/Pf)t

n t n t

Because the time span between Period 1 and Period 2 for
which data are available may not be exactly the same for
all countries, A is "normalized" for all countries by

dividing by V, the mean of V between the two periods:

U is a pure number which may be positive or negative.

1

A second index which we have chosen to measure the

46

extent of domestic versus international cost-price dis-
tortions relates to a specific policy device--exchange
control. As was discussed in the preceding section,
balance-of—payments difficulties have been both the moti-
vating cause for the adOption of, and a by-product of,
policies for I-S in many LDC's. Faced with balance-of-
payments deficits, many countries are reluctant to adjust
their exchange rates, but instead adopt exchange control
programs to ration scarce foreign exchange and keep
international obligations within limits. The result is a
persistent overvaluation of the exchange rate. Because it
discriminates against exports, the latter only intensifies
the foreign exchange scarcity. But as the bottleneck
created by the foreign exchange constraint becomes
increasingly felt, policymakers realize they must try to
increase export earnings; so they adopt a variety of
export promotion measures in the form of subsidies, rebates
to taxes, etc. These measures amount in effect to a de
facto devaluation of the currency and partially offset the
impact of overvaluation.

Since the true equilibrium exchange rate in a system
of trade and exchange controls is not known, the extent to
which the official exchange rate is overvalued relative to
the true shadow price of foreign exchange is hard to
estimate in practice. As a proxy, however, we have chosen
to use the ratio of the black market to the official

exchange rate:

47

n
X (eh/ef)t
t

112:;
I1

where eb black market exchange rate

ef official exchange rate

Pick's Currency Yearbook defines black market rates as the

 

"unofficial, usually illegal prices of particular currencies,
mostly in terms of U.S. dollar value. Listing the true or
'real' worth of monetary units, which are subject to
controls, these rates reflect unrestricted supply and
demand."20
Although U1 and U2 are both measures of cost-price
distortions associated with trade/growth policies, it is
important to note a fundamental difference in the type of
distortions they measure. The extent of overvaluation of
the exchange rate (U2) is determined largely by official
policy. Like the effective rate of protection, it is
more or less an objective measure of the amount of incen-
tives provided by official policy in an effort to effect
the particular pattern of resource allocation necessary for
I-S. By contrast, while the divergence over time between
domestic and international manufacturing price levels may
have been caused by official policy, the extent of this
divergence (which is what_U1 measures) depends not just
on the amount of incentives provided by official policy
but also on how and to what extent economic units within
and without the manufacturing sector respond to these incen-

tives as well as on endogenously determined changes in

48

relative productivities. Conceptually, then, U2 is the
more appropriate proxy for the level of effective rate of
protection.

Since exchange control is only one of several possible
protective devices, U2 can only be a partial measure of
the amount of incentives provided by I-S policies. Never-
theless, as would be suggested by numerous theoretical and
empirical studies,21 exchange control is a sufficiently
important policy instrument so that an examination of the
growth effect of overvaluation of the exchange rate becomes
important in and of itself. Of even greater importance,
perhaps, is an examination of the effect of the type of
distortion measured by U1. The relationship between
economic growth and inflation has been a central focus of
the debate about the import—substituting strategy of

industrialization.22

U1, as defined, can, strictly speaking,
be viewed as an index of the rate of domestic industrial

inflation relative to the rate of world industrial in-

flation.

FOOTNOTES--CHAPTER THREE

We consider it important to call attention to a certain
ambiguity that is unavoidable in the use of the word
"policy" in the discussion of import-substitution.

The word seems to us to have two distinct connotations
which should not be confusing as long as careful
attention is paid to the context in which it is used.
One can talk about the policy of (I-S) in the sense

of the strategy of replacing imports with domestic
production. Policies for import-substitution, on the
other hand, would mean policy devices--tariffs, quotas,
etc.--which foster I—S.

A. Hirschman, "The Political Economy of Import-
Substituting Industrialization in Latin America,"
Quarterly Journal of Economics, February, 1968;

 

reprinted in C.T. Nisbet, ed. Problems in Economic
Development, New York: Free Press, 1969, pp. 237-266.

 

Ibid., p. 240.

J. Bhagwati and Anne Kureger, "Exchange Control,
Liberalization, and Economic Development," American
Economic Review, Papers and Proceedings, May 1973,

 

p. 420.

Ibid., p. 420.

 

See R. Soligo and J. Stern, "Tariff Protection,
Import-Substitution and Investment Efficiency,"
Pakistan Development Review, Summer 1965.

 

Cf. Hirschman, 32. cit.

 

W. A. Lewis, "Economic Development with Unlimited
Supplies of Labour," Manchester School, May 1954; G.
Ranis and J. C. Fei, "A Theory of Economic Development,’
American Economic Review, Sept. 1961.

 

 

49

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

50

See especially, Raul Prebisch, "The Economic
Development of Latin America and its Principal
Problems," reprinted in Economic Bulletin for
LathaAmerica,U. N., Economic Commission for Latin

 

 

America, Feb. 1962; Hans W. Singer, "The Distribution
of Gains Between Investing and Borrowing Countries,"
reprinted in R. Caves and H.G. Johnson, ed. Readings
in International Economics, Homewood: R.D. Irwing,

 

Inc., 1968.

In a conference on inflation and growth in Latin
America, Dudley Seers was chosen to represent the
"structuralist" school of economists. See W. Beer
and I Kersternetzky Inflation and Growth in Latin
America, New Haven: Yale University Press, 1964.

A part of Seers' model of structural inflation will

be developed as part of the analytical framework in
Chapter Four, below.

See B. Balassa, et al., The Structure of Protection
in DevelOping Countries, Baltimore: The Johns-Hopkins

 

 

Press, 1971.

Effective rates ranging up to lOOOZ were found for
some industries within the East African Community.
See Richard Reimer, "Effective Rates of Protection
in East Africa," Eastern Africa Economic Review,
December, 1971.

 

S. Guisinger, "Negative Value Added and The Theory of
Effective Protection," Quarterlprournal of Economics,
August 1969.

 

See for example Soligo and Stern, loc. cit.

 

Dudley Seers, "A Theory of Inflation and Growth in
Underdeveloped Countries," Oxford Economic Papers,
June, 1962, p. 179.

Cf. Murray C. Kemp, "The Mill-Bastable Infant Industry
Dogma," Journal of Political Economy, Feb. 1960.

 

See Balassa, et al., pp. cit.

Cf. G. Haberler, "Some Problems in the Pure Theory of
International Trade," reprinted in Caves and Johnson
ed. Readings in International Economics.

 

Cf. Stephen Lewis, "The EffectschTrade Policy on
Domestic Relative Prices," American Economic Review,
March 1968.

 

 

20.

21.

22.

51

Pick's Currency Yearbook, Pick Publishing Co.,
1972, p. 13.

 

See for example Anne Krueger, "Some Economic Costs
of Exchange Control: The Turkish Case," The Journal
of Political Economy, October, 1966; Harry G. Johnson,

 

 

"The Welfare Costs of Exchange Stabilization," The
Journal of Political Economy, October, 1966; R. N.

 

 

C00per, "Devaluation and Aggregate Demand in Aid-
Receiving Countries," in J. Bhagwati, et al., ed.
Trade, Balance of Payments, and Growth, Amsterdam: 5“

 

North-Holland Co., 1971.

See Baer and Kersternetzky, 2p. cit.

 

 

 

CHAPTER FOUR

THE ANALYTICAL FRAMEWORK

Introduction

 

The basic purpose of the present study is to examine
empirically the effect on long-run economic growth of
export instability and of cost-price distortions due to
policies for import-substitution. Essentially, an index
of export instability and indices of these cost-price
distortions will be used to "explain" intercountry
variations in growth rates. If export instability and
protection-induced distortions affect long-run growth, they
can only be two of many other sets of factors that do so;
consequently, a statistically meaningful analysis requires
that the influence of these other factors be held constant.

The purpose of the present chapter is to identify the
variables which should be held constant in testing for the
growth effect of export instability and of protection-
induced cost-price distortions. To do this in a systematic
manner, three explicit models are utilized within the
framework of the theory of trade and growth implicit in the
hypotheses under investigation. The variables so selected
constitute the basic regression model employed in the

statistical analysis of the next chapter.

52

 

53

As was noted in Chapter One, the arguments supporting
the hypothesis that export proceeds instability is
detrimental to growth are based largely on the theory that
most developing countries, in the absence of foreign aid,
must rely for their deve10pment either directly on the
savings-investment process generated in the export sector,
or indirectly on the imports of development goods purchased
with exports. It is this same theory that underlies the
so-called "two-gap" model which has come into widespread
use during the last decade in the analysis of the growth
problems of developing countries.1 The model views the
development process in terms of two scarce factors, foreign
exchange and capital. Over the relatively short period, the
two factors are complementary and their availabilities
constitute separate limits to growth. With a given struc-
ture of demand and production, a trade constraint exists
where the earnings from exports are insufficient to purchase
the minimum level of imports required to achieve a given
rate of output growth. In the absence of external capital
to fill this (trade) gap, output growth is limited by the
availability of export earnings.

Similarly, a savings constraint arises where the level
of savings is insufficient to realize the level of invest-
ment necessary for the desired rate of output growth. In
the absence of an inflow of foreign capital to close this

(savings) gap, output growth is constrained by the

 

54

availability of domestic savings. In theory, at any given
point in time, only one of the two constraints is operative.
Empirical work by Chenery and various associates, based

on the above theoretical model, has suggested that the

trade gap is the limiting constraint for many LDC's at

T

this time.2 Theoretically, for the ex ante trade gap to T
exceed the ex ante savings gap implies that increased
savings cannot be converted into capital formation. This

assumption is on the face of it highly rigid, and some have is
3

 

questioned its general empirical validity.
Besides being used to estimate the foreign exchange

resources needed for projected rates of output growth,

the two-gap model has also been employed in a more disaggre-

gated form to spotlight the economic planning issues

implied in the divergence between the ex ante trade and

savings gaps.4 We are not concerned in the present study

directly with planning or programming issues; however,

since the two—gap theory is implicit in the arguments

supporting the hypothesis of a negative relationship

between export instability and economic growth, the two-

gap model provides a convenient and logical starting

framework for our analysis. The logic of the two-gap model

will be utilized to derive a reduced form relationship

between out-put growth and certain explanatory variables.

But because of some highly rigid assumptions implicit in

the two-gap model, the growth equation so derived will be

55

modified and extended to incorporate variables which give
it more general applicability.

Before proceeding to present the analytical frame-
work for the study, we digress briefly to note the nature
of the empirical literature on economic growth. A careful
examination of this literature shows that while there have
been extensive studies of economic growth for individual

5 very little quantitative analysis has been

countries,
done on a cross-section basis. Among the notable excep—
tions are (a) Kuznets,6 whose work involved the accumula-
tion of various types of data concerning growth, the cal-
culation of certain average and incremental ratios, and
the classification of countries into groups; and (b)

9 who have

Denison,7 Hagen and Hawrylyshyn,8 and Robinson,
attempted to identify the sources of economic growth by
disentangling the interrelationships between output and

such inputs as the labor force, the capital stock, and the
rate of technological change. The general procedure has
been to fit an aggregate production function and to assign
relative weights to the various factor inputs. Following
Solow's pioneering use of this approach in historical
studies of technical change in the United States, the

rates of growth of all the relevant factors are calculated
and then multiplied by their respective growth elasticities.

The difference between the actual and the estimated rates

of growth--the "Residual"--is attributed to technological

change.10

 

56

In the next section the production function model as
well as a model of structural inflation will be utilized to
modify the growth equation derived from the simple two-gap

model.

The Analytical Framework

 

The Two-Gap Model: The relationship between output

 

and exports under the trade-constraint hypothesis arises
from the dependence of investment on the imports of capital

goods made possible by foreign exchange available from

 

exports and external capital. The argument may be summa-

rized as follows:

(1) AYt = i It
k
(2) It = bME
(3) ME = clxt + czrt
where Y = output
I = investment

MK = capital goods imports

X = importing capacity of exports

F = net foreign capital inflow

k = the incremental capital-output ratio

t = time subscript
b, c1, c2 = constant parameters
Equation (1) relates capacity output to investment in
capital stock via an incremental capital-output ratio. It

follows from the well-known Harrod-Domar production function

57

(Y = i Kt, where K = capital stock) in which capital and

t
labor :re assumed to be employed in fixed proportions, and
the supply of labor is always sufficient to engage all
available capital; output growth is thus in effect deter-
mined only by the capital input. In equation (2), invest-
ment depends on capital goods imports which, in equation
(3) depend on real export earnings and external capital.

Solving equations (1), (2), and (3) simultaneously for

AYt and dividing throughout by Yt we get:
AYt

Yt Yt

(“Gr—- =a1§s+a235

Yt

where GY = the growth rate of output

 

31 = bcl
k
a2 = bcz

 

It is clear that (4) is a reduced form equation, and a1 and
a2 the reduced form coefficients which incorporate the
effects of the endogenous variables, I and MK. For given
values of X/Y and F/Y, the growth rate will be higher the
larger the capital imports coefficient, b, and the smaller
the incremental capital output ratio. In the two-gap
approach of looking at the equation, it is of course the
structural coefficients which are taken as given so that
changes in output result from variations in the exogenous

variables, X/Y and F/Y.

 

58

In the savings-constraint hypothesis, the relationship
between output and exports is established through the ex
pggg equality of investment to total (domestic plus
external) savings; and the dependence of some of domestic
savings on the surplus extracted from the export sector
through government taxes or public institutions such as

marketing boards:

(5) AYt = i It
k

(6) It St + Ft

(7) St = 31 (Yt - Xt) + 32Xt
= SlYt + (32 - 51)Xt
where S = gross domestic savings

31 = the rate of saving out of the non-export sector
52 - the rate of saving out of the export sector
Solving equations (5), (6), and (7) simultaneously for

AYt and dividing throughout by Yt, we get:

 

(8) GY=£E=b1+b2x_t+b3_F_t
where bl = :l
k
b2 = (82 - 81)
k
b3=i
k

Again, (8) is a reduced form equation, and b2 and b3 the
reduced form coefficients. For given values of the exog-

enous variables, the growth rate of output will be higher,

 

59

the smaller the incremental capital—output ratio, and the
greater the positive difference between the rates of saving
out of the export and non-export sectors. The hypothesis
that the rate of saving out of the export sector is greater
than from the non-export sector, first suggested by Maizels
in the context of the two-gap analysis, received some
support from his time-series analysis of eleven British
Commonwealth countries.11 A more recent study of twenty-
two countries by Lee also tended to support it.12

It is interesting to note that, viewed as linear
regression models, equations (4) and (8) are identical. In
other words, the trade-constraint and savings-constraint
hypotheses give rise to equivalent reduced form relation-
ships between the growth rate of output on the one hand, and
X/Y and F/Y on the other. In terms of economic inter-
pretation, however, there is a difference, relating to the
coefficient of X/Y in the savings-constraint hypothesis. A
significant, negative coefficient could mean either or both
of two things: (a) that the Maizels hypothesis mentioned
earlier regarding the differential rates of saving out the
export and non-export sectors does not hold, and (b) that,
given the structural relationships envisaged in the model
(namely, that capital formation is the only source of growth
and that external capital can be used to take the place of
domestic savings), if imports have been used for consumption

‘purposes rather than for capital formation, then investment

 

60

and exports become variables competing for scarce growth
resources.

The growth model we have derived so far, i.e., either
equation (4) or (8), is one in which the rate of growth is
explained by only the relative availabilities of real
export earnings and net foreign capital inflow. This rs
follows necessarily from the rigid assumptions of the model:
(a) the strict factor proportions production function which

considers capital the only scarce domestic factor, and (b)

 

that while foreign capital can be used to supplement domestic
capital in the case of a savings constraint, domestic saving
cannot be used to relieve a foreign exchange constraint.
These assumptions are not unrealistic for the empirical
analysis of the short-run growth Options facing many LDC's,
particularly the labor-surplus countries. But for the
purpose of a long-run growth analysis with general appli-
cability, the model must be modified to take into account
the possibility of (a) some flexibility in the use of
domestic savings, and (b) substitutability between capital
and labor. These modifications are made in the next two
sub-sections respectively by (l) incorporating the long-run
‘relationship between output growth and the capacity to
import, and (2) replacing the factor proportions assumption

witfii the more general neo-classical production function.

61

A Model of Structural Inflation: As pointed out
earlier, the assumption of a lack of substitutability
between domestic and foreign savings embodied in the two-
gap model has been criticized by several authors. But in
actual fact, Chenery and Strout, in their article, which
appears to be the most complete statement to date of the
two-gap philosophy, did recognize the extreme restrictive-
ness of this assumption and accepted the possibility that

in the long run, domestic savings could be substituted for
13

 

foreign savings through import-substitution. But if one

assumes, as did Chenery and Strout, that maximum foreign aid
is forthcoming, then there is never an immediate pressure

to institute import-substitution; hence the authors did not
dwell at length on this question.

In a recent article, Susan Cochrane14 has pointed out
the strong similarities between the two-gap model and
Dudley Seers' model of structural inflation in LDC's.15
Both models recognize the importance for output growth of
the demand for imports; both are also pessimistic about the
prospects for future export growth. However, Seers, in whose
nuniel emphasis is placed on the long-run trend rather than
the short-run variations in exports, assumed, unlike
Chenemy and Strout, that no foreign aid will be forthcoming.
(In View of the actual experience of many LDC's, this would

seem to be the more realistic assumption.)

With this assumption, Seers proceeded to derive the

62

long-run relationship between output growth and imports as
follows. Income growth, with the urbanization that
invariably accompanies it, causes a very rapid increment in
the demand for manufactured consumer goods. Given the high

income elasticity of demand for these goods and the changing

f7

tastes due to urbanization, this demand grows faster than
income. Since the desired consumer goods can be imported
or produced domestically, Seers considered two situations:

First, one in which a constant percentage, m, of these ’4

 

manufactured consumer goods is imported, and second, one in
which an attempt is made to reduce m through import-
substitution.

In the first case, if the demand for manufactured
consumer goods grows faster than income, and m is constant,
the consumption of imported consumer goods rises more
rapidly than income:

A(mC)t > AYt
(mC)t in:
Similarly, if the economy is unable to substitute domestic
savings for imported components, then m', the import require—
ment for domestic production of imported consumer goods,
will remain constant. Therefore importation of intermediate
goods will rise faster than income:

A(m'(l - m)C)t > AYt
(m'(1 - m)C)t Ha:

 

Seers then concluded that since there is no reason to expect

other types of imports to grow more slowly than gross output,

63

total imports (M) will also grow faster than output.16
Thus

AMt > AYt

Mt Yt

Ignoring capital inflows or taking them to be constant, this
conclusion implies that for long-run growth without

inflation to prevail, exports must grow faster than output:

> AYt

M Yt

If export growth should lag behind this equilibrium rate of
growth, then an attempt must be made to reduce imports
through import-substitution.

From the immediately foregoing, we see that in the
long run, the rate of growth of exports becomes a critical
determinant of the rate of economic growth;17 so we modify

our growth model accordingly to get:

Y Y

where GX the growth rate of real exports

bi = constant parameters

The Production Function Model: The simplest dynamic
model of growth is derivable from a general aggregate
production function:

(10) Y = f(K,L)
where L = labor force

Differentiating with respect to time, and dividing throughout

F

 

64

(11> dY._=5_Y.d_K-i+§3-d_L.i
dt Y 6K dt Y 6L dt Y
= dY . . . dK + 6Y . . . dL

6K Y E dt 6L Y 3 dt

(where d and 6 represent total and partial derivatives
respectively)

Rewriting (11) with more convenient notational symbols, we

get:
(12) GY = aGK + bGL
where GY = E: . i = growth rate of Y
dt Y
GK = if . i = growth rate of K
dt K
CL = if . i = growth rate of L
dt L
GY K .
a = __ . _ = elasticity coefficient of Y with re-
6K Y
spect to K
dY L
b = __ . _ = elasticity coefficient of Y with re-
5L

spect to L
Because of data limitations, the investment rate, I/Y, is
customarily used in actual estimation in place of the
grrnvth rate of capital stock, GK' For the same reason,

the growth rate of total p0pu1ation, CD,

for: the growth rate of the labor force, GL.18

The production function model can now be written as

is used as a proxy

Fa.

 

65

(13) CY = a' I + b'GN

By a straight-forward combination of equations (9) and (13),

we get

(14) GY=81;+Ban+B3Gx+BA:+BS;

where 81 = constant parameters.
In equation (14), the roles of domestic capital and labor,

and substitutability between the two, are explicitly

allowed for. This completes our modification of the

 

embryonic growth equation derived initially from the
simple short-run, two-gap framework. Equation (14) is the

basic regression model to be used in the statistical anal-

ysis of Chapter Five.

FOOTNOTES--CHAPTER FOUR

Among the most important works are Hollis B. Chenery
and M. Bruno, "Development Alternatives in the Open
Economy: The Case of Israel," The Economic Journal,
June, 1962; Chenery and A. Strout, "Foreign Assistance
and Economic Development," American Economic Review,
September 1966; McKinnon, "Foreign Exchange Constraints
in Economic Development and Efficient Aid Allocation,"
The Economic Journal, June 1964; and J. Vanek,

 

 

 

Estimating Resources Needs for Economic DevelOpment,
New York: McGraw-Hill, 1967.

Chenery and Strout, loc. cit., and Chenery and P.
Eckstein, "Development Alternatives for Latin America,‘
Journal of Political Economy, July/August 1970.

 

See H. Bruton, "The Two Gap Approach to Aid and Develop-
ment; Comment, " American Economic Review, June, 1969;
and R. Nelson, "The Effective Exchange Rate; Employ-
ment and Growth in a Foreign Exchange-Constrained
Economy," Journal of Political Economy, May/June 1970.

 

 

See Chenery and Bruno, loc. cit; and Vanek, 2p. cit.

 

 

These have been mostly quantitative analyses in the
context of linear programming models. Among the most
notable are Chenery and A. McEwan, "Optimal Patterns
of Growth and Aid: The Case of Pakistan," Pakistan
Development Reviews, (Summer, 1966); H. Bruton,

 

"Productivity Growth in Latin America," American
Economic Review, December 1967; and M. Bruno,

 

"Estimation of Factor Contribution to Growth Under
Structural Disequilibrium," International Economic

Review, February 1968.

Simon Kuznets, "Quantitative Aspects of the Economic
Growth of Nations," I through X, Economic Development
and Cultural Change, various issues from October 1956

 

 

through January 1967; also Modern Economic Growth:
RateL Structure and Spread, Yale University Press,

 

 

1966.

66

 

10.

11.

12.

l3.

14.

15.

16.

17.

67

E.F. Denison, Why Growth Rates Differ, Brookings
Institute, 1957.

 

E.E. Hagen and O. Hawrylyshyn, "Analysis of World
Income and Growth, 1955-65," Economic Development
and Cultural Change, October 1969, Part II.

 

 

Sherman Robinson, "Sources of Growth in Less Developed
Countries," The Quarterly Journal of Economics,
August 1971.

 

R.M. Solow, "Technical Change and the Aggregate Pro-
duction Function" The Review of Economics and
Statistics, August 1957; "Technical Progress, Capital
Formation, and Economic GrowthJ'American Economic
Review, September 1962.

 

 

 

Alfred Maizels, Exports and Economic Growth of
Developing Countries, Cambridge University Press,
1968, Chap. 3.

 

 

J.K. Lee, "Exports and the Propensity to Save in
LDC's," Economic Journal, June 1971.

 

Chenery and Strout, loc. cit., p. 697.

 

 

Susan Cochrane, "Structural Inflation and the Two-
Gap Model of Economic DevelOpment," Oxford Economic
Papers, November 1972.

 

Dudley Seers, "A Theory of Inflation and Growth in
Underdeveloped Economies," Oxford Economic Papers,
June 1962; "Inflation and Growth: The Heart of the
Controversy," in W. Baer and I. Kevstenetzky eds.
Inflation and Growth in Latin America, New Haven:
Yale University Press, 1964.

 

 

 

Seers, "A Theory of Inflation . . . ," loc, cit.,
p. 177.

Being essentially an "orthodox-money or gold standard"
model, as Seers himself refers to it, this model
explicitly ignores monetary implications. In
principle, of course, policymakers could deliberately
attempt to restrict the growth in money supply to a
rate lower than the growth rate of exports, thereby
causing income to grow at a lower rate than is poten-
tially possible with the given rate of growth of
export earnings. However, Seers considers this un-
likely since "unskilled wage—earners and existing firms
are assumed to have some defensive strength," Ibid.,
p. 181.

18.

68

The use of I/Y has some firm theoretical justifi-
cation if one assumes a constant capital—output ratio.

Thus

x|>< mlcx
n7:

where h

and since

H

ER? H?
average capital-output ratio

dK by definition.

5?

The use of CD as a proxy for GL is rationalized on the

assumption that the labor force grows in proportion

with total population.

Cf. Hagen, loc. cit.

 

CHAPTER FIVE

STATISTICAL ANALYSIS OF THE RELATIONSHIP BETWEEN
ECONOMIC GROWTH, EXPORT INSTABILITY,
AND PROTECTION-INDUCED

COST-PRICE DISTORTIONS

Introduction

 

In this chapter we test statistically for the existence
and significance of a negative relationship between economic
growth on the one hand, and on the other, export insta-
bility and cost-price distortions due to policies for
import-substitution. Two indices measuring aspects of these
distortions were developed in Chapter Three.

The reasons for expecting export instability to affect
growth adversely were outlined in Chapter One. In Chapter
Three, it was noted that many economists believe that
industrialization programs were embarked upon in many LDC's
in an effort to escape possible adverse consequences of
international trade instability, both short-run and long—
run. Industrialization was pursued through the use of
various protective devices to encourage domestic manu-
facturing in replacement of imports. Also discussed were
various ways in which protection in furtherance of a

strategy of import-substituting industrialization could

69

7O

adversely affect long-run growth. The commencement of
import-substitution, it was noted, instead of reducing
import requirements, tends actually to escalate them. Con-
tinuing fluctuations in export earnings, together with a
lagging trend in export growth, serve only to intensify the
pressure on the capacity to import. Faced with deepening
balance-of—payments difficulties, the authorities may react
in ways which result in a persistent overvaluatipn of the
exchange rate. _Since the latter discriminates against
exports, it further exacerbates the foreign exchange con-
straint. As the bottlenecks created by the foreign
exchange scarcity become increasingly felt, policymakers
are forced to resort to further interventions in the economy
in an attempt to increase export earnings through various
export promotion measures. The cumulative result of this
spiralling process is bound to be a progressive divergence
over time between domestic and international cost-price
ratios. The indices developed in Chapter Three were designed
to measure (a) the extent of overvaluation of the exchange
rate, and (b) the change over time in the excess of domestic
over world price valuation of domestic value-added in
manufacturing. The export instability index to be used in
this study was discussed in Chapter Two.

The procedure chosen to test for the growth effect of
export instability and of protection—induced cost—price

distortions is to include the indices measuring them as

71

additional variables in a basic regression model that
explains growth. This model, developed in Chapter Four,
incorporates the essential determinants of growth in LDC's
implicit in the theoretical arguments relating to the effects
of export instability and to the process of import-

substitution.

The Data

 

Data on the expenditure components of GDP in constant
(1960) U.S. dollars were available for 65 developing coun-
triesu' The data were supplied by the Statistical Office of
the United Nations, and the countries appear to span the

1 For most of the countries in the

entire Spectrum of LDC's.
sample, the series were long enough so that the values of
the variables in the basic regression mode1--whether growth
rates or mean ratios--as well as the export instability
index, could be computed for the fifteen year period 1956-

70. Data on p0pulation came from various issues of the

United Nations Demographic Yearbook.

 

We were able to obtain data on index numbers of manu-
facturing production and of industrial wholesale prices for
a total of thirty-five of the sixty-five countries. (See
Table A-4 in the Appendix.) Both sets of data came from

various issues of the United Nations Statistical Yearbook.

 

The price indices of world exports of manufactures were

taken from various issues of the United Nations Yearbook of

 

International Trade Statistics.

 

72

For most of the thirty-five countries, U1 was computed
for the eleven-year time span, 1958-68, that falls exactly
in the middle of the fifteen year period (1956-70) for
which growth rates were computed for all thirty-five
countries. The base year, 1963, for all the index numbers
also falls exactly in the middle of this period. This
eleven-year span was chosen not just for its obvious
statistical convenience but also because it was the longest
period within the fifteen-year Span for which unbroken
series of the index numbers could be obtained for most of
the countries. Within the eleven year Span, the first
three years, 1958-60, were taken to be Period 1 while the
last three years, 1966-68, were Period 2. The use of the
average of a few years at the beginning and at the end of
the eleven year span as Periods l and 2 was considered
superior to using just the terminal years since the latter
involves the risk of biases that may result if individual
years were exceptiona1. (However, for a few countries for
which the time Span for which data were available was
shorter than eleven years, the terminal years were used as
Periods 1 and 2.) Although the use of the average of a few
years rather than a single year for each period reduces the
length of the time Span over which change is measured, we
do not consider this factor important for the length of time
involved here. It is to be noted that the period over which

change is measured should, in any case, not be too long or

73

else the possibility of substantial cyclical variation in
policy measures within the time span arises.

U2 was computed for the above thirty-five countries
from data on official and black market exchange rates given

in various issues of Pick's Currency Yearbook. The index

 

was computed for the ten year period, 1959-68.

As was noted in Chapter Three, Balassa and associates
have computed effective rates of protection for Six develop-
ing countries--Brazi1, Chile, Mexico, Malaya, Pakistan, and
the Philippines--and one developed country, Norway, used for
comparison in their analysis.2 Before proceeding with our
own analysis, we believe it might be of some interest to
compare the ranking of those seven countries in terms of
effective rates of protection with the rankings obtained
with U1 and U2. The figures are shown in Table 5-1. There
were no data available for the computation of U1 for Malaya,
so the rankings are for the six remaining countries. The
Spearman rank correlation coefficient between U1 and the
effective rate of protection is .74, not significant at the
5% level. On the other hand, that between U2 and the
effective rate of protection is .94, which is Significant
at the 1% level. Besides computing effective rates of pro-
tection, Balassa et al. also attempted to estimate the Egg
effect of protection in terms of the welfare costs to the
economy. Using the computed effective rates of protection,

together with assumed price elasticities of supply and

74

demand for exports and imports, and various assumptions
about the nature of long-run costs of production in
different industries, they computed the net "cost" of
protection (as a proportion of GNP) as a summation of (a)
the static cost of protection of import-substitutes: excess
costs plus above-normal profits and wages in industries

that would not survive under free trade; (b) the dynamic
cost of protection of import-substitutes: excess costs

plus above—normal profits in industries that would become
competitive under free trade; (c) the consumption effect:
consumer surplus on the increased consumption of imports;
(d) the terms of trade effect: reductions in export prices
in the event of free trade; and (e) the cost of increased
exports under free trade: the rise of the cost of exports
under free trade under the assumption that export industries
are subject to increasing costs.

The figures for the net cost of protection are given
in column (4) of Table 5-1. The rank correlation coeffi-
cient between U1 and the net cost of protection is .92,
significant at the 1% level; U2 and the net cost of pro-
tection give a rank correlation coefficient of .76, which
is non-Significant at the 5% level. These correlation
figures seem to be consistent with the statements made in
Chapter Three regarding the difference in the nature of the
types of distortion measured by U1 and U2. U2, as compared
to U1, it will be recalled,more appropriately falls into

the same category of measures as the level of effective

75

Table 5 — l

 

Measures of The Effect of Protection

 

 

 

(1) (2) (3) (4)
Effective Net Cost of
Rate of Protection
Protection (Z of GNP)
Country U1 U2 (Z)

1. Brazil 5.92 1.2054 127 9.5

2. Chile 2.65 1.4018 158 6.2

3. Mexico 0.73 0.9992 32 2.5

4. Malaya --- --- 11 -0.4

5. Pakistan 0.99 1.6439 188 6.2

6. Philippines 0.60 1.1403 53 3.7

7. Norway .44 1.0010 9 1.8

Source: Columns (3) and (4): Balassa et. al. The Structure

 

of Protection in Developing Countries, pp. 54, 82.

 

76

rate of protection, which is more or less a direct index of
the amount of ex ante incentives for import-substitution
provided by official policy; by contrast, U1, as defined,
is more properly a measure of the pgp cost-price effect
resulting from official policy. Based on this very limited
sample of countries, it would appear then that both U1 and
U2 are reliable indices of the type of distortion each is
supposed to measure.

Further discussions about the computation of the input
variables used in the following analysis are presented in

the Appendix.

The Regpession Results

 

The export instability index, as well as the values of
the variables in the basic regression model, could be
computed for all of the Sixty-five countries for which
national income accounts data were available. But U1 and
U2, as already noted, could be computed only for those
thirty-five countries for which adequate industrial statis-
tics were obtainable. The problems of export instability
and import-substitution, as much of the theoretical analysis
of this study has Shown, are closely interrelated; hence,
proper empirical procedure would require that hypotheses
relating to both be tested simultaneously with the same
sample and for the same time period. This can be done only
with the thirty-five-country sample for which the values of

all relevant variables could be computed. But in order to

77

make use of all available information, we Shall start by
testing for the Sign and Significance of the export
instability index alone in the Sixty—five-country sample
and then proceed to the thirty-five-country sample where
all variables are considered simultaneously.

The export instability index was computed using the
"Standard error of estimate" formula discussed in Chapter
Two. It was noted there that measures of export instability
are influenced by the type of trend fitted to the data.
Massell (1970) who used the same formula computed indices
based on an exponential rather than a linear trend.

Although he noted that the exponential trend gave a better
fit for most of the countries in his sample for the time
period used, he argued that there is in any case some
justification for considering the exponential trend the

more appropriate one on a priori grounds. This is because,
according to him, "countries tend to plan in terms of the
growth rate, not in terms of absolute increments."4 We

find this theoretical argument somewhat tenuous. In order
to determine which is the more appropriate type of trend for
our analysis, we also fitted an exponential as well as a
linear trend to the series on real exports. Although the
exponential trend gave a better fit for about two-thirds of
the sixty-five countries, the "goodness of fit" was
approximately the same for both types of trend in almost all
cases. By way of establishing a firmer basis for choice,

we proceeded to compute and compare instability indices

78

based on both types of trend. The instability index based
on the linear trend showed a relatively high correlation
with the growth rate of exports (which was computed by
fitting an exponential function). The correlation coeffi-
cients were +.49 and +.81 in the sixty-five- and thirty-
five-country samples respectively. The corresponding fig-
ures for the index based on the exponential trend were

+.05 and +.26 respectively.5

On purely empirical grounds,
therefore, the index based on the exponential trend was
chosen as the more appropriate for our samples.

The results of ordinary least squares (O.L.S.) re-
gressions performed in this study are shown in Table 5-2,
below. In parentheses beneath the corresponding regression
coefficients, t-Statistics are Shown. The regression

6 Double

residuals were inspected and found satisfactory.
asterisks indicate regression coefficients significant at
the 52 level or better in a one-tailed test of significance.
Single asterisks indicate those significant at the 10%
level.

Line (1) shows the result of regression with just the
basic model (i.e., equation (14), Chapter Four). The
coefficient of determination, corrected for degrees of
freedom (RE) is significant at better than the 12 level
using an F-test. The growth rate of the importing capacity
are

of exports, G and the growth rate of population, G

X’ n’

positive and significant at the 1% level. The domestic

79

investment ratio, I/Y, and the ratio of exports to output,
X/Y, are positive and significant at the 5% level while

the ratio of net foreign capital inflow to output, F/Y, is
positive but non-significant. Regression (2) Shows the
result of bringing in the export instability variable,
V(X), into the equation. There is scarcely any change from
regression (1). V(X), contrary to a priori expectations,
is positive, though highly non—significant, with a t-value
less than unity.

From now on, until we bring in U1 and U2, the analysis
will concentrate on the export instability variable. The
cross-section analysis just performed has considered each
country as a sample point, thereby resting implicitly on
the assumption that inter-country differences with respect
to vulnerability to export instability is purely quanti—
tative. Although the total sample consists of only LDC's,
these countries vary widely in terms of their levels of
economic development; hence it is plausible to expect
significant qualitative differences among them in terms of
vulnerability to export instability. The term, "economic

' encompasses a wide variety of economic as

development,'
well as non-economic attributes. The question then is what
particular attributes should we expect, a priori, to result
in qualitative differences in terms of vulnerability to
export instability? It will be recalled from Chapter One
that Massell (1970) had argued that the level of per capita

income--"a rough measure of the level of economic

80

SOphistication"--might reflect such qualitative differences
with respect to intercountry variations in the incidence of
export instability. His reasoning was (a) that a more
sophisticated country tends to produce and export commo-
dities that are Skill- and research-intensive, and that
the demand for such goods is relatively stable; and (b) that
a high per capita income is "usually associated with greater
flexibility," A more flexible country would be better
able to Shift resources among products in response to or in
anticipation of exogenous changes in market conditions.
But the author found per capita income to be insignificant
in his multiple regression analysis.

Massell's analysis was of intercountry variations in

the incidence of export instability; we are concerned here

 

with intercountry variations in vulnerability to export

 

instability. If the attributes measured by per capita
income appear not to be an important determinant of the
degree of instability in a country's exports, it does not
follow that they mayrun:be important in determining the
extent of the country's vulnerability to instability; hence,
thoroughness requires that we consider the effects of such
factors.

The theoretical grounds on which Massell had eXpected
the level of per capita income to be a Significant factor
in his analysis seem to be relevant to the present analysis
also; but they would now lead to theoretically conflicting

expectations. First, being relatively more oriented to the

81

production of skill- and research-intensive goods is
synonymous with being more industrialized. And being more
industrialized would ordinarily also imply being more
flexible, particularly in the sense of being more able,
through countercyclical action, to moderate the impact of
export fluctuations. But in terms of the growth model
which forms the basis of the present study, being more
industrialized has a more direct implication suggesting
greater vulnerability to a given degree of export insta-
bility. As was discussed in Chapter Three, it has been
suggested that most currently less developed countries
engaging in import-substitution are at a stage of indus-
trialization where progress with import-substitution brings
with it an acceleration in the demand for imports. If this
is the case, then the more industrialized a currently less
developed country is, the greater is the degree of its
dependence on imports likely to be. And the greater the
degree of an economy's dependence on imports, the greater
its vulnerability to a given degree of instability in the
importing capacity of exports.

It would appear then that while a more industrialized
LDC may possess the greater flexibility necessary to
moderate the impact of a given degree of export instability,
it may at the same time be more vulnerable by virtue of its
relatively greater dependence on imports. To the extent

that the level of per capita income is an indicator of the

82

characteristics under discussion, we utilize it in our
analysis to take account of their net effect.

The question now is the proper functional form in which
to bring per capita income into the equation. Since the
empirical distribution of per capita income more closely
approximates a log-normal rather than a normal distribution,
we shall, following Massell (1970), introduce per capita
income into the equation in a log-linear rather than the
linear form. Line (3) shows the result of adding log(y)
[where y = per capita GDP] to the equation. There is a
slight reduction in the corrected coefficient of deter-
mination; V(X) remains positive and non-significant; and
per capita income is itself non-Significant. Either the
net effect of the characteristics discussed earlier is
insignificant or the mere addition of per capita income as
an explanatory variable in the equation does not adequately
take account of those characteristics. Indeed, with regard
to the level of industrial development, the level of per
capita income may be a particularly unreliable indicator
since, as Chenery has noted, a country "having a continuing
comparative advantage in primary production may . . . reach
a high level of [per capita] income without an increase in

the share of industry in total output."7

But even if the level of per capita income were an
adequate measure of the level of industrial development,

merely adding it as an explanatory variable in the equation

83

would still not be sufficient to take account of structural
differences among countries. This is because such a pro-
cedure implicitly assumes the existence of gradual and
continuous relationships among variables across all sample
points; an assumption which is hardly warranted when
account is taken of discontinuities in the process of
development and structural change.

In his theoretical analysis of the effect of trade on
growth (which is the subject of our inquiry), Staffan
Burenstam Linder took as his starting point the existence
of two distinct groups of LDC's to which two different
models are applicable.8 While not arguing that LDC's can be
classified into "two neat and easily recognizable groups,"
he makes a distinction between "developing" and "backward"
LDC's which can be "regarded as representing two typical,
extreme forms on the scale." The former are the countries
where a determined drive toward industrial develOpment is
under way; and it is to these that the foreign exchange gap
model (which forms the analytical framework for this Study)
is applicable: "For only when an underdeveloped nation
reaches the stage where there is determination and realism
in its development effort does the requirement of imports to
operate and expand the production apparatus arise; and not
until then does the [constraint]--import requirements

"9

exceeding export earnings-~become important. It is in

such countries that trade has "a potentially important

84

leverage effect on capacity utilization." The more backward
LDC'S are, by and large, those without a Significant import-
competing sector, and which have been "able either to earn
more foreign exchange in relation to their national income
or to satisfy their needs for input imports at a lower
relative level of foreign exchange earnings."

It is clear from the immediately foregoing that it is
the "developing" subgroup of LDc'S which is implicit in the
hypotheses regarding the effects of export instability and
protection-induced cost-price distortions. That being the
case, proper empirical procedure requires that the other
countries be excluded from our sample; failure to do so
invites the risk of at least obscuring the relationship
under investigation.

The task now is to identify the countries to be ex-
cluded from the sample. A useful guide in this connection
is provided by a recent effort by the United Nations'
Economic and Social Council (Committee for Development
Planning) to identify the "least deve10ped among the

developing countries,' which are deserving of Special

assistance in overcoming the obstacles to social and

11

economic development. The rationale for this effort was

stated as follows:

The Committee is convinced of the need for
special measures in favour of the least devel-
Oped among the developing countries. While
developing countries as a group face more or
less the same general problems of underdevelop-
ment, the difference between the poorest and
the relatively more advanced among them is

 

85

quite substantial. While the basic framework

of the International Development Strategy

concerns all developing countries, the capacity

of these to benefit from general development

measures varies widely. The least developed

among them cannot always be expected to benefit

fully or automatically from such general

measures adopted in favour of all developing

countries. Some Special supplementary measures

are therefore called for to remove the handi-

caps which limit the ability of the least

developed countries to derive significant

advantages from the Second United Nations

Development Decade.12

Besides a low level of per capita income, the features

judged by the committee to characterize the least deve10ped
countries to a relatively high degree include the predom-
inance of agriculture and/or primary activities; low level
of education and insufficient adaptation of the education
system to development needs; weak over-all administrative
and governmental organization, a consequence of the lack of
trained manpower; rudimentary economic infrastructure, such
as communications, power generation facilities, water works,
harbors and roads; and low Standards of health and inade-
quate health facilities. The committee also mentioned lack
of basic economic statistics as being itself an indication
of low level of development. While recognizing that a two-
fold classification of LDC's into "least developed" and
"other" is somewhat arbitrary, given the multidimensional
complexity of economic and social develOpment, and that
different concepts of "least development" may be relevant

in different contexts, the committee nevertheless con-

sidered it desirable, as a matter of practical necessity, to

86

establish concrete criteria for the identification of

"least developed" countries. Ideally, a composite index
incorporating all the major strategic variables would be
desirable; however, the information presently available
makes the construction of such an index extremely difficult.

The committee accordingly adOpted a technique based on a

very few highly significant and widely available indicators."
The following three indicators were chosen: (1) per
capita GDP-~"a rough and ready indicator of the productive
capacity of an economy andcflfits ability to provide needed
services." It also serves as "a general indicator of the
dimensions of poverty and over-all development"; (2) the
Share of manufacturing in total output--an indicator of the
extent of structural transformation of the economy; and
(3) the literacy rate, which indicates the Size of the base
for enlarging trained and skilled human resources. The
committee considered that countries having all three of the
following characteristics "Should almost certainly be
classified as least developed": per capita GDP of $100
or less; share of manufacturing in total GDP of 10% or less;
and literacy rate--proportion of the literate persons in
the age group of fifteen years and more--of 202 or less.13
In order to capture border line cases, the committee, with
respect to the level of per capita income and the manu-
facturing ratio, ad0pted effective cut-off points of $120

and 122 respectively. These figures, like any others that

could be chosen, are quite aribtrary. However, they are so

87

clearly weighted toward the low extreme of the scale of
development that they provide a strong prima facie test of
low level of underdevelopment. On the basis of the above
criteria, the committee classified the following twenty-
five countries among the least developed of the LDC's:
Afghanistan, Dahomey, Ethiopia, Guinea, Haiti, Mali,
Malawi, Niger, Sudan, Uganda, Upper Volta, Tanzania; Bhutan,
Botswana, Burundi, Chad, Laos, Lesotho, Maldives, Nepal,
Rwanda, Sikkim, Somalia, Western Samoa, and Yeman.14

Of these, only the first twelve are included in our
total sample. Of the three criteria stated above, only per
capita income and the manufacturing ratio are directly
relevant to the concept of "least development" implicit in
our study. On the basis of Simultaneously having a per
capita GDP of $120 or less, and a manufacturing ratio of
122 or less, the following additional eight countries qual-
ify for exclusion from our sample: Burma, Kenya, Mada-
gascar;jMauritania,Nigeria, Pakistan, Thailand, and Togo.

Equations-(4) and (5) in Table 5-2 are the results of
regressions with the remaining forty-five countries. There
is a Slight increase in the corrected coefficient of
determination. V(X) acquires the expected negative Sign
but remains non-significant. Log(y) is also negative and
non-Significant.

We now proceed to analysis with the thirty—five-country

sample. Regressions (6) through (11) are the results of

88

adding U1 and U2, as well as log(y), in various combi-
nations to the basic model. Both U1 and U2 are everywhere
positive and non-Significant, with a t-ratio less than
unity. Equation (12) shows the result of adding just the
export instability variable, V(X), to the basic model.
There is a slight increase in the corrected coefficient of
determination and the negative coefficient of V(X) becomes
significant at the 102 level. When per capita income is
brought into the equation in regression (13), the coeffi-
cient of V(X) becomes Significant at the 5% level. Bringing
in U1 and U2 into the equation in regression (14) has the
effect of reducing the explanatory power of the model, and
both variables remain non-significant.

Of the above thirty—five countries, two--Pakistan and
Thailand--meet the modified United Nations test of "least
development" (i.e., excluding the literacy test) while a
third, EthiOpia, meets the full test. DrOpping these three
countries from the sample, regressions (14), (12), and (13)
were rerun with the remaining thirty-two countries to obtain
equations (15), (16), and (17) respectively. Both U1 and
U2 remain positive and non-significant. But the explanatory
power of the model as well as the negative t-ratio for the
coefficient of V(X) increase from their previous levels.

All of the foregoing statistical results indicate (a)
that there is a lack of association between export insta-

bility and economic growth among LDC's in general, although

p gin-1|!
fl 1 ML .2 _ .

,1 . 1

B , . ..

89

there is some evidence suggesting that a negative relation-
shiptexistsamong the more industrially advanced LDC's; and
(b) that the degree of cost-price distortions associated
with protection is not systematically related to the rate

of economic growth.

 

9()

Aaon.v Ammm.v Ammm.v «afimww.av «aaqqm.mv ««Amwm.uv Amom.v Aqmn.v

H.a on. was.1 ama.su was. osa.s NmN. mam. «mm.fl Omw.fl
A~H.av AmNa.V «sammm.av «sflmaa.mv Anam.mv Ammfl.v Asma.v
m.os an. om~.nu SSH. smm.H mu. wma. omw. Nam.

Amsn.v Aaoa.v Amas.v «aflflm.av asfiama.sv ssAHSo.mv «fifinm.av Aako.av
n.HH mm. SON. me.H mos. Hmm.H kmfl. am~.fl amo.a mm~.~u

Aaqm.v Anqq.v eeAmhw.Hv «aAnmo.qv «sAmmm.mv «rammo.mv Aenn.fiv
e.MH em. wmw. wHH. wa.H emH. mmN.H wo~.m coo.HI

«aammw.av «samuw.sV «sAsaa.mV seaflso.Nv Amms.flv

 

«.ma an. Hus. MNS.H omH. ma~.H on.w 0mm.au
cause-» «a a as AsvaH Axvs a x :o o H ucwumnou

wo I SHnSfiumv us summon mmcowmmoumom .m.A.o

Nlm manmw

0218
3161111173

91.

 

Anna.v Aaam.v Auss.v Ammo.v mmmw.v «SAHHs.sV «sawmm.wv rsasmm.av Amam.v

a.a No. Nam. Hso. asa.u Hoe. mam.a was. SSN.H ma.mH mao.~u
Amma.v Aoom.v Amao.v Amms.v ssAoaS.sV ssAHam.Nv «sAomw.Hv Amnm.av

m.m no. Has. «no. moo. mom.H «as. omN.H mam.~a NNA.N-
Asao.v qu.v aNao.v Awam.v «samom.sV «*Asam.~v ssfiflam.av Asaa.v

mo. Nmm. amfi.a mao.u amo.H awn. sNN.H ~ss.SH sao.~-

Amsm.v Asmo.v Amam.v ssxaaa.sv sixmao.m seamso.mv Amsm.flv

N.HH so. Nam. mso.u OHH.H mas. mH~.H aw0.m~ www.ms
Asma.v has.v ANAH.V Ammo.v «mamas.sv «saamm.mv ««Ammm.av Amso.v

~.m me. «no. oow.- mqo. mum.fi own. SAN.H smm.sa aao.H-
Ammm.v Amss.vs+nmho.qv asaafia.mv *«Ammw.sv A“ m.HV

0.0“ «a. Hmo. kmo. asm.H «as. HSH.H mmq.NH oak.Ha

IN... hm X a KIA..-
osuamlm «m «p as AavaH Axvs a x o o H newsmaou

 

Assumaucoov usm manna

mm

m
m

mm

mm

mm

2:

.0H

.m

€92

 

Amom.V*Amoa.Hv Afloa.v Amflm.v ssfisaq.nv *«ANSB.NV «sAmam.Hv Amwm.v
n.0H mo. mn~.a oms.oau «ma. mos.a mom. H00.H mam.H was.
«Ammo.av Amka.v Amnm.v «saaas.mv sshssa.m «sansw.av “was.v
H.~H mo. msa.mn has. mtm.a mow. Noo.H am..HH mom. u
Aomm.v ANNN.V Amao.v {Aamo.av Anon.v Aomm.v «saomo.mv «aAmmo.~v «Amoa.flv Asoo.v
«.5 no. mum. mso. Nm~.- Nao.OH: Nma. msa.H HAN. H0H.H osa.~H mac. .
Amms.v a~ma.v Asma.V«ASHo.HV Amws.v Amqo.v «*Amom.qv «sAoqo.mV «sAama.H Ammo.v
m.a as. mmm. omo. mfl.m- ohm.0Hu um. mom.H mom. sq~.fl mmw.sH ASN. a
Amma.HV««Aomm.Hv Ammo.v momc.v «aAHNo.nv «annfla.mV «xfiomu.~v Amom.v
o.oa so. mmm.- mow.0Hu Nun. mmH.H com. SAN.” cmq.mfl mam.
*Amaq.fl Awss.v Arms.v ssfimso.mv «sfimoo.mV 4«ANNS.MV Aroo.v
”.mfl mm. maa.m- «ma. mmH.H mom. mmH.H snm.mfi mxw. u
n S. x .. e.
owummum NM NS Hp Aavmoa Axv> m x o :0 H ucmumsou

Awmscfiuuoov mum manna

mm

mm

mm

mm

ozrg
oldmus

.na

.oH

.qH

(\I
I4

10.

11.

12.

13.

14.

FOOTNOTES--CHAPTER FIVE

The countries are shown in Table A-1 of the Appendix.

Balassa, et al., pp. cit.

Ibid., p. 82.

 

Massell (1970), Ice. cit., p. 619.

 

Indices based on the linear trend are shown under the
column V(X)' in Appendix Table A-3 while indices based
on the exponential trend are shown as V(X) in Table
A-2.

Over 952 of the standardized residuals fell within two
standard deviations from zero.

Chenery, "Patterns of Industrial Growth," loc. cit.,
p. 624.

Staffan B. Linder, Trade and Trade Policy for Develop:

 

ment, New York: Praeger Co., 1967.

Ibid., pp. 3-4.

 

Ibid., p. 152.

 

I am indebted to Professor M.E. Kreinin who suggested
the use of the U.N. criteria for identifying the
countries to be excluded from the sample.

Committee for Development Planning, Official Records
of the Economic and Social Council, Slst Session,
Supplement 7, Section II, p. 12.

 

 

Ibid., p. 16

 

93

CONCLUSIONS

In this study we set out to examine empirically the
claims that (a) export instability, and (b) protection-
induced cost-price distortions have been detrimental to
the growth of LDC's. The analysis was undertaken within
a theoretical framework suggested by the theory of trade
and growth implicit in the arguments supporting such
claims. The statistical results (a) confirmed the find-
ings of previous studies of a lack of association between
export instability and economic growth among LDC's in
general, although there was some evidence suggesting that
a negative relationship might exist among the more indus-
trially advanced LDC's; and (b) indicated that the degree
of cost-price distortions associated with protection is
not systematically related to the rate of economic growth.

The implications of the latter finding with respect
to the effect of cost-price distortions call for some
careful interpretation.

Under the classical comparative-cost doctrine, free
trade leads to optimum allocation of productive resources,
and it is assumed that this static efficiency in allocation
is a precondition for dynamic efficiency in growth. Neo-
classical modifications of this theory Show that even in the
presence of domestic distortions due to, say, externalities

94

95

and factor market imperfections, the adOption of appro-
priate (tax-cum-subsidy) policies will still permit the
simultaneous maintenance of internal and external equilib-
rium, and thus allow the economy to remain on the optimal
growth path.1 With regard to the type of rigidities argued
to give rise to the savings and trade gaps, Bruton has
maintained that these are not "structural" (in the sense of
being the result of institutional or exogenous factors) but
the result of bad policy; hence, again, the adoption of
appropriate (balance of payments and expenditure) policies
will remove the gaps and permit growth along the optimal
path.2
In opposition to Bruton, Chenery writes: "The effects
of an overvalued rate or other misguided policies are
incorporated in the existing distorted structure of produc-
tion and trade. The fact that it requires a reallocation
of investment and other changes extending beyond the Short
run to expand the trade limit makes the problem structural,

"3 To the extent that such rigidities

whatever its origin.
are a fact of life in the developing countries, argues
Linder, a reformulation, not a modification, of the con-
ventional trade theory and policy for such countries is
called for; and the foreign exchange gap model is the
logical outcome of assuming such rigidities.4 Under such a
model, trade/growth policy no longer presumes the ability

of the price mechanism to bring about a socially acceptable

rate of growth and income distribution; hence "forced"

96

import-substitution and/or export promotion may be
necessary to achieve an accelerated rate of growth.

With respect to inflation, the type of cost-price
distortion usually considered by prOponents of the orthodox
theory to be most symptomatic of bad policy, Seers has
noted: "It is scarcely possible to claim that policy weak-
nesses were the central cause of inflation. To support this
it would be necessary to explain the fundamentally different
eXperiences of different countries in terms of levels of
national competence, which would be hard to sustain in an
area where power has been in the hands of peOple who Show
certain obvious cultural Similarities. It would be necessary
to explain why inflation accelerated (or decelerated) in
several different countries in the same periods, and this
would require rather elaborate psychological cycles."5
More directly, he states: "It is meaningless to set up a
hypothesis that inflation helps or hinders growth. Growth
and inflation are interrelated but not in any simple way."6
An implication of all of the foregoing arguments is that
observed cost-price distortions such as measured in this
study can no longer be taken to necessarily reflect in any
systematic manner the extent of deviations from optimal
policy, but are rather the more or less random indications
of the extent of structural disequilibria--socia1, political
and economic-- in individual countries. The empirical

findings of this study support these "structuralist" argu-

ments.

97

It is important to note that the conclusion to which
this study lends support says nothing about the relative
merits of an import-substitution biased as against an

export promotion oriented development strategy. AS noted

 

in the text, no theoretically meaningful distinction, in a
dynamic general equilibrium framework, can be made between
forced export promotion and forced import-substitution in
terms of their long-run growth effects. As a partial
approximation, however, the optimum resource allocation
dictum--equalizing the marginal costs of earning and saving
foreign exchange—-provides a valid basis for comparison.
Studies of the type reported by Krueger and Bhagwati would
seem to be the most feasible way of making such comparisons.
If further studies along those lines should definitely Show
that the existing structure of world trade and the institu-
tional setting in developing countries tend to produce an
asymmetry in favor of export promotion in terms of the
actual returns to the two types of policies, then the
obvious implication would be that policy should err on the
Side of allowing a higher marginal cost for earning than

for saving foreign exchange.

FOOTNOTES--CONCLUSIONS

See G. Haberler, "Some Problems . . .," loc. cit;

E. Hagen "An Economic Justification of PFEFecTTEnism,"
Quarterly Journal of Economics, November, 1958.

 

H. Bruton, "The Two Gap Approach to Aid and Develop-
ment: Comment," loc. cit.

 

H. Chenery,"The Two Gap Approach to Aid and Develop-
ment: A Reply to Bruton," Ibid., p. 447.

S.B. Linder, op. £11.

Dudley Seers, "A Theory of Inflation . . ., loc. cit.,
p. 189.

 

Ibid., p. 191.

 

Krueger and Bhagwati, loc. cit.

 

98

STATISTICAL APPENDIX

SOURCES AND METHODS

Sources

The following were the sources of data used in the
study:

U.N. Statistical Office: Expenditure components of GDP in
constant (1960) U.S. dollars.

U.N. Demographic Yearbook, various issues: Population.

 

U.N. Yearbook of National Income Accounts, various issues:
Industrial origin of GDP at factor cost (used to compute
ratio of manufacturing to total output, RMFG).

U.N. Yearbook of International Trade Statistics, various
issues: Index numbers of world exports of manufactures.

U.N. Statistical Yearbook, various issues: Index numbers
of manufacturing value-added; index numbers of industrial
wholesale prices.

 

Pick's Currency Yearbook, 1968: Official and black market
exchange rates.

 

Input Variables

 

Net foreign capital inflow, defined for the purposes of
this study as the balance of payments on current account,
was measured as the difference between the export and import
of goods and services components of GDP in constant U.S.
dollars.

Indices of industrial wholesale prices given for
various countries varied in coverage from Single items such

as "Textiles" to broader groupings such as "Finished Goods."

99

100

For each country the best available index was chosen.

Growth rates, mean ratios, and the export instability
index were computed for each country for the period shown in
Table A-1. RMFG was computed as the average for the period
1958-68, except where data availability required the use of
a shorter period. Growth rates were computed by fitting an

expenential function of the type

V A (l + r)t

to the time series by least squares regressions; where r is
the estimated growth rate of the variable, V. The export
instability index was computed using the "standard error of
estimate" formula (Ig) discussed in Chapter Two. The values
of the input variables used in the study are shown in

Tables A-2, A-3, and A-4.

Definition of Terms

 

Y = real GDP

H
II

real gross domestic fixed capital formation

X = real exports

F = real net foreign capital inflow

GY = growth rate of real GDP

G = growth rate of total population

Gx = growth rate of real exports

V(X) = instability index based on an exponential trend
V(X)‘ = instability index based on a linear trend

RMFG 8 share of manufacturing in total output

U1 and U2 are defined in the text

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

Country

Afghanistan
Argentina
Bolivia
Brazil
Burma
Cambodia
Cameroon
Ceylon
Chile
Colombia
Costa Rica
Dahomey

Dominican
Republic

Ecuador
Egypt

El Salvador
Ethopia
Ghana

Guatemala

101

Table A-1

 

Total Sample

Period

1960-70
1956-70
1956-70
1956-70
1956-70
1960-70
1956-70
1956-70
1956—70
1956-70
1956-70

1960-70

1956—70
1956-70
1956-70
1956-70
1956—70
1956-70

1956-70

(m1.

Mean GDP

863
13838
462
25562
1643
752
636
1590
4664
5205
579

177

795

1048

5234

712

1002

1433

1254

Mean Pop.
U.S.$)(millions)

15.22

21.84

10.68

17.10

Mean GDP/
Capita

(ml. U.S.$)
57
630
111
326
68
122
124
148
555
301
406

74

233

183
253

45
193

293

 

102

Table A-1 (Continued)

Country Period Mean GDP Mean Pop. Mean GDP/
(ml. U.S.$)(millions) Capita
(ml. U.S.$)
20. Guyana 1956-70 186 0.62 296
21. Guinea 1960-70 268 3.51 77
22. Haiti 1956-70 317 4.25 75
23. Honduras 1956-70 450 2.06 216
24. India 1956-70 35536 465.61 76
25. Iran 1956-70 5835 23.64 240
26. Iraq 1956-70 2110 7.64 271
27. Israel 1956-70 3393 2.37 1389
28. Jamaica 1956-70 762 1.73 435
29. Jordan 1956-70 373 1.84 198
30. Kenya 1956-70 909 8.92 101
31. Lebanon 1956—70 962 2.30 413
32. Madagascar 1956-70 621 4.79 107
33. Malawi 1956-70 206 3.69 55
34. Malaysia 1956-70 2739 7.62 355
35. Mali 1960-70 317 4.53 70
36. Mauritania 1960-70 129 1.05 120
37. Maurituis 1956-70 175 0.72 244
38. Mexico 1956-70 15370 40.34 373
39. Morocco 1956-70 2032 12.76 158
40. Niccaragua 1956-70 481 1.58 299
41. Niger 1960-70 281 3.53 79
42. Nigeria 1956-70 3534 46.59 75

43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.

58.

59.
60.
61.
62.
63.
64.

65.

Table A-1 (Continued)

Country

Pakistan
Panama
Paraguay
Peru
Phillipines

Senegal

Sierra Leone

Singapore
South Korea
Sudan

Syria
Taiwan
Tanzania
Thailand
Togo

Trinidad &
Tobago

Tunisia
Uganda
Upper Volta
Uruguay
Venezuala
Zaire

Zambia

Period

1956-70
1956-70
1956-70
1956-70
1956-70
1960-70
1956-70
1960-70
1956—70
1956-70
1956-70
1956—70
1956—70
1956-70

1956-70

1956-70
1956—70
1956-70
1960-70
1956-70
1956-70
1956-70

1956-70

103

Mean GDP

9590
566
338

2531

7146
646
297

1189

4722

1241

1088

2473
753

3315

147

623
917
679
211
1273
9523
1380

723

Mean POp.
(m1. U.S.$)(millions)

109.13

1.18

1.94

11.09

30.68

11.68

11.19

29.19

Mean GDP/
Capita
(m1. U.S.$)
87
468
173
226
231
186
127
640
120
95
215
205
67
111

92

674
204
92
43
482
1145
92

200

104

Table A-2

 

Input Data

 

10.

ll.

12.

13.

14.

15.

l6.

17.

18.

19.

GY GX GN V(X)
Afghanistan 2.93 8.50 2.10 11.5
Argentina 3.77 4.52 1.63 7.0
Bolivia 4.84 6.10 2.77 11.0
Brazil 5.78 4.05 3.12 7.8
Burma 4.25 -7.33 2.06 24.9
Cambodia 3.12 ~3.54 2.34 19.6
Cameroon 5.83 7.15 2.27 11.8
Ceylon 4.54 1.69 2.45 3.5
Chile 4.49 5.31 2.48 3.5
Colombia 4.93 3.62 3.20 4.7
Costa Rica 7.00 10.00 3.56 11.6
Dahomey 4.53 16.78 2.45 23.8
Dominican
Republic 3.54 1.13 3.58 11.8
Ecuador 4.72 2.73 3.32 5.5
Egypt 6.12 3.12 2.53 14.5
El Salvador 5.62 7.14 3.48 8.1
EthOpia 2.42 7.34 1.87 12.4
Ghana 3.41 3.16 2.84 13.5
Guatemala 5.01 9.67 2.92 6.7

20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.

42.

Guyana
Guinea
Haiti
Honduras
India

Iran

Iraq
Israel
Jamaica
Jordan
Kenya
Lebanon
Madagascar
Malawi
Malaysia
Mali
Mauritania
Maurituis
Mexico
Morocco
Niccaragua
Niger

Nigeria

105

CY
4.59

Table A-2 (Continued)

Gx
5.99

-0134

11.71
23.26

3.76

3.14

V(X)

10.4

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

106

Table A-2 (Continued)

G G

Y X
Pakistan 5.56 5.02
Panama 7.25 8.80
Paraguay 4.08 2.32
Peru 5.35 5.99
Phillipines 4.53 6.53
Senegal 1.37 .13
Sierra Leone 5.45 5.33
Singapore 7.08 5.02
South Korea 7.91 27.98
Sudan 3.00 4.31
Syria 4.94 6.53
Taiwan(China) 9.21 19.11
Tanzania 4.24 5.69
Thailand 7.83 8.59
Togo 6.88 10.50
Trinidad &

Tobago 5 74 8.89
Tunisia 4.26 3.19
Uganda 4.62 5.17
Upper Volta 2.39 14.11
Uruguay .66 .242
Venezuala 5.31 6.16
Zaire 1.87 -2.47
Zambia 6.97 6.83

2.58

V(X)

10.4

12.4
12.0
13.6
16.0
15.4
16.4
15.3
12.3

13.2

m

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

Afghanistan
Argentina
Bolivia
Brazil
Burma
Cambodia
Cameroon
Ceylon
Chile
Colombia
Costa Rica
Dahomey

Dominican
Republic

Ecuador
Egypt

El Salvador
Ethopia
Ghana

Guatemala

107

Table A-3

 

Input Data

 

I/Y

.119
.189
.169
.180
.138
.159
.129
.152
.170
.195
.172

.156

.187
.144
.158
.132
.126
.179

.119

X/Y

.111
.009
.170
.059
.125
.120
.229
.284
.137
.140
.224

.174

.196
.172
.158
.222
.111
.261

.143

F/Y

.150
.005
.101
.002
.006
.045
.012
.006
.023
.016
.061

.074

.050
.004
.036
.022
.027
.004

.012

V(X)'

11.90

19.30

21.70

10.00

12.90

 

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

108

Table A-3 (Continued)

Guyana
Guinea
Haiti
Honduras
India

Iran

Iraq
Israel
Jamaica
Jordan
Kenya
Lebanon
Madagascar
Malawi
Malaysia
Mali
Mauritania
Maurituis
Mexico
Morocco
Niccaragua
Niger

Nigeria

I/Y

.229
.070
.060
.165
.165
.163
.179
.263
.218
.164
.166
.211
.124
.122
.132
.171
.316
.160
.197
.119
.180
.107

.127

X/Y

.503
.206
.144
.225
.054
.211
.392
.179
.372
.143
.331
.263
.163
.192
.527
.158
.400
.471
.100
.235
.255
.158

.182

F/Y
.015
.009
.033
.037
.021
-.062
-.157
.184
-.001
.293
.021
.099
.033
.117
-.151
.076
.088
-.025
.014
-.002
.042
.023

.014

 

43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.

58.

59.
60.
61.
62.
63.
64.

65.

Table

Pakistan
Panama
Paraguay
Peru
Phillipines
Senegal
Sierre Leone
Singapore
South Korea
Sudan

Syria
Taiwan(China)
Tanzania
Thailand
Togo

Trinidad &
Tobago

Tunisia
Uganda
Upper Volta
Uruguay
Venezuala
Zaire

Zambia

109

A—3 (Continued)

I/Y

.134
.191
.150
.224
.172
.119
.128
.116
.188
.142
.135
.232
.158
.202

.121

.236
.204
.136
.123
.143
.205
.188

.231

X/Y
.070
.341
.144
.219
.132
.234
.304
1.392
.074
.208
.226
.62
.276
.203

.254

.700
.073
.250
.102
.163
.343
.304

.604

F/Y
.024
.051
.024
.052
.042
.041
.038
-.128
.096
-.000
.061
.062
-.022
.030

.002

-.045
5.314
-.014

.087

.010
-.137
-.262

-.156

V(X)'

56.0
11.30
11.80

33.00

11.50

14.70

 

110

Table A-4

 

Input Data

 

10.
ll.
12.

13.

14.

15.

16.

17.

18.

19.

RMFG U1 U2

Afghanistan .11 -—— --—
Argentina .32 2.87 1.0678
Bolivia .14 76 1.0747
Brazil .26 5.92 1.2054
Burma .10 --- —--
Cambodia .10 --- ---
Cameroon .04 --- ---
Ceylon .08 44 2.1013
Chile .26 2.65 1.4018
Colombia .18 1.16 1.5589
Costa Rica .16 .33 1.2243
Dahomey .05 --- ---
Dominican

Republic .17 .66 1.3178
Ecuador .16 .79 1.1572
Egypt .22 1.11 1.8416
El Salvador .17 .71 1.1754
EthOpia .07 .61 1.1880
Ghana .06 1.43 1.8599
Guatemala .14 .11 1.0800

 

7“"

.1, _ p .
1r _
M

v ' ”I ~ _ p

20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.

42.

Guyana
Guinea
Haiti
Honduras
India

Iran

Iraq
Israel
Jamaica
Jordan
Keyna
Lebanon
Madagascar
Malawi
Malaysia
Mali
Mauritania
Mauritius
Mexico
Morocco
Niccaragua
Niger

Nigeria

111

Table A-4 (Continued)

RMFG

.10
.06
.12
.14
.14
.28
.09
.24
.15
.08
.11
.12
.05
.07
.09
.08
.03
.19
.28
.14
.12
.06

.08

.74
.76
.73
.36

.68

.73
.34

.28

1.0000
1.5245
1.0878
1.1080

1.1621

.9992
1.1608

1.2092

112

Table A-4 (Continued)

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

RMFG U1 U2

Pakistan .11 .99 1.6439
Panama .15 .92 1.0000
Paraguay .16 .34 1.1025
Peru .17 1.35 1.0555
Phillipines .18 .60 1.1403
Senegal .08 --- —-—
Sierra Leone .05 --- ---
Singapore --- --- -—-
South Korea .15 2.95 1.2608
Sudan --- —-- ---
Syria .12 .87 1.1059
Taiwan(China) .20 .94 1.1148
Tanzania .06 --- ---
Thailand .12 .79 1.0122
Togo .ll --- —-—
Trinidad &

Tobago .13 —_- ---
Tunisia .14 .61 1.8005
Uganda .08 --- ---
Upper Volta .06 —-- ---
Uruguay .23 .58 1.0531
Venezuala .12 .55 1.1657
Zaire .16 ___ ___
Zambia .08 1.24 1.1287

 

10.

11.

12.

BIBLIOGRAPHY

BOOKS

Balassa, B., et. al., The Structure of Protection
in Developing Countries. Baltimore: The Johns—
Hopkins Press, 1971.

 

 

Baer, N., and Kersternetzky, 1., Inflation and Growth
in Latin America. New Haven: Yale University
Press, 1964.

 

 

Coppock, Joseph D., International Economic Instability.
New York: McGraw-Hill Book Co. Inc., 1962.

Denison, E.F., Why Growth Rates Differ, Brookings
Institute, 1957.

 

International Bank for Reconstruction and Development,
Supplementary Financial Measures, TD/B/43, 1965.

 

Johnson, Harry G., International Trade and Economic
Growth. Cambridge: Harvard University Press, 1958.

 

, Economic Policies Towards Less Developed

 

Countries. Washington, D.C.: The Brookings

 

Institution, 1967.

Kindleberger, Charles P., Foreign Trade and the
National Economy. New Haven: Yale University
Press, 1962.

 

 

, Economic Development, 2nd ed. New York:

 

 

McGraw-Hill Book Co., 1965.

Kreinin, M.E., International Economics, Harcourt,
Brace, Jovanovich, Inc., 1971.

 

Linder, Staffan 8., Trade and Trade Policy for Develop-
ment, 3rd ed. London: McMillan Co., 1960.

 

McBean, Alasdair 1., Export Instability and Economic
Development. Cambridge: Harvard University Press,
1966.

 

 

113

 

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

114

Maizels, Alfred, Exports and Economic Growth of
Developing Countries. Cambridge: Cambridge
University Press, 1968.

 

Malinvaud, E., Statistical Methods of Econometrics.
Amsterdam: North-Holland Publishing Co., revised
ed., 1970.

Meier, Gerald M., International Trade and Development.
New York: Harper and Row, Publishers, 1963.

 

Michaely, Michael, Concentration in International
Trade. Amsterdam: North-Holland Publishing Co.,
1962.

 

Mills, F.C., Statistical Methods. New York: Holt and
Co., revised ed., 1938.

 

Nisbet, C.T., ed., Problems in Economic Development.
New York: Free Press, 1969.

Pick's Currency Yearbook, Pick Publishing Co.

 

United Nations, Dept. of Economic Affairs, Relative
Prices of Exports and Imports of Under-Developed
Countries. Lake Success: United Nations, 1949.

 

 

, Measures of International Economic Stability.

 

New York: United Nations, 1952 (E/2156, ST/ECA713)

, Instability in Export Markets of Under-
Developed Countries. New York: United Nations,
1952 (1952. II A. I)

 

 

 

, Dept. of Economic and Social Affairs.
International Compensation for Fluctuation in
Commodity Trade. New York: United Nations, 1961.

 

 

 

, Committee for Development Planning, Official

 

Records of the Social and Economic Council, Slst

Session, Supplement 7.

, Report by the Secretary General of the United

 

Nations Conference on Trade and Development.
Towards a New Trade Policy for Development. New
York, 1964.

, Report of the United Nations Conference on
Trade and Development on its Second Session.
TD/L/37, 1968.

 

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

115

Vanek, Jaroslav, Estimating Foreign Exchange Resource
Needs for Economic Development: Theory, Method
and a Case Study of Columbia. New York: McGraw—
Hill Co., 1967.

 

 

ARTICLES

Bhagwati, J. and Krueger, A., "Exchange Control
Liberalization and Economic Development," American
Economic Review, Papers and Proceedings, May 1973.

 

 

Brainard, William C. and Cooper, Richard N., "Uncer-
tainty and Diversification in International Trade,"
Studies in Agricultural Economics, Trade and
DevelOpment, VIII, No. 3. (1968), pp. 256-85.

 

Bruno, N., "Estimation of Factor Contribution to
Growth Under Structural Disequilibrium," Inter-
national Economic Review, February 1968.

 

Bruton, H., "Productivity Growth in Latin America,"
American Economic Review, December 1967.

 

, "The Two Gap Approach to Aid and Development:

Comment," American Economic Review, June 1969.

 

Chenery, H.B., and Taylor, Lance, "Development Patterns:
Among Countries and Over Time," The Review of
Economics and Statistics, Vol. L, No. 4. (November
1968). '

 

 

, "Patterns of Industrial Growth," American

 

 

Economic Review, September 1960.

 

, and Bruno, M., "Development Alternatives in

 

an Open Economy: The Case of Israel," The
Economic Journal.

 

, and Strout, A., "Foreign Assistance and

 

Economic Development," American Economic Review,
Vol. 56. (September 1966).

 

, and Ecstein, Peter, "Development Alternatives
for Latin America," Journal of Political EcOnomy,
July/August, Supplement, 1970.

 

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

116

, and McEwan, "Optimal Patterns of Growth and

Aid: The Case of Pakistan," Pakistan Development

Review, Summer 1966.

, "The Two Gap Approach to Aid and Development:

 

A Reply to Bruton," American Economic Review,
June 1969.

 

Cochrane, 8., "Structural Inflation and the Two Gap
Model of Economic Development," Oxford Economic
Papers, November 1972.

 

Cooper, R.N., "Devaluation and Aggregate Demand in
Aid-Receiving Countries," in J. Bhagwati, et.al.,
eds., TradepiBalance of Payments and Growth,
Amsterdam: North-Holland Co., 1971.

 

Dorrance, 6.3., "The Terms of Trade," Review of

Economic Studies. 1948/49, pp. 50-56.

 

 

Erb, G.F., and Schiavo-Campo, 3., "Export Instability,
Level of Deve10pment and Economic Size of Less
Developed Countries," Oxford University Institute
of Economics and Statistics Bulletin. November
1969.

 

Fleming, Marcus J., and Lovasy, Gertrude, "Fund
Policices and Procedures in Relation to the
Compensatory Financing of Commodity Fluctuations,"
IMF Staff Papers, VIII, No. 1. (November 1960),
pp. 1-76.

 

Guisinger, "Negative Value Added and the Theory of
Effective Protection," Quarterly Journal of
Economics, August 1969.

 

Haberler, G., "Some Problems in the Pare Theory of
International Trade," reprinted in Caves and
Johnson, eds., Readings in International Economics,
Irwing, Inc., 1968.

Hagen, E.E., "An Economic Justification of Protection-
ism," Quarterly Journal of Economics, November
1968.

 

Hirschman, A., "The Political Economy of Import-
Substituting Industrialization in Latin America,"
Quarterly Journal of Economics, February 1968.

 

 

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

117

, and Hawrylyshyn, "Analysis of World Income

and Growth, 1955-65," Economic Development and

Cultural Change, October 1969, Part 2.

 

Johnson, H.G., "The Welfare Costs of Exchange
Stabilization," The Journal of Political Economy,
October 1966.

Kemp, M., "The Mill-Bastable Import Industry Dogma,"
Journal of Political Economy, February 1960.

 

Kenen, Peter, and Voivoidas, C., "Export Instability
and Economic Growth," Kyklos, Fasc. 4, 1972.

Krueger, A., "Some Economic Costs of Exchange Control:
The Turkish Case," The Journal of Political

Economy, October 1966.

 

Kuznets, 8., "Economic Growth of Small Nations,"
E. A. G. Robinson, ed., Economic Consequences of
the Size of Nations, New York, MacMillan, 1960.

 

 

, "Quantitative Aspects of the Economic Growth
of Nations: IX, Level and Structure of Foreign
Trade: Comparisons of Recent Years," Economic
Development and Cultural Chapge, October 1964,
Part II.

 

 

, "Quantitative ASpects of the Economic Growth

of Nations: II. Industrial Distribution of

National Product and Labor Force," Economic

Development and Cultural Change, No. 5. (July
1957) Supplement.

 

 

Lee, J.K., "Exports and the Propensity to Save in
LDC's," Economic Journal, June 1971.

 

Lewis, S., "The Effects of Trade Policy on Domestic
Relative Prices," American Economic Review,

March 1968.

 

Lewis, W.A., "Economic Development with Unlimited
Supplies of Labor," The Manchester School, 22,
(May 1954).

 

Massell, E.F., "Export Instability and Economic
Structure," American Economic Review, Vol. LX.
(September 1970).

 

, "Export Concentration and Fluctuations in
Export Earnings: A Cross-Section Analysis,"
American Economic Review, LIV, No. 2. (March 1964),
pp. 47-63.

 

 

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

118

McKinnon, R., "Foreign Exchange Constraints in Economic
Development and Efficient Aid Allocation,"
The Economic Journal, June 1964.

 

Nelson, R., "The Effective Exchange Rate; Employment;
and Growth in a Foreign Exchange Constrained
Economy," Journal of Political Economy, May/

June 1970.

 

Nurkse, Ragnar, "Trade Fluctuations and Buffer Policies
of Low-Income Countries," Kyklos, XI, Fasc. 2.
(1958), pp. 141-54. F

Prebisch, R., "The Economic Development of Latin
America and its Principal Problems," reprinted
in Economic Bulletin for Latin America, U.N.
Economic Commission for Latin America, February
1962.

 

 

‘mvan Mt: Il‘n!

Ranis, G., and Fei, J.C., "A Theory of Economic
Development," American Economic Review, September
1961.

 

Reimer, R., "Effective rates of Protection in East
Africa," Eastern Africa Economic Review, December
1971.

 

Robinson, 8., "Sources of Growth in Less Developed
Countries," Quarterly Journal of Economics,
August 1971.

 

Seers, Dudley, "A Theory of Inflation and Growth in
Under-Developed Countries," Oxford Economic Papers,
XIV, No. 2. (June 1962), pp. 173-95.

 

Singer, Hans, "The Distribution of Gains Between
Investing and Borrowing Countries," American
Economic Review, Vol. 40. (May 1950).

 

 

Soligo, R.,and Stern, J., "Tariff Protection, Import-
Substitution, and Investment Efficiency,"
Pakistan Development Review, Summer 1965.

 

Solow, R.M., "Technical Change and the Aggregate
Production Function,"The Review of Economics and
Statistics, August 1957.

 

 

, "Technical Progress, Capital Formation, and

 

Economic Growth," American Economic Review,
September 1962.