7/30/2019 1847208223
1/265
7/30/2019 1847208223
2/265
Trade Liberalisationand
The Poverty of Nations
7/30/2019 1847208223
3/265
Other Books by A.P. Thirlwall
Growth and Development: with Special Reference to DevelopingEconomies (1972; Eighth Edition, 2006)
Inflation, Saving and Growth in Developing Countries (1974)Regional Growth and Unemployment in the UK (with R. Dixon,
1975)Financing Economic Development (1976)Keynes and International Monetary Relations (edited, 1976)Keynes and Laissez Faire (edited, 1978)Keynes and the Bloomsbury Group (edited with D. Crabtree, 1980)
Balance of Payments Theory and the UK Experience (1980; FourthEdition with H. Gibson, 1992)
Keynes as a Policy Adviser (edited, 1982)Keynes and Economic Development (edited, 1987)Nicholas Kaldor (1987)Deindustrialisation (1989 with S. Bazen: Third Edition, 1997)Further Essays on Economic Theory and Policy, Volume 9 by
Nicholas Kaldor (edited with F. Targetti, 1989)The Essential Kaldor (edited with F. Targetti, 1989)
European Factor Mobility: Trends and Consequences (edited withI. Gordon, 1989)
The Performance and Prospects of the Pacific Island Economies inthe World Economy (1991)
Keynes and the Role of the State (edited with D. Crabtree, 1993)Economic Growth and the Balance of Payments Constraint (with J.
McCombie, 1994)The Economics of Growth and Development: Selected Essays of
A.P. Thirlwall, Volume 1 (1995)
Causes of Growth and Stagnation in the World Economy (theMattioli Lectures by N. Kaldor) (edited with F. Targetti, 1995)
Macroeeconomic Issues from a Keynesian Perspective: SelectedEssays of A.P. Thirlwall, Volume 2 (1997)
Economic Dynamics, Trade and Growth: Essays on HarrodianThemes (edited with G. Rampa and L. Stella, 1998)
The Euro and Regional Divergence in Europe (2000)The Nature of Economic Growth: An Alternative Framework for
Understanding the Performance of Nations (2002)Trade, the Balance of Payments and Exchange Rate Policy in
Developing Countries (2003)Essays on Balance of Payments Constrained Growth: Theory and
Evidence (with J. McCombie, 2004)
7/30/2019 1847208223
4/265
Trade Liberalisationand
The Poverty of Nations
A.P. ThirlwallProfessor of Applied Economics, University of Kent,Canterbury, UK
and
Penlope Pacheco-Lpez
Economic Consultant
Edward ElgarCheltenham, UK Northampton, MA, USA
7/30/2019 1847208223
5/265
A.P. Thirlwall and Penlope Pacheco-Lpez 2008
All rights reserved. No part of this publication may bereproduced, stored in a retrieval system or transmitted in anyform or by any means, electronic, mechanical or photocopying,recording, or otherwise without the prior permission of thepublisher.
Published byEdward Elgar Publishing LimitedThe Lypiatts15 Lansdown RoadCheltenhamGlos GL50 2JAUK
Edward Elgar Publishing, Inc.William Pratt House9 Dewey CourtNorthamptonMassachusetts 01060USA
A catalogue record for this bookis available from the British Library
Library of Congress Control Number: 2008932908
ISBN 978 1 84720 822 4 (cased)
Printed and bound in Great Britain by MPG Books Ltd, Bodmin,Cornwall
7/30/2019 1847208223
6/265
Contents
Acknowledgements viPreface vii
1 The theory and measurement of tradeliberalisation 1
2 Trade liberalisation, tradeperformance and economic growth 62
3 Trade liberalisation and international
inequality 1074 Trade liberalisation, poverty anddomestic inequality 145
5 Trade strategy for development 174
Bibliography 211Index 239
v
7/30/2019 1847208223
7/265
Acknowledgements
The authors are very grateful to Dr Ha-JoonChang, Professor Robert Wade and Professor Dani
Rodrik for allowing them to use sections of booksand articles written by them in abridged form:
Ha-Joon Chang (2002), Kicking Away the Ladder:Development Strategy in Historical Perspective(London: Anthem Press).
R. Wade (2004), Is Globalisation Reducing Povertyand Inequality?, World Development, April.D. Rodrik (2001), The Global Governance of Trade: As
if Development Really Mattered (New York:UNDP).
We would also like to thank all the team at
Edward Elgar Matthew Pitman, Elizabeth Clackand Nep Elverd for all their help in seeing the
book through the press.
vi
7/30/2019 1847208223
8/265
Preface
The purpose of this book is to argue that trade lib-eralisation, with the ultimate objective of free
trade, is not in the interest of poor developingcountries at their present stage of economic andsocial development. History, economic theoryand contemporary statistical evidence does notsupport the view that trade liberalisation, ormaking economies more open, necessarily raises
the growth of living standards, reduces poverty orleads to a more equal, fairer, distribution of incomebetween countries in the world or between peoplewithin countries. The orthodox theory of trade lib-eralisation is therefore being sold by internationalinstitutions to developing countries on a falseprospectus.
The book is the last of a trilogy of short, readablebooks which attempt to provide an alternativeapproach to that of orthodox neoclassical tradeand growth theory for an understanding of thegrowth and development process, with particularreference to poor developing countries. The first
book, The Nature of Economic Growth: An AlternativeFramework for Understanding the Performance ofNations (Thirlwall, 2002) presents a structural,demand-constrained model of growth, in contrastto the one-sector supply-oriented growth model of
vii
7/30/2019 1847208223
9/265
orthodox neoclassical theory, in which the balance
of payments or foreign exchange is identified asthe major constraint on growth, and the structureof production is crucial because different goodshave different production and demand character-istics. The second book, Trade, the Balance ofPayments and Exchange Rate Policy in DevelopingCountries (Thirlwall, 2003) focuses on the neglect
of the balance of payments (or monetary) conse-quences of trade in orthodox trade and growththeory and casts doubt on whether exchange ratedevaluation, or depreciation of the currency, is anefficient balance of payments adjustment weapon.
This new book continues to question the ortho-
doxy proposed by mainstream academic econo-mists, and policymakers in international institutionssuch as the International Monetary Fund (IMF),World Bank and World Trade Organization (WTO),that trade liberalisation is the panacea, or magicbullet, that will launch poor developing countries onto a higher growth path consistent with balance of
payments equilibrium. We argue that what wewitness in the world economy today, in an era ofunprecedented trade liberalisation, resembles muchmore closely the predictions of non-orthodoxmodels of growth and development that highlightcumulative processes and polarisation between
people and countries, rather than the equalising,equilibrating predictions of orthodox neoclassicaltrade and growth theory.
The prevailing consensus in the years since theSecond World War has been that if countries are to
viii Trade liberalisation and the poverty of nations
7/30/2019 1847208223
10/265
develop and prosper they must liberalise their
trade, almost at any cost. Free trade has become amantra. The orthodoxy has a long and distin-guished ancestory that goes back at least to AdamSmiths famous book,An Inquiry into the Nature andCauses of the Wealth of Nations (1776) and DavidRicardos On The Principles of Political Economy andTaxation (1817), but it was given renewed impetus
after the Second World War following the greatdepression and the beggar-thy-neighbour policies(protectionism) experienced during the 1930s. Atthe Bretton Woods Conference convened in 1944,which established the IMF and the World Bank, theworlds major economies committed themselves to
a liberal, free trade, order, endorsed in the 1980sby the so-called Washington Consensus, and re-inforced by the establishment of the WTO in 1995.
There is no doubt that there are real resourcegains from trade, as Smith and Ricardo originallyshowed, but there is nothing in the doctrine of freetrade, or in the practice of liberalising trade, to
guarantee an equal distribution of the gains fromtrade among participating countries or betweenindividuals within countries. Indeed, in certain cir-cumstances, which are not so unusual, such asterms of trade deterioration and balance of pay-ments problems, countries may suffer absolute
real income losses from specialisation according tothe law of comparative advantage. Nor is thereanything in the doctrine of free trade that guaran-tees a permanent increase in the rate of growth ofoutput and living standards. On the contrary, it is
Preface ix
7/30/2019 1847208223
11/265
an interesting and pertinent fact that despite exten-
sive trade liberalisation, the economies of manycountries, particularly those in Africa and LatinAmerica, have performed very poorly in the last20 years or so, and world income inequality hasworsened ignoring the fast growth of China. TheGini ratio, as a measure of income distributionacross countries of the world (international
inequality) and across peoples of the world (globalinequality), is as high today (if not higher) as it wasin the 1960s. Only a handful of poor developingcountries (mainly in Asia) have narrowed the rela-tive income gap between themselves and the richdeveloped countries. The vast majority of develop-
ing countries continue to lag behind, and manyAfrican countries are absolutely poorer today thanthey were 30 years ago. Does this mean that thesecountries do not trade enough, that their tradeneeds to be more liberal, or is there somethingwrong with the doctrine of free trade? It is alsointeresting and significant that while international
organisations and rich developed countries preachfree trade to developing countries, the rich coun-tries themselves do not practice what they preach,and never did so historically when they wereattempting to industrialise and develop (Chang,2002). The basic problem with the free trade ortho-
doxy is that it brushes to one side the different pro-duction and demand characteristics of differenttypes of goods, and assumes continuous fullemployment so that in the reorganisation ofresources that specialisation and trade bring about
x Trade liberalisation and the poverty of nations
7/30/2019 1847208223
12/265
there is no permanent unemployment; everyone
finds a job. In practice, however, some goods (forexample, land-based activities) are producedunder conditions of diminishing returns, whileothers are produced under conditions of increasingreturns. This makes a profound difference to thegrowth of living standards in countries specialisingin these different types of activity. Also, some
goods (for example, primary commodities) have alow income elasticity of demand in world trade,while others have a high income elasticity ofdemand. This makes a significant difference to therate at which exports and imports grow in differentcountries specialising in different types of activity.
There is a crucial link between the structure of pro-duction and trade and the economic performancethat can be expected from trade liberalisation,which the orthodoxy ignores. In many circum-stances, trade liberalisation may not be optimal fordeveloping countries, at least without the imple-mentation of complementary measures such as
export promotion policies, strategic investment instructural change, infant industry protection andsupportive exchange rate policies. The speed andsequencing of liberalisation measures are alsoimportant to bear in mind.
In the book we attempt to bring together in a
readable and coherent way the vast literature thatnow exists on the theory, practice and impact oftrade liberalisation for the benefit of students,policymakers, and all those interested in what ishappening to developing economies as the world
Preface xi
7/30/2019 1847208223
13/265
economy becomes more liberalised and globalised
(to use the fashionable terminology). The book hasthree essential messages:
1. Trade liberalisation is neither a necessary norsufficient condition for the more rapid devel-opment of poor countries, when developmentis defined in its broad sense (Goulet, 1971,
1995) of not only providing life sustenance forits people, but also self-esteem and freedom. Iftrade liberalisation does not reduce poverty,that is not development. If trade liberalisationleaves countries more dependent on others
because trade is unequal, that is not develop-
ment. If pressure to liberalise trade reducescountries freedom to pursue their own eco-nomic policies, that is not development.
2. Neo-liberal economics, which puts faith in tradeliberalisation as the route to development, has
become like a religion or creed, which not onlyignores perfectly respectable economic theory
which shows that there are several legitimateeconomic arguments for protection, but also dis-regards history and the facts of experience ofhow todays developed countries progressed inthe 19th century and the first half of the 20thcentury, and the miserable contemporary expe-
rience of many of todays developing countriesthat have been forced by multilateral institu-tions to liberalise prematurely or too quickly.
3. The extensive empirical research on the impactof trade liberalisation on exports, imports,
xii Trade liberalisation and the poverty of nations
7/30/2019 1847208223
14/265
the balance of payments and growth does not
support the view that liberalisation, at leastsince the 1980s, has been associated with asuperior economic performance of poor coun-tries, or a narrowing of the degree of inter-national income inequality between countriesand peoples of the world. The view to the con-trary, espoused by neo-liberal economists and
international institutions, is largely dependenton the inclusion of China and India in thesample of poor countries taken by investiga-tors. These are two poor populous countries,which have grown very fast in the last 20 years,and where trade as a proportion of GDP has
grown very rapidly, and they naturally make ahuge difference to the statistical results relatingto trade and growth and trade and incomeinequality. But China and India are just two ofover 150 poor, developing countries. If they areexcluded from the sample of poor, globalisingeconomies, a different picture of the world is
painted one in which international and globalinequality is rising, where the number ofpeople living on less than $2 a day is rising, andglobalising countries are performing no betteron average than those which are not liberalis-ing their economies so fast. Individual case
studies show a variety of experience.
We argue, therefore, that the rhetoric of trade liber-alisation, as a necessary condition for developmentand poverty reduction, is not matched by the reality.
Preface xiii
7/30/2019 1847208223
15/265
The choice is not between autarchy and free trade,
but the sensible management of trade: to achieve abalance between the growth of exports and importsso as to avoid balance of payments deficits; to allowpolicy space for poor countries to identify andencourage new areas of comparative advantage; toenable the State to intervene to promote self-dis-covery (Hausmann and Rodrik, 2003); and to allow
a judicious mix of tariffs and subsidies for infantindustry protection. This is the way rich countriesdeveloped historically, and it is the way that poordeveloping countries today should be allowed todevelop. In an environment oflaissez-faire, laissez-passer, structural change is impossible.
In Chapter 1 we briefly rehearse the theory oftrade liberalisation, and the benefits that it is sup-posed to confer on countries, and then outline thevarious meanings and measures of trade liberal-isation that investigators have used in theirresearch. Then we challenge the orthodoxy in aformal way, and consider history.
Chapter 2 focuses on the trade performance ofcountries that have liberalised, but not just theeffect of liberalisation on exports which is the con-ventional approach. The effect of liberalisation onimports and the balance of payments is also con-sidered, which can act as a constraint on economic
performance. We survey the vast amount ofresearch that has been conducted on trade liberal-isation and economic growth, both within coun-tries using time-series data, and across countriesusing cross-section and panel data.
xiv Trade liberalisation and the poverty of nations
7/30/2019 1847208223
16/265
In Chapter 3 we discuss unorthodox models that
predict growing inequality with trade and we giveevidence of the degree of income inequality in theworld economy between rich and poor countries,using measures of international inequality (whichtake the average per capita income of eachcountry) and global inequality (which takesaccount of the distribution of income within coun-
tries as well).Chapter 4 examines research on the link
between trade liberalisation in developing coun-tries and poverty, wage inequality, and the per-sonal distribution of income using householdsurvey data. Orthodoxy predicts a narrowing of
the domestic income distribution within poorcountries, but the reality is different.Chapter 5 draws some policy conclusions on
how trade policy should proceed in the interests ofthe poorer, developing nations of the world, andof poor people within poor countries. There arenational issues to consider, and also international
issues relating to how international economic rela-tions are organised and conducted relating totrade in goods and services, capital flows andfinancial arrangements between rich and poorcountries, in the interests of a fairer distribution ofincome in the world economy between countries
and people.
Preface xv
7/30/2019 1847208223
17/265
7/30/2019 1847208223
18/265
1. The theory andmeasurement of tradeliberalisation
Politicians and economists who promise that trade liber-alisation will make everyone better off are being disin-genuous. Economic theory (and historical experience)suggests the contrary. (Stiglitz, 2006)
Introduction
To consider the relationship between trade liberal-isation and economic development, it is impor-tant to state at the outset what we mean by theprocess of economic development. The traditionalmeasure of economic development, and what thetheory of trade liberalisation focuses on, is the
standard of living measured by the average levelof per capita income of the citizens of a country.The static gains from trade liberalisation are sup-posed to raise the level of income in the process ofresource reallocation according to the law of com-parative advantage; and the dynamic gains fromtrade, associated with more competition, the flowof new knowledge and a faster rate of capital accu-mulation, are supposed to raise the growth ofincome. The focus on national income, or on thegrowth of income, as a measure of development
1
7/30/2019 1847208223
19/265
is not enough, however, for two main reasons.
Firstly, what happens to the distribution of incomewithin countries matters for the measurement ofeconomic welfare. Trade liberalisation does notguarantee that all sectors of society will benefitequally. On the contrary, in the very process of lib-eralisation itself some groups will gain, but others,previously protected, will undoubtedly lose either
relatively or absolutely through unemploymentor a reduction in their real income. While totalincome and average per capita income may rise, adistribution-weighted measure of per capitaincome which gives a higher weight to the lossesof already poor people than to the gains of already
rich people may show a loss of welfare. Secondly,what happens to the international distribution ofincome is also important. Trade liberalisation doesnot guarantee an equal distribution of the gainsfrom trade between countries. Some countries will
benefit more than others, depending on their com-petitive strength in world markets, and move-
ments in the terms of trade. If rich countries gainmore than poor countries, the international distri-
bution of income will worsen. The gains of somerich countries may even be at the expense ofother (poorer) countries that lose absolutely. Asthe German Chancellor, Otto Von Bismark, once
remarked free trade is a policy for the strong!The eminent French sociologist, Denis Goulet,defines economic and social development in termsof three core components, which he calls life sus-tenance, self-esteem and freedom (Goulet, 1971,
2 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
20/265
1995). Life sustenance refers to the provision of
basic human needs such as food, shelter and cloth-ing. A country that cannot provide basic needs forall its people cannot be considered as developed.A country in which poverty is increasing is notdeveloping. As Stiglitz argues in his powerful
book, Making Globalization Work (2006), if eco-nomic growth is not shared throughout society,
then development has failed development isabout transforming the lives of people, not justtransforming countries. Trade liberalisation thatcreates poverty is anti-developmental. Self-esteemrefers to feelings of independence and self-worth.A country is not fully developed if it lacks self-
esteem because it is unable to conduct economicrelations on equal terms. Trade liberalisation thatmakes a poor country more dependent on richercountries for international aid, or on internationalinstitutions, such as the IMF, for loans, is not devel-oped. Likewise, an economy that comes to bedominated by multinational corporations in key
sectors of the economy, which destroy local indus-tries, or which has its banking sector taken over byforeign banks, cannot be said to be developing.Freedom is about countries and people being ableto decide their own destiny. No country or personsare free if they cannot chose, because they are
dependent on others for aid and assistance andeconomic policymaking is outside their democra-tic control (see also Sen, 1999). Trade liberalisationthat locks countries into unfavourable economicstructures based on static comparative advantage,
The theory and measurement of trade liberalisation 3
7/30/2019 1847208223
21/265
which perpetuates slow growth and poverty, is not
a development strategy.The key issue from a development perspective,therefore, is not whether trade liberalisationimproves the growth performance of a country,although that is important, but whether it reducespoverty; whether it raises the income of poorpeople relative to the rich, and whether it enhances
the ability of poor countries to catch up with therich. These are the issues to constantly bear inmind in discussing the practice and merits of tradeliberalisation. But first let us discuss the theoryand measurement of liberalisation, and some ofthe theoretical challenges to the orthodoxy.
The Classical Foundations of Trade
Liberalisation
The doctrine that trade enhances welfare andgrowth has a long and distinguished ancestrydating back at least to the great classical econo-
mists of Adam Smith (172390), David Ricardo(17721823) and John Stuart Mill (180673). In hisfamous book,An Inquiry into the Nature and Causesof the Wealth of Nations (1776), Smith stressed therole of trade as a vent, or outlet, for surplus domes-tic production, and as a means of widening the
market, thereby allowing greater specialisation ordivision of labour. Specialisation, in turn, stimu-lates capital accumulation and promotes learning
by doing. Remember Smiths famous dictum thatspecialisation is limited by the extent of the
4 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
22/265
market, but that is only one side of the coin. The
extent of the market is determined, in turn, by thedivision of labour because it is the latter that raisesproductivity and purchasing power. As well asstatic economies of scale from specialisation, there-fore, there is also the prospect that more trade willenhance growth by inducing so-called dynamiceconomies of scale associated with capital accu-
mulation, technical progress embodied in capital,and the spread of knowledge. To quote Smithdirectly:
Between places foreign trade is carried on, they all of themderive two distinct benefits from it. It carries out thatsurplus part of the produce of their land and labour forwhich there is no demand among them, and brings back inreturn for it something else for which there is a demand. Itgives a value to their superfluities, by exchanging them forsomething else, which may satisfy a part of their wantsand increase their enjoyments. By means of it, the narrow-ness of the home market does not hinder the division oflabour in any particular branch of art or manufacture frombeing carried to the highest perfection. By opening a moreextensive market for whatever part of the produce of theirlabour may exceed the home consumption, it encouragesthem to improve its productive powers and to augment itsannual produce to the utmost, and thereby to increase thereal revenue and wealth of the society. (p. 338)
In 1817, Ricardo published his On The Principlesof Political Economy and Taxation in which the basis
of trade is not vent for surplus commoditiesexpanding the market and permitting specialisa-tion, as in Smith, but differences in the productiv-ity of labour (or relative costs) between countriesin the production of different goods. This leads to
The theory and measurement of trade liberalisation 5
7/30/2019 1847208223
23/265
the idea of comparative advantage. A country
may have an absolute productivity (cost) advan-tage in the production ofevery good, but accord-ing to Ricardos remarkable theorem1 it will stillpay a country to specialise in those commoditiesin which it has a comparative advantage, that is inthose commodities in which relative labour pro-ductivity is the highest or for which the opportu-
nity cost of production is lowest. Ricardo is notexplicit about what determines relative differ-ences in productivity and costs, but clearlyresource endowments will be the major determi-nants: natural resources, labour, human capitaland the level of technology. To illustrate the gains
from trade, according to the law of comparativeadvantage, Ricardo used the example of Englandspecialising in cloth and Portugal in wine, so weshall do the same. England has a natural advan-tage in the production of cloth, and Portugal has anatural advantage in the production of wine.Suppose that in England the opportunity cost
ratio between cloth and wine is 10:1 (with clothmeasured in yards and wine in bottles), whereasin Portugal it is only 4:1. This means that ifEngland wanted to produce 1000 bottles of wine itwould have to sacrifice 10 000 yards of cloth,whereas if Portugal produced 1000 bottles of wine
it would sacrifice only 4000 yards of cloth. On theother hand, if England produced 1000 yards ofcloth, it would sacrifice only 100 bottles of wine,whereas if Portugal wanted to produce 1000 yardsof cloth it would sacrifice 250 bottles of wine.
6 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
24/265
Clearly, the opportunity cost of producing wine in
England is greater than in Portugal, and theopportunity cost of producing cloth in Portugalis higher than in England. If there was aninternational rate of exchange between Englandand Portugal between the two internal rates ofexchange of 10:1 and 4:1, both countries could
benefit from specialising in what they are best at
producing in an opportunity cost sense, andexchanging goods at a more favourable rate oftransformation internationally than domestically.For example, if the international rate of exchangeof cloth for wine was, say, 7:1, England could nowget more wine per unit of cloth sacrificed and
Portugal could get more cloth per unit of wine.Both countries would benefit from specialisingand trading the surplus of production overdomestic consumption. To use technical language,
by trading, the consumption possibility frontierwould lie outside the domestic production possi-
bility frontier for both countries. Ricardos
theorem is a very powerful one, and has beenextremely influential because it lies at the heart ofthe free trade doctrine that countries will always
benefit if they liberalise trade. Those who invokeRicardo, and the law of comparative advantage, insupport of trade liberalisation, however, often
forget (intentionally or otherwise) to mention itsimplicit assumptions and flaws which consider-ably weaken its appeal. Two main points will beemphasised here: first its static nature and itsindifference to the types of goods that countries
The theory and measurement of trade liberalisation 7
7/30/2019 1847208223
25/265
specialise in, and secondly, the assumptions of
continuous full employment and balanced trade.The law of comparative advantage is staticbecause the resource (consumption) gains from acountry moving from self-sufficiency to specialisa-tion and then trade are once-for-all. They do notrecur. Once a country has liberalised, and theprocess of resource allocation between activities has
taken place, there are no further gains. In otherwords, the comparative cost doctrine by itself hasno predictions about economic growth. The gainsfrom trade are not dynamic; they do not necessarilylaunch a country on a higher growth path. Thegrowth of economies through time depends on what
countries specialise in producing, not on the processof specialisation itself, because some activities havehigher potential growth rates than others. Thetheory of comparative advantage has nothing to sayabout productivity growth and its sources. It maybenefit Portugal to specialise in wine in a staticsense,but labourproductivity growth in viniculture
may be inherently lower than in cloth because thereis not so much scope for economies of scale or tech-nical progress. Indeed, viniculture is almost cer-tainly a diminishing returns activity because landsuitable for wine production is in fixed supply. Inthis case, Portugal is condemned to a slower growth
of its economy than England which specialisesin cloth, which is an increasing returns activitybecause all factors of production are variable.
What has been said in this example of Portugalis true for many developing countries today that
8 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
26/265
still specialise in the production of primary com-
modities which are land-based activities subject todiminishing returns. More than 50 per cent ofthe export earnings of developing countries comefrom primary commodities (and 70 per cent inAfrica). They specialise in these commodities forreasons of nature, and also for reasons of history.Many primary producing developing countries
today were once colonies of developed countriesthat were used as suppliers of raw materials, andthe processing of the primary commodities in situwas forbidden or discouraged (see later). Thestatic gains from this pattern of specialisation may
be offset by dynamic losses (in a relative sense
compared to other countries). Not even the staticgains may be that large. Dowrick (1997), in asurvey of trade and growth, makes the point thatalthough there are theoretical reasons why tradeliberalisation should increase welfare, it hasproved rather more difficult to come up with goodreasons why such welfare gains should be size-
able. He quotes typical estimates of the benefits oftrade liberalisation of about 1 per cent of GDP, andraises the question of whether a radical readjust-ment of an economy is worth such a small once-for-all gain. The answer must be no if only staticgains are on offer, with the prospect of future
dynamic losses. The case for trade liberalisationneeds to be made on the basis of dynamic gains,but, as indicated above, some countries, pursuingstatic comparative advantage, may get locked intoa productive structure of low growth activities.
The theory and measurement of trade liberalisation 9
7/30/2019 1847208223
27/265
This is why trade liberalisation needs to be com-
bined, at the very least, with a trade strategy whichseeks out new (dynamic) areas of comparativeadvantage, or promotes self-discovery to use theterminology of Hausmann and Rodrik (2003) (seelater). Trade liberalisation should not be regardedas an end in itself or as a substitute for a develop-ment strategy.
The second point to emphasise about the com-parative cost doctrine is that the welfare gainsfrom trade rest on the assumption of continuousfull employment and balanced trade in theprocess of resource reallocation. In the aboveexample of Portugal specialising in wine and
England in cloth, the free trade doctrine assumesthat all cloth workers in Portugal can find work inwine-making, and all wine-workers in Englandcan find work in the cloth industry. If this is notthe case, then the resource gains from specialisa-tion will be offset by welfare losses from theunemployment of resources. Indeed, if countries
have unemployed resources, the whole doctrineof comparative advantage based on the concept ofopportunity cost loses meaning, because whenunemployment exists the opportunity cost ofusing resources is zero. A contemporary examplewould be what is happening to farmers in Mexico
who cannot compete with the United States (US)maize growers under the terms of the NorthAmerican Free Trade Agreement (NAFTA) signedin 1994. They lose their jobs and have no alterna-tive employment. US producers, and Mexican
10 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
28/265
consumers, gain; Mexican farm workers lose
(who also consume less of other products), and anabsolute loss of welfare is possible. It was unem-ployment (combined with the free movement ofcapital) in the 1930s that convinced Keynes thatfree trade between countries is not necessarilyoptimal. In 1933, he published two articles, in theNew Statesman and Nation (8 and 13 July) entitled
National Self Sufficiency in which he argued letgoods be homespun and finance be primarilynational. In conditions of unemployment, theeconomic benefit of national self-sufficiency for acountry can outweigh the costs. Sir John Hicksalso recounts (Hicks, 1959) how unemployment in
the interwar years undermined his belief in freetrade:
The main thing that caused so much liberal opinion inEngland to lose faith in free trade was the helplessness ofolder liberalism in the face of massive unemployment, andthe possibility of using import restrictions as an element inan active programme of fighting unemployment. One, of
course, is obliged to associate this line of thought with thename of Keynes. It was this, almost alone, which ledKeynes to abandon his early belief in Free Trade. (p. 48)
One of the major causes of unemployment is theinability of countries to grow in line with theirproductive potential because of balance of pay-
ments difficulties. There is nothing in the doctrineof comparative advantage that guarantees bal-anced trade. Ricardo recognised this full well andrelied on the gold standard mechanism to producean equilibrium with the relative prices of goods
The theory and measurement of trade liberalisation 11
7/30/2019 1847208223
29/265
rising in surplus countries with gold inflows and
falling in deficit countries with gold outflows. Butthe international gold standard never worked thatway (Thirlwall, 2003), as Keynes (1936) recognisedin his General Theory in defending the doctrine ofmercantilism. In practice, surplus countries hadlow interest rates which stimulated investmentand growth, while deficit countries had higher
interest rates leading to unemployment. Whatadjusted the balance of trade was changes inincome (output) not relative prices (see alsoHarrod, 1933). Those worried today about theunemployment and balance of payments con-sequences of globalisation may find some of
Keyness arguments prescient and worth revisit-ing (see also Stiglitz, 2002).In 1848, John Stuart Mill published his Principles
of Political Economy which reiterated Adam Smithspoint about trade widening the market for goods,thereby permitting the division of labour: Acountry which produces for a larger market than its
own can introduce a more extended division oflabour, can make greater use of machinery, and ismore likely to make inventions and improvementsin the process of production. Mill also stressed therole of trade as a conduit for the international dis-semination of ideas and technology. There are three
main forces at work here. Firstly, a domestic buyerof an imported good may imitate the productiontechnique, or adapt the new technique if it ispatented. Secondly, there may be a direct exchangeof ideas between importer and exporter. Thirdly,
12 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
30/265
new imported goods may stimulate ideas for
further varieties of goods which increase welfare.New growth theory (or endogenous growththeory), which incorporates trade, as pioneered byGrossman and Helpman (1991a, 1991b), incorpor-ates many of these original ideas of Mill. But Millrecognised that the growth effects of trade dependon what a country specialises in; whether natural
resource activities or manufacturing industries. Acountry with abundant natural resources may finditself at a disadvantage because it is the productionof industrial goods, and particularly research-intensive goods, that produces technical dyna-mism. This is one aspect of what is now called in
the economic development literature the curse ofnatural resources which has been documented,among others, by Gylfason (2001) and Sachs andWarner (2001) who show that countries with 10 percent more of their labour force engaged in primaryproduction grow on average at 1 per cent less thanother countries, controlling for differences in other
country characteristics. One explanation is slowerproductivity growth in land-based activities com-bined with greater scope for rent-seeking behav-iour and corruption, particularly when naturalresources are in the hands of the State. The scenariois very different for countries specialising in indus-
trial activities. In the new growth theory modelsof Grossman and Helpman, referred to above,in which technical change is endogenous, tradeaffects growth by providing access to importedinputs and new ideas, and by increasing the size of
The theory and measurement of trade liberalisation 13
7/30/2019 1847208223
31/265
the market. They show that trade protection can
raise the long growth rate of an economy if itencourages investment in research-intensive activ-ities in which a country has a comparative advan-tage. Protection will also keep expected profitshigher than under free trade competition.
The neoclassical development of the doctrineof comparative advantage, associated with the
names of two Swedish economists, Eli Heckscher(1919) and Bertil Ohlin (1933), attributes differ-ences in relative costs of production not todifferences in natural resources or levels of tech-nology, but to differences in relative factor endow-ments, that is differences in the amount of capital
per unit of labour. The HeckscherOhlin (HO)theorem states that poor countries with an abun-dance of labour and scarce capital should find itrelatively cheaper to produce and export labour-intensive goods, while richer countries with morecapital and a relative shortage of labour shouldfind it cheaper to produce and export capital-
intensive commodities. This process should leadto greater wage equality in poor countries, and togreater inequality in richer countries. The link
between trade and wages is through changes inproduct prices. This linkage, formalised by Stolperand Samuelson (1941), says that in poor countries
the prices of labour-intensive products will rise,shifting resources to those sectors and raisingthe demand, and therefore wages, for unskilledlabour. Contrariwise, the price of skill-intensiveproducts will fall, reducing the demand for skilled
14 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
32/265
labour and reducing wages. The opposite happens
in rich countries, where the demand for unskilledlabour falls, and the demand for skilled labourrises. A corollary of the StolperSamuelson (SS)theorem is that protection increases the real wageof the scarce factor, because a tariff raises the priceof import-competing goods, so that a reduction inprotection will widen inequality in rich countries
where unskilled labour is the scarce factor andreduce inequality in poor countries if it is skill-intensive activities that are protected the most.Samuelson (1948, 1949) later extended the SStheorem to show rigorously how trade shouldequalise factor prices across countries, without
any factor movements.2
On the surface, neoclassical theory is appealingand sounds plausible, but in practice some ofthe basic underlying assumptions of the modelare weak, and the empirical evidence does notsupport it. The wage gap between unskilled andskilled labour in poor countries has not narrowed
with the process of trade liberalisation (see later,Chapter 4), and the wage gap between rich andpoor countries has not narrowed through time ifanything, it has increased. The weakest assump-tion of all is that technology is the same acrosscountries. But if the productivity of labour is lower
in poor countries than in rich countries, abundantlabour in poor countries is not necessarily cheaplabour in an economic sense. The money wagemay be lower, but what determines the economiccost of labour is the efficiency wage or the wage
The theory and measurement of trade liberalisation 15
7/30/2019 1847208223
33/265
cost per unit of output (wL/Y), which is the money
wage (w) relative to labour productivity (Y/L).This consideration is one of the major explanationsof what is called the Leontief Paradox, named afterWassily Leontief (1953) who first discovered that,contrary to the prediction of the HeckscherOhlintheorem, US exports were relatively labour-intensive compared to import substitutes.3
A major factor explaining differences in the levelof labour productivity between countries is differ-ences in the amount of human capital embodied inlabour in the form of education and skill training.Trade will not equalise factor prices if the effi-ciency wage is lower in rich countries than in poor
countries. A second point to bear in mind is thatthe theorem takes only two groups of countries poor developing countries and rich developedcountries but when a poor country liberalises itstrade, it will export and import more to and from
both developed and other developing countries.As the demand for labour-intensive exports from
poor countries rises in developed countries, wagesin poor countries may rise, but may fall as a resultof competitive imports from other poor develop-ing countries. Mexico, for example, benefits fromexporting labour-intensive products to the US, butnow suffers from labour-intensive imports from
China. What happens to the wages of unskilledlabour in poor countries is the outcome of abalance of forces. Finally, the HeckscherOhlintheorem ignores the effect of long-term capitalmovements between countries, and particularly
16 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
34/265
the impact of multinational investment in poor
countries. When trade is freed, it tends to stimulatethe flow of foreign direct investment to developingcountries which adds to the demand for skilled, aswell as unskilled, labour. If the demand for skilledlabour rises by more (relative to the supply), thiswill increase the wages of skilled relative tounskilled labour in poor countries, also defying
the predictions of the HeckscherOhlin model. Aswe shall see in Chapter 4, income and wageinequality is increasing in both rich and poorcountries, driven by technological progress andtrade which demands skilled labour.
Free Trade Enthusiasm in the Modern Era
Despite the putative arguments for free trade laiddown by classical and neoclassical economictheory it was never seriously practised by coun-tries (except by Britain post-1850) until after theSecond World War with the establishment of the
General Agreement on Tariffs and Trade (GATT) in1947 and the general commitment by developedcountries to the freeing of international trade inthe wake of the protectionism and beggar-thy-neighbour policies practised in the inter-wardepression years. Even so, the process of trade lib-
eralisation took a long time to gather momentum.Many developed countries maintained quite hightariff levels, and non-tariff barriers, until the early1970s, and many developing countries, freed fromtheir colonial past, went down the protectionist
The theory and measurement of trade liberalisation 17
7/30/2019 1847208223
35/265
route, particularly in Latin America and parts of
South Asia. It is only in the last 30 years or sothat both developed and developing countrieshave made concerted efforts to liberalise trade
between themselves, under pressure from avariety of sources. Dornbusch (1992) identifiesfour major pressures on developing countries:(i) the intellectual swing in favour of free markets;
(ii) institutional pressures from the World TradeOrganization (the successor to GATT) establishedin 1995, the World Bank, the IMF and otherinternational organisations; (iii) greater awarenessof the efficiency losses from restricting trade; and(iv) the alleged, or apparent, poor economic per-
formance of countries pursuing protection.The tide of opinion in favour of free marketshad its intellectual roots in the ideas of econo-mists such as Friedrich Von Hayek (1949) andMilton Friedman (1990), and gathered pace withthe disillusion with planning, particularly aspractised by the former Soviet Union and Eastern
European countries. The role of the State in eco-nomic affairs came under increased scrutiny byboth economic theoreticians and practitionersalike. Anti-statism became the fashion, and in theinternational arena it led to the development ofthe so-called Washington Consensus a term
originally coined by John Williamson of theInstitute for International Economics in 1989 torefer to an agenda of desirable economic reformsin Latin America (see Williamson, 1993). Thereforms quickly came to be seen as a model for
18 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
36/265
the wider developing world, and was already
embodied in the thinking of the World Bankand IMF.The Washington Consensus extols the virtues of
free markets and free trade for the achievementof more rapid economic progress, and includesamong its major objectives: the privatisation ofState enterprises; trade liberalisation, particularly
the replacement of quantitative restrictions ontrade with low and uniform tariffs, and opennessto foreign direct investment. This is the routedown which most developing countries have nowgone. The institutional pressure from the WorldBank and IMF for trade reform is well summarised
in a direct quote from the IMF itself which statesthat the main goal of IMF-supported programmesin developing countries has been to improve theeconomic efficiency by creating a transparent andneutral system of incentives that eliminates anti-export bias, direct impediments to trade, and eco-nomic distortions caused by the trade regime
(IMF, 1998).The greater awareness of the efficiency losses
from restricting trade came from major empir-ical studies across countries sponsored by theOrganisation for Economic Co-operation andDevelopment (OECD) and the National Bureau of
Economic and Social Research in the US showingthe apparent wastefulness of import substitutionregimes (see, for example, Little, et al., 1970; andsummaries of this and other studies by Bhagwati,1978 and Krueger, 1978). The East Asian experience
The theory and measurement of trade liberalisation 19
7/30/2019 1847208223
37/265
of an outward-oriented development strategy
based on exports, rather than a trade policy basedon import substitution, was also regarded assalutary. Interestingly, however, economic successlinked to export performance has not always beenbased on free trade and laissez-faire. It is true thatthe economies of Japan, South Korea, Taiwan,Singapore, Hong Kong, Malaysia, Indonesia and
Thailand have recorded some of the highest exportgrowth rates in the world since 1965 (averagingover 10 per cent per annum) and some of thehighest GDP growth rates (averaging nearly 6 percent per annum), but some of these countries havebeen very interventionist pursuing export promo-
tion and import substitution at the same time.Japan and South Korea are clear examples, not tomention, today, modern China which is farremoved from a laissez-faire, free trade economy,but which has the fastest rate of growth of exportsand GDP in the world. China pursues an indepen-dent development strategy sometimes called the
Beijing Consensus (in contrast to the WashingtonConsensus), which makes it cautious about freetrade and the free movement of internationalcapital, although not about attracting long-termforeign direct investment. China is fortunate to belarge enough (and stable enough) to plough its
own furrow. Many developing countries, unfortu-nately, are either too small and vulnerable or toounstable to resist the orthodoxy because they aredependent on loan support from the IMF andWorld Bank.
20 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
38/265
The premise of poor economic performance
of developing countries under protective traderegimes, as a motive for liberalisation, is not factu-ally accurate. After the Second World War, whilethe developed countries started to dismantle tariffsand non-tariff barriers to trade under the auspicesof GATT, the developing countries were allowedconsiderable policy space and made no firm com-
mitments to free trade. During the period from1950 to 1979, the developing countries had theirbest economic performance in recent economichistory (Chang, 2002). During colonialism, Asiaand Africa grew extremely slowly, but in the post-colonial era until 1979 average per capita income
grew at 3 per cent per annum, more than doublethat achieved by developed countries in the 19thcentury during their development process. Not allthe success achieved during this period can beattributed, of course, to trade protection, but it issignificant that in the period of neo-liberalism post-1980 the growth performance of many developing
countries, particularly in Africa and Latin America,has been worse than in the pre-liberalisation era. InAfrica as a whole, between 1980 and 2000, livingstandards actually fell, and per capita incomegrowth in Latin America has averaged only 1.5 percent per annum. Trade liberalisation has not deliv-
ered the performance promised by its advocates.As Rodrik (2004) says of Latin America:
Latin America [during the 1990s] grew more slowly notonly compared with other parts of the world but also
The theory and measurement of trade liberalisation 21
7/30/2019 1847208223
39/265
compared to its own performance in the 1960s and 1970s.That is a striking empirical fact, the importance of which ishard to downplay. After all, Latin America of the 1960sand the 1970s is a region of import substitution, macro-economic populism, and protectionism, while LatinAmerica of the 1990s is a region of openness, privatisationand liberalisation. The cold fact is that per capita eco-nomic growth performance has been abysmal during the1990s by any standards. (p. 3)
The Case for Protection
Having rehearsed the orthodox arguments forfree trade, and highlighted the recent political,institutional pressures on developing countries toembrace trade liberalisation (sometimes on spuri-
ous grounds), it is now time to give equal weightto the perfectly respectable economic argumentsfor protection, and then to look at the historicalexperience of countries now developed. Did theygain their industrial strength and economic poweron the basis of free trade or protection? There is aninteresting story to tell.
The first point to make is that for free trade ingoods and services to maximise social welfare itneeds to be shown, or proved, that the demandcurve for goods measures the true social benefit ofproduction and that the supply curve measuresthe true social cost. If this is the case, a restriction
on trade, in the form of a tariff for example, willcause production and consumption distortionswhich reduce welfare, as illustrated in Figure 1.1.
DD and SS represent the domestic demand andsupply of the good, respectively. Pw is the world
22 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
40/265
price of the good, and S1S2 represents importdemand at world prices to fill the gap betweendomestic demand and supply. Now suppose a per-centage tariff (t) is imposed on the good raising theprice to Pw(1t). This produces two distortions:firstly, a production distortion, or loss of producer
surplus, measured by the shaded area under thesupply curve, because costs have risen; and sec-ondly a consumption distortion, or loss of con-sumer surplus, measured by the shaded areaunder the demand curve, because prices haverisen. These shaded areas under the supply and
demand curves representing losses of producerand consumer surplus are often called Harbergertriangles (Harberger, 1954) named after the distin-guished Chicago economist Arnold Harberger. Ifthe supply and demand curves, however, do not
The theory and measurement of trade liberalisation 23
S3 S4 S2S1 Quantity
Price
Pw(1 + t)
Pw
D
D
S
S
Figure 1.1 Production and consumption distortionsfrom a tariff
7/30/2019 1847208223
41/265
represent the true social cost and benefit of pro-
duction and consumption, the loss of producerand consumer surplus from restricting trade andincreasing domestic output may be exaggerated.In particular, the producer surplus (which is lost)may not properly measure the benefits of prod-ucing a good domestically because of variousdomestic distortions and market failures, such as
positive externalities and technological spilloversfrom one industry to another or the existence ofunemployment. In fact it is perfectly possiblefor the marginal social benefit from protectingindustry, and increasing domestic productionfrom S1 to S3 in Figure 1.1, to more than offset this
loss of consumer and producer surplus. HarryJohnson (1964) identifies four major distortionsparticularly in the context of developing countrieswhich constitute an economic argument for protec-tion (so a countrys real income will be above whatit would otherwise be), although not necessarilyan argument for tariffs if the distortions are dom-
estic (see later):
1. The infant industry argument for protection:allowing industries to reach their optimum size interms of average costs of production, so that theycan compete with foreign competition in domes-
tic and foreign markets. This argument goes backat least to Mills Principles of Political Economy(1848), and to the writings of the German econo-mist Friedrich List (17891846) in the 1840s. Millpromoted the idea of temporary infant industry
24 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
42/265
protection to stimulate industry. He writes: the
only case in which, on mere principles of politicaleconomy, protecting duties can be defensible, iswhen they are imposed temporarily . . . in thehope of naturalizing a foreign industry.4 ForList, infant industry protection was just oneof a series of measures recommended for theindustrial development of countries, recognising
that economic development requires a combina-tion of capital accumulation, technical progressand learning capabilities from both home andabroad. List was not the ardent protectionist thatneo-liberals claim (Sai-wing Ho, 2005); he had amuch broader vision of the growth and develop-
ment process, and argued, in fact, that protectionshould be very selective: it should attempt tostimulate only those industries which have anassured home market and appear to have the bestchance of success. List anticipated the idea maderigorous by Krugman (1984) of import protectionas export promotion because a large domestic
market, in the presence of economies of scale anddecreasing costs, makes exports more competi-tive. Turning back to Figure 1.1, the supply curvethere is drawn upward sloping representing anincrease in the cost of producing additional unitsof output, whereas if there are economies of scale,
and fixed costs can be spread over a larger volumeof output, costs will fall and the supply curve willbe downward sloping. There is no loss of pro-ducer surplus, but a gain. In recent times econo-mists such as Stiglitz (2006) and Chang (2002)
The theory and measurement of trade liberalisation 25
7/30/2019 1847208223
43/265
have argued forcibly and persuasively that devel-
oping countries need to protect themselves toswitch their productive structure from reliance onprimary production to the production of manu-factured goods. Stiglitz argues not just for infantindustry protection but infant economy protection.In other words, it may be better to impose auniform tariff on all manufactured good imports
to give time for an infant economy to establisha manufacturing base. This avoids the rent-seeking behaviour of special interest groups thatlobby for preferential protection for their ownindustry, and at the same time provides revenuefor governments to spend for development pur-
poses such as infrastructure, education andhealth. Stiglitz (2006) puts the simple but endur-ing case for protection very well when he writes:
Without protection, a country whose static comparativeadvantage lies in, say agriculture, risks stagnation; itscomparative advantage will remain in agriculture, withlimited growth prospects. Broad based industrial protec-
tion can lead to an increase in the size of the industrialsector which is, almost everywhere, the source of innova-tion; many of these advances spill over into the rest of theeconomy, as do the benefits from the development of insti-tutions, like financial markets, that accompany the growthof an industrial sector. Moreover, a large and growingindustrial sector (and the tariffs on manufactured goods)provide revenues with which the government can fund
education, infrastructure, and other ingredients necessaryfor broad-based growth. (p. 72)
In fact, the effect on imports of uniform revenue-raising tariffs on all industrial imports would be
26 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
44/265
little different from an equivalent depreciation of
the currency or temporary industrial subsidies but with the latter policies there would be norevenue.
Chang (2002), in his famous book Kicking Awaythe Ladder (evocative words taken from List)reminds us of the historical experience of the nowdeveloped countries (see below), and the time it
takes for industries to reach their most efficientscale to compete both domestically, with foreigncompetition, and internationally, because learningand absorption of technology is a gradual process.Before the Second World War, the comparativeadvantage of Japan and South Korea was in rice
production, yet today these countries have two ofthe most successful steel industries in the world,based not on free trade but protection over anumber of years. The examples could be multi-plied; on the other hand, protection can go wrong.Inappropriate industries may be chosen, or pro-tection be granted for too long so that the infant
industries never grow up. The conditions forsuccess are firstly that the choice of infant indus-tries needs to be realistic; secondly that infantindustry protection needs to be combined with anexport strategy in order to provide the foreignexchange to pay for imported inputs and to widen
the market; and thirdly that the rent-seekingactivities of government officials and politiciansmade possible by protective policies need to becontrolled (as Stiglitzs proposal would do). AnneKrueger (1997), an ardent free-trader, argues that it
The theory and measurement of trade liberalisation 27
7/30/2019 1847208223
45/265
is impossible to identify infant industries, but the
facts and history are against her. But, in any case,even if the infant industries are relatively ineffi-cient, it may be more beneficial for an economy tohave some industry than no industry at all becauseof knowledge externalities and spillovers that inrecent years have been highlighted by so-callednew trade theory (see below). Krueger (1998)
further claims that the traditional infant industryargument of dynamic gains outweighing staticlosses has now been overturned. This is simplynot true.2. The externalities argument for protection: wherethe social benefit of increased domestic production
is greater than the private benefit due to spillovers.There are both micro and macro considerations(Ocampo and Taylor, 1998), both relating to pro-ductivity effects. In a free market, resources maynot shift naturally or easily from low-return tohigh-return sectors of an economy because of bar-riers to mobility. If the manufacturing sector is the
major source of productivity growth in the earlystages of development, because it is an increasingreturns activity, promoting industry, or not pro-tecting it, has implications for the growth of thewhole economy. At the micro level there arethe technological spillovers from one industry to
another to consider, arising from research anddevelopment (R&D), knowledge creation, learn-ing by doing and the production of skilled labour.Without the protection of industry, a substantialsource of productivity growth may be lost.
28 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
46/265
These considerations are the foundations of
new trade theory, pioneered by Krugman (1986)and Grossman and Helpman (1991b) in the 1980sand 1990s, which uses increasing returns andimperfect competition to explain why large spatialdifferences in economic development exist withinand between countries, when orthodox tradetheory predicts convergence. It is essentially the
division of labour, or specialisation, that leads toincreasing returns and the concentration of manu-facturing in particular geographic locations whichthen generates externalities, both pecuniaryand technological. In addition, with knowledgespillovers from innovating to following firms, new
trade theory extends the infant industry argumentto protecting not only industries, but the process ofinnovation itself. Grossman and Helpman (1991b)show how trade liberalisation can have negativeeffects on R&D by displacing innovative activities.Without protection, countries will tend to spe-cialise in low-technology products. Trade restric-
tions give the opportunity for countries to developa more complex, sophisticated, interconnectedindustrial base. This is what the historical evi-dence shows. All this means that ifthere is a short-term cost of protecting infant industries,5 it isoutweighed by the long-term gains from a more
dynamic industrial structure. As in new growththeory, which endogenises technical change,protection will promote growth if the forces ofcomparative advantage push resources towardsactivities with superior growth potential; that is
The theory and measurement of trade liberalisation 29
7/30/2019 1847208223
47/265
activities that promote R&D, innovate and
produce goods higher up the value chain. Despitethe logic of the argument, however, the pioneers ofnew trade theory have been reluctant to endorseprotection; witness, for example, the remarkableclaim by Krugman (1993a) that by emphasisingthe virtues of free trade, we also emphasise ourintellectual superiority over the unenlightened
who do not understand comparative advantage.The fear of protection seems to be rent-seeking
behaviour and that governments lack the neces-sary information for optimal policy intervention.To quote Krugman (1987) again: Free trade rulesare best for a world whose politics are as imperfect
as its markets. However, this did not stop manycountries in South East Asia adopting remarkablysuccessful strategic industrial and trade strategieswhich have propelled them into industrial giantsto challenge the manufacturing base of Europeand North America. Perhaps it is these countriesthat Krugman (1987) had in mind when he also
declared free trade is notpassbut it is an idea thathas irretrievably lost its innocence . . . it can never
be asserted as the policy that economic theory tellsus is always right.3. The domestic distortion argument for protection:for example, in the labour market where the social
cost of using labour is less than the private costbecause of unemployment. This is a classic argu-ment for intervention that goes back to AdamSmith.6 Poor developing countries have pervasiveunemployment, so that the social (opportunity)
30 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
48/265
cost of using labour is less than the private cost(the wage). The argument is illustrated in Figure1.2, which is a modification of Figure 1.1.
S1S1 is the social cost of production, and SS is theprivate cost of production (as before) which is tper cent above the social cost. Under free trade,domestic producers will produce up to Q1 where
the private cost equals the world price of the good,Pw. Demand is Q4, so Q1Q4 will be imported. Q1Q2imports could be replaced, however, by additionaldomestic production Q1Q2, with real savings equalto the area A, if domestic producers are eitheroffered tariff protection which brings the domestic
price to Pw(1t), or a subsidy oft per cent whichreduces the private cost to the social cost. It can beseen immediately in the example, however, thatthe welfare effects of tariffs and subsidies are notthe same. In the case of a tariff, t, the price of the
The theory and measurement of trade liberalisation 31
Q2 Q3 Q4Q10 Quantity
Price
Pw(1 + t)
Pw
SD
S1
S1
S A
B
Figure 1.2 Protection when social cost of labour isless than the private cost
7/30/2019 1847208223
49/265
good rises and there is a loss of consumer sur-
plus, B, which may be greater than the welfaregain ofA. In the case of a subsidy, however, thereis no loss of consumer surplus, and the welfaregain is unequivocal. In this respect, a subsidy tolabour is first best (see later).4. The optimal tariff argument for protection: whereinternational distortions cause the domestic rate of
transformation between goods to diverge from theforeign rate of transformation due, for example, tomonopoly power in international trade, so that ifa country is a big buyer, it can turn the terms oftrade in its favour by restricting imports. This gaincould be in excess of the efficiency loss from the
production and consumption distortions from atariff illustrated in Figure 1.1. The argument goesback at least to Robert Torrens (1833). In practice,the optimum tariff argument is not a serious onefor developing countries because they are not bigenough buyers of products individually. But thereis another terms of trade argument, and that is
that the price of primary commodities relative toindustrial goods has been falling by at least 1 percent per annum on average for the last 100 yearsor more (see Cashin and McDermott, 2002). Thisrepresents a substantial real income loss for coun-tries specialising in the production and export of
primary commodities. If trade liberalisation re-inforces traditional patterns of specialisation, thereal income loss for poor developing countries isperpetuated. There is a case for protection to alterthe structure of production in favour of commodi-
32 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
50/265
ties that have more favourable terms of trade in
world markets. This was one of the major argu-ments originally made by Prebisch, as discussedearlier, in advocating import substitution policiesin Latin America. The avoidance of terms of tradeloss may offset the costs of protection.7
Another economic argument for protection,not considered by orthodox theory, is the balance
of payments argument. Johnson (1964) calls thisargument a non-argument (like the terms oftrade argument) because of the classical/neoclassi-cal presumption that the balance of payments isself-adjusting through relative price, or exchangerate, changes. If trade liberalisation, however,
causes the balance of payments to deterioratebecause imports rise faster than exports, and it isincome that falls to restore balance of paymentsequilibrium (not relative price changes), the realresource gains from trade may be offset by thewelfare losses from the underutilisation of domes-tic resources. In other words, if trade liberalisation
leads to balance of payments problems which causeunemployment, the social cost of labour is less thanthe private cost of labour, and we have seen abovethat in these circumstances there is a legitimate eco-nomic argument for protection. We will examine inChapter 2 the evidence on the balance of payments
consequences of trade liberalisation. Interestingly,Article XVIII of GATT allowed member countries totemporarily restrict trade in goods and servicesfor balance of payments purposes, just as therewas a scarce currency clause in the Articles of
The theory and measurement of trade liberalisation 33
7/30/2019 1847208223
51/265
Agreement of the IMF that would have allowed
countries to discriminate against goods from coun-tries in balance of payments surplus but the clausehas never been used.
Despite the body of trade theory which legit-imises protection, Dani Rodrik (1988), the wellknown trade and development economist fromHarvard, is correct that the arguments have still
not penetrated the vast literature on trade policy indeveloping countries which continue to reflect theorthodox view that trade liberalisation is optimal,as if written in tablets of stone. Yet, as Rodrik says,the market imperfections of the type analysedabove would appear, if anything, to be more
prevalent and serious in developing countriesthan in developed countries. One of the mostardent exponents and advocates of the free tradedoctrine, even for developing countries, is thedistinguished trade theorist Jagdish Bhagwati.Krugman says of Bhagwati (in Subramanian, 2005)it is he, above all else, who has played a large if
subtle role in keeping protectionism from becom-ing respectable.
Bhagwatis Lament
In his polemical book, Free Trade Today, Bhagwati
(2001) bemoans (unnecessarily) the lack of econo-mists willing to put the argument for free trade.He recognises many of the distortions discussedabove, but is adamant that they should be tackleddomestically rather than through trade restric-
34 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
52/265
tions. Indeed, it was Bhagwati and Ramaswami
(1963) who first showed rigorously how domesticmarket distortions are not an argument againstfree trade, but for correcting the distortionsdirectly. In the labour market, if the social cost oflabour is less than the private cost, the first-bestpolicy is a wage subsidy plus free trade, as we sawin Figure 1.2. In the product market, if there is a
production externality, the first-best policy wouldbe a production subsidy plus free trade. In the caseof international distortions, such as monopolypower in trade, where a countrys terms of tradedepend on its volume of trade, the optimum tariffargument applies, and this is the first-best policy
but this is the only case for tariffs that Bhagwaticoncedes. He is also concerned with the directlyunproductive profit-seeking activities (DUP) (aphrase coined by him) which arise when tariffs,quotas and licenses are used to control trade (seealso, Tullock, 1967). These need to be taken ser-iously, but should not be exaggerated. As Ocampo
and Taylor (1998) say, if quotas cover only a frac-tion of imports, and imports are a fraction of GDP,associated rents and rent seeking outlays cannot
be huge. Moreover, if there is not full employ-ment, or the balance of payments is a constraint ongrowth, the additional income flows associated
with a larger volume of domestic output bythe imposition of import controls will generatewelfare gains. It is also possible that rents in theright hands may be invested productively and notconsumed.
The theory and measurement of trade liberalisation 35
7/30/2019 1847208223
53/265
If subsidies to labour and production are first
best when market imperfections and distortionsare domestic, and protection is second best, theissue arises of how subsidies are financed. If theyare financed by taxes on wage income and profits,this adds further distortions in an economy. Inpractice, a large fraction of tax revenue in devel-oping countries is raised from trade taxes. Tariffs
are being used to pay subsidies. Reducing tradetaxes, to make trade freer, may seriously under-mine a countrys fiscal position and make it moredifficult to pursue desirable domestic policies topromote development which are also welfareenhancing. This is where issues of the distribution
of income and consumption become important toconsider. The welfare effects of imposing tariffs onnecessary consumption goods for the poor, andon necessary inputs for industrial development,are not the same as the imposition of tariffs orquotas on luxury consumption goods for the rich.Bhagwatis plea for greater freedom of trade is
silent on these distributional issues.Indeed, none of Bhagwatis attempted rehabilita-
tions of free trade address directly the importantissues of infant industry protection, the need forstructural change in poor countries, or the distri-butional consequences of trade liberalisation and
its effects on the poor (see later for the evidence).On the other hand, he is honest when he makes theimportant distinction between the effect of trade oncurrent welfare on the one hand, and the effect oftrade ongrowth on the other; they are not the same.
36 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
54/265
In most standard growth models, the effect of trade
on growth is ambiguous. In the Harrod growthmodel, for example (see Thirlwall, 2002), freetrade may reduce the capitaloutput ratio, raisinggrowth; but may also reduce the savings ratio,reducing growth. In the neoclassical Solow model(1956), trade cannot affect the steady-state growthrate, although it can be altered if the steady state is
not treated as an exogenous constant. Only in thenew growth theories of Grossman and Helpman(1991a, 1991b) does trade have the potential to raisethe growth rate permanently through learning andspillover effects, but only if they are continuous.Bhagwati remarks, interestingly:
Those who assert that free trade will . . . lead necessarilyto greater growth either are ignorant of the fine-nuances ofthe theory and the vast quantity of literature to the con-trary on the subject at hand or are nonetheless basing theirargument on a different premise; that is, that the prepon-derant evidence on the issue (in the post-war period) sug-gests that freer trade tends to lead to greater growth afterall. In fact, where theory includes several models that canlead in different directions, the policy economist is chal-lenged to choose the model that is most appropriate to thereality she confronts. And I would argue that, in thepresent instance, we must choose the approaches thatgenerate favourable outcomes for growth when trade isliberalised. (pp. 423)
In Chapter 2 we shall examine the evidence onthe relationship between trade liberalisation andeconomic growth, but first it is useful and salutaryto look at the experience of the now-developedcountries to see whether they developed on the
The theory and measurement of trade liberalisation 37
7/30/2019 1847208223
55/265
basis of free trade or protection. Unfortunately for
the free trade doctrine, it is the latter that turns outto be the case, so the developed countries (andinternational institutions) that preach free trade todeveloping countries are preaching what theynever practised themselves.
Protection in Historical Perspective
The best historical description of the role of pro-tection in the early industrialisation phase of thenow-developed countries is given by Ha-JoonChang, the Cambridge economist, in three fascin-ating books: Kicking Away the Ladder: Development
Strategy in Historical Perspective (2002); WhyDeveloping Countries Need Tariffs? (2005); and BadSamaritans: Rich Nations, Poor Policies and the Threatto the Developing World (2007). In this section werely heavily on the evidence in these books.8
The current developed countries of the world,including Britain, the United States and the coun-
tries of continental Europe and Scandinavia did notdevelop their economies on the basis of free trade.On the contrary, they heavily protected theirdomestic industries, and also did their utmost toprevent the countries that they colonised from com-peting with them. Britain started to protect and
foster industries as early as the late 15th centurywhen Henry VII took the deliberate decision tochallenge the successful woollen manufacturingindustry of Belgium and Holland, which wasreliant on the export of British wool. He taxed the
38 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
56/265
export of raw wool and banned export of some
types of unfinished cloth in order to encourage pro-cessing at home. Henry VIII continued the protec-tionist policy, and by the middle of the reign ofElizabeth I, Britain had sufficient processing capac-ity to ban the export of wool entirely, which ruinedthe cloth industry of the Low Countries. Britain firstbecame rich on its woollen industry nurturedby the
State. Serious protection of new manufacturingindustries started with Robert Walpole in 1721,using tariffs, subsidies, tariff rebates on importedinputs and other protective devices all of whicharedeemedtobedamagingtodevelopingcountriestoday. In the early 19th century, Britain imposed
some of the highest tariff rates on manufacturedgoods in the world, averaging 4555 per cent.Britain also prevented its colonies from produc-
ing manufactured goods. William Pitt the Elder, theBritish Prime Minister from 1766 to 1768, is quotedby Friedrich List (1885) as saying that the coloniesshould not be permitted to manufacture so much
as a horsenail. All sorts of devices were resorted toin the 18th century to keep the colonies as produc-ers of primary commodities, giving subsidies toproduction, and reducing tariffs on raw materialimports into Britain. A law passed in 1699 forbadethe export of processed wool products from the
English colonies, including Ireland. In 1700, allcotton goods from India were prohibited. In the1720s, Walpole gave export subsidies and abol-ished import duties on raw materials produced inthe American colonies so that their comparative
The theory and measurement of trade liberalisation 39
7/30/2019 1847208223
57/265
advantage stayed in primary products. Some man-
ufacturing activities were even prohibited, such ashigh value-added steel products in America. Theuse of tariffs by the colonies was either bannedor, where used for revenue purposes, a tax wasimposed on the industry concerned to neutralise itscompetitive advantage. In other countries notcolonised by Britain, unequal treaties were signed
which took away the tariff autonomy of the coun-tries and set binding tariffs that countries couldnot exceed, typically about 5 per cent in countriessuch as Brazil, China, Japan, Siam (now Thailand)and Persia (now Iran). With regard to Europe,Britain also tried to protect itself against competi-
tion, although to less effect. The export of sometypes of machinery embodying new technologywas banned, and for over sixty years from 1719 to1782 there was a ban on the emigration of skilledlabour from Britain. Those who defied the ban, anddid not return within six months, had their posses-sions confiscated and citizenship withdrawn.
Britains industrial revolution gathered momen-tum in the mid-18th century, when protection stillprevailed. It would be a rewriting of history, there-fore, to argue that Britain started its developmentprocess on the basis of free trade. Britain did notstart dismantling its structure of protection until
the repeal of the Corn Laws in 1846, but by then ithad already attained technological superiorityover all other countries in the world. From then on,Britain preached free trade, but as List (1885)remarked, such preaching was like kicking away
40 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
58/265
the ladder up which one has climbed oneself so
that no-one else can reach the top. List comments:It is a very common clever device that where anyone hasattained the summit of greatness, he kicks away the ladderby which he has climbed up, in order to deprive others ofthe means of climbing up after him. In this lies the secret ofthe cosmopolitan doctrine of Adam Smith, and the cos-mopolitan tendencies of his great contemporary William
Pitt, and of all his successors in the British Governmentadministrations. Any nation which by means of protectiveduties and restrictions on navigation has raised her manu-facturing power and her navigation to such a degree ofdevelopment that no other nation can sustain free competi-tion with her, can do nothing wiser than throw away theseladders of her greatness, to preach to other nations the ben-efitsoffreetrade,and todeclare inpenitent tonethatshe hashitherto wandered in the paths of error, and has now for thefirst time succeeded in discovering the truth. (pp. 2956)
The United States followed Britains protection-ist route at the end of the 18th century, contrary toAdam Smiths advice in the Wealth of Nations. Hereis what Smith had to say:
Were the Americans, either by combination or by anyother sort of violence to stop the importation of Europeanmanufactures, and, by thus giving a monopoly to such oftheir own countrymen as could manufacture the likegoods, divert any considerable part of their capital intothis employment, they would retard instead of accelerat-ing the further increase in the value of their annualproduce, and would obstruct instead of promoting the
progress of their country towards real wealth and great-ness. (pp. 3478)
If the United States had followed Adam Smithsadvice, it would have remained an economic
The theory and measurement of trade liberalisation 41
7/30/2019 1847208223
59/265
backwater for a long time, instead of becoming the
richest industrialised country in the world. In the19th century, the US economy was the fastestgrowing in the world, and also the most protec-tionist. Paul Bairoch (1993) has described theUnited States as the mother country and bastionof modern protectionism. It was the US TreasurySecretary, Alexander Hamilton in 1791, who first
coined the term infant industry, and who firstargued the case for industrialisation by protectionusing tariffs, subsidies and other means, recognis-ing that without protection it would be impossiblefor America to compete against more advancedcountries, notably Britain. List, in his classic book
The National System of Political Economy, first pub-lished in German in 1841, claims that he first learntthe infant industry argument for protection whilein exile in the US in the 1820s. The US first imposedtariffs on industrial goods in 1789. Protection con-tinued to increase in the 19th century and by 1870,import tariffs accounted for more than 50 per cent
of the value of imports. Protection continued in theearly 20th century, and was even strengthened inthe 1930s with the SmootHawley tariff whichraised the average tariff on manufactured goods tonearly 50 per cent. According to Bairoch (1993) noother country implemented a more protectionist
policy to promote its industry than the UnitedStates. Only after the Second World War did it startto liberalise its trade, having already establishedindustrial supremacy, and was able to kick awaythe ladder, as Britain had done a century earlier.
42 Trade liberalisation and the poverty of nations
7/30/2019 1847208223
60/265
German industrial policy in the 19th century
was heavily influenced by the views of List. Hebelieved that import duties should not only beused to protect industry but also to promote it,supported by the State. Germanys average tariffrate on industrial goods was not as high as inthe US, but the German State actively promotedindustry by assigning monopoly rights, establish-
ing industrial cartels, providing export subsidies,importing industrial experts and skilled labour,establishing large banks and making large invest-ments in coal production and railway and roadconstruction (Skarstein, 2007).
Japan was prevented from using tariffs up to
1911 because of the unequal treaties signed, asreferred to earlier. But after 1911, Japan embarkedon a comprehensive development strategy, a majorpart of which included substantial tariff protection,combined with subsidies to key (infant) industriesand State investment in infrastructure. Just beforethe First World War, Japans average tariff on man-
ufactured imports was 30 per cent. The protection-ist stance continued after the Second World War,with tariffs on car imports, for example, of nearly40 per cent. Protectionism in the 1950s and 1960swas combined with the highest GDP growth rate ofany country in the world. If Japan had listened
to the free-traders, it would have no industrialbase.The average tariff rates on manufactured goods
for selected developed countries in their earlystages of development are shown in Table 1.1.
The theory and measurement of trade liberalisation 43
7/30/2019 1847208223
61/265
44 Trade liberalisation and the poverty of nations
Table 1.1 Average tariff rates on manufactured productsfor selected developed countries in their earlystages of development (weighted average, in
percentages of value)1
Country 18202 18752 1913 19