Report No: IDP - 142
July 1994
Regional Trading Arrangements and Beyond
Exploring Some Options for South Asia
Theory, Empirics and Policy
T.N. Srinivasan
Al~I
A4
.~ ~ -
The World Bark
South Asia Region
Office of the Chiecf Econiomist
The views presented here are those of the author,
and they should not be interpreted as reflecting those of the World Bank.



SOUTH ASIA REGIONAL SERIES
Title                                      Author             Date              Oriainator
IDP 111   How ComposWion of Public Expenditure
Affects Competitiveness: The Case of
Bangladesh                                 K. Matin           March 1992        P. Mitra (80419)
IDP121   Labor Retrenchment and Redundancy
Compensation in State Owned Enterprises:
The Case of Sri Lanka                      A. Fiszbein        December 1992    G. Nankani (84641)
IDP126   Some Guidelines for the Appraisal of
Large Projects                             W. Jack            February 1993     A. Estache (81442)
IDP109   Reforming Higher Secondary Education iL
in South Asia: The Case of Nepal           H. Abadzi          May 1993          H. Abadzi (80375)
IDP1 27   Some Lesons for South Asia from
Developing Country Experience with
Foreign Direct Investment                  M. Fry             June 1993         A. Estache (81442)
IDP129   Quasi-Fiscal Deficits: Latin American
Lessons for South Asia                     C. A. Rodriguez    August 1993       A. Estache (81442)
IDP131   The Impact of Rural Infrastructure on
Rural Poverty: Lessons for South Asia      E. Goldstein       June 1993         G. Nankani (84641)
IDP1 34 Infrastructure and Industial Policy in
South Asia: Achieving the Transiton to
a New Regulatory Environment               P. Seabright       June 1993         A. Estache (81442)
IDP142   Regional Trading Arrangements ard Beyond
'Exploring Some Options for Soutl Asia
Theory, Empirics and Policy                T.N. Sriniasan     July 1994         G. Nankani (84641)
Note: The denominations, the classifications, the boundaries, and the colors used in the Regional map on the cover do not
imply on the part of The World Bank and its affiliates any judgment on the legal or other status of any territory,
or any endorsement or acceptance of any boundary. The purpose is to give the readers an overview of the Region.



REGIONAL TRADING ARRANGEMENTS AND BEYOND
EXPLORING SOME OPTIONS FOR SOUTH ASIA
THEORY, EMPICS AND POLICY
T. N. Srinivasa
July 1994



FOREWORD
South Ash's move towd freer markets and freer trade has a new momntum. As in other regions
during the 1980s and before, each county in South Asia has, in its own way, receatly begm to restruc   the role
of the stae, liberalize and deregulate markets, and let the private sector take a larger, more acdve economic role.
The pace may be slower than in other regions, for many important areas of each economy have yet to be reformed.
But in the minds of most South Asians, the sustainability of reforms is no longer in doubt. Indeed most would deem
the reforms irrevesible. There nevertheless is a differene with other regions. Despite the recent conclusion of
the Urguay Round negotiations, regionalism - regional trading arrangements and the other cooperaion possibilities
they open up-remains nascent in South Asa.
Economists are divided on regionalism. Some see it as improving the chances of freer trade
worldwide over time and others, as an obstacle to mialis. But all agree that regionalism has a political
momejtum different from that in the past. And all agree that regionalism today has some reassuring aspects. It is
a part of outward-oriented trade strategies. It tends to involve both developing and developed countries (South-
North and not just South-South ties), and frequently to be open to pottial new members. And in some cases it
is much deeper, addrng not only trade in goods, but also hlealizai  of trade in sevices, movemes of labor
and capital, harmonization of regulatory regimes, and the coorination of domestic uolicies that influence
intenal co        vens.
This study begins the process of explorng the scope for beneficial regionalism in and involvnmg South
Asia. Intmregional trade and the choice of countries in regional anangemnts are important dimensions of the trend
towards regionism  In other regions, the growth of trade was accmpanied by even faster growth of intra regional
tade, prticularly for the Euopen Community, East Asia, and most recently in Latin America In South Asia, by
contra, intraregional trade has been (and remains) veiy modest. In other regions, the scope for South-South and
South-North regionaism is being advanced (Mercosur and NAFTA respectively) or acfively explored (Chile with
NAFTA, increased efforts at cooperation among the 18 couies of the Asia Pacific Economic Coopertion
(APEC)). In South Asia, the South Asan Asocito for Regiona Cooperion (SAARC) is gradually becommg
more active, and has recently agreed on a frmework for promoting prefernt  regioa tradg aragemnts
(SAARC Preferential Tding Arrangement-SAPTA). But the momentum is considerably less than elsewhere. It
is clear that the South Asian coutries, now puring uniateral liberalization wihin GATT's famework of
mulatealim , are on a respectable and defensible path. But the trends in regionalsm elsewhere raise the quesdon:
Would there be any befts to Soth Asia from a more acaive stance in the light of regionalsm 's
spreaad?
An unequivocal answer demands further analysis. Moreover, poltcal econmy considerations
heavily ifluence the scope for regionalism and these differ considerably between South Asia and other regions.
There is no obvious Nordhem economy in a South-Nor.h regional tie-up for South Asia. And the countries of Souh
Asia, despite their strong shared historical heritage, are divided by many differences over borders, religion and other
maters. As a fist step, however, exploring the likely outcomes from economic cooperation in South Asia could
cotibute to the beghinn   of a debate on whether a more active stance on regionalim would benefit South Asia.
This paper reports the findins of a reserch project by Professor T.N. Srinivasan ( Yale University,
Economic Growth Center) on behalf of the Office of the Vice-President, South Asia Region, World Bank and
finaed by the Bank's Researh Commiue (RP0# 678-22). The project uses the available empirical techniques
to explore the scope for icased trade beween South Asian countries, following tariff reductions, and also betwee
them and the EC, the NAFTA counries and an East Asian group (including Japan). Despite the limit of the data
and the available tecniques for simulating tade growth as taiffs are reduced, the study's findings should smulate
fither consideaion of options for regionalism in South Asi  Skeptics will be quick to point to the futlity of
regionalism for countries that have many differences still under resolution. Opimists wil see the possbilities for
economic cooperation as a steppmg stne to better poltc relations.



..
The five main conclusions:
*     Unilateral trade liberalization still promises subantial gains. The contnuation of the
unilate  trade liberazation efforts by each of the Souti AsiMe countries would continue to
yield substanti  benefts. From a strategic perspecive, however, it might be easier to
mcrease its momentum if liberalization is part of a coordinaed effort in all of South Asia,
with liberalized access beig exeded to the rest of the world on a most favored nation
principle.
*     Potential gains In intraregonal South Asian trade are substantial. The low transport costs
between counties in the subcontnent come into play ruce tariff begin to be reduced. The
simulations suggest some orders of magnitude. If all tariffs on intra-South Asi trade are
removed, total trade inces by between 3% of GDP (Indi) and 59% of GDP (Nepal) - and
in between for the other countries. For more modest tariff reductions, say 50%, trade
increases bee 1% of GDPfo.rlnidiaand 9% of GDP for Nepal. The increases are much
larger, in proportional terms, for the smaller counries  (Nepal, Sri Laka, and
Bangladesh, in decling order of gains) than for the larger counties (Pakistan and India,
agai in delining order of benfits).
*     Potea gins in tade for South Asia- across the Big Tree trading blocs - are
greaest wih the EC but......
For the South Asia region as a whole, analysis based on historical data suggest that
a tie-up with the European Comm ty would be more beneficia than with either
NAPTA or with an East Asian group,
*     -8a StgLe tie    with East Asia ma, be e_n more benfia.
Giving more weight to the reoent growdh of trade betwn    South and East Asia and
recognizin that East Asia is the fastest growing region of the world, preiminy
analysis and qualitative judgments suggest that a staegic tie-up with an East Asian
group is likely to be more beneficial. Such a tie-up might be more consistent with
today's realpolitk, becas  the EC continues to seek ways to ac
reionalism with Eastern Europe and the states of the former Soviet Union.
*           Asia mar do better by    a     _ ((one or more) of the Big Three as a aro.
The rive gas to individuat South Asian countries from linking up with the Big
Three vary considerably.
For Sri Lak,  Nepal and Bangaesh the poteti gains are similar across the Big
Three, and indeed are domated by hose in an intra-South Asian grouping. But for
India and Paistan, the historcal domne of the EC and the prospective benefits
of a link with a East Asian group are clearly more beneficial. This suggest that the
subcontinet may do better if i aproaches one (or more) of the Big Tbree as a
group.
We still are far from drawing more than preliminary proposons on the benefits of regionalism in
South Asia. Even so, the findings here sggest hat a hesitant samce by South Asia may be ill-advised. The  are
likely to be import  ganms from reducig tariff and non tariff barrers wihin South Asia, as transport cost
economies come into greater play. There also are likely to be gins from a tie-up with one of the Big Three, the
most beneficial cadidae probably being the East Asian group.



India's and Pakistan's expressions of interest in joining in APEC are useful steps. And building on
the progress with SAPTA, it would be worthwhile for the counries of the region to expand the agenda of
regionalism and to explore other options. biueed, the gains from increased intra-regional trade may well include
non economic benefits: increased economic linkages might well help in reducing political tensions.
The possibility of a tie-up with an East Asian group draws attention to the nk between trade and
foreip direct investment (E:DI), as captured in the 'flying geese hypothesis. The origin of this hypothesis is the
flow of FDI from Japan into other East Asian countries, and from the North-East Asian countries to the South-East
Asian countries. It draws atention to an observed patten in changes in the location of production across countries
over time: 'lead country firms combine their technological and marketing advantages with lower production costs
in "follower" countries, resulting in increasing FDI and trade among them over time. Under the right
encouragement, this may very well be applicable to South Asia. By working together, the counties of South Asia
might propel their ongoing trade reforms in a coordiatd fashion. And using the lowest tariff among the agreeing
countries as the basis for coordiation, they could create more trade and systematically attract FDI to the region.
It is often easier to lock in reforms when the efforts of countries are coordinated than when each country proceeds
on its own.
A concluding observation: the size and fast growth of the South Asian region can be exploited to
generte. faster growth in trade and FDI. The world economy is becoming increasingly integrated economically and
financally. It is also _teating polaization between fast-growing and lagging economies. And it is characzed
by an increase in private capital flows to emerging economies and stagnating public concessional assistance.
Growth of trade and inermational ivestments have an important role to play in South Asia's e3forts to attain
susaable growth and reduce poverty, and open regionalism may create new opportumnies. South Asia cannot
afford to be hesitant.
Gobind T. Nankai
Chief Economist, South Asia Regica
World Bank



TABLE OF CONTENTS
Foeword      .......    .  .  .  .  . .  .  .  ..  .  .  .  .  .  .  .  .   i
1.  Introduction ......                         ;                          1
1.1 Recent Revival of Regionaism .                                   1
1.2  Regional ntegation and South Asia .                             6
2.   Preferential Trading and Trade Liberalization Among
Neighbours: A Model and Illustative Simulations .13
3.   South Asia's Iernatonal Trade: An Econometric Model
and Countrfaal Simulations of the Effects of Trade
Blocs .17
4.   ConclusionsandPoicyplicai   ....................... 24
Annex I                               .31
Rerences  .......................................... 32
List of Tables
Table 1    Direction of Imports by South Asian Countries
1980 and 1990  ................................. 35
Table 2    Direction of Eports from the South Asian Countries
1980and1990 .36
Table 3 Total Trade .37
Table 4    Simulated Effects of PTA - Expced Values  ............ . . 38
TableS    Siulated Effects of PTA (50% tariffreduction) .39
Table 6    Umguay's round's Taiff Reduion: Simulated
Effects .40



REGIONAL TRADING ARRANGEMENT AND BEYOND:
EXPLORING SOME OPTONS FOR SOUTH ASE -
T1HEORY, EMPIRICS AND POLICY
T. N. 9Sriva
1. INTRODUCTION
1.1 Recent Revival of Resionaism
Since the end of the Second World War ther have been a number of regional integion (RQ.A) or
preferal trading aeements (PTA)', the most endn of which have been the European Union (EU) and the
Eropean Free Trade Area  Util the collapse of the Soviet Union and soaist regimes in Eastn Europe, the
Council of Muual Economic Assistance represented a trading bloc of counties other tham market economies. In
the sixes there were several PTAs in Latin America and Afica, includig the Latin Arcan Free Trade Area
of 1960 (a  replaced by the Ltin American bJgration Association in 1980), the Cental American Common
Market (1960), and the East African Common Market of 1961. The Association of South East Asian Naions was
formed in 1967, the Andean Pact in 1969 and the Caribbean Community and Common Market in 1973. Faroutan
(1993) reports sven or eight regional groupings in Sub-Saharam Africa, of which the oldest, South African Customs
Union, dates back to 1910, the rest having been formed in the swventies and ealy eighties. Not all of these were
ful-fledged customs unions or even free trade areas, and even those which were itended to become such
aangements did not in faar do so. Some, such as the East African Common Market, collapsed within a few years
of their formation.
Ineengly, regionalism has become popur  again in the late eighties, even as the Uruguay Round of
muliatera negotiations under the auspices of the GATT were underay. The goals of these negotiations were
furtier liberalization of global trade and extension of GATT discipnes into other areas such as trade in services
'I wish to thk Carlos Braga, Gustavo Canonero, Rmi Coumto, Gobind Nankan and Arvind Panagaiya for
their comments on an earier draft. Commens of participants at semis at the University of Michigan, the
University of Mimesota and the World Bank are gratefly adknowledged.
'See Annex 1 for definitions of various terms de ng economic assocat    among counries.



-2-
and trade-related investme  measurs and intellec    property. The most r-lbrated and recent of the regional
agreemes is the North Ameican Free Trade Areement (NAPTA) between Canada, Mexico and the United States
(USA), signed in 1992 and approved by the legislatures of all thee cow ries by the end of 1993. It superseded
and was preceded by the Canada-U.S. Free Trade Agreement of 1989. Mers, an agreemet between Argenftna.
Brazil, Paraguay and Umuguay to form a common market, was signed in 1991. APEC, the Asia Pacific Ecoomic
Cooperation, fonmed in 1989 held a summit of the leades of its 15 members in November 1993 at Seatle, USA.
It was fonned as a consultative body dealing with such issues as hamonisaion of customs data and excng of
informaion on marine pollution and so on. Although the report of an Eminent Persons Group appointed by APEC
presened at the summit meetmg called for APEC to become an Asian Pacific Economic Community and to take
specific steps towards the establihment of an Asian Free Trade Area, the mems appear to be deeply divided
about the nature and degree of future cooperation. Yet there is a clamour to join APEC-Mexico, Chile and Papua
New Guinea have already applied for membership and India, Pastan. Macao, Sri Lanka, Rusia, Ecuador and Peru
have expressed an inrest in applying (Fin al-irnes, November 15, 1993, page 15).
The increasing popuaity of regionaism in spite of faiures of many such atempts in the past, is attrbuted
by Andeson and Bladihurst (1993) to two events, one on either side of the Atantic Ocean. The first was the
admission of Greece, Portugal and Spain io the European Commumity and the agreement to remove aU intnal
barers on the movemet for goods, services, capital and labour by 1992. The second was the conclusion of the
Canada-U.S. Free Trade Agreemen of 1989. Although the two events were largely indende  of eah other and
had little to do with the failure to concude the Uruay Round by December 1990 as orgialy envisaged,
Anderson and Blaclhurst suggest, plausbly, that the rest of the world saw the two events and the difficulties e'
concluding the Uruguay Romnd negotiations, as indiat  that the future of the multilatal and liberal fmwork
of GATr was uncertain. This led them to seek membesip in the existig or newly established PTAs.
John Whalley (1993, p. 352) also atibutes a key role, among factors explaining the enthusism for PTAs,
to the "search for safe-haven trade agreements by smaller counties who now, more tham ever before, wish to secure
marets of large neighbouring tradig partners becase of the fear of highe trade baiers in the futre. Besides
this fear on the part of smal countries, Whaley sees also a straegic motive on the part of some large counries



-3-
which, by threatening to negotiae, or actually negouasing, regional arrangmenpts may be able to force other
reuctant large powers to make concessons multilaterally wbich they are apparently unwflling to make in the
Uruguay Round negotiatons. Since in recent times growth of trade between neghbouring countries has exceeded
that of global trade, Whaley suggests that regional PTAs are now being viewed more positively as accelerating
further the already dynamic porton of world trade. Peuoni and Whaley (1993) confirm with numerical simulations
from their general equilibrimn trade model of tariff retaliation that recent regional trade agreements are indeed safe-
haven or insurace anmangements sought by smaller countries who have paid prma i the form of non-trade
Oncessions to large countries (e.g. side agreement signed by Mexico with respect to labour and environment as
part of NAFTA mainly to molfy opponents of NAFTA in the U.S.). They find that the value of msch agr
to large countries rises as the risk of global trade conilic' rises since the enhanced risk will naturally lead to larger
pma etactable from smaller counties. Viewed i this light, the recent proliferation of PIAs is a serious threat
to the global trading system.
Baldwin (1993) notes a prima facie puzzling aspect of the reoent enthusiasm for regional trade
liberalization, namely the wlingess of European coutries outside of the EC to open their ariculture, services
and goods markets to EC fims while they were reluctant tc do so in the context of mulital lbeaization under
the Uruguay Round. He offers a ratione for this bdeviour using a political economy model in which an
idiosyncraic shock, such as closer integration of the members of the EC under Europe-199, theates the
fortunes of exports in non-member couies who then lobby for membership in the EC. If the goverments in
some non-member countries were on the margin of indifference between joming EC and staying outside pnor to
the "shockl of Europe 92, they migbt be pushed by domestic export lobby towards joining after the shock. Once
some non-members join, those still outside come mxuer pressure on two counts-from loss of exports to domestic
producers in a better integrated EC and also their marlets in erswhile non-members. This sets off a 'domino'
effect inducing more and more non-members to apply for membship and, in doing so, they become willing to offer
concessions to EC, which they were reluctat to offer in a multi-led contex.
Hallet and Braga (1994) suggest an additioral inducme  for joining a PTA with one's major trang
partners: it is a device for pre-commimen for ensuring proper behaviour of such parmers. They view adhering



-4-
to GAT?T disciplines as "commitment to a coopealive regime In which there is limited advantage in unilateral
libeazation.-but the advantages of multilateral libeaizaion accrue to all only so long as all "play the game"...
But, like any cooperative regime, there is often little sanction against individuals who revert to their best non-
coopeative policies-and quite possible none at all agaist those who form a coalition with market power" (p. 10)
"Moreover, GATI's effectiveness is further complicated by the fact that many trade restictions
which are now commonly used do not prply fa wider the disciplines of the GAT  treaty.
This allows many governments to maintain a reasonably GATT-cosent face to the world while
operatig policy instuaents which have a protectonist impact in practice.
Against this bacWound, it seems clear that countries with their own interests in bloc-wise
noncooperative behavior could easily find ways of doing so without apparently violatig GATT
disciplines (e.g., tbrough ai-dumping actions that disciminate against firts of non-member
countries). Once countries realize that is the case, the most powerful reason for wanting to fotm
free trade blocks within the GAIT system becomes apparent. It is a defensive move, since there
is no other way that counties can credibly commit to behave in a properly cooperative mer
within GAIT. It is therefore better to withdraw into a smaller coalition of your immedie
partners where an explicit agreem   (and the implicit threat of expulsion and loss of market
access in general, rahr of access to particuar maikets) can "lock in" liberal trading poLcies and
market access where it matters." (p. 11)
7hey then ask "Is there anyhing itrsic to Regional Inegration Agrements (RlAs) that make them work better
on pre-commitment devices than a global regime suck as the GATT system" and answer affirmatively by adducing
seeral reasons mainly to do with raising the costs and reducing the benefits from deviaion from cooperation.
Bhagwati, a committed warrior in the cause of a muldlueral process of trade liberalization, in his critical
overview of reonism and   lil        , raises a mber of searching questions as to the compatibility of the
liberal-multilaeral-non-iscrimnatory framework for the world tradig system with discriminatory PTAs
"is the immediate impact of p entl trading blocs...to reduce rather ihan increase world
welfre?
Regardless of the immediate effect, will regionalism lead to non-discrimitomy fee trade
for all, through contied expansion of regional blocs unil universal free trade is reached, or wil
it fragment the world economy? And will, in any event, such a dynamic time-path show that
regionalism will get us closer to the "  of multilateral free trade for all than mulltism as
the ooce  of trade negotiadon will?" (Bhagwati (1993), pp. 31-32, emphasis in the original).
To esure a negative response to his first question, Bbagwai would min*imie the potential world-welfare
reducing immediate effects of tading blocs by replacing the estg Artile XXIV of the GAT! on PTAs by ruling
out aU PTAs other than customs unions (CUs) and insistng that the common extnal tariff of ary CU on any item



-5.-
be Lie Ims tariff of any union member on that item prior to the union. While recognizing the practical problems
of its implementation, Bhagwati also endorses a simple aggregative test proposed by Jobn McMillan (1991), for the
admissibility of a CU or a PTA viz. 'does the bloc rsult in less trade between member countries and outsider
counties'? This test is based on balancing the classic Vinerian benefits of trade creation against the costs of trade
diversion.
On Bhagwati's sec=nd question, there is very little2 by way of analytical and empirical results in the
literature other than the well-known proposition of Kemp and Wan that a common external tariff could be found
that makes at least some residents of a CU better off while not hurting any citizen of non-members, as compared
to the pre-union situation. Thus, the Kemp-Wan proposition shows the existence of a dynamic path towards
multilatral free trade for all through CUs that is mnoticalty mceing in world welfare. But there is no reason
to believe that such a path would in fact be foItowed. Bhagwati finds unpersu=s-;e the argument dwat regionalism
is a quiker, more efficient and more certain process towards multilateral free trade for all. While judging the
recent revival of  egionasm to be unforne, he nonetheless recognizes that only tume will tell whether
regionaism has been a sanguine or a malign force.
The Umguay Round has been completed with the signing of the Final Act in Marcesh in April 1994.
The only uncertin    remaining are about the process of its ratficato by the relevant authorities in the signatory
countries, leaving aside any ambiguities and problems a careful analysis of the Act might reveal. To the extent that
21n his recenty completed doctoral dissertation at Stanord, Pbilip Levy examines whether inceives for
mltilatea trade liberalization are blunted by the possibility of concluding regional trade agreemts in a political
economy model of trade policy detminaon. More precisely, he considers two periods dunng the second of which
an opportuity for multilatera libbalia   arises and asks whether two counies oonduding bilater  trade
agreement in the first period will retain any interest In multilateral liberaiation in the second period. In a standard
Hechscher-Ohlin-Samuelson model with median voter politics, the answer is in the affrmative since the only effect
of trade liberalizaon, bilateral or muiSlatr, is the Stlper-Samueso effect on factor prices induced by terms
of trade changes. However, if the model is of the diffetiated-product-monopolsticcompetition Wpe, there is an
additiona effect to consider, namely, the expansion of the varieties of the product available for consumpdon with
tade lealization. Jn such a model, concluding a bilateral agreement in the first period might result in the two
countr  losing iterest in multilateral l lizai  in the second period. For example, consider two countries
identical in factor endowmens to each other but differlg from those m the rest of the world. Opening up of trade
between them clearly has only a variety epansion effect and would be beneficial. But opeing up of their trade
with the rest of the world bas both effects which could go in opposite directions. If the negative Stolper-amuelson.
effect more than offses the vaiety expansion effect, cleady the two counies could not be intested in expandig
their trade bloc muldlately.



-6-
the enthusiam  for regionalisni refleced unrainty about the conclusion of the Uruguay Round, since this
unceutaiy is no longer there, is it time to pronounce "Regionaism is dead and long live the new World Trade
Organisation!"? A recent paper of the World Bank (1994a) suggests tbat it is premate to pronounce the death of
regionalism. The reason is that
the "new regionaism  is qualitatively different from earlier cycles of regional integration. Most
new RIAs go beyond conventional trade arngements, addreing not only trade in goods, but also
the liberaization of trade in services, movements of labor and capital, the harmonizato of
regulatory regimes, and the codation of domestic policies that influence internaial
competitveness. In this contet, they can be characterized as expeimens in mdeep integration."
The concem with economic intgion rather than trade preferences per se is not new. What
is new is the ation given to reglatory barriers as obstacles to "deep integration" and the focus
on market incenives rather than on centred planning as the driing force behid the integration
process. The advent of North-South arangements based on reciprocity and of South-South RIMs
designed as a complement to outward-looking trade strategies are other hallmr  of the new
regionaism. (p. i)
1.2 Reeional hel ation and South Asia
A mber of studies have appeared since the recent revival of regionaim includig the two volumes of
papers presented at two confeences, one organized by the GATT (Anderson and Blackhuist (1993) and the other
by the World Bank (deMelo and Pagaiya (1993)). A number of researchers inside and outside the World Bank
have also wrte on various aspects of reglism    and PTAs: Bagwell and Staiger (1993a, 1993b), Erzan and Yeats
(1992), Frankel et al. (1993), Frankel and Wei (1993), Krugman (1991), Levy (1993), ludema (1992), Panagarya
(1993), Perroni and Wballey (1993), Safadi and Yeats (1993) and Whalley (1992, 1993). These studies either
address genera issues (e.g. welfare consquences of regiona inegrton, the politica economy of the pursuit of
regional intgrtion and its consequences for the process of mulinaton lbeization, implications of the fornation
of PTAs for multilateral cooperation and so on) or experien  of specific regiona trading arrangements (e.g. Lat
American Free Trade Area, ASEAN and so on).



-7 
While the genaer issues and specfic expedenc  are indeed relevant to our research on South Asia, their
significance is very much tempered by several facts, one of the most important being the Asian context. Although
the trade and growth performance of South Asia in the eighties is superior compared to most Sub-Saharan African
and Latin American eronomies, it pales in comparison to that of East and Southeast Asian economies. It has been
noted widely that the latter are currntly the most dynamic economies in the entire world. South Asian economies
could conceivably benefit more from closer integration with East and Southeast Asia in comparison to a narow
South Asian integration (Section 3 below offers some suggestive simlations to this effect) or even closer integration
with industrialized regional blocs such as the EC or NAFTA. Yet, political support for such a purely Asian bloc
is not strong. Besides, as long as East Asia cones to be as dynamic as it has been without such a bloc, it may
have little interest in joiing an Asian bloc (E. Morimer, Financial Times, Januay 21, 1992, p. 13). In any case,
since histcrically South Asian economies have far more trade with indusriazed countries (icluding Japan) than
with oher Asian ountries (Tables 1 and 2), empirical analysis based on past trade flows is not of much help in
assessing the potential effects of closer integration of South Asian economies with those of East and Southeast Asia.
The economies of South Asia are in many ways quite different from other developing economies (except
possibly ASEAN economies) that have formed in the past or are presently contemplating preferential tading blocs.
Unlike most of them, prior to the 1947 partition of the sub-contnet, Bangladesh, India and Pakistan, being parts
of the same country, constutd  a large common mret. Indeed, they were integrated monetarily as well. These
facts might be potentialy relevant since the trade and financial flows that exsted prior to 1947 could be revived in
a possible South Asian Free Trade Area.
The South Asian countries formed the South Asian Association for Regional Cooperation (SAARC) nealy
a decade ago to streng   regiona reations and promote greater coopation in all sphees. Yet historical tensions
among major member countries (e.g. lidia and Pakstan) bave neither dsipaWd nor become less ite. It is
conceivable that promoting freer movement of goods, services, people and capit in the region might also facilitate
the resolution of political and territorial disputes. The 'peace dividend', i.e. the resources that could be saved from
a reduction of military expendtr  i the region followmg the settlement of regional disputes, is likely to be
substial. But, as with all potentialpeace dividds, wheter the potential willbe transled intoreality is not easy



-8-
to predict.
The literature also does not adequately address at least three imrtant features of economic integraon
among a geographically (and otherwise) large economy and its neighbours, as India and the other South Asian
economies haWen to be. First, the large economy of India could be viewed as a common mtcet of its many
Stawes. Tbis in turn means that in the formation of the market, potential intenational trade with neighbours could
have been diverted to  itjW trade among States nce tansport costs (as a proportion of pnce) on internal trade
could be significantly less than the tariff rates applicable to international trade. Clearly, regional itgration could
tten replace intnal trade in the large country with international trade. For example, with integation, trade flows
bween the souther states (Tamllndu and Kerala, for example) of India and Sri lanka, the eastern se  (Assm,
Bihar and West Bengal) and Bangladesh, northwestern smtes (Hayana, Punjab and Rajasthan) and Pakistan and so
on could exeed signfnly the pre-winteon trade-flows between the relevant pairs of countries. S   trade
among eighbours could easily include (new or expanded) trade in goods and services (e.g. electricity) that are
usually trated as non-traded. For example, among Banglaesh, India and Nepal such trade could be pottially
significant. Trd in the pre-integration sitation with significant trade bamers of differing intensity, iUegal trade
between neighbours uully flourishes. And if the internal bariers of the neighbs differ substantially, then it
is also possible that goods fom the rest of the world enter the low bai  country legaly and then are smuggled
ino the high barrier neighbou. Clearly regio  inration by eliminating trade barriers among neighbours would
subally reduce the icives for illegal trade flows. These three reasons alone suggest that an extension of
the exisng analydcal literature on reginal integration is desirable m consderig such integion of South Asia
Srinvasan and Canonero (1993a) formulate a bare-bones analydcal model intended to capture the fis two effects.
Secdon 2 mmaiz  their findings from  meical simulations using the model.
Even in the absence of regional integration, if nighbouring economies within a region liberalize their
foreign trade as a whole, and in different sectors, at differem speeds, such differcs could have significant impacts
on trade flows (legal and ilegal). P_rst for the same reasons emphasized in the traditonal ltrre on economic
integration, namely scale and scope economies, spill-over externalities and so on, liberaizaton among neighbous
even without integraion could expand markets and ths induce better utilization of existing capacities as weU as



-9 -
create incentives for new investment. Put another way, if tanport and transacion costs associaed with trade with
the rest of the world are substantial, expansion of the regional market following liberalization could be more
important for exploitation of scale economies than iegraton with the global economy.
Łecond, if a high barrier country liberalizes faster (resp. more slowly) than her low barrier neighbour, so
that the difference in the heights of the two banriers comes down (resp. goes up), clearly the incentives to smuggle
imports from the rest of the world into the high barrier couny trough its neighbour would be reduced (resp.
iucred).
ird, relative speeds (particularly if they vary among sectors) of libealization could also affect relative
sizes of potental flows of foreign direct investment (FDI). If. for example, a country with a large domestic market
suc as India maintain a relatively bigh barier by going slowly in its lib ion, compared to Palkstan, a tariff-
jumping type of foreign investm   flow could be enhanced. The reason is that foreign investors could not only
hope to supply the large domestic market of India but also export to Pakista, since by assumption the Pakistan
bars are lower and going down faster as compared to India's. The welare cnseq   of such FDI could be
negative for India and positive for PaksM, if the trade baiers in both conies protect capital-intenisive industries.
Besides the FDI from the rest of the world into the region, intra-regiona FDI (i.e. investment from one of the
counties of the region in another) could also be induced by liberalization per se (regardless of relative speods).
Because such flows are a response to reductions in barriers rather han in ed by the need to 'jump" high barriers
to reach a protected domesdc market, they are likely to be wefae-enhancing for the hosting as well as investing
nations.
An analysis of the effects of unilatera liberalization possibly at diferent speeds and of regional integration
on FDI flows would involve the identif on of sectors with a potential to atract FDI from the rest of the world
as well as from within the region and an analysis of the direct as well as indirect (through, for example,
technological spill-over) growth effects of such imvestment. As is well known, FDI flows to South Asia have not
been substatial historically and are as yet to increase in a major way in response to libezation. PDI flows to
South Asia are muh lowor compared to such flows to Malsia and Thaiand, let alone China. As such, an analysis
based on history and recent past is unlikely to prove lumiating in assessing future flows. While recognizing their



- 10-
importance, no attempt at such an analysis is made bere and is left for a fte occasion.
It is important to quatify also the impacts of the fmatdon of other trading blocs and their possible
expansion on South Asian countries, whether or not they form a bloc of their own (or join with other Asian blocs
such as ASEAN). For example, Raed Safadi and Alexander Yeats (1993) analysed the likely loss to South Asia
of possible trade-diversion resulting from NAFTA and concluded that it would be marginal and concentrated on a
naow range of products includng textiles and clothing. Andson and Snape (1994), note that even though the
share of intra-regiond trade in world trade has been increasing, so has beet the share of extra-regional trade, the
reason beig the increasing share of GDP being traded. They ask whether an enlargement of NAFTA and EU
would slow or reverse this trend. They conclude tIat on balae  the fears of trade and mvestment diversions are
probably exaggerated. But they also see cause for concern tat such expanon might erode the GATT-rules based
multilateral trading system.
The more difficult task is to anticipate the impact on a potential South Asian Bloc (whose stucture could
range all the way from some loose prefeential trading aangement to ful economic integraton) of other actual and
potentia blocs with yet to be decided suc . An attempt has been made (Srinivasan and Canonero (1993b)) to
assess such nipacts, however imprecise they might be, using a gravity-type quaative model explaining bilateral
trade flows among South Asian countries and the rest of the world, grouped into relevant countries and regions.
Section 3 smmaris the results of this exacise.
Sine the research for this paper was completed early in the fall of 1993, the Final Act embodying the
results of the Uruguy Round has been signed. If it is ratified by the signatories by mid-1995 as expected, a new
organisation called the World Trade Orgaiation will come into operation replacig the GATr and the various parts
of the Act will be implemented over a ten perod from thn. TIhe Final Act is 550 pages long and the text,
partcularly the passages in legal terms, is not exacy a transparent document. The parts that are of primary ilterest
from the perective of some or aU of South Asian countries are those that relae to textiles and clothig,
agriculture, services, trade-related intellectual property rights and investment measures, of which textiles and
clothing is the most significant in a quattative sens in South Asia's trade.
The quotas on textiles and cloing imports, currently negotiated bilterly under the Multifibre



Arrgement (MFA), will be eliminaed i three stages over a perod of ten years from January 1, 1995. During
each stage, annual growth in the import quotas on those products that are stll under restraint will be increased by
no less than 16% over MFA 1994 in Stage 1, by no less than 25% over Stage 1 in Stage 2, and finally by no less
than 27% over Stage 2 in Stage 3. Although a  aster phase-out of MFA would have been better, stll substantial
oppornmities for increasing tetle and clothing exports from South Asia emerge. However, since the competition
from otber exporters will be intense, unless South As=an exportes are competitve in quality and cost, they might
not be able to maintain, let alone increase, their share in the growing eport maiets.
With respect to agriculture, the two major decisions are tariffication, i.e. conversion of exng restrictions
into tariffs and reducton of such trfs by 36% (resp. 24%) in the case of developed (resp. develoing) countries,
the reduction being achieved over a period of six (resp. ten) years in the case of developed (resp. developing)
countries. Also, domesdc support measures (other than those wbich have a minimal impact on foreign trade) and
export subsidies are to be reduced as well. In principle, the existing quantitative restictions in some South Asian
countries (e.g. Idia) on agricultura trade and domestic support measures (e.g. subsidies on fertisers, irrigation,
credit etc.) could attract these provisions. Yet given the "special treatment clause that allows, under certain
conditions, a country to maintain import restrictions up to the end of the ten year implenttion perod and the
"special and differenta  tratment applicable to developing countries, the Final Act does not appear to call for
many changes in exstg agricultural trade and subsidy polices in Souh Asia, though it does not rule out such
changes being brought by the countries themselves in thei own inte. The reduction of export subsidies in
indusrilied counies could conceivably raise the cost of agrculual imports, particularly food, into South Asia,
but available stdies suggest that these are unikely to be qunatively significant.
It is likely that implementing the Final Act's provisions regrding intellectual property nghts might have
some adverse consequences i the short nm, for example by raising the price of some essent  drugs. Also,
exendiag patent protection to seeds and plant varieies could also have similar adverse effects. But existing
knowledge precludes any firm estimation of the possible rise in prices, and indeed establishing that long-term
bnefts, if any, would outweigh short-rnm costs.
The completion of the Uruguay Round does not mean that the exsting pref a trading aangements



- 12 -
are not important or that the trading world will not end up in four likely blocs, viz the enlarged EC, an extended
NAFTA including South American countries, an East-South-East Asian bloc and the rest. But it certaiety means
that for the time being the threat of trade wars breaking out has diminished significanly. Nevertheless it is essential
that South Asia does not become part of a fourth group of margmalzed counies. From a political economy
perspective one has to analyze the prospect of South Asian countries: (i) becoming associates of the EC (as some
of the African countries are at present); (ii) negotiating a free trade agreement with NAFTA or alternatively with
USA alone; (iii) joining an East-South-East Asian group or even the Asian Pacific Nations for Economic
Cooperation group that is formmg. The model discussed in Section 3 sheds some light on the quantitative impact
of some of the choices.
Briefly stated, the model suggests that a naw   South Asin PTA, while it benefits the smaller counries
of the region to a modest ement, is unlikely to yield substantial benefits to the laer counries. A larger Asian PTA
is likely to benefit all. In any case, even withoutjoining a narrow or wider PTA, if each of the countries coninue
and extd their unieral liberalization currently underway, they are likey to gain significantly more. However,
fom a political economy perspective, coordinated liberalization by all comtries of the region together (with the
iberalized access to each market being extded to extra-regional trading partners on a MFN basis) is likely to find
greater political support in each country. Such coordination serves to blunt the fear among producers in a couty
that unilaterally liberalizes faster than its neighbours that they will lose in their home matret to producers in
neighbouring c   em wihout at the same time  _g  greater access to makets of those countries.
Even if one ignored the polical benefits and focused only on economic ones, coordiated iberliation
extended to all tradig partes on a MFN basis is lily to be the best policy. In fact, the World Bank (1994b)
reports on a oDmparism (using a general equilbrium model) of four policy options of trade lberalization in East
Asia. The four options were: the formation of a small group that is open (i.e. a modest extension of unilateral
libaization), the formation of a large group that is open (the analogue of coordinad liberalization euended on
a MFN basis to all others), a small group that is diny  (the analogue of a narrow South Asian PTA), a
Iarge group that is i m        (the analogue of an Asian PTA). In the East Asian conte, the second of the
above options was foumd to be by far the best. The results of Section 3 suggests that this conclusion is likely to



- 13 -
be valid in the South Asian context as well.
Section 4 concludes the paper by pulling togetber the reslts of Sections 2 and 3 and attempts to answer
three basic questions: (1) Whether, and If so, what kind of regional trading armgement is appropriate for South
Asia; (it) Does the fact that one of the liberalizing eoonomies, viz. India, is much larger and more diversified than
others in the region have any further implications with respect to the desirability of regional preferential trding
arrangement that included India and to the pace of their liberalization relative to India's in the absence of such an
armgement, and (iii) are there any regional trade policy implications flowing from the answers to the first two
questions?
2.                      TRADING    TRADE T   RALUMON AMONG NEIGHBOURS: A
MQODEL AND ILLUSTRATJVE SMLATh)NS                           !
It was poined out earlier that in the context of trade liberalization among neighbours the inteaction
between  olicv detrmined barriers (e.g. tariffs, quotas and other nontariff bariers) and natural barriers (e.g.
transport costs, linguistic and institutional differences) to trade could be important. In particular, Lbiiation by
a large country of its emenal trade with its neighbours on all sides, could mean substitution of internal trade among
distant regions withn the country by external trade of each region with a foreig neighbour closer to it than another
region. Also, goods and services not traded among neighbours prior to lbalization could be traded after
significant hlazaton. In South Asi, these considerations could be of some simptance sine it icludes a
geographicaly lrge and diversified country, viz. India, which already trades (or could potentially trade) with
neighbours in East (Bangladesh), North West (Pakdstan) and South (Sri Lana).
Snivasan and Canonro (1993a) illustrate the interaction between transport costs and tariff in the context
of trade liberalization among neighbours using two simple (bare-bones) models, one focusing on trnsport cost
wihin a large country and the other on non-raded goods. In the first model, there are three countries and a rest
of the world (ROW). One of the counties has two regions, each of the regions being a neighbor of one of the other
two countries and distant from the other region and country. Transportion costs between neighbors are assumed



- 14 -
to be zero and those between distant points ae of the Samuesonian 'melting iceberg' type. lTus, only a faction
of a unit of a commodity leaving an origin arrives at a distant destination. Each country or region is endowed with
an exogenously specified amount of 'its' commodity, but it consumes an aggregate made up of its commodity and
the commodities of all other countries and regions, including the ROW. Thus there is no production in this model
and as such, trade policy affects only conumption and welfare. The amount used of each commodity in making
up a unit of the aggre    is detrmined by minimizing costs. In the initial situation, each country has a tariff on
its exnal trade. Preferential tade liberalization consists of the removal of utiffs on trade among the three
countries while leaving the tariffs on the rest of the world intact. Of course, unilateral but non-preferntil trade
liberalization could be viewed as reducing the eal tariff of each country, possibly m different proportions.
Under the assumption that each country and region is a price taker in markets of the ROW, the model is
first solved numerically for a base case competitive trading equilibrium. Each region of the large country is
specified to be larger than its neighbors and the distance between each regional country and the ROW is twice as
muich as the distamce between one counry in the region and another. Preference parameter values and tariff rates
are chosen suc that the initial trade among regional counries is small. Besides, a unique ad-valorem tariff rate
is set for each country, whh tariff rates being equal for all regional countries, while ROW is assumed completely
open to internatonal tade. Then the changes in welfare and trade structure ex enced by initially proteced
econOmies after joining a prefrential trade angement are simulated and the sensivity of the results to
exog aousy imposed parameters are tested.
The second model, borrows the Ricarian teuchnology with a contimmm of goods from Dobbuwch et al.
(1977) and extends it to have three  uries and two diffentated sets of goods. Two of the countries are from
the sme geographical region and therefore transport costs (again, of the Samuelsonian 'melting iceberg' tpe)
between them are lower than with respect to the third country. The regiona counties do not produce the same set
of goods while the third country is able to produce any good from both sets. Given labor endowmen, one solves
for a competitive trading equilrum, and the associated relative wages, total expenditr for each county, as well
as the pattern of production, in paticular the traded and non-traded ranges of goods. The umercal exercises done
with the first model are repeated for the second, thus simulatg the effects on welfare, patterns of production, and



shares in production of traded and non-traded goods, once a preferentl trade arragement is introduced.
The simulations (see Srinivasm and Canonero (1993a) for details) with the first model show that, Ars, a
unilateral move to free trade with aU  artners leads to much laraer welfare Iains comnared to free trade only with
retional partners. Thus, for countries which are price takers in world markets PTAs with some parmers is inferior
to free trade with aU partners. Second, any county's choice of the extent of its unilateral and uniform reduction
in tariff does not harm the other member's situations. This kind of indedice is a generic chaactric of the
model. It comes from the assumption, appropriate in this context, that the regional economies are price-takers in
international makets.   ,  the welfare improving effect of substiuting inmal trade that is subject to real and
unremovable transport costs by international trade with neighbors, oce the policy-rated barier of tariffs is
removed, is eviden. Thus. the interaction between tariffs and trans-ort costs is imwortant. Fourth, when transport
os  withi the region are much lower tan those between the region and ROW, significant gains from preferential
trade lberii on occur. iM   and finaly, there are no gains to non-reciprocity in regional trade liberlization,
i.e. there is no gain from not matcbing the reduction in tariff on one's exports offered by one's neighbors. Thus,
there is no advanage to having a slower rate of lbelization compared to one's neighbours.
The simulation resu  from Model 2 can be summaized as follows: f= as in the case of model 1, the
trade effects from PTA are quantively more significant than welfare effects, again for the same rean, namely
the feasibility of substution   consumption among goods whose prices are affected by the awrangement. S
in conast with the fir model in which all goods are in atnaly taded, in the second, the trade effects also
include the change in status of some goods from being non-taded to being traded, not only between counties of
the region, but also between them and the rest of the world, once the arangement comes into force. And these
effects are found to be significant. Thid, as in the first model, the welfae effects of unilatera lberalization of all
trade are higher than those of prfental libeization of trade within the region. Fo_uh, initl tariff levels and
the distance of the counties of the region from the rest of the world matter, again as in the first model: high tariffs
and a large distce preclude significant effects on reginal trade with the introduction of PTA. Ff. even in this
simple model, a small country is more open than a large one, and the reduction in non-traded goods following the
PTA is higher the smaller the country. As is to be exected, when two of the regional counries increase in size



- 16-
relative to the third, the range of goods not traded with the rest of the world increases.
The simulations from the two models, wlde adequa  to demonsrate the trade and welfare effects of
prefereial trade lbaliatio, have to be viewed with caution for several reasons. Frst and most obvious, the
results naturally are specific to the partiwlar models. However, given the models, the sensitivity of the results to
variations in the numerical values of its parameters have been analyzed. Second, the models are static: as
mentioned in the introuction, there could be important and ieesdAg dynamic effects aising from changes in the
returns to domestic and foreign investment as well as technological transfers consequen to liberaiation, preferenaal
or otherwise. _Mrd, in the first model there is no productio and in the second there are no scale economies or
diseconomies in production and pure compeition rules in both. To the extt there are significant scale economies
that get exploited and maket power of firms get redued with trade liberation, such effects are not catred m
the models. However, it is arguable that except possibly in iioaly non-traded infrastucture scale economies
are unlikely to be important in most other goods.
Used with appropiae caution and awareness of their above-mentioned limitations, the two models do have
a few policy implications for South Asian counries. Given the substantil ransport costs associated with internal
trade within the large onries of the region, and also with trade between each of the countries of the region and
their major trading parners outide the region, preferential liberalization of trade within the region could lead to
a significant increase in intra-regional trade that is beneficial to all pamers. However, even larger gains could be
achbieved by lberalizbg trade with all pares. The second model also sggests that removal of trade bariers
wib the region (prefernly or on most favored nation basis) could generate trade in goods and services, the
demand for which is currently bemg  ntely met by domestic supply in each country. Smce some of these would
be ifra-structural services (e.g. power) whose production involves scale economies (not modeled) and whose limited
and erratic supply inhbits efficient prodctiou of other goods, a more rational investmt, production and trade
pattern could lead not only to a once-and-for-all improvement in efficiency of production of all goods and services
but also a more rapid growth. The growth enhancing effect could be significant: the prospect of supplying to a
larger market foilownng trade liberalization attrats direct foi  investm  into infst uct  sectors.



-17-
3.   SOUTH  ASIA'S INIE RNATIONAL  RODE: AN ECONOMETRIC  MOD3L AND
COUNTERFACTUAL SIMULATIONS OF TBE EF FE  OF TRADE BLOCS
Srinivasan and Canonero (1993b) broaden the line of research initiated by Safadi and Yeats (1993) who
analyzed the likely impact of the formation of the North American Free Trade Area on South Asia by considering
other potential preferental tading arrangements. In particular, they evaluate the effects of South Asia Regional
Inigration (among Bangladeh, India, Nepal, Pakistan and Sri Lanka), as well as preferential trade arrangements
(PTAs) of mdividual South Asian counties with some non-regional partners like USA, NAFTA as a whole, EEC
or an Asiam Group.
The tool used for empmical analysis is an exded version of the simple and well-known Gravity Model
of Biateral Trade, such as the one used by Prankel et al. (1993). The model estmated is speified in equation (1)
Leg BT2IC'4+d Log (GNP *Gj)+d2   g L      .GNPPC*GNPPC)                   (1)
+d3'D,+   Log T,5+aLog Tj+i6 Log RERj+uj.
where BTuj is bilateral trade betwee countes-or group of counties-i and j in the year t and the supscript c
indicates the commodity group. GNP and GNPPC are the respective Gross National Product and per capita National
Prodt of both parmers in year t, Dj is the distance in kilometers between the rlevant cities, r73 is (one plus)
the tariff imposed by country i in its tade on commodity c with j and REXR is an index of real excage rate
between the two trading paes. BT, GNP, GNPPC are measued in units of thousand 1987 US dollars, using
the GDP defator for the US to deflate nominal values.
The uilization of the panel featues of the data set required special trtment of the enror term in (1). A
general expression for uij, could be the following:
4j, e fj + p4 + r{,                               (2)
here, e and 9 represent the individual contry effects, j a temporal effect and ; a purely random effect. If e and



- 18 -
5 were fixed effects, thenby using ounty dummies the problem is easily solved. Onthe other hand, if these effects
were random, the best  gy to esdmate equation (1) would be by using Generalized Least Squares, taking ino
account the contribution of the variance of country effects to the vaiance of 4  In this case, a commo
problem that aries is the possible conrlation between the individual country effects and the explanatory variables.
A piminary esmation of (1) using country dummies dmnstated that the assumption of fixed effects is
inappropriate: the residuals from this estmation still showed the presence of Individual effects. Therefore equaton
(1) was reesdmated assuming random effects, using the lags of GNP and GNPPC as inuments and tested for their
exogeneity.
The raw data on trade are from the United Nations' COMTRADE databa  (at the Inernational Computing
Center in Geneva), where imports and exports are reported in cunt USA dollars for each country and all its
tading pas. The data are available amnually and by commodity using SlTC classification (Revisions 1 and 2).
With this basic information the bilateral trade (BT = imports + exports) on 10 commodities for 21 trading
partners was calculated. The ten composite commodities ae: Coffee and Tea, Textile Fibres, Leather and Dressed
Fur., Textle Yam and Clothing, where South Asia concenuae its exports, and other aggregates such as Non-Fuel
Prmies, Fuels, Machinery and Transport Eqpmet, Other Manfactur  and Total Trade. On the other hand,
13 individual countries were selected (the five from South Asia and the  major partners: USA, Canada Mexico,
Japm, Hong Kong, Korea, Singapore and China) and 8 goups (EEC 6, EFTA, Africa, Asia, Middle East, Lati
America, OECD and European counties not considered elsewhere). Both commodity and country aggregation were
meamt to enhance the most import   atures of the ac    trade patn of South Asia Countries.
In general the variable Da was calculated as the distce between the major city (usually, but not always,
the capital) of country i and that of j. When i (or j) represened a group of countries, Do represetd the average
of the disances of two major cities (or the city at the geographic centre) of i to the major city in j. In the case of
ndia, the distmce of the closest among its tee major ports (Bombay, Calcutta and Madras) to the major city of
its trading partner was used as the relevant distance. Tiis was done to cpre the trade-off between internal trade
and exmer  trade with her neighbours.



- 19 -
Turing to the tariff vaiable TV, the figures for South Asian countries are the implicit effective tariffs
calculated from tariff revenues. These are average tariffs and include import and export duties. The data on tariffs
imposed by developed countries and Mexico come from the UNCTAD-World Bank Software for Market Analysis
and Restrictions on Trade (SMART), where they are available desaggregated by commodity. The tariff stucmre
is the reported one for Japan and Mexico (again effective tariff) but not for Canada. EEC and USA. Although the
actual tariffs imposed by the latter are not completely similar, there probably are compensating differences in their
NTB coverage. It was therefore decided to assign the sane tariff structure as rep-esented by the average of their
repoted tariffs.
One final remark should be made about the tariffs. Although tariff are considered to affect all traders
uniformly, nevertheless trade not involving South Asian counties is treated as unaffected by tariffs. In other words,
TL and TJ are set at unity whenever neiher i nor j are Bangladesh, India, Nepal, Pakdst or Sri Laka Basically
t-is procedure implies that the  _magtes of such tariffs, as compared to those assocated with South Asia
countries, to be small enough to be negligible.
Finally the variable REXR aims to account for significant changes in real exchge rate in the South Asian
counties. REXR is an index with base 1965 = 1 for the real exchange rate of the domesdc currencies to the USA
dollar. REXR is calculated using the nomial ehange rate and GDP defators from IMF's Interational _inanci.-I
Sics (exhange rate is market value, period average -series rf-). It is the source also for the GNP and
Popuation figures. Thus, REXR is set at 1 if both iandj are non South Ain. If either i or jare from South Asia
and not both, REXRU is that REXR of the South Asian comnty. In those obsevatios where both i and j are from
South Asia, REXR0 is an average of the two REXRs.
The technicalities of estimaton and detailed results are prented in Srinivasan and Canonro (1993b).
Briefly, anmual data from 1968-1991 were used, generaing 5175 observaion bilateral-trade flows among 218 pais
(13 countries and 8 groups) of trading partners. The stmated equation for total trade is given in Table 3. It shows
the remarkable explanatory power of the gravity model despite its weak grounding in economic theory. All
xariables have the expected signs and are significant at 5% level. The bigger the trade partners, the more significant
is the bilateral trade, the greater the distance between the trading coies and the higher the tariff baiers the



-20-
smaller Is the value of bilateral trade. Only the coefficients of GNPC, the product of GNP per capita of the two
patners, have ambiguous signs (i.e. positive in some and negive in otier equations. See Srinivasan and Canonero
(1993b) for details). They are statistically signfict in some but not al cases.
These findings are not inconsistent with received theory under which a higher GNP per capita indicates a
higher degree of development, and therefore more specialization, which in tum is more important for manufacred
goods. That both significant negative coefficients relate to trade In primay goods is consistent with this hypothesis.
'ecause the dependent vaiable is the sum of imports and exports, the effect of the rea exchnge rate on
it is ambiguous in theory because its effec is expected to be positive on exo u and negative on imports (actually
ttese are the e3xcted signs in the loig mn, while on an annual basis trade-exchange rate relationship for some
c odies may not always follow the long-nm diredon, principally because of contual rigidities). However,
the real change rate in fact showed positive corration with bilateral trade in every commodity, although this
stastic  relatio is not significant for the primary goods. Obviously, the explanations for the signs and sizes of
the estmated coefficients have to be found In the likely influence of the real exchage rate on the demand and
pradion of the relvant commodities. Be that as it may, the significance of real exchange rate on total trade for
non-prmary goods indicates that empirically this effect is positive.
Equipped with this set of results the efect of prefial aangemems involving South Asian counies
was simulated. An uwD bound for this effect is to conder the total elimation of tariff baniers between those
couries or grwup of counties for which such an arra%,zam is contemplated. Inerdiate values of the effbct
could be obtaed by partally reducing the barriers without completely dminatg them.
n inteing the smulation results, it is important to keep in mind that the simulations are driven by the
estimated coefficients x4 and % for the tariff variables T1 and T2 respectively. Basically the impact of preferentil
trade arangemerts (PTA) is meaued by the pr tonate change in the USA doLar vahle of trade they create.
The higher the iniial tariff level on trade between pares, tim greater the final effect of such arrangements since,
by definition, they liminmae the tariffs. One should, therefo, expect the impact to be increasing in the estimsed
valu of c4 and as. However, the tariff is only one among many elements that determine the impact of PTA on
tade. In assg the impact of pfentl tariff reducos, two other features of the model have to be kept in



- 21 -
mind.
First, the series T, mainly represent tariffs imposew  y Canada. EEC, Japan, Mexico and USA on imports
from South Asian countries, while series T2 are tariffs imposed by the latter on trade with the former group. Since
tariffs T2 are initially higher than TX, the higher the coefficient of T2, in absolute values, the greater the impact of
preferetial arrangement.
Second since a4 and %5 in equation (1) are elasticities indicating the proportionate response of bilateal
trade to changes in tariffs, the initia tariff level as well as the itia trade level are relevant for daeminng the
absolute changes in trade following a PTA.
Once the forces driving the simulations are undestood, their limitations also become apparent. F,,
possible terms of trade effects associated with the creation of trade are not taken into accout. As such, the
simulated results almost ceray overestimate the tue impact. On the other hand, the static fmework of the
execise does not consider many important dynamic aspects of trade liberalization and these could reinforce the
short-run trade creation thus u    img the true long-run impact. The long-rum effect is very difficult to
forecast without a more general dynamic setup that genrates terms of trade effects and allows for scale economies,
miveSme, and spillovers of tecbnology etc. Ihe basic information to quantify aU these is not available: for
example, some price dasticities could be approximated but mformation on scale economies simply does not exist.
However, experience with the few studies that tfied to estimate dynamic effects in trade, indicate that short-inm
effects are likely to overestimate the tre long-run effects. &ond, the measure of trade barriers, viz, an indicator
for average tariffs, does not fully capture the effect of many non-tariff barriers to trade. 1dinl even if the
simuions conrecy measure the impact on trade creation, it should be realized that this impact is not the only
factor to take into account in evaluating PTAs. For instance, the negative effect on bilateral trade with countries
not enteing the arrangement is not assessed at all in the simulations. Moreover, none of the indicator from the
simuations could be viewed as a welfare measure, thus making the comparis   of difrent scenarios rather
incomplete. In other words, the results serve only the limited purpose providing an esfimate of the potenti effects
on bilateral tade between each South Asian country and its parner in the simulated PTA.
A summary indicaor of all the simuations with respect to complete eliinaon of tariffs on all



.22 -
commodidies by partner countries ar presented in Table 4 relating to total trade. Table 5 presents the results of
a 50% reduction in tariffs along the sme lines. Table 6 conepons to the simulated effect of the estimated 25%
reduction of tariffs of all commodities by all cotraies. Tables 5 and 6 also include East Asia (Hong Kong, Japan,
Korea and Singapore) as a region.
The numerical values in Table 4 of the expected increase in trade with complete liberalition ona
arefectil basis are very large, in fact too large, to be totally credible. However, these should not be viewed as
forecasts but only as indicatig the broad orders of magnitde of potentW increases.3 It is clear that the greatest
rovortional increase expected in bilateral trade would come from regional integration. The counties in the region
have all higher tariffs than other counries and therefore a larger impact from reduction in their tariff should be
expected. For example the trade between Bangladesh and the other South Asian countries is expected to increase
by 9.5 times and almost the same is the case of Pakistan (8.9). For Sri Lanka the exped proportional increase
is about 10.3 ts, for India 12.8 times, while for Nepal this increase is even higher, at 17.2 times. However,
given the initial trade patten of these counmes, regional inwgration leads to a gater increase in trade for
angladesh and Nepal than for the other South Asian countries.
For the two smal economies of Bangladesh and Nepal, regional integration seems to be the most powerful
preferential arangement for trade creation. The values of expected incrases in trade are incredible. Bagladesh's
new trade with the region would account for US $ 4.6 billion, ceeding its actual atade of $ 3.8 billion by
17%, and accounting for a whopping 21.1% of its ONPI
In the case of Nepal the trade epected to be creaed in the region is around US $ 1.7 billion. Though this
is maller than that for Bangladesh, it is not less impressive consdeing the (economic) size of Nepal. The new trade
would be almost three times the actual total trade of Nepal and 58.5% of its GNP!.
The effects of regional intration on the large economies inthe region, viz. India or Pakistan, are namrally
very different from those of Bangldesh and Nepal. For India and Pakistan regional  tegration is important, but
3Using a prat equilibrium model (i.e. assuming that terms of trade and incomes do not dcange as liblization
takes place) and values of trade elasucities found in the litertatue,. Canonero (1994) estimated the expected net trade
expansion following the formation of a South Asian and other PTAs. These estimates, as is to be expected, are
considerably smaler thanl those in Table 2.



-23 -
their much largea trade with the Eurea Communtie and USA, maies integration with the latr more attcive.
Both countries would achieve the greatest impact on their tade by integrating their economes with EEC, for india
bilateral trade would increase by 2 times its actual total trade, while for Pak    the coespondmg figure is around
0.95. Translating these effects into US dollars, new blater trade between India and EEC would amount to US $
85 billion, reprenting 30% of India's GNP! For Pakis  these figures would be US $ 13 billion and 30%.
The simulations in Table S for a 50% reduction of initial tariffs, again on a preferential basis as compared
to the conresponding figures in Table 4, indicate the expansion of trade is considerably less than half that from
complete, i.e. 100%, reduction. This is, of course, to be expected since in the grvity model bilateral tade flow
is a convex function of the tariff rate so that the mnal expanson from a reduction in tariffs is an increasing
finction of the reduction. Also, reducing tars by 50% on trade with South Asian partners exands trade to a
lesser (resp. greater) exten than the same reduction of tariffs on trade with East Asia in the case of India, Patn
and Sri lanka (resp. Bangladesh (marginally) and Nepal). Thus the advanages of South Asian regional integration
for smaler economies of the region are seen once agin. In the case of Bangladesh, Palisan and Sri Lanka, the
expxet  epansion of trade with East Asia emeeds that from a reduction in 50% of tarffs on trade with USA.
Given that East Asian economies are the most dynamic in the world, thes results are informative.
Turnng now to Table 6, the results represent trade exansion following a non-referenia 25% reducion
in tariffs on all commodities by all countries. The figare of 25% was chosen to represemt in an aggregate sense the
exected reduction in various tariff and no-taiff bari once the Uruguay Rounin agwement is fully enfoed.
Since the reduction covers the entire trading world, a Rest-of-the World group is added to the groups of Tables 2
and 3. It is clear that even a mods t   by 25% in tariffs, given that it is universal, results in a significant
expansion of trade. For example, for India, total trade with USA and Canada increases by 9.4% of its 1990-91
value, with East Asia by 8.2% and the rest of the world by 20.5%. On the oher hand, for the land-locked and
smallest economy, viz. Nepal, the gain in trade volume is lagest (14.39%) with the Rest of the World, trade with
its neighbours in South Asia being a close second at 12.9%-trade with EEC and East Asia expands by approximate
equal percentages of 10.5% and 10.7 respectively.



-24-
4. CONCLUSIONS AND POLICY DMIlICTIONS
Research on the prospects of regional integradon in South Asia was intated by the South Asian Region
of the World Bank (World Bank, 1993). After reviewing the received theory and empirical evidence, this study
idenified the following four p-condiions that significantly enhance the probability of successfhl prefernal trade
arrangemens in the fora of a free trade area (PTA) among a group of countries.
"First, the gMFTA tariffs should be bha. Second, the members of the FTA should be Wmor
tadi  Ras before enterng into an arrangement. Third, there should be coplenenrity in
dand. For the South Asian counies this by and large means entering an arrngement vwth
countries that have different economic strucures. Fourth, the diffences in economic sure
should be based on the 'true' _2ixdm     of the counties involved. This mean that
aangements with counties that have ubstantally different factor endowments are superior to
those with similr endowmes' (World Bank (1993), p. 16, emphasis in original).
The study concluded tat only the first condition was met in the case of a possible South Asian Free Trade Area.
Clearly, even after recent libalizati, tariff and non-tariff bawriess to extral tade conue to be very
high in South Asia relative to the average rates of protection in the rest of the developing world. Tables S and 6,
repnrduced from World Bank (1993) show that, except Nepal which exportc to (resp. imports from) Idia a large
but declinig share of its exports (resp. imports), South Asian counties trade little with each other but with
industralized countries particularly in North America and Europe. lheir export composition is similar, with an
ovewhein  and inesig share accounted for by labour inteive    m        . Although the composifti  of
in uoasdifered somewhat, sdtll it iS laly the case that trade betwee South Asian countries is cmpetitive than
complementay, although given the eae of dtiortions n al markets in these eonomies, it is hard to infer 'true'
competiiveness fotm the observed trade paens. While viewing the prospects of a South Asian Free Trade Area
as dim, World Bank (1993) suggested that "it may be worthwhile to explore the oosbility of exiandine trade with
the rest of Asia.. If South Asia could secure access to EC by formn an PTA. it would clearly benefit them' (p.
20 emphasis in original). While offering no direct evidence in terms of data on factor endowments (human capital,
physical capital, land and natual resources) the report nevertheless concluded that
In fact, the South Asian counries have vety similar factor ndowments and their grest
resource is undoubtedly their large labor forces. Tle scope for tade based on 'true'
competitiveness wthin South Asia is therefore not large. But while the differences between South



- 25 -
Asia and the industrial countries is clearly the largest. there is much scope for trade with the other
Asian develonina countries based on difeces in factor endoment." (ibid, p. 24, emphasis
in original)
The presen paper and the two companion papers (Srinivasan and Canonero (1993a, 1993b) confirm in some
importat respects and modify in others the conclusions reached in World Bank (1993). Given their exisig trade
pattems (Tables 5 and 6), for large economies such as lidia and Pakistan, while regional integraion yields some
gains, the principal gains seem to come from PTAs with the EC and USA. On the other hand, regional integration
leads to greater gains in the value of their trade for the smal economies of Bangladesh and Nepal. There are
significant differences in the impact of PTAs on trade in different commodity groups. For all South Asian countries
regional integraion would clearly create more trade on textile fibres than any other preferenial trade arrangement.
In this commodity the tarff reductions would be significant and the atal trade within the region is strong enough
to foster likely future trade. On the other hand, the actual trade pattern indicates that trade in clothing could receive
an extraordinary impulse if South Asian countries integrate with EEC or USA, with the increased trade beig similar
for the large and small ecoaomies in the region.
It would appear that a substantial part of the trade gains in a PrA with EC and USA arise from the
eimination of their barriers to imports of textiles and apparel from South Asia. This is in consonance with the
findings of Safadi and Yates (1993) that the potential for a NAFTA-induced diversion of exports is mamily
concentrated in textiles and clthing, since NAPTA keeps the barrers against imports from South Asia while
elmiatig them with respect to trade among members. As they corectly note whether Mexico's accession to
NAFTA will induce a major displacement of South Asia's exports to the US wil depnd to a great extent on how
compeitive Mexico could be and to what extent the 'nules of origin' restrction of NAFTA wil affect Mexican
exports of clothing. Be that as it may, as MFA is phased out as part of a successfW Uruguay Round agement,
whetier South Asia is able to expand its exports will depend on its comptitiveness relative to other supplies in East
Asia  Safadi and Yates estimate substa     gams to South Asia from a scess  Uruguay Round.
World Bank (1993), Safadi and Yates (1993) and Srinivas  and Canonero (1993b) all address only the
trade impacts of ib n and notwelareimpacts. As iswellknown(Bbagwati (1971,1973), Kirman (1973)),
joinig a customs union, even it is trade-diverting, cIod be welfare improving for a country. As such, welfare



-26 -
effects should be dstmgushed from effects on trade flows. While Srinivas  and Canonero (1993a) do address
welfare effect, they do so with simuladons from an illustrtive model and not with an econometric model of South
Asian trade. Yet their conclusions that uniateral liberaliztion of all trade are higher than those of preferential trade
liberaliation within the region is suggestive. If it is, as is very lkely, supported by an empirical welfare analysis
of Souih Asian beralization altematively, not only would it be important to support and exnd on-goig unilateral
trade liberalization in each of the South Asian countries but also to create a favourable global environment for liberal
trade by successfully concluding the UruguW Round.
Tuming now to some policy issues, the conclusions reached in World Bank (1993) are by and large
confirmed by the present study where the two studies overlap. These are: Ast, the gains from a narrowly conceived
South Asian PTA are unlikely to be substantial for the larger conries of the rion while the smaller economies
could gain from greater integration with tbeir neighbours. S, the gains from a larger Asian PTA could be
sgificant for South Asia. 1hir, while unilatera llization by each of the South Asian coumties, at whatever
speed and sectoral composition, would yield substnti beefs to each, from a political pepective it might be
easier to sell such hiberalization as a pat of a coordid ibe ti  in al of South Asia. with the liberalized
access then being extended to the rest of the world on a most favoured nation principle. This idea, proposed by
Panagriya (1993), is worth pursuing. Results presen  in Table 4 could be interpreted as an approximation tothe
effects of such a liberalization, because it envisages liberalization by all counties rather than just by South Asian
counies. Yet tbis modest liberalization expands trade within South Asia by as high as 12.9% and 8.3%
respectively for Nepal and Bangbdesh. For the others it is much more modest. Itis likely that if only South Asian
countries liberalized in a coordiated and multilateal way and to a greater extent, then ntra regional trade will
exand even more. Fourth, the gains from liberized access to induslhzed couaies through arrangemens
rangig from association along the lines of LOME conventon between EC and some African counties to joing
NAFTA and other free trade areas as members should be pursued.
In this comection the following reco      made in the report of Bhagwati and Sriivasan (1993) on
India's Economic Reforms are apropriate for other South Asiam conties as well.
lFbally, if we are to move to an outward-oriented strategy, seeking to exloit trade and



-27 -
investmet opportnities provided by the world economy, te we must take appropriate action
to ee that these opportunities are available to us maimaly. Othewinse, we would be
operaing with one blade of the scissor and ignoring the other.
Towards this end, we recommend t1at bidia play a constcive role in midn the
Uruguay Round successful since the alteive is the strong assertng their might as with the
Super 301 and Special 301 actions of the United States against other nations (including us) and
the breakdown of the multilatr trade discipline that, despite its weakss, provides something
like a nile of law benefiting us.
More importan, we recommend that India now actively start exploring the options of
joining in the free-trade blocs that are in place (EC and NAFTA) and which might emerge (in
Asia). Else, it stands in danger of loSing out to her countries, members of such blocs, in trade
access and in atact  investment.' (Bhagwati and Srinivasan (1993), p. vi.)
Their remaks regarding an Asian trading arrangement are perdnent as well:
*The possibility that we should explore with the greatest energy, however, is with Asia At the
moment, there is no Asian block, and many in the West thnk that one may not materiase. On
the other hand, it is hard to imagine anyone taking seriously the formation of the US-Canada free
tade area over a decade ago, and its extension to Mexico would have been regaded as altogether
utopian. The speed with which old assumptos are shifted by new possibilities and realities is
truly darlt.
As it happens, the sense that the world trading system is being already 'carved up' into
blocs by the EC and the United States is already leading Maaysia to sponsor such an Asian bloc.
Tbe anti-American rhetoric in which such a proposal tends to emanate from Mlaysia handicaps
the acceptance of the idea by Japan. For, Japan has natrally no desire to make the Asian bloc
an offensive rather than a defensive move: Japan cannot afford to, and sensibly will not wish to,
alienate the United States.
But the sentiment for an Asian bloc is now stronger and growing. India must beome
an acdve proponent of the idea, seeking membeship as and when the idea mmiais. This will
need patient diplomacy since the curme t position is that fndia is not perceived as a 'natural
member of such an Asian bloc and our membership even of the APEC, which is only a looser
body of Asian and Pacific nations for economic coopa, is still peiding.
We wi need to woo the ASEAN nations and we wi also need to work actively on Japan
itself, using both economic and political carrots to do so. Essential to our success will be
commitment to an Asian identity (which need not excde multiple affilations and identies).
Without this commitme, our membesi may run into difficulty just as Bria's less-than-total
enthusiasm for enty into the EC contributed to repeated French vetoes on the Britsh applicadon
for membehip.
Only by getng ready diplomatically for exploring these policy options, and pursuing
them urgendy, can we expect to safeguard our ecoomic inteests in the evolving economy. Our
thiking and policies have to be reshaped to suit the rapiy changig world economic world
economic scene.
The exploration of these tradg posiblities and choices woud also make more credible
our commitment to the refoms initated since 1991, both at home and abroad. In tum, that would
benefit the reform process itsef, creating a virtuous circle where reforms lead to intimate
engagemt into the evolving trade regimes and that  pggement in tmn encourages foreign



-28-
investmt and imterest in India's economic and reinforces the success of the reforms in providing
benefits.* (Bhagwat and Srinivasm (1993), p. 69.)
It is absolutely clear from the debates in the U.S. Congress on NAFTA, it will not be easy to dispel
misperceptions about the nature of such arrangements and the fear that US and other industrialized couies will
be hurt in terms of manufatring employment, wages, labour and environmental standards by entering into such
arrangements. These fears, unfounded they might be in fact, bave been exloited by protectionist interests tO
demand a linkage betwee market access and adhce to their conception of labour and envi l standards.
Hence while it might be in the interest of South Asia to seek membship of or association with NAFTA and EC,
such an expansion of membeship might not be politically saleable in the legislatures of the member countries of
the latter. ITe most ratical stra   for SoAth Asian countries would be to move on all  nts min order: first
unilaeraly liberalize trade, preferably coornating such a step with eah ote, second, vigorously push for the
radfication and implemation of the Final Act of Uruguay Round while firmly resisting attempts to link market
access to adhece to labour and evionmental sadrds and, _   and finaly, explore posibilities of mmbeshp
in APEC, EC and NAPTA.
Coordinated liberalization within South Asia together with a preferenl trade arrangement (PTA) with an
industrialized co     group such as the EC or NAFTA could be politially useful in two ways: first it makes it
easier to lock-In the libelization in each of the couies in that any reversal will jeopardie the PTA. Second,
as noted earlier, not only regional trade will exand in response to coordinated liberaization but, by the same tokea,
in the absence of coordiaon in a county liber g its economy faster than its neighbours, political opposition
to liberalization could arse, if its imports from neighbours rnse faster      its exports to them. For example, the
study of World  stue  for Development Economics Research (WIDER (1993)) on Indo-Sri Lanka Economic
Cooperation draws aention to the prevailing percepdons in Sri T Ln  that *theme is a lack of a 'level playing field'
in trade between the two counries, that India was enjoying a 'free ride' in the liberazed Sri I  ka*  maket, and
that mutual trade relations could not be built up when one side was obtaining all the benefts...Currently, Sri Lana
is a reavely open mairet for Indian goods but the reverse does not apply..." (p. 16). While recognizing that
ujilje liberalization on an MFN basis by both counries would be in the best intereStS of both countries, the



-29-
report expresses some scepticism about India reducing her import baries on an MPN basis as rapidly and
exensively as Sri Lhka as done, primarily becase of domestic prsure from import competmg industies. Given
these perceptions, the report advocates a policy of reciprocal trade preferene between india and Sri  ka.
However, it is more plausible to argue that the domestic constraints In India are better counteed by the prospect
of profitable opportunities opening up with coordinaed liberalization among all her South Asian neighbours than
by a liberal argement with Sri T Lnka
To conclude, the analysis in this paer suggests that the most preferred option for South Asia is to l1balize
their exunal trade in a coordinated mnner and exteding the liberalized market access to the rest of the trdng
world on a MFN basis. Such a move will not only take advantage of lower trasport costs on  nra-reaional as
compared to       i     -ade, but also rationalize regional investment in infr cture. As is well-known, in
South Asia the inadequacy and inefficient functioning of publicly owned and operated 8afrastructurat activites have
m the past constaned growth and, unless vastly improved, will in the ftue constrain the response to economic
refDrMs in other sectors includig foreign trade. As such, improvement in their availability and functioning as a
consequence of coordinaed lialization would have significat efficiency enhancing effects on the whole economy
in each of the countries.
The sgested coordinated liberaliation should be viewed as an integral part of extending and deepening
the ongoig economic refom process in South Asia. As a part of this process several sectors of the economy,
inuding ftose such as Ifrasucture that were prevously closed tO pnvate investment and operation, have been
opened up to domestc and foreg prive ivestment. Indeed in some couties of the region (e.g. India) this has
resulted in a boom in foreign portfolio investment in the domestic market for equities and in domestic firms
borrowing foreign capital markets. Contrac   agreements for some foreign direct invesutnat (PDI) into power
generation and other infrastructure bave also been recently concluded. Yet, in contrast to flows of ootfolio
investment, E-1 flows, while much larger in volume compared to the pre-form  period, are still relatively smal
compared to what other economies in East and South East Asia have bem able to atract. Portfolio investments by
foreign investors are a mixed blessing: on the one hand they augment domestic resources, but on the other they tend
to cause an undesirable appreciation of the excnge rate if imports do not go up rapidly enough andlor if



-30 -                          .
complemenraiy policy reforms are not accelered. FDI flows are less volatile and in part create their own demand
for imports. Although the analysis in this paper has been confined to the effect of alenative forms of tade
liberalization including PTAs, on trade flows, it is clear that from the perspective of accelerating growth, and hence
of achieving the overarhing objective of alleviation of massive poverty in the region, increasing investment, private
and public, domestic and foreign, in physical and humnm capital is vital. The World Bank's (1994b) proposal in
the context of East Asia for *a concerted program for li zation of FDI policies ... and for mobilization of PDI
from other iarts of the retion and the world' (p. Si, emphasis added) is relevant for South Asia as well. Designing
such a programme for FDI would be facilitated by cooperation in designing and implementing a programme of
oordinated trade liberalization.
The importance of maitinng an open and lberal global trang system for the success of economic
reforms and twade liberaiaion in South Asia can hardly be ov rasized. The successful conclusion of the
Uruguay Round is certainly a major step in that direction. In this conext, cooperation among South Asia counties,
firs in matters of trade  liberizion, and then with respect to FDI, could pave the way for a much broader and
equally, if not more, important polical agreement to reduce tensions and hosilities in the region, thereby releasing
significant resources for development and poverty alleviation. Ideed, one could go furter: expanded trade and
financial flows among couies would create a constituency in each country having a via mterest in m g
tension- free politcal relationshi   the regin Tbus, close economic rlations and cooperation will help to -lock-
in- peaceful political relations.



- 31 -. 
Definiton of terms commonly used in describing economic associaton among countries (from the loosest to the
closest).
Economic Cooneration Aermnts (ECA): Agremets betweep two or more countries to coopate in specific
areas of common economic interest, e.g. agreements to avoid double taxation of individuals and enterris
that engage in aconomic activties m more than one country that subscribe to the agreement.
Prefrntial Tradine AnWit (PTA): An ageement to extend prefrential access to each other's market in
saecfied amods and non-factor services, the preferenti terms, by definiion not being extded to non-
members.
Free Trade Area (FTA): An agreement to remove all bariers to trade in all goods and non-factor services among
members, with each member being free to set its own tarffs on trade with non-members.
oms Union: Same as PTA but with members levying            rates of tarift on trade with non-members.
Common Market (CM): Same as CU but with *ee movement of factors among members.
Ecoomic Union (EU): Same as CM but with al1 economic policies being hamonized among members.
Rewiona lnearaton Ament: Any one of the above but with members coming from a well-dened geographical
reg11.



- 32 -
Anderson, Kym and Richard Blackhurst, editors (1993), Regional Inteeration and the Global Trating Svstem, New
Wvrk: Harvester/Wheatsheaf.
Anderson, Kym and Richard Snape (1994), "European and American Regionaism: Effects On and Options For
Asia," Discussion Paper 983, London: Centre for Economic Policy Research.
Bagwell, Kyle and Robert Staiger (1993a), "Multilateral Tariff Cooperation During the Formation of Regional Free
Trade Areas," NBER Working Paper No. 4364.
(1993b), "Multilateral Tariff Cooperation During the Formation of Customs Unions." Par preseed at the
Universities Research Conference on bIernational Trade Rules and Institutions, December 4, 1993, NBER,
Cambridge, MA.
Baldwin, Richard (1993), 'A Domino Theory of Regionalism," Centre for Ecoomic Policy Research, Discussion
Paper No. 857, processed.
Bbagwati, Jagdish (1993), "Regionalism and Milalism: An Overview," in Jaime De Melo and Arvind
Panagariya, editors, New Dimensions in Resional Intearation, Cambridge: University Press.
(1971), "Trade-Diverting Customs Unions and Welfare-Improvement: A Clarification," Economic Joumal
81, Sepember, 580-587.
(1973), "A Reply to Professor Kirman," conomic Journal 83, Sepmber, 895-897.
Bbagwati, Jagdish and T. N. Sriivasan (1993), IdiW's Economic Reforms, New Delhi: Ministry of Finance.
Canonero, Gustavo (1994), "Preferetial Trade Amagements: Partal Equlibrium Estmates for South Asian
Countries," (World Bank (processed).
deMelo, Jaime and Arvind Panagariya (1993), "Introdction," in Jaime deMelo and Arvind Panagriya, editors,
New Dimersios in Resilond Ingation Cambridge: University Press.
Dornbusch, R., S. Fisdher and P. Samnelson (197), "Comparative Advantage, Trade and Paymes in aRicardiar
Model with a Contimnmmof Goods," American Economic Review, December 1977, 823-839.
Erzan, Refik and Alexander Yeats (1992), "ree Trade Agreements with the United States: What's In It for Latin



- 33 -
America?," World Bank: Policy Research Working Paper.
Faroutan, Faezeh (1993), 'Regional Integration in Sub-Saharan Africa: Past Experience and Future Prospects," in
Jaime deMelo and Arvind Panagariya, editors, New Dimensions in Re,ional Integration, Cambridge:
University Press.
Frankel, Jeffrey, Emesto Stein and Shang-jin Wei (1993), 'Coninental Trading Blocs: Are They Natural, or Super-
Natural?" procsed).
Frankel, Jeffrey and Shang-Jin Wei (1993), -Trade Blocs and Cumneny Blocs," NBER Working Paper No. 4335.
Hallet, Andrew and C. A. Primo Braga (1994), 'The New Regionalism and the Threat of Protectionism,"
processed.
Kirman, A. P. (1973), 'Trade Diverdng Customs Unions: A Comment," Economic Journal, 83 (331), 890-893.
Krugman, Paul (1991), 'The Move to Free Trade Zones," Federal Reserve Bank of Kansas City,
Economic Review 76(6), November-December, pp. 5-25.
Levy, Philip (1993), "A Political-Economic Analysis of Free Trade Agreements," Stanford University (processed).
Ludema, Rodney (1992), "On the Value of Prferential Trade Agemems in Multilateral Negotiations" (processed).
McMilan, John (1991), "Do Trade Blocs Foster Open Trade?", Univesity of Califormia at San Diego (processed).
Panagarya, Arvind (1993), "Should East Asia Go Regioal? No and Maybe," World Bank (processed).
Peroni, Carlo and Jobn Whalley (1993), "The New Regionalism: Trade LUbadization or Insurance?" Paper
presened at the Univesties Research Conferen  on  tnational Trade Rules and Instiions, December
4, 1993, NBER, Cambridge, MA.
Safadi, Raed and Alexander Yeats (1993), "The North Amer!c  Free Trade Agreement: Its Effect on South Asia,"
World Bank: Policy Research Woring Paper.
Srinivasm, T. N. and Gustavo Canonero (1993a), "Liberalization of Trade Among Neigbhbors: Two lusative
Models and Simlations", South Asia Region Discussion Paper Series, Supplement II to IDP# 142.
_ Prefereatial Trade Airangements: Estmatg the Effects on South Asian Counties", South Asia Region
Discssion Paper Series, Supplement m to IDP# 142.
World Bank (1993), "Regionalsm and South Asia's Trade," South Asia Region Discussion Pmcr Series,



-34 -
Supplement I to IDP# 142, World Basn, Washingto, D.C.
World Bank (1994a), "Consequences of Trading Blocs,* processed.
World Bank (1994b), "Building on the Uruguay Round: East Asian Leadership in Liberalization,' Discussion Paper.
Wblaey, John (1993), 'Regiona Trade Arrangemes in North America: CUSTA and NAFTA,' in Jaime deMelo
and Arvind Panagriya, editors, New Dimensions in Reaional Integaon, Cambridge: University Press.
Whalley, Jobn (1992), 'Expanding the North American Free Trade Agreement,' Institute for Policy Reform
Worldng Paper No. 42.
World nsitute for Developmet Economics Research (WIDER) (1993), Indo-Sri Lanka Economc Cooratoon:
Facilitatng Trade Exansion tbrouh a Re cal Prefere   Scheme, Study Group Series No. 9, Helsinki,
Finland: United Nadons UniversityM/WIIER.



- 35 -
TABLE 1 DIRECTION OF IMPORTS BY SOUTH ASIN COUNTRIES. 10 AND 1990.
(Percentage share of totl merhadise  mports)
.t..                                                           ;*INDIAKRTA  |.NGLADESH  S. .A.  NEPAL'..
____.__   .__ ;_ 19i8 .|  19¶980 _t 199|0 _   i' 1980                      *199      ' 1980       1990       198   11990
India        _0                                    1          5          6          4          4        46         31
Pakistan            a          0                                3          3          1          2         0          0
amngladesh           O          0          1          1                                0          0         0          1
Sri Lanka            0          O          1         10                    0                          __0             0
Nepal              0    0          0          0               0          0          0          0
South Asia           0          0          2         2          8           8          5          7       46         33
ASEAN                3          3         4          5          4          8,         3          9          2         2
Asian NIEs           2          6          6          8          9         25         Is         19        15         18
Japan              5          8        14         12         13         14         14         12         22       13
USA              11         11         10        13         13          10         7          8          4         1
Canada             2          1.         2          1         4           4          1                    0         0
EC12             is8        2          24        22          18         19         16         15         8   _   14
CPE               9          6         1        7            9          6          3          6          1         8
,~~ -                                            -                                -          -                 . .._.
Other             49         37        37         30         22          7         34         23          2        11
Wortd _ _01                            100       100        100         100       100        100        100       1001
Note: Data after 1990 is not available yet for mm of the countries.
Source: LU CtOTRADŁ Data Base.



- 36 -
TABL_E2 DIRECTION OF EXPORTS FROM TEE SOUTH ASIAN COUNTRES. 1980 AND 1990.
(Percenage share of ota  me dmfe expors)
India                                     I ;  1    1          1           1         3          1        53        25
Paksta               0          0             .    .           10          4    -    3          2         3          0
I         2          3         2         .                      O          1      4            0
Sri Lanka            I          I         1                      1        0         .                    0          0
NIl                                                                                   0         0 0    a   _   o   o   o
Sout Asia            3         3                                           5         7          3        60      2_
AStIli               2          3         2         3           1          0          1         1         1         0
As_ian mi_s          0          0         9        13           2          6         4          4         4          1
jaomn               10          9        11         8          3          6          3          5         3         1
USA- -              11         is         9         12         12        36          11        26         3        29
canda                1          I         1         2          1          2           1         2         0          1
EC12                23         28        27        32         128                    2         26        24        33
CPE                 2          18   !     °         4         1-3          4          9         4         0          1
Other               28         23        35        22         40           13        42        2          6         10
-. : -  -                                                           -  -    a  .  --
Vorld              100        100       100       100         100        100       1O0        100   |  100        100
Souwr: UN  TRADE Data ease



- 37 -
Table 3. TOTAL TRADz
Panel of Annual Data   1968-1991                                W#Obs. 5175
INsTRUMsTAL VAR LIES - 30 PAN   FEATURES
Variable                      Coeff       Std Error       T-Stat     Signif
0.72735026   0.01715866      42.38968  0.00000000
GNPPC                       0.33445465   0.02252389      14.84888  0.00000000
D                          -0.80508512   0.04660275    -17.27549  0.00000000
TI                         -5.74306686   0.41535882    -13.82676  0.00000000
T2                         -5.45346591   0.37926731    -14.37895  0.00000000
REXR                        0.78663776   0.09355628       8.40818  0.00000000
Constant                  -21.99658331   0.82354954    -26.70948  0.00000000
R**2- 0.54                     R**2 Adj.- 0.54                      DW-0.74
ANALYSIS OF VARUGNCE:
Source      Sum of Squares Degree.  Mean Square   F-Statistic  Sig.Level
INDIV       15259.147996383      217    70.318654361        24.399 0.00000000
ERROR       14286.355410242    4957      2.882056770
TOTAL       29545.503406625    5174
INSTRUMENTAL VARIABLES - RANDOM EFFECTS
Variable                      Coeff       Std Error       T-Stat     Signif
!,m==mminmumm._
GNP                         0.82954569   0.05252544      15.79322  0.00000000
GNPPC                       0.01079169   0.09054588       0.11918  0.90513363
D                          -1.08263825   0.16999122      -6.36879  0.00000000
Ti                         -3.90443785   1.55071295      -2.51783  0.01183788
T2                        -4.65837217   1.19281557       -3.90536  0.00009529
REXR                        0.87676670   0.07449218      11.76992  0.00000000
Constant                 -19.15797828   2.49129123       -7.68998  0.00000000
R**2-  0.007                   R**2 Adj.- 0.007                     DW-1.23
WALD TEST
F(7)m 2159 Significance Level 0.000
J



- 38 -
Table 4,.   SMHULATED EFFECTS OF PTA - Expected Values -
Total Trade
Bangladesh    India     Nepal  Pakistan   Sri Lanka
POTENTIAL PARTNERS
USA                      3.3        8.2         6.1        4.4        3.8
2,496.8   46,140.9      436.2    7,619.3    2,155.6
65.4%     116.8%      69.3%       54.8%      51.9%
11.4%      16.4%      15.0%       17.8%      34.2%
Canada                   3.3        8.2         6.1        4.4        3.8
352.4   3,803.6        20.1       714.8      194.6
9.2%       9.6%       3.2%        5.1%       3.5%
1.6%       1.3%       0.7%        1.7%       2.3%
Mexico                   3.3         8.2        6.1         4.4        3.8
6.1      569.5        1.5        26.2       79.1
0.2%       1.4%       0.2%        0.2%       1.4%
0.0%       0.2%       0.1%        0.1%       0.9%
EEC                      3.3        8.2         6.1        4.4        3.8
2,973.8   85,682.5      755.4   13,109.1    3,532.7
77.9%     216.8%     120.1%       94.2%      64.2%
13.5%      30.4%      26.0%       30.7%      42.3%
Japan                     1.8        5.0        3.7         2.6        2.2
682.1   18,869.1      293.9    4,548.5       943.8
17.9%     47.7%       46.7%       32.7%      17.1%
3.1%       6.7%      10.1%       10.6%      11.3%
SAS                      9.5       12.8        17.2         8.9       10.3
4,642.0   8,528.9    1,696.6    3,002.8    2,998.1
121.7%      21.6%     269.7%       21.6%      54.5%
21.1%       3.0%      58.5%        7.0%      35.9%
NOTE: The information on each cell is the following:
row 1: Proportional Increase of Bilateral Trade.
row 2: Value of Such Increase in Millions USA $.
row 3: Ratio of Row 2 over Total Trade.
row 4: Ratio of Row 2 over GNP.
*Bilateral Trade, Total Trade and GNP are average of 1990-91 values,
except for Nepal where 1989-90 values were used.



- 39 -
Table S   SZMUL&TED EFFECTS OF PTA (50% tariff reduction)
- Expected Values - Total Trade
Bangladesh    India     Nepal    Pakistan   Sri Lanka
POTENTIA  PARTNERS
USA                    1.0        1.7        1.5         1.2        1.1
745.4    9,623.4      103.4    2,047.6       809.1
19.5%      24.4%      16.4%       14.7%      14.7%
3.4%       3.4%       3.6t        4.8t       9.7t
Canada                 1.0        1.7        1.5         1.2        1.1
105.2      793.3        4.8       192.1       55.2
2.8%       2.0%       0.8t        1.4%       1.0%
0.5%       0.3%       0.2%        0.4%       0.7%
Mexico                 1.0        1.7        1.5         1.2        1.1
1.8      118.8        0.4         7.0       22.4
0.0%       0.3%       0.1%        0.1%       0.4%
0.0%       0.0%       0.0%        0.0%       0.3%
BEC                    1.0        1.7        1.5         1.2        1.1
887.8   17,870.3      179.0    3,S22.9    1,001.0
23.3t      45.2t      28.5%       25.3%      18.2%
4.0%       6.3%       6.2%        8.2%      12.0%
Japan                  0.6        1.2        1.0         0.8        0.7
23S.5    4,604.4       81.1    1,418.0       309.7
6.2t      11.7%      12.9%       10.2*       5.6%
1.1%       1.6%        2.8%       3.3%       3.7%
Korea                  1.2        2.0        1.7         1.5        1.3
348.6    1,711.3       22.2       736.4      276.8
9.1%       4.3%       3.S%        5.3%       5.0%
1.6%       0.6%       0.8%        1.7t       3.3%
Hong Rong              0.6        1.1        0.9         0.7        0.6
67.2      806.8        7.4       277.8       69.6
1.8%       2.0%       1.2%        2.0%       1.3t
0.3%       0.3%       0.3t        0.7%       0.8*
Singapore              0.6        1.1        0.9         0.7        0.6
241.1    1,631.2       62.4       267.6      171.6
6.3%       4.1%       9.9%        1.9%       3.1%
1.1%       0.6%       2.1%        0.6%       2.1%
East Asia           892.5    8,753.7       173.0    2,699.8       827.7
23.4%      22.2%      27.5%       19.4%      15.0%
4.1%       3.1%       6.0O        6.3%       9.9%
SAS                    1.8        2.2        2.6         1.8        1.9
903.7    1,456.1      253.4       599.9      564.0
23.7%       3.7%      40.3%        4.3%      10.2%
4.1%       0.5S       8.7%        1.4%       6.8%
NOTE: The information on each cell is the following:
row 1: Proportional Increase of Bilateral Trade.
row 2: Value of Such Increase in Millions USA $.
row 3: Ratio of Row 2 over Total Trade.
row 4: Ratio of Row 2 over GMP.
*Bilateral Trade, Total Trade and GNP are average of 1990-91 values,
except for Nepal where 1989-90 values were used.



- 40 -
Table 6 URUGW Y ROMND'S TARIFF REDUCTIONs SIWMLATED EFFECTS
- Bxpected Values - Total Trade
Bangladesh    India      Nepal    Pakistan   Sri Lanka
POTENTIAL PARTNERS
USA                    0.4         0.6        0.5         0.5         0.4
299.0    3,421.3        38.3       791.1      318.5
7.8%       8.7%        6.1%       5.7%        5.8%
1.4%       1.2%        1.3%       1.9%        3.8t
Canada                 0.4         0.6        0.5         0.5         0.4
42.2      282.0         1.8       74.2        21.7
1.1%       0.7%        0.3%       0.5%        0.4%
0.2%       0.1%        0.1%       0.2%        0.3%
Mexico                 0.4         0.6        0.5         0.5         0.4
0.7       42.2         0.1        2.7         8.8
0.0%       0.1%        0.0%       O.Ot        0.2%
O.0t       0.0%        0.0%       0.0%        0.1%
EEC                    0.4         0.6        0.5         0.5         0.4
356.2    6,353.3        66.3    1,361.1       394.0
9.3%      16.1%       10.5%       9.8%        7.2%
1.6%       2.3%        2.3%       3.2%        4.7%
Japan                  0.3         0.5        0.4         0.3         0.3
100.2    1,736.5        31.9      580.9       129.3
2.6%       4.4%        5.1%       4.2%        2.3%
0.5%       0.6%        1.1%       1.4%        1.5%
Korea                  0.5         0.7        0.6         0.5         0.5
134.3      584.7         7.9      273.4       104.7
3.5%       1.5%        1.3%       2.0t        1.9%
0.6%       0.2%        0.3%       0.6%        1.3%
Hong Kong              0.2         0.4        0.4         0.3         0.3
28.9      307.2         2.9      114.8        29.3
0.8t       0.8%        0.5%       0.8%        0.5%
0.1%       0.1%        0.1%       0.3%        0.4%
Singapore              0.2         0.4        0.4         0.3         0.3
103.5      621.1        24.7      110.6        72.3
2.7%       1.6%        3.9%       0.8%        1.3%
0.5%       0.2%        0.9%       0.3%        0.9*
East Asia            367.0    3,249.5        67.4    1,079.8       335.6
9.6%       8.2%       10.7%       7.8%        6.1%
1.7%       1.2%        2.3*       2.5%        4.0%
SAS                    0.6         0.7        0.8         0.6         0.7
314.8      485.9        81.1      211.0       194.2
8.3%       1.2%       12.9%       1.5t        3.5%
1.4%       0.2%        2.8%       0.5%        2.3*
Rest of World          0.2         0.4        0.4         0.3         0.3
282.0    8,087.8        89.9    2,100.3       826.0
7.4%      20.5%       14.3%      15.1%       15.0%
1.3*       2.9*        3.1%       4.9%        9.9t
NOTE: The informption on each cell is the following:
row 1: Proportional Increase of Bilateral Trade.
row 2: Value of Such Increase in Millions USA $.
row 3: Ratio of Row 2 over Total Trade.
row 4: Ratio of Row 2 over GNP.
*Bilateral Trade, Total Trade and GNP are average of 1990-91 values,
except for Nepal where 1989-90 values were used.