<%BANNER%>

The EU-ACP Economic Partnership Agreements (EPA) for Malawi

Permanent Link: http://ufdc.ufl.edu/UFE0042275/00001

Material Information

Title: The EU-ACP Economic Partnership Agreements (EPA) for Malawi Trade Effects and Fiscal Impacts for Alternative Policy Options
Physical Description: 1 online resource (85 p.)
Language: english
Creator: Thindwa, Innocent
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: acp, agreement, cotonou, creation, diversion, epa, eu, fiscal, liberalization, malawi, preferential, trade
Food and Resource Economics -- Dissertations, Academic -- UF
Genre: Food and Resource Economics thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: THE EU-ACP ECONOMIC PARTNERSHIP AGREEMENTS (EPA) FOR MALAWI: TRADE EFFECTS AND FISCAL IMPACTS FOR ALTERNATIVE POLICY OPTIONS Name: Innocent Lwafyo Thindwa Mobile and Email: +265-(0)888-366-263 and ithindwa@ufl.edu Department: Food and Resources Economics Chair: John J. VanSickle Degree: Master of Science Graduation Date: August 2010 This study undertook to quantify the potential trade flow and tariff revenue effects for Malawi if the country liberalizes its trade along the Economic Partnership Agreements (EPA) framework with the EU, pursuant to the June 2000 Cotonou Agreement. Study results demonstrate that other than restricting trade liberalization to the EU which is not beneficial, Malawi stands to benefit from a broader liberalization framework that includes other trading partners. The results also provide useful insight to trade negotiators and policy makers into the product categories that would be most hit by the policy change. In particular, changes on trade flows and revenue reductions in the three key agricultural products of tea, sugar and tobacco were found to be insignificant. This finding does not support the long standing fear among ACP countries that an EPA with the EU is bound to result into an influx of cheap EU agricultural imports.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Innocent Thindwa.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Vansickle, John J.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042275:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042275/00001

Material Information

Title: The EU-ACP Economic Partnership Agreements (EPA) for Malawi Trade Effects and Fiscal Impacts for Alternative Policy Options
Physical Description: 1 online resource (85 p.)
Language: english
Creator: Thindwa, Innocent
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: acp, agreement, cotonou, creation, diversion, epa, eu, fiscal, liberalization, malawi, preferential, trade
Food and Resource Economics -- Dissertations, Academic -- UF
Genre: Food and Resource Economics thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: THE EU-ACP ECONOMIC PARTNERSHIP AGREEMENTS (EPA) FOR MALAWI: TRADE EFFECTS AND FISCAL IMPACTS FOR ALTERNATIVE POLICY OPTIONS Name: Innocent Lwafyo Thindwa Mobile and Email: +265-(0)888-366-263 and ithindwa@ufl.edu Department: Food and Resources Economics Chair: John J. VanSickle Degree: Master of Science Graduation Date: August 2010 This study undertook to quantify the potential trade flow and tariff revenue effects for Malawi if the country liberalizes its trade along the Economic Partnership Agreements (EPA) framework with the EU, pursuant to the June 2000 Cotonou Agreement. Study results demonstrate that other than restricting trade liberalization to the EU which is not beneficial, Malawi stands to benefit from a broader liberalization framework that includes other trading partners. The results also provide useful insight to trade negotiators and policy makers into the product categories that would be most hit by the policy change. In particular, changes on trade flows and revenue reductions in the three key agricultural products of tea, sugar and tobacco were found to be insignificant. This finding does not support the long standing fear among ACP countries that an EPA with the EU is bound to result into an influx of cheap EU agricultural imports.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Innocent Thindwa.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Vansickle, John J.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042275:00001


This item has the following downloads:


Full Text





THE EU-ACP ECONOMIC PARTNERSHIP AGREEMENTS (EPA) FOR MALAWI: TRADE
EFFECTS AND FISCAL IMPACTS FOR ALTERNATIVE POLICY OPTIONS




















By

INNOCENT LWAFYO THINDWA


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2010


































2010 Innocent L. Thindwa































To my late mother, I dedicate this piece of work with great humility as I reflect on her
untiring efforts to see me educated even in the face of glaring challenges within an apparently
harsh environment as she ploughed on to see me through the journey that not even herself knew
the destination. How much I long to have had her alive this day to witness the fruits of her
diligence and commitment to a cause noble enough as I bare testimony today. May Her Soul
Rest In Eternal Peace!!









ACKNOWLEDGEMENTS

I benefited from the contributions of many people in the course of my studies at the

University of Florida and throughout this thesis project. Foremost among these is my major

professor and academic advisor, Dr. John J. VanSickle, who from the very start of my studies

provided me with invaluable moral and academic support to the level I never envisaged. Dr.

VanSickle was particularly instrumental in providing me with the required momentum to finish

the research project when I was back in Malawi and fraught with several family responsibilities

that might otherwise have compromised my completion of the project. He physically visited me

in March, 2010 to fulfill this noble cause, affording me constructive advisory comments on the

written drafts and handy techniques of handling the project data. I do really appreciate his

contribution, without which this thesis would not have been possible.

I also wish to commend Dr. Rick Weldon, who demonstrated to be a very useful

committee member of my project. I am also grateful to Dr. James Stems, who without much ado

spared time to sit in my defense far in Malawi.

Dr. Walter Bowen dedicated himself as a very compassionate and helpful administrator

of the USAID program that provided financial support for my studies at the University of

Florida. Throughout my studies, Walter remained ready and committed to listen and offer

solutions to any challenges that those of us supported by the program could bring to his attention.

The financial support provided by the USAID is, to say the least, greatly appreciated.

I am also indebted to my Malawian colleagues; Fiskani, Lucy, Pearson, Bonnet, Jonathan

and Davie who we shared part of our stay together at the University of Florida. The moral and

academic support I got from Fiskani, Lucy, Pearson and Bonnet during the course work is

greatly appreciated.









I also thank GOD for his continual guidance and for having maintained my incredibly

loving fiancee Alice Mbisa with patience and the spirit of continued moral support to me

throughout my studies.









TABLE OF CONTENTS

page

A C K N O W L E D G E M E N T S ......... ................. ............................................................................4

L IST O F T A B L E S ................................................................................................... 8

LIST O F A B B R EV IA TIO N S ......... ................. .................................................................. 9

ABSTRAC T ................................................... ............... 12

CHAPTER

1 INTRODUCTION ............... ................. ........... .............................. 14

Problematic Situation........... ...................... .. .... ..... ............ ....16
Problem Statem ent ............................................................... .... ..... ........ 18
Research Objectives.......... ......... ..................................20

2 L ITE R A TU R E R E V IE W ........................................................................ .. .......................2 1

Liberalization versus Protectionism and Economic Performance...............................22
Development and Growth of the Theory of Preferential Trade Agreements ......................27
Methodologies for the Empirical Analysis of the Impacts of Preferential Trade
A g re em en ts ............................................................................. 3 2
G rav ity M o d els ................................................................................. 3 2
Empirical Application of the Gravity M odel........................................ ............... 35
Computable General Equilibrium Models (CGE) .....................................................37
Application of CGE Models in Empirical Work.................................39
P artial E quilibrium M odel (PE ).......................................................................... ...... 4 1
Overview of the V erdoorn M odel ..................................................... ...................42
Em pirical A application of PE M odels ........................................ ......................... 44
Analysis of Fiscal Im pacts ...................................................... .......... .............. .. 45

3 M O D E L D E V E L O PM E N T ......................................................................... .....................48

4 DATA REQUIREMENTS, SOURCES AND TREATMENT .............. ................ 54

5 E M P IR IC A L R E SU L T S ............................................................................. .....................59

Trade E ffects......................... .. ....... ..... ........ ........... ................................. 60
Commodities Most Affected by the Contemplated Changes in Trade Policy........................66
Three Earm arked A agricultural Products.......................................................................... ...67










6 SUMMARY AND CONCLUSIONS......................................................... ............... 70

S u m m ary ................... ...................7...................0..........
C o n c lu sio n s............................. .... .. ..... .. ........................................7 2
Study Weaknesses and Suggestions for Further Research ............................................... 75

APPENDIX: MOST AFFECTED PRODUCTS BY TRADE AND TARIFF REVENUE
IM P A C T S ...........................................................................................7 7

L IST O F R E R E F E N C E S ..................................................................................... ....................8 1

B IO G R A PH IC A L SK E T C H .............................................................................. .....................85









LIST OF TABLES


Table page

4- 1 A ssum ed V alues of E lasticities ......... ............... ............................................................58

5- 1 Trade Effects from Tariff Elimination by Malawi on EU Imports Only .........................62

5- 2 Trade Effects from Tariff Elimination by Malawi on Imports from EU and SADC
M em b er State s .......................................................................... 6 3

5- 3 Baseline Levels of Import Duties and Impact on Government Revenue......................65

5- 4 Sum m ary of Static Fiscal Im pacts ............................................................................... 66

5- 5 Impacts under the EU liberalization on three key agricultural commodities...................68

5- 6 Impacts under the EU & SADC liberalization on three key agricultural commodities.....68

A- 1 List of commodities with the highest Trade Effects as a result of Tariff Elimination
on E U Im ports ................ .................................... ...........................77

A- 2 List and levels of commodities with highest Tariff Revenue Changes as a Result of
Tariff Elim nation on EU Im ports.......... ............................. ............... ............... 78

A- 3 List of commodities with the highest Trade Effects as a Result of Tariff Elimination
on im ports from EU & SAD C countires.................................................. ..... .......... 79

A- 4 List and levels of commodities with highest Tariff Revenue Changes as a Result of
Tariff Elimination on Imports from EU & SADC countries .....................................80












ACP

APEC

ASEAN

CACM

CARICOM

CER

CGE

CIF

DPRU

EBA

EC

ECOWAS

EFTA

EPA

ERP

ESA

EU

FTA

FTAA

GATT

GCC

GDP


LIST OF ABBREVIATIONS

African, Pacific and Caribbean

Asia-Pacific Economic Cooperation

Association of Southeast Asian Nations

Central American Common Market

Caribbean Community and Common Market

Closer Economic Relations

Computable General Equilibrium (models)

Cost, Insurance and Freight

Development and Policy Research Unit

Everything But Arms

European Commission

Economic Community of West African States

European Free Trade Association

Economic Partnership Agreement

Effective Rates of Protection

Eastern and Southern Africa

European Union

Free Trade Area

Free Trade Area of the Americas

General Agreement for Tariffs and Trade

Gulf Cooperation Council

Gross Domestic Product









GNP

GTAP

GSP

HS

IADB

ISI

LAFTA

LAIA

LDC

MERCOSUR


MK

NAFTA

NBER

NSO

OECS

PE

PTA

RTA

SADC

SPARTECA

TC

TD

UEMOA


Gross National Product

Global Trade Analysis Project

Generalized System of Preferences

Harmonized System

Inter-American Development Bank

Import Substitution Industrialization

Latin America Free Trade Area

Latin America Integration Association

Least Developed Countries

Mercado Comun del Sur, Mercado Comum do Sul (Southern Common Market (of
America)

Malawi Kwacha

North American Free Trade Area

National Bureau of Economic Research

National Statistical Office

Organization of Eastern Caribbean States

Partial Equilibrium

Preferential Trade Area

Regional Trade Area

Southern Africa Development Cooperation

South Pacific Regional Trade and Economic Cooperation Agreement

Trade Creation

Trade Diversion

West African Economic and Monetary Union









USAID United States Aid for International Development

US$ United States Dollar

WTO World Trade Organization









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Mater of Science

THE EU-ACP ECONOMIC PARTNERSHIP AGREEMENTS (EPA) FOR MALAWI: TRADE
EFFECTS AND FISCAL IMPACTS FOR ALTERNATIVE POLICY OPTIONS

By

Innocent Lwafyo Thindwa

August 2010

Chair: John J. VanSickle
Major: Food and Resources Economics

Malawi has, since 1975, been enjoying non-reciprocal trade preferences from the EU bloc

along with all other ACP countries pursuant to the successive Lome Convention Agreements.

This framework was not consistent with the existing World Trade Organization (WTO) rules and

regulations. As such, the EU and ACP were pressured to come up with new World Trade

Organization (WTO) compatible trading arrangements. To this effect, the Cotonou Agreement

was conceived by the two parties and signed in June, 2000. Under the Cotonou Agreement, the

non-reciprocal preferential market access for ACP economies was scheduled to be replaced by a

string of Economic Partnership Agreements (EPAs) meant to progressively liberalize trade in a

reciprocal way, leading to the establishment of free trade areas (FTAs) between the EU & ACP

regional groups in accordance with the relevant WTO rules. The glaring reality for the ACP

countries in general, and Malawi in particular, of opening their markets to EU imports has

created debate on the anticipated implications.

This research employed the partial equilibrium model to quantify the trade effects and

fiscal impacts for Malawi of an EPA with the EU, compared to those obtainable in a

contemplated broader liberalization scenario that included implementation of the EPA with the









EU concurrently with the formation of the regional Free Trade Area (FTA) among the Southern

Africa Development Cooperation (SADC) group of countries.

Results from the study demonstrate that Malawi stands to benefit more from a broader

liberalization framework than one restricted to the EU. The results also provide useful insight

into the types of product categories that would be most hit by the policy change. In particular,

trade and fiscal effects to the three key agricultural products of tea, sugar and tobacco were

found to be insignificant.









CHAPTER 1
INTRODUCTION

The post World War II period saw the rejuvenation of efforts aimed at promoting

multilateral trade liberalization among countries. The climax of this effort was in 1947, when the

General Agreement on Tariffs and Trade (GATT) was established with a view to promoting free

trade through elimination and/or reduction of trade barriers among countries. Following the

classical arguments of Adam Smith and David Ricardo, free trade is considered to be welfare

improving as each country tends to allocate their productive resources in areas that they have the

most comparative advantage. As such, trade liberalization is considered a desirable vehicle

towards the attainment of sustainable economic growth and development, leading to

improvement in living standards for the populace.

Parallel to the movement towards multilateral trade liberalization, was the emergence and

proliferation of the concept of regional trading arrangements. These were (and still are)

configured through various forms and names depending on the varying degree and/or stage of

economic integration, such as free trade area, customs union, common market, economic union,

and complete economic integration. This feature of the global economy stimulated a lot of debate

and theorization among economists during the entire second half of the twentieth century on the

desirability and relationship of regionalism vis-a-vis multilateralism as an approach to free trade.

One of the offshoots from this debate has been the development of the theory of second best,

referring to regional trade arrangements as opposed to the superior case of multilateral trade

liberalization. It is worth noting though that the proliferation of the regional free trade

arrangements is not without legal basis in the international trade rules and regulations as Article

XXIV of the GATT sanctions their formation.









At the heart of these multilateral and regional trade arrangements has been the feature of

preferential treatment, usually sought by the developing countries, which calls for special

considerations and exemptions from reciprocal trade liberalization measures. Thisfair treatment

plea falls within the larger body of arguments by some against equal treatment of developed and

developing countries in the pursuit of open trade policies on account of the notion that smallness

of the economies for the later makes them more vulnerable to external shocks. The proponents of

this paradigm have advanced the argument that openness and liberalization do have significant

negative effects on poverty, unemployment and a host of other issues (Weisbrot and Baker 2002;

Grunbaum 2007). Free trade proponents dispute this and argue that these small and low income

countries need not be treated differently if they are to realize significant positive economic

impacts from trade.

These arguments and counter-arguments for open trade are reflective of the generally

observed fact that trade liberalization does not only create challenges to the country undertaking

such policy but also offers them opportunities to attain superior levels of growth and prosperity

through improved competitiveness and enhanced efficiencies (Grunbaum 2007). This is exactly

the situation in which the African, Caribbean and Pacific (ACP) group of countries find

themselves now as they face the glaring reality of opening their borders to imports from the

European Union countries under the 2000 Cotonou Agreement between the EU on the one side

and ACP countries on the other. Malawi is an integral member of this group of countries, itself

being within the African countries sub-group. With an estimated annual gross domestic product

(GDP) level of about US$ 4,268 million in 2008, the country is categorized as one of the least

developed countries (LDC). As an LDC, Malawi benefits significantly from the EU trade

preferences under the previous trading arrangement as well as the on-going EU unilaterally









determined Everything But Arms (EBA) arrangement. The country's major export commodity

under the EBA arrangement is sugar, which is sold at an EU predetermined high price in pursuit

of subsidizing sugar production within the EU farming realm.

Problematic Situation

The ACP countries have benefited from preferential treatment from the EU since 1975,

when the Lome I Convention was signed in Lome, Togo. Under this agreement, the ACP

countries were accorded non-reciprocal preferential market access to the European Union

countries. The period that ensued saw successive renewal of the initial Lome I Convention,

through Lome II in 1979; Lome III in 1984; and Lome IV in 1989 that remained in force for the

entire decade until 1999. All trade preferences, articles of trade and development aid schemes

under these renewals were characterized by the same non-reciprocal principle. At the dawn of

the twenty-first century, the wave of globalization and the associated efforts to strengthen the

multilateral approach to trade and economic processes warranted the need for adapting the EU-

ACP trade arrangements to ensure their compatibility with contemporary WTO rules and

regulations. Besides, the section of developing countries that falls outside the ACP group had not

been supportive of the discriminatory EU-ACP preferential trade agreements under the Lome

Conventions. As such, the WTO accorded a final waiver to the EU and ACP countries running

only up to December 2007, after which the requisite trade cooperation was required to be made

WTO compatible. To this end, the two parties commenced negotiations in September 1998 that

culminated into the new ACP-EU Partnership Agreement, signed in Cotonou, Benin on 23rd June

2000.

One of the major innovations of the new ACP-EU Partnership Agreement, commonly

known as the Cotonou Agreement, was the introduction of new fundamental principles with

respect to trade between the European Union and the ACP countries relative to the Lome









Convention. In particular, the non-reciprocal preferential market access for ACP economies was

scheduled to be replaced by a string of Economic Partnership Agreements (EPAs) meant to

progressively liberalize trade in a reciprocal manner. In fact, Article 36.1 of the Cotonou

Agreement empowers parties to the Agreement "to conclude new World Trade Organization

(WTO) compatible trading arrangements, removing progressively barriers to trade between them

and enhancing cooperation in all areas of trade" (The Cotonou Agreement 2000, Article 36.1). It

was envisaged that the progressive removal of barriers to trade would result in the establishment

of Free Trade Agreements (FTAs) between the EU and ACP regional groups in accordance with

the relevant WTO rules and help to enhance the existing regional integration efforts among the

ACP. The EPAs were to be negotiated from September 2002 to 31st December 2007 to enable

the new trading arrangements to enter into force by 1st January 2008 (The Cotonou Agreement

2000, Article 37.1). However, this deadline was not met and to date, only the Caribbean sub-

group and a couple of African countries have managed to sign a fully complete EPA. Since the

substantive issues under this process were discussed within seven configured regional groupings,

Malawi together with fifteen other countries formed the Eastern and Southern Africa (ESA)1

configuration and launched the EPA negotiations with the EC on 7th February, 2004. The ESA

group identified six clusters of issues to negotiate an EPA with the EC, which includes the

following; Development Issues, Agriculture, Market Access, Fisheries, Trade in Services, and

Trade-Related Issues.

The formation of EPAs and elimination of EU's long cherished trade preferences for

ACP will obviously have far reaching implications on development strategies followed by these

countries. In essence, these ACP countries have for a long time tailored their productive efforts


1 Other group members are: Burundi, Comoros, DR Congo, Djibouti, Eritrea, Ethiopia, Kenya, Madagascar, Mauritius,
Seychelles, Rwanda, Sudan, Uganda, Zambia and Zimbabwe









towards sectors favored by the preferences, thereby creating a potential for distortions in

resource allocation. It is clear that the anticipated change in the trade framework, which would

invariably result in a Free Trade Area (FTA) will necessitate some appropriate restructuring of

the development strategies for the ACP so that concentration is placed on areas for which each

one of these countries has the most competitive advantage. This restructuring process, otherwise

known as a rationalization process, involves costs that the ACP countries need to be prepared to

incur. The process will practically mean transferring resources, including laying off workers,

from sectors deemed inefficient to those for which the country has comparative advantages. It is

generally feared that these changes may bring about increased unemployment, provoke

heightened economic insecurity and political instability (Keck and Piermartini 2005). There is

also a very clear sense of despair among ACP countries with respect to the level of competition

that domestic producers will be subjected to after the FTA. While competition is lauded as a

positive phenomenon in that it promotes efficiency in allocation of productive resources,

commentators in ACP countries have expressed concerns over the level of EU production

subsidies in the agricultural sector and argue that such competition will invariably be unfair.

Problem Statement

The delay by ACP countries in finalizing the EPA regional frameworks and the

consequential formation of the FTAs between the EU and ACP countries is clear testimony of

the significant concerns among ACP countries with the proposed reciprocal trading arrangement.

While the nature of concerns is similar for most of the ACP countries, it is undeniable that the

extent of the envisaged impacts, both positive and negative, will vary from country to country. In

this regard, country-specific analyses of the potential impacts of the proposed FTA with the EU

are of colossal importance because outcomes from such work are bound to help the concerned

country to chart the best way forward as regards its national development strategies vis-a-vis









trade policy. This is precisely the type of research that this thesis was formulated to carry out for

Malawi.

The fact has already been alluded to above that trade policy is usually analyzed on the

basis of the link it has to economic growth and development for the country. A large body of

literature on the analysis of economic integration and its shallower forms such as the proposed

Free Trade Area (FTA) suggest that their formation tends to breed various economic effects,

particularly relating to the direction of trade which is largely influenced by the underlying

advantages for countries involved. "An expansion in trade can be had when less efficient

domestic production is substituted by imports from a more efficient member of the agreement"

(Grunbaum 2007, p.19). A further expansion in trade flows can be expected as the domestic

country imports more products from member countries at preferential terms at the expense of

excluding imports, which in fact can be less costly, from non-member countries (Viner 1950;

Pomfret 1988; Grunbaum 2007).

The trade liberalization argument can not be complete without consideration of the fiscal

impacts associated with the immediate loss of tariff revenue. This loss directly affects the ability

of the liberalizing government to deliver public services to its citizenry and negate on-going

efforts towards economic growth, development and poverty alleviation. The situation tends to be

more serious for countries where the proportion of revenue from import tariffs is higher than

where it is lower. The situation for ACP countries is such that in most countries, revenue derived

from import tariffs constitutes a substantial portion of government revenue. In fact for Malawi,

the Development Policy Research Unit of the University of Cape Town (DPRU) estimates the

contribution of tariff to total government revenue as standing at about 22 percent (DPRU Policy

Brief No. 01/P8, 2001). The situation currently ruling presents us a rare opportunity to carry out









research that will enable us to get quantitative estimates of the potential trade effects and

associated fiscal impacts of the proposed EU-EPA for Malawi. For Malawi, trade in three key

agricultural products is of paramount importance. Exports of tobacco, tea and sugar together

account for over 70 percent of the country's total annual exports. As such, it will be interesting in

this study to move a step further and isolate the potential impacts that the EPA with EU will have

on sectors involved in the production and processing of these three key products.

Research Objectives

The move towards EPA and FTA poses a huge challenge for ACP countries such as

Malawi. The consideration of potential benefits of improved and more secure access to EU

markets by Malawian exports is blurred with the recognition of expected loss in revenue from

customs duty and increased competition for the domestic industries. As debate on the likely

effects continues and time for implementing the EPAs is already surpassed, very few country-

specific quantitative analyses of the situation have been undertaken, ex ante. This study is

precisely aimed at addressing this apparent gap for one particular country, Malawi. As such, the

overall objective of this research is to evaluate the potential trade effects and fiscal impacts of an

EPA trade arrangement regime for Malawi.

Specifically, the proposed research is aimed at attaining the following objectives;

* Conduct a quantitative analysis of the potential trade and fiscal effects for Malawi of the
proposed EPA and/or FTA with the EU. Trade creation, trade diversion and fiscal impacts
will all be estimated.

* Identify commodity categories that show highest trade effects, particularly investigating
the trade effects to imports in the three key agricultural products of tobacco, tea, and sugar.

* Based on outcomes from above analyses, provide alternative policy options for the
government and stakeholders on what course of action should be pursued in respect of the
upcoming EPA/FTA.









CHAPTER 2
LITERATURE REVIEW

The essence of the argument for international trade is with respect to its presumed link to

economic growth and development. Classical trade theory suggests a positive relationship

between increased open trade and superior economic performance and that protectionism is

inversely related to economic growth and development. Freer trade regimes result in more rapid

economic growth through direct effects of trade that operate via dynamic advantages of

increased capacity utilization and more efficient investment projects as well as indirect effects

through accelerated export growth (Edwards 1993). On the other hand, there are some

researchers who object to this proposition and argue that empirical evidence does not strongly

support the classical view. They also question the validity of the direction of the casual

relationship between export growth and economic development as insinuated in the classical

argument.

As stated in the first chapter, the formation and proliferation of preferential trade

agreements and/or regional trade agreements has become an important feature of the global

economy. This feature has been considered by some as a catalyst for attainment of the more

desired multilateral free trade regime, yet others see it as an obstacle to the same. This chapter

starts by providing a review of the theoretical and empirical literature on the link between trade

and economic performance. We further endeavor to provide the foundations of the development

and growth of the theory of preferential trade agreements before reviewing its empirical

literature. We end the chapter with a presentation of the major analytical methodologies that

have been used in the empirical analysis of preferential trade agreements, including citation of

the empirical studies on which those methodologies have been applied.









Liberalization versus Protectionism and Economic Performance1

The notion that freer international trade stimulates growth and development dates back to

the days of Adam Smith. A host of later economists worked to augment and popularize the free

trade notion in the two centuries that followed. Nonetheless, the post World War II decades of

1950s, 1960s, and 1970s saw increased emphasis on the pursuit of protectionist development

strategies, especially among developing countries. A major impetus for such an orientation in the

developing world was the policy of Import Substitution Industrialization (ISI), embraced by

many of these countries in Latin America, Africa and East Asia. The urge to develop and sustain

their manufacturing sectors provided for the pursuit of the ISI strategy which promoted

protectionism in line with the infant industry argument for industrialization. It was believed that

successful economic development would be attained through rapid development of the local

industry to ensure self-sufficiency and insulate the domestic economy from external

vulnerability. One of the major consequences of the ISI euphoria was the preoccupation by most

of the development economists of the time with the design of planning models based on ISI. In

spite of the apparent dominance of the protectionist paradigm, "a small group of academics

embarked, independently, on major empirical investigations aimed at assessing the consequences

of alternative trade regimes" (Edwards 1993, p. 1359). The pioneer work to this effect was

undertaken by Little, Scitovsky, and Scott (1970) and Balassa (1971), who calculated effective

rates of protection (ERP) in a score of developing countries and linked these to the countries'

overall economic structure and performance2


1 This section draws heavily on the work of Edwards (1993) who provides an excellent synthesis of the studies
undertaken earlier by various authors on empirical analysis of the link between outward looking and inward looking
trade policies with economic performance
2 An excellent review of these studies is provided by Edwards (1993) and summarized and cited by Grunbaum
(1997)









The pursuit of ISI by the developing countries was criticized for its disregard of

fundamental economic principles such as that of comparative advantage in determining what to

produce. As a consequence of this disregard, some fundamental distortions in resource allocation

were argued to have been identified by some authors. For instance, the Little, Scitovsky, and

Scott (1970) study concludes "that the policies followed in most of the developing world after

World War II had excessively encouraged industrialization at the cost of reducing the incentives

for expanding agriculture and exports" (Edwards 1993, p. 1362). As a consequence of these

inward looking policies, those developing countries experienced a ray of economic problems

including "a worsening income distribution, a reduction in savings, an increase in the rate of

unemployment and a very low rate of capacity utilization" (Edwards 1993, p. 1362).

African countries were a major segment of the LDCs that pursued inward looking trade

policies in the hope of achieving positive economic growth and self-reliance. However, instead

of attaining these positive socio-economic benefits, implementation of these protectionists

policies only resulted in severe crises for most of the African countries as "market incentives

were seriously distorted, food production plummeted, GNP per capital fell by almost one percent

per year during the 1970s, corruption became rampant, and shortages were generalized"

(Edwards 1993, p. 1370). This appears to suggest that adoption of liberalized trade policies

would have helped these countries to improve their economic performance. As such, in its 1981

report, the World Bank recommended that these countries undertake the implementation of

reforms to their economic systems to make them more liberal and more open to international

trade.

Notwithstanding their novel pioneering work on the empirical link between trade policy

orientation and economic performance, the policy recommendation for more open trade in the









developing countries by Little, Scitovsky, and Scott (1970) and Balassa (1971) was heavily

criticized by Paul Streeten (1971) for lack of additivity and inconsistency in their arguments for

freer trade. Edwards (1993) also pointed out that the authors only used single-period snapshots of

the protection levels in the specific countries to derive their conclusions other than attempting to

estimate evolving protection levels which could have permitted them to analyze liberalization

episodes over time and the link between alternative protection levels and growth in those

particular historical settings. Edwards (1993) further identifies a computational weakness in the

two studies, that while using the same technique, the Little, Scitovsky, and Scott (1970) study

calculated a 49 percent effective rate of protection to the manufacturing sector for the Philippines

in 1965 while the Balassa (1971) study arrived at a rate of 61 percent for the same sector in the

same country during the same year.

In a manner that served to address the second weakness in the above studies, one of the

classic works was undertaken by Anne Krueger (1978) and Jagdish Bhagwati (1978). In the

NBER study that they co-directed, Krueger (1978) and Bhagwati (1978) formally derived a trade

liberalization index based on the degree of bias against exports. With the help of this index,

countries could be determined to be at varying stages of liberalization, starting from a scale of I

as the most protected to V as the most liberalized. Assigning dummy variables for these stages,

and using country-specific pooled data for traditional and non-traditional exports, Krueger

(1978) undertook a formal econometric analysis of the link between trade liberalization and

economic growth. She estimated two equations, one for exports and the other for real GNP for

each one of the ten countries in the sample. The first was dependent on the real effective

exchange rate, dummy variables for the trade liberalization regime, and time trend variable while

the later was dependent on the exports index, trade liberalization regime dummy and a time trend









variable. The results showed significant positive effects on export growth for the coefficient on

the dummy variable for phases four and five of the liberalization ladder, while no direct

significant effects were discerned from the same phases of liberalization on GNP growth. The

model did produce a significant positive effect on GNP growth for the export coefficient. This

suggests existence of an indirect as opposed to the theoretically surmised direct positive link

between trade liberalization and economic growth.

Krueger's finding of lack of direct effect of trade liberalization on economic performance

did not go without challenge from other researchers. Balassa (1982) in his study concluded that

countries with intense trade liberalization policies displayed high rates of economic performance.

In the Balassa (1982) study, both quantitative restrictions to trade and tariff barriers were

incorporated into the methodology used to categorize the country's trade regime. This was done

to address what Balassa (1982) pointed out as one of the shortfalls of the Krueger (1978) study

i.e. that use of quantitative restrictions alone in her study to determine trade regimes was in fact

leaving out a major component of bias against exports inflicted by high tariff rates. Balassa

(1982) used export growth rates as a proxy for trade regime, which turned out to be one of the

sources of criticisms leveled against his study. As Edwards (1993) puts it, "it is not clear whether

it is exports growth that causes output expansion" (p.1373), or that output expansion spurs

exports growth.

In the years that followed, in the 1990s, a number of influential studies were undertaken

to establish the link between trade openness and economic performance. The approach for these

later studies had been different and quite flexible in searching for alternative ways of

liberalization and establishing their impact of economic performance. As Grunbaum (2007)

observes that "while early empirical studies related to the role of exports and growth, more









recent studies on trade policy and economic performance have searched for alternative measures

of openness and their relationship to economic outcomes" (Grunbaum 2007, p. 61). Grunbaum

cites the works of such authors as Dollar (1992); Sachs and Warner (1995); Harrison (1996);

Edwards (1998); Frankel and Romer (1999) as some of the most influential in undertaking cross-

country econometric studies on the subject. In most of these studies, the results have tended to

suggest the existence of a positive link between more open trade and higher rates of economic

growth.

Notwithstanding the foregoing, the view that trade liberalization is directly linked to

higher levels of economic growth has also been criticized by a number of authors who refute it

and argue that there is no firm empirical evidence establishing such a direct link.

The work of Jeffrey Sachs (1978), and Deepak Lal and Sarath Rajapatirana (1987) stand out in

literature as some of the leading representatives of this criticism. Most recently, Rodriguez and

Rodrik (2001) carried out a critical review of some of the works of earlier authors which

supported the positive link between open trade and economic performance. By using original

data sets, Rodriguez and Rodrik (2001) were able to "replicate and analyze the measures of

openness used, disaggregated these into tariff and non-tariff components to test statistical

significance, extended and modified the empirical models to obtain additional results"

(Grunbaum 2007, p. 61). Their findings suggest that those results in the original studies had

been greatly over-stated and that the purported link was in fact very weak.

Authors have acknowledged existence of a couple of challenges associated with these

cross-country econometric studies, which make generalization of results tricky. These include

"poor data, inappropriate methodology and a weak theoretical framework lacking the ability to

establish clear relationships between variables" (Grunbaum 2007, p.61).











Notwithstanding the quandary portrayed in the literature reviewed above about the

significance of the link between trade liberalization and economic performance, it is important to

recognize that a more open trade regime brings with it elements that promote economic growth

and development. Whether that happens directly or indirectly is beside the point. "Trade imposes

rigorous competition on domestic producers thereby enhancing competitiveness, fosters foreign

direct investment, alters the political economy by reducing rent seeking behavior, constrains

government's ability to manipulate macroeconomic policies and subjects and binds economic

actors to the discipline of international markets and the international environment" (Grunbaum

2007, p.62). In addition, trade liberalization allows for feedback mechanisms to evaluate the

performance and effects of policies (Berg and Krueger 2003; Grunbaum 2007).

Development and Growth of the Theory of Preferential Trade Agreements

After the Second World War, the international community agreed on a set of rules and

principles aimed at promoting multilateral trade under what was known as the new international

economic order, under the regulation of the GATT. However, it soon became apparent that

attaining free trade at that level was slow and difficult. As such, a number of countries started

forming smaller groups, configured on the basis of either geographical vicinity or political and/or

historical perspectives to pursue regional economic integration in a preferential manner. This

feature of global economy has come to be known as regionalism, with its peak at development

experienced in the 1950s and the 1990s. As such, Bhagwati and Panagariya (1996) have termed

the two stages as the old regionalism of the fifties and the new regionalism of the nineties. The

proliferation of these regional trade agreements and/or customs unions in the period after the

1950s spurred a lot of interest and debate on their desirability and relationship to the multilateral

trade agenda. Some free trade proponents argued that regionalism was a necessary stepping stone









to attainment of the most desirable multilateral trade agreement while others abhorred the feature

for its tight grip at restrictions to trade against non-members to the agreement. At the same time,

some protectionists liked the regional trade agreements for increasing the degree of preference

for favored countries and maintaining restrictions against non-member countries while others

disliked them for dismantling protection accorded to domestic producers against suppliers from

favored countries.

The work of Jacob Viner (1950) stands out as the earliest and most influential on this

topic, laying out the primary theoretical basis for the analysis of regional and/or preferential

trade arrangements. The most significant contribution of Viner's work was his observation that

welfare effects from any form of regional economic integration are not unambiguous a priori. As

such, no general conclusion is possible on whether such economic integration arrangements are

welfare enhancing or damaging. Central to the development of his theoretical framework was the

formulation of the concepts of trade creation and trade diversion. Viner's original definitions for

these two concepts have been refined over time by various subsequent researchers. Thus, we

define trade creation as the substitution of domestic production with cheaper imports from a

member country while trade diversion is the substitution of cheaper imports from non-member

countries with more costly goods from a member country (El-Agraa 1997; Gordon 2007). As

such, "trade creation reflects a shift from an inefficient to an efficient source of supply while

trade diversion is movement from an efficient supplier to an inefficient one" (Gordon 2007,

p.34). In his refinement of the definition, Robson (1980) distinguished between two components

of trade creation. The first one of these is "a production effect reflecting savings from the

reduction of domestically produced goods and second, a consumption effect reflecting the gains

in consumer surplus as high cost consumption goods were substituted for lower cost goods"









(Grunbaum 2007, p.65). Similarly, Grunbaum (2007) summarizes that trade diversion reflects to

a shift in the source of imports from a non-union low cost producer to a more costly union

member.

The theoretical framework3 developed by Jacob Viner and all refinements made

thereafter provided a fertile basis for the empirical analysis of regional economic integration.

With this tool in hand, one would proceed to estimate the trade effects and welfare outcomes of a

preferential trade agreement simply by estimating the extent and comparing the size of trade

creation and trade diversion. The agreement is regarded as welfare enhancing and therefore

advantageous in the event that trade creation exceeds trade diversion (Grunbaum 2007). Since

only a sub-set of the country's trading partners is included in the preferential trade agreement, a

three country construct, involving the first country, member countries in the agreement and non-

member countries in the same becomes necessary. It is in this setting that the analysis proceeds

to simulate and/or evaluate the trade flows and production and consumption effects. The

economic integration agreement so being evaluated would have predictable impacts on resource

allocation, economies of scale, terms of trade, factor productivity, economic growth and stability,

and the distribution of income (Robson 1980; Grunbaum 2007). Nonetheless, this analysis is

based on a number of restrictive assumptions that include the following4; perfectly elastic supply

for imports; perfect competition for factor and product markets; factor mobility within countries

but not among countries; zero transportation costs; tariffs as the only available policy tool; prices

accurately reflecting requisite opportunity costs; balanced trade; and full employment of

resources.


3 This section draws heavily from the work of Grunbaum (2007), who provides an excellent synthesis of the
Vinerian theoretical framework and assumptions thereof, as refined by Robson (1980).
4 This list of the Vinerian assumptions was summarized by Robson (1980) and cited by Grunbaum (2007).









Apart from the later work of Robson (1980) cited above, previous authors such as R.G.

Lipsey (1960); Harry G. Johnson (1960); and C.A. Cooper and B.F. Massell (1965) also helped

to advance the Vinerian framework with their respective refinements to his original ideas. In his

influential survey article that aimed at further refining Viner's trade creation and trade diversion

concepts, Lipsey (1960) observed that Viner assigned a positive value to trade creation and a

negative value to trade diversion. As such, Lipsey (1960) maintains that the suggested inclusion

of the consumption effects in their definitions would invalidate such a motive by Viner.

Viner's influential insights uncovered what came to be known in literature as the theory

of second best, literally referring to partially liberalized regimes characterizing preferential trade

arrangements. At the same time, the Vinerian theory and all literature from subsequent authors in

the 1960s seemed to have failed to explain the motive behind these regional and/or preferential

trade agreements. As Pomfret (2003) observes literature of the 1960s "failed to explain why

countries would form a customs union, when they could realize all the trade creation benefits and

avoid any trade diversion costs by reducing tariffs in a non-preferential manner". Johnson

(1960); and Cooper-Massell (1965) separately provided a formal treatment of this question and

suggested that formation of the preferential trading arrangements is largely motivated by political

other than economic reasons since their regimes can not breed superior economic effects to

unilateral trade liberalization. They cited the pursuit of ISI in Latin America and Africa, and

maintenance of peace and harmony between France and Germany in the EU case as two

examples of such political reasons for PTAs formation. Technically, Johnson (1960) "advocated

that the measurement of trade creation and trade diversion should include production and

consumption effects as changes in import demands were a consequence of the formation of a

customs union and the tariff reduction or elimination" (Grunbaum 2007, p. 67). Cooper-Massell









(1965) on the other hand observed that measurement of the welfare effects from a customs union

has to include the impacts of a tariff reduction which results in consumer surplus gains and not

just those of the pure trade diversion. This means that if the former effect is larger than the later,

then the regime is beneficial to the country. Notwithstanding this inclusion, Cooper-Massell

(1965) maintained that a negative welfare effect would be had from these preferential trade

arrangements.

Meade (1965); Mundell (1964) and Corden (1972) also provided influential

augmentations to the Vinerian framework5. The initial theoretical framework developed by

Jacob Viner had a number of gaps that he himself knew would come up later and be addressed

by other researchers. As Grunbaum (2007) observes "Although Viner was aware that economies

of scale, imperfect competition, and terms of trade issues would arise, he left them unattended"

(p. 66). Indeed Meade (1965) led the task by providing a critique of Viner's model and extending

the same in a general equilibrium framework. He assumed existence of infinite supply elasticities

and demand elasticities of zero, providing an allowance for multi-product production in all

countries. "Meade's work provided a general static framework of analysis for integration

agreements that admitted substitution of goods both in demand and supply and allowed for

simultaneous adjustments in related factor and goods markets in trading countries" (Grunbaum

2007, p.67). Using a three country model, Mundell (1964) addressed the dynamics that arise in

measuring welfare effects by accounting for changes in the terms of trade that come with the

formation of preferential trade arrangements through changes in tariffs and relative prices.

Investigating the relevance of the notions of trade creation and trade diversion in the face of scale

economies, Corden (1972) did maintain their relevance. However, he suggested that such


5 Grunbaum (2007) excellently synthesizes the contributions from these authors and this brief discussion draws from
his work.









analyses should incorporate cost reduction and trade suppression effects that do emerge from

scale economies resulting from the regime.

Methodologies for the Empirical Analysis of the Impacts of Preferential Trade Agreements

Literature cites three major quantitative analytical techniques that researchers use to

empirically evaluate the effects of entering some form of an economic integration by countries.

These include the gravity models, computable general equilibrium (CGE) models and the partial

equilibrium (PE) models. Naturally, each one of these methods has its respective merits and

demerits such that choice of the appropriate methodology has to be informed by consideration of

the costs and benefits to the researcher based on the specific situation at hand, including data

requirements considerations. This tends to help in balancing the inherent trade offs. It is also

clear though that gravity models are employed to carry out ex-post analyses of the economic

integration agreements while CGE and PE models are used to conduct ex-ante analyses that

simulate "how today's economy will look in future as a consequence of a specified set of policy

changes" (Piermartini and Teh 2005, p.1). The one general thing is that in all cases, the analysis

undertakes to quantify the sizes of trade creation and trade diversion as a result of the policy

changes thus implemented, such as formation of a PTA.

Gravity Models

The gravity models in trade policy derive their name from the 1687 Newtonian Law of

Universal Gravitation6 which held that the attractive force between two objects is positively

related to their mass and inversely related to the distance between them. Economists have used

the same concept to develop an analogous functional relationship of the gravity trade model. As

such, bilateral flows are considered to be positively related to the size of the countries'


6 This section draws on the work of Keith Head (2003), Piermartini and Teh (2005) and an excellent synthesis of the
model evolution and structure by Grunbaum (2007).









economies and inversely related to the distance between them (Head 2003; Grunbaum 2007).

This formulation was first proposed by Tinbergen (1962). As Piermartini and Teh (2005) observe

that "the first empirical study of trade using the gravity model was probably Tinbergen's7 (1962),

although there was no explanation for the use of the model nor for showing how it was related to

theoretical explanations of international trade" (p.38). Thus, Grunbaum (2007) summarizes the

basic Tinbergen gravity model as follows;

T,j = M I1 MI2DP3 (2-1)

where

Ti = bilateral trade flow between country i and country j,

Mi = economic mass interpreted as GDP of country i,

Mj = economic mass of country j,

Di = distance between countries i and j,

Ei = the standard error term.

The above formulation is usually transformed into its requisite logarithmic form to

estimate the following;

InT, = flo + fInAM + l2InA fl3D, + E, (2-2)

Literature suggests that GDP, GDP per capital or population size can all be used to

capture economic mass measured by M, and Mj in this model while Head (2003) observes that

distance is almost always measured using the "great circle" formula, which approximates the

shape of the earth as a sphere and calculates the minimum distance along the surface. Thus using

this formula, the distance between the capitals or commercial cities of the two economies is



7 Other authors such as Head (2003) and Grunbaum (2007) makes a similar observation. However, Grunbaum cites
the work of Sandberg (2003), who observes that Poyhonen (1963) was simultaneously working on a similar type of
the model.









calculated and used in the D, variable in the model. Grunbaum (2007) explains that economic

size determines the ability of the country to engage in trade as trade flows between countries is

largely a function of supply conditions in the country of origin and demand conditions in the

country of destination. Head (2003) gives a number of reasons to justify use of distance in the

model. These include the fact that distance serves well as a proxy for transport costs that include

freight and marine insurance; it indicates the time required for shipment which is crucial,

especially for perishable goods; it is a proxy for communication costs which involves personal

contacts between managers and customers to exchange information on the exchange to be made.

Distance also captures synchronization costs involving time of delivery of inputs into the

production process; it may also be a good proxy for the transactions cost of doing business; and

the fact that more distant areas are likely to be more culturally different suggests that trade

between such areas may be lower compared to geographically close areas/countries.

As noted above, the original gravity model was criticized for its lack of touch with the

standard theoretical bases for international trade as posited in the famous Ricardian and

Heckscher-Ohlin models. The former emphasized differences in comparative advantages among

countries to explain trade flows and pattern while the later dwelt on the variations in factor

endowment among countries. This criticism became one of the major bases for further

refinements of the standard gravity model as proposed by Tinbergen. Piermartini and Teh (2005)

observe that Anderson (1979) was probably the first author to establish the theoretical basis for

the gravity models. They observe that Anderson (1979) did this by constructing a model "where

goods were differentiated by country of origin and where consumers have preferences defined

over all the differentiated products" (Piermartini and Teh 2005, p. 38).









While the standard gravity model has been lauded to have performed significantly well in

explaining trade, Head (2003) notes that "there is a huge amount of variation in trade" (p.9) that

these models cannot explain. As such, researchers have chosen to include a number of extra

variables that have been observed to be relevant in explaining trade flows. Some of these

variables include; dummies to capture the effects of country adjacency, whether a country is

landlocked or not, whether two countries speak the same language or share some colonial

history, whether the country is a member of some PTA or RTA, customs union, whether they

share the same currency and so on. As Head (2003) observes, authors include these variables,

albeit there is weak theoretical justification for doing so. The fact that the two major variables in

the standard gravity model, GDP and distance tends to fit the data well, increases the temptation

to include any of these seemingly relevant variables. It is important though that choice of which

variables to include in the model must be guided by the specific issue at hand. As Piermartini and

Teh (2005) observe that one needs to proceed carefully in analyzing the theoretical questions at

hand which should guide the choice of appropriate regressors to be used in the empirical

estimation method.

Empirical Application of the Gravity Model

As alluded to above, gravity models have been extensively used to evaluate the effects of

preferential trading arrangements (PTA), simply by adding an intra-bloc and an extra-bloc8

dummy variable to capture such effects in the standard model. The analysis proceeds on the

presumption that the "normal trade volume" between countries is explained by two variables,

economic size of the trading partners and distance between them, such that any significant

effects from the PTA will manifest in either increased (trade creation) or reduced (trade


8 This nomenclature, intra-bloc and extra-bloc, draws from Piermartini el at (2005) usage









diversion) volumes between these partners as compared to the "normal volume". Thus, trade

creation is captured by a significant positive coefficient for the intra-bloc dummy while trade

diversion is captured by a significant negative extra-bloc dummy variable coefficient.

Frankel (1997) applied the traditional log-linear gravity model on levels of variables to

study the trade effects of a number of RTAs, including EC, NAFTA, EFTA, Adean, ASEAN,

and MERCOSUR. Using total trade as the dependent variable, the model was augmented by two

dummies for intra-bloc and extra-bloc trade. The study established the existence of net trade

creation effects for EC, MERCOSUR and Adean, and significant trade diversion for NAFTA and

EFTA. Furthermore, the intra-bloc dummies were found to have no significant effects for Adean,

NAFTA and EFTA. A similar model was used by Soloaga and Winters (2001) but with two

extra-bloc dummies, one for imports and the other for exports to study the trade effects of EC,

NAFTA, EFTA, Adean, ASEAN, CACM, MERCOSUR and other blocs. The results of the

model on levels showed existence of negative intra-bloc dummies for the EU, EFTA and

ASEAN; and positive intra-bloc trade impacts for CACM, ANDEAN and MERCOSUR. The

intra-bloc dummy for NAFTA was not statistically significant. The estimates on the first

differences of the variables showed that EU and EFTA were net trade diverting, while

MERCOSUR and CACM were net trade creating. Dee and Gali (2003) used the model with the

augmentation similar to the one by Soloaga and Winters (2001) to study the impacts of Andean,

APEC, EFTA, EC, GCC, LAFTA/LAIA, MERCOSUR, SPARTECA, CER, AFTA and a set of

bilateral agreements. However, Dee and Gali (2003) constructed what they called a Member

Liberalization index (an index of the coverage of the RTA), such that the three dummies would

take the value of this index whenever the RTA was in force. Results from their analysis found









that nearly all RTAs studied were found to have net trade diverting effects, while net trade

creating results were found only for Andean, LAFTA/LAIA, US-Israel and SPARTECA9.

Nilsson's (2002) study on the effects of the preferences accorded by the EU to ACP

countries under the Lome Convention compared to the EU GSP arrangements provides a good

example of how the gravity model can be augmented to captured a ray of other variables than

those in the standard model. Dummies were included to capture the effects of cultural and

historical ties and the significance of colonial linkage between former colonies and their

European colonizers. The study found that exports for developing countries were significant and

larger under the Lome Convention as compared to the EU GSP arrangement and that historic ties

significantly explain trade linkages for selected European countries and their former colonies.

Computable General Equilibrium Models (CGE)'1

CGE trade models exploit the computer capability to construct a rigorous analytical

framework that takes into account all market linkages, retains optimization assumptions thereby

preserving its consistency with the hallmarks of the general equilibrium theory of the economy

(Grunbaum 2007). Apart from their theoretical consistency, the ability of the CGE models in

arriving at precise numerical estimates has also been cited in literature as one of the model

strengths. This is easily attained by CGE models since "the workings of an economy and the

changes that would follow specific policy implementation can be simulated, as CGE models act

to emulate the functions of laboratory experiments" (Piermartini et al (2005).

It is the ability of CGE models to capture the overall picture of the changes that emanate

from requisite policy changes that elevates their importance in ex-ante analyses of trade policies



9 Piermartini et al (2005) provide a summary of these and more such studies in their work
10 This section draws from the work of Busse et al (2005); Piermartini and Teh (2005) and Grunbaum (2007)









such as enforcement of RTAs and PTAs. As Grunbaum (2007) observes, "overall aggregate

trade, terms of trade effects, factor prices, trade creation and trade diversion within an economy-

wide model can be estimated; as can be obtained estimates of inter-sectoral linkages, prices,

wages, and exchange rates that lead to equilibrium in product and factor markets, as well as

balance of trade figures" (p. 79). As such, alternative scenarios of the specific policy change can

be conjectured and their requisite potential parametric outcomes simulated.

While CGE models may be more suitable to analyze the overall trade and welfare effects

of changes in trade policy, their data requirements are significant and can be a major hurdle for

studies involving developing countries. These models require what is known as the "social

accounting matrix", which involves input and output data for the entire economy and their

requisite inter-sectoral linkages plus their associated contributions to output; government fiscal

and/or budget accounts, disaggregated into consumption, investment, government expenditures,

balance of payments; as well as data on the volumes and values of imports and exports

disaggregated by their requisite composition, origin and destination (Grunbaum 2007). Besides,

these data matrices must be appropriately arranged into revenues and expenditures, balanced and

standardized (Piermartini and Teh 2005; Grunbaum 2007). Apart from the social accounting

matrix construct, CGE models require exogenous variables that capture the behavior and

response of producers and consumers to any changes in incomes and relative prices resulting

from the policy changes thus effected. The parameters most often needed are: elasticities of

substitution related to the responsiveness of producers to changes in relative prices of factors of

production; consumer demand and income elasticities; and Armington elasticities, which

determine the substitutability between domestic and imported products. Usually such elasticities

are borrowed from previous econometric studies and may be adapted into the current model









(Grunbaum 2007). This considerable amount of good quality data can be rarely obtainable in

developing countries, let alone the availability of prior econometric studies in those countries

from which the stated elasticities could be borrowed. Otherwise, the persistence to applying the

model in studies involving those countries may result in use of dubious quality data.

As indicated above, the nature of the data employed by the CGE model is significantly

aggregated. This level of aggregation has been criticized as a potential source of loss of detail in

fundamental relationships that may be obtained. Indeed as Grunbaum argues, "complex

simulation models where large amounts of data inputs produce precise outputs can be deceiving

as precise sources of certain results are not clearly identifiable" (Grunbaum 2007, p 81). Pundits

further attack users of the model for their tendency to choose values of the requisite elasticities

arbitrarily and the Armington elasticities from outdated studies. In view of these concerns, some

have suggested that systematic validation of CGE simulations through ex-post evaluation is

necessary to enhance the confidence and the predictive potential of the analytical results derived

from the initial studies (Piermartini and Teh 2005).

Application of CGE Models in Empirical Work

In spite of the limitations cited above, CGE models have been used quite extensively to

analyze the trade effects of not only RTAs but also multilateral trade agreements, such as the

Uruguay and Doha rounds of trade agreements under the WTO. In fact, CGE models have the

ability to isolate the trade creation and trade diversion effects from such policy changes by sector

which allows computation of sectoral welfare effects other than aggregate.

Grunbaum (2007) cites the work of Kerkala, Niemi, and Vaittinen (2000) who used a

multiregional general equilibrium model to examine the consequences for African ACP countries

of a post-Lome world. Their work simulates the effects of entering into force of WTO

compatible trading arrangements under the proposed EPA as compared to the ones obtainable









under the EU GSP system. The authors employed the Global Trade Analysis Project (GTAP)

model to simulate the potential outcomes which suggested negative welfare effects for requisite

African countries. Results also showed increased trade volumes under the EPA and declining

volumes under the EU GSP system. Grunbaum summarizes that "in both cases world welfare

increased, however, positive effects were limited to the EU while they were absent in ACP

countries" (Grunbaum 2007).

Another study on the potential effects of reciprocal trade liberalization under the Cotonou

Agreement for the West African Economic and Monetary Union (UEMOA) against the EU was

conducted by Wolf (2002). Wolf (2002) employed the CGE tool to quantify the gains from this

policy change and compare them to the losses in tariff revenue that would likely follow to these

UEMOA countries. The results from this study showed that liberalization under the Cotonou

Agreement would have significant negative effects on tariff revenue generated by the UEMOA

countries.

Keck and Piermartini (2005) used an applied general equilibrium model covering 15

regions and 9 sectors to simulate the impact of signing EPAs with the EU for the Southern

African Development Community (SADC) countries. The standard GTAP model was extended

to include the elimination of textile quotas, EU enlargement to 25 members as well as tax

revenue sharing and a common external tariff among SADC countries. Simulation of outcomes

for various liberalization scenarios was undertaken. The issue of tariff revenue loss was also

tackled in their study, including calculation of requisite tax replacement values. The results

demonstrated that EPAs with the EU are welfare-enhancing for SADC overall, leading also to

substantive increases in real GDP. They also found that, for most countries, further gains could

arise from intra-SADC liberalization. While the possibility of the EU entering an FTA with









other countries, such as MERCOSUR, was found to reduce estimated gains, these gains still

remained positive. In terms of sectoral level effects, the largest expansions in SADC economies

was found in the animal agriculture and processed food sectors, with less attractive levels in the

manufacturing following EU-SADC liberalization.

Piermartini and Teh (2005) summarize a number of studies that used CGE models to

simulate the potential benefits from trade liberalization under the Uruguay and Doha Trade

Agreements, noting their numeric benefits in dollars to world welfare11.

Partial Equilibrium Model (PE)12

By their nature, PE models focus on a detailed analysis of only one particular market or

sector, holding all other factors that can affect the same constant. In trade policy analysis, PE

models have been used extensively to evaluate the effects of discriminatory tariff modification

such as that which would occur under the EU-ACP EPA arrangement. In his quest to quantify

trade effects of the then newly established European Community, Petrus J. Verdoorn (1960)

developed a partial equilibrium model based on the Vinerian theoretical framework that remains

significantly influential even today. Verdoorn's original model has been modified by a number

of researchers. Grunbaum (2007) observes that two basic types of PE models have so far been

used in the analysis of preferential trade liberalization arrangements. Of these, the first one

assumes trade in a homogenous commodity while the second assumes trade in a differentiated

product and existence of infinite supply elasticities. Grunbaum (2007) explains that under the

first scenario, a reduction in tariff tends to spur expansion of trade flows, only to be limited by

the corresponding supply elasticities. This implies that some significant growth in trade flows

1 These studies are not covered here since our focus is on regional trade arrangements.
12 This section draws heavily on the work of Grunbaum (2007); Piermartini (2005) and Busse et al (I2 rl4). However,
we only provide a brief overview of the Verdoorn model in this section, leaving a detailed derivation in the next
chapter









can be observed if the supply elasticity of a beneficiary country such as the EU bloc in our case

is high. Under the second scenario, it is the degree of substitutability among goods that limit

trade flows expansion. As such, "a high elasticity of substitution leads to a substantial increase

in trade flows and likewise a low elasticity of substitution leads to a small increase in trade

flows" (Grunbaum 2007, p. 84).

Overview of the Verdoorn Model

The Verdoorn model follows the second of the two types of PE models stated above such

that product differentiation between supplying countries is assumed and that these products are

imperfect substitutes in use. Busse, et al (2004) argues that "this assumption seems reasonable

for African countries, since the majority of African imports consist of manufactured goods" (p.

17). The model further inherits the usual partial equilibrium analysis assumptions including the

following: no repercussions on exchange rates or incomes as a result of changes in trade flows,

iso-elastic import-demand functions, and existence of infinite supply elasticities (Busse, et al

2004). Besides these assumptions, "the model requires knowledge of import demand elasticities

and the elasticities of substitution between preferred and non-preferred imports" (Grunbaum

2007, p.84). As such, trade creation13 is captured as follows;

TC = M [At/( + t)] (2-3)

where

TC = Trade Creation
Mp =Imports from preferences beneficiary
S= Import demand elasticity
t = tariff




13 This follows Grunbaum (2007) specification of the same









In this formulation, it is further assumed that "the substitutability between imports from

preferred sources and domestic production is equal to the substitutability of all imports and

domestic production" (Grunbaum 2007, p.85). On the other hand, trade diversion is formulated

as follows;

TD = Mp a2 (o- E)[At/(l+t)] (2-4)

where

TD = Trade Diversion
MP =Imports from preferences beneficiary
a2 =Share of imports from non-preference beneficiary
a =Elasticity of substitution
S= Import demand elasticity
t = tariff

One of the major variants to the original Verdoorn (1960) model is that developed by

Baldwin and Murray (1977)14, which utilizes data for domestic production on top of data on

imports from non-preferred suppliers. Thus, while the trade creation formulation is similar to the

Verdoorn model, the Baldwin and Murray model does have a different trade diversion

framework, which is captured as follows;

TD = Mp E[At / (1+t)] [MN/MD] (2-5)

where

TD= Trade Diversion
MP=Imports from Preferences Beneficiary
MN=Imports form Non-Preferred country
MD=Domestic Production
s = Import Demand Elasticity


14 This section is a synthesis of Grunbaum's (2007) work, which identifies the Baldwin and Murray (1977) model as
the major variant of the original Verdoorn model









t= Tariff
In this formulation, the authors assumed that the substitutability between imports from a

preference beneficiary source and those from a non-beneficiary source was equal to the

substitutability between imports from a preference beneficiary source and domestic production

(Sawyer and Sprinkle 1989; Grunbaum 2007).

Despite their detail, PE models have been criticized for a number of reasons. One such

matter at issue relates to the choice of elasticities used in the model. As Grunbaum (2007)

observes, "the values for the elasticities used are chosen arbitrarily on the basis of estimates often

considered unreliable" (Grunbaum 2007, p.87). Secondly, any change in trade flows is bound to

result in repercussions in all other sectors of the economy. As such, the ceterisparibus

assumption used in PE models tends to be considerably unrealistic. Pundits criticize PE models

for failure to take into account these resultant inter-sectoral linkages and factor markets

dynamics. As alluded to before, PE models have the advantage at being detailed in their

analyses. These models permit researchers to conduct analyses at highly disaggregated levels,

thus allowing for detailed analysis even by tariff line (Grunbaum 2007). This level of detail

allows analysts to be able to pin-point those specific products and trading partners that may show

significant effects of the alternative policy simulations. As such, PE models have been used

extensively in empirical analyses of discriminatory trading arrangements.

Empirical Application of PE Models

Busse et al (2004) applied the Verdoorn model to estimate the trade and budget effects of

the EPA on ECOWAS countries and Mauritania. Their study was aimed at addressing the fears

among African countries of entering the proposed EPAs without clear understanding of the

associated costs and benefits from the same. The results showed that imports from the EU in the

ECOWAS countries were expected to increase in the range running from 5.2 percent for Guinea-









Bissau to 20.8 percent for Nigeria. Importantly, Busse et al (2004) further reports that "Trade

creation exceeds trade diversion (in absolute levels) in all scenarios and for all West African

countries" (p.26). However, the budgetary effects were found to be very severe for some of the

West African countries. Cape Verde and Gambia were cited as the two countries to be severely

impacted by the impending tariff revenue decline. The authors went a step further to identify

those product categories that were found to have highest budgetary effects as well as trade

creation and trade diversion effects for each of the West African countries covered. This, they

sorted out by both absolute and relative changes.

Max Grunbaum (2007) used the Verdoorn version of the PE model to project the effects

of trade liberalization in the Organization of Eastern Caribbean States (OECS) countries along

the EU-ACP proposed EPAs which are essentially geared to create an FTA between these two

trading blocs. He undertook to quantify trade creation, trade diversion and fiscal impacts for

three alternative liberalization scenarios. The alternative scenarios were contemplated dependent

on scope of discrimination to other trading partners as EPAs come into force. The results

demonstrated that trade liberalization results in small to modest positive trade and negative fiscal

impacts. The other importantly indication from the results was that they "provided empirical

support to theoretical arguments and policy suggestions that, for small countries such as the

OECS, broader trade liberalization is superior to limited regional trade agreements" (Grunbaum

2007, p.12).

Analysis of Fiscal Impacts

As we pointed out in the first chapter, the issue of tariff revenue loss stands out as one of

the major fears related to trade liberalization, especially among the developing countries. It is

more of an issue among these countries since tariff revenue accounts for a significant proportion

of total government revenue. The concern is that loss of such a significant proportion of revenue









may suffocate the government's ability to deliver services to the citizenry and increase

impediments towards economic development. This fear is exacerbated by consideration of a

myriad of other short-term adjustment costs associated with liberalization such as "increased

unemployment, reduction in national output, elimination of certain domestic industries and

possible macroeconomic instability" (Grunbaum 2007, p.91. It is such fears as these that have

inspired interest among researchers on carrying out empirical studies on the anticipated levels of

tariff revenue loss resulting from liberalization and derive comparisons with the anticipated

levels of benefits from the same to establish an overall net position.

A survey of literature on this subject reveals two distinguished themes along which

discussion and research has dealt with this issue in the past. While the first of these has focused

on the measurement and evaluation of the relative importance of tariff revenues as a proportion

of aggregate government revenue, the second theme has aimed at exploring alternative tax

systems and/or fiscal reforms that could help compensate for tariff revenue losses (Grunbaum

2007). To understand the fiscal effects of tariff reduction in pursuit of trade liberalization, one

has to understand the ways through which such a policy change impacts tariff revenue.

Grunbaum (2007) reviewed the 2004 Report of the Inter-American Development Bank (IADB)

and summarizes five ways through which tariff reduction affects tariff revenue. These include;

direct effects or losses due to the reduction of a given tariff line; indirect effects or revenue

decline from taxes imposed on CIF plus tariffs that are forgone due to the tariff rate decline;

elasticity effects which affect the revenues depending on whether they cause an increase or

decrease in the volume of trade of certain products; substitution effects which result in trade

diversion from the displacement of imports from partners not facing tariffs; and induced effects

that are changes in total tax revenues resultant from consumption and production patterns borne









of economic structures post liberalization (Grunbaum 2007). Appreciation of these five channels

demonstrates how complex the resultant net outcome from trade policy change can be. As

literature shows, trade liberalization can in fact result in increased tariff revenue generated by the

government depending on a given situation. A host of factors do affect the net impact from trade

policy change for a country. Some of these factors include the initial conditions such as tariff

structures and general restrictiveness to trade, domestic tax including reforms to quantitative

restrictions, reduction in the number of tariff levels, and exchange rates and exchange rate

regimes. In general, the more restrictive the initial conditions, the more the losses in tariff

revenue, and the tighter the government is on the exchange rate regime the higher the costs from

liberalization.

The level of development for the country is found to be more important in determining

the country's ability to offset loss of tariff revenue due to liberalization. The work of Khattry and

Rao (2002) was central at establishing this perplexing phenomenon. They argued that structural

limitations characterizing low-income countries make the gradual substitution of trade taxes by

domestic sources of revenue very difficult (Grunbaum 2007, p.93). The low levels of

development among these countries tend to impact negatively on the breadth and levels of

domestic tax revenue collections. As such, the potential for these governments to offset lost

revenue from tariff with domestic collections is very minimal.









CHAPTER 3
MODEL DEVELOPMENT

This study intends to utilize the partial equilibrium (PE) model to simulate the trade

effects and fiscal impacts of trade liberalization for Malawi under the EPA arrangement with the

EU. The Verdoorn (1960) version of the PE model, initially introduced in chapter two above will

be used to undertake this empirical work. The choice of this model has been dictated by data

considerations and the ex-ante nature of the analysis. As insinuated in the previous chapter, ex-

ante analyses can be conducted by either using the computable general equilibrium (CGE) or PE

models. However, the enormous data requirements for the CGE models aspects of some of

these data do not seem readily available for Malawi has left us with the appropriate PE model.

Notwithstanding their simplicity as compared to CGE models, PE models have been extensively

used for analyses of this type, from which progressive policy recommendations have emanated.

Specification of the Empirical Model': The Verdoorn model is based on a couple of

restrictive assumptions. First, the model applies the Armington assumption of product

differentiation such that imported products from different countries are assumed to be imperfect

substitutes in use. For Malawi, we find this assumption to be reasonable as the majority of the

country's imports are manufactured goods. The Verdoorn model is based on the partial

equilibrium analysis and therefore makes the following other assumptions; trade flows changes

do not have any repercussions on the incomes and exchange rates; iso-elastic import demand

elasticities; and infinite supply elasticities. The last assumption seems particularly appropriate in

the case of Malawi vis-a-vis the EU as the latter's exports to the former must account for a

significantly small proportion of its total exports. On the other hand, exports from other smaller



1 The empirical model in this section is developed based on the one specified by Busse et al (2ki 14) and Grunbaum
(2007).









Southern African Development Corporation (SADC) countries to Malawi may not support this

assumption. However, we are captivated to maintain the assumption on the understanding that

exports to Malawi from these small SADC countries do not constitute significant proportions of

their total production. Nonetheless, we do recognize that those elasticities are less than infinite.

To conduct an ex-ante analysis of trade effects, one would consider a particular category

of commodities, say M, and proceed to model the consumer's behavior in the importing country.

Following this reasoning; the utility function the consumer is assumed to maximize can be

specified as follows;

U= f [f(Mp, MN), MD] (3-1)

where

fp = denotes a separable homogenous branch of the aggregate utility function

Mp = Imports of commodity M from preference beneficiary sources

MN= Imports of commodity M from non-preference beneficiary sources

MD = Domestic production of commodity M

The assumption of homogeneity in the specification is quite consequential in that it

implies that total imports of M, given by [Mp+MN] are substituted equally for domestic

production. The formal implication of this separability of the utility function is that the Verdoorn

model differentiates between the sources of imports, those from preference beneficiaries and

those from non-beneficiaries, labeled herein as Mp and MN (Grunbaum 2007).

There are two other crucial assumptions based on which the Verdoorn model is

developed. First, that the demand function for the preference donor, Malawi in our case, for a

single category of products such as M, is given by the following specification;

Mp + MN = PpI~p P "1 (3-2)









where


Pp = the price of imports of M from the preference beneficiary sources

PN = price of imports of M from the non-beneficiary sources

F = elasticity of import demand

ap = Mp/(Mp + MN) = Share coefficient of preferred imports

aN = MN/(MP + MN) = Share coefficient of non-preferred imports

ap + aN = 1

Second, the definition of the elasticity of substitution (C) between preferred and non-

preferred imports can be given as follows;

-k= y7 ; where y is a parameter (3-3)


As Malawi strikes off tariffs (t) under the EPA arrangement, only imports from the

preference beneficiary countries will be affected, thus reducing their price Pp while leaving the

price for non-preferred imports unchanged, such that OPV Q. We proceed to differentiate

Equation 3-3, divide the result by Equation 3-3 and using Px -= 0, we derive the following;


O g (3-4)


After totally differentiating Equation 3-2 and dividing the result by Equation 3-2, we

derive;


I4p+Mt PP
S= p + EloP, 10og$}8a) (3-5)

Substituting definitions for a, and a,, following Equation 3-2 into Equation 3-5, we

obtain the following;








aF + (1 p)" = (F, !ap + clogP6a, (3-6)

By using Equation 3-4, we proceed to rearrange Equation 3-6 to obtain the expression for

the derivative of ap as follows;

Sa6 = ap( aj)o pp (3-7)

We next eliminate from Equation 3-4 by multiplying with (1 a,) and rearranging

the equation to obtain the following;
--i- )- (l-- ,}) = (1l- ,) (3-8)
S(1 a,) (1 c} a(1 a) (3-8)
Mp ME PP



Next, we insert Equation 3-9 into Equation 3-6 and use Equation 3-7 to substitute the

6a, and obtain,

= (a(1 ap) + ap + logP, a,( a-)aJ] (3-10)

Since logP is close to zero, if F9 W Py, then Equation 3-10 can be expressed as

follows;

a = I(1 -c a)+ E I ( ) (3-11)

Now, the price for preferred imports can also be expressed in terms of export prices as

follows;

PP = P(1 +t) (3-12)

Where P is the export price exclusive of the tariff (t). The total derivative of Equation

3-12 gives the following;









4 = sP(1 +t) + Pft (3-13)

We divide Equation 3-13 by Equation 3-12 to obtain the change in preferred import

prices as follows;

S =,Z M + (3-14)


By considering that &SP = 0 and assuming infinite supply elasticities, Equation 3-14

culminates into;

-F =- (3-15)
Pr 2+

We now utilize Equation 3-11 and Equation 3-15 to obtain the expression for the total

expansion in imports from the preferred sources resulting from preferential liberalization as

follows

8 = (0C1-p) +a, +s(a) (3-16)

Since as = (1 ap), the total expansion in imports above can be re-written as

S( ))( (3-17)

As indicated in chapter two above, the total change in imports given by Equation 3-17

comprises the trade creation (TC) and trade diversion (TD) components. The former refers to the

change in trade flow between the preference donor and preference beneficiary as consumers in

the importing country substitute cheap imports for domestic production while the later concept

captures the amount of preferred imports substituting those imports coming from non-preference

beneficiary sources (Grunbaum 2007). We therefore separate Equation 3-17 into these two

components as follows;









TC = Mse (3-18)

And;

TD = Mrpaa a)( ) (3-19)

We finally deal with the expected change in tariff revenue resulting from preferential

trade liberalization of the sort we are concerned with. This change is given by the sum of the

following: i) preferred imports multiplied by the preferential tariff rate, and ii) the amount of

non-preferred imports multiplied by the tariff rate applicable to imports from non-beneficiary

sources (Grunbaum 2007). Thus, we express the change in import duty revenue (5ID) as follows;

sID = M,t, + TDtN (3-20)

where

ID = import duties

tp = tariff rate applicable on preferred imports

t = tariff rate applicable on non-preferred imports

In this study, Equation 3-18, Equation 3-19 and Equation 3-20 will be utilized to estimate

the trade effects and tariff revenue impacts of trade liberalization under the EPA/PTA

arrangement for Malawi vis-a-vis the EU.









CHAPTER 4
DATA REQUIREMENTS, SOURCES AND TREATMENT

The simulation and evaluation of the impacts of Malawi entering the EU-ACP Economic

Partnership Agreement will be carried out using the Verdoorn's (1960) Partial Equilibrium

model. As indicated in the previous chapter, use of the import demand elasticities (e) allows us to

employ import data without relying on domestic production data. These domestic production

data, if were to be used, have to be captured at a highly disaggregated level such as the eight-

digit level of aggregation (i.e.; HS-8) which is being used in this study. As is the case with most

other developing countries, such type of data were not readily available for Malawi so that use of

import data was particularly necessary in this study. At the eight-digit level of data aggregation,

the simulations permitted us to assess the effects of liberalization at product level and identify

those products most impacted by the contemplated trade policy changes. To this effect, we

managed to collect the national eight-digit HS tariff and trade schedules from the Malawi

National Statistical Office (NSO)1. These data were obtained for the baseline year of 20062

The raw data obtained from the NSO indicated that in 2006, Malawi imported goods and

services from the 27 EU countries using 1,602 tariff lines compared to the total number of 3,956

lines utilized during the same year at the HS eight-digit level of aggregation.

Malawi has three major trade-related taxes, of which import duties are just one type. The

other two types include excise duties and import surtaxes. At every eight-digit tariff line utilized





1 These data are initially captured by the Malawi Revenue Authority, a Government Agency mandated to collect
taxes. However, it is the National Statistical Office that has the legal right to process and disseminate all statistical
information for the country. As such, the NSO obtains these data sets from the MRA on a monthly basis
2 The choice of this year was based on the planned deadline for EPA effectiveness which was initially scheduled to
start in January, 2008. As such, the capture data were the most recent years for which the NSO had complete data
sets as required by the study prior to the EPA signing deadline.









during the baseline year for which data was obtained, the requisite CIF3 import value was

captured along with its import duty, excise duty and the import surtax that were levied by the

Malawi Revenue Authority. These data sets were obtained in the local currency, the Malawi

Kwacha. As such, we converted all the Malawi Kwacha values into United States Dollars using

the official average annual exchange rate as reported by the NSO and the Reserve Bank of

Malawi. The rate reported and applied in this study was MK139.34 to US$ 1.00 for the baseline

year of 2006.

In line with the spirit of the study, two different scenarios of the scope of liberalization

were contemplated, with each scenario changing the definition of which (or how many) countries

constitute the "preferred source" of imports compared to the rest of the world. First, we assumed

a scenario whereby Malawi enters into an EPA arrangement with the EU without any ancillary

agreements with other trading partners outside the EU realm. For this scenario, the trade and

tariff data were organized in such a manner that at every tariff line, imports from all the 27 EU

countries were categorized as preferred imports as distinguished from those coming from

elsewhere. Secondly, we built on the first scenario by contemplating a situation whereby Malawi

correspondingly removes tariff barriers on imports from fifteen (15) other countries from the

Southern Africa Development Corporation (SADC) region4. This is not a far-fetched assumption

as negotiations for trade liberalization within this region have been on-going since the early

1990s. To capture this scenario, we added all imports from these fifteen (15) member states into

the basket of preferred imports from the EU. The significance of the second scenario can be well

appreciated if one considers the fact that a significantly higher proportion of Malawi's trade is

3 Cost, Insurance and Freight aggregate value
4 The fifteen Member States of SADC include; Angola, Botswana, Democratic Republic of Congo, Lesotho,
Madagascar, Malawi, Mauritius, Mozambique, Namibia, Seychelles, South Africa, Swaziland, Tanzania, Zambia
and Zimbabwe.









conducted within the SADC region. In fact, during the baseline year, Malawi's imports from

SADC accounted for about 58.2 percent of total imports.

The debate on whether Malawi must move forward and implement the EPA agreement

with EU is of little value at this point because such a decision was already made at an ACP level,

even though the country has not moved forward with the implementation of the agreement.

Meanwhile, Malawi's trade with EU countries utilizes the Everything But Arms (EBA) trading

arrangement which is 100 percent unilateral such that the EU can decide to terminate it anytime

that they want to force Malawi along with other ACP least developed countries to migrate to the

EPA arrangement. In such an event, Malawi would be left with just two options, enforcing the

EPA or start utilizing the Generalized System of Preferences (GSP) framework. However, the

preferences under the EPA agreement are far more rewarding to Malawi than those offered under

GSP framework. To this extent, our interest in the current study was to explore alternative ways

through which Malawi could migrate to the EPA framework in the most rewarding manner. As

such, analysis of the import tariff removal along the two broad liberalization options explained

above was intended to provide us with vital information regarding the costs and benefits of a

more encompassing liberalization framework for the country as compared to a lean targeted

liberalization scenario. To this effect, trade creation, trade diversion and fiscal impacts for both

of the two contemplated liberalization scenarios were computed. It was envisaged that based on

such results, the study will derive and describe alternative trade policy options for the

government.

The last bit of data as per the applicable model in this study was with respect to two

crucial variables for the elasticity of import demand and the elasticity of substitution. Knowledge

of these elasticities is an important ingredient in ex-ante analyses of trade policy reforms.









Unfortunately, realistic figures for these two variables at the eight-digit level of data aggregation

are non-existent for most of the developing countries, let alone Malawi. As such, strong

assumptions about the levels of these elasticities were made. First, an extensive survey of the

empirical studies of import demand elasticities was made and one current such study was the one

conducted by Sebastian Vollmer, et al (2009)5. In their study, Vollmer and colleagues employed

empirical trade imports data to estimate the values of import demand elasticities for nine African

countries6. The same data was used to simulate the trade effects of an EPA for each one of the

nine countries. However, the calculations were neither done at each tariff line nor at each HS

chapter but rather blocked into two major product categories including Non-manufactured goods

(HS chapters 01 to 24) and manufactured goods (HS chapters 25 to 97). In this study, we

analyzed the socio-economic realities and trade structure of Malawi vis-a-vis each one of the said

countries. It was concluded that among the nine, Mozambique shares the most similarities with

Malawi. Uganda was found to be next to Mozambique in being similar to Malawi as compared to

the rest of the countries. As such, the import demand elasticities calculated by Vollmer and

friends for Mozambique were adapted for use in this study. The adaptation was largely with

respect to the assumed values of elasticities for the non-manufacturing category, which were

adjusted upwards by 39 percent to take into account the move from the HS-0 to HS-2 category of

the level of aggregation to the HS-3 to HS-8 category of the level of aggregation7. The basis for

the 39 percent adjustment follows the empirical work of Kee, Nicita, and Olarreaga (2004) who

estimated elasticities for different countries at the three and six digit level and found that on


5 Vollmer, et al (2009); EU-ACP Economic Partnership Agreements: Empirical Evidence for Sub-Saharan Africa
6 These Includes the following: Botswana, Cameroon, Cote d'Ivoire, Ghana, Kenya, Mozambique, Namibia,
Tanzania, and Uganda.

7 Vollmer et all (2009) categorized the levels into these two and only calculated the elasticity of Import Demand for
the Non-Manufacturing at the HS 0 to HS 2 levels.









average, the latter estimates were 39 percent higher than the former. The same principle was

applied by Grunbaum (2007) in his dissertation research. A single (and not two) adjustment was

made in line with Vollmer et al's categorization of the HS chapters in their calculations. For

sensitivity analysis, we used both of the values calculated by Vollmer and colleagues, i.e.; the

one for Mozambique and that for Uganda. The major results reported in this study are based on

the Mozambique elasticity value since Malawi is far more similar to Mozambique than Uganda

in terms of their GDP per capital as well as import structure. For the same reason, we included the

mid-level value of import demand elasticity employed by Grunbaum's study.

Since Vollmer et al (2009) did not calculate the levels of elasticity of substitution, we

employed those levels used in Grunbaum's study. This may at first sight appear to be a strong

assumption, but as Vollmer et al (2009) observe, "many of the elasticity estimates across import

categories and across these developing countries have very similar magnitudes"8. We also

employed three levels of magnitudes for this variable to provide useful insights on the sensitivity

of the simulations with respect to such changes. As such, the assumed values of elasticities

employed in this study are as detailed in Table 4-1 below.


Table 4-1. Assumed values of elasticities
Product category Elasticity of Import Demand Elasticity of Substitution
Vollmer study based Grunbaum Grunbaum Study Based
study based
Mozambique Uganda Grunbaum Low Mid High
basis basis Mid
Agricultural goods -1.65271 -0.72558 -0.97 1.39 2.78 4.17
(01-24)
Manufactured & -1.005 -0.684 -1.53 2.50 4.17 4.17
Semi-Man goods
(25- 97)


8 Vollmer et al (2009): EU-ACP Economic Partnership Agreements: Empirical Evidencefor Sub-Saharan Africa; pg
14. World Development Report: Background Paper. University of Gottingen, Germany. pg 14









CHAPTER 5
EMPIRICAL RESULTS

Utilizing the model developed in chapter three and the data described in the previous

chapter, we simulated trade effects and fiscal impacts under two broad liberalization scenarios.

The first of these reflects the outcome of Malawi's possible EPA with the EU while the second

builds on the first by including an FTA with the rest of SADC member states. One major

assumption that has to be pointed out from the onset is with respect to interpretation of the

results vis-a-vis what would realistically be obtainable. Under normal circumstances, the actual

implementation of the proposed EPA between Malawi and the EU would be expected to be

phased out over some considerable time period such that parties to the agreement would be

reducing tariffs on imports on a gradual basis. It is not unusual for this period to take ten years or

more. Besides, even full implementation of such agreements does not necessarily result into 100

percent reduction of existing tariffs on all products as parties tend to retain some level of

protection for some strategic commodities. An examination conducted by the WTO's Committee

on Regional Trade Agreements suggests that FTAs typically cover between 80 and 95 per cent of

the trade between FTA members (WTO 2002; Busse et al 2004). As such, trade effects and fiscal

impacts are also expected to be gradual in that the full impact is only felt after complete

elimination of tariff on imports is attained. The outcome of the simulations in this study assumes

full and immediate elimination of import tariffs by Malawi during the first year of

implementation. The results therefore capture an end-period and/or long term static effect of

complete tariff liberalization through the implementation of the EPA and/or SADC FTA by

Malawi.









Trade Effects

Simulation results indicating the trade effects for the two broad scenarios are presented in

Table 5-1 and Table 5-2 below. As discussed in the previous chapters, values for trade creation,

trade diversion and total trade effects were calculated within the realms of each one of the two

contemplated trade liberalization frameworks. In particular, Table 5-1 presents the trade effects

of an EPA with the EU while maintaining the status quo on all other trading partners. As

highlighted in Table 5-1, the simulated net trade creation values were negative, and significantly

so, for all assumed elasticity levels. This implies that an EPA with the EU alone would result in

significant levels of trade diversion for Malawi, shifting the structure of imports source away

from non-EU trading partners and towards EU member states. The simulations further illustrate

that implementation of the EPA would result in significant increases in EU imports into Malawi,

with the least projected increase of 16.5 percent for the lowest elasticity value and 26.9 percent

for the mid-range of the most favored elasticity value9 in this study. These huge increases are

largely reflective of the significantly high pre-simulation tariff rates on EU imports. As

compared to the baseline total imports value, the simulations demonstrated that ex-post imports

from the EU would increase by percentage levels in the range of 2.4 to 4.7 percent, with an

increase of about 4 percent based on the most favored elasticity value in this study.

Table 5-2 presents the trade effects of implementing a broader trade liberalization policy,

which includes SADC member states by way of implementing a SADC FTA concurrently with

the implementation of the EPA with the EU. As indicated in the table, the absolute values of

simulated trade effects were much higher in this case as compared to the first scenario. Most



9 As indicated in chapter 4, the elasticity values based on Mozambique figures as calculated in the Vollmer, et al
study are recommended in this study. Other values provide useful insights into deducing the sensitivity of the
simulations to changes in elasticity values.









importantly though is the fact that for all but two elasticity scenario sets, trade creation values

were significantly higher than trade diversion figures. As such, for all but those two scenarios,

trade effect simulations resulted in substantially higher positive net trade creation values that

ranged from about US$ 20 million for the low Ugandan based elasticity scenario case to US$ 85

million for the low-case of the Grunbaum elasticity scenario level. The total net trade creation

value for the mid-case of the most favored elasticity value in this study was US$ 27 million. In

terms of expost trade flow effects, the simulations showed that imports from the EU and SADC

region would increase by about 20.2 percent from the baseline year value, based on the mid

scenario case of the study preferred elasticity value. Variation of the elasticity value along the

other scenarios demonstrated that imports would increase in the range of 12.7 to 25.7 percent as

compared to the baseline year import value. As the table further demonstrates, the increases in

imports ranged from about 9.4 percent to 18.9 percent vis-a-vis total imports obtainable during

the baseline year. It is worth noting that large increases in imports were simulated under this

trade liberalization scenario as compared to the case where only EU imports benefited from a

preferential tariff regime. This is largely because of the fact that about half of Malawi imports

come from within the SADC region such that inclusion of the same in tariff liberalization was

found to have significantly encouraged increased intra-region trade.

As insinuated in the previous chapters, economists view trade creation as welfare

improving since consumers in the importing country tend to substitute more costly domestically

produced goods with low cost imports from the beneficiary country. On the other hand, trade

diversion entails displacement of the low cost imports from non-beneficiary trading partners by

the expensive imports from beneficiary countries, which would be EU in our case. As such, and

on the basis of the simulation results describe above, it was noted that an EPA with the EU alone










Table 5- 1. Trade effects from tariff elimination by Malawi on EU imports only
Basis of Import Scenario Trade creation Trade Net trade Total trade Total trade Total trade
Demand Elasticity setting for in US$ diversion in creation in US$ effects in US$ effects as a effects as a
the level of (TC) US$ (TC TD) (TC +TD) percentage percentage
Elasticity of (TD) of preferred of total
Substitution imports imports
Mozambique based Low 12,501,056.99 23,686,280.19 (11,185,223.20) 36,187,337.18 20.51 3.04
Mid 12,501,056.99 34,934,612.13 (22,433,555.14) 47,435,669.12 26.89 3.98
High 12,501,056.99 35,802,198.19 (23,301,141.20) 48,303,255.19 27.38 4.05

Uganda based Low 8,000,182.30 21,112,259.31 (13,112,077.02) 29,112,441.61 16.50 2.44
Mid 8,000,182.30 32,360,591.25 (24,360,408.95) 40,360,773.55 22.88 3.39
High 8,000,182.30 33,228,177.31 (25,227,995.02) 41,228,359.61 23.37 3.46

Grunbaum based Low 17,064,266.00 26,523,566.25 (9,459,300.25) 43,587,832.25 24.71 3.66
Mid 17,064,266.00 37,771,898.18 (20,707,632.18) 54,836,164.19 31.08 4.60
High 17,064,266.00 38,639,484.25 (21,575,218.25) 55,703,750.26 31.58 4.68











Table 5- 2. Trade effects from tariff elimination by Malawi on imports from EU and SADC member states
Basis of Import Scenario Trade creation Trade Net trade Total trade Total trade Total trade
Demand Elasticity setting for in US$ diversion in creation in US$ effects in US$ effects as a effects as a
the level of (TC) US$ (TC TD) (TC +TD) percentage percentage
Elasticity of (TD) of preferred of total
Substitution imports imports
Mozambique based Low 102,763,870.98 51,600,199.21 51,163,671.77 154,364,070.18 17.46 12.84
Mid 102,763,870.98 76,106,790.11 26,657,080.87 178,870,661.09 20.23 14.88
High 102,763,870.98 77,943,951.09 24,819,919.89 180,707,822.07 20.44 15.03

Uganda based Low 66,416,462.37 46,017,395.46 20,399,066.91 112,433,857.84 12.72 9.35
Mid 66,416,462.37 70,523,986.37 (4,107,523.99) 136,940,448.74 15.49 11.39
High 66,416,462.37 72,361,147.35 (5,944,684.97) 138,777,609.72 15.69 11.54

Grunbaum based Low 142,798,799.74 57,824,479.58 84,974,320.15 200,623,279.32 22.69 16.69
Mid 142,798,799.74 82,331,070.49 60,467,729.25 225,129,870.22 25.46 18.73
High 142,798,799.74 84,168,231.47 58,630,568.27 226,967,031.20 25.67 18.88









would result in net welfare losses since trade diversion was greater in all scenarios than the

welfare improving trade creation values. Conversely, an EPA with the EU that is

implemented together with a FTA with the SADC region was seen to be welfare improving

in all but two cases of mid and high assumed elasticity values based on the Ugandan

figures as calculated by Vollmer and colleagues. The significant positive net trade creation

simulation results obtained under the second liberalization framework would lead to

increased overall welfare in Malawi.


Fiscal Impacts
Table 5-3 below captures the contribution of trade taxes in general and import

duties in particular towards total government revenue in Malawi. As indicated, the

contribution of revenue from trade taxes towards total government revenue was about 36.6

percent, inclusive of grants and 40.2 percent excluding grants in the year 2006. During the

same year, import duties accounted for 11.2 percent and 12.5 percent of total government

revenue, including grants and excluding grants, respectively. As such, elimination of tariff

barriers on imports from the EU and/or EU and SADC at the same time is bound to have a

significant impact on government revenue.

Using equation 3-20 developed in chapter three above, fiscal impacts of the two

trade liberalization scenarios were calculated. It is important to point out that simulation

results in this study only quantified the direct and/or static levels of changes in ex-post

revenue as a result of lost revenues due to tariff elimination under the two scenarios. This

part of fiscal impact is clearly outlined as reported in Table 5-4 below. As indicated in

Table 5-4, loss of tariff revenue under the EPA arrangement with the EU alone is

significantly lower compared to the scenario that includes an FTA arrangement with









SADC. The revenue losses under the EU EPA were simulated to range from 11.2 percent to

12.1 percent of the baseline total tariff revenue. In terms of comparison to total government

revenue, the highest of these values was only 4.8 percent of total government revenue

excluding grants and 5.2 percent of total government revenue including grants for the

baseline year.

Table 5- 3. Baseline levels of import duties and impact on government revenue
Total government revenue excluding grants (US$) 489,708,626.38
Total government revenue including grants (US$) 537,861,346.35
Total trade taxes levied 196,594,338.65
Trade taxes as percentage of government revenue excluding grants 40.15
Trade taxes as percentage of government revenue including grants 36.55
Total import duties levied (US$) 61,299,189.69
Import duties as percentage of government revenue excluding grants 12.52
Import duties as percentage of government revenue including grants 11.19

The fiscal impact results tend to increase substantially when the SADC FTA

arrangement is included in the liberalization framework, with percentage reductions of up

to 30.5 percent of total revenue during the baseline year. The fact that imports from SADC

form about half of the total imports for the country helps to explain these drastic changes in

revenue.

It is important to point out that trade creation brings with it other revenue in form of

other trade taxes that are not affected by the tariff elimination policy. Considering the large

values of trade creation under the EU plus SADC FTA tier as described above, one would

expect to get extra revenue that would work towards offsetting the static figures reported

above. As such, the total changes in end-period levels of revenue would be significantly

lower than those reported in Table 5-4 above because of the extra revenue expected from

other trade taxes.









Table 5- 4. Summary of static fiscal impacts
Elasticity scenario Tariff elimination to EU imports Tariff elimination to EU and
setting only SADC imports
As a percent As a percent
Change (US$) in of baseline Change (US$) in of baseline
tariff revenue level tariff revenue level
Mid Mozambique -24,610,959.11 11.6 161,921,417.70 -30.1
Mid Uganda 23,748,488.89 -11.2 159,549,735.08 -29.7
Mid Grunbaum 25,624,080.19 12.1 164,123,067.17 -30.5

Commodities Most Affected by the Contemplated Changes in Trade Policy

One of the major advantages of the Partial Equilibrium Model for simulating the

likely trade effects and fiscal impacts of an EPA and/or an FTA using highly disaggregated

data such as the HS-eight digit utilized in this study is that it allows identification of

commodity categories most affected by the contemplated trade policy change. Knowledge

of such information would be crucial for those engaged in trade negotiations as it would

provide necessary guidance on which product categories are more sensitive to warrant

special treatment during implementation of the EPA framework. In this study, an effort was

made to capture the simulated trade and tariff effects at each individual tariff line along

with their associated HS-code and descriptions. The results column was thereafter sorted a

descending order to identify those product categories most affected by the contemplated

trade policy changes. This permitted us to isolated twenty commodity categories most

affected by the policy change under the two liberalization scenarios were identified. Tables

A-i through A-4 in Appendix A present results of these commodity categories along with

their simulated values for trade effects and tariff revenue changes.

Table A-i and Table A-2 present lists of the most sensitive commodities to an

impending EPA with the EU by Malawi in terms of the trade effects and fiscal impacts,

respectively. As indicated, simulation results demonstrated that some of the most sensitive









products include the following; clothing materials, four-wheel drive motor vehicles, solid

milk and cream products, various iron sheet materials, electrical appliances and telephony

and telegraphic apparatus. The tables further detail the associated values of estimated trade

effects (Table A-i) and those for the estimated changes in tariff revenue (Table A-2). It is

important to note that most of the commodities that were found to have high trade effects

levels were also identified to have high absolute tariff revenue changes.

The other two tables, Table A-3 and Table A-4 present the same information as the

one in Tables A-i and Table A-2 but for the second liberalization scenario that includes a

FTA with the rest of the SADC member states. Under this second scenario, some of the

product categories with highest trade effects include; distillates and other fuels, four wheel

drive motor vehicles, rubber tires used on busses and lorries, worn clothing materials, flat-

rolled iron and steel products, crude soybean oil, cement clinkers and tobacco cigarettes.

Associated values of trade effects are detailed in the tables.

Three Earmarked Agricultural Products

One of the major objectives of this study was to look at the trade effects and fiscal

impacts of the EPA with EU on key agricultural commodities. The three agricultural

products of tobacco, tea and sugar were earmarked for such a close evaluation of the extent

to which trade policy change affects their ex-post trade flow levels and tariff revenues.

Table 5-5 presents the impact of an EPA with EU on tea, sugar and tobacco trade flows and

tariff revenue contributions. Table 5-6 presents the same information for a broader

liberalization scenario that includes the SADC FTA. For each liberalization scenario, all

the disaggregated trade and fiscal effects at every individual tariff line relating to each one

of the three targeted agricultural products were isolated and aggregated into a product total

effect as captured in Table 5-5 below. These simulations utilized the same elasticities that









were employed in the entire study for all tariff lines relating to agricultural products from

HS- chapters 01 through 24. The ideal way was to use product specific elasticities of

import demand and substitution. These were however not available for Malawi, forcing us

to employ the sector wide elasticities.


Table 5- 5. Impacts under the EU liberalization on three key agricultural commodities
Commodity Tariff revenue As percent of Avg trade As percent of
change (US$) total effects (US$) total
Tea -69,497 -0.28 125,806 0.26
Sugar -8,882 -0.04 27,498 0.06
Tobacco 0 N/A 0 N/A

The basis for the choice of these products was two-fold. First, pundits in most of the

ACP countries in general and Malawi in particular have argued that an EPA with the EU

would invariably lead to suffocation of the agricultural sectors in the ACP countries which

is the major source of livelihood and economic development in these countries. Second, the

three agricultural commodities form the bulk of Malawi exports, with tobacco exports

accounting for over 60 percent of export earnings. As such, investigating the sensitivity of

these products to policy changes was deemed crucial.

Table 5- 6. Impacts under the EU & SADC liberalization on three key agricultural
commodities
Commodity Tariff revenue As percent of Avg trade As percent of
change (US$) total effects (US$) total
Tea -324,160 -0.20 365,706 0.20
Sugar -312,977 -0.19 709,613 0.40
Tobacco 2,642,279 1.63 2,361,947 1.32

As it turns out, simulation results demonstrate that trade effects and fiscal impacts

on tea and sugar were found to be quite insignificant under both liberalization frameworks.

Tobacco was found to have some modestly high trade effects and revenue changes only

under the second liberalization framework. Under the EPA arrangement with EU, no









effects were found. Zero effects were estimated under the EPA arrangement because all of

the tobacco imports into Malawi come from non-EU countries of Mozambique, Zimbabwe,

Zambia, Kenya, Mauritius and the United Arab Emirates.

The empirical results presented above do not support the long-standing fear among

policy makers and trade specialists in ACP countries that an EPA with the EU is bound to

stifle the local agricultural sector with an influx of cheap EU competing agricultural

imports. In fact, of the twenty commodities identified to be most affected by the EU EPA,

only the two product categories of milk and cream in solid form (both HS 04021000 and

HS 04022100) were part of the list, being projected to increase in ex-post imports from the

EU by 2.7 and 2.3 percent, respectively, vis-a-vis baseline levels. Under the EPA plus

SADC FTA framework, crude soybean oil (HS-15071000), milk and cream in solid form

(HS 04022100), crude palm oil (HS-15111000) were the only agricultural products within

the list of most affected product categories in terms of projected increases in imports.









CHAPTER 6
SUMMARY AND CONCLUSIONS

Summary

Malawi has been party to the EU-ACP trade agreements since the first Lome

Agreement in 1975. Meanwhile, these ACP countries are expected to implement the 2000

Cotonou Agreement that calls for migration away from the non-reciprocal trading

framework to a reciprocal trading arrangement between the two parties. As such, Malawi

along with other ACP group of countries is expected to liberalize their trade through

elimination of tariffs on EU imports. This has spurred concerns among ACP countries who

fear that such an arrangement will subject domestic producers to unfair competition from

the highly subsidized agricultural imports from the EU for instance. As such, it is generally

feared that these changes may bring about increased unemployment, provoke heightened

economic insecurity and political instability. In most cases, these fears have not been

substantiated with empirical evidence as very limited empirical country-specific studies

have so far been done to quantify the potential effects.

It is on the basis of the foregoing that the major objective of this research was to

conduct a quantitative analysis of the likely trade effects and fiscal impacts of an EPA trade

arrangement regime with Malawi. It was deemed crucial to quantify potential levels of

trade creation, trade diversion and fiscal impacts to the country emanating from an EPA

with the EU. Ancillary to this effort was the need to explore alternative trade policy

arrangements, particularly the broadening of the liberalization framework. We further

undertook to identify those commodities that would be most affected by the contemplated

policy changes in trade.









In order to enrich our basis for the empirical study, an extensive survey of the

requisite literature on the theoretical and empirical works related to the subject matter was

conducted. It was noted that classical trade theory argues for a positive correlation between

free trade policy and increased economic growth and development through direct dynamic

advantages of increased capacity utilization and more efficient investment projects as well

as indirect effects of accelerated export growth. It was observed however that other

researchers object to the classical perspective and argue that empirical evidence does not

strongly support such a view. These researchers also question the validity of the direction

of the casual relationship between export growth and economic development as purported

by the classical theorists. In spite of the arguments and counter-arguments among

researchers on the existence and/or strength of the direct link between open trade and

positive economic development, it was observed that open trade does bring with it elements

that spur competition among producers and imposes rationality in productive resource

allocation, thereby increasing economic growth and development. On the other hand

however, unambiguous evidence of poor economic performance in closed regimes was

observed and noted.

The evolution of the notion of preferential trading agreements was noted to have

emanated from the frustration during the 1990s by many countries to attain multilateral free

trade. As such, countries started configuring themselves into regional groupings to pursue

regional free trade policies. This influenced economists, led by Jacob Viner (1950) to start

studying and laying the earliest basis on the theory of regional free trade agreements. Viner

(1950) observed that welfare effects from any form of regional economic integration are

not unambiguous a priori. He further developed the concepts of trade creation and trade









diversion associated with the formation of free trade areas. In this study, these two concepts

were central to the analyses and conclusions to be made. Viner's work spurred a lot of

theoretical as well as empirical work on the study of free trade areas.

Literature identifies three major quantitative analytical techniques that researchers

employ to conduct empirical analyses of the effects of entering some form of an economic

integration by countries. These include the gravity models, computable general equilibrium

(CGE) models and the partial equilibrium (PE) models. Strengths and weaknesses for each

one of these models were highlighted. This thesis utilized the Verdoorn (1960) version of

the PE model to quantify the trade effects and fiscal impacts of the EU EPA with Malawi.

Central to the fears among developing countries to enter into free trade regimes are

the fiscal implications of such policies. Most of these countries fear that through

elimination of import tariffs, the government would experience huge revenue losses and

compromise the delivery of vital services to the populace. Consideration of several short

term policy migration costs to the local economies such as increased unemployment,

reduction in national output, elimination of certain domestic industries and possible

macroeconomic instability was observed to have exacerbated these fears. Research into

this aspect has been centered on two themes, including the evaluation of the relative

importance of tariff revenue with respect to total government revenue and exploration of

the fiscal reforms and alternative tax schemes that could offset tariff revenue losses.

Conclusions

The Verdoorn (1960) version of the partial equilibrium models was employed to

estimate and evaluate the trade effects and fiscal impacts of an EU EPA on the Malawi

economy. This model is based on an array of rather restrictive assumptions such as the

Armington assumption of product differentiation, and those applicable to most partial









equilibrium analyses including; non-existence of repercussions on the incomes and

exchange rates from changes in trade flows; existence of iso-elastic import demand

elasticities; and infinite supply elasticities. Notwithstanding these restrictive assumptions, it

was noted that this model has been extensively applied in similar works by various

researchers and academicians alike. In this study, highly disaggregated trade data at HS-

eight digit were utilized to conduct the analyses.

The major conclusion that was drawn from this study is that Malawi stands to

benefit more from trade if a broader liberalization framework is adopted along with the

EPA enforcement. As reported above, trade liberalization of tariffs towards EU and SADC

result in significantly high positive net trade creation levels for all but two elasticity values.

On the other hand, liberalization that is restricted to EU imports results in negative net

trade creation values. This conclusion stands firm even when one considers the major

weakness of this study, namely; that the values of elasticities employed were quite

arbitrary. This is the case because even as we varied the magnitude of the elasticity values,

a broader liberalization framework resulted in higher positive levels of net trade creation

values as compared to the negative net trade creation values obtained under the

liberalization framework restricted to the EU EPA. The breadth of liberalization need not

necessarily be restricted to EU and SADC countries. There are increasingly high levels of

imports from the eastern countries of Japan, China and India. Inclusion of these countries

in the tariff liberalization framework is bound to result in increased values of trade creation.

While not supporting the conventional fears about trade liberalization in most of the

developing world, the results from this research do conform to the theoretical arguments

for broader liberalization even for small countries like Malawi.









Similar results were obtained by Grunbaum (2007) who found that when

liberalization by the Organization of Eastern Caribbean States (OECS) with the EU alone

was considered, trade diversion exceeded trade creation in all cases and overall trade

effects and fiscal impacts were very small. When the larger Caribbean, the North American

Free Trade Area (NAFTA), and the Free Trade Area of the Americas (FTAA) were all

included in the liberalization framework, net trade effects were positive and larger, as were

the fiscal impacts.

Our results also agree with those obtained by Keck and Piermartini (2005) who

used an applied general equilibrium model covering 15 regions and 9 sectors to simulate

the impact of signing EPAs with the EU for the SADC countries under various

liberalization scenarios. The results demonstrated that EPAs with the EU were welfare-

enhancing for SADC overall, and that for most countries, further gains could arise from

intra-SADC liberalization.

Greenway and Milner (2003) also found comparable results when they analyzed the

impact of EPA formation for CARICOM countries. Employing two-digit level of data

aggregation, Greenway and Milner (2003) contemplated three liberalization frameworks in

their study including the one restricted to the EU, the second combining the EU and the

United States and the last one being full multilateral liberalization. Their results concluded

that the EPA formation with the EU resulted into net welfare losses by the CARICOM

countries involved where as the one under the last two scenarios resulted in positive

welfare gains with the full multilateral liberalization resulting in the highest welfare gains.

The other major conclusion of this study is with respect to the effects of EPA

formation on the domestic agricultural sector, mostly the fear of an influx of cheap EU









imports. Against these conventional fears, results from this research demonstrate that very

minimal effects on agricultural imports from the EU would be felt, save for imports of milk

and cream in solid form. There were particularly no major effects on imports of tea, sugar

and tobacco, which are currently the major agricultural export commodities for Malawi.

The fiscal impacts from trade liberalization do seem be significant if considered in

isolation. However, the simulation results only capture static effects at 100 percent tariff

elimination scenario. With the high positive trade creation values obtained from the

simulation results, one would expect a significant offsetting effect if dynamic effects of

extra revenue from increased trade are included. Besides, policy makers may consider

exploring other tax systems that will help to broaden the country's tax base. In fact this

subject has been at the centre of discussions within the country, considering the high levels

of dependence on trade taxes for government revenue under the current system.

This research has laid down a very useful foundation into research on trade effects

and fiscal impacts of trade liberalization at a country specific level that employs

disaggregated data at the level utilized herein. Until this work, we do not know of any

effort directed towards empirical quantification of such effects in the country. The results

from this thesis will provide useful guidance to the on-going debate with respect to the pros

and cons of the EU-EPA formation. The identification of products most affected by the

policy change will exonerate fears of unfair competition from EU agricultural imports

while alerting those engaged in the most affected products to explore remedial measures.

Study Weaknesses and Suggestions for Further Research

One of the major weaknesses of the Partial Equilibrium models and particularly the

one carried over into this thesis is with respect to the choice of elasticities of import

demand and substitution. These are very crucial variables in the Verdoorn (1960) model









and yet were strong assumptions were made in the process of determining their values

since they were not available at any disaggregated level. The values chosen were quite

reasonable considering that most of the previous researchers have not employed similarly

adapted values. Nonetheless, the ideal scenario is to use the study data to estimate elasticity

values at the same level of data aggregation, preferably every tariff line or HS chapter

applicable. Through this rigorous process one would get empirically deduced product

specific elasticity values to employ in the analysis of trade and fiscal effects of any form of

trade liberalization that might be contemplated. Notwithstanding its enormous financial and

technical demands, such research is highly recommended for Malawi as the results from

the same would help solidify the foundation laid in this research. In particular, the levels of

trade and fiscal effects simulated for each product category such as those relating to the

three agricultural products of tea, sugar and tobacco reported above would be much more

realistic if one employs such product specific elasticities. In terms of data availability, the

country's National Statistical Office has made significant strides in building requisite

databases. It would also be important to build in these studies an aspect of dynamic effects

of liberalization by way of employing the computable general equilibrium models, which

would take into account economy-wide repercussions of the policy shift and adjust the

trade and fiscal effects accordingly.










APPENDIX
MOST AFFECTED PRODUCTS BY TRADE AND TARIFF REVENUE IMPACTS

Table A- 1. List of commodities with the highest trade effects as a result of tariff
elimination on EU imports
Average As
Number HS code Item description trade effects percent of
(US$) total


1 63090000 Worn Clothing and Other Worn Articles
2 85175000 Apparatus for Carrier-Current line System or for
Digital line
3 87032310 Other Motor Vehicle Four Wheel Drive
4 72104900 Flatrolled Iron/Steel, Width >=600MM,Zinc
Plated/Coated(EXC.ELECT
5 04021000 Milk and Cream in Solid Forms Of =<1.5% FAT
6 85173000 Telephonic or Telegraphic Switching Apparatus
7 04022100 Milk and Cream in Solid Forms OF >1.5% FAT,
Unsweetened
8 87033390 Other (Four-Wheel Drives)
9 34012000 Soap in Other Forms, NES
10 63079000 Made up Articles (INCL. Dress Patterns), NES
11 49119990 Other Printed Matter N.E. S
12 85252090 Other Transmission Apparatus
13 87033310 Four-Wheel Drive Vehicle With Diesel Engine Of CC
>=2500CC
14 87032331 Other Vehicles With Compression-Ignition Internal
Combusion DI
15 85252020 Communication Transmitters, Transceivers and
Ancillary Apparatus
16 87033210 Other Vehicles Of Cylinder Capacity Not Exceeding
2000CC Four
17 85178000 Electrical Apparatus For Line Telephony or Line
Telegraphy, NE
18 87089990 Other Specialized Parts For Tractors
19 87032321 Four-Wheel Drive Vehicle with Spark-Ignition Engine
Of CC 1500-3000CC
20 85042100 Liquid Dielectric Transformers, Power Handling
Capacity =<650K


2,559,500.98 5.82


1,429,916.75
1,366,699.75

1,359,846.10
1,179,355.59
1,046,776.36
997,329.54

994,556.33
943,934.71
760,917.67
633,467.28
628,227.38
569,137.68

558,099.76

489,950.36

480,025.02

457,214.30

415,642.43
394,068.76

393,063.54











Table A- 2. List and levels of commodities with highest tariff revenue changes as a result
of tariff elimination on EU imports
Number HS code Item description Change in tariff As percent of
revenue (US$) total


1 63090000 Worn Clothing and Other Worn Articles
2 87032310 Other Motor Vehicle Four Wheel Drive
Apparatus for Carrier-Current line System
3 85175000 or for Digital line
Milk and Cream in Solid Forms OF >1.5%
4 04022100 FAT, Unsweetened
5 49119990 Other Printed Matter N.E.S
6 87033390 Other (Four-Wheel Drives)
Flatrolled Iron/Steel,Wid.>=600MM,Zinc
7 72104900 Plated/Coated(EXC.ELECT
Other Vehicles With Compression-Ignition
8 87032331 Internal Combusion DI
Other Vehicles Of Cylinder Capacity Not
9 87033210 Exceeding 2000CC Four
10 34012000 Soap in Other Forms, NES
11 85252090 Other Transmission Apparatus
Office Machines, NES(INCL. Coin-
12 84729000 Sorting/Counting/Wrapping Mach
13 87089990 Other Specialised Parts for Tractors
Four-Wheel Drive Vehicles with Spark-
14 87032321 Ignition Engine Of CC 1500-3000CC
Vehicles with Spark-Ignition Engine of
15 87032400 Cylinder Capacity >=300
16 21021000 Active Yeasts
Four-Wheel Drive Vehicle with Diesel
17 87033310 Engine Of CC >=2500CC
Communication Transmitters, Transceivers
18 85252020 and Ancillary Apparatus
19 48115100 Bleached Weighing More Than 150g/m2
Generating Sets with Compression Ignition
20 85021100 Engines, =<75 KVA


2,473,559.29
1,456,770.59


10.05
5.92

5.66

4.36
3.24
2.90


1,393,168.88

1,073,772.67
798,375.10
713,339.23

692,189.40

543,359.65

499,694.83
485,323.56
455,135.26

447,812.58
375,652.93

365,959.29

360,275.74
338,056.80

335,065.24

326,255.29
290,088.29

286,270.84


2.03
1.97
1.85











Table A- 3. List of commodities with the highest trade effects as a result of tariff
elimination on imports from EU & SADC counties
Average trade As percent


Nu


effects (US$) of total
19,329,820.98 11.28


imber HS code Item description
1 27101199 Distillates and other fuels nes(including
diesel oils,gas oil
2 87033310 Four-Wheel Drive Vehicles with Diesel
Engine Of CC >=2500CC
3 40112000 New Pneumatic Tyres, Of Rubber Of a Kind
Used on Buses or Lorries
4 63090000 Worn Clothing and Other Worn Articles
5 72104900 Flatrolled Iron/Steel, WID.>=600MM,Zinc
Plated/Coated(EXC.ELECT
6 87032310 Other Motor Vehicles Four Wheel Drive
7 87033210 Other Vehicles Of Cylinder Capacity Not
Exceeding 2000CC FOUR
8 15071000 Crude Soya-Bean Oil
9 25231000 Cement Clinkers
10 87042110 GVW Not Exceeding 2.99 Tonnes
11 27101919 Gases and Lubricating Oils
12 39011000 Polyethylene Having a Specific Gravity
<0.94, In Primary Forms
13 24022000 Cigarettes Containing Tobacco
14 34012000 Soap In Other Forms, NES
15 04022100 Milk and Cream in Solid Forms Of >1.5%
FAT, Unsweetened
16 85252090 Other Transmission Apparatus
17 15111000 Crude Palm Oil
18 87029020 Other Motor Vehicles for the Transport of
Ten or More Persons
19 85175000 Apparatus for Carrier-Current Line Systems
or For Digital Line
20 87033390 Other (Four-Wheel Drives)


3.00

2.35

2.13
1.94


5,138,646.05

4,024,071.99

3,654,699.51
3,320,894.25

3,317,019.91
3,219,850.34

2,984,694.08
2,274,888.34
2,241,255.55
2,126,549.46
2,066,537.93

1,928,517.27
1,728,086.55
1,712,645.59

1,662,007.39
1,650,016.29
1,644,461.20

1,598,042.62

1,426,783.77


0.97
0.96
0.96

0.93

0.83











Table A- 4. List and levels of commodities with highest tariff revenue changes as a result
of tariff elimination on imports from EU & SADC countries
As
Number HS Code Item Description Change in tariff percent
revenue (US$) of total


Distillates and other fuels nes(including
1 27101199 diesel oils, gas oil
Four-Wheel Drive Vehicles with Diesel
2 87033310 Engine Of CC >=2500CC
Other Vehicles with Cylinder Capacity Not
3 87033210 Exceeding 2000CC FOUR
4 87032310 Other Motor Vehicles Four Wheel Drive
New Pneumatic Tyres, Of Rubber of a Kind
5 40112000 Used on Buses or Lorries
6 63090000 Worn Clothing and Other Worn Articles
7 25231000 Cement Clinkers
8 27101919 Gases and Lubricating Oils
9 24022000 Cigarettes Containing Tobacco
10 15071000 Crude Soya- Bean Oil
11 87042110 GVW Not Exceeding 2.99 TONNES
Dyed Plain Weave Fabrics, <85% Polyester
12 55142100 Fibres + Cotton, >170
Flatrolled Iron/Steel,WID.>=600MM,Zinc
13 72104900 Plated/Coated(EXC.ELECT
Four-Wheel Drive Vehicles with Spark-
14 87032321 Ignition Engine Of CC 1500-3000CC
Apparatus for Carrier-Current Line Systems
15 85175000 or for Digital Line
16 87033390 Other (Four-Wheel Drives)
Dyed 3 or 4 Thread Twill (Inc. Cross
17 52093200 Twill), With >=85% COTTON
18 27101159 Other Kerosene
Other Motor Vehicles for the Transport of
19 87029020 Ten or More Persons
Mixtures/With Basis of/Odorifer's Substitutes
20 33021000 Incl. Alcohol Solutions


26,528,783.18

8,203,313.08

3,545,208.91
3,478,076.27

2,970,721.26
2,700,024.24
2,659,825.23
2,550,591.47
2,348,131.82
2,089,339.84
2,007,778.07

1,882,079.78

1,784,902.09

1,735,524.63

1,630,463.53
1,469,126.14

1,449,647.51
1,402,972.46

1,309,333.56

1,286,655.85


16.38

5.07

2.19
2.15

1.83
1.67
1.64
1.58
1.45
1.29
1.24

1.16

1.10

1.07

1.01
0.91

0.90
0.87

0.81

0.79









LIST OF REREFENCES


Andriamananjala, S; P. Brenton; J.E. Von Uexkill and P. Walkenhorst. 2009. Assessing the
Economic Impacts of an Economic Partnership Agreement on Nigeria. World Bank
African Region Policy Research Working Paper, Washington, D.C.

Balassa, B. 1971. The structure ofprotection in developing countries. Baltimore, MD: Johns
Hopkins University Press.

1982. Development strategies in semi-industrial economies. NY and London: Oxford
University Press.

Baldwin, R.E. and T. Murray. 1977. "MFN Tariff Reductions and Developing Country Trade
Benefits Under the GSP." The Economic Journal 87(345):30-46.

Berg, A. and A. Krueger. 2003. "Trade, Growth, and Poverty: A Selective Survey." IMF
Working Paper no. 03/30, International Monetary Fund, Washington, D.C.

Bhagwati, J. 1978. Anatomy and consequences of exchange control regimes. Cambridge, MA:
NBER.

Bhagwati, J. and A. Panagariya. 1996. "Preferential Trading Areas and Multilateralism-Strangers
Friends, or Foes?" In J. Bhagwati and A. Panagariya,.eds. The Economics of Preferential
Trade Agreements. Washington D.C: American Enterprise Institute, pp.1-78.

1999. "Preferential Trading Areas and Multilateralism-Strangers, Friends or Foes?"
In J. N. Bhagwati, et al., eds. Trading Blocs: Alternative Approaches to Analyzing
Preferential Trading Arrangements. Cambridge, Massachusetts, MIT Press.

Busse, M; A. Borrmann and H. Grossmann. 2004. The Impact of ACP/EU Economic Partnership
Agreement on ECOWAS Countries: An Empirical Analysis of the Trade and Budget
Effects. Study Final Report to the Hamburg Institute of International Economics;
Hamburg, Germany.

Cooper, C.A. and B.F. Massell. 1965a. "A New Look at Customs Union Theory." Economic
Journal 75(300):742-747.

1965b. "Toward a General Theory of Customs Unions for Developing Countries."
Journal of Political Economy 73(5):461-476.

Corden, W.M. 1972. "Economies of Scale and Customs Unions Theory." Journal of Political
Economy 80(3, Part 1):465-75.

Dee P. and J. Gali. 2003. "The Trade and Investment Effects of Preferential Trading
Agreements" NBER Working Paper No. 10160.









Development Policy Research Unit of the University of Cape Town. 2001. What are the potential
benefits and pitfalls of a free trade area in the Southern African region? DPRU Policy
Brief No. 01/P8. Cape Town, South Africa

Dollar, D. 1992. "Outward-Oriented Developing Economies Really Do Grow More Rapidly:
Evidence form 95 LDCs, 1976-1985." Economic Development and Cultural Change
40(3):523-544.

Edwards, S. 1993. "Openness, Trade Liberalization, and Growth in Developing Countries."
Journal of Economic Literature 31(3): 1358-1393. Web accessed:
http://www.jstor.org/stable/2728244. June, 26th 2009.

El-Agraa, A.M. 1997. Economic integration worldwide. New York, St. Martin's Press.

Europe Website. 2001 Development and Relations n i/h African, Caribbean and Pacific States;
(Accessed in July, 2009)
http://ec.europa.eu/development/geographical/cotonou/cotonou20001_en.cfm

Frankel, J.A. and D. Romer. 1999. "Does Trade Cause Growth?" The American Economic
Review 89(3):379-399.

Frankel, J. A. 1997. Regional Trading Blocs, Washington, D. C.: Institute for International
Economics.

Gordon, M.G. 2007. Impact of the CARICOM Economic Integration Arrangements on the
Economic Gains of Selected CARICOM Countries. PhD dissertation Report

Greenway, D. and C. Milner. 2003. "A Grim REPA?" Research Paper No.2003/30, Leverhulme
Centre on Globalization and Economic Policy, University of Nottingham.

Grunbaum, Max. 2007. "Evaluation of Trade and Tariff Revenue Effects for the Organization of
Eastern Caribbean States under Alternative Trade Liberalization Scenarios". PhD
dissertation Report.

Harrison, A. 1996. "Openness and Growth: A Times-Series, Cross-Country Analysis for
Developing Countries." Journal of Development Economics 48:419-447.

Head, K. 2003. "Gravity for Beginners." Faculty of Commerce, University of British Columbia.
http://www.econ.ubc.ca/nmalhotra/490/Articles/KHead20on%20gravity.pdf July 6th
2008.

Johnson, H.G. 1960. "The Economic Theory of Customs Union." Pakistan Economic Journal
X(1): 14-32.

Keck, A. and R. Piermartini. 2005. The Economic Impact of EPAs in SADC Countries. WTO,
Economic Research and Statistics Division Working Paper, Geneva, Switzerland









Kerkala, L., J. Niemi, and R. Vaittinen. 2000. "Renegotiating the Lome Convention-Trade Policy
Schemes and Their Effects for African Regions." Helsinski School of Economics and
Business Administration.

Khattry, B. and J. M. Rao. 2002. "Fiscal Faux pas?: An Analysis of the Revenue Implications of
Trade Liberalization." World Development 30(8): 1431-1444.

Krueger, A.O. 1978. Foreign trade regimes and economic development: Liberalization attempts
and consequences. Cambridge, MA: NBER.

Lal, D. and S. Rajapatirana. 1987. "Foreign Trade Regimes and Economic Growth in Developing
Countries," World Bank Res. Observer, July 1987, 2(2), pp. 189-217.

Little, I., T. Scitovsky, and M. Scott. 1970. Industry and trade in some developing countries.
London & NY. Oxford University Press for OECD.

Lipsey, R.G. 1960. "The Theory of Customs Unions: A General Survey. In R. Pomfret ed.
Economic Analysis of Regional Trading Arrangements. Massachusetts: Edward Elgar
Publishing Ltd.

Malawi Government. 2008. Annual Economic Report 2008: Government Printer, Lilongwe,
Malawi

Meade, J.E. 1955. The Theory of Customs Unions. Amsterdam: North-Holland.

Michaely, M., D. Papageorgiou, and A. Choksi. 1991. "Liberalizing foreign trade: Lessons of
experience in the developing world". Vol. 7. Oxford and Cambridge, MA: Basil
Blackwell.

Mundell, R.A. 1964. "Tariff Preferences and the Terms of Trade." Manchester School of
Economic and Social Studies XXXII(1): 1-13.

Nilson, L. 2002. "Trading Relations: Is the Roadmap form Lome to Cotonou Correct?"
AppliedEconomics 34(4):439-453.

Partnership Agreement Between the Members of the African, Caribbean and Pacific Group of
States of the one Part, and the European Community and its Member States, of the Other
Part, Signed in Cotonou, Benin on 23rd June, 2000.
http://www.acpsec.org/en/conventions/cotonou/accordl.htm

Piermartini, R. and R. Teh 2005. Demystifying Modelling Methods for Trade Policy. Discussion
Paper no. 10, World Trade Organization, Geneva.

Pomfret, R. 2003. Economic Analysis of Regional Trading Arrangements. Massachusetts:
Edward Elgar Publishing Ltd.

Robson, P. 1980. The Economics ofInternational Integration. London: George Allen & Unwin.









Rodriguez, F. and D. Rodrik 2000. "Trade Policy and Economic Growth: A Skeptic's Guide to
Cross-National Evidence. In B. Bernanke and K. Rogoff. eds. Macroeconomics Annual
2000. Cambridge, MA, NBER. pp.261-325.

Sachs, J.D., A. Warner, A. Aslund and S. Fischer. 1995. "Economic Reform and the Process of
Global Integration." Brookings Papers on Economic Activity 1995(1): 1-118.

Sawyer, W.C. and R.L. Sprinkle 1989. "Alternative Empirical Estimates of Trade Creation and
Trade Diversion: A Comparison of the Baldwin-Murray and Verdoorn Models."
Weltwirtschaftliches Archiv 126:61-73.

Soloaga, I. and L.A. Winters. 2001. "Regionalism in the Nineties: What Effect on Trade?" North
American Journal of Economics and Finance 12, 1: 1-29.

Streeten, P. "Review of Little, et al.," Econ. J., Mar. 1971, 81(1), pp. 144-48.

Tinbergen, J. 1962. .\, qping the World Economy: Suggestions for an International Economic
Policy. New York: The Twentieth Century Fund.

Verdoorn, P.J. 1960. The Intra-Block Trade of Benelux. In E. A. G. Robinson ed. Economic
Consequences of the Size ofNations. London: Macmillan & Co.

Viner, J. 1950. The Customs Union Issue, in R. Pomfret ed. Economic Analysis of Regional
Trading Arrangements. Massachusetts: Edward Elgar Publishing Ltd.

Vollmer, S.; I. Martinez-Zarzoso; F.D. Nowak-Lehmann; and N.H. Klann. 2009. EU-ACP
Economic Partnership Agreements: Empirical Evidence for Sub-Saharan Africa. World
Development Report 2009 Background Paper, University of Gottingen, Gottingen,
Germany.

Weisbrot, M. and D. Baker. 2002. The Relative Impact of Trade Liberalization on Developing
Countries. Center for Economic Policy Research Briefing Paper, Washington, D.C.

Wolf, S. 2002. "From Preferences to Reciprocity General Equilibrium Effects of a Free Trade
Area Between The EU and the UEMOA." Third Annual Conference on Global Economic
Analysis, Bonn, Germany.

WTO (2002), Committee on Regional Trade Agreements, Coverage, Liberalization Process and
Transitional Provisions in Regional Trade Agreements, WTO Document.
WT/REG/W/46, June 2004, Internet Posting: (Accessed in May, 2009)

http://www.wto.org/english/tratope/regione/regcome.htm.









BIOGRAPHICAL SKETCH

Innocent Lwafyo Thindwa was born in Karonga, Malawi on 4th August, 1977. He

received a Bachelor of Social Science, majoring in economics, from the University of Malawi,

Chancellor College in 2003. In that same year, Innocent joined the Malawi Government's

Economic Common Service, working as an Economist. In March 2007, he was promoted to the

position of Senior Economist before a further promotion to the position of Principal Economist

in June, 2007.

In August 2008, he joined the Food and Resources Economics Department of the

University of Florida to pursue a Master of Science degree program under the sponsorship of the

United States Agency for International Development (USAID). He completed his degree in

August 2010.





PAGE 1

THE EU-ACP EC ONOMIC PARTNERSHIP AGREEMENTS (EPA) FOR MALAWI: TRADE EFFECTS AND FISCAL IMPACTS FOR ALTERNATIVE POLICY OPTIONS By INNOCENT LWAFYO THINDWA A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2010 1

PAGE 2

2010 Innocent L. Thindwa 2

PAGE 3

To m y late mother, I dedicate this piece of work with great humility as I reflect on her untiring efforts to see me educated even in the face of glaring challenge s within an apparently harsh environment as she ploughed on to see me through the journey that not even herself knew the destination. How much I long to have had her alive this day to witness the fruits of her diligence and commitment to a cause noble enough as I bare testimony today. May Her Soul Rest In Eternal Peace!! 3

PAGE 4

ACKNOW LEDGEMENTS I benefited from the contributions of many pe ople in the course of my studies at the University of Florida and thr oughout this thesis project. Fore most among these is my major professor and academic advisor, Dr. John J. VanSickle, who from the very start of my studies provided me with invaluable moral and academic support to the level I never envisaged. Dr. VanSickle was particularly instru mental in providing me with th e required momentum to finish the research project when I was back in Malawi and fraught with several family responsibilities that might otherwise have compromised my comple tion of the project. He physically visited me in March, 2010 to fulfill this noble cause, afford ing me constructive advisory comments on the written drafts and handy techniques of handling the project data. I do really appreciate his contribution, without which this thesis would not have been possible. I also wish to commend Dr. Rick Weldon, who demonstrated to be a very useful committee member of my project. I am also gratef ul to Dr. James Sterns, who without much ado spared time to sit in my defense far in Malawi. Dr. Walter Bowen dedicated himself as a very compassionate and helpful administrator of the USAID program that provided financial support for my studies at the University of Florida. Throughout my studies, Walter remain ed ready and committed to listen and offer solutions to any challenges that those of us s upported by the program could bring to his attention. The financial support provided by the USAID is, to say the least, grea tly appreciated. I am also indebted to my Malawian collea gues; Fiskani, Lucy, Pearson, Bonnet, Jonathan and Davie who we shared part of our stay togeth er at the University of Florida. The moral and academic support I got from Fiskani, Lucy, Pearson and Bonnet during the course work is greatly appreciated. 4

PAGE 5

I also thank GOD for his continual guidance and for having maintained my incredibly loving fiance Alice Mbisa with patience and the spirit of continued moral support to me throughout my studies. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGEMENTS.............................................................................................................4LIST OF TABLES................................................................................................................. ..........8LIST OF ABBREVIATIONS.......................................................................................................... 9ABSTRACT...................................................................................................................................12 CHAPTER 1 INTRODUCTION................................................................................................................. .14Problematic Situation.......................................................................................................... ....16Problem Statement.............................................................................................................. ....18Research Objectives............................................................................................................ ....202 LITERATURE REVIEW.......................................................................................................21Liberalization versus Protectionism and Economic Performance..........................................22Development and Growth of the Theory of Preferential Trade Agreements.........................27Methodologies for the Empirical Analysis of the Impacts of Preferential Trade Agreements..................................................................................................................... ....32Gravity Models................................................................................................................3 2Empirical Application of the Gravity Model...................................................................35Computable General Equilibrium Models (CGE)...........................................................37Application of CGE Models in Empirical Work.............................................................39Partial Equilibrium Model (PE).......................................................................................41Overview of the Verdoorn Model...................................................................................42Empirical Applicati on of PE Models..............................................................................44Analysis of Fiscal Impacts......................................................................................................453 MODEL DEVELOPMENT....................................................................................................484 DATA REQUIREMENTS, SOURCES AND TREATMENT..............................................545 EMPIRICAL RESULTS........................................................................................................59Trade Effects...........................................................................................................................60Commodities Most Affected by the Cont emplated Changes in Trade Policy........................66Three Earmarked Agricultural Products.................................................................................67 6

PAGE 7

6 SUMMARY AND CONCLUSIONS.....................................................................................70Summary.................................................................................................................................70Conclusions.............................................................................................................................72Study Weaknesses and Suggestions for Further Research.....................................................75APPENDIX: MOST AFFECTED PRODUCTS BY TRADE AND TARIFF REVENUE IMPACTS...............................................................................................................................77LIST OF REREFENCES...............................................................................................................81BIOGRAPHICAL SKETCH.........................................................................................................85 7

PAGE 8

LIST OF TABLES Table page 41 Assumed Values of Elasticities..........................................................................................5851 Trade Effects from Tariff Elimination by Malawi on EU Imports Only...........................6252 Trade Effects from Tariff Eliminati on by Malawi on Imports from EU and SADC Member States.................................................................................................................. .6353 Baseline Levels of Import Duties and Impact on Government Revenue...........................6554 Summary of Stat ic Fiscal Impacts.....................................................................................6655 Impacts under the EU liberalizati on on three key agricultural commodities.....................6856 Impacts under the EU & SADC liberaliz ation on three key agricultural commodities.....68A1 List of commodities with the highest Tr ade Effects as a result of Tariff Elimination on EU Imports.................................................................................................................. ..77A2 List and levels of commodities with highest Tariff Revenue Changes as a Result of Tariff Elimination on EU Imports......................................................................................78A3 List of commodities with the highest Trade Effects as a Result of Tariff Elimination on imports from EU & SADC countires............................................................................79A4 List and levels of commodities with highest Tariff Revenue Changes as a Result of Tariff Elimination on Imports from EU & SADC countries.............................................80 8

PAGE 9

LIST OF ABBREVIATIONS ACP African, Pacific and Caribbean APEC Asia-Pacific Economic Cooperation ASEAN Association of Southeast Asian Nations CACM Central American Common Market CARICOM Caribbean Community and Common Market CER Closer Economic Relations CGE Computable General Equilibrium (models) CIF Cost, Insurance and Freight DPRU Development and Policy Research Unit EBA Everything But Arms EC European Commission ECOWAS Economic Community of West African States EFTA European Free Trade Association EPA Economic Partnership Agreement ERP Effective Rates of Protection ESA Eastern and Southern Africa EU European Union FTA Free Trade Area FTAA Free Trade Area of the Americas GATT General Agreement for Tariffs and Trade GCC Gulf Cooperation Council GDP Gross Domestic Product 9

PAGE 10

GNP Gross National Product GTAP Global Trade Analysis Project GSP Generalized System of Preferences HS Harmonized System IADB Inter-American Development Bank ISI Import Substitution Industrialization LAFTA Latin America Free Trade Area LAIA Latin America Integration Association LDC Least Developed Countries MERCOSUR Mercado Comn del Sur, Mercado Comm do Sul (Southern Common Market (of America) MK Malawi Kwacha NAFTA North American Free Trade Area NBER National Bureau of Economic Research NSO National Statistical Office OECS Organization of Ea stern Caribbean States PE Partial Equilibrium PTA Preferential Trade Area RTA Regional Trade Area SADC Southern Africa Development Cooperation SPARTECA South Pacific Regional Trad e and Economic Cooperation Agreement TC Trade Creation TD Trade Diversion UEMOA West African Economic and Monetary Union 10

PAGE 11

USAID United States Aid for International Development US$ United States Dollar WTO World Trade Organization 11

PAGE 12

Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Mater of Science THE EU-ACP ECONOMIC PARTNERSHIP AGREEMENTS (EPA) FOR MALAWI: TRADE EFFECTS AND FISCAL IMPACTS FOR ALTERNATIVE POLICY OPTIONS By Innocent Lwafyo Thindwa August 2010 Chair: John J. VanSickle Major: Food and Resources Economics Malawi has, since 1975, been enjoying non-recipr ocal trade preference s from the EU bloc along with all other ACP countries pursuant to the successive Lom Convention Agreements. This framework was not consistent with the ex isting World Trade Organization (WTO) rules and regulations. As such, the EU and ACP were pressured to come up with new World Trade Organization (WTO) compatible trading arrangements. To this effect, the Cotonou Agreement was conceived by the two parties and signed in June, 2000. Under the Cotonou Agreement, the non-reciprocal preferential market access for ACP economies was scheduled to be replaced by a string of Economic Partnership Agreements (EPAs) meant to progressively liberalize trade in a reciprocal way, leading to the establishment of free trade areas (FTAs) between the EU & ACP regional groups in accordance with the relevant WTO rules. The glaring reality for the ACP countries in general, and Mala wi in particular, of opening th eir markets to EU imports has created debate on the anti cipated implications. This research employed the partial equilibrium model to quantify the trade effects and fiscal impacts for Malawi of an EPA with the EU, compared to those obtainable in a contemplated broader liberaliza tion scenario that included impl ementation of the EPA with the 12

PAGE 13

EU concurrently with the for mation of the regional Free Trade Area (FTA) among the Southern Africa Development Cooperation (SADC) group of countries. Results from the study demonstrate that Malawi stands to benefit more from a broader liberalization framework than one restricted to the EU. The results also provide useful insight into the types of product categorie s that would be most hit by th e policy change. In particular, trade and fiscal effects to the three key agricu ltural products of tea, sugar and tobacco were found to be insignificant. 13

PAGE 14

CHAP TER 1 INTRODUCTION The post World War II period saw the rejuve nation of efforts aimed at promoting multilateral trade liberalization among countries. Th e climax of this effort was in 1947, when the General Agreement on Tariffs and Trade (GATT) wa s established with a view to promoting free trade through elimination and/or reduction of trade barriers among countries. Following the classical arguments of Adam Smith and David Ri cardo, free trade is consid ered to be welfare improving as each country tends to allocate their productive resources in areas that they have the most comparative advantage. As such, trade lib eralization is considered a desirable vehicle towards the attainment of sustainable econom ic growth and development, leading to improvement in living standards for the populace. Parallel to the movement towards multilatera l trade liberalization, was the emergence and proliferation of the concept of regional trading arrangements. These were (and still are) configured through various forms and names depending on the varying degree and/or stage of economic integration, such as free trade ar ea, customs union, common market, economic union, and complete economic integration. This feature of the global economy stimulated a lot of debate and theorization among economists during the entire second half of the tw entieth century on the desirability and relationship of regionalism vis--vis multilateralism as an approach to free trade. One of the offshoots from this debate has been the development of the theory of second best, referring to regional trade arrangements as oppos ed to the superior case of multilateral trade liberalization. It is worth noting though that the proliferation of the regional free trade arrangements is not without legal basis in the international trade rules and regulations as Article XXIV of the GATT sanctions their formation. 14

PAGE 15

At the hea rt of these multilateral and regional trade arrangeme nts has been the feature of preferential treatment, usually sought by the developing countries, which calls for special considerations and exemptions from recipr ocal trade liberalization measures. This fair treatment plea falls within the larger body of arguments by some agains t equal treatment of developed and developing countries in the pursu it of open trade policies on account of the notion that smallness of the economies for the later ma kes them more vulnerable to exte rnal shocks. The proponents of this paradigm have advanced the argument th at openness and liberaliz ation do have significant negative effects on poverty, unemployment and a hos t of other issues (Weisbrot and Baker 2002; Grunbaum 2007). Free trade proponents dispute this and argue that these small and low income countries need not be treated differently if they are to realize significant positive economic impacts from trade. These arguments and counter-arguments for op en trade are reflectiv e of the generally observed fact that trade liberal ization does not only create chal lenges to the country undertaking such policy but also offers them opportunities to attain superior levels of growth and prosperity through improved competitiveness and enhanced e fficiencies (Grunbaum 2007). This is exactly the situation in which the African, Caribbean and Pacific (ACP) group of countries find themselves now as they face the glaring reality of opening their borders to imports from the European Union countries under the 2000 Cotonou Agreement between the EU on the one side and ACP countries on the other. Malawi is an in tegral member of this group of countries, itself being within the African count ries sub-group. With an estimated annual gross domestic product (GDP) level of about US$ 4,268 million in 2008, the country is categorized as one of the least developed countries (LDC). As an LDC, Mala wi benefits significantly from the EU trade preferences under the previous trading arrangement as well as the on-going EU unilaterally 15

PAGE 16

determ ined Everything But Arms (EBA) arrange ment. The countrys major export commodity under the EBA arrangement is sugar, which is sold at an EU predetermined high price in pursuit of subsidizing sugar production wi thin the EU farming realm. Problematic Situation The ACP countries have bene fited from preferential treatme nt from the EU since 1975, when the Lom I Convention was signed in Lom, Togo. Under this agreement, the ACP countries were accorded non-reciprocal preferen tial market access to the European Union countries. The period that ensued saw successive renewal of the initial Lom I Convention, through Lom II in 1979; Lom III in 1984; and Lom IV in 1989 that remained in force for the entire decade until 1999. All trade preferences, articles of trade and development aid schemes under these renewals were characterized by the sa me non-reciprocal princi ple. At the dawn of the twenty-first century, the wave of globalization and the associated efforts to strengthen the multilateral approach to trade and economic proc esses warranted the need for adapting the EUACP trade arrangements to ensure their comp atibility with contemporary WTO rules and regulations. Besides, the section of developing countries that fa lls outside the ACP group had not been supportive of the discriminatory EU-ACP preferential trade agreements under the Lom Conventions. As such, the WTO accorded a final waiver to the EU and ACP countries running only up to December 2007, after which the requisite trade cooperation was required to be made WTO compatible. To this end, the two parties commenced negotiations in September 1998 that culminated into the new ACP-EU Partnershi p Agreement, signed in Cotonou, Benin on 23rd June 2000. One of the major innovations of the new ACP-EU Partnership Agreement, commonly known as the Cotonou Agreement was the introduction of new fundamental principles with respect to trade between the European Union and the ACP countries relative to the Lom 16

PAGE 17

Convention. In particular, the non-reciprocal preferential m ark et access for ACP economies was scheduled to be replaced by a string of Economi c Partnership Agreements (EPAs) meant to progressively liberalize trade in a reciprocal manner. In fact, Article 36.1 of the Cotonou Agreement empowers parties to the Agreemen t to conclude new Wo rld Trade Organization (WTO) compatible trading arrang ements, removing progressively barriers to trade between them and enhancing cooperation in all areas of trade (The Cotonou Agreem ent 2000, Article 36.1). It was envisaged that the progressive removal of ba rriers to trade would result in the establishment of Free Trade Agreements (FTAs) between the EU and ACP regional groups in accordance with the relevant WTO rules and help to enhance the existing regional integration efforts among the ACP. The EPAs were to be negotiated from September 2002 to 31st December 2007 to enable the new trading arrangements to enter into force by 1st January 2008 (The Cotonou Agreement 2000, Article 37.1). However, this deadline was not met and to date, only the Caribbean subgroup and a couple of African countries have managed to sign a fully complete EPA. Since the substantive issues under this pro cess were discussed within seven configured regional groupings, Malawi together with fifteen other countries formed the Eastern and Southern Africa (ESA)1 configuration and launched the EPA negotiations with the EC on 7th February, 2004. The ESA group identified six clusters of issues to ne gotiate an EPA with the EC, which includes the following; Development Issues, Ag riculture, Market Access, Fisheries, Trade in Services, and Trade-Related Issues. The formation of EPAs and elimination of EUs long cherished trade preferences for ACP will obviously have far reaching implications on development strategies followed by these countries. In essence, these ACP countries have for a long time tailored their productive efforts 1 Other group members are: Burundi, Comoros, DR Congo, Djibouti, Eritrea, Ethiopia, Kenya, Madagascar, Mauritius, Seychelles, Rwanda, Sudan, Ug anda, Zambia and Zimbabwe 17

PAGE 18

towards sectors favored by the preferences, th er eby creating a potential for distortions in resource allocation. It is clear that the anticipated change in the trade framework, which would invariably result in a Free Trade Area (FTA) will necessitate some appropriate restructuring of the development strategies for the ACP so that c oncentration is placed on areas for which each one of these countries has the most competitive a dvantage. This restructuring process, otherwise known as a rationalization process, involves costs that the ACP countri es need to be prepared to incur. The process will practically mean transf erring resources, including laying off workers, from sectors deemed inefficient to those for which the country has comparative advantages. It is generally feared that these changes may bring about increased unemployment, provoke heightened economic insecurity and political in stability (Keck and Piermartini 2005). There is also a very clear sense of despair among ACP co untries with respect to the level of competition that domestic producers will be subjected to afte r the FTA. While compet ition is lauded as a positive phenomenon in that it promotes effici ency in allocation of productive resources, commentators in ACP countries have expressed concerns over the le vel of EU production subsidies in the agricultural sector and argue th at such competition will invariably be unfair. Problem Statement The delay by ACP countries in finalizi ng the EPA regional frameworks and the consequential formation of the FTAs between th e EU and ACP countries is clear testimony of the significant concerns among ACP countries with the proposed reciprocal trading arrangement. While the nature of concerns is similar for most of the ACP countries, it is undeniable that the extent of the envisaged impacts, both positive and ne gative, will vary from country to country. In this regard, country-specific analyses of the pot ential impacts of the prop osed FTA with the EU are of colossal importance because outcomes fr om such work are bound to help the concerned country to chart the best way forward as regard s its national developmen t strategies vis--vis 18

PAGE 19

trade po licy. This is precisely th e type of research that this thes is was formulated to carry out for Malawi. The fact has already been all uded to above that trade polic y is usually analyzed on the basis of the link it has to economic growth a nd development for the country. A large body of literature on the analysis of ec onomic integration and its shallo wer forms such as the proposed Free Trade Area (FTA) suggest that their formation tends to br eed various economic effects, particularly relating to the di rection of trade which is larg ely influenced by the underlying advantages for countries invol ved. An expansion in trade can be had when less efficient domestic production is substituted by imports from a more efficient member of the agreement (Grunbaum 2007, p.19). A further expansion in trade flows can be expected as the domestic country imports more products from member countri es at preferential te rms at the expense of excluding imports, which in fact can be less co stly, from non-member countries (Viner 1950; Pomfret 1988; Grunbaum 2007). The trade liberalization argument can not be co mplete without consideration of the fiscal impacts associated with the immediate loss of tari ff revenue. This loss directly affects the ability of the liberalizing government to deliver public services to it s citizenry and negate on-going efforts towards economic growth, development and poverty alleviation. The situation tends to be more serious for countries wher e the proportion of revenue from import tariffs is higher than where it is lower. The situation for ACP countries is such that in most countries, revenue derived from import tariffs constitutes a substantial por tion of government revenue. In fact for Malawi, the Development Policy Research Unit of the University of Cape Town (DPRU) estimates the contribution of tariff to total government revenue as standing at about 22 percent (DPRU Policy Brief No. 01/P8, 2001). The situation currently ruling presents us a rare opportunity to carry out 19

PAGE 20

20 research that will enable us to get quantitativ e estimates of the potential trade effects and associated fiscal impacts of the proposed EU-E PA for Malawi. For Malawi, trade in three key agricultural products is of paramount importance. Exports of tobacco, tea and sugar together account for over 70 percent of the co untrys total annual exports. As such, it will be interesting in this study to move a step further and isolate the potential impacts that th e EPA with EU will have on sectors involved in the pr oduction and processing of these three key products. Research Objectives The move towards EPA and FTA poses a huge challenge for ACP countries such as Malawi. The consideration of potential benefits of improved and more secure access to EU markets by Malawian exports is blurred with th e recognition of expected loss in revenue from customs duty and increased competition for the domestic industries. As debate on the likely effects continues and time for implementing the EPAs is already surpassed, very few countryspecific quantitative analyses of th e situation have been undertaken, ex ante This study is precisely aimed at addressing this apparent gap for one particular country, Malawi. As such, the overall objective of this research is to evaluate th e potential trade effects a nd fiscal impacts of an EPA trade arrangement regime for Malawi. Specifically, the proposed research is aimed at attaining the following objectives; Conduct a quantitative analysis of the potential trade and fiscal effects for Malawi of the proposed EPA and/or FTA with the EU. Trade creation, trade diversi on and fiscal impacts will all be estimated. Identify commodity categories th at show highest trade effects, particularly investigating the trade effects to imports in the three key agricultural products of tobacco, tea, and sugar. Based on outcomes from above analyses, pr ovide alternative policy options for the government and stakeholders on what course of action should be pursued in respect of the upcoming EPA/FTA.

PAGE 21

CHAP TER 2 LITERATURE REVIEW The essence of the argument for international tr ade is with respect to its presumed link to economic growth and development. Classical trade theory suggests a positive relationship between increased open trade and superior economic performance and th at protectionism is inversely related to economic grow th and development. Freer trade regimes result in more rapid economic growth through direct effects of tr ade that operate via dynamic advantages of increased capacity utilization and more efficient investment projects as well as indirect effects through accelerated export growth (Edwards 1993). On the other hand, there are some researchers who object to this proposition and argue that empirical ev idence does not strongly support the classical view. They also question the validity of the di rection of the casual relationship between export growth and economic development as insinuated in the classical argument. As stated in the first chapter, the forma tion and proliferation of preferential trade agreements and/or regional trade agreements has become an important feature of the global economy. This feature has been considered by some as a catalyst for attainment of the more desired multilateral free trade regi me, yet others see it as an obst acle to the same. This chapter starts by providing a review of the theoretical and empirical lite rature on the link between trade and economic performance. We further endeavor to provide the foundations of the development and growth of the theory of preferential trade agreements before reviewing its empirical literature. We end the chapter with a presentati on of the major analy tical methodologies that have been used in the empirical analysis of pr eferential trade agreements, including citation of the empirical studies on which thos e methodologies have been applied. 21

PAGE 22

Liberaliz ation versus Protectionism and Economic Performance1 The notion that freer international trade stimul ates growth and development dates back to the days of Adam Smith. A host of later econom ists worked to augment and popularize the free trade notion in the two centuries that followed. Nonetheless, the post World War II decades of 1950s, 1960s, and 1970s saw increased emphasis on the pursuit of protectionist development strategies, especially among developing countries. A major impetus for such an orientation in the developing world was the policy of Import Substitution Industrialization (ISI), embraced by many of these countries in Latin America, Africa and East Asia. The urge to develop and sustain their manufacturing sectors provided for the pursuit of the ISI strategy which promoted protectionism in line with the infant industry argument for industrializat ion. It was believed that successful economic development would be attain ed through rapid deve lopment of the local industry to ensure self-suffi ciency and insulate the dom estic economy from external vulnerability. One of the major consequences of the ISI euphoria was th e preoccupation by most of the development economists of the time with the design of planning models based on ISI. In spite of the apparent dominance of the prot ectionist paradigm, a small group of academics embarked, independently, on major empirical inves tigations aimed at assessing the consequences of alternative trade re gimes (Edwards 1993, p.1359). The pioneer work to this effect was undertaken by Little, Scitovsky, an d Scott (1970) and Balassa ( 1971), who calculated effective rates of protection (ERP) in a sc ore of developing countries a nd linked these to the countries overall economic structure and performance2. 1 This section draws heavily on the work of Edwards (199 3) who provides an excellent synthesis of the studies undertaken earlier by various authors on empirical analysis of the link between outward looking and inward looking trade policies with economic performance 2 An excellent review of these studies is provided by Edwards (1993) and summarized and cited by Grunbaum (1997) 22

PAGE 23

The pursuit of ISI by the de veloping countries was critic ized f or its disregard of fundamental economic principles such as that of comparative advantage in determining what to produce. As a consequence of this disregard, some fundamental distortions in resource allocation were argued to have been identified by some au thors. For instance, the Little, Scitovsky, and Scott (1970) study concludes that the policies followed in most of the developing world after World War II had excessively enc ouraged industrialization at the cost of reducing the incentives for expanding agriculture and exports (Edw ards 1993, p. 1362). As a consequence of these inward looking policies, thos e developing countries experienced a ray of economic problems including a worsening income distribution, a redu ction in savings, an increase in the rate of unemployment and a very low rate of cap acity utilization (Edwards 1993, p.1362). African countries were a major segment of the LDCs that pursued inward looking trade policies in the hope of achieving positive economic growth and self-reliance. However, instead of attaining these positive soci o-economic benefits, implementa tion of these protectionists policies only resulted in severe crises for most of the African countries as market incentives were seriously distorted, food production plummeted, GNP per capita fell by almost one percent per year during the 1970s, corruption became rampant, and shortages were generalized (Edwards 1993, p. 1370). This appears to suggest that adoption of liberalized trade policies would have helped these countries to improve their economic performance. As such, in its 1981 report, the World Bank recommended that thes e countries undertake the implementation of reforms to their economic systems to make them more liberal and more open to international trade. Notwithstanding their novel pioneering work on the empirical link between trade policy orientation and economic performance, the policy recommendation for more open trade in the 23

PAGE 24

developing countries by Little Scitovsky, and Scott (1970) and Balassa (1971) was heavily criticized by Paul Streeten (1971) for lack of additivity and inc onsistency in their argum ents for freer trade. Edwards (1993) also pointed out that the authors onl y used single-period snapshots of the protection levels in the specifi c countries to derive their conc lusions other than attempting to estimate evolving protection levels which could have permitted them to analyze liberalization episodes over time and the link between altern ative protection levels and growth in those particular historical settings. Edwards (1993) further identifies a computational weakness in the two studies, that while using the same techni que, the Little, Scitovsky, and Scott (1970) study calculated a 49 percent effective rate of protectio n to the manufacturing sector for the Philippines in 1965 while the Balassa (1971) study arrived at a rate of 61 percent for the same sector in the same country during the same year. In a manner that served to address the sec ond weakness in the above studies, one of the classic works was undertaken by Anne Kruege r (1978) and Jagdish Bhagwati (1978). In the NBER study that they co-direc ted, Krueger (1978) and Bhagwati (1978) formally derived a trade liberalization index based on the degree of bias against exports. With the help of this index, countries could be determined to be at varying st ages of liberalization, starting from a scale of I as the most protected to V as the most liberalized. Assigning dummy variables for these stages, and using country-specific pooled data for traditional and non-traditional exports, Krueger (1978) undertook a formal econometri c analysis of the link betw een trade liberalization and economic growth. She estimated two equations, on e for exports and the other for real GNP for each one of the ten countries in the sample. The first was dependent on the real effective exchange rate, dummy variables for the trade liber alization regime, and time trend variable while the later was dependent on the exports index, tr ade liberalization regime dummy and a time trend 24

PAGE 25

variable. The results showed significant positiv e effects on export growth for the coefficient on the dumm y variable for phases f our and five of the liberaliz ation ladder, while no direct significant effects were discerned from the same phases of liberalization on GNP growth. The model did produce a significant positive effect on GNP growth for the export coefficient. This suggests existence of an indirect as opposed to the theoretically surmised direct positive link between trade liberalization and economic growth. Kruegers finding of lack of direct effect of trade liberalization on economic performance did not go without challenge from other researchers. Balassa ( 1982) in his study concluded that countries with intense trade liber alization policie s displayed high rates of economic performance. In the Balassa (1982) study, both quantitative rest rictions to trade a nd tariff barriers were incorporated into the methodology used to categorize the country s trade regime. This was done to address what Balassa (1982) pointed out as one of the shortfalls of the Krueger (1978) study i.e. that use of quantitative rest rictions alone in her study to determine trade regimes was in fact leaving out a major component of bias against exports inflicted by high tariff rates. Balassa (1982) used export growth rates as a proxy for trade regime, which turned out to be one of the sources of criticisms leveled against his study. As Edwards (1993) puts it, it is not clear whether it is exports growth that causes output expans ion (p.1373), or that output expansion spurs exports growth. In the years that followed, in the 1990s, a number of influe ntial studies were undertaken to establish the link between trade openness and economic performance. The approach for these later studies had been different and quite flex ible in searching for alternative ways of liberalization and establishing their impact of economic performance. As Grunbaum (2007) observes that while early empiri cal studies related to the role of exports and growth, more 25

PAGE 26

recen t studies on trade policy and economic performa nce have searched for alternative measures of openness and their relationship to economic outcomes (Grunbaum 2007, p. 61). Grunbaum cites the works of such authors as Dollar ( 1992); Sachs and Warner (1995); Harrison (1996); Edwards (1998); Frankel and Romer (1999) as some of the most influential in undertaking crosscountry econometric studies on the subject. In most of these studies, the results have tended to suggest the existence of a posit ive link between more open trade and higher rates of economic growth. Notwithstanding the foregoing, the view that tr ade liberalization is directly linked to higher levels of economic growth has also been criticized by a number of authors who refute it and argue that there is no fi rm empirical evidence establishing such a direct link. The work of Jeffrey Sachs (1978), and Deepak Lal and Sarath Rajapatirana (1987) stand out in literature as some of the leadi ng representatives of this criticism. Most recently, Rodriguez and Rodrik (2001) carried ou t a critical review of some of the works of earlier authors which supported the positive link between open trade and economic perfor mance. By using original data sets, Rodriguez and Rodrik (2001) were able to replicate and analyze the measures of openness used, disaggregated these into tariff and non-tariff components to test statistical significance, extended and modified the empirical models to obtain additional results (Grunbaum 2007, p. 61). Their findings suggest that those results in the original studies had been greatly over-stated a nd that the purported link wa s in fact very weak. Authors have acknowledged existence of a c ouple of challenges associated with these cross-country econometric studies, which make generalization of results tricky. These include poor data, inappropriate methodol ogy and a weak theoretical fram ework lacking the ability to establish clear relationships between variables (Gr unbaum 2007, p.61). 26

PAGE 27

Notwithstanding the quandary portrayed in th e literature reviewed above about the significan ce of the link between trade liberalization and economic performance, it is important to recognize that a more open trade regime brings w ith it elements that promote economic growth and development. Whether that happens directly or indirectly is beside the point. Trade imposes rigorous competition on domestic producers there by enhancing competitiveness, fosters foreign direct investment, alters the political economy by reducing rent seeking behavior, constrains governments ability to manipulate macroeconomic policies and subjects and binds economic actors to the discipline of international markets and the international environment (Grunbaum 2007, p.62). In addition, trade liberalization allows for feedback mechanisms to evaluate the performance and effects of policies (Berg and Krueger 2003; Grunbaum 2007). Development and Growth of the Theory of Preferential Trade Agreements After the Second World War, the internationa l community agreed on a set of rules and principles aimed at promoting multilateral trad e under what was known as the new international economic order, under the regulation of the GAT T. However, it soon became apparent that attaining free trade at that level was slow and difficult. As su ch, a number of countries started forming smaller groups, configured on the basis of either geographical vicinity or political and/or historical perspectives to pursu e regional economic integration in a preferential manner. This feature of global economy has come to be known as regionalism, with its peak at development experienced in the 1950s and the 1990s. As suc h, Bhagwati and Panagariya (1996) have termed the two stages as the old regionalism of the fif ties and the new regionalism of the nineties. The proliferation of these regional trade agreements and/or customs unions in the period after the 1950s spurred a lot of interest and debate on thei r desirability and relationship to the multilateral trade agenda. Some free trade proponents argued that regionalism was a necessary stepping stone 27

PAGE 28

to attainm ent of the most desirable multilateral trade agreement while others abhorred the feature for its tight grip at restrictions to trade against non-members to the agreement. At the same time, some protectionists liked the regional trade agre ements for increasing the degree of preference for favored countries and mainta ining restrictions ag ainst non-member countries while others disliked them for dismantling protection accorded to domestic producers against suppliers from favored countries. The work of Jacob Viner (1950) stands out as the earliest and most influential on this topic, laying out the primary th eoretical basis for the analysis of regional and/or preferential trade arrangements. The most significant contribu tion of Viners work was his observation that welfare effects from any form of regional econo mic integration are not unambiguous a priori. As such, no general conclusion is possible on whethe r such economic integration arrangements are welfare enhancing or damaging. Central to the de velopment of his theoretical framework was the formulation of the concepts of trade creation and trade diversion Viners original definitions for these two concepts have been refined over time by various subsequent researchers. Thus, we define trade creation as the s ubstitution of domestic production with cheaper imports from a member country while trade diversion is the substitution of cheaper imports from non-member countries with more costly goods from a me mber country (El-Agraa 1997; Gordon 2007). As such, trade creation reflects a shift from an in efficient to an efficient source of supply while trade diversion is movement from an efficien t supplier to an ineffi cient one (Gordon 2007, p.34). In his refinement of the definition, Robson (1980) distinguished between two components of trade creation. The first one of these is a production effect reflect ing savings from the reduction of domestically produced goods and se cond, a consumption effect reflecting the gains in consumer surplus as high cost consumption goods were substituted for lower cost goods 28

PAGE 29

(Grunbaum 2007, p.65). Similarly, Grunba um (2007) summarizes that tr ade diversion reflects to a shift in the source of imports from a non-uni on low cost producer to a more costly union member. The theoretical framework3 developed by Jacob Viner and all refinements made thereafter provided a fertile basis for the empirical analysis of regi onal economic integration. With this tool in hand, one would proceed to esti mate the trade effects and welfare outcomes of a preferential trade agreement simply by estimati ng the extent and comparing the size of trade creation and trade diversion. The agreement is regarded as welfare enhancing and therefore advantageous in the event that trade creation exceeds trade diversion (Grunbaum 2007). Since only a sub-set of the countrys trad ing partners is include d in the preferential trade agreement, a three country construct, involvi ng the first country, member count ries in the agreement and nonmember countries in the same becomes necessary. It is in this setting th at the analysis proceeds to simulate and/or evaluate the trade flow s and production and consumption effects. The economic integration agreement so being evaluated would have predictable impacts on resource allocation, economies of scale, terms of trade, fa ctor productivity, economic growth and stability, and the distribution of income (Robson 1980; Grunbaum 2007). Nonetheless, this analysis is based on a number of restrictive a ssumptions that include the following4; perfectly elastic supply for imports; perfect competition for factor and pr oduct markets; factor mob ility within countries but not among countries; zero transpor tation costs; tariffs as the onl y available policy tool; prices accurately reflecting requisite opportunity costs; balanced tr ade; and full employment of resources. 3 This section draws heavily from the work of Grunbaum (2007), who provides an excellent synthesis of the Vinerian theoretical framework and assumptions thereof, as refined by Robson (1980). 4 This list of the Vinerian assumptions was summarized by Robson (1980) and cited by Grunbaum (2007). 29

PAGE 30

Apart from the later work of Robson (1980) cited above, previous authors such as R.G. Lipsey (1960); Harry G. Johnson (1960); and C. A. Cooper and B.F. Massell (1965) also helped to advance the Vinerian framework with their resp ective refinements to his original ideas. In his influential survey article that aimed at further refining Viners trade creation and trade diversion concepts, Lipsey (1960) observed that Viner assi gned a positive value to trade creation and a negative value to trade diversion. As such, Lipsey (1960) maintains that the suggested inclusion of the consumption effects in their definitions wo uld invalidate such a motive by Viner. Viners influential insights uncovered what cam e to be known in literature as the theory of second best, literally referring to partially lib eralized regimes character izing preferential trade arrangements. At the same time, the Vinerian theo ry and all literature from subsequent authors in the 1960s seemed to have failed to explain the mo tive behind these regional and/or preferential trade agreements. As Pomfret (2003) observes l iterature of the 1960s f ailed to explain why countries would form a customs union, when they c ould realize all the trad e creation benefits and avoid any trade diversion cost s by reducing tariffs in a non-preferential manner. Johnson (1960); and Cooper-Massell (1965) se parately provided a formal tr eatment of this question and suggested that formation of the preferential trading arra ngements is largely motivated by political other than economic reasons since their regime s can not breed superior economic effects to unilateral trade liberalization. They cited the pu rsuit of ISI in Latin America and Africa, and maintenance of peace and harmony between France and Germany in the EU case as two examples of such political reasons for PTAs formation. Technically, Johnson (1960) advocated that the measurement of trad e creation and trade diversion should include production and consumption effects as changes in import demands were a consequence of the formation of a customs union and the tariff reduction or elim ination (Grunbaum 2007, p. 67). Cooper-Massell 30

PAGE 31

(1965) on the other hand observed that m easuremen t of the welfare effects from a customs union has to include the impacts of a tariff reduction which results in consumer surplus gains and not just those of the pure trade diversion. This means that if the former effect is larger than the later, then the regime is beneficial to the countr y. Notwithstanding this inclusion, Cooper-Massell (1965) maintained that a negative welfare effect would be had from these preferential trade arrangements. Meade (1965); Mundell (1964) and Corden (1972) also provided influential augmentations to the Vinerian framework5. The initial theoretical framework developed by Jacob Viner had a number of gaps that he hims elf knew would come up later and be addressed by other researchers. As Grunbaum (2007) obser ves Although Viner was aware that economies of scale, imperfect competition, and terms of trad e issues would arise, he left them unattended (p. 66). Indeed Meade (1965) led the task by providing a critique of Viners model and extending the same in a general equilibrium framework. He assumed existence of infinite supply elasticities and demand elasticities of zer o, providing an allowance for multi-product production in all countries. Meades work provided a general sta tic framework of analysis for integration agreements that admitted substitution of goods both in demand and supply and allowed for simultaneous adjustments in related factor and goods markets in tradi ng countries (Grunbaum 2007, p.67). Using a three country model, Mundell ( 1964) addressed the dynami cs that arise in measuring welfare effects by accounting for change s in the terms of trade that come with the formation of preferential trade arrangements through changes in tariffs and relative prices. Investigating the relevance of the notions of trade creation and tr ade diversion in the face of scale economies, Corden (1972) did maintain their re levance. However, he suggested that such 5 Grunbaum (2007) excellently synthesizes the contributions from these authors and this brief discussion draws from his work. 31

PAGE 32

analyses should incorporate cost reduction a nd trade suppression effects that do em erge from scale economies resulting from the regime. Methodologies for the Empirical Analysis of the Impacts of Preferential Trade Agreements Literature cites three major quantitative anal ytical techniques that researchers use to empirically evaluate the effects of entering some form of an economic integration by countries. These include the gravity models, computable ge neral equilibrium (CGE) models and the partial equilibrium (PE) models. Naturally, each one of these methods has its respective merits and demerits such that choice of the appropriate methodology has to be informed by consideration of the costs and benefits to the researcher base d on the specific situation at hand, including data requirements considerations. This tends to help in balancing the inherent trade offs. It is also clear though that gravit y models are employed to carry out ex-post analyses of the economic integration agreements while CGE and PE models are used to conduct ex-ante analyses that simulate how todays economy will look in future as a consequence of a specified set of policy changes (Piermartini and Teh 2005, p.1). The one gene ral thing is that in all cases, the analysis undertakes to quantify the sizes of trade creation and trade divers ion as a result of the policy changes thus implemented, such as formation of a PTA. Gravity Models The gravity models in trade policy derive their name from the 1687 Newtonian Law of Universal Gravitation6 which held that the attractive fo rce between two objects is positively related to their mass and invers ely related to the distance betw een them. Economists have used the same concept to develop an analogous functiona l relationship of the gr avity trade model. As such, bilateral flows are considered to be pos itively related to the size of the countries 6 This section draws on the work of Keith Head (2003), Piermartini and Teh (2005) and an excellent synthesis of the model evolution and structure by Grunbaum (2007). 32

PAGE 33

econom ies and inversely related to the dist ance between them (Head 2003; Grunbaum 2007). This formulation was first proposed by Tinbergen (1962). As Piermartini and Teh (2005) observe that the first empirical study of trade using the gravity model was probably Tinbergen's7 (1962), although there was no explanation for the use of th e model nor for showing how it was related to theoretical explanations of international trade (p.38). Thus, Grunbaum (2007) summarizes the basic Tinbergen gravity model as follows; Tij = Mi 1 Mj 2Dij 3ij (2-1) where Tij = bilateral trade flow betw een country i and country j, Mi = economic mass interprete d as GDP of country i, Mj = economic mass of country j, Dij = distance between countries i and j, ij = the standard error term. The above formulation is usually transforme d into its requisite logarithmic form to estimate the following; In Tij = 0 + 1In Mi + 2In Mj 3Dij + ij (2-2) Literature suggests that GDP GDP per capita or population size can all be used to capture economic mass measured by Mi and Mj in this model while Head (2003) observes that distance is almost always measured using the great circle formula, which approximates the shape of the earth as a sphere and calculates the minimum distance along the surface. Thus using this formula, the distance between the capitals or commercial cities of the two economies is 7 Other authors such as Head (2003) and Grunbaum (2007) makes a similar observation. However, Grunbaum cites the work of Sandberg (2003), who observes that Poyhonen (1963) was simultaneously working on a similar type of the model. 33

PAGE 34

calculated and used in the Dij variable in the model. Grunbaum (2007) explains that economic size determines the abilit y of the country to engage in trade as trade flows between countries is largely a function of supply condi tions in the country of origin and demand conditions in the country of destination. Head (2003) gives a number of reasons to justify us e of distance in the model. These include the fact that distance serves well as a proxy for transport costs that include freight and marine insurance; it indicates the time required for shipment which is crucial, especially for perishable goods; it is a proxy for communication costs which involves personal contacts between managers and customers to exch ange information on the exchange to be made. Distance also captures synchronization costs invo lving time of delivery of inputs into the production process; it may also be a good proxy fo r the transactions cost of doing business; and the fact that more distant areas are likely to be more culturally differe nt suggests that trade between such areas may be lower compared to geographically close areas/countries. As noted above, the original gr avity model was criticized for its lack of touch with the standard theoretical bases for international trade as posited in the famous Ricardian and Heckscher-Ohlin models. The former emphasized differences in comparative advantages among countries to explain trade flows and pattern while the later dwe lt on the variations in factor endowment among countries. This criticism became one of the major bases for further refinements of the standard gravity model as proposed by Tinbergen. Pi ermartini and Teh (2005) observe that Anderson (1979) was probably the first author to esta blish the theoretical basis for the gravity models. They observe that Anderson (1979) did this by constructing a model where goods were differentiate d by country of origin and where consumers have preferences defined over all the differentiated products (Piermartini and Teh 2005, p. 38). 34

PAGE 35

While the standard gravity m odel has been la uded to have performed significantly well in explaining trade, Head (2003) notes that there is a huge amount of variation in trade (p.9) that these models cannot explain. As such, researcher s have chosen to include a number of extra variables that have been observe d to be relevant in explaining trade flows. Some of these variables include; dummies to cap ture the effects of country ad jacency, whether a country is landlocked or not, whether two countries speak the same language or share some colonial history, whether the country is a member of some PTA or RTA, customs union, whether they share the same currency and so on. As Head (2003) observes, authors include these variables, albeit there is weak theoretical justification for doing so. The fact that th e two major variables in the standard gravity model, GDP and distance tend s to fit the data well, increases the temptation to include any of these seemingly relevant variables. It is import ant though that choice of which variables to include in the mode l must be guided by the specific issue at hand. As Piermartini and Teh (2005) observe that one needs to proceed carefully in analyzing the theoretical questions at hand which should guide the choice of appropriate regressors to be used in the empirical estimation method. Empirical Application of the Gravity Model As alluded to above, gravity models have been extensively used to evaluate the effects of preferential trading arrangement s (PTA), simply by adding an intra-bloc and an extra-bloc8 dummy variable to capture such effects in the standard model. The analysis proceeds on the presumption that the normal trade volume betw een countries is explained by two variables, economic size of the trading partners and dist ance between them, such that any significant effects from the PTA will manifest in either increased (trade creation) or reduced (trade 8 This nomenclature, intra-bloc and extra-bloc, draws from Piermartini el at (2005) usage 35

PAGE 36

diversion) volum es between these partners as compared to the normal volume. Thus, trade creation is captured by a significant positive coe fficient for the intra-bloc dummy while trade diversion is captured by a signi ficant negative extra-bloc du mmy variable coefficient. Frankel (1997) applied the trad itional log-linear gravity model on levels of variables to study the trade effects of a number of RTAs including EC, NAFTA, EFTA, Adean, ASEAN, and MERCOSUR. Using total trade as the dependent variable, the model was augmented by two dummies for intra-bloc and extra-bloc trade. The study established the existence of net trade creation effects for EC, MERCOSUR and Adean, a nd significant trade diversion for NAFTA and EFTA. Furthermore, the intra-bloc dummies were found to have no significant effects for Adean, NAFTA and EFTA. A similar model was used by Soloaga and Winters (2001) but with two extra-bloc dummies, one for imports and the othe r for exports to study th e trade effects of EC, NAFTA, EFTA, Adean, ASEAN, CACM, MERCOSUR and other blocs. The results of the model on levels showed existence of negativ e intra-bloc dummies for the EU, EFTA and ASEAN; and positive intra-bloc trade imp acts for CACM, ANDEAN and MERCOSUR. The intra-bloc dummy for NAFTA was not statistically significant. The estimates on the first differences of the variables showed that EU and EFTA were net trade diverting, while MERCOSUR and CACM were net trade creating. D ee and Gali (2003) used the model with the augmentation similar to the one by Soloaga and Winters (2001) to study the impacts of Andean, APEC, EFTA, EC, GCC, LAFTA/LAIA, MERCOS UR, SPARTECA, CER, AFTA and a set of bilateral agreements. However, Dee and Gali (20 03) constructed what they called a Member Liberalization index (an i ndex of the coverage of the RTA), such that the three dummies would take the value of this index whenever the RTA was in force. Results from their analysis found 36

PAGE 37

that ne arly all RTAs studied we re found to have net trade dive rting effects, while net trade creating results were found only for And ean, LAFTA/LAIA, US-Israel and SPARTECA9. Nilssons (2002) study on the effects of the preferences accorded by the EU to ACP countries under the Lom Convention compared to the EU GSP arrangements provides a good example of how the gravity model can be augmente d to captured a ray of other variables than those in the standard model. Dummies were in cluded to capture the e ffects of cultural and historical ties and the signif icance of colonial linkage betw een former colonies and their European colonizers. The study f ound that exports for developing c ountries were significant and larger under the Lom Convention as compared to the EU GSP arrangement and that historic ties significantly explain trade linkages for selected Eu ropean countries and their former colonies. Computable General Equilibrium Models (CGE)10 CGE trade models exploit the computer capa bility to construct a rigorous analytical framework that takes into account all market linkages, retains optimization assumptions thereby preserving its consistency with the hallmarks of the general equilibrium theory of the economy (Grunbaum 2007). Apart from their theoretical c onsistency, the ability of the CGE models in arriving at precise numerical estimates has also been cited in literature as one of the model strengths. This is easily atta ined by CGE models since the workings of an economy and the changes that would follow specific policy implem entation can be simulated, as CGE models act to emulate the functions of laboratory e xperiments (Piermartini et al (2005). It is the ability of CGE models to capture the overall pictur e of the changes that emanate from requisite policy changes that elevates their importance in ex-ante analyses of trade policies 9 Piermartini et al (2005) provide a summary of these and more su ch studies in their work 10 This section draws from the work of Busse et al (2005); Piermartini and Teh (2005) and Grunbaum (2007) 37

PAGE 38

such as enforcem ent of RTAs and PTAs. As Grunbaum (2007) observes, overall aggregate trade, terms of trade effects, factor prices, trade creation and trade diversion within an economywide model can be estimated; as can be obtained estimates of inter-sectoral linkages, prices, wages, and exchange rates that lead to equilibrium in product a nd factor markets, as well as balance of trade figures (p. 79). As such, alternat ive scenarios of the specific policy change can be conjectured and their requisite pote ntial parametric outcomes simulated. While CGE models may be more suitable to an alyze the overall trade and welfare effects of changes in trade policy, their data requirements are significant and can be a major hurdle for studies involving developing count ries. These models require wh at is known as the social accounting matrix, which involves input and out put data for the entire economy and their requisite inter-sectoral linkages pl us their associated contributions to output; government fiscal and/or budget accounts, disaggregated into cons umption, investment, government expenditures, balance of payments; as well as data on the volumes and values of imports and exports disaggregated by their requisite composition, origin and destina tion (Grunbaum 2007). Besides, these data matrices must be appropriately arranged into revenues and expenditures, balanced and standardized (Piermartini and Teh 2005; Grunbaum 2007). Apart from the social accounting matrix construct, CGE models require exogenous variables that captu re the behavior and response of producers and consumers to any cha nges in incomes and relative prices resulting from the policy changes thus eff ected. The parameters most ofte n needed are: elasticities of substitution related to th e responsiveness of producers to changes in relative prices of factors of production; consumer demand and income elas ticities; and Armington elasticities, which determine the substitutability be tween domestic and imported products. Usually such elasticities are borrowed from previous econometric studies and may be adapted in to the current model 38

PAGE 39

(Grunbaum 2007). This considerable amount of g ood quality data can be rarely obtainable in developing countries, let alone the availability of prior econometr ic studies in those countries from which the stated elasticities could be borro wed. Otherwise, the pers istence to applying the model in studies involving those countries may result in use of dubious quality data. As indicated above, the nature of the data employed by the CGE model is significantly aggregated. This level of aggregation has been cri ticized as a potential so urce of loss of detail in fundamental relationships that may be obt ained. Indeed as Grunbaum argues, complex simulation models where large amounts of data inputs produce precise outputs can be deceiving as precise sources of certain re sults are not clear ly identifiable (Grunbaum 2007, p 81). Pundits further attack users of the model for their tendency to choose values of the requisite elasticities arbitrarily and the Armington elasti cities from outdated studies. In view of these concerns, some have suggested that systematic validation of CGE simulations through ex-post evaluation is necessary to enhance the confidence and the predicti ve potential of the anal ytical results derived from the initial studies (Piermartini and Teh 2005). Application of CGE Models in Empirical Work In spite of the limitations cite d above, CGE models have been used quite extensively to analyze the trade effects of not only RTAs but also multilateral trade agreements, such as the Uruguay and Doha rounds of trade agreements under the WTO. In fact, CGE models have the ability to isolate the trade creation and trade dive rsion effects from such policy changes by sector which allows computation of sectoral welfare effects other than aggregate. Grunbaum (2007) cites the work of Kerkala, Niemi, and Vaittinen (2000) who used a multiregional general equilibrium model to examin e the consequences for African ACP countries of a post-Lom world. Their work simulates the effects of entering into force of WTO compatible trading arrangements under the proposed EPA as compared to the ones obtainable 39

PAGE 40

under the E U GSP system. The authors employed the Global Trade Analysis Project (GTAP) model to simulate the potential outcomes which suggested negative welfare effects for requisite African countries. Results also showed increased trade volumes under the EPA and declining volumes under the EU GSP system. Grunbaum su mmarizes that in both cases world welfare increased, however, positive effects were limited to the EU while they were absent in ACP countries (Grunbaum 2007). Another study on the potential effects of reciprocal trade liberalization under the Cotonou Agreement for the West African Economic and Monetary Union (UEMOA) against the EU was conducted by Wolf (2002). Wolf (2002) employed th e CGE tool to quantify the gains from this policy change and compare them to the losses in tariff revenue that would likely follow to these UEMOA countries. The results from this study showed that liberaliz ation under the Cotonou Agreement would have significant negative eff ects on tariff revenue generated by the UEMOA countries. Keck and Piermartini (2005) used an applied general equilibrium model covering 15 regions and 9 sectors to simulate the impact of signing EPAs with the EU for the Southern African Development Community (SADC) countries. The standard GTAP model was extended to include the elimination of textile quotas, EU enlargement to 25 members as well as tax revenue sharing and a common external tariff among SADC countries. Simulation of outcomes for various liberalization scenarios was undertak en. The issue of tariff revenue loss was also tackled in their study, including calculation of requisite tax replacement values. The results demonstrated that EPAs with the EU are welfar e-enhancing for SADC overall, leading also to substantive increases in real GDP. They also found that, for most countries, further gains could arise from intra-SADC liberalization. While th e possibility of the EU entering an FTA with 40

PAGE 41

other countries, such as MERCOSUR, was found to reduc e estimated gains, these gains still remained positive. In terms of sectoral level effects, the largest expansions in SADC economies was found in the animal agriculture and processed f ood sectors, with less attractive levels in the manufacturing following EU-SADC liberalization. Piermartini and Teh (2005) summarize a numbe r of studies that used CGE models to simulate the potential benefits from trade liberalization under the Uruguay and Doha Trade Agreements, noting their numeric bene fits in dollars to world welfare11. Partial Equilibrium Model (PE)12 By their nature, PE models focus on a detailed analysis of only one pa rticular market or sector, holding all other factors that can affect the same constant. In tr ade policy analysis, PE models have been used extensively to evaluate the effects of discriminatory tariff modification such as that which would occur under the EU-A CP EPA arrangement. In his quest to quantify trade effects of the then newly established Eu ropean Community, Petr us J. Verdoorn (1960) developed a partial equilibrium model based on th e Vinerian theoretical framework that remains significantly influential even t oday. Verdoorns original model has been modified by a number of researchers. Grunbaum (2007) obs erves that two basic types of PE models have so far been used in the analysis of preferential trade liber alization arrangements. Of these, the first one assumes trade in a homogenous commodity while the second assumes trade in a differentiated product and existence of infinite supply elasti cities. Grunbaum (2007) explains that under the first scenario, a reduction in tariff tends to spur expansion of trade flows, only to be limited by the corresponding supply elasticities This implies that some significant growth in trade flows 11 These studies are not covered here since our focus is on regional trade arrangements. 12 This section draws heavily on the work of Grunbaum (2007); Piermartini (2005) and Busse et al (2004). However, we only provide a brief overview of the Verdoorn model in this section, leaving a detailed derivation in the next chapter 41

PAGE 42

can be observed if the supply elasticity of a bene ficiary country such as the EU bloc in our case is high. Und er the second scenario, it is the degree of substitutability among goods that limit trade flows expansion. As such, a high elasticity of substitution leads to a substantial increase in trade flows and likewise a low elasticity of substitution leads to a small increase in trade flows (Grunbaum 2007, p. 84). Overview of the Verdoorn Model The Verdoorn model follows the second of the tw o types of PE models stated above such that product differentiation between supplying countries is assume d and that these products are imperfect substitutes in use. Busse, et al (2004) argues that this assumption seems reasonable for African countries, since the majority of Af rican imports consist of manufactured goods (p. 17). The model further inherits the usual partial equilibrium analysis assumptions including the following: no repercussions on exchange rates or in comes as a result of changes in trade flows, iso-elastic import-demand functions and existence of infinite s upply elasticities (Busse, et al 2004). Besides these assumptions, the model re quires knowledge of import demand elasticities and the elasticities of substitution between pr eferred and non-preferre d imports (Grunbaum 2007, p.84). As such, trade creation13 is captured as follows; TC = MP [ t/(1+t)] (2-3) where TC = Trade Creation MP =Imports from preferences beneficiary = Import demand elasticity t = tariff 13 This follows Grunbaum (2007) specification of the same 42

PAGE 43

In this form ulation, it is further assumed that the substitutability between imports from preferred sources and domestic production is equa l to the substitutability of all imports and domestic production (Grunbaum 2007, p.85). On the other hand, trade diversion is formulated as follows; TD = MP 2 ( )[ t/(1+t)] (2-4) where TD = Trade Diversion M P =Imports from preferences beneficiary 2 =Share of imports from non-preference beneficiary =Elasticity of substitution = Import demand elasticity t = tariff One of the major variants to the original Verdoorn (1960) model is that developed by Baldwin and Murray (1977)14, which utilizes data for domestic production on top of data on imports from non-preferred suppliers. Thus, while th e trade creation formula tion is similar to the Verdoorn model, the Baldwin and Murray mode l does have a different trade diversion framework, which is captured as follows; TD = MP [ t / (1+t)] [MN / MD] (2-5) where TD= Trade Diversion M P =Imports from Preferences Beneficiary M N =Imports form Non-Preferred country M D =Domestic Production = Import Demand Elasticity 14 This section is a synthesis of Grunbaums (2007) work, which identifies the Baldwin and Murray (1977) model as the major variant of the original Verdoorn model 43

PAGE 44

t = Tariff In this formulation, the authors assumed that the substitutability between imports from a preference beneficiary source and those from a non-beneficiary source was equal to the substitutability between imports from a preference beneficiary source and domestic production (Sawyer and Sprinkle 1989; Grunbaum 2007). Despite their detail, PE models have been criticized for a number of reasons. One such matter at issue relates to the choice of elasti cities used in the m odel. As Grunbaum (2007) observes, the values for the elas ticities used are chosen arbitraril y on the basis of estimates often considered unreliable (Grunbaum 2007, p.87). S econdly, any change in trade flows is bound to result in repercussions in all othe r sectors of the economy. As such, the ceteris paribus assumption used in PE models tends to be consid erably unrealistic. Pundits criticize PE models for failure to take into account these result ant inter-sectoral linkages and factor markets dynamics. As alluded to before, PE models ha ve the advantage at being detailed in their analyses. These models permit researchers to co nduct analyses at highly disaggregated levels, thus allowing for detailed analysis even by tariff line (Grunbaum 2007). This level of detail allows analysts to be able to pin-point those specific products and trading part ners that may show significant effects of the altern ative policy simulations. As such, PE models have been used extensively in empirical analyses of discriminatory trading arrangements. Empirical Application of PE Models Busse et al (2004) applied the Verdoorn mode l to estimate the trade and budget effects of the EPA on ECOWAS countries and Mauritania. Their study was aimed at addressing the fears among African countries of ente ring the proposed EPAs without clear understanding of the associated costs and benefits from the same. The results showed that imports from the EU in the ECOWAS countries were expected to increase in the range running from 5.2 percent for Guinea44

PAGE 45

Bissau to 20.8 percent for Nigeria. Importantly, Busse et al (2004) further reports that Trade creation exceeds trade dive rsion (in a bsolute levels) in all scenarios and for all West African countries (p.26). However, the budge tary effects were found to be very severe for some of the West African countries. Cape Verd e and Gambia were cited as the two countries to be severely impacted by the impending tariff revenue decline. The authors went a step further to identify those product categories that were found to have highest budgetary eff ects as well as trade creation and trade diversion effects for each of the West African countries covered. This, they sorted out by both absolute and relative changes. Max Grunbaum (2007) used the Verdoorn version of the PE model to project the effects of trade liberalization in the Organization of Eastern Caribbea n States (OECS) countries along the EU-ACP proposed EPAs which are essentially geared to create an FTA between these two trading blocs. He undertook to quantify trade cr eation, trade diversion an d fiscal impacts for three alternative liberalization s cenarios. The alternative scenarios were contemplated dependent on scope of discrimination to other trading part ners as EPAs come into force. The results demonstrated that trade liberalization results in small to modest positive trade and negative fiscal impacts. The other importantly indication from the results was that they provided empirical support to theoretical arguments and policy suggestions that, for small countries such as the OECS, broader trade liberalizati on is superior to limited regional trade agreements (Grunbaum 2007, p.12). Analysis of Fiscal Impacts As we pointed out in the first chapter, the i ssue of tariff revenue loss stands out as one of the major fears related to trade liberalization, especia lly among the developing countries. It is more of an issue among these countries since tariff revenue accounts for a significant proportion of total government revenue. The co ncern is that loss of such a significant proportion of revenue 45

PAGE 46

m ay suffocate the governments ability to deliv er services to the citizenry and increase impediments towards economic development. This fear is exacerbated by consideration of a myriad of other short-term adjustment costs asso ciated with liberalizatio n such as increased unemployment, reduction in national output, e limination of certain domestic industries and possible macroeconomic instability (Grunbaum 2007, p.91. It is such fears as these that have inspired interest among researchers on carrying out empirical studies on the anticipated levels of tariff revenue loss resulting from liberalization and derive comparisons with the anticipated levels of benefits from the same to establish an over all net position. A survey of literature on this subject reveals two distinguished themes along which discussion and research has dealt with this issue in the past. While the first of these has focused on the measurement and evaluation of the relativ e importance of tariff revenues as a proportion of aggregate government revenue, the second th eme has aimed at exploring alternative tax systems and/or fiscal reforms that could help compensate for tariff revenue losses (Grunbaum 2007). To understand the fiscal effe cts of tariff reduction in pursuit of trade liberalization, one has to understand the ways through which such a policy change impacts tariff revenue. Grunbaum (2007) reviewed the 2004 Report of the Inter-American Development Bank (IADB) and summarizes five ways through which tariff reduction affects tariff revenue. These include; direct effects or losses due to the reduction of a given tariff line; indi rect effects or revenue decline from taxes imposed on CIF plus tariffs th at are forgone due to the tariff rate decline; elasticity effects which affect the revenues depending on whether they cause an increase or decrease in the volume of trad e of certain products; substituti on effects which result in trade diversion from the displacement of imports from partners not facing tariffs; and induced effects that are changes in total tax revenues resultant from consump tion and production patterns borne 46

PAGE 47

47 of economic structures post liber alization (Grunbaum 2007). Appreci ation of these five channels demonstrates how complex the resultant net ou tcome from trade policy change can be. As literature shows, trade liberaliza tion can in fact result in increased tariff revenue generated by the government depending on a given situation. A host of factors do affect the net impact from trade policy change for a country. Some of these factor s include the initial conditions such as tariff structures and general restrictiveness to trad e, domestic tax including reforms to quantitative restrictions, reduction in the number of tariff le vels, and exchange rate s and exchange rate regimes. In general, the more restrictive the initial conditions, the more the losses in tariff revenue, and the tighter the governme nt is on the exchange rate re gime the higher the costs from liberalization. The level of development for the country is found to be more important in determining the countrys ability to offset loss of tariff reve nue due to liberalization. The work of Khattry and Rao (2002) was central at establ ishing this perplexing phenomenon. They argued that structural limitations characterizing low-income countries make the gradual substitution of trade taxes by domestic sources of revenue very diffi cult (Grunbaum 2007, p.93). The low levels of development among these countries tend to impact negatively on the breadth and levels of domestic tax revenue collections As such, the potential for these governments to offset lost revenue from tariff with domestic co llections is very minimal.

PAGE 48

CHAP TER 3 MODEL DEVELOPMENT This study intends to utilize the partial equi librium (PE) model to simulate the trade effects and fiscal impacts of trade liberalizati on for Malawi under the EPA arrangement with the EU. The Verdoorn (1960) version of the PE model, initially introduced in chapter two above will be used to undertake this empirical work. The choice of this model has been dictated by data considerations and the ex-ante nature of the anal ysis. As insinuated in the previous chapter, exante analyses can be conducted by either using the computable general equilibrium (CGE) or PE models. However, the enormous data requirements for the CGE models aspects of some of these data do not seem readily av ailable for Malawi has left us with the appropriate PE model. Notwithstanding their simplicity as compared to CGE models, PE models have been extensively used for analyses of this type, from which progressive policy recommendations have emanated. Specification of the Empirical Model1: The Verdoorn model is based on a couple of restrictive assumptions. First, the model applies the Armington assumption of product differentiation such that imported products from di fferent countries are a ssumed to be imperfect substitutes in use. For Malawi, we find this assu mption to be reasonable as the majority of the countrys imports are manufactured goods. Th e Verdoorn model is based on the partial equilibrium analysis and therefore makes the following other assumptions; trade flows changes do not have any repercussions on the incomes a nd exchange rates; iso-elastic import demand elasticities; and infinite supply el asticities. The last assumption s eems particularly appropriate in the case of Malawi vis--vis the EU as the latters exports to the former must account for a significantly small proportion of its total exports. On the other hand, exports from other smaller 1 The empirical model in this section is developed based on the one specified by Busse et al (2004) and Grunbaum (2007). 48

PAGE 49

Southern African Developm ent Corporation (SADC ) countries to Malawi may not support this assumption. However, we are captivated to ma intain the assumption on the understanding that exports to Malawi from these small SADC countri es do not constitute significant proportions of their total production. Nonetheless, we do recognize that those elasticities are less than infinite. To conduct an ex-ante analysis of trade effect s, one would consider a particular category of commodities, say M, and proceed to model the consumers behavior in the importing country. Following this reasoning; the utility function the consumer is assumed to maximize can be specified as follows; U = f [ fp(MP, MN), MD] (3-1) where fp = denotes a separable homogenous bran ch of the aggregat e utility function MP = Imports of commodity M from preference beneficiary sources MN = Imports of commodity M from non-preference beneficiary sources MD = Domestic production of commodity M The assumption of homogeneity in the specifi cation is quite consequential in that it implies that total imports of M, given by [MP+MN] are substituted equally for domestic production. The formal implication of this separabili ty of the utility functi on is that the Verdoorn model differentiates between the sources of imports, those from preference beneficiaries and those from non-beneficiarie s, labeled herein as MP and MN (Grunbaum 2007). There are two other crucial assumptions based on which the Verdoorn model is developed. First, that the dema nd function for the preference donor Malawi in our case, for a single category of products such as M, is given by the following specification; MP + MN = (3-2) 49

PAGE 50

where PP = the price of imports of M from the preference beneficiary sources PN = price of imports of M from the non-beneficiary sources = elasticity of import demand P = MP/(MP + MN) = Share coefficient of preferred imports N = MN/(MP + MN) = Share coefficient of non-preferred imports P + N = 1 Second, the definition of the elasticity of substitution ( ) between preferred and nonpreferred imports can be given as follows; = ; where is a parameter (3-3) As Malawi strikes off tariffs (t) under the EPA arrangement, only imports from the preference beneficiary countries will be affected, thus reducing their price PP while leaving the price for non-preferred im ports unchanged, such that We proceed to differentiate Equation 3-3, divide the resu lt by Equation 3-3 and using we derive the following; = = (3-4) After totally differentiating Equation 3-2 and dividing th e result by Equation 3-2, we derive; = + (3-5) Substituting definitions for and following Equation 3-2 into Equation 3-5, we obtain the following; 50

PAGE 51

+ = + log P (3-6) By using Equation 3-4, we proceed to rearrange Equation 3-6 to obtain the expression for the derivative of as follows; = (3-7) We next eliminate from Equation 3-4 by multiplying with and rearranging the equation to obtain the following; = (3-8) = (3-9) Next, we insert Equation 3-9 into Equation 3-6 and use Equation 3-7 to substitute the and obtain, = (3-10) Since log P is close to zero, if then Equation 3-10 can be expressed as follows; = (3-11) Now, the price for preferred imports can also be expressed in terms of export prices as follows; = (3-12) Where is the export price exclusive of the tariff ( t). The total derivative of Equation 3-12 gives the following; 51

PAGE 52

= + (3-13) We divide Equation 3-13 by Equation 3-12 to obtain the change in preferred import prices as follows; = + (3-14) By considering that = 0 and assuming infinite su pply elasticities, Equation 3-14 culminates into; = (3-15) We now utilize Equation 3-11 and Equation 3-15 to obtain the expression for the total expansion in imports from the preferred sources resulting from preferential liberalization as follows = (3-16) Since = (1 ), the total expansion in imports above can be re-written as = (3-17) As indicated in chapter two above, the tota l change in imports given by Equation 3-17 comprises the trade creation (TC) and trade diversion (TD) components. The former refers to the change in trade flow between the preference don or and preference beneficiary as consumers in the importing country substitute cheap imports for domestic production while the later concept captures the amount of preferred imports substituting those imports coming from non-preference beneficiary sources (Grunbaum 2007). We therefore separate Equation 3-17 into these two components as follows; 52

PAGE 53

TC = (3-18) And; TD = (3-19) We finally deal with the expected change in tariff revenue resulting from preferential trade liberalization of the sort we are concerned with. This change is given by the sum of the following: i) preferred imports multiplied by the preferential tariff rate and ii) the amount of non-preferred imports multiplied by the tariff rate applicable to imports from non-beneficiary sources (Grunbaum 2007). Thus, we expre ss the change in import duty revenue ( ID) as follows; ID = + TD (3-20) where ID = import duties tariff rate applicable on preferred imports = tariff rate applicable on non-preferred imports In this study, Equation 3-18, Equation 3-19 and Equation 3-20 will be utilized to estimate the trade effects and tariff revenue impact s of trade liberaliza tion under the EPA/PTA arrangement for Malawi vis--vis the EU. 53

PAGE 54

CHAP TER 4 DATA REQUIREMENTS, SOURCES AND TREATMENT The simulation and evaluation of the impacts of Malawi entering the EU-ACP Economic Partnership Agreement will be carried out us ing the Verdoorns (1960) Partial Equilibrium model. As indicated in the previous chapter, use of the import demand elasticities ( ) allows us to employ import data without relying on domestic production data. These domestic production data, if were to be used, have to be captured at a highly disaggregated level such as the eightdigit level of aggregation (i.e.; HS-8) which is being used in this study. As is the case with most other developing countries, such type of data were not readily available for Malawi so that use of import data was particularly nece ssary in this study. At the eight -digit level of da ta aggregation, the simulations permitted us to assess the effects of liberalization at product level and identify those products most impacted by the contemplated trade policy changes. To this effect, we managed to collect the nationa l eight-digit HS tariff and tr ade schedules from the Malawi National Statistical Office (NSO)1. These data were obtained for the baseline year of 20062. The raw data obtained from the NSO indica ted that in 2006, Malawi imported goods and services from the 27 EU countries using 1,602 ta riff lines compared to the total number of 3,956 lines utilized during the same year at th e HS eight-digit level of aggregation. Malawi has three major trade-related taxes, of which import duties are just one type. The other two types include excise du ties and import surtaxes. At every eight-digit tariff line utilized 1 These data are initially captured by the Malawi Revenue Authority, a Government Agency mandated to collect taxes. However, it is the National Statistical Office that has the legal right to process and disseminate all statistical information for the country. As such, the NSO obtains these data sets from the MRA on a monthly basis 2 The choice of this year was based on the planned dead line for EPA effectiveness which was initially scheduled to start in January, 2008. As such, the capture data were the most recent years for which the NSO had complete data sets as required by the study prior to the EPA signing deadline. 54

PAGE 55

during the baseline year for which data was obtained, the requisite CIF3 import value was captured along with its import dut y, excise duty and the import su rtax that were levied by the Malawi Revenue Authority. These data sets were obtained in the local currency, the Malawi Kwacha. As such, we converted a ll the Malawi Kwacha values into United States Dollars using the official average annual exchange rate as reported by the NSO and the Reserve Bank of Malawi. The rate reported and applied in this study was MK139.34 to US$ 1.00 for the baseline year of 2006. In line with the spirit of the study, two different scenarios of the scope of liberalization were contemplated, with each scenario changing the definition of which (or how many) countries constitute the preferred source of imports compar ed to the rest of the world. First, we assumed a scenario whereby Malawi enters into an EPA arrangement with the EU without any ancillary agreements with other trading partners outside the EU realm. For this scenario, the trade and tariff data were organized in such a manner that at every tariff line, imports from all the 27 EU countries were categorized as preferred imports as distinguished from those coming from elsewhere. Secondly, we built on the first scenario by contemplating a situation whereby Malawi correspondingly removes tariff barriers on imports from fifteen (15) other countries from the Southern Africa Development Corporation (SADC) region4. This is not a far-fetched assumption as negotiations for trade liberalization within this region have been on-going since the early 1990s. To capture this scenario, we added all imports from these fi fteen (15) member states into the basket of preferred imports from the EU. The significance of the second scenario can be well appreciated if one considers the fact that a si gnificantly higher proportion of Malawis trade is 3 Cost, Insurance and Freight aggregate value 4 The fifteen Member States of SADC include; Angola, Botswana, Democratic Republic of Congo, Lesotho, Madagascar, Malawi, Mauritius, Mozambique, Namibia, Se ychelles, South Africa, Swaziland, Tanzania, Zambia and Zimbabwe. 55

PAGE 56

conducted w ithin the SADC region. In fact, durin g the baseline year, Malawis imports from SADC accounted for about 58.2 pe rcent of total imports. The debate on whether Malawi must move forward and implement the EPA agreement with EU is of little value at th is point because such a decision wa s already made at an ACP level, even though the country has not moved forward with the implementation of the agreement. Meanwhile, Malawis trade with EU countries u tilizes the Everything But Arms (EBA) trading arrangement which is 100 percent unilateral such th at the EU can decide to terminate it anytime that they want to force Malawi along with other ACP least devel oped countries to migrate to the EPA arrangement. In such an event, Malawi woul d be left with just two options, enforcing the EPA or start utilizing the Gene ralized System of Preferences (GSP) framework. However, the preferences under the EPA agreement are far more rewarding to Malawi th an those offered under GSP framework. To this extent, our interest in the current study was to e xplore alternative ways through which Malawi could migrate to the EPA framework in the most rewarding manner. As such, analysis of the import tariff removal along the two broad liberaliz ation options explained above was intended to provide us with vital information regarding the costs and benefits of a more encompassing liberalization framework for th e country as compared to a lean targeted liberalization scenario. To this effect, trade creation, trade divers ion and fiscal impacts for both of the two contemplated liberalization scenarios were computed. It was envisaged that based on such results, the study will derive and descri be alternative trade policy options for the government. The last bit of data as per the applicable model in this study was with respect to two crucial variables for the elasticity of import de mand and the elasticity of substitution. Knowledge of these elasticities is an important ingredient in ex-ante analyses of trade policy reforms. 56

PAGE 57

Unfortunately, realistic figures for these two variab les a t the eight-digit level of data aggregation are non-existent for most of the developing c ountries, let alone Malawi. As such, strong assumptions about the levels of these elasticities were made. Firs t, an extensive survey of the empirical studies of import demand elasticities was made and one current such study was the one conducted by Sebastian Vollmer, et al (2009)5. In their study, Vollmer and colleagues employed empirical trade imports data to estimate the valu es of import demand elasticities for nine African countries6. The same data was used to simulate the trade effects of an EPA for each one of the nine countries. However, the calculations were neither done at each tariff line nor at each HS chapter but rather blocked into two major pr oduct categories including Non-manufactured goods (HS chapters 01 to 24) and manufactured goods (HS chapters 25 to 97). In this study, we analyzed the socio-economic realities and trade stru cture of Malawi vis--vis each one of the said countries. It was concluded that among the nine, Mozambique shares the most similarities with Malawi. Uganda was found to be next to Mozambique in being similar to Malawi as compared to the rest of the countries. As such, the impor t demand elasticities calculated by Vollmer and friends for Mozambique were adapted for use in this study. The adaptation was largely with respect to the assumed values of elasticities for the non-ma nufacturing category, which were adjusted upwards by 39 percent to take into account the move from the HS-0 to HS-2 category of the level of aggregation to the HS-3 to HS-8 category of the level of aggregation7. The basis for the 39 percent adjustment follows the empirical work of Kee, Nicita, and Olarreaga (2004) who estimated elasticities for different countries at the three and six digit level and found that on 5 Vollmer, et al (2009); EU-ACP Economic Partnership Agreements: Empirical Evidence for Sub-Saharan Africa 6 These Includes the following: Botswana, Cameroon, Cote d'Ivoire, Ghana, Kenya, Mozambique, Namibia, Tanzania, and Uganda. 7 Vollmer et all (2009) categorized the levels into thes e two and only calculated the elasticity of Import Demand for the Non-Manufacturing at the HS 0 to HS 2 levels. 57

PAGE 58

average, the latter estim ates were 39 percent higher than the former. The same principle was applied by Grunbaum (2007) in his dissertation research. A single (and not two) adjustment was made in line with Vollmer et als categorization of the HS chapters in their calculations. For sensitivity analysis, we used both of the values calculated by Vollmer and colleagues, i.e.; the one for Mozambique and that for Uganda. The ma jor results reported in this study are based on the Mozambique elasticity value since Malawi is far more similar to Mozambique than Uganda in terms of their GDP per capita as well as import structure. For the same reason, we included the mid-level value of import demand elasticity employed by Grunbaums study. Since Vollmer et al (2009) di d not calculate the levels of elasticity of substitution, we employed those levels used in Grunbaums study. Th is may at first sight appear to be a strong assumption, but as Vollmer et al (2009) observe many of the elasticity estimates across import categories and across these developing countries have very similar magnitudes8. We also employed three levels of magnitudes for this variab le to provide useful insights on the sensitivity of the simulations with respect to such changes. As such, the assumed values of elasticities employed in this study are as detailed in Table 4-1 below. Table 4-1. Assumed valu es of elasticities Product category Elasticity of Import Demand Elasticity of Substitution Vollmer study based Grunbaum study based Grunbaum Study Based Mozambique basis Uganda basis Grunbaum Mid Low Mid High Agricultural goods (01 24) -1.65271 -0.72558 -0.97 1.39 2.78 4.17 Manufactured & Semi-Man goods (25 97) -1.005 -0.684 -1.53 2.50 4.17 4.17 8 Vollmer et al (2009): EU-ACP Economic Partnership Agreements: Empirical Evidence for Sub-Saharan Africa ; pg 14. World Development Report: Background Paper. University of Gottingen, Germany. pg 14 58

PAGE 59

CHAP TER 5 EMPIRICAL RESULTS Utilizing the model developed in chapter thr ee and the data described in the previous chapter, we simulated trade effects and fiscal impacts under two broad lib eralization scenarios. The first of these reflects the outcome of Mala wis possible EPA with the EU while the second builds on the first by including an FTA with th e rest of SADC member states. One major assumption that has to be pointed out from the onset is with re spect to interpretation of the results vis--vis what would realistically be obt ainable. Under normal circumstances, the actual implementation of the proposed EPA between Mala wi and the EU would be expected to be phased out over some considerable time period such that parties to the agreement would be reducing tariffs on imports on a gradua l basis. It is not unusual for this period to take ten years or more. Besides, even full implementation of such agreements does not necessarily result into 100 percent reduction of existing tariffs on all produc ts as parties tend to retain some level of protection for some strategic commodities. An examination conducted by the WTOs Committee on Regional Trade Agreements suggests that FTAs typically cover betwee n 80 and 95 per cent of the trade between FTA members (WTO 2002; Busse et al 2004). As such, tr ade effects and fiscal impacts are also expected to be gradual in th at the full impact is only felt after complete elimination of tariff on imports is attained. The outcome of the simulations in this study assumes full and immediate elimination of import ta riffs by Malawi during the first year of implementation. The results therefore capture an end-period and/or long te rm static effect of complete tariff liberalization through the im plementation of the EPA and/or SADC FTA by Malawi. 59

PAGE 60

Trade Effects Sim ulation results indicating the trade effects for the two broad scenar ios are presented in Table 5-1 and Table 5-2 below. As discussed in the previous chapters, va lues for trade creation, trade diversion and total trade e ffects were calculated within the realms of each one of the two contemplated trade liberalization frameworks. In particular, Table 5-1 pres ents the trade effects of an EPA with the EU while maintaining th e status quo on all other trading partners. As highlighted in Table 5-1, the simulated net trade creation values were nega tive, and significantly so, for all assumed elasticity levels. This implies that an EPA with the EU alone would result in significant levels of trade divers ion for Malawi, shifting the stru cture of imports source away from non-EU trading partners and towards EU me mber states. The simulations further illustrate that implementation of the EPA would result in significant increases in EU imports into Malawi, with the least projected increase of 16.5 percent for the lowest elasticity value and 26.9 percent for the mid-range of the most favored elasticity value9 in this study. These huge increases are largely reflective of the significantly high pr e-simulation tariff rates on EU imports. As compared to the baseline total imports va lue, the simulations demonstrated that ex-post imports from the EU would increase by percentage levels in the range of 2.4 to 4.7 percent, with an increase of about 4 percent based on the most favored elasticity value in this study. Table 5-2 presents the trade effects of impl ementing a broader trade liberalization policy, which includes SADC member states by way of implementing a SADC FTA concurrently with the implementation of the EPA with the EU. As indicated in the table, the absolute values of simulated trade effects were much higher in this case as compared to the first scenario. Most 9 As indicated in chapter 4, the elasticity values based on Mozambique figures as calculated in the Vollmer, et al study are recommended in this study. Other values provide useful insights into deducing the sensitivity of the simulations to changes in elasticity values. 60

PAGE 61

61 importantly though is the fact that for all but two elas ticity scenario sets, trade creation values were significantly higher than trade diversion figures. As such, for all but those two scenarios, trade effect simulations resulted in substantia lly higher positive net trad e creation values that ranged from about US$ 20 million for the low Uga ndan based elasticity scenario case to US$ 85 million for the low-case of the Grunbaum elasticity scenario level. The to tal net trade creation value for the mid-case of the most favored elas ticity value in this study was US$ 27 million. In terms of ex post trade flow effects, the simulations show ed that imports from the EU and SADC region would increase by about 20.2 percent fr om the baseline year value, based on the mid scenario case of the study prefe rred elasticity value. Variation of the elasticity value along the other scenarios demonstrated that imports would increase in th e range of 12.7 to 25.7 percent as compared to the baseline year import value. As the table further demonstrates, the increases in imports ranged from about 9.4 percent to 18.9 pe rcent vis--vis total im ports obtainable during the baseline year. It is worth noting that large increases in imports were simulated under this trade liberalization scenario as compared to th e case where only EU imports benefited from a preferential tariff regime. This is largely because of the fact that about half of Malawi imports come from within the SADC regi on such that inclusion of the same in tariff liberalization was found to have significantly encouraged increased intra-region trade. As insinuated in the previous chapters, economists view trade creation as welfare improving since consumers in the importing country tend to substitute more costly domestically produced goods with low cost imports from the beneficiary country. On the other hand, trade diversion entails displacement of the low cost imports from non-benefici ary trading partners by the expensive imports from beneficiary countries, which would be EU in our case. As such, and on the basis of the simulation result s describe above, it was noted that an EPA with the EU alone

PAGE 62

Table 51. Trade effects from tariff e limination by Malawi on EU imports only Basis of Import Demand Elasticity Scenario setting for the level of Elasticity of Substitution Trade creation in US$ (TC) Trade diversion in US$ (TD) Net trade creation in US$ (TC TD) Total trade effects in US$ (TC +TD) Total trade effects as a percentage of preferred imports Total trade effects as a percentage of total imports Mozambique based Low 12,501,056.99 23,686,280.19 (11,185,223.20) 36,187,337.18 20.51 3.04 Mid 12,501,056.99 34,934,612.13 (22,433,555.14) 47,435,669.12 26.89 3.98 High 12,501,056.99 35,802,198.19 (23,301,141.20) 48,303,255.19 27.38 4.05 Uganda based Low 8,000,182.30 21,112,259.31 (13,112,077.02) 29,112,441.61 16.50 2.44 Mid 8,000,182.30 32,360,591.25 (24,360,408.95) 40,360,773.55 22.88 3.39 High 8,000,182.30 33,228,177.31 (25,227,995.02) 41,228,359.61 23.37 3.46 Grunbaum based Low 17,064,266.00 26,523,566.25 (9,459,300.25) 43,587,832.25 24.71 3.66 Mid 17,064,266.00 37,771,898.18 (20,707,632.18) 54,836,164.19 31.08 4.60 High 17,064,266.00 38,639,484.25 (21,575,218.25) 55,703,750.26 31.58 4.68 62

PAGE 63

63 Table 52. Trade effects from tari ff elimination by Malawi on imports from EU and SADC member states Basis of Import Demand Elasticity Scenario setting for the level of Elasticity of Substitution Trade creation in US$ (TC) Trade diversion in US$ (TD) Net trade creation in US$ (TC TD) Total trade effects in US$ (TC +TD) Total trade effects as a percentage of preferred imports Total trade effects as a percentage of total imports Mozambique based Low 102,763,870.98 51,600,199.21 51,163,671.77 154,364,070.18 17.46 12.84 Mid 102,763,870.98 76,106,790.11 26,657,080.87 178,870,661.09 20.23 14.88 High 102,763,870.98 77,943,951.09 24,819,919.89 180,707,822.07 20.44 15.03 Uganda based Low 66,416,462.37 46,017,395.46 20,399,066.91 112,433,857.84 12.72 9.35 Mid 66,416,462.37 70,523,986.37 (4,107,523.99) 136,940,448.74 15.49 11.39 High 66,416,462.37 72,361,147.35 (5,944,684.97) 138,777,609.72 15.69 11.54 Grunbaum based Low 142,798,799.74 57,824,479.58 84,974,320.15 200,623,279.32 22.69 16.69 Mid 142,798,799.74 82,331,070.49 60,467,729.25 225,129,870.22 25.46 18.73 High 142,798,799.74 84,168,231.47 58,630,568.27 226,967,031.20 25.67 18.88

PAGE 64

would result in net welf are losses since trade diversion was greater in all scenarios than the welfare improving trade creation values. C onversely, an EPA with the EU that is implemented together with a FTA with the SA DC region was seen to be welfare improving in all but two cases of mid and high assu med elasticity values based on the Ugandan figures as calculated by Vollmer and colleagues. The significant posi tive net trade creation simulation results obtained under the second liberalization framew ork would lead to increased overall welfare in Malawi. Fiscal Impacts Table 5-3 below captures the contribution of trade taxes in general and im port duties in particular towards total governme nt revenue in Malawi. As indicated, the contribution of revenue from trade taxes to wards total government revenue was about 36.6 percent, inclusive of grants and 40.2 percen t excluding grants in the year 2006. During the same year, import duties accounted for 11.2 perc ent and 12.5 percent of total government revenue, including grants and excluding grants, respectively. As such, elimination of tariff barriers on imports from the EU and/or EU and SADC at the same time is bound to have a significant impact on government revenue. Using equation 3-20 developed in chapter three above, fiscal impacts of the two trade liberalization scenarios were calculated. It is important to point out that simulation results in this study only quantif ied the direct and/or static levels of changes in ex-post revenue as a result of lost revenues due to tariff elimination under the two scenarios. This part of fiscal impact is clearly outlined as reported in Table 5-4 below. As indicated in Table 5-4, loss of tariff revenue under the EPA arrangement with the EU alone is significantly lower compared to the scenario that includes an FTA arrangement with 64

PAGE 65

SADC. The revenue losses under the EU EPA we re simulated to range from 11.2 percent to 12.1 percent of the baseline tota l tariff revenue. In terms of comparison to total government revenue, the highest of these values was only 4.8 percent of total government revenue excluding grants and 5.2 percent of total gove rnment revenue including grants for the baseline year. Table 53. Baseline levels of impor t duties and impact on government revenue Total government revenue ex cluding grants (US$) 489,708,626.38 Total government revenue including grants (US$) 537,861,346.35 Total trade taxes levied 196,594,338.65 Trade taxes as percentage of government revenue excluding grants 40.15 Trade taxes as percentage of govern ment revenue including grants 36.55 Total import duties levied (US$) 61,299,189.69 Import duties as percentage of govern ment revenue excluding grants 12.52 Import duties as percentage of govern ment revenue including grants 11.19 The fiscal impact results tend to increase substantially when the SADC FTA arrangement is included in the liberalization framework, with percentage reductions of up to 30.5 percent of total revenue during the baseline year. The fa ct that imports from SADC form about half of the total imports for the c ountry helps to explain th ese drastic changes in revenue. It is important to point out that trade crea tion brings with it othe r revenue in form of other trade taxes that are not affected by the tariff elimination policy. Considering the large values of trade creation under the EU plus SADC FTA tier as desc ribed above, one would expect to get extra revenue that would work towards offsetting the static figures reported above. As such, the total changes in end-peri od levels of revenue would be significantly lower than those reported in Table 5-4 above because of the extra revenue expected from other trade taxes. 65

PAGE 66

Table 54. Summ ary of static fiscal impacts Elasticity scenario setting Tariff elimination to EU imports only Tariff elimination to EU and SADC imports Change (US$) in tariff revenue As a percent of baseline level Change (US$) in tariff revenue As a percent of baseline level Mid Mozambique 24,610,959.11 11.6 161,921,417.70 30.1 Mid Uganda 23,748,488.89 11.2 159,549,735.08 29.7 Mid Grunbaum 25,624,080.19 12.1 164,123,067.17 30.5 Commodities Most Affected by the Contemplated Changes in Trade Policy One of the major advantages of the Part ial Equilibrium Model for simulating the likely trade effects and fiscal impacts of an EPA and/or an FTA usi ng highly disaggregated data such as the HS-eight digit utilized in th is study is that it a llows identification of commodity categories most affected by the c ontemplated trade polic y change. Knowledge of such information would be crucial for t hose engaged in trade negotiations as it would provide necessary guidance on which product cat egories are more sensitive to warrant special treatment during implementation of the EPA framework. In this study, an effort was made to capture the simulated trade and tari ff effects at each individual tariff line along with their associated HS-code and descriptions The results column was thereafter sorted a descending order to identify t hose product categories most af fected by the contemplated trade policy changes. This permitted us to isolated twenty commodity categories most affected by the policy change under the two liber alization scenarios were identified. Tables A-1 through A-4 in Appendix A present results of these commodity categories along with their simulated values for trade effects and tariff revenue changes. Table A-1 and Table A-2 present lists of the most sensitive commodities to an impending EPA with the EU by Malawi in terms of the trade effects and fiscal impacts, respectively. As indicated, simulation results demonstrated that some of the most sensitive 66

PAGE 67

products include the following; clothing m aterials, four-wheel drive motor vehicles, solid milk and cream products, various iron sheet ma terials, electrical ap pliances and telephony and telegraphic apparatus. The tables further detail th e associated values of estimated trade effects (Table A-1) and those for the estimated changes in tariff revenue (Table A-2). It is important to note that most of the commodities that were found to have high trade effects levels were also identified to have high absolute tariff revenue changes. The other two tables, Table A-3 and Table A-4 present the same information as the one in Tables A-1 and Table A-2 but for the second liberalization scenar io that includes a FTA with the rest of the SADC member stat es. Under this second scenario, some of the product categories with highest tr ade effects include; distillates and other fuels, four wheel drive motor vehicles, rubber tires used on busses and lorries, worn clot hing materials, flatrolled iron and steel products, crude soybean oil, cement clin kers and tobacco cigarettes. Associated values of trade effects are detailed in the tables. Three Earmarked Agricultural Products One of the major objectives of this study was to look at the trad e effects and fiscal impacts of the EPA with EU on key agricu ltural commodities. The three agricultural products of tobacco, tea and sugar were earmarked for such a close evaluation of the extent to which trade policy change affects their ex -post trade flow levels and tariff revenues. Table 5-5 presents the impact of an EPA with EU on tea, sugar and tobacco trade flows and tariff revenue contributions. Table 5-6 pres ents the same information for a broader liberalization scenario that includes the SADC FTA. For each liberalization scenario, all the disaggregated trade and fiscal effects at ev ery individual tariff lin e relating to each one of the three targeted ag ricultural products were isolated and aggregated into a product total effect as captured in Table 5-5 below. These simulations utilized the same elasticities that 67

PAGE 68

were em ployed in the entire study for all tariff lines relating to agricultural products from HSchapters 01 through 24. The ideal way wa s to use product specific elasticities of import demand and substitution. These were howev er not available for Malawi, forcing us to employ the sector wide elasticities. Table 55. Impacts under the EU liberaliz ation on three key ag ricultural commodities Commodity Tariff revenue change (US$) As percent of total Avg trade effects (US$) As percent of total Tea 69,497 0.28 125,806 0.26 Sugar 8,882 0.04 27,498 0.06 Tobacco 0 N/A 0 N/A The basis for the choice of these products wa s two-fold. First, pundits in most of the ACP countries in general and Malawi in partic ular have argued that an EPA with the EU would invariably lead to suffocation of the ag ricultural sectors in the ACP countries which is the major source of livelihood and economic de velopment in these countries. Second, the three agricultural commodities form the bulk of Malawi exports, with tobacco exports accounting for over 60 percent of export earnings. As such, investigating the sensitivity of these products to policy changes was deemed crucial. Table 56. Impacts under the EU & SADC liberalization on th ree key agricultural commodities Commodity Tariff revenue change (US$) As percent of total Avg trade effects (US$) As percent of total Tea 324,160 0.20 365,706 0.20 Sugar 312,977 0.19 709,613 0.40 Tobacco 2,642,279 1.63 2,361,947 1.32 As it turns out, simulation results demonstrate that trade effects and fiscal impacts on tea and sugar were found to be quite insign ificant under both liberalization frameworks. Tobacco was found to have some modestly high trade effects and revenue changes only under the second liberalization framework. Under the EPA arrangement with EU, no 68

PAGE 69

69 effects were found. Zero effects were estimated under the EPA arrangement because all of the tobacco imports into Malawi come from non-EU countries of Mozambique, Zimbabwe, Zambia, Kenya, Mauritius and the United Arab Emirates. The empirical results presented above do not support the long-standing fear among policy makers and trade specialists in ACP c ountries that an EPA with the EU is bound to stifle the local agricultural sector with an influx of cheap EU competing agricultural imports. In fact, of the twenty commodities id entified to be most affected by the EU EPA, only the two product categories of milk a nd cream in solid form (both HS 04021000 and HS 04022100) were part of the list, being projected to increase in ex-post imports from the EU by 2.7 and 2.3 percent, respectively, vis-vis baseline levels. Under the EPA plus SADC FTA framework, crude soybean oil (H S-15071000), milk and cream in solid form (HS 04022100), crude palm oil (HS-15111000) were the only agricultura l products within the list of most affected pr oduct categories in terms of proj ected increases in imports.

PAGE 70

CHAP TER 6 SUMMARY AND CONCLUSIONS Summary Malawi has been party to the EU-ACP trade agreements since the first Lom Agreement in 1975. Meanwhile, these ACP countries are expected to implement the 2000 Cotonou Agreement that calls for migration away from the non-reciprocal trading framework to a reciprocal trading arrangement between the two parties. As such, Malawi along with other ACP group of countries is expected to liberali ze their trade through elimination of tariffs on EU imports. This has spurred concerns among ACP countries who fear that such an arrangement will subject domestic producers to unfair competition from the highly subsidized agricultural imports from the EU for instance. As such, it is generally feared that these changes may bring about increased unemployment, provoke heightened economic insecurity and political instability. In most cases, these fears have not been substantiated with empirical evidence as ve ry limited empirical c ountry-specific studies have so far been done to quantify the potential effects. It is on the basis of the foregoing that the major objective of this research was to conduct a quantitative analysis of the likely trade effects and fi scal impacts of an EPA trade arrangement regime with Malawi. It was deemed crucial to quantify potential levels of trade creation, trade diversion and fiscal impacts to the country emanating from an EPA with the EU. Ancillary to this effort was the need to explore alternative trade policy arrangements, particularly the broadening of the liberalization framework. We further undertook to identify those commodities that wo uld be most affected by the contemplated policy changes in trade. 70

PAGE 71

In order to enrich our basis for the empi rical study, an extens ive survey of the requis ite literature on the theoretical and empi rical works related to the subject matter was conducted. It was noted that classical trade theory argues for a positive correlation between free trade policy and increased economic growth and development through direct dynamic advantages of increased capacity utilization and more efficien t investment projects as well as indirect effects of accelerated export growth. It was observed however that other researchers object to the classical perspectiv e and argue that empirical evidence does not strongly support such a view. Th ese researchers also question th e validity of the direction of the casual relationship between export grow th and economic development as purported by the classical theorists. In spite of the arguments and counter-arguments among researchers on the existence and/or strength of the direct link be tween open trade and positive economic development, it was observed th at open trade does bring with it elements that spur competition among producers and im poses rationality in productive resource allocation, thereby increasing economic growth and development. On the other hand however, unambiguous evidence of poor econom ic performance in closed regimes was observed and noted. The evolution of the notion of preferential trading agreements was noted to have emanated from the frustration during the 1990s by many countries to attain multilateral free trade. As such, countries started configuring themselves into regional groupings to pursue regional free trade policies. Th is influenced economists, led by Jacob Viner (1950) to start studying and laying the earliest ba sis on the theory of regional free trade agreements. Viner (1950) observed that welfare effects from a ny form of regional economic integration are not unambiguous a priori. He further develope d the concepts of tr ade creation and trade 71

PAGE 72

diversion associated with the form ation of free trade areas. In this study, these two concepts were central to the analyses and conclusions to be made. Viners work spurred a lot of theoretical as well as empirical work on the study of free trade areas. Literature identifies three major quantitativ e analytical techniques that researchers employ to conduct empirical analyses of the e ffects of entering some form of an economic integration by countries. These include the gr avity models, computable general equilibrium (CGE) models and the partial equilibrium (PE) models. Strengths and weaknesses for each one of these models were highlighted. This thesis utilized the Ver doorn (1960) version of the PE model to quantify the trade effects and fiscal impacts of the EU EPA with Malawi. Central to the fears among developing count ries to enter into free trade regimes are the fiscal implications of such policies. Most of these countries fear that through elimination of import tariffs, the government would experience huge revenue losses and compromise the delivery of vital services to the populace. Consideration of several short term policy migration costs to the local economies such as increased unemployment, reduction in national output, elimination of certain domestic industries and possible macroeconomic instability was observed to have exacerbated these fears. Research into this aspect has been centered on two them es, including the evaluation of the relative importance of tariff revenue with respect to total government revenue and exploration of the fiscal reforms and alternative tax scheme s that could offset tariff revenue losses. Conclusions The Verdoorn (1960) version of the partia l equilibrium models was employed to estimate and evaluate the trade effects and fi scal impacts of an EU EPA on the Malawi economy. This model is based on an array of ra ther restrictive assumptions such as the Armington assumption of product differentiation, and those applicable to most partial 72

PAGE 73

equilibrium analyses includi ng; non-existence of reperc ussions on the incomes and exchange rates from changes in trade flow s; existence of iso-elastic import demand elasticities; and infinite supply elasticities. Notwithstanding these restrictive assumptions, it was noted that this model has been extens ively applied in similar works by various researchers and academicians alike. In this study, highly disaggregated trade data at HSeight digit were utilized to conduct the analyses. The major conclusion that was drawn from this study is that Malawi stands to benefit more from trade if a broader libera lization framework is adopted along with the EPA enforcement. As reported above, trade lib eralization of tariffs towards EU and SADC result in significantly high positive net trade crea tion levels for all but two elasticity values. On the other hand, liberalization that is restricted to EU im ports results in negative net trade creation values. This conclusion stands firm even when one considers the major weakness of this study, namely; that the va lues of elasticities employed were quite arbitrary. This is the case because even as we varied the magnitude of the elasticity values, a broader liberalization framewor k resulted in higher positive levels of net trade creation values as compared to the negative net trade creation values obtained under the liberalization framework restricted to the EU EPA. The breadth of liberalization need not necessarily be restrict ed to EU and SADC countries. Ther e are increasingly high levels of imports from the eastern countries of Japan, Ch ina and India. Inclusion of these countries in the tariff liberalization framework is bound to result in increased valu es of trade creation. While not supporting the conven tional fears about trade liber alization in most of the developing world, the results from this resear ch do conform to the theoretical arguments for broader liberalization even for small countries like Malawi. 73

PAGE 74

Sim ilar results were obtained by Gru nbaum (2007) who found that when liberalization by the Organization of Eastern Ca ribbean States (OECS) with the EU alone was considered, trade diversion exceeded trade creation in all cases and overall trade effects and fiscal impacts were very small. When the larger Caribbean, the North American Free Trade Area (NAFTA), and the Free Trade Area of the Americas (FTAA) were all included in the liberalization framework, net trad e effects were positive and larger, as were the fiscal impacts. Our results also agree with those obtai ned by Keck and Piermartini (2005) who used an applied general equilibrium model covering 15 regions and 9 sectors to simulate the impact of signing EPAs with the EU for the SADC countries under various liberalization scenarios. The results demonstr ated that EPAs with the EU were welfareenhancing for SADC overall, and that for mo st countries, further gains could arise from intra-SADC liberalization. Greenway and Milner (2003) also found compar able results when they analyzed the impact of EPA formation for CARICOM count ries. Employing two-digit level of data aggregation, Greenway and Milner (2003) contem plated three liberalization frameworks in their study including the one rest ricted to the EU, the second combining the EU and the United States and the last one being full multila teral liberalization. Th eir results concluded that the EPA formation with the EU result ed into net welfare losses by the CARICOM countries involved where as the one under th e last two scenarios resulted in positive welfare gains with the full multilateral liberalizat ion resulting in the highest welfare gains. The other major conclusion of this study is with respect to the effects of EPA formation on the domestic agricultural sector, mostly the fear of an influx of cheap EU 74

PAGE 75

im ports. Against these conventiona l fears, results from this research demonstrate that very minimal effects on agricultural imports from the EU would be felt, save for imports of milk and cream in solid form. There were particularly no major effects on im ports of tea, sugar and tobacco, which are currently the major agricultural export commodities for Malawi. The fiscal impacts from trade liberalization do seem be significant if considered in isolation. However, the simulation results only capture static effects at 100 percent tariff elimination scenario. With the high positive trade creation values obtained from the simulation results, one would e xpect a significant offsetting effect if dynamic effects of extra revenue from increased trade are incl uded. Besides, policy makers may consider exploring other tax systems that will help to broaden the countrys tax base. In fact this subject has been at the centre of discussions within the coun try, considering the high levels of dependence on trade taxes for government revenue under the current system. This research has laid down a very usef ul foundation into research on trade effects and fiscal impacts of trade liberalization at a country specific level that employs disaggregated data at the level utilized he rein. Until this work, we do not know of any effort directed towards empirical quantificati on of such effects in the country. The results from this thesis will provide useful guidance to the on-going debate with respect to the pros and cons of the EU-EPA formation. The identi fication of products most affected by the policy change will exonerate fears of unfair competition from EU agricultural imports while alerting those engaged in the most affected products to explore remedi al measures. Study Weaknesses and Suggestions for Further Research One of the major weaknesses of the Partial Equilibrium models and particularly the one carried over into this thesis is with re spect to the choice of elasticities of import demand and substitution. These are very cruc ial variables in the Verdoorn (1960) model 75

PAGE 76

and yet were strong assumptions were m ade in the process of determining their values since they were not available at any disaggregated level. Th e values chosen were quite reasonable considering that most of the prev ious researchers have not employed similarly adapted values. Nonetheless, the ideal scenario is to use the study data to estimate elasticity values at the same level of data aggregation, preferably every tariff line or HS chapter applicable. Through this rigorous process one would get empirically deduced product specific elasticity values to employ in the analysis of trade and fiscal effects of any form of trade liberalization that might be contemplat ed. Notwithstanding its enormous financial and technical demands, such resear ch is highly recommended for Ma lawi as the results from the same would help solidify th e foundation laid in this researc h. In particular, the levels of trade and fiscal effects simu lated for each product category su ch as those relating to the three agricultural products of tea, sugar a nd tobacco reported above would be much more realistic if one employs such product specific el asticities. In terms of data availability, the countrys National Statistical Office has made significant strides in building requisite databases. It would also be important to build in these studies an aspect of dynamic effects of liberalization by way of employing the comp utable general equilibrium models, which would take into account economy-wide repercus sions of the policy shift and adjust the trade and fiscal effects accordingly. 76

PAGE 77

APPENDIX MOST AF FECTED PRODUCTS BY TRADE AND TARIFF REVENUE IMPACTS Table A1. List of commodities with the highest trade effects as a result of tariff elimination on EU imports Number HS code Item description Average trade effects (US$) As percent of total 1 63090000 Worn Clothing and Other Worn Articles 2,559,500.98 5.82 2 85175000 Apparatus for Carrier-Current line System or for Digital line 1,429,916.75 3.25 3 87032310 Other Motor Vehicle Four Wheel Drive 1,366,699.75 3.11 4 72104900 Flatrolled Iron/Steel, Width >=600MM,Zinc Plated/Coated(EXC.ELECT 1,359,846.10 3.09 5 04021000 Milk and Cream in Solid Forms Of =<1.5% FAT 1,179,355.59 2.68 6 85173000 Telephonic or Telegraphic Switching Apparatus 1,046,776.36 2.38 7 04022100 Milk and Cream in Solid Forms OF >1.5% FAT, Unsweetened 997,329.54 2.27 8 87033390 Other (Four-Wheel Drives) 994,556.33 2.26 9 34012000 Soap in Other Forms, NES 943,934.71 2.15 10 63079000 Made up Articles (INCL. Dress Patterns), NES 760,917.67 1.73 11 49119990 Other Printed Matter N.E.S 633,467.28 1.44 12 85252090 Other Transmission Apparatus 628,227.38 1.43 13 87033310 Four-Wheel Drive Vehicle With Diesel Engine Of CC >=2500CC 569,137.68 1.29 14 87032331 Other Vehicles With Compression-Ignition Internal Combusion DI 558,099.76 1.27 15 85252020 Communication Transmitters, Transceivers and Ancillary Apparatus 489,950.36 1.11 16 87033210 Other Vehicles Of Cylinder Capacity Not Exceeding 2000CC Four 480,025.02 1.09 17 85178000 Electrical Apparatus For Line Telephony or Line Telegraphy, NE 457,214.30 1.04 18 87089990 Other Specialized Parts For Tractors 415,642.43 0.95 19 87032321 Four-Wheel Drive Vehicle with Spark-Ignition Engine Of CC 1500-3000CC 394,068.76 0.90 20 85042100 Liquid Dielectric Transformers, Power Handling Capacity =<650K 393,063.54 0.89 77

PAGE 78

Table A2. List and levels of commodities with highest ta riff revenue changes as a result of tariff elimination on EU i mports Number HS code Item description Change in tariff revenue (US$) As percent of total 1 63090000 Worn Clothing and Other Worn Articles 2,473,559.29 10.05 2 87032310 Other Motor Vehicle Four Wheel Drive 1,456,770.59 5.92 3 85175000 Apparatus for Carrier-Current line System or for Digital line 1,393,168.88 5.66 4 04022100 Milk and Cream in Solid Forms OF >1.5% FAT, Unsweetened 1,073,772.67 4.36 5 49119990 Other Printed Matter N.E.S 798,375.10 3.24 6 87033390 Other (Four-Wheel Drives) 713,339.23 2.90 7 72104900 Flatrolled Iron/Steel,Wid.>=600MM,Zinc Plated/Coated(EXC.ELECT 692,189.40 2.81 8 87032331 Other Vehicles With Compression-Ignition Internal Combusion DI 543,359.65 2.21 9 87033210 Other Vehicles Of Cylinder Capacity Not Exceeding 2000CC Four 499,694.83 2.03 10 34012000 Soap in Other Forms, NES 485,323.56 1.97 11 85252090 Other Transmission Apparatus 455,135.26 1.85 12 84729000 Office Machines, NES(INCL. CoinSorting/Counting/Wrapping Mach 447,812.58 1.82 13 87089990 Other Specialised Parts for Tractors 375,652.93 1.53 14 87032321 Four-Wheel Drive Vehicles with SparkIgnition Engine Of CC 1500-3000CC 365,959.29 1.49 15 87032400 Vehicles with Spark-Ignition Engine of Cylinder Capacity >=300 360,275.74 1.46 16 21021000 Active Yeasts 338,056.80 1.37 17 87033310 Four-Wheel Drive Vehicle with Diesel Engine Of CC >=2500CC 335,065.24 1.36 18 85252020 Communication Transmitters, Transceivers and Ancillary Apparatus 326,255.29 1.33 19 48115100 Bleached Weighing More Than 150g/m2 290,088.29 1.18 20 85021100 Generating Sets with Compression Ignition Engines, =<75 KVA 286,270.84 1.16 78

PAGE 79

Table A3. List of commodities with the highest trade effects as a result of tariff elimination on imports from EU & SADC countires Number HS code Item description Average trade effects (US$) As percent of total 1 27101199 Distillates and other fuels nes(including diesel oils,gas oil 19,329,820.98 11.28 2 87033310 Four-Wheel Driv e Vehicles with Diesel Engine Of CC >=2500CC 5,138,646.05 3.00 3 40112000 New Pneumatic Tyres, Of Rubber Of a Kind Used on Buses or Lorries 4,024,071.99 2.35 4 63090000 Worn Clothing and Other Worn Articles 3,654,699.51 2.13 5 72104900 Flatrolled Iron/Steel, WID.>=600MM,Zinc Plated/Coated(EXC.ELECT 3,320,894.25 1.94 6 87032310 Other Motor Vehicles Four Wheel Drive 3,317,019.91 1.94 7 87033210 Other Vehicles Of Cylinder Capacity Not Exceeding 2000CC FOUR 3,219,850.34 1.88 8 15071000 Crude Soya-Bean Oil 2,984,694.08 1.74 9 25231000 Cement Clinkers 2,274,888.34 1.33 10 87042110 GVW Not Exceeding 2.99 Tonnes 2,241,255.55 1.31 11 27101919 Gases and Lubricating Oils 2,126,549.46 1.24 12 39011000 Polyethylene Having a Specific Gravity <0.94, In Primary Forms 2,066,537.93 1.21 13 24022000 Cigarettes Containing Tobacco 1,928,517.27 1.13 14 34012000 Soap In Other Forms, NES 1,728,086.55 1.01 15 04022100 Milk and Cream in Solid Forms Of >1.5% FAT, Unsweetened 1,712,645.59 1.00 16 85252090 Other Transmission Apparatus 1,662,007.39 0.97 17 15111000 Crude Palm Oil 1,650,016.29 0.96 18 87029020 Other Motor Vehicles for the Transport of Ten or More Persons 1,644,461.20 0.96 19 85175000 Apparatus for Carrier-Current Line Systems or For Digital Line 1,598,042.62 0.93 20 87033390 Other (Four-Wheel Drives) 1,426,783.77 0.83 79

PAGE 80

80 Table A4. List and levels of commodities with highest ta riff revenue changes as a result of tariff elimination on imports from EU & SADC countries Number HS Code Item Description Change in tariff revenue (US$) As percent of total 1 27101199 Distillates and other fuels nes(including diesel oils, gas oil 26,528,783.18 16.38 2 87033310 Four-Wheel Drive Vehicles with Diesel Engine Of CC >=2500CC 8,203,313.08 5.07 3 87033210 Other Vehicles with Cylinder Capacity Not Exceeding 2000CC FOUR 3,545,208.91 2.19 4 87032310 Other Motor Vehicles Four Wheel Drive 3,478,076.27 2.15 5 40112000 New Pneumatic Tyres, Of Rubber of a Kind Used on Buses or Lorries 2,970,721.26 1.83 6 63090000 Worn Clothing and Other Worn Articles 2,700,024.24 1.67 7 25231000 Cement Clinkers 2,659,825.23 1.64 8 27101919 Gases and Lubricating Oils 2,550,591.47 1.58 9 24022000 Cigarettes Containing Tobacco 2,348,131.82 1.45 10 15071000 Crude SoyaBean Oil 2,089,339.84 1.29 11 87042110 GVW Not Exceeding 2.99 TONNES 2,007,778.07 1.24 12 55142100 Dyed Plain Weave Fabrics, <85% Polyester Fibres + Cotton, >170 1,882,079.78 1.16 13 72104900 Flatrolled Iron/Steel,WID.>=600MM,Zinc Plated/Coated(EXC.ELECT 1,784,902.09 1.10 14 87032321 Four-Wheel Drive Vehicles with SparkIgnition Engine Of CC 1500-3000CC 1,735,524.63 1.07 15 85175000 Apparatus for Carrier-Current Line Systems or for Digital Line 1,630,463.53 1.01 16 87033390 Other (Four-Wheel Drives) 1,469,126.14 0.91 17 52093200 Dyed 3 or 4 Thread Twill (Inc. Cross Twill), With >=85% COTTON 1,449,647.51 0.90 18 27101159 Other Kerosene 1,402,972.46 0.87 19 87029020 Other Motor Vehicles for the Transport of Ten or More Persons 1,309,333.56 0.81 20 33021000 Mixtures/With Basis of/Odorifer's Substitutes Incl. Alcohol Solutions 1,286,655.85 0.79

PAGE 81

LIST OF REREFE NCES Andriamananjala, S; P. Brenton; J.E. V on Uexkll and P. Walkenhorst. 2009. Assessing the Economic Impacts of an Economic Partnership Agreement on Nigeria. World Bank African Region Policy Research Wo rking Paper, Washington, D.C. Balassa, B. 1971. The structure of protection in developing countries Baltimore, MD: Johns Hopkins University Press. 1982. Development strategies in semi-indus trial economies. NY and London: Oxford University Press. Baldwin, R.E. and T. Murray. 1977. "MFN Tari ff Reductions and Developing Country Trade Benefits Under the GSP." The Economic Journal 87(345):30-46. Berg, A. and A. Krueger. 2003. "Trade, Growt h, and Poverty: A Selective Survey." IMF Working Paper no. 03/30, Internationa l Monetary Fund, Washington, D.C. Bhagwati, J. 1978. Anatomy and consequences of exchange control regimes Cambridge, MA: NBER. Bhagwati, J. and A. Panagariya. 1996. Preferen tial Trading Areas and Multilateralism-Strangers Friends, or Foes? In J. Bhag wati and A. Panagariya,.eds. The Economics of Preferential Trade Agreements Washington D.C: American Enterprise Institute, pp.1-78. 1999. Preferential Trading Areas and Multila teralism-Strangers, Friends or Foes? In J. N. Bhagwati, et al., eds. Trading Bl ocs: Alternative Approaches to Analyzing Preferential Trading Arrangements. Cambridge, Massachusetts, MIT Press. Busse, M; A. Borrmann and H. Grossmann. 2004. The Impact of ACP/EU Economic Partnership Agreement on ECOWAS Countries: An Empiri cal Analysis of the Trade and Budget Effects. Study Final Report to the Hamburg Institute of International Economics; Hamburg, Germany. Cooper, C.A. and B.F. Massell. 1965a. "A New Look at Customs Union Theory." Economic Journal 75(300):742-747. 1965b. "Toward a General Theory of Cust oms Unions for Developing Countries." Journal of Political Economy 73(5):461-476. Corden, W.M. 1972. "Economies of Scale and Customs Unions Theory." Journal of Political Economy 80(3, Part 1):465-75. Dee P. and J. Gali. 2003. The Trade and I nvestment Effects of Preferential Trading Agreements NBER Working Paper No. 10160. 81

PAGE 82

Developm ent Policy Research Unit of the Univers ity of Cape Town. 2001. What are the potential benefits and pitfalls of a free trade area in the Southern African region? DPRU Policy Brief No. 01/P8. Cape Town, South Africa Dollar, D. 1992. "Outward-Oriented Developing Economies Really Do Grow More Rapidly: Evidence form 95 LDCs, 1976-1985." Economic Development and Cultural Change 40(3):523-544. Edwards, S. 1993. "Openness, Trade Liberalization, and Growth in Developing Countries." Journal of Economic Literature 31(3):1358-1393. Web accessed: http://www.jstor.org/stable/2728244. June, 26th 2009. El-Agraa, A.M. 1997. Economic integration worl dwide. New York, St. Martin's Press. Europe Website. 2001 Development and Relations with African, Caribbean and Pacific States ; (Accessed in July, 2009) http://ec.europa.eu/development/g eographical/cot onou/cotonou20001_en.cfm Frankel, J.A and D. Romer. 1999. "Does Trade Cause Growth?" The American Economic Review 89(3):379-399. Frankel, J. A. 1997. Regional Trading Blocs, Wa shington, D. C.: Institute for International Economics. Gordon, M.G. 2007. Impact of the CARICOM Economic Integration Arrangements on the Economic Gains of Selected CARICOM Countries. PhD dissertation Report Greenway, D. and C. Milner. 2003. A Grim REPA? Research Paper No.2003/30, Leverhulme Centre on Globalization and Economic Policy, University of Nottingham. Grunbaum, Max. 2007. Evaluation of Trade and Tari ff Revenue Effects for the Organization of Eastern Caribbean States under Alternative Trade Liberalization Scenarios. PhD dissertation Report. Harrison, A. 1996. "Openness and Growth: A Time s-Series, Cross-Country Analysis for Developing Countries." Journal of Development Economics 48:419-447. Head, K. 2003. Gravity for Beginners. Faculty of Commerce, University of British Columbia. http://www.econ.ubc.ca/nmalhotra/4 90/Articles/KHead20on%20gravity.pdf July 6th, 2008. Johnson, H.G. 1960. "The Economic Theory of Customs Union." Pakistan Economic Journal X(1):14-32. Keck, A. and R. Piermartini. 2005. The Economic Impact of EPAs in SADC Countries. WTO, Economic Research and Statistics Divi sion Working Paper, Geneva, Switzerland 82

PAGE 83

Kerkala, L., J. Niem i, and R. Vaittinen. 2000. R enegotiating the Lom Convention-Trade Policy Schemes and Their Effects for African Re gions. Helsinski School of Economics and Business Administration. Khattry, B. and J. M. Rao. 2002. "Fiscal Faux pas?: An Analysis of the Revenue Implications of Trade Liberalization." World Development 30(8):1431-1444. Krueger, A.O. 1978. Foreign trade regimes and economic de velopment: Liberalization attempts and consequences. Cambridge, MA: NBER. Lal, D. and S. Rajapatirana. 1987. "Foreign Trad e Regimes and Economic Growth in Developing Countries," World Bank Res. Obse rver, July 1987, 2(2), pp. 189-217. Little, I., T. Scitovsky, and M. Scott. 1970. Indus try and trade in some developing countries. London & NY. Oxford University Press for OECD. Lipsey, R.G. 1960. "The Theory of Customs Un ions: A General Survey. In R. Pomfret ed. Economic Analysis of Regional Trading Arrangements. Massachusetts: Edward Elgar Publishing Ltd. Malawi Government. 2008. Annual Economic Report 2008 : Government Printer, Lilongwe, Malawi Meade, J.E. 1955. The Theory of Customs Unions. Amsterdam: North-Holland. Michaely, M., D. Papageorgiou, and A. Choksi. 1991. Liberalizing foreign trade: Lessons of experience in the developing world. Vol. 7. Oxford and Cambridge, MA: Basil Blackwell. Mundell, R.A. 1964. "Tariff Preferences and the Terms of Trade." Manchester School of Economic and Social Studies XXXII(1): 1-13. Nilson, L. 2002. "Trading Relations: Is the Roadmap form Lome to Cotonou Correct?" AppliedEconomics 34(4):439-453. Partnership Agreement Between the Members of the African, Caribbean and Pacific Group of States of the one Part, and the European Co mmunity and its Member States, of the Other Part, Signed in Cotonou, Benin on 23rd June, 2000. http://www.acpsec.org/en/conventions/cotonou/accord1.htm Piermartini, R. and R. Teh 2005. Demystifying M odelling Methods for Trade Policy. Discussion Paper no. 10, World Trade Organization, Geneva. Pomfret, R. 2003. Economic Analysis of Regional Trading Arrangements. Massachusetts: Edward Elgar Publishing Ltd. Robson, P. 1980. The Economics of International Integration. London: George Allen & Unwin. 83

PAGE 84

84 Rodriguez, F. and D. Rodrik 2000. Trade Poli cy and Economic Growth: A Skeptic's Guide to Cross-National Evidence. In B. Bernanke and K. Rogoff. eds. Macroeconomics Annual 2000. Cambridge, MA, NBER. pp.261-325. Sachs, J.D., A. Warner, A. Aslund and S. Fisc her. 1995. "Economic Reform and the Process of Global Integration." Brookings Papers on Economic Activity 1995(1):1-118. Sawyer, W.C. and R.L. Sprinkl e 1989. "Alternative Empirical Es timates of Trade Creation and Trade Diversion: A Comparison of the Baldwin-Murray and Verdoorn Models." Weltwirtschaftliches Archiv 126:61-73. Soloaga, I. and L.A. Winters. 2001. Regionalism in the Nineties: What Effect on Trade? North American Journal of Economics and Finance 12, 1: 1-29. Streeten, P. "Review of Little, et al.," Econ. J., Mar. 1971, 81(1), pp. 144-48. Tinbergen, J. 1962. Shaping the World Economy: Suggestions for an International Economic Policy. New York: The Twentieth Century Fund. Verdoorn, P.J. 1960. The Intra-Block Trade of Benelux. In E. A. G. Robinson ed. Economic Consequences of the Size of Nations London: Macmillan & Co. Viner, J. 1950. The Customs Union Issue in R. Pomfret ed. Economic Analysis of Regional Trading Arrangements. Massachusetts: Edward Elgar Publishing Ltd. Vollmer, S.; I. Martinez-Zarzoso; F.D. No wak-Lehmann; and N.H. Klann. 2009. EU-ACP Economic Partnership Agreements: Empirical Evidence for Sub-Saharan Africa. World Development Report 2009 Background Paper, University of Gottingen, Gottingen, Germany. Weisbrot, M. and D. Baker. 2002. The Relative Impact of Trade Liberalization on Developing Countries. Center for Economic Policy Res earch Briefing Paper, Washington, D.C. Wolf, S. 2002. From Preferences to Reciprocity General Equilibrium Effects of a Free Trade Area Between The EU and the UEMOA. Th ird Annual Conference on Global Economic Analysis, Bonn, Germany. WTO (2002), Committee on Regional Trade Agreements, Coverage Liberalization Process and Transitional Provisions in Regional Trade Agreements, WTO Document. WT/REG/W/46, June 2004, Internet Posting: (Accessed in May, 2009) http://www.wto.org/english/tr atop_e/region_e/regcom _e.htm

PAGE 85

BIOGR APHICAL SKETCH Innocent Lwafyo Thindwa was born in Karonga, Malawi on 4th August, 1977. He received a Bachelor of Social Science, majori ng in economics, from the University of Malawi, Chancellor College in 2003. In that same year Innocent joined the Malawi Governments Economic Common Service, working as an Econom ist. In March 2007, he was promoted to the position of Senior Economist before a further pr omotion to the position of Principal Economist in June, 2007. In August 2008, he joined the Food and Resources Economics Department of the University of Florida to pursue a Master of Sc ience degree program under the sponsorship of the United States Agency for International Development (USAID). He completed his degree in August 2010. 85