Trade creation and trade diversion effects in the EU-South Africa Free Trade Agreement.

CHAPTER 1

INTRODUCTION

The conclusion of the Uruguay Round Agreement in 1994 and subsequent creation of the

World trade organization led to a proliferation of overlapping preferential trade and/or

integration initiatives in nearly all corners of the globe. A number of countries and South

Africa have been in a variety of Trade Agreements. The main objective of this study is to

investigate the effects of Regional Trade Agreements on bilateral and export trade. The

study focuses on the positive impact on member countries (trade creation) and the

negative impact on non-member countries (trade diversion).

Regional Trade Agreements

Regional trade agreements involve a group of countries deciding to pursue free

trade internally, while maintaining tariffs against the rest of the world. Under a customs

union, the countries involved choose a common external tariff with the rest of the world,

whereas under a free trade area the countries maintain different tariffs on imports from

the rest of the world. The analysis of customs union dates back to Viner (1950), who

introduced the terms “trade creation” versus “trade diversion”. Trade creation refers to a

situation where two countries within the customs union begin to trade with each other,

whereas formerly they produced the good in question for themselves. In international

trade terms it means the countries go from autarky (in this good) to trading with zero

tariffs, and they both gain. Trade diversion, on the other hand, occurs when two countries

begin to trade within the union, but one of these countries had formerly imported the

good from outside the union. The importing country formerly had the same tariffs on all

other countries, but purchased from outside the union because that was lowest. After the

union, the country switches its purchases from the lowest – price to a higher – price

country, in this case there is negative efficiency effect.

The possibility of trade diversion identified by Viner (1950) and any changes in

the terms of trade need to be evaluated empirically before judging actual agreements.

More generally, the issue of major concern is, whether regional trade agreements help or

hinder the movement towards global free trade through multilateral negotiations. The

idea that increasing returns might be a reason for trade between countries was well

recognized by Bertil Ohlin (1933) and also Frank Graham (1923), and has been the

motivation for policy actions. When dynamic effects such as the realization of economies

of scale and increased investment and technology flows are considered, the presumption

is more likely that the partners will benefit from the union, and that outside world may

also gain.

The European Union is an example of a single market project that has caused

excitement both within and outside Europe and has had important consequences for

international trade. Important economic integration is occurring in the African region

with the implementation of COMESA (Common Market for East and Southern Africa)

SACU (South Africa Customs Union), Southern Africa Development Community,

commission for East African Cooperation, Cross-Border initiative and Indian Ocean

Commission. Many African countries belong to several regional groupings. With

multiple groupings regulations can conflict, strategies can differ, and political difficulties

can abound (Sharer, 1999).

Southern African Customs Union (SACU)

In terms of the agreement, each member of the Southern African customs union

imposes a similar form of import control to that imposed by South Africa and each issues

import permits where necessary for the import of goods into its territory. An importer

who is in possession of an import permit for the import of goods into one member state

may not use that import permit for the importation of goods into another member state.

Common Market for East and Southern Africa (COMESA)

COMESA was established in Lusaka on 21 December 1981. The original treaty

called for the gradual reduction and eventual elimination of customs duties and non-tariff

barriers. The history of COMESA began in December 1994 when it was formed to

replace the former Preferential Trade Area (PTA) which had existed from the earlier days

of 1981. COMESA (as defined by its Treaty) was established ‘as an organization of free

independent sovereign states which have agreed to co-operate in developing their natural

and human resources for the good of all their people’ and as such it has a wide-ranging

series of objectives which necessarily include in its priorities the promotion of peace and

security.

However, due to COMESA’s economic history and background its main focus is

on the formation of a large economic and trading unit that is capable of overcoming some

of the barriers that are faced by individual states. COMESA’s current strategy can thus be

summed up in the phrase ‘economic prosperity through regional integration’. With its 20

member states, population of over 374 million and annual import bill of around US$32

billion COMESA forms a major market place for both internal and external trading. Its

area is impressive on the map of the African Continent and its achievements to date have

been significant.

European Union (EU)

European Union was brought into existence by a series of multilateral treaties

between sovereign states. The treaty of Rome, which came into force in January 1985,

established the European Economic Community (EEC) between the original six member

states, Belgium, France, Germany, Italy, Luxembourg and the Netherlands. The treaty of

the European Union, commonly known as the Maastricht Treaty identifies the goal of

Economic and Monetary Union but also covers areas such as visa policy, industry policy,

education, culture etc.

In over fifty years since Jacob Viner’s seminal article on customs unions, a vast

literature has amassed around the potential effects of preferential free trade agreements

(PTAs). Theoretical research in this area has generally suggested that welfare should

improve for members if more trade is created within the union relative to the trade

diverted from outside, with the effect on the rest of the world being ambiguous.

Empirical tests of these hypotheses have generally found positive welfare gains from

PTAs. However, existing studies may be biased because they do not sufficiently account

for the general equilibrium implications of PTA formation and spatial correlation among

bilateral trades across countries. As shown in this paper, the bias caused by these effects

is not trivial. In fact, after correcting for them, the estimates found here suggest that the

consequences of PTAs are likely to be sizeable.

This study focuses on the roles of four PTAs (The EU, COMESA, SACU and

EU-SOUTH AFRICA FTA) on trade patterns among 38 countries for which data is

available. The empirical test is designed to address the key question of Trade Creation

and Trade Diversion effects of the PTAs.

Historical Background

During the crisis years of apartheid, Britain was always more concerned with

safeguarding its interests in South Africa rather than exerting pressure which the

existence of those interests gave it, as a lever to alter the conditions that put the interests

at risk in the first place. After a long battle in the South Africa-Europe negotiations, the

European Union finally agreed that affirmative action criteria should be allowed to apply

in tenders for the supply of computers and other equipment for the South African

parliament: preference would be given to black tenderers or tenderers involved in

subcontracting or partnership agreements with black entrepreneurs. The European Union,

the stronger partner in the two-way negotiations, has in its Common Agricultural Policy

(CAP) the largest social-political-economic programme in the world. (Bozzoli, 1999 p.

195-197)

Agreement appeared to have been reached during February 1999 although full

details were not made public. The European Union had lowered its exclusion list of South

African agricultural exports from 46 to 38 per cent so that over ten years the European

Union was committed to lower tariffs on 62 per cent of South Africa's agricultural

exports. Free trade was to cover 'substantially all trade' - about 95 per cent - without

excluding any sector while, in its turn, South Africa agreed to drop tariffs on 81 per cent

of European Union agricultural imports over 12 years. South Africa would also be able to

export 60,000 tons of canned fruit at 'favorable conditions' to the European Union and

about 32 million liters of wine at 50 per cent rebates at the most favored nation rate.

However, five European countries rejected the deal as too favorable to South Africa.

These countries were Spain, Portugal, France, Italy and Greece, each of which is a major

fruit exporter.

Following this offer, South Africa requested and obtained access to GSP

preferences, and called for a long-term agreement under terms as close as possible to the

Lomé Convention. The European Union rejected this request and offered in its place a free

trade agreement and a qualified accession to Lomé (excluding the trade aspects of the

Convention). The negotiations for the Free Trade Agreement were formally opened in

June 1995 and were still on going at the time the study was completed (June 1998).

Problem Statement

To seek reasons for the treatment only in trade relations contradicts other South

African behavior. If trade was that important, why was South Africa consistently

unenthusiastic about giving favorable trade treatment to European Union, to the extent of

suppressing the opportunity whenever it was possible? Even during the apartheid era,

numerous import restrictions were imposed on European products. It was only in 1985

that the Republic decided to treat The European Union as a most favored nation and

stopped applying article 35 of the WTO to the already established major trading partner

(Bell, 2004, 45-49). The EU gives specific reference to internal support and export

subsidies and improved market access to the main export market could be beneficial for

the South African agricultural sector.

Through the process of gathering information about the Trade creation and trade

diversion effects in the EU-South Africa FTA, specific characteristics and influences

have been identified that help to further explain what comprises the total trade effect of a

free trade agreement. Once identified, the effects of trade creation and trade diversion can

be quantified to determine the total trade effect on South Africa resulting from the free

trade agreement. A successful estimation and quantification of trade creation and trade

diversion effects can provide assistance to future investigations with respect to free trade

agreements of this nature. The purpose of this study is to estimate the trade creation and

trade diversion effects in the EU-South Africa FTA and their participation in other

Regional and Preferential Trade Agreements using the gravity model of bilateral trade

flows.

Justification

The effects of trade creation and trade diversion in the bilateral free trade

agreement are important to study for several reasons. Free trade agreements are very

important for developing countries especially in a situation of an import-based economy.

With free trade, trade flows will be smooth for both trading partners and eventually

improve the welfare and consumption level of the populace at large, depending on the

total effect. For this reason trade creation and trade diversion effects play an important

role in identifying trade patterns.

In the empirical literature of studies that have been conducted in the area of trade

creation and trade diversion in various free trade agreements, there has been much

information and insight added to understand the trade creation and trade diversion

effects of free trade areas. Most of the research that has been published has focused

largely on the area of Customs Union type of preferential trade agreements. Viner (1950)

showed that regional trade agreements could be beneficial or harmful to the participating

countries because the preferential nature of these trade deals generated both trade creation

and trade diversion effects. The empirical work on the subject has proven to be

challenging, in that it could not answer “even the most basic issue regarding preferential

trading agreements: whether trade creation outweighs trade diversion” Clausing (2001,

p.678). This study deals specifically with FTA between two trading partners EU-South

Africa. The purpose for choosing the EU and South Africa is to broaden the

understanding of how countries benefit from free trade even when protectionists are of

the contrary opinion.

Tinbergen (1962) and Pöyhönen (1963) were the first authors applying the gravity

equation to analyze international trade flows. Since then, the gravity model has become a

popular instrument in empirical foreign trade analysis. The model has been successfully

applied to flows of varying types such as migration, foreign direct investment and more

specifically to international trade flows. According to this model, exports from country i

to country j are explained by their economic sizes (GDP or GNP), their populations,

direct geographical distances and a set of dummies incorporating some kind of

institutional characteristics common to specific flows.

This study is important because it offers a detailed analysis of the trade creation

and trade diversion effects in the EU-South Africa FTA using the gravity model of

bilateral trade to estimate trade flows from South Africa to the EU. This analysis is

differentiated on the basis of large country versus small country trade partnership and by

the fact that it will provide estimates of whether trade creation and trade diversion are

lower among trade partners that sign agreements than among those that decline the

option. The implications of this study can be far reaching and can project the impact of

the FTA between South Africa and the EU on the bilateral trade flows between the two.

Study Objective

Free trade agreements have a substantial impact on the participant countries in terms of

welfare, consumption, production and trade flows. This study will put emphasis on the

trade creation and trade diversion effects of the EU-South Africa FTA. The main

objective of this study is to investigate the trade creation and trade diversion effects in the

EUSAFTA. The focus area is on their participation in other regional and preferential

trade Agreements (EU, COMESA, SACU, and EUSAFTA) and the effects of trade

creation and trade diversion on trade volume.

The rest of this work summarizes the existing theoretical and empirical literature

on trade creation and trade diversion effects in the free trade agreements and a discussion

of the methods and procedures used which includes the gravity model used in this study.

The next section presents a discussion of the model, data and variables included in the

model. In conclusion, there will be a discussion of results.

Thesis Organization

This study is segmented into five chapters. Chapter one comprises of the

introduction, problem statement, justification, objectives, and estimation techniques.

Chapter two comprises of theoretical and empirical review of international trade theory to

support the analytical methods used in this study. Chapter three discusses the

methodologies employed. Chapter four discusses the data and variables used for the

analysis, along with the estimated results from the models used for this study. Chapter

five includes the summary and conclusions.

CHAPTER 2

LITERATURE REVIEW

When it comes to estimating and analyzing trade creation and trade diversion

effects in the trade among countries, and between member countries and non-members,

the theoretical literature showed that the formation of free trade areas, customs unions, or

other preferential trading blocs had uncertain effects on economic welfare. Viner (1950)

showed that regional trade agreements could be beneficial or harmful to the participating

countries because the preferential nature of these trade deals stimulates both trade

creation and trade diversion. The empirical work on the subject has proven to be

challenging that it could not answer “even the most basic issue regarding preferential

trading agreements: whether trade creation outweighs trade diversion” Clausing (2001,

p.678).

A Standard Gravity Model

The first formulations of the gravity equation are found in Timbergen (1962),

Pöyhönen (1963) and Pulliainen (1963). Linnemann (1966) and many other authors such

as Aitken (1973) and Leamer (1974) extended its use. According to Deardorff (1984), the

empirical success of the gravity equation is due to the fact that it can explain some real

phenomenon the conventional factor endowment theory of international trade cannot: the

trade between industrialized countries, the intra-industry trade and the lack of dramatic

reallocations of resources when trade liberalization processes have taken place.

Tinbergen (1962) and Pöyhönen (1963) were the first authors applying the gravity

equation to analyze international trade flows. Since then, the gravity model has become

popular instrument in empirical foreign trade analysis. The model has been successfully

applied to flows of varying types such as migration, foreign direct investment and more

specifically to international trade flows. According to this model, exports from country i

to country j are explained by their economic sizes (GDP or GNP), their populations,

direct geographical distances and a set of dummies incorporating some kind of

institutional characteristics common to specific flows.

Theoretical support of the research in this field was originally very poor, but since

the second half of the 1970s several theoretical developments have appeared in support of

the gravity model. Anderson (1979) made the first formal attempt to derive the gravity

equation from a model that assumed product differentiation. More recently Deardorff

(1995) has proven that the gravity equation characterizes many models and can be

justified from standard trade theories. The differences in these theories help to explain the

various specifications and some diversity in the results of the empirical applications.

More recent discussion by Deardoff (1984), Learmer and Levinohn (1995), and

Helpman (1999) show that the gravity model has a relatively long history. It differs from

most other theories (including traditional theory) in that it tries to explain the volume of

trade and does not focus on the composition of that trade. The model uses an equation

framework to predict the volume of trade on a bilateral basis between any two countries.

The particular equation form has some similarity to the law of gravity in physics, which

has resulted in the term gravity model being applied. These foundations were

subsequently developed by, among others, Anderson (1979) and Bergstrand (1985), who

derived gravity models from models of monopolistic competition, and Deardorff (1998),

who demonstrated that the gravity model can be derived within Ricardian and Heckscher-

Ohlin frameworks.

Trade patterns have also been investigated using gravity-type equations. The trade

overlap (i.e. two-way trade within industries) is examined in Bergstrand (1989) and

shares rather than trade volumes, which depart slightly from the bulk of work on gravity

equations. Gravity models have been extensively used to address the issue of the impact

of trade policies on trade flows like the impact of regional trade agreements. Consider

that two countries i and j sign a regional agreement. Introduce one dummy: 1 for «both

in» (i and j in the agreement) and 0 otherwise. If the parameter estimate is positive and

significant there is trade creation due to regionalism. Bilateral trade flows are considered

between these countries, in a symmetric manner.

The gravity model has been used frequently to analyze bilateral trade flows

between countries. The equation used is similar in all studies and has the following

specifications;

Xij = α0 + α1(Yi) + α2(Yj) + α3(Ni) + α4(Nj) + α5(Dij) + α6(Aij) + α7 (Pij) (2.1)

Where Xij irepresents the value of the trade flow from country i to country j; Yi and Yj

are the values nominal GDP in i and j; Ni and Nj are the size of population in both

countries; Dij is the physical distance from the economic center of country i to that of

country j ; Aij represent any other factor affecting trade among i and j either positive or

negative; and Pij is trade preferences among the countries.

The GDP of the exporting country measures productive capacity, while that of the

importing country measure, absorptive capacity. These two variables are expected to be

positively related to trade. Physical distance and country adjacency dummies are proxies

for transportation costs. The most commonly used of the other variables affecting trade,

are dummies for the Preference agreement in which countries participate; total population

of importing and exporting countries as well as their per capita income levels. Population

is used as measure of country size, and since larger countries have more diversified

production and tend to be more self sufficient, it is usually expected to be negatively

related to trade. As noted by Bergstrand (1989), there is an inconsistency in this

argument, as larger populations allow for economies of scale, which are translated into

higher exports.

Linnermann (1966), Aitken (1973) and Sapir (1981) used the same general

specification, but also included exporter and importer populations. Microeconomic

foundations of this alternative specification are discussed in Bergstrand (1984).

Bergstrand (1984) addressed the argument by critics that this approach is “loose” and

does not explain the multiplicative functional form and other issues in developing further

the microeconomic foundations of the gravity equation.

In a study using 1988 data just before the Canada - U.S. FTA was signed,

McCallum (1995) estimated a gravity model where the dependent variable was exports

from each Canadian province to other provinces and to U. S. states. Exports depend on

the province or state GDP’s, hence the estimated regression is

ln Χ јі = α + β1 ln Υi + β2 ln Υj + Χ δij + ρ ln dij (2.2)

Where δij irepresents an indicator variable that equals unity for trade between two

Canadian provinces and zero otherwise and dij is the distance between any two provinces

or states. The results showed negative relationship between distance and trade. The

results also show that cross provincial trade was some 22 times larger than cross-border

trade in 1988 and 15.7 times larger in 1993.

Apart from international trade flows, gravity models have achieved empirical

success in explaining various types of inter-regional and international flows, including

capital flows and labor migration (Vandekamp 1977). Evenett and Keller (1998) along

with Deardoff (1988) evaluated the usefulness of gravity models in testing alternative

theoretical models of trade. Apart from the dummy variables, other exogenous regressors

used are dummies for wars, conflicts, natural disasters, etc. Krueger (1999) also includes

a dummy for remoteness to take into account the fact that some countries are further

away from most of their trading partners than other countries.

According to the theorem (Krugman, 1980) with two countries trading, the larger

market will produce a greater number of products and be a net exporter of the

differentiated good. Helpman (1987), Wei, (1996), Soloaga and Winters (1999), Limao

and Venables (1999) and Bougheas et al, (1999) among others, contributed to the

refinement of the explanatory variables considered in the analysis and to the addition of

new variables.

According to the generalized gravity model of trade, the volume of exports

between a pair of countries, Xij, is a function of their incomes (GDPs), their populations,

their geographical distance and a set of dummies. Numerous empirical studies showed

that trade flows follow the physical principles of gravity: two opposite forces determine

the volume of bilateral trade between countries - the level of their economic activity and

income, and the extent of impediments to trade. National borders are among these

impediments, even for industrialized countries (MaCallum, 1995).

CHAPTER 3

METHODOLOGY

Economic Theory

Theoretical studies regarding the microeconomic foundations of the

gravity equation (Anderson (1979), Bergstrand (1985 and 1989) and Helpman and

Krugman (1985, ch. 8) provide rigorous explanations for the log linear form. Mátyás

(1997) and (1998), Chen and Wall (1999), Breuss and Egger (1999), and Egger (2000)

improved the econometric specification of the gravity equation. Second, Berstrand

(1985), Helpman (1987), Wei, (1996), Soloaga and Winters (1999), Limao and Venables

(1999) and Bougheas et al, (1999) among others, contributed to the refinement of the

explanatory variables considered in the analysis and to the addition of new variables.

According to the generalized gravity model of trade, the volume of exports

between pairs of countries, Xij, is a function of their incomes (GDPs), their populations,

their geographical distance and a set of dummies accounting for Regional and

Preferential trade Agreement membership. More recently, Deardoff (1984), Learmer and

Levinohn (1995), and Helpman (1999) show that the gravity model has a relatively long

history. It differs from most other theories (including traditional theory) in that it

explains the volume of trade and does not focus on the composition of that trade. The

model uses an equation framework to predict the volume of trade on a bilateral basis

between any two countries.

Other variables are often introduced, such as population size in the exporting and

or importing country (as related to market size or economies of scale) or a variable to

reflect an economic integration arrangement (such as a free trade area) between the two

countries. These foundations were subsequently developed by, Anderson (1979) and

Bergstrand (1985) among others, who derived gravity models from models of

monopolistic competition, and Deardorff (1998), who demonstrated that the gravity

model could be derived within the Ricardian and Heckscher-Ohlin frameworks.

Traditionally, the gravity model uses distance to model transportation costs.

However, Bougheas et al., (1999) showed that transport costs are a function not only of

distance but also of public infrastructure. They augmented the gravity model by

introducing additional infrastructure variables (stock of public capital and length of

motorway network). Their model predicts a positive relationship between the levels of

infrastructure and the volume of trade, which is supported using data from European

countries. They took a further step in this direction by introducing a new infrastructure

index (taking information on roads, paved roads, railroads and telephones) and

differentiating between exporter and importer infrastructure as explanatory variables of

bilateral trade flows.

According to Sanso et al., (1989, 155-166) the basic formulation of the gravity

equation is as follows:

Mij = A+Yiβ1+Υіβ2+Liβ3+Ljβ4+Dijβ5 (3.1)

Where Mij represents the current value of exports from country i to country j, A

represents constant, Y represents the current value of income, L represents population,

and Dij represents distance between countries i and j.

One of the characteristics of the equation is its general validity, since it is equally

applicable to any pair of countries. It is also symmetrical because it provides the trade

flows in both directions by changing country i variables for country j ones. Other dummy

variables indicating membership to an economic area or neighborhood, indicators of

protection levels, or any relevant variables can be added to the equation. Sanso et al.,

went further to consider the following model to be estimated:

Mij = Ft +Yit+Yjt+Yjt+ Dij,+EECijt,+EFTAijt+ NEARij (3.2)

Where Mijt represents the current value of sales from country i to country j in period t,Yit

represents the current value of country i per capita income in period t,Yjt represents the

current value of country j per capita income in period t, Yjt represents current value of

country j income in period t, Dij represents distance from country i to country j, EECijt

represents the dummy variable that shows Whether both countries i and j are integrated

into the EEC in period t, EFTAijt represents the dummy variable that shows whether both

countries i and j are integrated into the EFTA in period t, and NEARij represents the

dummy variable that shows whether both countries i and j have a common frontier.

Using total export and total import, annual data from 1964 to 1987 between each

pair of countries, GDP and population of every country, distance between countries, and

dummies of membership in the EEC and EFTA are used to estimate the gravity model in

log-linear format (Sanso et al., 1989). There are three reasons, which justify the addition

of these variables to the basic formulation. First, they usually appear in models that use

the equation with developed countries, as is our case. Second, they are perfectly

compatible with the spirit inspiring the gravity equation. Finally, the inclusion of EEC

and EFTA enables us to assess the evolution through time of the evolution of both trade

agreements.

There are a large number of empirical applications in the literature of international

trade, which have contributed to the performance of the gravity equation. Some of them

are closely related to this study. For nearly thirty years the gravity equation has been

frequently and successfully used to aid in the understanding of bilateral trade flows

across countries as well as to analyze commercial policy measures. The formulation is a

log linear function upon a set of well-defined variables. The explanatory variables

include the incomes and populations of both countries and the distance between them.

Almost all of the empirical works use the log linear form and include these variables, but

they add others according to their particular circumstances.

Theoretical Linkages to the Gravity Model

Economic theory gives several indications as to the factors that affect trade. These

factors include income, transaction costs and trade agreements. Higher income countries

trade more; transaction costs and trade agreements are determinants of export potentials

in the gravity model. Thus various combinations of microeconomic variables, such as

income and geographic distance, are powerful predictors of trade potentials. Hence,

gravity equations have been used extensively in the modeling of international trade flow.

It is common to augment the basic gravity model through additional bilateral variables.

For instance variables are added to account for common language, common border,

common colonial history, and common currency. The impact of income and transaction

costs on trade can also be explained in the partial equilibrium model.

Regional trading agreements are generally perceived to be potentially beneficial

in a trade sense. In the short-run, member countries benefit, as long-run trade diversion

does not outweigh immediate trade creation effects. Long-run gains occur through the

channels of increased efficiency through specialization, economies of scale, increased

trade, and investments.

Impact of Income on Trade (Exporting Country)

Figure 3.1 represents the effects of changes in the income of the exporting country

in the partial equilibrium model. The free – trade equilibrium is at E, the intersection of

the exporting country’s excess supply and the importing country’s excess demand. An

increase in the income of the exporting country shifts the domestic demand curve

outward from D to D’, indicating an increase in demand and shifting the excess supply in

the rest of the world upwards from ES to ES’x indicating a decrease in excess supply.

A decrease in the income of the exporting country shifts the domestic demand

curve from D to the left D’’ and the excess supply curve shifts downward from ES to

ES’’x indicating a decrease in excess supply. Greater income in the exporting country

means a greater capacity to consume and hence decreases its supply exports to the

importing country. Conversely, lesser income in the exporting country means a lesser

capacity to consume and hence increase its supply of exports to the importing country.

The implications of greater income on domestic price are that, will increase from

Pd to Pd’’. Conversely, the implication of lesser income on domestic price is that there

will be an increase in domestic price from Pd to Pd’’.

The implications of greater income on quantity demanded domestically are that, it will

increase from Qd to Qd’. Conversely, lesser income will lead to a decrease in quantity

demanded domestically from Qd to Qd’’.

Impact of Income on Trade (Importing Country)

Figure 3.2 represents the effects of changes in the income of the importing

country in the partial equilibrium model. The free – trade equilibrium is at E, as discussed

in the previous section. An increase in the income of the importing country shifts the

domestic demand outward from D to D’’. This causes the rest of the world excess

demand curve to shift outward from ED to ED’m indicating an increase in demand for

imports.

A decrease in the income of the importing country causes the excess demand

curve to shift left from ED to ED’’m indicating a decrease in demand for imports. Greater

income in the importing country indicates greater capacity to demand imports from the

exporting country, while decreased income has the opposite effect.

The implication of greater income on price is that the domestic market price will

rise from Pd to Pd’. Conversely, lesser income will lead to a fall in domestic market

price from Pd to Pd’’. The implication of greater income in the exporting country on

quantity demanded in the domestic market is that, there will be an increase from Qx to

Qx’. Conversely, lesser income in the exporting country will lead to a decrease in

quantity demanded in the domestic market from Qx to Qx’’.

The implication of greater income in the importing country is that there will be an

increase in domestic price from Pm to Pm’. Conversely, lesser income in the importing

country will lead to a decrease in domestic price from Pm to Pm’’. The implication of

greater income in the importing country on domestic quantity demanded is an increase

from Qm to Qm’ indicating an increase in quantity demanded. Lesser income will lead to

a decrease in quantity demanded domestically from Qm to Qm’’.

Transaction Costs

Figure 3.3 represents the effects of changes in transaction costs on trade.

Transaction costs include factors related to transportation, handling costs, common

language and colonial ties. Greater distances between partner countries lead to increased

transportation costs, which lead to an upward shift of the excess supply curve. These

changes in transaction costs are seen in Figure 3.3 by a shift in ES. Conversely, lesser

distances between partner countries lead to a downward shift in the excess supply. The

consequence of greater distances (greater transportation costs) is to reduce the volume of

trade, while lesser distances will enhance trade.

A decrease in handling costs and potential ease of transportation as a result of

common borders between partner countries will lead to a downward shift of the excess

supply curve and an increase in trade between both countries. Another factor that can

reduce handling costs is common language between trading countries. If trading partners

speak a common language, there will be a shift of the excess supply to the right, resulting

in increased trade.

Colonial ties can be positive or negative depending on the countries involved.

Colonial ties between Canada and Great Britain as well as the Caribbean and other

European countries are typically viewed as positive. Contrary to this, the colonial ties

between Great Britain and countries such as South Africa may result in a less positive

relationship. Positive colonial ties will lead to increased trade between countries, whereas

negative colonial ties will lead to decreased trade between countries with such ties.

Effects of Economic Integration: Trade Creation and Diversion Effects

Economic integration implies preferential treatment for member countries as

opposed to nonmember countries. Since this type of arrangement can lead to shifts in the

pattern of trade between members and nonmembers, the effects must be judged on the

basis of each individual country.

Rest of the world

Price Price ES'x

D S ES

Pm ESx''

Pd' D''

Pw Pw E

Pd''

Pd Pd

EDm

Qx'' Qx Qx' Quantity Qw Quantity

Exporting Country

Figure 3.1 Impact of Income on Trade (exporting country)

Rest of the world Importing country

Price Price

D S

ESx

D''

Pm Pm

Pw'

Pw E Pw

Pd'

ED'm

EDm

ED''m

Qd Qw Quantity Qm Quantity

Figure 3.2 Impact of Income on Trade (importing country)

Exporting country Rest of the world

Price Price ES'x

D S ES

Pm ESx''

Pw Pw E

Pd Pd

EDm

Qx Quantity Qw Quantity

Figure 3.3 Impact of Transaction costs on Trade

While integration represents a movement to free trade on the part of member

countries, at the same time it can lead to the diversion of trade from lower-cost

nonmember source (which still faces the external tariffs of the group) to a higher - cost

member country source (which no longer faces any tariffs). These two effects of

economic integration are called trade creation and trade diversion. These terms were

initially used by Jacob Viner (1950), who defined trade creation as taking place whenever

economic integration leads to a shift in product origin from a domestic producer whose

resource costs are higher to a member producer whose resource costs are lower. This shift

represents a movement in the direction of the free- trade allocation of resources and thus

is presumably beneficial for welfare. Trade diversion takes place whenever there is a shift

in product origin from a nonmember producer whose resources costs are lower to a

member country producer whose resources costs are higher. This shift represents a

movement away from free-trade allocation of resources and could reduce welfare.

Figure 3.4 and 3.5 represent trade creation and trade diversion effects of a

Customs Union. Before the economic integration, the price of the good in country A is

PA ( PB plus the tariff). With integration between A and B, the tariff is removed, and A

now imports (Q4 – Q1) rather than (Q3 – Q2) from B. Q2 - Q1 of the increased imports

displace previous home production, and Q4 – Q3 reflect the greater consumption at the

new price PB facing country A’s consumers. The trade effect is the sum of areas b and d.

In general trade creation means that a free trade area creates trade that would not have

existed otherwise. As a result, supply occurs from a more efficient producer of the

product.

Before the union with country B, country A has a tariff on imports of the good.

Thus country C’s tariff – inclusive price in A’s market is PA = PC (1 + t). Before the

union, A imports (Q3 – Q2) from C. When the union is formed with B, country A imports

(Q4 – Q1), all coming from partner B, which no longer faces a tariff. In general trade

diversion means that a free trade area diverts trade that existed otherwise. The net trade

effects for A is the difference between areas b + d (a positive effect due to the lower price

in A) and area e (a negative effect due to lost tariff revenue by A that is not captured by

A’s consumers). Producers in the importing country suffer losses as a result of the free

trade area. The decrease in the price of their product on the domestic market reduces

producer surplus in the industry. The price decrease also induces a decrease in output of

existing firms, a decrease in employment, and a decrease in profit.

In addition to the trade effects of economic integration, it is likely that the

economic structure and performance of participating countries may evolve differently

than if they had not integrated economically. Reducing trade barriers brings about a more

competitive environment and possibly reduces the degree of monopoly power that was

present prior to integration. In addition, access to larger union markets may allow

economies of scale to be realized in certain export goods. These economies of scale may

result internally to the exporting firm in a participating country as it becomes larger, or

they may result from a lowering of costs of inputs due to economic changes external to

the firm. In either case they are triggered by market expansion brought about by

membership in the union. The realization of economies of scale may also involve

specialization on particular types of a good, and thus, trade may increasingly become

intra-industry trade rather than inter-industry trade.

It is also possible that integration will stimulate greater investment in the member

country from both internal and foreign sources. Investment can result from structural

changes, internal and external economies and geographic markets now open to producers.

Furthermore, foreigners may wish to invest in productive capacity in a member country

in order to avoid being choked out of the union by trade restrictions and a high common

external tariff. Economic integration at the level of the common market may lead to

dynamic benefits from increased factor mobility. If both capital and labor have the

increased ability to move from areas of surplus to areas of scarcity, increased economic

efficiency and correspondingly higher factor incomes in the integrated area will emerge.

The trade creation effect is caused by the extra output produced by the member

countries. This extra output is generated due to the freeing up of trade between them.

Increased specialization and economies of scale should increase productive efficiency

within member countries. The trade diversion effect exists because countries within

trading blocs, protected by trade barriers, will now find they can produce goods more

cheaply than countries outside the trade bloc. Production will be diverted away from

those countries outside the trade bloc that have a natural comparative advantage to those

within the trading bloc.

Preferential trade arrangements are often supported because they represent a

movement in the direction of free trade. If free trade is economically the most efficient

policy, it would seem to follow that any movement towards free trade should be

beneficial in terms of economic efficiency. Whether a preferential trade arrangement

raises a country's welfare and raises economic efficiency depends on the extent to which

the arrangement causes trade diversion versus trade creation. The theoretical literature

showed that the formation of free trade areas, customs unions, or other preferential

trading blocs had uncertain effects on economic welfare. Viner (1950) showed that

regional trade agreements could be beneficial or harmful to the participating countries.

Price DA

PA = PB (1 + t)

a c

Q1 Q2 Q3 Q4 Quantity

Figure 3.4 Trade Creation effects of a customs union

Price DA

b d PA = PC (1+ t)

a c

Q1 Q2 Q3 Q4 Quantity

Figure 3.5 Trade Diversion effect of a customs union

A Standard Gravity Model

Gravity models were first applied to international trade by Tinbergen (1962) and

Pöyhönen (1963), who proposed that the volume of trade could be estimated as an

increasing function of the national incomes of the trading partners, and a decreasing

function of the distance between them. Although the gravity model became popular

because of its perceived empirical success, it was also criticized because it lacked

theoretical foundation.

The basic assumption of the gravity is that holding every other variable constant,

particular countries tend to have rich trading partners. Distance has an influence on trade

flows. Trade becomes cheaper when trading countries are nearby each other. An increase

in distance decreases trade flows, the more the transportation costs between trading

partners. There are also economic and political integrations like the EU, COMESA,

SACU, and EUSAFTA etc that create trade preference in selected countries. Dummy

variables are usually added to capture participation in Regional and Preferential Trade

Agreements. Whalley (1998) noted that the benefits from this form of integration might

be quite large, particularly in small country cases.

As a result of these factors and various Regional and Preferential Trade

Agreements between countries and their trading partners, the following specification of

the gravity model is considered in this study:

Yi = A + Xi1c1 + Xi2c2 + Di1c3 + Di2c4 + Di3c5 + Di4c6 (3.3)

Where Yi represents trade flows (exports or imports) between country 1 and country “i”,

Xi1 represents the GDP of country “i”, and Xi2 represents the Distance between country 1

and country “i”.

Dummy variable (Dij) indicate to which Regional and Preferential Trade

Agreement a particular country belongs: Di1 represents 1 – Membership in the EU (EU-

25), Di1 represents 0 – Otherwise, Di2 represents 1 – Member of COMESA (Common

market for the East and Southern Africa), Di2 represents 0 – Otherwise, Di3 represents 1 –

Member of SACU (South Africa Customs Union), Di3 represents 0 – Otherwise, Di4

represents 1 – Member of EUSA (EU – 25 and South Africa), and Xi1 represents 0 –

Otherwise.

Based on the gravity equations put forth by Whalley (1998), Sanso et al., (1989)

above specification of the gravity model (1), the following bilateral trade equations are

estimated. Four Regional preferential trade agreements as well as one overall Trade

Creation and Trade Diversion dummies are explained in both equations.

The (Bilateral Trade Model) is specified as follows:

LOG (Xij) = a0 + a1LOG (GDPi) + a2LOG (DISTij) + a3 (EUcij)

+ a4 (EUdij) + a5(COMEScij) + a6(COMESdij)

+ a7(SACUcij) + a8(SACUdij) + a9(EUSAcij)

+ a10 (EUSAdij) + a11 (PTAcij) + a12(PTAdij)

+ a13Σ (Sij)

(3.5)

Where PTAcij represents preference dummy (trade creation), PTAdij represents

Preference dummy (trade diversion), PTAcij represents 1 – If both countries belong to

any Preferential Trade Agreement, PTAcij represents 0 – otherwise, PTAdij represents 1

– If one of them belong to any Preferential Trade Agreement, PTAdij represents 0 –

Otherwise, Σ (Sij) represents other variables that affects trade like History, language,

land locked, borders etc., Σ (Sij) represents 1 – Colonial ties, Common language,

Landlocked, Common borders; and Σ (Sij) = 0 – Otherwise.

The (Export Model) is specified as follows

LOG (Xi) = b0 + b1 LOG (GDPj) + b2 LOG (DISTij) + b3 (EUcij)

+ b4 (EUdij) + b5 (COMEScij) + b6(COMESdij)

+ b7 (SACUcij) + b8 (SACUdij) + b9(EUSAcij)

+ b10 (EUSAdij) + b11 (PTAcij) + b12(PTAdij)

+ b13Σ(Sij)

(3.6)

Where, PTAcij represents preference dummy (trade creation), PTAdij represents

Preference dummy (trade diversion, PTAcij represents 1 – If both countries belong to any

Preferential Trade Agreement, PTAcij represents 0 – otherwise PTAdij represents 1 – If

one of them belong to any Preferential Trade Agreement, PTAdij represents 0 –

Otherwise, Σ (Sij) represents other variables that affects trade like History, language, land

locked, borders etc., Σ (Sij) represents 1 – Colonial ties, Common language, Landlocked,

Common borders; and Σ (Sij) represents 0 – Otherwise

The measure of the geographical distance between countries is defined as the

distance between capital cities. For neighboring countries this distance is defined as the

distance between their capital city and geographical center. The relationship between

trade flows and the various explanatory variables will be estimated by ordinary least -

squares (OLS) regression methods. The variables are measured in the following units:

Trade flows (Xij) measured in millions of dollars; GDP measured in millions of dollars;

Distance equals Thousands of miles; History equals 1 if they have colonial ties and 0

otherwise, Language represents 1 if they speak a common language and 0 otherwise,

Landlocked represents 1 If landlocked and 0 otherwise, Borders represents 1 If they share

a common border and 0 otherwise.

Preference dummy PTA, PTAcij equals 1 If both countries are members of any

and 0 otherwise, PTAdij represents 1 If one of them is a member and 0 otherwise, EUcij

represents 1 If both countries are members and 0 otherwise, EUdij represents 1 if one of

them is a member and 0 otherwise, COMESAcij represents 1 If both countries are

members and 0 otherwise, COMESAdij equals 1 If one of them is a member and 0

otherwise, SACUcij represents 1 If both countries are members and 0 otherwise,

SACUdij represents1 If one of them is a member and 0 otherwise,

EUSAcij represents 1 If both countries are members and 0 otherwise, and

EUSAdij represents 1 if one of them is a member and 0 otherwise

The basic gravity model has been expanded to include other variables that can

explain the trade creation and trade diversion effects of Regional Preferential trade

agreement. The resulting models shown in equations 3.5 and 3.6 represent an expanded

version of the basic gravity model. These models will be empirically estimated in log-

linear format in the next chapter.

CHAPTER 4

EMPIRICAL ANALYSIS

Estimation Techniques

Following the theoretical literature on the use of the gravity trade model, the

model used in this study will be the log-log form of the gravity equation, using standard

OLS regression analysis. The current gravity model is an adaptation of the model used by

some other variables like physical distance (measured as the great circle distance between

capital cities), common borders, and language. In order to capture the trade effects of

various PTA’s some interesting variables will be introduced notably dummies for each

PTA. These dummies are proxies for the two main trade effects – trade creation and

diversion. The welfare effects associated with trade creation and trade diversion are not

captured by these dummies, as noted by Viner (1950), the reason being that the

dependent variable exports from country i to country j measures the bilateral export flows

as against welfare. These dummies capture changes in volumes of trade among PTA

members as well as between them and non-members. The efficiency gains or losses

associated with changes in export volumes are the major factor that links changes in

export volumes with welfare.

The formulation is a log linear function upon a set of well-defined variables. The

explanatory variables are the incomes and populations of both countries and the distance

between them. Almost all the empirical works use the log linear form and these variables,

but they add others according to their particular circumstances.

The Data

The data is of 1998 trade data from the IMF direction of trade database.

The model is estimated for a cross-section dataset of 39 countries comprised of EU-25

(Austria, Belgium-Luxembourg, Bulgaria, Czechs Republic, Denmark, Estonia, Finland,

France, Germany, Sweden, Hungary, Ireland, Italy, Netherlands, United Kingdom, Poland,

Greece, Portugal, Spain, Cyprus, Latvia, Malta, Slovakia, Slovenia, and Lithuania),

COMESA member countries that hold 75% of trade share (South Africa, Egypt,

Kenya, Malawi, Mauritius, Zambia, Zimbabwe), SACU member countries (South

Africa, Botswana, Lesotho, Namibia and Swaziland), and the United States.

Information on four Regional and Preferential Trade Agreements used in the

study was from the WTO database. The data set consists of bilateral trade and

exports respectively.

The current GDP is expressed in purchasing power parity (PPP) values. PPP

values are in theory preferable as noted by Sirnivassan (1995); however PPP values are

subject to significant measurement errors. Yet, the risk of significant alterations of the

regression estimated is small, as shown by Linnemann (1966) and Frankel (1997). The

physical distances between countries were calculated using Indo (2005), web based

calculations of country distance based on the computer program. The information on

history, languages, landlocked and common border is based on Central Intelligence

Agency World Fact book.

The Variables

A total bilateral trade flow for country pairs and exports in log form is the

dependent variable for this study. Table 4.1 lists the variables considered in the gravity

model analysis. The tables include both dependent and independent variables. The

dependent variables are bilateral trade flows and export respectively. The independent

variables are bilateral exports, exports from one of the country members, GDP of the

importing country and distance from the economic centers of country pairs.

Table4.1. Gravity model variables

Symbol Variable Expected Sign

LNXij The logarithm of bilateral trade flows

c1LN GDPi The logarithm of GDP for country i (+)

c2LNGDPj The logarithm of GDP for country j (+)

c3 LNdistij The logarithm of distance between countries (-)

c4EUcij EU membership trade creating dummy (+)

c5EUdij EU membership trade diverting dummy (+/-)

c6COMEScij COMESA membership trade creating dummy (+)

c7COMESdij COMESA membership trade diverting dummy (+/-)

c8SACUcij SACU membership trade creating dummy (+)

c9SACUdij SACU membership trade diverting dummy (+/-)

c10EUSAcij EUSA membership trade creating dummy (+)

c11EUSAdij EUSA membership trade diverting dummy (+/-)

c12PTAcij A trade creating Preference dummy (+)

c13PTAdij A trade diverting Preference dummy (+/-)

c14HISTORY Colonial ties dummy (-)

c15LANDLOCKED Landlocked dummy (-)

c16LANGUAGE Common language dummy (+)

c17BORDERS Adjacency dummy (+)

eij Normal distribution error term

Table 4.2.Definition of Variables and Data Sources

Variable Source

Trade Flows IMF direction of trade (1998)

GDP World Bank Development Indicator (2004A)

Distance Programme developed by John A. Byers

History Central Intelligence Agency World Factbook

Languages Central Intelligence Agency World Factbook

Landlocked Central Intelligence Agency World Factbook

Borders Central Intelligence Agency World Factbook

PTAs World Trade Organization database

EU World Trade Organization database

COMESA World Trade Organization database

SACU World Trade Organization database

Results

The main objective of this study is to estimate the trade creation and trade

diversion effects in the EU-South Africa bilateral trade agreement using the gravity

model. This section of the study examines the estimated gravity model. In particular it

examines whether the factors indicated in the gravity equation make a significant

contribution to an explanation of the bilateral and export trade. The applied regression

method used in determining the significance of variables within the model was the

standard OLS using the SAS.

Countries have developed more active foreign trade relations with other countries

where total GDP is higher. Distance negatively influences trade flows. Nearby country

partners have developed more active foreign trade relation with each other. Participation

in the EU, COMESA, SACU, EUSA etc., and influences trade flows and leads to trade

creation on one hand, and stimulates trade with non-members on the other hand, resulting

in minimal trade diversion from non-member countries.

The estimated regression equations are as follows:

LOG (Bilateral trade) = -7.7551E-12 + 0.7647*LOG (GDP)

– 1.09*LOG (DIST) + 3.0*(EUcij) + 0.93*(EUdij)

+ 3.297*(COMEScij) + 2.462*(COMESdij) – 2.558*(SACUcij)

– 0.061*(SACUdij) + 2.31*(EUSAcij + 1.38*(EUSAdij)

+ 3.85*(PTAcij) + 9.187*(PTAdij) - 0.328*(HISTORY)

+0.279*(LANDLOCKED) + 0.0131*(LANGUAGE)

+ 0.684*(BORDERS)

(4.1)

Where the coefficients represent α

LOG (EXPORT) = 1.39611E-12 + 0.696*LOG (GDP)

- 1.131*LOG (DIST) + 2.235*(EUcij) + 0.980*(EUdij)

+2.551*(COMEScij) +1.723*(COMESdij) - 2.276*(SACUcij)

– 0.286*(SACUdij) + 2.199*(EUSAcij) + 1.009*(EUSAdij)

+ 4.746*(PTAcij) + 9.7995*(PTAdij) – 0.447*(HISTORY)

- .362*(LANDLOCKED) + 0.017*(LANGUAGE)

+ 0.4694*(BORDERS)

Where the coefficients represent β

(4.2)

Interpretation of Model Coefficients

Two models were estimated, one for bilateral trade and the other for exports (See

tables 4.3 and 4.4). Dummy trap problem was encountered in this study. However, an

interaction country with respect to the United States was introduced to the model as non-

members so as to nullify the effect. The estimated coefficients on GDP, distance and the

dummy variables have the expected signs and are significantly different from zero in both

regressions. The positive and significant coefficients of GDP indicate that richer

countries usually trade more compared to poor ones.

All coefficients of variables that make up the bilateral trade had the expected

signs and were significant at the 10 per cent level except the (SACUdij), which were

insignificant but had the expected signs. The negative sign of the coefficient also,

indicate that participation in the integration schemes does not stimulate enough mutual

trade with member countries. This sign suggests that this economic integration

arrangement is not sufficiently strong to influence trade with non-member countries

positively and significantly.

Bilateral Trade Model

Table 4.3 shows the results of the estimations based on bilateral trade between

countries. The GDP coefficient was positive as expected and statistically significant at

the 1 per cent level as expected. This is as a result of the positive relationship between

trade and the income of trading countries. The coefficient of distance variable (dist) was

negative and significant at the 1 per cent level indicating that transportation costs act as a

constraint to trade. A country faces higher trading costs if the port is not located in the

economic center. Distance is negatively related to trade.

The coefficient of the European Union Trade Creation dummy variable (EUCij)

was positive and statistically significant at the 5 per cent level, indicating the trade

creating effects of participating in a regional trade arrangement. This suggests that

economic integration scheme is sufficiently deep and strong to influence the mutual trade

between member countries positively and significantly. The coefficient of the European

Union Trade Diversion dummy variable (EUDij) was positive and statistically significant

at the 10 per cent level, suggesting that the trade diverting effects of the EU is minimal

compared to the trade creation.

The coefficient of the Common Market for the East and Southern Africa Trade

Creation dummy variable (COMESACij) was positive and statistically significant at the 1

per cent level, indicating that participation in the regional trade arrangement influences

trade flows and leads to trade creation. The coefficient of the Common Market of the East

and Southern Africa Trade Diversion dummy variable (COMESADij) was positive and

statistically significant at the 1 per cent level, this suggests that these arrangements are

important to these members performance and the impact was so strong that it generated

trade between these countries and non-member countries resulting from the fact that the

demand increasing income effect of Regional and Preferential Trade Agreements out -

weighed trade diversion from non-member countries.

The coefficient of the Southern African Customs Union Trade Creation dummy

variable (SACUCij) was negative and statistically significant at the 5 per cent level for

the simple reason that most of the participating countries originated from the Southern

Africa region. The negative sign of the coefficient also, indicate that participation in the

integration schemes does not stimulate enough mutual trade with member countries. The

coefficient of the Southern African Customs Union Trade Diversion dummy variable

(SACUDij) was negative but not statistically different from zero, suggesting that this

economic integration arrangement is not sufficiently strong to influence trade with non-

member countries positively and significantly. The negative signs of the (SACUCij) and

(SACUDij) coefficients are in line with the hypothesis tested. Although there are other

factors responsible for the (SACUCij) and (SACUDdij) negative signs. These factors are;

participation of smaller countries that are very similar, overlapping memberships of other

regional trade agreements like Comesa, conflicting regulations, different strategy and

objectives that can result to negative trade creation and trade diversion effects

The coefficient of the EU-South African Trade Creation dummy variable

(EUSACij) was positive and statistically significant at the 1 per cent level resulting from

the strong trade creating effects of the free trade agreement between the EU and South

Africa. It indicates that participation in the trade agreement stimulates a high volume of

trade between South Africa and the EU. The EU- South African Trade Diversion dummy

variable (EUSADij) was positive and statistically significant at the 5 per cent level

resulting from the strong impact of the trade agreement which gave rise to trade between

the trading partners and non-member countries. This also, indicates that the overall effect

of trade agreements is positive but can also lead to stronger trade stimulation with non-

member countries.

The coefficient of the preference Trade Creation dummy variable (PTAcij) was

positive and statistically significant at the 10 per cent level indicating that participation in

preferential trade agreements induce trade flows, stimulate mutual trade and leads to

trade creation. It also suggests that Regional and Preferential Trade Agreements had a

positive and statistically significant effect on overall trade. The coefficient of the

preference Trade Diversion dummy variable (PTAdij) was positive and highly significant

at the 1 per cent level, resulting from the overall effect of preferential trade agreements

which gave rise to additional trade which did not have enough trade diverting effect with

non-member countries especially, with the presence of the United States in the trade

matrix. The demand increasing income effect of regional and preferential trade

arrangements outweighs any trade diverting effect, which resulted to high volume of

trade with non-member countries.

The coefficient of the History dummy variable (history) was negative and

significant at the 10 per cent level but had the expected signs indicating that common

colonial links between the Great Britain and South Africa as a result of the apartheid era

also had a negative impact on trade as expected. The coefficient of the Landlocked

dummy variable (landlocked) was negative and not significantly different from zero as

expected, indicating that inaccessibility to sea or ocean transportation hinders countries

ability to trade with each other.

The coefficient of the language dummy variable (language) was positive but not

significantly different from zero. The sign was as expected because common language

amongst countries facilitates trade negotiations. Most of the countries considered in the

study speak English as a common language, which is responsible for the positive sign.

There is a positive relationship between language and trade. The coefficient of the

borders dummy variable (borders) was positive and statistically significant at the 5 per

cent level indicating that trade tends to increase if countries shared a common land border

(i.e. there is a positive relationship between the adjacency variable and trade between

countries).

The R2 coefficients for the estimated equation was 0.5725 and is at satisfactory

levels for cross section analysis, although they are somewhat lower than those obtained in

other previous gravity equation applications to trade. This indicates that 57 per cent of

variation in bilateral trade flows is explained by the variables used in the model. The F

value is 110.06 indicating that all the variables are relevant to the model.

Export Model

Table 4.4 shows the results of the estimations based on exports trade from one

country to the other. The GDP coefficient was positive and statistically significant at the

1 per cent level as expected. This is as a result of the positive relationship between trade

and the income of the recipient countries. The coefficient of distance variable (dist) is

negative and statistically significant at the 1 per cent level significant indicating the trade

barrier effects of transportation costs. A country faces higher trading costs if the port is

not located in the economic center.

The coefficient of the European Union Trade Creation dummy variable (EUCij)

was positive and statistically significant at the 5 per cent level, indicating the trade

creation effects of participation in a regional trade arrangement. This suggests that

economic integration scheme is sufficiently deep and strong to influence the mutual trade

between member countries positively and significantly. The coefficient of the European

Union Trade Diversion dummy variable (EUDij) is positive and statistically significant at

the 10 per cent level, suggesting that the trade diverting effects of the EU is minimal

compared to the trade creation.

The coefficient of the Common Market for the East and Southern Africa Trade

Creation dummy variable (COMESACij) was positive and statistically significant at the 1

per cent level indicating that participation in the regional trade arrangement enhances

exports and has the trade creating capabilities. The coefficient of the Common Market for

the East and Southern Africa Trade Creation dummy variable (COMESADij) was

positive and statistically significant at the 1 per cent level, this suggests that these

arrangements are important to these members performance and the impact was so strong

that it generated trade between these countries and non-member countries resulting from

the demand increasing income effect of RPTAs which out-weighs the trade diverting

effect.

The coefficient of the Southern African Customs Union Trade Creation dummy

variable (SACUCij) was negative and statistically significant at the 1 per cent level for

the simple reason that most of the participating countries originated from the Southern

Africa region. The negative sign of the coefficient also, indicate that participation in the

integration schemes does not stimulate enough mutual trade. The coefficient of the

Southern African Customs Union Trade Diversion dummy variable (SACUDij) was

negative but not statistically different from zero, suggesting that this economic

integration arrangement is not sufficiently strong to influence trade with non-member

countries significantly. The negative signs of the (SACUCij) and (SACUDij) coefficients

are in line with the hypothesis tested. Although there are other factors responsible for the

(SACUCij) and (SACUDij) negative signs. These factors are; participation of smaller

countries that are very similar, overlapping memberships of other regional trade

agreements like Comesa, conflicting regulations, different strategy and objectives that

can result to negative trade creation and trade diversion effects.

The coefficient of the EU - South African Trade Creation dummy variable

(EUSACij) was positive and statistically significant at the 1 per cent level resulting from

the strong trade creating effects of the free trade agreement between the EU and South

Africa. It indicates that participation in the trade agreement stimulates a high volume of

trade between South Africa and the EU. The coefficient of the EU - South African Trade

Diversion dummy variable (EUSADij) was positive and significant at the 5 per cent level,

resulting from the strong impact of the trade agreement towards generating trade between

the trading partners and non-member countries. This also, indicates that the overall effect

of trade agreements is positive but can also lead to stronger trade relations with non-

member countries, which led to minimal trade diverting tendencies.

The coefficient of the preference Trade Creation dummy variable (PTAcij) was

positive and statistically significant at the 10 per cent level indicating that participation in

preferential trade agreements induce trade flows, stimulate mutual trade and leads to

trade creation. It also suggests that Regional and Preferential Trade Agreements PTAs

had a positive and statistically significant effect on overall trade. The coefficient of the

preference Trade Diversion dummy variable (PTAdij) was positive and statistically

significant at the 1 per cent level, resulting from the overall effect of preferential trade

agreements which gave rise to additional trade which did not have enough trade diverting

effect with non-member countries especially, with the presence of the U.S in the trade

matrix. The demand increasing income effect of regional and preferential trade

arrangements outweighs any trade diverting effect, which resulted to high volume of

trade with non-member countries.

The coefficient of the History dummy variable (history) was negative and

significant at the 1 per cent level, but had the expected signs indicating that common

colonial ties between the Great Britain and South Africa coupled with apartheid had a

negative impact on trade as expected. The coefficient of the Land locked dummy variable

(landlocked) was negative and not significantly different from zero as expected,

indicating that lack of access to sea or ocean transportation hinders countries ability to

trade with each other.

The coefficient of the Language dummy variable (language) was positive but not

significantly different from zero. The sign was as expected because common language

amongst countries facilitates trade negotiations. Most of the countries considered in the

study speak English as a common language which is responsible for the positive sign.

There is a positive relationship between language and trade. The coefficient of the

borders dummy variable (borders) was positive and statistically significant at the 5 per

cent level indicating that trade tended to increase if countries shared a common land

border (i.e. there is a positive relationship between the adjacency variable and trade

between countries).

The R2 coefficient for the estimated equation was 0.5646. This indicates that 56

per cent of variation in export trade is explained by the variables used in the model. (The

explanation for the export trade is the same as in the bilateral trade). The F value is

106.57 indicating that all the variables are relevant to the model.

Summary

The gravity model was used to estimate trade creation and trade diversion effects

in the EU and South Africa bilateral and exports trade. The results of the regression

analyses show that the explanatory variables explain more than 57% and 56% of the

variation in the dependent variables of the standard gravity equation used in this study.

The estimates for the entire group of countries confirm the hypothesis put

forward in the analysis that participation in a Regional Preferential Trade Agreement

leads to Trade Creation. All the regression coefficients have the expected sign, and most

including the dummy variables are statistically different from zero. Significant

coefficients for GDP in the analysis confirm that it has a positive relationship with trade.

The negative and statistically significant coefficients of the distance variable indicates the

trade barrier impact of transportation costs,

The significant coefficients of the preference dummy variables suggest that this

economic integration scheme is sufficiently deep to influence the mutual trade between

member countries. The positive sign of the preference dummy variables indicates that

participation in trade agreements stimulate mutual trade and lead to trade with other non

member nations because the income effect is sufficiently strong to create trade between

member countries.

The significance and size of the coefficients for the preferential trade agreements

suggest that these arrangements create trade with non-member countries compared to

diversion. The significance and size of the Trade Creation and Diversion dummy

coefficients (COMESAcij, COMESAdij, EUSAcij, EUSAdij SACUcij, PTAcij and

PTAdij), suggest that these arrangements are important to the performance of

participating countries.

The coefficients on the history dummy indicate that the impact of the British

Colonial links was weak as a result of recovery from the negative impact of apartheid era.

The coefficient of landlocked was negative and not significant. The coefficient on the

language dummy was positive but not significant. The coefficient on the borders was

positive and significant. The extent of trade flows can increase if countries share a

common land border (i.e. there is a positive sign on the adjacent variable).

The R2 coefficients for the estimated equation were 56% and 57% which were at

satisfactory levels for cross section analysis, although they were somewhat lower than

those obtained in some other gravity equation application to international trade.

Lastly, the overall statistical significance of trade creating and trade diverting

effects were tested for, as shown in tables 4.3 and 4.4 respectively. The hypothesis that

trade creation and trade diversion effects were zero was rejected at all statistical levels.

The trade creating effects were positive as expected. Furthermore, the trade diverting

effects were also positive and higher than the trade creating effects.

Table 4.3 Gravity Model Estimated Results (Log (Bilateral Trade) as

dependent Variable)

Variables Model Coefficient Standard Error

Intercept -7.7551E-12 2.47641

LNGDP 0.76147* 0.04129

LNDISTANCE -1.09102* 0.11538

EUCIJ 3.07695** 0.92809

EUDIJ 0.92739*** 0.61214

COMESACIJ 3.29731* 0.68448

COMESADIJ 2.46171* 0.28370

SACUCIJ -2.55774* 0.47356

SACUDIJ -0.06120 0.25682

EUSACIJ 2.30857* 0.51991

EUSADIJ 1.38073** 0.48727

PTACIJ 3.84644*** 2.75702

PTADIJ 9.18651* 2.72985

HISTORY 0.0131*** 0.18103

LANDLOCKED -0.27860 0.26143

LANGUAGE 0.01317 0.01278

BORDERS 0.68413** 0.20022

R-Square

Number of observations

F-Value

(overall) p-Value

0.5725

1331

106.57

<.0001

*=0.01

**= 0.05

***=0.1

(level of significance)

Table 4.4 Gravity Model estimated Results (Log (Exports) as Dependent

Variable)

Variables Model Coefficient Standard Error

Intercept 1.39611E-12 2.46066

LNGDP 0.69616* 0.04102

LNDISTANCE -1.13124* 0.11465

EUCIJ 2.23487** 0.11465

EUDIJ 0.98009*** 0.92020

COMESACIJ 2.55126* 0.68012

COMESADIJ 1.72349* 0.28189

SACUCIJ -2.27579* 0.47055

SACUDIJ -0.28629 0.25519

EUSACIJ 2.19948* 0.51660

EUSADIJ 1.00907** 0.48417

PTACIJ 4.74552*** 2.73949

PTADIJ 9.79951* 2.71249

HISTORY -0.44683* 0.17988

LANDLOCKED -0.36194 0.25976

LANGUAGE 0.01740 0.01270

BORDERS 0.48936** 0.19895

R-Square

Number of observations

F-Value

(overall) p-Value

0.5646

1331

106.57

<.0001

*=0.01

**= 0.05

***=0.1

(level of significance)

CHAPTER 5

SUMMARY AND CONCLUSIONS

Summary

The economic structure and the level of economic development differ between

these different groupings of countries. The difference in economic structure and

economic development among these countries may cause the effects of the variables such

as GDP, distance etc., on trade flows to differ from one group to another.

This study has considered trade creation and trade diversion effects of the EU,

COMESA, SACU and EUSAFTA regional and preferential trade agreements. The

objective of this study was accomplished using the standard gravity model of

international trade. Countries have developed more mutual foreign trade relations with

other countries where GDP is higher. Distance negatively influences trade flows. Nearby

country pairs have developed more active foreign trade relations with each other.

Participation in the regional and preferential trade agreements influences trade flows and

leads to trade creation.

The positive and significant GDP coefficients confirm that bilateral and export

trade is strongly affected by the trading partner’s incomes. The negative but statistically

significant coefficients of the distance variable indicate the trade barrier impact of

transportation costs, but the extent of trade flows can increase if the countries share a

common land border which was positive and statistically different from zero. The

coefficients on landlocked was negative but not statistically different from zero. History

coefficient was negative but statistically different from zero. Most preferential trade

agreement variables were statistically significant at the 5% and 10% levels of

significance, but SACU trade creating dummy variable coefficients were negative but

statistically significant at the 10% level of significance. The overall trade creation

coefficients were positive and statistically significant at the 5% and 10% levels of

significance except for SACU, which was negative but statistically significant at the 5%

level.

The overall trade diversion effects were positive and statistically significant at the

5% level except for SACU, which was negative and insignificant. The creation

coefficients indicate that preferential trade agreements create trade opportunities for

member countries. The positive effects of a trade-diverting dummy suggest that

additional trade due to preferential trade agreements does not lead to diverting trade with

non-member countries. Preferential trade agreements most likely stimulate demand for

imports and supply of exports from non participating countries by increasing countries

overall income.

Conclusions

The overall trade creation and trade diversion effects of (EUSAFTA) along with

other regional preferential trade arrangements were analyzed for bilateral and export

trade respectively using the gravity model. The overall effects of regional and preferential

trade agreements are positive and significant indicating that trade agreements, induce and

generate trade among member countries. The trade creating effects of (SACU) was

negative and statistically significant perhaps because these countries operate within the

same geographical area and have similar economies.

Overall trade-diverting effects were positive, suggesting that regional and

preferential trade agreements do not necessarily divert trade with non-member countries.

The rationale for this is that the trade creating impact of regional and preferential trade

agreements increases overall demand to such an extent that the income effect outweighs

the trade diverting effect of the agreements. It has been noted that the benefits of regional

and preferential trade agreements are larger for member countries than for non-member

countries. Overall, the results suggest that there were significant trade creation and trade

diversion effects in the European Union and South Africa preferential trade agreement.

Further Study and Limitations

This study showed that trade can be influenced positively when countries

participate in regional and preferential trade agreements as a result of trade creation

effects which lead to demand increasing income effects that out-weigh any trade

diversion effect with non-members. The usefulness of using cross-section data in a

gravity model to assess the effects of most regional and preferential trade agreement was

highlighted by this study. Contrary to the use of bilateral trade flows in estimating the

gravity model, the use of export of one of the trading country pairs was introduced in this

study. One implication of this study is the positive trade creation and trade diversion

effects of being a participating country in a regional and preferential trade agreement

contrary to the notion of negative trade diversion effects.

The limitations of the gravity model include the inability of the model to predict

the welfare effects of Regional and Preferential Trade Agreements. The second limitation

is the model’s dependence upon aggregated data as opposed to disaggregated data which

can help in analyzing the effects of trade agreements on specific commodities.

Trade creation and trade diversion effects in the EU-South Africa Free Trade Agreement.

Students also viewed