+ All Categories
Home > Documents > South Africa’s Trade Flows -A Gravity Model Analysis

South Africa’s Trade Flows -A Gravity Model Analysis

Date post: 18-Jul-2016
Category:
Upload: emmanuel
View: 20 times
Download: 1 times
Share this document with a friend
Description:
Isolated from the rest of the world for most of the 20th century, a post-1994 democratic South Africa (SA) has embarked on an outward oriented trade policy that has culminated into the signing of four major free trade arrangements, i.e, Southern Africa Development Community (SADC) in 1994; Trade and Development Cooperation Agreement (TDCA) with the European Union in 1999; African Growth and Opportunity Act (AGOA) with the USA in 2000 and the Brazil, Russia, India and China (BRIC) group in 2010. It is well documented that the resulting impact of Free Trade Agreement (FTA) can be positive or negative depending on the free trade area being studied. This study therefore surveys the literature on the dynamics of free trade and uses a modified gravity model framework to empirically test the trade effects on SA of EU, USA, SADC and BRIC. Using bilateral trade data between South Africa and 100 selected countries over a 15 year period the study shows that both the EU and SADC have had expansionary effects on trade for South African exporters as well as foreign exporters in the EU and SADC. The EU-SA trade agreement has increased South Africa’s imports more than it has increased its exports. However, the reverse is true for SADC; South Africa’s exports to SADC have increased more significantly than its imports from this FTA. The study finds no evidence for either trade creation or trade diversion effects for AGOA and BRIC. The study suggests that in the case of BRIC, the adjustment period is just too short for significant changes in trade patterns to emerge. The desired FTA impact should be given time before any meaningful conclusions can be made. On the other hand in spite of clear evidence, AGOA can produce better results for both USA and South Africa if full liberalization is considered.
319
UNIVERSITY OF KWAZULU-NATAL SOUTH AFRICA’S TRADE FLOWS: A GRAVITY MODEL ANALYSIS By Emmanuel Ndakaza 210555658 A dissertation submitted in fulfillment of the requirements for the degree of Master of Commerce (Economics)
Transcript
Page 1: South Africa’s Trade Flows -A Gravity Model Analysis

UNIVERSITY OF KWAZULU-NATAL

SOUTH AFRICA’S TRADE FLOWS: A GRAVITY MODEL ANALYSIS

By

Emmanuel Ndakaza

210555658

A dissertation submitted in fulfillment of the requirements for the degree of

Master of Commerce (Economics)

School of Accounting, Economics and Finance

College of Law and Management Studies

University of KwaZulu-Natal

Pietermaritzburg

Supervisor: Ms Vanessa Tang

November

2012

Page 2: South Africa’s Trade Flows -A Gravity Model Analysis

2

Declaration

I Emmanuel Ndakaza, hereby declare that

(i). The research reported in this dissertation except where otherwise indicated is

my original research.

(ii).This dissertation has not been submitted for any degree or examination at any

other university.

(iii). This dissertation does not contain other persons’ data, pictures, graphs or

other information, unless specifically acknowledged as being sourced from

other persons.

(iv). This dissertation does not contain other persons’ writing, unless specifically

acknowledged as being sourced from other researchers. Where other written

sources have been quoted, then:

a). Their words have been re-written but the general information attributed

to them has been referenced:

b). Their exact words have been used; their writing has been placed inside

quotation marks, and referenced.

(v).This dissertation does not contain text, graphics or tables copied and pasted

from the internet, unless specifically acknowledged, and the source being

detailed in the dissertation and in the reference sections.

Signature……

Date:

Page 3: South Africa’s Trade Flows -A Gravity Model Analysis

I

ACKNOWLEDGEMENTS

First I express my gratitude to Rev. Dr. Dennis Bailey and Mrs. Gillian Bailey for

your patient selfless support and motivation throughout my masters program. Den,

your editing has made my work more satisfying.

No words can explain my appreciation of Miss Vanessa Tang for your continued

guidance and dedication to my progress. Your interest in my work, expanse

knowledge of international economics and the enthusiasm with which you helped me

has made such a big difference to this work.

To Mich and Jono, thanks for your friendship and for those morning and afternoon

rides to and from Varsity! To Dan, Morgs, Mark, Maryke, Trish, and all “Hilton-

pharmacians” for your friendship and love!

To the almighty! I am lost for words!

Page 4: South Africa’s Trade Flows -A Gravity Model Analysis

II

DEDICATIONS

To my late father! I am sure you would be proud. May your soul rest in peace!

Page 5: South Africa’s Trade Flows -A Gravity Model Analysis

III

Abstract

Isolated from the rest of the world for most of the 20 th century, a post-1994

democratic South Africa (SA) has embarked on an outward oriented trade policy that

has culminated into the signing of four major free trade arrangements, i.e, Southern

Africa Development Community (SADC) in 1994; Trade and Development

Cooperation Agreement (TDCA) with the European Union in 1999; African Growth

and Opportunity Act (AGOA) with the USA in 2000 and the Brazil, Russia, India and

China (BRIC) group in 2010. It is well documented that the resulting impact of Free

Trade Agreement (FTA) can be positive or negative depending on the free trade area

being studied.

This study therefore surveys the literature on the dynamics of free trade and uses a

modified gravity model framework to empirically test the trade effects on SA of EU,

USA, SADC and BRIC. Using bilateral trade data between South Africa and 100

selected countries over a 15 year period the study shows that both the EU and SADC

have had expansionary effects on trade for South African exporters as well as foreign

exporters in the EU and SADC. The EU-SA trade agreement has increased South

Africa’s imports more than it has increased its exports. However, the reverse is true

for SADC; South Africa’s exports to SADC have increased more significantly than its

imports from this FTA. The study finds no evidence for either trade creation or trade

diversion effects for AGOA and BRIC.

Page 6: South Africa’s Trade Flows -A Gravity Model Analysis

IV

The study suggests that in the case of BRIC, the adjustment period is just too short for

significant changes in trade patterns to emerge. The desired FTA impact should be

given time before any meaningful conclusions can be made. On the other hand in

spite of clear evidence, AGOA can produce better results for both USA and South

Africa if full liberalization is considered.

Page 7: South Africa’s Trade Flows -A Gravity Model Analysis

V

Table of Contents

Declaration......................................................................................................................I

ACKNOWLEDGEMENTS..........................................................................................II

DEDICATIONS...........................................................................................................III

Abstract........................................................................................................................IV

Table of Contents.........................................................................................................VI

List of Figures...............................................................................................................X

CHAPTER I...................................................................................................................1

1.1. Introduction.........................................................................................................1

1.2. Background and Context.....................................................................................3

1.3. Problem Statement..............................................................................................5

1.4. Research and Study Objectives...........................................................................6

1.5. Contribution and Relevance of the Study...........................................................6

1.6. Structure of the Study..........................................................................................7

CHAPTER II..................................................................................................................8

EVOLUTION OF THE THEORY OF INTERNATIONAL TRADE...........................8

2.1. Introduction.............................................................................................................8

2.1.1. The classical theory of international trade...........................................................9

2.1.2. The Heckscher-Ohlinn (H-O) model.................................................................14

2.1.3. The modern theory of international trade...........................................................16

2.2. Summary...............................................................................................................24

CHAPTER III...............................................................................................................27

THE GAINS FROM TRADE......................................................................................27

3.1. The regionalism debate.........................................................................................27

3.2. Trade creation versus trade diversion...................................................................39

Page 8: South Africa’s Trade Flows -A Gravity Model Analysis

VI

3.2.1. Conclusion..........................................................................................................46

CHAPTER IV..............................................................................................................48

4. THE DETERMINANTS OF TRADE AND TRADE PATTERNS BETWEEN SOUTH AFRICA AND PARTNERS..........................................................................48

4.1. Introduction...........................................................................................................48

4.1.1. Trade openness...................................................................................................50

4.1.2. Intra Industry Trade............................................................................................52

4.1.3. Differences in product characteristics................................................................57

4.2. South Africa’s trade patterns with EU, SADC, BRIC and AGOA.......................61

4.3. Conclusion.............................................................................................................71

CHAPTER V................................................................................................................72

THE GRAVITY MODEL............................................................................................72

5.1. Introduction...........................................................................................................72

5.1.1. Limitations of the gravity model........................................................................75

5.2. The theoretical gravity equation............................................................................75

5.3. Methodology and Data..........................................................................................77

5.3.1. Selection of multilateral resistance variables and hypothesis............................81

5.3.2. Estimation technique..........................................................................................83

5.3.2.1. The Fixed effects Model.................................................................................84

5.3.3. Data and Variable description............................................................................89

CHAPTER VI..............................................................................................................91

ESTIMATION RESULTS...........................................................................................91

6.1. INTRODUCTION.................................................................................................91

6.2. ESTIMATION RESULTS....................................................................................92

6.2.2. Panel Estimations...............................................................................................96

Page 9: South Africa’s Trade Flows -A Gravity Model Analysis

VII

CHAPTER VII...........................................................................................................104

CONCLUSION AND RECOMMENDATIONS.......................................................104

7.1. Conclusion and Recommendations.....................................................................104

Bibliography...............................................................................................................108

Page 10: South Africa’s Trade Flows -A Gravity Model Analysis

VIII

List of Tables

Table 2. 1: Evidence of Intra-industry trade in computer parts between Germany and

US, UK, Japan, France, 2009-2010..............................................................................20

Table 4. 1: Intra Industry Trade between South Africa and EU (2009-2011)..............56

Table 4. 2: Trade in primary products between South Africa and Malawi, Zambia,

Zimbabwe, 2009-2010..................................................................................................62

Table 4. 3: Trade in primary products between South Africa and Malawi, Zambia,

Zimbabwe, Mozambique, 2009-2010..........................................................................63

Table 4. 4: Direction of Trade in both primary and secondary products between South

Africa and Germany, UK, US, 2009-2010...................................................................65

Table 4. 5: Trade in top seven exports between South Africa and USA 2010.............67

Table 6. 1: Cross section estimation results.................................................................92

Table 6. 2: Stage 1 Estimation results: Exports model.................................................97

Table 6. 3: Stage 1 estimation results: Imports model.................................................97

Table 6. 4: Stage 2: Fixed effects estimation results: Bilateral trade flows model....100

Table 6. 5: Stage 2 estimation results: Exports model...............................................101

Table 6. 6: Stage 2 estimation results: Imports Model...............................................102

Page 11: South Africa’s Trade Flows -A Gravity Model Analysis

IX

List of Figures

Figure 4.1: South Africa’s GDP, 1997-2010..............................................................49

Figure 4.2: Trade Openness index for South Africa, EU, BRIC, SADC, AGOA, 2001-

2010............................................................................................................................51

Figure 4. 3: Composition of exports from SADC, EU, BRIC and AGOA, 2009-2010.

....................................................................................................................................58

Figure 4. 4: potential markets for South African exports, 2009-2010.........................59

Figure 4. 5: Import sources and export destinations for South Africa in SADC, EU,

BRIC, and AGOA......................................................................................................69

Page 12: South Africa’s Trade Flows -A Gravity Model Analysis

1

CHAPTER I

1.1. Introduction

In the last two decades, the world has witnessed an expansion in international trade

and investment flows which can be attributed to a concerted international effort to

liberalize trade. By removing barriers to trade, countries can accelerate their

integration in the world economy and increase their participation in international trade

(Robles, 2008). For Africa, this shift towards increased liberalization can potentially

improve the economic growth and development prospects of African countries; thus

there are efficiency gains from free trade (Krugman, 2012).

In the case of South Africa, since the advent of democracy in 1994, the country has

become a major player in the global trading system and as a contracting party to the

General Agreement on Tariffs and Trade (GATT) and a member of the World Trade

Organization (WTO), a series of trade reforms have been implemented. South Africa

is also party to a number of bilateral and regional Free Trade Agreements (FTAs).

For instance, with the Southern African Development Community (SADC) in 1994;

Trade and Development Cooperation Agreement (TDCA) with the European Union in

1999; African Growth and Opportunity Act (AGOA) with the United States of

America (USA) in 2000 and the Brazil, Russia, India and China (BRIC) emerging

group in 2010. It is important to highlight that these free trade arrangements (besides

Page 13: South Africa’s Trade Flows -A Gravity Model Analysis

2

BRIC) “all came into force in 2000/2001 after a period of phased in aggressive trade

liberalization under the Uruguay Round” (Holden and McMillan, 2006: 112).

Whether the concluded FTAs benefit South Africa is an ongoing debate. FTAs ought

to benefit South Africa as long as there is the possibility of realizing the “gains from

trade”. Although the literature does not establish conclusive causality (Kono, 2002),

estimates based on the World Bank Development Indicators database (World Bank,

2012), international trade and investment flows have increased. Between 1999 and

2008 (after the signing of the TDCA, AGOA and SADC) South Africa’s exports grew

by 45.8 percent from 34 billion US$ to over 50 billion US$ dollars. In the same

period, imports have more than doubled from 31 billion US$ to over 63 billion US$

measured in 2000 constant US$ (World Bank, 2012). Also, over the same period, net

foreign direct investment inflows for South Africa have risen by more than 300 per

cent from 1.5 billion US$ to over 5.3 billion US$ (World Bank, 2012)

These trends, show that the volume of South Africa’s trade and investment have

increased substantially. However, it is interesting to note that South Africa’s demand

for imports has surpassed foreign demand for its exports. Such trade imbalances can

have serious negative effects on the country’s growth and balance of payments if such

trends persist in the long term (Baharumshah et al, 2003). Aiming to contribute to

ongoing FTAs debate, this study seeks to empirically assess using a modified gravity

Page 14: South Africa’s Trade Flows -A Gravity Model Analysis

3

model, South Africa’s trade flows with its major trading partners (EU, USA, SADC

and BRIC).

1.2. Background and Context

South Africa as with many developing countries, faces many social and economic

challenges such as poverty, income disparities, lower competitiveness in the industrial

sector and high unemployment (Perry, 2000). FTAS have proliferated Worldwide in

recent years, and more than 300 were in place by 2009 (Lester and Mercurio, 2009)

There are obvious benefits to FTAs including larger markets and economies of scale;

however it is important to highlight that the selective application of FTAs can be

economically inefficient through the process of trade diversion.

South Africa is party to many FTAs. In 2010, South Africa joined emerging

economies Brazil, Russia, India, China, in BRIC. The agreement strengthens

economic and political cooperation (that existed among the countries even before

2010). South Africa had bilateral trade ties with India (since 1994) with the

establishment of an inter-governmental committee that aimed to facilitate trade

relations between the two countries.

On the other hand, since 1996 China and South Africa have reestablished diplomatic

ties and have made commitments to improve their bilateral trade. According to the

Page 15: South Africa’s Trade Flows -A Gravity Model Analysis

4

United Nations Commodity Trade Statistics database (UNCOMTRADE, 2012), the

combined exports from South Africa to China and India have grown from 1.3 billion

US$ (current) since 2000 to 14.3 billion US$ (current) in 2010. In the same year

(2010) the combined GDP of China and India was over 4.2 trillion in constant 2000

US$ which is slightly less than a half of the EU’s total GDP. Also, the combined

BRIC market size should present South Africa with diverse opportunities to boost

trade and investments.

South Africa has also signed in 1999, the Trade, Development and Co-operation

Agreement (TDCA) with the European Union (EU). The initial objective was to

remove trade barriers between them and moving towards a FTA by 2012. Their

agreement also includes provisions on trade in services, and investment and

development. The EU remains South Africa’s major trading partner.

Another major SA trade agreement, AGOA, signed in 2000 with the USA initially for

an agreement period of eight years has been reviewed and extended to 2015. AGOA

aims at providing Sub Saharan African (SSA) countries market access to the US

market as well as boosting investment into SSA.

Within Africa, South Africa is a member of several regional groups such as SACU,

COMESA, and SADC. For instance, SADC alone has a combined population of 257.7

Page 16: South Africa’s Trade Flows -A Gravity Model Analysis

5

million people and its GDP is over 471.1 US$. By 2008, over 85 percent of all trade

in the SADC was free (SADC today, 2010)

According to Mthembu, (2008), many SSA countries are dependent on tariffs revenue

generating between a quarter and a third of their national revenue. Thus, it is

important to highlight the potential loss or tariffs revenues forgone for South Africa. It

is true that Free trade may create winners and losers, however more relevant to this

research is the overall net gains from trade.

Given the inconclusive nature of the literature in analyzing the economics of free

trade (Thirwall, 1995, Rodrik,2001 and Nandasiri, 2008), this study aims to examine

whether FTAs (the SADC, EU, AGOA and BRIC) have benefited South Africa.

1.3. Problem Statement

International trade with the EU, SADC, USA and BRIC provides South Africa with

diverse trading opportunities, granting free access to large industrialized markets and

a consumer base that is composed of a combined population of about 2.7 billion

people. On the other hand the literature, both theoretical and empirical on the

economics of free trade remains inconclusive. While several studies find trade

creation effects under preferential trading (Baldwin, 1993; Glick and Rose 2002),

others find the trade diversion effects to be stronger (Bhagwati, 1991; Bagwell and

Staiger, 2002; Panagariya, 1996). Hence, the value of this study with these in mind.

Page 17: South Africa’s Trade Flows -A Gravity Model Analysis

6

1.4. Research and Study Objectives

Against the above background, this study aims to:

i. Survey both theoretical and empirical literature on the economics of free trade

and provide a deeper analysis of the impact of FTAs.

ii. Using a modified gravity model framework to empirically test and provide

evidence for the “gains” from FTAs with the EU, US, SADC and BRIC for South

Africa.

iii. Draw lessons learned from the above objectives.

1.5. Contribution and Relevance of the Study

It is estimated that almost every country in the world belongs to an FTA or RTA. It is

therefore in South Africa’s interests not to be isolated. On the other hand, the issue of

“free trade” is still widely contested and a common ground on its implication on both

developing and developed economies is as yet to be reached. As South Africa expands

its trading arrangements, a study aimed at assessing the impact of the country’s major

trading agreements given that sufficient time has passed will add to the regionalism

debate.

1.6. Structure of the Study

Page 18: South Africa’s Trade Flows -A Gravity Model Analysis

7

This study is set out as follows. Chapter 1 describes the background of the study, its

objectives and problem statement. Chapter 2 reviews the literature on different

theories of international trade. Chapter three discusses the potential gains from free

trade as well as the issues of trade creation and trade diversion. Chapter four analyzes

the determinants of trade and trade patterns between South Africa and its trade

partners. Chapter five reviews the methodology used in this study. Chapter Six

presents and discusses the estimation results. Chapter 7 contains the conclusion and

recommendations of the study.

Page 19: South Africa’s Trade Flows -A Gravity Model Analysis

8

CHAPTER II

EVOLUTION OF THE THEORY OF INTERNATIONAL TRADE

2.1. Introduction

The main goal of international trade is to increase trade and welfare among countries

by encouraging exports, research and development, free movement of capital among

others. International trade also allows countries to reap from external economies of

scale particularly through reverse engineering due to the impossibilities by firms to

fully protect the knowledge they create. It is no coincidence therefore that most of the

countries that have developed at a faster rate in recent decades have done so following

a period of more liberalized economic policies. Such countries include China,

Singapore, and India among others.

Until the late 1970s, international trade was largely dominated by comparative

advantage (which underlines differences in tastes, technology and factor endowments

as the basis of international trade). This followed the failure of mercantilism and the

gold standard systems as the basis of international trade and wealth of Nations

respectively.

On the other hand, with globalization and changes in the way international business is

conducted, many developments motivated economists to search for better

explanations of international trade beyond those given by the traditional trade models.

Page 20: South Africa’s Trade Flows -A Gravity Model Analysis

9

As a result many new models of imperfect competition that allow for the role of

increasing returns to scale, and the importance of government intervention have been

developed.

Economists such as Linder (1961), Leontief (1953), Krugman and Venables (1997)

amongst others, have all contributed to the development of the new trade theory. Also

worth mentioning are papers by Dixit and Norman, (1980) and Kelvin (1980) that

established the notion that increasing returns through the effects of agglomeration are

as important as comparative advantage in explaining international trade.

The following sections will discuss the main ideas behind the classical and modern

theories of international trade.

2.1.1. The classical theory of international trade

The classical theory explains the conditions and benefits of free international

exchange of goods and services. The theory is quite simple, countries engage in

international trade because they are endowed with different natural resources. Free

trade is intended to take advantage of these differences. Hence international trade

arises due to differences in countries’ tastes, technology and factor endowments

(Krugman, 1987).

A common theme of the Literature in favor of free trade is that it facilitates economic

expansion (Samad 2011; Dhawan and Biswal 1999; Hachicha 2003). The view that

Page 21: South Africa’s Trade Flows -A Gravity Model Analysis

10

trade acts as an engine of economic growth dates back to Adam Smith’s concept of

absolute advantage.

In his book, the wealth of Nations (1776), Smith compares the running of Nations as

similar to running family affairs… “it is the maxim of every prudent master of family

never to attempt to make at home when it will cost him more to make than to buy… If

a foreign country can supply us with a commodity cheaper than we ourselves can

make it, better buy it of them with some part of the product of our own industry,

employed in a way in which we have some advantage (cited by Linder, 1961).

According to Smith, trade among countries increases specialization and productivity

of labor because each country directs its resources towards the production of goods

and services in which it incurs the lowest relative costs of production.

For example if the European Union can supply South Africa with a commodity at a

price cheaper than the local price, then it is more profitable for South Africa to trade

some of its local products that it produces efficiently in exchange for EU’s cheaper

goods.

Smith’s theory relies on the assumption that the price of a commodity is solely

determined by the cost of production. Since labor is the only factor of production,

price is determined by the amount of labor required to produce a unit of a given

product. Given that labor is considered to be mobile within a country but immobile

across nations, the differences in relative prices across nations result from the

Page 22: South Africa’s Trade Flows -A Gravity Model Analysis

11

differences in labor productivity. Therefore as long as there are differences in relative

prices, Smith argues that a country would make a profit by exporting the locally

cheaper good but which commands a relatively higher price abroad in exchange for

the commodity that is produced cheaply from abroad.

Following this line of argument, it can be shown that with liberalization, the price of

goods in which each country has an absolute advantage will command a price that is

higher than that reached under autarky but lower than the price abroad because the

importing countries will only import if the price for imports is lower than the local

price. While on the other hand, the price of the commodity in which a country is less

efficient will fall. This way, free trade is not a zero sum game as all participants

benefit. The exporters benefit because the price for their export is higher than the

previous local prices under autarky although still lower than the prices abroad.

Consumers also benefit in a way that the prices for imports are lower than the

previous local prices of the same commodity under autarky.

One of the major misgivings of Smith’s principle is that if a country or region (e.g. the

EU) incurs less relative costs of production in all products than another (e.g. South

Africa), then South Africa should not trade with the EU. Under such circumstances if

South Africa was to trade freely with the EU, it would end up only importing and

local production would suffer. This paper argues that this view of international trade

is somewhat erroneous. Even though it is logically correct that a country may face

Page 23: South Africa’s Trade Flows -A Gravity Model Analysis

12

higher production costs than the rest of the world, nonetheless factors such as

differences in factor endowments, technology, diversification and product

differentiation make it very difficult for a country to be less productive in all

commodities. As long as such differences ignored by Smith are present countries can

profitably trade with each other, even if some may not have any ‘absolute advantage’.

David Ricardo complimented Smith’s theory of absolute advantage by introducing the

theory of ‘comparative advantage’. In this theory Ricardo explicitly asserts that

countries can increase their welfare by producing specifically those goods in whose

production they incur the lowest opportunity costs of production. For example even if

the EU incurs relatively lower costs of production in all commodities, South Africa

would nonetheless benefit from trade if opportunity costs are considered. The EU

specializes in the production of a commodity in which it is more efficient while South

Africa specializes in the other product in which the EU is less efficient (but still more

efficient than South Africa). That is, comparative advantage ensures that each country

has a commodity in which it incurs the lowest opportunity cost over the other

(Suranovic, 2010).

The Ricardian Model indicates that when two countries are autarkists, each country

produces some of each product, production technologies are different, and the price of

the product in which each country has a comparative advantage is lower than that of

the same good in the other country (Feenstra and Taylor, 2008). Wages are also

Page 24: South Africa’s Trade Flows -A Gravity Model Analysis

13

relatively higher in industries that are more competitive (where productivity is

highest). When countries finally open to free trade, the difference in technology

stimulates trade. For example the price for a commodity in whose production the EU

has a comparative advantage is relatively higher in South Africa. Since transport costs

are zero, South Africa will import this commodity until the relative prices are equal.

This way, free trade increases welfare by replacing expensive domestic products with

cheaper imports.

According Thirwall, (1995) comparative advantage can also be expressed as benefits

of trade creation that result from the establishment of freer trade by means of FTAs

and Custom Unions. If a trade arrangement is characterized by progressive

investment, exchange of technological know-how and some degree of specialization,

the benefits of trade increase the production capabilities of partner economies. The

acquisition of higher scale production capacities enables the growth of National

incomes and general purchasing power (or economic expansion). Intuitively,

economic expansion means that given prices that existed before freeing of trade, a

country can demand and produce more goods and services today than before.

Based on the Ricardian model therefore, as long as there are differences in prices and

production technologies between the EU, SADC, BRIC or AGOA and South Africa,

both parties should benefit from opening up boarders to each other’s goods.

Page 25: South Africa’s Trade Flows -A Gravity Model Analysis

14

The weakness of the Ricardian theory is that countries cannot trade if they have

similar endowments and it neglects the influence of transport costs on trade flows

between countries.

Another criticism against the Ricardian model is that it neglects other reasons for

trade besides technology differences. Countries may trade with each other because of

proximity (which reduces transport costs) and availability of resources and or

markets. For example, South Africa may be better at the production of beef than India

but Indians may not like beef (differences in tastes), technological differences would

not result into trade in such a scenario. Therefore, based on the Ricardian model alone

to determine the viability of free trade relationship between South Africa and EU,

SADC and US would lead to misleading conclusions because the model ignores the

potential determinants of trade beside differences in technology and prices.

2.1.2. The Heckscher-Ohlinn (H-O) model

The H-O model was introduced in 1919 by two notable economists Eli Heckscher and

Bertil Ohlin. The model is an extension to Ricardo’s comparative advantage theory in

a way that it explains why one country may have a comparative advantage over

another. Ricardo’s model assumes differences in productivity but does not explain

why there may be differences in productivity. The H-O model simplifies that question

Page 26: South Africa’s Trade Flows -A Gravity Model Analysis

15

by stating that even if countries have identical technologies, they may have different

resource endowments (Feenstra and Taylor, 2008).

The model predicts that countries export goods for whose production they have

abundant resources and import goods for which they have scarce resources. Under

autarky, the price of the commodity whose factors of production are available in

abundance will command a lower price locally vis-à-vis the rest of the world and vice

versa.

Supposing that South Africa is endowed with land and labor and scarce capital, it can

be shown that it will tend to produce more land and labor intensive goods. Since land

and labor are abundant, the price for land and the wages paid to labor are lower vis-à-

vis the rest of the world. However since capital is scarce, the price of capital is high

and the price of the commodity produced therewith is higher in South Africa than in

other countries. The difference in resource endowments is sufficient enough to cause

different PPF curves such that the price ratios would differ across countries

(Dominick 2004). South Africa would be better off importing such a commodity from

abroad.

Extending this argument further, after opening up to trade the abundant under

employed resources in South Africa flow to the rest of the world thus forcing their

prices to increase locally and fall abroad. On the other hand, the unemployed capital

abroad will be imported thus pushing the prices of capital down in the local market.

Page 27: South Africa’s Trade Flows -A Gravity Model Analysis

16

Generally, countries benefit in a way that trade brings about a more equitable

distribution by lowering wages in the resource scarce industry and increasing wages

in the resource abundant industry.

2.1.3. The modern theory of international trade

The major difference between the classical theory and the modern theory of

international trade is that the latter acknowledges the idea that government

intervention can and is capable of improving market outcomes and, to some degree,

the theory assumes that international trade is driven by increasing returns to scale that

exist under imperfect competition.

The modern theory of international trade corrects weaknesses in the Ricardian Model

by giving factor and product prices as equal importance as labor productivity in the

determination of international trade. It also relaxes other limiting assumptions, for

example: labor is considered to be heterogeneous, transport costs to be present and

production costs are variable. This theory is more fitting because the assumptions

usually hold in the real world. In fact it is very difficult to imagine a situation where

labor is homogeneous as assumed by the Ricardian model.

Despite the flaws in the traditional trade models, developing empirical models of how

different entities behave under imperfect competition was and still is very

complicated because some firms may at times behave in ways that are obscure even to

Page 28: South Africa’s Trade Flows -A Gravity Model Analysis

17

themselves (Krugman, 1987). Nevertheless, the contributions of Krugman and

Venables (1997) in showing that international trade is characterized by increasing as

opposed to constant returns to scale coupled with the work of Wassily Leontief (1953)

in proving that international trade is mainly characterized by intra-industry trade

contributed much in motivating modern economists to moving away from traditional

models.

The presence of increasing returns to scale implies that countries can engage in trade

irrespective of where comparative advantage lies because increasing returns to scale

in international trade systematically enable firms to improve their competitiveness.

This is because international trade provides a sufficient market that allows production

to rise further and in turn lower production costs. As firms specialize in the

production of commodities, they experience a reduction in their average costs of

production as production rises. Therefore, even if countries are similar in many

aspects, the modern trade theory predicts that the presence of economies of scale

enables all firms to cut their costs sufficiently enough to compete under conditions of

free trade.

This new trade theory predicts imperfect as opposed to perfect competition in

international trade. According to Pomfret (1992), imperfect competition is present in

manufacturing and agriculture among other sectors where parastatals and large

competitive companies often handle international business. Differences in production

Page 29: South Africa’s Trade Flows -A Gravity Model Analysis

18

technology enable firms in the same industry to produce differentiated versions of the

same goods. This way, countries can import different versions of the same good at

different prices. For example, countries that produce cars also import cars that are

produced from elsewhere. This is mainly because each product is seen as slightly

different from another even if they virtually serve the same purpose.

The modern trade theory also assumes that the presence of transport costs is one

reason why countries may engage in international trade. Donald, R and Weinstein,

(1998) argue that the traditional theory ignores the importance of trade costs not

because they are small or irrelevant but because they believed that trade costs do little

to change the patterns of trade. On the contrary, new trade theories consider trade

costs to be an important ingredient in determining trade patterns in a way that they

affect markets in different ways. Krugman and Venables (1995) show how trade costs

affect international trade and the distribution of production.

First, the presence of trade costs affects international trade through the home market

effect. Trade costs persuade producers to relocate their plants closer to consumers of

their products (backward linkages). This has become a very popular strategy with

many multinational companies especially in the manufacturing, telecommunication

and automobile industries among others.

Page 30: South Africa’s Trade Flows -A Gravity Model Analysis

19

Second, the existence of trade costs may persuade producers to operate in large

markets where it is easier to access differentiated inputs (forward linkages) in order to

reduce their production expenses. Such agglomeration can have serious impacts on

welfare since in extreme cases exodus of companies may be experienced leaving

some countries less developed.

The role of trade costs in determining the patterns of international trade is

demonstrated by Du Plessis, (1987). He puts forward a simple example; “If Canada

needs fertilizers in Alberta and has a surplus in Quebec, it will certainly not haul it

4000 KMs to get it there; instead, it will export the Quebec fertilizer to eastern US

and import the fertilizer for Alberta from Western US”. These costs can be made

smaller by exploring foreign market even when factor endowments and technologies

are similar.

Today there is no sign of total world specialization; evidence shows that countries

import some of the products they produce/export. The classical models do not predict

the situation where a country imports and exports the same product simultaneously.

They ignore product differentiation and intra-industry trade as some of the reasons

why countries may trade in the same goods.

Table 2. 1: Evidence of Intra-industry trade in computer parts between Germany and US, UK, Japan, France, 2009-2010

Page 31: South Africa’s Trade Flows -A Gravity Model Analysis

20

Germany imports from

(Million US$)

Germany exports to

(Million US$)

Year 2009

France 260.25 1,048.68

Japan 323.26 62.88

UK 597.43 1,098.81

US 1,109.34 340.14

Year 2010

France 239.15 1,308.3

Japan 476.6 77.93

UK 655.56 1,009.46

US 1,094.24 393.77

Source: The United Nations Commodity Trade Statistics Database (UN COMTRADE, 2011)

The table 2.1 shows the trade in ‘computers and parts thereof’ between Germany and

four of the major world economies in 2009 and 2010. Despite being close substitutes,

data indicates that each country exports and imports some of the same product (in

other words the presence of intra-industry trade and lack of specialization). For

example in 2009 Germany imported computers worth $260 million and exported

computers worth $1 billion to the same country. In the same year, Germany imports

computers worth $1 billion from the US and exports $393 million worth of computers

to the same country. This kind of trade pattern is not explained by the classical trade

theories of perfect specialization.

Page 32: South Africa’s Trade Flows -A Gravity Model Analysis

21

A number of studies have attempted to test how well the classical theory explains

international trade today. In order to establish the structural basis of USA’s trade with

other countries, Leontief (1953) tested the H-O theory by comparing the quantities of

labor and capital required to produce two goods, one export and other competitive

import. Contrary to the predictions of the H-O model Leontief’s paradoxical findings

showed that despite the USA being capital abundant, its exports were largely labor

intensive and its imports were largely capital intensive. His findings were later

complimented by the finding of Baldwin (1971) who tested 1962 US trade data

showed that imports to the USA were 27% more capital intensive USA’s exports.

The Leontief paradox however was criticized mainly on statistical grounds citing that

choosing 1947 as the base year was erroneous since the country was still dealing with

disorganizations of the second world war. Even though Leontief (1956) finding of the

second test using 1951 data sets met wide criticism on grounds of his choice of

empirical methodology from many economists such as Brecx (1967), Merrett, (1965),

his research took international economic theory by surprise and initiated a great deal

of empirical and theoretical research on the subject.

Subsequent tests of the H-O model using Leontief’s procedure showed that

considering finder categories of the main factors of production shows that USA

exports were also intensive in skilled labor, research, engineering talent among others

in which the US was well endowed.

Page 33: South Africa’s Trade Flows -A Gravity Model Analysis

22

Another attempt was made by Linder (1961). According to Linder, trade flows

between two countries are a function of market homogeneity subject to distance

constraint. A country exports more of those commodities that are highly demanded

locally. This implies that countries with similar demand structures are more likely to

trade together more than those with different local demand structures.

On this relationship, Gray (1998) notes that Linder’s analysis is directly linked to

differentiated markets in international trade. This is partly because a product

manufactured in one country is more likely to have characteristic differences as

opposed to end-use differences from a product made in another country. Since the end

use is similar, differences in product characteristics can be a basis for international

trade.

According to Linder, traditional theories of international trade, specifically the H-O

theory only explains patterns of trade in natural resource intensive products but the

theory performs poorly in explaining trade in manufactured products since their

fabrication does not require unique factor intensity. For example the same product can

be produced using labor intensive techniques in a country where labor is cheap but

also the same product can be produced in another country using capital intensive

techniques where capital is cheaper. The bottom line to Linder’s analysis is that goods

Page 34: South Africa’s Trade Flows -A Gravity Model Analysis

23

that are abundantly produced locally do not depend on factor endowments but on the

structure of local demand.

Another point of divergence between the traditional international trade theory and the

new trade theory is that modern economists believe that endowments change over

time. At the fore front of this view is the product cycle theory.

The product cycle theory assumes that the allocation of industries whose products

warranties investment in technological innovation follows a number of stages;

Stage 1: Production is introduced in large markets (high income countries) that

have the necessary resources.

Stage 2: As countries open to trade, the industry develops a capacity for export

due to its comparative advantage as predicted by the H-O model.

Stage 3: With more trade, gradually other countries also begin producing the

product due to technology and capital spillovers.

Stage 4: Industries in the initial exporting country begin to lose competitive

advantage as the technological gap with its trade partners begins to shrink.

Stage 5: Finally, the same goods that were formerly exported begin to be

imported in the initial country in the form of intra-industry trade.

The product life cycle theory assumes that developing countries with a high number

of semi-skilled labor finally gain production advantage making their goods relatively

Page 35: South Africa’s Trade Flows -A Gravity Model Analysis

24

cheaper. Gains from trade therefore depend on the stage of the product cycle and

changes in endowments.

2.2. Summary

According to the Ricardian model, trade among countries takes place because of

technological differences whereas according to smith, differences in resource

endowments are the main reason why countries trade with each other. The Hecksher-

ohlin model emphasizes the differences in relative factor endowments as the reason

why countries trade with each other. All models predict that as long as the

corresponding assumptions are met, all countries are better off by trading than not.

On the other hand Tiiu (2000) notes that classical models of perfect specialization are

limiting in a sense that they explain trade only in specific items but cannot explain

why countries have stronger trade links with some countries more than others. Also

they ignore the possibility that endowments may change or be transferred over time

which may change the patterns of trade among countries. In a nutshell, the classical

models fail to explain why trade increases over time for some countries and fail for

others. It should be noted however that the classical models are more suitable in

explaining exchanges between developing and developed countries (Krugman et al,

2011) because of the certain differences in productivity and consumption patterns.

Intuitively therefore, the classical theory predicts profitable trade between South

Page 36: South Africa’s Trade Flows -A Gravity Model Analysis

25

Africa and EU; and between South Africa and the US. Given the difference in the

economic development between South Africa and SADC, the Ricardian and

Hecksher-Ohlin models predict profitable trade from such an FTA.

It should be noted that all the theories present models towards an understanding of

international trade. Each model presents some but not all reasons why countries trade

with one another. None of these theories has been proved to hold in all circumstances

for every country. It is not automatic that countries with different resources and factor

endowments will simultaneously benefit from trading among and between them.

There is no single model that encompasses all the variables that may encourage or

hamper trade among a given set of nations. They are only theories that try to predict

the likely causes and outcomes of international trade under a set of assumptions. The

actual knowledge about how countries succeed after liberalization can only be

obtained by carrying out case specific analyses of the countries of interest.

Since it is not guaranteed that South Africa automatically benefits from SADC, BRIC,

AGOA, TDCA; this study aims at providing an empirical study about the trade

creation and the trade diversion effects that different trade agreements have had on

partner countries.

Page 37: South Africa’s Trade Flows -A Gravity Model Analysis

26

CHAPTER III

THE GAINS FROM TRADE

3.1. The regionalism debate

One of the five strategic objectives of South Africa’s Department of Trade and

Industry (DTI) is “(to) build mutually beneficial regional and global relations to

advance South Africa’s trade…” (DTI, 2012). Yet it is clear over the past decade that

South Africa’s efforts towards building regional relations have far surpassed efforts

towards global economic integration despite the latter promising much wider market

prospects than any RTA can create. Such behavior in international relations has been

at the heart of the ongoing debate about the wisdom of a regional approach to global

trade liberalization.

Bhagwati (1991) divides regionalism into two periods. He writes that the “first

regionalism” was inspired by the work of Viner (1950). With the additional

contribution of economists such as Meade (1955) and Lipsey (1957) towards the

development of the Vinerian theory, the model started to get worldwide recognition

especially after the formation of the European community. Later, other economists

such as Cooper and Massel (1965) attempted to show that FTAs can effectively

reduce the costs of industrialization when extended to the developing world.

Beyond the European Community, the adoption of PTAs lacked enthusiasm

especially because of the reluctance of the USA to endorse PTAs in its trade policy.

Page 38: South Africa’s Trade Flows -A Gravity Model Analysis

27

Instead, until the late 1970s, governments in developing countries adopted the

strategy of import substitution as a base for the development of their economies. The

strategy involved the replacing of imports with local production and sheltering local

production from foreign competition by means of strict controls against foreign

investment and high import duties. Even though the strategy registered some success

in countries such as Brazil, Argentina and Mexico (Blouet and Blouet, 2005), by the

early 1980s many countries had accumulated huge debts.

Since then, the IMF and the World Bank have encouraged governments in developing

countries to open up their borders to foreign investment, and they continue to provide

financial incentives to those governments that undertake liberalization and

privatization. Today, the WTO allows the formation of free trade zones as a means

towards quicker regional integration. The “second regionalism” began when

proliferation of FTAs started to gain momentum particularly after the signing of the

Canada-US free trade area (CUFTA) in 1987. Bhagwati (1991) claims that the change

of heart on the part of the US was partially a result of the impossibilities and

complications involved in agreeing on a common multilateral agreement that could be

supported by all nations.

Today, there are more than 300 bilateral trade agreements (BTAs) worldwide. It is

estimated that at least each country belongs to a either a Regional Trade Agreement or

BTA (Lester and Mercurio, 2009). Therefore, the move towards regional integration

as opposed to multilateral liberalization is an important discussion. Yet even at this

Page 39: South Africa’s Trade Flows -A Gravity Model Analysis

28

rate of proliferation, the overall effect of RTAs on multilateral trade is yet to be

universally agreed upon. The question whether free trade blocks can effectively

replace or enhance multilateralism has yet to be answered. Instead of committing to

multilateral trade under WTO, countries are individually negotiating trade deals with

each other, and or with other regional trade zones. South Africa, for example, is

committed to multilateral trade through the World Trade Organization yet at the same

time has formally and aggressively negotiated for both bilateral and regional trade

agreements.

The implication of such economic undertakings has led to a worldwide discussion

with some analysts believing that such moves could halt multilateral negotiations,

while others believe that regionalism is a building block towards greater international

trade. Either way, proliferation of FTAs has reached another level in recent years.

This is because FTAs used to be agreements between neighboring states but in the last

15 years, even distant countries have established regional trade partnerships e.g. South

Africa and EU, EU and MERCOSUR, SACCU and MERCOSUR among others.

Those in support of FTAs argue that with such networks of bilateral agreements

between individual states and regional trade blocks, and between regional blocks

themselves, eventually the whole world will be united under the same trade

agreement.

Page 40: South Africa’s Trade Flows -A Gravity Model Analysis

29

Critics argue that RTAs are a form of protectionism that encourages countries to

remove barriers between themselves while imposing them on nonmembers. For

example, Krugman (1991), while analyzing the interaction between FTAs and

countries that trade under the MFN principle, shows that the non-cooperative behavior

of RTAs often leads to higher external tariffs.

Others argue that FTAs impede worldwide trade liberalization because governments

are focused only on securing personal trade deals and spend less energy on pursuing

multilateral liberalization under the WTO. Therefore, it is widely agreed upon by

many critics that FTAs are a big step backwards from the principles of

nondiscrimination towards which they claim to build. This is witnessed in the form of

barriers on trade imposed on nonmembers in terms of tariffs, and rules of origin that

are more likely to cause a negative welfare effect. Besides, the manner in which PTAs

are run calls for reciprocity; countries that are excluded from an RTA may be tempted

to resort to protectionism as a result of marginalization that RTAs may generate. This

erodes the possibility of global liberalization.

The idea that RTAs are a building block towards multilateralism is also contested by

Krishna (1998), citing the presence of lobbyists as a factor that prevents RTAs from

building towards global liberalization. His idea follows the same line of argument as

Grossman and Helpman (1994). According to them, lobbyists are predominantly

producers who influence government policies towards what favors their profits.

Page 41: South Africa’s Trade Flows -A Gravity Model Analysis

30

Together they argue that liberalization favors the export industry and consumers but

not import-competing industry that would rather scuttle any agreement to preserve

their protection.

Second, countries engage in RTAs only if the arrangement promises to divert trade in

their favor. From this, Grossman and Helpman (1994) assert that an FTA is more

likely to be backed and eventually formed if it promises more trade diversion effects

for the countries involved.

Krishna replicated Grossman and Helpman (1994) and considered three scenarios, the

first one being π1 (profits under WTO Most Favored Nations (MFN) tariff system), π2

being benefits under FTA and π3 being benefits under global free trade. According to

him, lobbyists can only support a move from MFN to FTA only if π2 > π1. Also a move

from MFN to global free trade can be supported only if π3> π1. However, when an

FTA is formed, switching towards a global free trade system is more likely to be

blocked by special interest groups because it only generates π3 – π2 which is smaller

than π3 - π1 because already π2> π1. He therefore concedes that once an FTA is formed,

the incentive to give access to a third partner is reduced because more liberalization

means more consumer benefits and less producer protection.

This view is seconded by Albertin (2008) who argues that as proliferation of FTAs

gains momentum, there is more evidence of reluctance by countries in pursuing

Page 42: South Africa’s Trade Flows -A Gravity Model Analysis

31

multilateral trade negotiations. Once a free trade agreement is in place, private entities

tend to form an “anti-multilateralism force” to protect their own industrial interests

that are otherwise guaranteed by preferential treatment.

Such views give an insight into a debate that has been going on for a long time. In

1950, in his book titled The Customs Issue, Viner Jacob was among the first

economists to initiate a discussion about the possibility of free trade areas diverting

trade rather than creating it. The debate about whether countries should pursue

preferential trade instead of global integration under the principle of most favored

nations is somewhat irrelevant. The reason is that preferential trade areas are mainly

intended to create new and bigger markets for participating countries. The bigger the

PTA becomes, the greater opportunities there are for partner countries. Yet, no

alternative can provide greater opportunities than multilateral trade under the WTO.

Therefore countries like South Africa would benefit more from multilateral free trade

than from preferential trading. Yet, sixty years after Viner articulated the economic

implications of preferential trading, the choice between PTAs and multilateralism

remains a big problem that many policy makers tend to ignore (Lester and Mercurio,

2009). Many have studied the importance and evolution of RTAs and derived

different conclusions.

Page 43: South Africa’s Trade Flows -A Gravity Model Analysis

32

According to Mensbrugghe et al (2005), initially countries rushed to negotiate

bilateral trade agreements in order to gain “first-mover” advantages with stronger

economies before others did. However, they note that today it is almost impossible to

obtain such gains because of the numerous networks of FTAs all over the world. In

2012, the WTO reports 511 RTAs to have been notified, of which 319 are already

operating. This means that it is virtually impossible for a country such as South Africa

to be a member to all of them. Therefore, by joining one or two FTAs South Africa

misses out on the benefits of trading freely with other FTAs, a scenario which may be

unfavorable in a sense that South Africa would be better off by having free access to

all markets without discrimination. It is argued that African countries’ membership to

numerous RTAs is a main reason why the continent has been very slow towards

integration (UNECA 2011).

Mensbrugghe et al (2005) show why countries should adopt preferential trading with

due caution. They performed a benchmark simulation on a number of global reform

scenarios and found out that globally, countries would experience a 0.8%

improvement in baseline income by 2015 if they pursued multilateral trade. On the

other hand, the study reports that developing countries would lose 0.4% in real

income if they pursued RTA arrangements in the same period. They ran a simulation

where all developing countries sign a BTA with major economies (precisely USA,

UK, Canada, Japan, EU, New Zealand and Australia) and reported that in such a

Page 44: South Africa’s Trade Flows -A Gravity Model Analysis

33

setting, developing countries would be worse off by $22 billion as compared to a

scenario where they would trade with one another without trade barriers.

It can be argued that the structure of barriers that are put in place when PTAs are

formed is one that allows free movement of goods and services within while at the

same time giving individual member countries the right to independently levy taxes

on imports from non-members. The expected result is that total trade among members

increases at the expense of non-members. Unfortunately, even for some members, the

FTA may divert trade from low cost sources to higher cost sources.

Such trade diversion happens mainly because a country dismantles tariffs for

members, thus making their exports cheaper as opposed to those from non-members.

If a non-member country is the low cost supplier, continued importation of goods

from a member country that enjoys duty exemptions means that the importing country

pays more for its imports than previously, in addition to losing tax revenues.

Even in light of such studies to the contrary, the trends in international trade today

suggest that countries are moving faster towards the formation of fragmented free

trade areas all over the world. This trend, it is believed, is partly because despite being

a second best strategy, it has been proved that preferential trading is easier to achieve.

Proponents of regionalism emphasize that RTAs create momentum for eventual

global liberalization through the initiation of intra-regional liberalization that

Page 45: South Africa’s Trade Flows -A Gravity Model Analysis

34

facilitates the integration of member countries into a more sophisticated multilateral

trading system. In other words, the growth in the number of available RTAs in the

world will eventually lead to the merger of different RTAs, thus accelerating global

trade liberalization faster than through any other means. However, this view is

challenged by Babatunde (2007) who argues that this does not apply to most of Africa

because of low intra-industry trade within RTAs, and dependence on primary and

mineral resources, in addition to lower levels of “structural complementarity”.

A report released by UNECA (2010) maintains Babatunde’s argument about the

weakness of Africa’s intra industry trade but argues that with the formation of more

RTAs, intra Africa trade today is greater than what it was a decade ago. The report

concludes that by forming RTAs, African countries can enhance regional trade

interactions that can eventually generate economic power enough to drive stronger

interactions and integration among different sub regions of Africa.

Some governments believe that opening up markets leads to greater efficiency and

economic growth, as evidenced in East Asia and other developed Nations of the west.

Competition ensures that inefficient entities are excluded from production while at the

same time presenting opportunities for competitive firms to realize their potential by

producing for a wider market. Even though multilateral trade promises a bigger

market than any FTA can generate, countries like South Africa that are still part of a

wider setting under the flag of WTO continue to negotiate FTAs for various reasons:

Page 46: South Africa’s Trade Flows -A Gravity Model Analysis

35

Multilateral negotiations under WTO take longer time due to the volume of

issues discussed and lack of mutual commitment from all parties. On the other

hand, FTAs are easier to establish as they involve fewer members and

considerably less issues to discuss.

From the developing world’s perspective, multilateral trade involves

unfavorable trade terms in the way that the trading countries are often of

unequal size. The resulting welfare effects from such trade are often negative

for weaker countries because of the differences in the bargaining powers on

the international market. It is also difficult for smaller countries to have a

significant influence on the proceedings of the WTO.

FTAs, however allow smaller and weaker countries to form a considerable

force, as a larger and strong unit that can effectively deal with stronger blocs.

From this point of view, increasing RTAs acts as an incentive towards

multilateral trade liberalization by enhancing awareness of the relative

interdependence between different trade blocs.

In addition, FTAs also have their own strengths. According to Shujiro (2002),

combinations of factors are behind the expansion of Regional Trade Agreements

(RTAs). For example, the need to secure markets for exports calls for a two-way

dismantling of trade barriers among countries. As more countries converge in

Page 47: South Africa’s Trade Flows -A Gravity Model Analysis

36

different regional trade units, countries that do not belong to any become

marginalized and as a result lose potential markets.

In their discussion of FTAs, Rosson et al (2003) distinguish between short term and

long term effects of FTAs. They explain that the short term effects are measured in

terms of creation and diversion of trade. They further argue that in the case where an

FTA is backed by full employment, it can improve the general welfare for member

countries by pushing down consumer prices and increasing baseline incomes. More

so, they predict that long term benefits such as “…increased competition, economies

of scale, stimulus investment and more use of economic resources” in FTAs are

expected to exceed the short term gains.

Also, dismantling of trade barriers and imposition of rules of origin means that

producers can benefit more by producing from within the region rather than from

outside. This attracts multilateral companies towards expanding production to larger

regional trade areas more than individual countries. For example, the figure below

presents South Africa’s FDI inflows since 1994.

Figure 3. 1: Flow of South Africa’s Foreign Direct Investment, 1994-2010

Page 48: South Africa’s Trade Flows -A Gravity Model Analysis

37

Source: calculations based on IMF data (IMF, 2011).

Figure 3.1 above clearly shows a remarkable difference in South Africa’s FDI inflows

before the year 2000 as opposed to the period beyond 2000. Since 2007, the trend line

reveals that FDI for South Africa has been increasing at an increasing rate even

though individual annual data shows a high level of fluctuation. Interestingly, even

during the financial crisis, FDI for South Africa stayed positive indicating that FDI

inflows are greater than Investment outflows. The data also reveals a very big

improvement in the value of FDI directed toward South Africa since 2005, with the

exception only being in 2006. Whether this is because of the free trade agreements

signed by South Africa a few years back remains to be empirically investigated but

what the study can say is that this period coincides with the period in which the

effects of the agreements should start to be sensed.

Page 49: South Africa’s Trade Flows -A Gravity Model Analysis

38

3.2. Trade creation versus trade diversion

The agenda of almost any free trade agreement is centred on promoting trade and thus

economic growth. Although it is true that when a country trades more, it can increase

its share of world exports and gain more influence on terms of trade in the world

market, some studies have shown that is not always the case. Opening up to trade

does not automatically guarantee economic success (Martin and Sunley (1996). ,

1996; Rodrik, 2001A; Rodrik, 2008).

Since it is possible for FTAs to divert trade from low cost producers, consumers may

end up paying the same prices as before the PTA in addition to losing tax revenues.

Therefore, benefiting from regional trade areas requires more than just removing

barriers to trade. It necessitates important “…human and institutional resources and

infrastructures…with respect to size and economic conditions” (Mina et al, 2005) of

each participating country. Many empirical studies have shown that the impact of

Free Trade Agreements on international trade flows is still mixed.

Thirwall, (1995) for example, points out that trade between developing and developed

countries has often resulted in trade diversion rather than creation. Rodrik (2001B),

amongst others, suggest that countries should only open up to freer trade when they

have a very strong local industry that can compete on the world market. It is only

Page 50: South Africa’s Trade Flows -A Gravity Model Analysis

39

when nations have a strong economic base that they can start benefiting from

international trade.

Contrary to Thirwall (1995) and Rodrik (2001A), Nandasiri, (2008) used an

augmented gravity model and analyzed panel trade data of 184 countries over a period

of nine years with reference to seven regional trade blocks namely, the Association of

South East Asian Nations (ASEAN), North American Free Trade Agreement

(NAFTA), the European Free Trade Association (EFTA), Dominican Republic –

Central America Free Trade Agreement (DR-CAFTA), Caribbean Community

(CARRICOM), South Asian Association For Regional Cooperation (SAARC) and the

European Union). He found that countries that formed a Free Trade Area with one or

more of these regional trade blocks gained more from trade liberalization than those

that didn’t. He concludes that in this case, FTAs created trade for member countries

and diverted trade for non-members.

Bhagwati and Panagariya (1996) explain the cause of such developments; they argue

that RTAs unambiguously give preferential treatment to partner countries while the

same time discriminating against nonmembers. In this way, internal trade within the

RTA expands to the detriment of the rest of the world.

On the other hand, some studies have produced positive results in favor of preferential

trading. For example, trying study the impact of NAFTA on trade flows between the

Page 51: South Africa’s Trade Flows -A Gravity Model Analysis

40

USA and Mexico, Susanto et al (2007) constructed import demand functions for the

two countries and included time dummies to capture whether the removal and

reductions of tariffs after the implementation of the free trade agreement affected

exports from Mexico to the USA. The results showed that following reductions in

tariffs, Mexico’s exports to the USA increased significantly. They concluded that

NAFTA has created trade for both Mexico and the United States of America.

Besides empirical studies, another reason that is frequently put forward by those in

support of preferential trading is that countries that are within each other’s proximity

are natural trading partners (Krugman, 1991). Their lower transport costs resulting

from proximity make them more likely to benefit from trading with each other than if

they pursue protection policies. The argument of countries being natural trade partners

is based on a number of presumptions. First, it is assumed that if countries are within

each other’s proximity, their volumes of trade are naturally higher, notably because of

low transport costs. Secondly, since trade is naturally high, trade creation effects are

greater than diversion effects with the creation of an FTA.

This view is challenged on the other hand by those who believe that neighborliness

does not necessarily mean natural trade partnership. They argue that trade creation

does not depend on transport costs and benefits accruing from proximity only but it

also relies heavily on the importing country’s GDP (purchasing power) and the

exporter’s price (Panagariya, 2000).

Page 52: South Africa’s Trade Flows -A Gravity Model Analysis

41

Supposing three countries, South Africa (home country) imports a particular

commodity from Zimbabwe and Kenya at a fixed price (Pzim+t and Pken+t); where”

t” represents import tax.

As Viner (1950) illustrates, initially both Kenya and Zimbabwe would face the same

tax rates. Assume further that the prices in Kenya are lower than the price of the same

commodity in Zimbabwe. If South Africa and Zimbabwe join an FTA (SADC for

example) while South Africa maintains the same level of tariff against Kenya, the

implication of this move on South Africa’s welfare can be represented as below:

Figure 3. 2: Trade creation and trade diversion effects of an FTA (Vinerian Approach)

Page 53: South Africa’s Trade Flows -A Gravity Model Analysis

42

Source: Panagariya(2000)

Ekenya and Ezimbabwe represent the quantities which each country is willing to

supply under free trade. E’kenya and E’zimbabwe are quantities that both countries

are willing and able to supply under a common tariff ‘t’. It can be seen that in both

scenarios, Kenya supplies more at a relatively lower price. Hence South African

importers purchase all imports from Kenya. However, with the creation of an FTA

between South Africa and Zimbabwe, the price faced by Zimbabwean exporters is

Pzim (that is free of import duty ‘t’) whereas the tariffs against Kenya are left intact.

The exporting price for Kenya is Pkenya+t. This makes Zimbabwean goods cheaper

1

2

3

CD

E Zimbabwe

E’ Zimbabwe

E’ Kenya

E Kenya

MPTAz MPTAk

M0

Pzim+t

Pken+t

Pzim

Pken

Page 54: South Africa’s Trade Flows -A Gravity Model Analysis

43

than Kenyan ones. As a result, total imports increase from M0 to MPTAz. This has

the following implications:

i. Since Zimbabwean duty free imports are now cheaper in contrast with Kenyan

goods, the source of imports is entirely diverted from Kenya to Zimbabwe thus

making Kenya worse off.

ii. The increase in exports is less than it would have been had the PTA been formed

between South Africa and Kenya. The gains in welfare would have been

1+2+3+C+D less the loss in taxes equal area 1+2. Net welfare gains would have

been 3+C+D which is unquestionably positive. But since the FTA is between

South Africa and Zimbabwe, the gain in welfare is the area 1+3 and the loss in

taxes is area 1+2. The net welfare gain is area 3-2 which may or may not be

positive. This move sees South Africa forego the gain in welfare totaling area

C+D.

iii. If Zimbabwe does not have the capacity to satisfy the increase in demand for its

exports, South Africa continues to import a fraction of the goods from Kenya as

well. Given that South Africa cannot have two different local prices for the same

commodity, the local price will be set to equal Kenyan import prices. The

implication is shown by Panagariya (2000)

Figure 3.3: Trade diversion effects of an FTA (Vinerian Approach)

Page 55: South Africa’s Trade Flows -A Gravity Model Analysis

44

Source: Panagariya (2000)

The removal of import duties against Zimbabwe shifts the supply curve from

E’zimbabwe to Ezimbabwe. Imports from Zimbabwe increase from MTZIM to MZIM but

local demand dictates total imports totaling M0. Kenya supplies the difference

between MZIM and M0 at price Pkenya+T. This sees South African consumers gain

no additional welfare while at the same time the government of South Africa loses

tariff revenues equaling area 1+2+3+4. However, Zimbabwean exporters gain area

1+2+3 in form of producer surplus. The net loss to Zimbabwe and South Africa is

area 4.

E’zimbabwe

Ezimbabwe

E’kenya

Ekenya

MTZIM MZIM M0

Pkenya

Pkenya+T1

2

3

4

Page 56: South Africa’s Trade Flows -A Gravity Model Analysis

45

As shown using the Vinerian approach above, FTAs have the potential to create as

well as divert trade. When a PTA is formed with the most efficient supplier, the loss

in tax revenue is surpassed by the gains in welfare to consumers in the form of

reduced local prices and increased quantity. This way the agreement strengthens

instead of diverting trade.

3.2.1. Conclusion

A number of studies have tried to assess the trade creating and diverting effects of

various FTAs and their results have often been mixed. As a result, feelings about the

importance of FTAs to member and nonmember countries have been mixed. It is

generally understood that the effects of an FTA are not necessarily uniform across all

involved countries. It is possible that an FTA may create trade for one country and

divert it from another. The results are specific to only the trade arrangement being

studied and therefore cannot be generalized to fit all trade blocks. There is a need for

a thorough investigation of this subject from South Africa’s perspective. South

Africa’s strategy of negotiating FTAs with other economies appears to be a second-

best option. But since universal liberalization is unattainable in the foreseeable future,

negotiating free trade deals is better than not. The question whether the move has

been detrimental to South Africa’s trade remains to be empirically examined.

Page 57: South Africa’s Trade Flows -A Gravity Model Analysis

46

Page 58: South Africa’s Trade Flows -A Gravity Model Analysis

47

CHAPTER IV

4. THE DETERMINANTS OF TRADE AND TRADE PATTERNS BETWEEN

SOUTH AFRICA AND PARTNERS

4.1. Introduction

It is argued that countries open up to free trade in order to exchange goods they have

in surplus for those that are scarce. More importantly, trade can enhance economic

growth by exposing domestic production to foreign technology. This gives domestic

producers an opportunity to learn through imitation of foreign technologies, thus

accelerating the transmission of technology (Grossman and Helpman, 1991) by means

of importing hi-tech commodities, as witnessed in East Asian Countries (Alesina et al,

2005).

However, trade can create winners and sometimes losers (Viner 1950). Nevertheless,

Baldwin (2004) argues that generally, open economies as opposed to autarkists

experience more rapid economic growth. As a prescription therefore, countries with

an objective of reducing poverty could benefit more from dismantling trade barriers

against goods and services as wide market access tends to attract FDI.

Linking free trade agreements to economic progress is not a straightforward approach.

The study argues that the relationship between the two is an indirect one. It can be

understood by measuring the incremental effects of the FTA in question on members’

Page 59: South Africa’s Trade Flows -A Gravity Model Analysis

48

exports. Since a wide body of literature has shown a causal relationship between

exports and GDP (Dhawan and Biswal, 1999; Awokuse (2003); Jordaan and

Hinaunye, (2007), accordingly, this study argues that the effect that an FTA has on

exports reflects its indirect effect on GDP.

Figure 4.1: South Africa’s GDP, 1997-2010

Source: calculations based on World Bank data (World Bank, 2012)

In Figure 4.1 (panel A). shows that South Africa’s GDP and exports have been

moving in the same direction since 1997. In the aftermath of the signing of 3

important free trade agreements, South Africa realized a significant increase in the

volume of exports between 2001 and 2004 whose value is estimated to be around

US$38 billion per year compared to the average of $33.5 billion realized in the

previous periods. Beyond 2004, the average value of exports was US$45 billion.

More still, between 2001-2004, average GDP was US$38.2 billion an increase of over

US$16.3 billion from its pre-2000 values.

Page 60: South Africa’s Trade Flows -A Gravity Model Analysis

49

By 2010, South Africa’s average GDP was around US$45 billion. On the other hand,

although Figure 4.1 (panel B). shows that the average growth rates of GDP and

exports are higher in the periods after the signing of AGOA, TDCA and SADC, it

also reveals that the exports growth rate had started falling steadily even before the

2007 financial crisis; that is from 8.5% in 2005, 5.9% in 2007 to -19% in 2009. Apart

from the three years (i.e. 2005-2007), South Africa’s export growth rates do not seem

to be any different from those that existed before 2000.

Some of the main factors that may hinder South Africa’s export growth could be the

degree of openness of the economy and its trade partners, the level of intra industry

trade, comparative advantage over other trade partners in the production of major

commodities, as well as the marketable characteristics of South African commodities

relative to those from its trade partners. But most notably, gains from international

trade depends on how open countries are.

4.1.1. Trade openness

A higher degree of openness ensures that countries can trade more since the level of

barriers to trade is very low. As countries slash down import duties levied on other

countries’ exports, the trade openness index for the import-duty-reducing country

increases. The main advantage of such a move is that it encourages reciprocity from

beneficiaries.

Page 61: South Africa’s Trade Flows -A Gravity Model Analysis

50

Figure 4.2: Trade Openness index for South Africa, EU, BRIC, SADC, AGOA, 2001-2010

Source: calculations based on World Bank Development Indicators (World Bank, 2012)

Even though the relationship between trade openness and economic growth is widely

contested in growth literature, there is near consensus about the existence of a positive

correlation between trade flows and growth (Yannikaya 2002). For a trade

arrangement to be profitable for member countries, exports and imports need to be of

significant importance to the respective economies’ GDP. In other words, the benefits

from trade could be reduced if member countries are overly self-reliant (i.e. lower

levels of openness). Therefore the main strength of the trade openness index used

above is that it takes into account the trade flows and GDP (the best known proxy for

economic growth).

Page 62: South Africa’s Trade Flows -A Gravity Model Analysis

51

Figure 4.2 shows that among the four trade arrangements considered, the SADC and

the EU are the most open, followed by BRIC. AGOA is the most closed trade

arrangement of all. BRIC has made the most significant improvement in opening up

to trade than any other. In 2001, the trade openness index for the BRIC economies

was 39.3%. However, with the formation of BRIC in 2010, the openness index was

67%, a 70% increase from its 2001 level.

It can therefore be argued that South Africa is more likely to trade more with SADC,

BRIC and the EU as opposed to trading with the United States. Nonetheless, despite

conforming to free trading, South Africa’s economy still remains more closed relative

to SADC, EU and BRIC. This is a potential limiting factor against the gains that

South Africa could reap from trading with a more open approach.

4.1.2. Intra Industry Trade

The level of intra industry trade represents the extent to which countries are able to

trade with each other in products that lie in the same industry. A higher intra industry

trade index value means that countries are able to exchange similar goods that are

differentiated by branding and inclusion of different additional attributes. Product

differences are attributed to production technology differences across countries in a

way that products from each country are quite different. Such trade is more common

in vehicles, wine, electronics, cosmetics and clothing etc. For example, the success of

South Africa’s products in any of the FTAs mentioned partly depends on the ability of

Page 63: South Africa’s Trade Flows -A Gravity Model Analysis

52

South African exporters to differentiate their commodities from competing substitutes

in the same trade zone.

The table 4.1 presents the top South Africa’s import sources and export destinations in

the European Union for selected commodities. In both exports and imports, nine top

destinations and suppliers were considered. The data revealed that apart from trade in

minerals, South Africa is worse off in all other industries considered by the study.

There are signs of significantly high intra industry trade in the vehicle industry and

considerably low trade in automatic data processors and minerals.

In 2010, the intra industry trade index in the vehicle industry rose to 79%, its highest

point in 3 years, and then fell to 70% in 2011; a percentage that is still higher than the

2009 index.

Trade in electronics is weak with the rate at which South Africa and the EU demand

each other’s electronic equipment declining from 27% in 2009 to 23% in 2011. This

pattern favors the EU because intra industry trade data show that the volume of trade

in this industry has been progressively increasing, whereas the share of South Africa’s

exports in total trade has been progressively declining. It is tricky to reach a

conclusion on such a development because even though the share of South Africa’s

exports in this industry has been falling, South Africa’s total exports in the same

industry have been increasing. The pattern is that both the EU’s and South Africa’s

exports have been increasing but EU’s exports have been increasing at a higher rate

Page 64: South Africa’s Trade Flows -A Gravity Model Analysis

53

than South Africa’s. As long as South Africa’s exports do not suffer in the process,

the TDCA is not a bad idea.

A completely different pattern emerges when trade in automatic data processors and

minerals is considered. The European Union highly dominates the data processor

industry. Trade is mainly a one way traffic, with the EU supplying and South Africa

importing. Unlike trade in electronics where the EU dominates but both parties’

exports are increasing, in data processors, the intra industry trade index has been

significantly low (9% in 2009) and falling (to 7% in 2011); worse still, South Africa’s

exports in this industry have fallen from US$16.2 million in 2009 to US$11.4 million

in 2011. The share of South Africa’s data processor exports in total intra industry

exchange fell from 5% in 2009 to 3.4% in 2011. This shows that South African

export-oriented and import-competing companies are struggling against European

imports. If such trends persist in the long term, continued free trade may see South

Africa’s data processor exports wiped out of the EU market.

On the other hand, the situation is the extreme opposite when trade in minerals is

considered. The index of intra industry trade returns very low rates. In 2009, the IIT

index was 8%. It fell to 5% in 2010 and rose to 11% in 2011. South Africa, in this

case, is the main supplier with the EU providing less competition as an exporter and

providing more market for South Africa’s exports. On average, South Africa has been

providing 95% of all minerals traded with the EU. The value of South Africa’s

Page 65: South Africa’s Trade Flows -A Gravity Model Analysis

54

mineral exports increased progressively from US$2.018 billion in 2009 to over

US$3.493 billion in 2011, an increase of over US$1.475 billion in just two years.

Page 66: South Africa’s Trade Flows -A Gravity Model Analysis

55

Table 4. 1: Intra Industry Trade between South Africa and EU (2009-2011)

Vehicles other than railw

ay

Electrical, Electronic equipm

ent

Automatic data

Processing machines

and units thereof

Pearls, Precious stones, M

etals, Coins

Year

Country

Exports

Country

Imports

Intra industry trade index

Country

Exports

Country

Imports

Intra industry trade index

Country

Exports

Country

Imports

Intra industry trade index

Country

Exports

Country

Imports

Intra industry trade index

Ger 620.3 Ger 1403.9 Fr 118.3 Ger 705.1 Spain 5.7 Czech 76.6 UK 1352.1 Ger 36.1

Pol 106.5 UK 257.3 Ger 63.1 UK 240.5 Neth 2.5 UK 54.1 Ger 345.5 Bel 26.4

Fr 98.8 Spain 127.3 UK 30.4 Fr 195.6 UK 2.2 Ire 49.7 Bel 298.0 It 9.3

Bel 58.5 It 121.0 Neth 24.5 Swed 169.7 Ger 1.8 Ger 40.3 It 7.7 Ire 5.7

UK 58.0 Fr 118.4 Bel 20.2 It 160.8 Cyp 1.5 Hung 33.4 Ire 6.9 UK 5.4

Spain 56.1 Bel 49.1 Spain 14.2 Fin 107.1 Fr 1.0 Neth 31.1 Neth 5.2 Fr 0.7

It 26.1 Pol 48.1 Swed 8.4 Bel 106.7 Rom 0.9 Fr 24.3 Swed 1.2 Den 0.4

Port 22.8 Aust 30.6 Ltvia 8.1 Neth 104.1 Ire 0.4 Pol 20.0 Spain 0.9 Spain 0.4

Fin 16.9 S.vakia 25.7 Greece 6.8 Hung 102.6 Bel 0.3 It 2.9 Aust 0.9 Neth 0.1

Total 1064.1 2181.4 0.66 294.0 1892.3 0.27 16.2 332.2 0.09 2018.5 84.6 0.08

Ger 1314.7 Ger 1703.8 Fr 139.6 Ger 696.4 Fr 4.3 Czech 111.6 Aust 1853.6 Ger 29.3

Fr 143.2 UK 511.7 Ger 61.4 Hung 363.4 UK 4.2 Ire 90.7 Bel 618.9 Bel 19.2

UK 129.3 Spain 211.0 Bel 36.5 UK 218.8 Neth 2.8 UK 61.7 Bulg 449.2 It 13.0

Bel 121.5 Fr 160.3 UK 33.6 Fr 173.3 Cyp 1.4 Ger 40.3 Cyp 36.5 Ire 9.9

Spain 76.1 It 148.2 Neth 25.5 It 156.2 Ger 1.2 Hung 30.5 Czech 20.5 UK 2.7

Pol 60.6 Bel 110.7 Swed 8.0 Fin 142.1 Swed 0.9 Neth 29.4 Den 2.8 Fr 1.0

Swed 42.8 Swed 73.5 Greece 2.5 Aust 122.7 Czech 0.3 Fr 29.3 Eston 2.6 Spain 0.7

Fin 32.9 Pol 60.3 Spain 2.4 Swed 118.2 Den 0.2 Pol 19.7 Fin 1.8 Den 0.6

Hung 31.8 S.vakia 33.5 Pol 2.4 Bel 97.4 Pol 0.2 Swed 2.5 Fr 0.4 Swed 0.5

Total 1952.9 3013.0 0.79 312.0 2088.5 0.26 15.6 415.7 0.07 2986.3 77.0 0.05

Ger 1280.6 Ger 2136.4 Fr 66.7 Ger 694.5 Fr 4.1 Czech 113.5 UK 1903.0 UK 87.5

Fr 164.7 UK 678.2 Ger 64.5 Hung 503.7 UK 2.7 Fr 40.8 Ger 809.0 Bel 53.2

UK 150.7 Spain 213.8 Bel 58.9 UK 219.4 Ger 2.0 Ger 39.7 Bel 607.3 Ger 42.7

Bel 146.5 Fr 204.0 Neth 44.2 Ire 171.9 Neth 1.5 UK 39.5 It 137.1 It 15.3

Spain 90.7 It 187.0 UK 34.5 Rom 171.7 Bel 0.4 Hung 28.1 Ire 32.0 Ire 9.5

Pol 61.4 Swed 149.1 Swed 30.4 Fr 170.7 Cyp 0.2 Neth 26.2 Swed 2.3 Fr 1.0

Hung 54.0 Bel 111.8 Hung 5.8 It 165.0 Lux 0.2 Ire 15.7 Fr 1.0 Spain 0.5

Swed 44.2 Pol 77.0 It 3.2 Fin 159.6 Ire 0.1 Pol 12.5 Den 0.8 Aust 0.4

Ire 40.3 Czech 51.6 Lith 3.1 Swed 136.5 It 0.1 Swed 7.9 Neth 0.7 Den 0.4

Total 2033.0 3808.9 0.70 311.2 2393.2 0.23 11.4 323.9 0.07 3493.3 210.5 0.11

2009

2010

2011

Source: Calculations based on the United Nations Commodity Trade Statistics Database (UN

COMTRADE, 2012)

Page 67: South Africa’s Trade Flows -A Gravity Model Analysis

56

4.1.3. Differences in product characteristics

South Africa can arguably benefit more from trading freely with SADC, EU, BRIC

and AGOA if its products have recognizable differences from those of other

countries. Administratively, however, it is an insurmountable task to clearly

determine those differences as this would involve analyzing thousands of

commodities. The United Nations statistical division, through the Standard

International Trade Classification (SITC) categorizes all possible products into nine

major categories, from which this study will compare the nature of products exported

by South Africa against those from other countries. This enables the identification of

products in high demand elsewhere which South Africa can supply.

The tables below show the nine product categories as classified by the UN and the

share they account for in each trade arrangement’s exports in 2009 and 2010.

As stated earlier, the concern should be whether all countries export similar products.

Even though the possibility of intra industry trade reduces such concerns, the extent to

which countries can mutually benefit from trading freely is limited by high

substitutability of traded commodities. Secondly, by exporting similar products,

developed countries are more likely to benefit from intra industry trade than

developing ones since they have more means and better know-how for engineering

characteristic differences in products.

Page 68: South Africa’s Trade Flows -A Gravity Model Analysis

57

Figure 4.4 compares South African exports against products imported by SADC, EU,

BRIC and AGOA. The beneficial pattern for South Africa would be one where each

major export category for South Africa has high demand from SADC, EU, BRIC and

AGOA.

Figure 4. 3: Composition of exports from SADC, EU, BRIC and AGOA, 2009-2010.

Source: Calculations based on the United Nations Commodity Trade Statistics Database (UN

COMTRADE, 2012)

Figure 4.3 shows that manufactured goods, machinery and transport equipment, crude

materials and minerals make up the biggest share of South African exports. Together

they accounted for over 80% of South Africa’s total exports between 2009 and 2010.

The same products are highly exported by the EU and BRIC. On the other hand,

machinery and transport equipment are the biggest composition of the USA whereas

manufactured goods dominate SADC’s exports.

Page 69: South Africa’s Trade Flows -A Gravity Model Analysis

58

Figure 4. 4: potential markets for South African exports, 2009-2010.

Source: Calculations based on the United Nations Commodity Trade Statistics Database (UN

COMTRADE, 2012)

The graph 4.3 shows that for each commodity category that South Africa produces in

plenty, there are at least two other trade blocs that export almost as much. Yet it

appears that all countries export insignificant amounts in commodities such as animal

and vegetable oils, and beverages and tobacco, and other products that are not

mentioned anywhere in the above categories. This would make someone believe that

any country would benefit from exporting more of those commodities. On the other

hand, a close examination of Figure 3.4 reveals that actually these commodities make

up an insignificant share of total imports in any trade arrangement considered in this

study.

Page 70: South Africa’s Trade Flows -A Gravity Model Analysis

59

Unexpectedly, the commodities that have been shown in Figure 4.3 to be the main

exports for most of the countries are still the biggest composition of imports in the

same countries. For example, manufactured goods were 15.7% of BRIC total exports

in 2009-2010 and in the same period, the same products accounted for 15.1% of

BRIC’s total imports. Machinery and transport equipment were 35.6% of AGOA total

exports and in the same period, USA imported the same products amounting to 36.4%

of its total imports.

Figure 4.4 reveals that:

South Africa’s exports of manufactured goods are high enough (34.1%), given

that, manufactured goods make a modest composition of all country groups’

imports (making 11.31% of BRIC total imports; 10% for AGOA; 17.58% for

SADC and 12.4% for the EU).

South Africa can gain more by increasing its machinery and transport

equipment exports, given a huge market that is available in all trade blocs

considered. Machinery imports accounted for the lion’s share in imports of

BRIC (36.6%); USA (36.46%); EU (30.09%); and SADC (27.42%) in 2009

and 2010, yet they only amounted to 19.43% of South Africa’s exports in the

same period.

Chemicals, mineral fuels and lubricants and miscellaneous manufactured

articles are major imports in the EU, SADC, AGOA and BRIC but whose

exports share is still very low in South Africa.

Page 71: South Africa’s Trade Flows -A Gravity Model Analysis

60

Therefore, trade with the EU, SADC, BRIC and AGOA seems to be centered on trade

in related products. All countries’ exports are quite similar, yet, there is also evidence

that countries tend to import goods that are similar to those that they export.

For South Africa to benefit from its free trade agreements, resources must be directed

in a way that capitalizes on the market opportunities provided by the EU, SADC,

AGOA and BRIC. The only way to do that is by diversifying and or increasing

exports of commodities that fall under the following product categories;

Chemicals and related products

Machinery and transport equipment

Manufactured goods classified chiefly by material

Mineral fuels, lubricants and related materials

Production of crude materials should be aimed at targeting markets provided

especially by BRIC and SADC, whereas miscellaneous manufactured articles should

be tailored to suit mainly USA and EU demand.

4.2. South Africa’s trade patterns with EU, SADC, BRIC and AGOA

One of the main concerns of trade between countries that are at different levels of

development is the pattern of trade that is likely to arise. Some countries fear that

trading with more efficient partners may generate “uncompensated losses to import

competing firms” (Suranovic 1997). Even though this may be true, most of the time

Page 72: South Africa’s Trade Flows -A Gravity Model Analysis

61

when countries trade with one another compensation schemes somehow appear in the

sense that superior countries export more of the sophisticated goods whereas the

weaker partners tend to export more of the products that require less technology and

capital. Besides, Suranovic (1997) argues that uncompensated losses may also appear

under protectionism in form of “…higher consumer prices and lost opportunities to

some individuals in the economy”.

Table 4.2: Trade in primary products between South Africa and Malawi, Zambia, Zimbabwe,

2009-2010

South Africa imports from South Africa Exports toYear 2009

live trees and other plantsMalawi 66% 34%Zimbabwe 86% 14%Zambia 61% 39%Year 2010

live trees and other plantsMalawi 53% 47%Zimbabwe 87% 13%Zambia 47% 53%

oil seeds and oleaginous fruitsMalawi 60% 40%Zimbabwe 54% 46%Zambia 71% 29%

Source: Calculations based on the United Nations Commodity Trade Statistics Database (UN COMTRADE,

2012).

The percentage share of South Africa’s imports/exports in a particular item in tables 4.2 and 4.3 is calculated as: %share of imports/exports of commodity X = imports/exports of commodity X from country Y/Exports + imports of commodity X between Y and South Africa.

Page 73: South Africa’s Trade Flows -A Gravity Model Analysis

62

Table 4. 3: Trade in primary products between South Africa and Malawi, Zambia, Zimbabwe, Mozambique, 2009-2010.

South Africa imports from

South Africa Exports to

Year 2009Iron and SteelMalawi 0.28% 99.72%Mozambique 1.68% 98.32%Zimbabwe 10.77% 89.23%Zambia 5.11% 94.89%Electrical machinery equipment and other parts thereofMalawi 1.88% 98.12%Mozambique 2.00% 98.00%Zimbabwe 3.40% 96.60%Zambia 22.07% 77.93%Vehicles other than railwayMalawi 0.25% 99.75%Mozambique 0.92% 99.08%Zimbabwe 6.69% 93.31%Zambia 0.68% 99.32%Year 2010Iron and SteelMalawi 0.34% 99.66%Mozambique 1.01% 98.99%Zimbabwe 11.31% 88.69%Zambia 6.39% 93.61%Electrical machinery equipment and other parts thereofMalawi 0.80% 99.20%Mozambique 0.22% 99.78%Zimbabwe 2.75% 97.25%Zambia 22.94% 77.06%Vehicles other than railwayMalawi 0.25% 99.75%Mozambique 0.21% 99.79%Zimbabwe 0.47% 99.53%Zambia 0.65% 99.35%

Source: Calculations based on the United Nations Commodity Trade Statistics Database (UN

COMTRADE, 2012)

Page 74: South Africa’s Trade Flows -A Gravity Model Analysis

63

The tables (4.2 and 4.3) above present trade flows of specific commodities between

South Africa and three selected SADC countries. Table 4.2 presents trade in

agricultural products. The table shows that South Africa does poorly when it comes to

trade in agricultural products with Malawi, Zambia or Zimbabwe. In trade in live trees

and other plants, South Africa accounts for only 34% of total exports with Malawi,

14% of total exports with Zimbabwe and 39% of total exports with Zambia in 2009.

The trade pattern is similar when trade in oil seeds and oleaginous fruits is considered

in both 2009 and 2010.

On the other hand, South Africa’s shares in the export of products that require

elevated production technology far outweigh any of its selected SADC partners. In

trade of either iron and steel, electrical machinery equipment or vehicles, South Africa

accounted for over 96% of all total exports in 2009 and 2010.

From Table 4.3 , it can be concluded that although South Africa faces unfavorable

terms of trade in the selected agricultural products in the SADC, it compensates for

those trade deficits by registering trade surpluses in the industrial sector.

On the other hand South, Africa’s trade pattern with some of its top trade partners in

the EU and US is different from that of SADC countries considered above. While it

has been shown that South Africa exports more capital intensive commodities to the

comparatively smaller SADC countries, here, South Africa is more dominant in the

Page 75: South Africa’s Trade Flows -A Gravity Model Analysis

64

export of agricultural products. For example in trade of trees and other plants, South

Africa accounted for 97% of total trade with Germany, 99.87% of total trade with the

UK and 94% of total flows with the USA in 2009. Conversely, the share of South

Africa’s exports in products that necessitate elevated production technology declines

considerably when contrasted against trade in the same products with the selected

SADC countries. For example, South Africa’s exports of electrical machinery and

equipment made only 8% of trade in that particular sector with Germany, 11% in the

same sector with the UK, and only 9% with the US in 2009. The same trade pattern

existed in trade of vehicles other than railway in 2009 and 2010.

Table 4. 4: Direction of Trade in both primary and secondary products between South Africa and Germany, UK, US, 2009-2010

South Africa imports from South Africa Exports toYear 2009live trees and other plantsGermany 3% 97%UK 0.3% 99.7%US 6% 94%oil seeds and oleaginous fruitsGermany 9% 91%UK 11% 89%US 61% 39%Electrical Machinery and equipment and parts thereofGermany 92% 8%UK 89% 11%US 91% 9%Vehicles other than railwayGermany 69% 31%UK 82% 18%US 20% 80%Year 2010

Page 76: South Africa’s Trade Flows -A Gravity Model Analysis

65

live trees and other plantsGermany 3% 97%UK 0.4% 99.6%US 7% 93%oil seeds and oleaginous fruitsGermany 12% 88%UK 15% 85%US 55% 45%Electrical Machinery and equipment and parts thereofGermany 92% 8%UK 87% 13%US 88% 12%Vehicles other than railwayGermany 56% 44%UK 80% 20%US 24% 76%

Source: Calculations based on the United Nations Commodity Trade Statistics Database (UN

COMTRADE, 2012)

It can be shown from Table 4.4 that the USA exported more oil seeds to South Africa

than what the latter supplied to the former. Interestingly, in the trade of vehicles, the

value of South African vehicles exported to the US (76% of total vehicle exports

between South Africa and USA) far exceeded the value of vehicles exported by USA

to South Africa. On the other hand, when the top seven exports from one of the two

countries to another are considered, it is clear that South Africa only has a competitive

advantage over the US in the export of low-tech commodities (e.g. primary products,

mining metals and minerals) as shown below.

Page 77: South Africa’s Trade Flows -A Gravity Model Analysis

66

Table 4. 5: Trade in top seven exports between South Africa and USA 2010

Year 2010 US top imports from South Africa (millions of dollars)

US top exports to South Africa (millions of dollars)

Platinum sponge 767.7 small automobiles 10.4Medium-size automobiles

699.2Civilian aircraft and parts

94.1

Medium-size automobiles, smaller engine

618.9Calcined petroleum coke 17.3

Non-Industrial diamonds, Processed

533.2Road tractors for semi-trailers

1.6

Rhodium345.3

vehicle parts and accessories

6.2

Palladium166

Tunneling Machinery parts including rock cutters

9

Unsaturated acyclic hydrocarbons 117.2

reception and transmission equipment

8.2

Source: Workman (2010)

The trade structure as discussed above reveals that South Africa neither dominates nor

is dominated in its trade with the EU, SADC and AGOA. Trade Patterns show that

international trade has automatically generated compensation schemes that allow each

trade partner to dominate in one industry and be dominated in the other. For example,

South Africa is better in industrial exports in its trade with the SADC whereas the

SADC does well in the agricultural exports. In its trade with the EU and AGOA, the

roles are reversed, whereby South Africa has been seen to dominate in agricultural

exports and its partners dominate in the industrial sector.

Page 78: South Africa’s Trade Flows -A Gravity Model Analysis

67

The implication of such trade structure is that more often, the country that is more

competitive in the industrial sector reaps more from bilateral relations than economies

whose exports are mainly agricultural products. Free trade in this case may lead to

deterioration in terms of trade for countries that are not competitive in the industrial

sector.

The figure below shows the apportionment of South Africa’s trade flows among

SADC, AGOA, BRIC and the EU. In 2010, the EU was South Africa’s top trade

partner with total trade flows worth US$44.4 billion. This amounted to over 45.5% of

South Africa’s total trade with the EU, SADC, BRIC and AGOA. Despite being

signed in 2010, BRIC importance in South Africa’s trade far surpasses that of the

United States with the latter accounting for 28.6% and the former only 13.3% of

South Africa’s total trade flows with the EU, SADC, AGOA and BRIC. SADC’s

share was the lowest with only goods worth around 12 billion dollars exchanged

between SADC member countries and South Africa. This is partly due to a

comparatively lower purchasing power of SADC countries, given their low gross

domestic product as opposed to the EU, AGOA and BRIC.

Page 79: South Africa’s Trade Flows -A Gravity Model Analysis

68

Figure 4. 5: Import sources and export destinations for South Africa in SADC, EU, BRIC, and AGOA

Source: Calculations based on the United Nations Commodity Trade Statistics Database (UN COMTRADE, 2012)

The above trade agreements give South Africa’s producers free access to a combined

market of over US$42 trillion which is more than 35 times bigger than that provided

by the whole of Africa. It is important to note that among the four PTAs, South Africa

faces unfavorable terms of trade with the EU and BRIC. The trade deficit with the EU

amounts to over US$7.1 billion, which is around 15.9% of total trade with the EU,

Page 80: South Africa’s Trade Flows -A Gravity Model Analysis

69

whereas the trade deficit with BRIC is over US$ 3 billion, that is over 13% of total

bilateral trade flows.

On the other hand, South Africa earns a trade surplus against both the USA and

AGOA. Its trade surplus with the United States amounts to US$1.26 billion. As

expected, South Africa faces the most favorable terms of trade from its trade with

SADC. Its trade surplus is over US$5.31 billion, which is over 43.4% of total bilateral

flows with SADC. In liquid terms therefore, it is very clear that trade with SADC is

more important to South than any other.

Since trade with the EU and BRIC enables South Africa to import goods that it

doesn’t have the capacity to produce, a current account deficit in this case may not be

totally detrimental. On the contrary, free trade allows South Africans to access such

goods at comparatively lower prices. More still, the importation of capital goods may

also imply that the terms of trade are more likely than not to correct themselves in the

long run.

Page 81: South Africa’s Trade Flows -A Gravity Model Analysis

70

4.3. Conclusion

This discussion has revealed several concerns regarding the fact that the products

traded between South Africa and its partners are quite similar, thus increasing the

pressure of competition. The major concern as demonstrated is that South Africa’s

exports are still very low in areas that promise potential markets. This raises

speculation about whether South Africa’s move for trade liberalization was not a right

move at a wrong time. Rodrik (2001B) argues that, liberalization should be a gradual

process, where countries liberalize in only those areas against whose imports they can

compete effectively and in which they have built enough supply capacity.

However, understanding the over-all impact of free trade agreements warrants more

than just graphical analysis. Applying analytical econometric techniques can reveal

more information and important details that visual observation of charts derived from

raw data cannot. Besides, the product categories used in this chapter have been either

very narrow or too broad; they have included too few or too many product groups,

which when considered exclusively may lead to different conclusions. In an attempt

to address this observation, the next chapter discusses how a gravity model can be

used to overcome such limitations.

Page 82: South Africa’s Trade Flows -A Gravity Model Analysis

71

CHAPTER V

THE GRAVITY MODEL

5.1. Introduction

According to Tiiu (2000) the classical models of perfect specialization (reviewed in

chapter 2) are limiting in a sense that they only explain trade in specific items but

cannot explain why countries have stronger trade links with some countries than

others. Tiiu further argues that since they ignore the possibility that endowments may

change and be transferred over time, they fail to explain why trade is created and or

increases over time for some and fail for others.

Unlike the more theoretical models by Adam Smith and David Ricardo, the gravity

equation which is a monopolistic competition model (Feenstra et al 2001) that allows

a more empirical approach of explaining trade patterns through econometric analysis.

Its application to the prediction of bilateral trade flows was pioneered by Tinbergen in

1962. Since then, the model has gained popularity due to its high explanatory power

of international trade flows. The gravity equation is not limited to analyzing trade

between countries only; it has also been used to analyze trade between regions

(Filippini and Molini 2003, Breuss and Egger 1999), as well as flows of specific

products (Kangas and Niskanen 2003, Pelletiere and Reinert 2003, Jayasinghe and

Sarker 2007) among others. Recent studies have extended gravity analysis to account

for factors that affect international trade for example rules of origin (Augier et al

Page 83: South Africa’s Trade Flows -A Gravity Model Analysis

72

2004), border effects (Nitsch 2000), FDI (Gopinath and Echevenia 2004), rights and

democracy effects (Sama and kucera (2006) to mention but a few.

The model predicts a baseline for trade flows as being determined by the size of GDP,

population size, and distance between countries. According to the model, interaction

among large clusters is likely to be stronger than between smaller ones. More still, a

unit is more likely to interact with neighboring units than with those located far away.

Simply put, the model can be easily explained as implying that the level of trade

between countries depends on the exporting country’s capacity to supply (GDP), and

on the Market (GDP) available in the importing country and the distance (transport

costs) between them (Bergeijk and Brakman, 2010).

In the real world, there is likely to be large trade flows among big economies than

among smaller ones. Also, neighboring countries are likely to trade with each other

more than they trade with countries that are far away keeping other factors affecting

trade flows unchanged. For example, the level of interaction between South Africa

and Zimbabwe should be higher than the interaction between South Africa and France

assuming that France and Zimbabwe are of the same economic size. Therefore, the

model adds distance as a significant factor in determining the level of trade flows.

The popularity of the model was dwarfed for a long time because of its weak and

ambiguous micro-economic foundation. This ambiguity is partly because the model

Page 84: South Africa’s Trade Flows -A Gravity Model Analysis

73

can be derived from many different international trade models (Helpman and

Krugman 1985, Evenett and keller 2002) “thus offering no scope to test between

theories” (Bergeijk and Brakman, 2010). On the other hand, Bergeijk and Brakman

argue that this ambiguity should give policy makers confidence in the robustness of

the model since it doesn’t depend on the vision of a specific theory. Thus, the

possibility of deriving the gravity equation from many models of international trade

provides it with rather than limit “more theoretical foundation than any other model”

(Baldwin 2006).

Although initially the model lacked sound theoretical justifications, over time it has

been enriched with better theoretical underpinnings and estimation techniques. There

are many empirical applications of the gravity model that have contributed the

improvement of its specification and popularity. Among them include, Helpman and

Krugman, (1985); Feenstra, (2002); Breuss and Egger, (1999); Wei (1996);

Bergstrand, (1985); and Anderson (1979).

The technique has been applied in a number of fields especially for ex-post analysis of

migration flows, FDI flows, Currency unions, WTO membership, and RTA

membership among others. Thus gravity equation has been used more frequently with

quite a remarkable success in the prediction of international trade flows and the

evaluation of the success of different trade arrangements. This is mainly because, the

Page 85: South Africa’s Trade Flows -A Gravity Model Analysis

74

model can be used account for factors that are regularly ignored by the traditional

models yet highly important in explaining trade patterns among countries.

5.1.1. Limitations of the gravity model

The gravity model has some important limitations notable among which include;

The baseline gravity model is likely to produce biased results because it

ignores some variables that are important in the prediction of trade flows. For

example distance alone is not enough to explain limitations to trade.

The dummy variables used may falsely attribute increased trade to an FTA

because they tend to be correlated with other variables such as diplomatic

efforts, technological diffusion, security that may be responsible for the

growth in regional trade among others.

5.2. The theoretical gravity equation

We use a gravity model to explain bilateral trade flows (Xij) in terms of incomes of

the importing and exporting countries, and the distance between them. The basic

gravity model explaining the relationship between trade flows, distance and economic

size is, in the general form expressed as follows:

Trade = Xij = GDPi, GDPj, Dij (5.1)

Page 86: South Africa’s Trade Flows -A Gravity Model Analysis

75

In other words trade flows between two countries (Xij) are expressed as a function of

the characteristics of the exporting country (GDPi) and the importing country (GDPj)

and the degree of limitations/distance between them (Dij).

Each country produces a differentiated product and every country demands other

country’s goods because of the principle of comparative advantage (Feenstra and

Taylor 2008). These differences can be a result of differences in technology, natural

endowments, tastes and preferences, among others.

However, it is not completely true that trade depends only on the factors put forward

by the gravity model; it also depends on factors that a researcher deems to be relevant.

In order to account for these other factors, the gravity model is re-written as:

Xij = GDPi, GDPj, Dij, Vt (5.2)

Where Vt is a vector representing trade determinants that are ignored by the gravity

model. Introducing logs on both sides of equation 5.2 above transforms it into a

linear function.

LnXij = βLnK + U, U~N (0,σ2) (5.3)

Where K = The Vector of independent variables

β = The Vector of Parameters to be estimated

U = White noise error term.

Page 87: South Africa’s Trade Flows -A Gravity Model Analysis

76

The main purpose of logs is to reduce effects of inflationary tendencies and generally

make data as closely comparable as possible.

5.3. Methodology and Data

Since it is well documented that preferential trading may negatively or positively

affect trade, Laszlo (1997) and Anderson and Wincoop (2003) demonstrate how

multilateral resistance terms (longitudinal and binary) can be added to the model to

remove much of the bias that may result from the omitted variables. The selection

process for variables to be included in the gravity model therefore follows the

theoretical work of Anderson and Wincoop (2003). According to them, gravity

models that include distance as the only cost to trade; are often more likely to produce

biased results and most importantly, the results cannot be used for comparative statics

exercises. They show that to correctly specify the gravity equation, the researcher has

to account for other trade costs that are not specific to only the bilateral barriers

between the countries under consideration. These costs are known as multilateral

resistances. The intuition is that trade between a given set of countries depends on

their trade costs with all trade partners in the ‘rest of the world’ (Deardorffs, 2012);

this way, trade will be higher for those countries that have relatively low trade

barriers.

As Anderson and Wincoop (2003) demonstrate, excluding such variables often results

in overestimation of country specific coefficients, and thus overstating the overall

Page 88: South Africa’s Trade Flows -A Gravity Model Analysis

77

impact of trade arrangements on trade flows. The correct specification of the gravity

model including multilateral resistance terms is:

LnXij = αi + πj + λt + β2LnGDPjt+ β4LnPj + β3LnDij +…+ uijt (5.4)

Where

Pj = Population of country j

Xij = trade flows between country i and j.

GDPJt = GDP of country i at time t.

i = 1,…,i-1, i+1,…N+1. Where N+1 is the rest of the world (ROW).

πj = Country effect, J = 1,…,N+1.

λt = time effect, t=1, …,t.

Uijt = White noise disturbance term.

The relationship among variables is direct; the larger GDPj is, the larger Xij. There is

an inverse relationship between Dij and Xij. In other words, the original gravity

model assumes that trade flows between two countries depend positively on the levels

of output of the countries in concerned and negatively correlate with the distance

between countries.

On this relationship, studies by Baier and Bergstrand (2004), Egger and Larch (2008)

provide evidence of the existence of such a relationship between GDP and the level of

trade flows. They find that the effects of trade creation are greater in countries that

Page 89: South Africa’s Trade Flows -A Gravity Model Analysis

78

have larger GDP and even greater in those whose GDP is high and comparatively

similar in size.

The gravity model is outstanding in a way that it doesn’t ignore the limitations to

trade. It uses distance between countries (Dij) as one of the costs of trade and its

effect on trade flows is negative. The longer the distance between countries the lesser

trade there is among them. Distance has been a vital variable among the determinants

of trade ever since the inspiring works of Rinbergen 1962, Anderson 1979, and

Krugman 1997 (cited by Nuno, 2010).

The specification of trade costs is empirically not straightforward because they are

numerous and sometimes consist of subcomponents that are very difficult to quantify.

For example, many studies use actual distance between countries, while others may

use actual data on shipping costs. Even though the use of actual shipping costs data

sounds more convincing in a way that it may sometimes be cheaper to transport goods

from South Africa to USA; than from South Africa to Liberia (depending on the

quality of infrastructure), the method is very difficult to apply. This is because

transport costs vary from one commodity to another and from one Transport

Company to the next. Also data on such costs are very rare on the international level

if not absent at all.

Page 90: South Africa’s Trade Flows -A Gravity Model Analysis

79

Another way to measure costs is by using the exchange rate (EXCHij) as a proxy for

changes in prices between countries. The exchange rate is added to the gravity model

as an impediment to trade, since volatility in its rates affect trade flows (Egger 2002).

It is also one of the reasons why differences in relative prices exist between countries.

The study also adopts population of trade partners (Pj) as one of the variables that can

determine the level of trade flows between countries. The relationship between

population and trade flows can be explained in two different ways. In the first case,

the coefficient is expected to have an inverse relationship with trade flows because

larger populations mean self-sufficiency. This means that the country can effectively

absorb most of its produce, thus reducing the amount of surplus left for export.

However Bergstrand, (1985) notes that this relationship is not always true because

larger populations allow the accumulation of economies of scale which in turn boost

the export sector. This study maintains Bergstrand’s position.

5.3.1. Selection of multilateral resistance variables and hypothesis

As discussed earlier, costs to trade are an important ingredient in international trade in

that they help to measure or predict the ease of access of one country to markets in

other countries. Therefore this study finds it is erroneous to predict international trade

flows without taking into account the potential trade costs. This is one of the major

weaknesses of the traditional theories of trade.

Page 91: South Africa’s Trade Flows -A Gravity Model Analysis

80

Distance is a major proxy of trade costs and it enters the model in various ways. The

traditional way in the above sections shows it affects trade flows as an approximation

of transport costs and time. In this section, it represents intangible mental distance of

trade partners. Intangible distance may include differences in cultures, language,

border structure and technology among others; that increase with distance (Bergeijk

and Brakman 2010).

Differences in languages are considered to impede trade in that it complicates the

grounds for doing Business. Exporters and Importers prefer to trade with people who

they can communicate with easily to those with whom they find it difficult to

communicate.

Border structure can have serious effects on the volume of trade between a given set

of countries .Generally, countries that have access to the sea, also have a natural

strategic advantage over land locked ones in terms of lower cross-continental

transport costs (Gwartney et al, 2000). This study includes access to the sea as one of

the factors that determine the level of bilateral flows between countries.

Finally, the study assumes that after controlling for GDP, Population, Exchange rate

and availability of a coastline, South Africa’s trade flows should be higher with

countries with which it shares a common border than those that it doesn’t. The

concept of natural trade partnership predicts such a trend in a way that it makes it

Page 92: South Africa’s Trade Flows -A Gravity Model Analysis

81

easier for South Africa to penetrate closer markets than those located far away as

shown by Frankel et al (1993)

Therefore the final econometric gravity model to be estimated in this study is in the

general form of:

LnXij = β0 + β2LnGjt + β4LnPjt + β5LnDij + β6EXCHij + π0DSADC + π1DAGOA + π2DEU +

π3LANGUAGEj + π4 DBRIC + π5DLANDLOCKED + π6DBORDER + λ0FTAij + Uijt (5.7)

Where;

- FTAj is a structural dummy representing the period since 2001 to 2011. It

measures if there has been a structural change in trade flows of all countries

since the signing of the FTAs.

FTAj = 1 if period beyond year 2000, 0 otherwise

- DEU, DSADC and DAGOA are dummies representing EU, SADC and

AGOA respectively

DEU = 1 if a country is member of EU, 0 otherwise

DSADC = 1 if a country is a member of SADC, 0 otherwise

DAGOA = 1 if a country is a member of AGOA, 0 otherwise.

DBRIC = 1 if a country is a member of BRIC, 0 otherwise

Page 93: South Africa’s Trade Flows -A Gravity Model Analysis

82

- EXCHij is the exchange rate between country i and i. Its effect on trade flows

is expected to be negative.

- DLANDLOCKED is a dummy representing the boarder structure of a given

country.

DLANDLOCKED = 1 if a country j borders an ocean, 0 otherwise

- DBORDER is a dummy representing proximity to South Africa

DBORDER = 1 if a country shares a border with South Africa, 0 otherwise.

- DLANGUAGE = 1 if a country speaks the same official language as South

Africa, 0 otherwise.

5.3.2. Estimation technique

The statistical techniques used to estimate the gravity equation could result to more or

less accurate parameters (Washington et al, 2003).

Traditionally, the ordinary least squares method (OLS) has been the common

technique for estimating the gravity model coefficients. Even though OLS is still

used, some researchers have made known the flaws in methodology and modeling of

the gravity equation using OLS (Henderson and Millimet 2008, Feenstra et al 2001,

Anderson and Wincoop 2003)

5.3.2.1. The Fixed effects Model

Recently, researchers have opted for more advanced techniques such as fixed effects

and random effects models as opposed to the OLS approach. The main reason is that

Page 94: South Africa’s Trade Flows -A Gravity Model Analysis

83

results obtained by the latter are biased because the method does not control for

heterogeneity among regressors.

Since it is very difficult to know the source of heterogeneity bias, FEM enables the

use of dummy variables to control for the possible factors that may be correlated with

the volumes of trade between countries. By introducing such dummies, the FEM

avoids biases that may arise due to omission of variables, - especially those that do

not change over time.

The strength of FEM is that being a technique of panel data analysis; (a combination

of serial and cross sectional data), it gives more information, allows more variability

by increasing the degrees of freedom and also takes heterogeneity into account by

expressing each variable as a deviation from its mean value. A basic FEM model is

estimated as:

LnXijt = β0i + β2LnGDPjt + β4LnPjt + β5LnEXCHijt – β3LnDij + Ut. (5.8)

The subscript i on intercept β0i suggests that the intercepts for every country may be

different. The coefficient lack subscript t to suggest that even though the intercepts

may be different, they do not vary over time. Inferring from the slope coefficient, it is

clear that the model assumes that the slope coefficients do not differ across countries.

Page 95: South Africa’s Trade Flows -A Gravity Model Analysis

84

In order to allow for heterogeneity among countries, the “differential intercept dummy

technique” is used. It calls for the introduction of a differential intercept dummy for

each cross sectional unit. In this case SADC, EU, BRIC and AGOA dummies should

be introduced. However, in order to avoid the dummy variable trap (a presence of

perfect collinearity among the regressors), only one dummy for each category is

introduced.

LnXij = β0 + β2LnGDPjt –β4LnPjt + β6EXCHij + π0DSADC + π1DAGOA + π2DEU + π4

DBRIC + Uijt (5.9)

Since a differential intercept dummy is introduced for each trade arrangement, all

other time invariant variables are excluded from the model. This is because, the

dummy variable representing each country does not change with time and so does

distance, membership to a given agreement, borderline, language among others.

Therefore all heterogeneity that exists in such variables is absorbed into the

differential intercept dummies. If added, the model experiences a problem of perfect

collinearity among time invariant variables, thus making it impossible to identify the

effects of each time invariant variable on the dependent variable.

The FEM discussed above ignores time invariant variables that may be equally

important in the prediction of trade flows. Although the model eliminates any bias

that may result from the omission of such variables, it is impossible to know their

Page 96: South Africa’s Trade Flows -A Gravity Model Analysis

85

specific bearing on the dependent variables because their total effects are absorbed

into the differential intercept. One way to overcome this problem is to use the Error

Component Model (ECM) commonly known as the Random Effects Model (REM).

Instead of assigning an intercept to every cross-sectional unit, the REM only estimates

parameters that describe the distribution of the intercepts.

The REM assumes that the cross-sectional units being studied are a sample from a

large population of units that have a common mean value. How much each country’s

intercept differs from the rest is reflected as a random deviation from the common

intercept.

On the other hand, the REM requires that the random effects be uncorrelated with the

individual effects, which is often very unlikely.

The other way (and one adopted by this study) is to estimate separate regressions, one

including country specific effects and the other taking time invariant variables and

Independent variables (Martinez and Nowak, 2003)

The gravity model is therefore estimated in two stages:

- Stage one estimates bilateral trade flows as a function of the original gravity

equation variables plus all time invariant variables.

- Stage two accounts for unobserved country specific effects.

Thus the equations to be estimated in this study are:

Stage 1

Page 97: South Africa’s Trade Flows -A Gravity Model Analysis

86

- LnXij = β0 + β2LnGDPjt + β4LnPjt + β5LnDij + β6EXCHij + ¥it+ Uijt (5.10)

Where

¥it is a vector of all time invariant variables in the regression.

The study will break equation (5.10) into import and export functions. Doing so

allows the researcher to capture variability in respective data that may be missed in

case of aggregate flows. Estimation of the STAGE 1 export and import function

follows the same procedure as the bilateral flows model. The equations to be

estimated are

Export model:

LnEXPij = β0 + β2LnGDPjt + β4LnPjt + β5LnDij + β6EXCHij + ¥it+ Uijt (5.11)

Imports model:

LnIMPij = β0 + β2LnGDPjt + β4LnPjt + β5LnDij + β6EXCHij + ¥it+ Uijt (5.12)

Stage 2 estimation

LnXij = β0 + β2LnGDPjt + β4LnPjt + β6EXCHij + π0DSADC + π1DAGOA + π2DEU +

π4 DBRIC + Uijt (5.13)

The exports model

LnEXPij = β0 + β2LnGDPjt + β4LnPjt + β6EXCHij + π0DSADC + π1DAGOA + π2DEU

+ π4 DBRIC + Uijt (5.14)

Page 98: South Africa’s Trade Flows -A Gravity Model Analysis

87

The imports model

LnIMPij = β0 + β2LnGDPjt + β4LnPjt + β6EXCHij + π0DSADC + π1DAGOA + π2DEU

+ π4 DBRIC + Uijt (5.15)

5.3.3. Data and Variable description

The study uses bilateral trade data between South Africa and 100 countries for a

period of 16 years starting from 1995-2010. It is important to note that the study does

not consider intra bloc trade. The data includes observations for all 27 EU countries,

10 SADC countries, all 4 BRIC countries, the USA and 58 other countries that do not

belong to any of the trade agreements under consideration. The countries in the latter

group were chosen basing on the size of their National Income and data availability.

Data on exports, imports and the exchange rate were obtained from UNCTAD STAT,

a trade database of the United Nations Conference on Trade and Development and are

measured in constant 2000 USD. The study initially intended to use data provided by

UNCOMTRADE but later found that UNCOMTRADE does not provide trade

statistics for South Africa prior to 2000. UNCTAD on the other hand contains data

prior to 2000 collected from a number of databases including UNCOMTRADE,

Directions of Trade Statistics (DOTS), UN DESA Statistics Division among others

(UNCTAD, 2012A).

Page 99: South Africa’s Trade Flows -A Gravity Model Analysis

88

Data on population and GDP were obtained from World Development Indicators

(2012) and are also measured in constant 2000 USD. Data on distance was obtained

from time and date (2012) expressed in number of kilometers from Johannesburg to

the capital city of a given country. All data on exports, imports, and GDP is expressed

in current US dollars. The exchange rate is presented as a unit of National currency

per US dollar (UNCTAD, 2012B)

Other data on border structure, language, and participation in a given trade agreement

were obtained from Wikipedia and CIA fact book. To avoid inflation biases resulting

from inflationary tendencies, all data has been logged where applicable.

Page 100: South Africa’s Trade Flows -A Gravity Model Analysis

89

CHAPTER VI

ESTIMATION RESULTS

6.1. INTRODUCTION

This chapter presents the results obtained from the regressions as specified in the

previous chapter. The main objective of this research is to investigate the effects of

South Africa’s trade arrangements on its trade flows. It is important to note that the

trade flows considered are between South Africa and its trade partners but trade flows

among partner countries is neglected (i.e. the study does not considered trade flows

between UK vs France, France Vs Germany, Zimbabwe vs Zambia, Zambia vs

Mozambique…etc.). Therefore, the study does not estimate intra-bloc or extra-bloc

trade. In this case, trade creation and trade diversion effects are measured from South

Africa’s perspective and not from the whole of a trade arrangement under

consideration. In this case, a statistically significant and positive or negative EU,

SADC, BRIC, or AGOA dummy coefficient will imply that the respective trade

arrangement has created or dampened trade between South Africa and its trade

partners and in that order.

Another important point to note is given by Kanda and Jordaan (2010) about the

objectivity of studies that investigate trade creation and trade diversion using the

gravity model. He argues that the concept of trade creation and trade diversion must

not be confused with welfare analysis. This is because the dummies employed by the

Page 101: South Africa’s Trade Flows -A Gravity Model Analysis

90

gravity equation only represent bilateral trade flows. Yet, linking the gravity equation

to welfare is only possible by analysing the efficiency losses or gains associated with

changes in exports. This is out of the scope of this study.

To capture best how different factors considered in this study affect trade flows,

results are first presented on a yearly basis (cross-section analysis) to capture how

trade flows have been responding to changes in the independent variables. Then a

Ordinary Least Square and then fixed effects technique will be applied to all 100 cross

sections over a total period of 15 years.

6.2. ESTIMATION RESULTS

The table below shows how bilateral trade flows have been responding to different

variables over the last 16 years. The importance of such cross sectional analysis is that

it helps in identifying the changes in the strengths of relationships between a set of

variables over the years. Identifying such changes involves taking note of changes in

signs and size of coefficients.

Apart from Border, all independent variables have the expected signs. The R2 are high

enough for us to conclude that the variables included convincingly explain the

changes in the dependent variable. The cross sectional results show that contrary to

popular belief, sharing a common border with South Africa does not affect in any way

the level of bilateral trade interactions between any given country with South Africa.

Page 102: South Africa’s Trade Flows -A Gravity Model Analysis

91

The border effect was only statistically significant in 1999-2004. In that period, the

border effect was negative implying that countries that share a border with South

Africa possess special features that lower their bilateral trade flows with South Africa

lower than the average country from the rest of the World

Table 6. 1: Cross section estimation results

Dependent Variable: Bilateral trade flows

Year Constant GDP DistancePopulatio

nExchange

rateLanguage Border R2 F stat

1995 14.283 1.548*** -3.577*** -0.216* 0.079 1.131** -1.984 0.76 46.74

1996 13.916 1.450*** -3.304*** -0.276* 0.065 1.191** -1.662 0.77 46.13

1997 12.622 1.347*** -2.866*** -0.174 0.051 1.154*** -1.623 0.80 53.48

1998 12.258 1.303*** -2.781*** -0.144 0.021 1.140*** -1.254 0.81 57.06

1999 14.652 1.472*** -3.253*** -0.307** 0.063 1.046** -1.593 0.81 57.33

2000 14.119 1.481*** -3.149*** -0.295** 0.065 1.115*** -.729* 0.85 103.76

2001 13.663 1.471*** -3.172*** -0.249** 0.092** 1.017*** -1.841** 0.87 90.93

2002 13.743 1.341*** -2.843*** -0.229** 0.083** 1.191*** -1.503** 0.86 82.12

2003 12.924 1.357*** -2.896*** -0.174** 0.121*** 0.966*** -1.588** 0.86 83.56

2004 11.921 1.413*** -2.935*** -0.171** 0.116** 1.073*** -1.565** 0.84 71.32

2005 11.851 1.371*** -2.814*** -0.167 0.094** 1.057*** -1.410 0.8 54.18

2006 12.038 1.290*** -2.637*** -0.142 0.072 1.003*** -1.220 0.8 53.67

2007 10.130 1.259*** -2.434*** -0.093 0.081** 0.989*** -0.678 0.83 65.12

2008 11.249 1.266*** -2.598*** -0.085 0.054 1.102*** -0.967 0.82 62.37

2009 11.075 1.243*** -2.539*** -0.093 0.075 1.21*** -0.813 0.82 60.02

2010 10.395 1.243*** -2.528*** -0.058 0.073 1.185*** -0.779 0.80 53.67

* significance at 10% level; ** significance at 5% level; *** significance at 1% level.

The results also show that higher populations in partner countries have a negative net

effect on bilateral trade flows with South Africa. The population coefficient was

negative and statistically significant in 1995-2004 except for 1997-1998. This means

Page 103: South Africa’s Trade Flows -A Gravity Model Analysis

92

that in this period the intensity of trade was smaller trade between South Africa and

highly populated countries than with less populated countries. The possible

explanation is that countries with higher populations also have local market that are

big enough to stimulate investment in the production of a variety of goods, thus

reducing the need for South African exports. However, beyond 2004 the impact of

population on bilateral trade flows shrunk in a way that in the last six years the net

effect of population on bilateral flows has been nil.

The impact of the exchange rate on trade flows has also been mixed. The results show

that the depreciation of the dollar only had a significant impact on bilateral trade

flows between 2001 and 2007 (with the exception of 2006). In other periods however

changes in the value of the dollar have had no significant effects on South Africa’s

trade flows.

The strongest and perhaps the most important variables that have influenced South

Africa’s bilateral trade flows over the past sixteen years have been GDP, Distance and

Language. Distance as the major proxy for trade costs has a strong negative impact on

South Africa’s trade flows as predicted by the gravity model. Countries that are

located far away trade less with South Africa than an average country. The

coefficients of the distance variable have been declining in size since 1995. This is

very promising for South Africa’s trade relations since it means that the constraints to

Page 104: South Africa’s Trade Flows -A Gravity Model Analysis

93

trade have been decreasing, thus slowly making it relatively easier to do business this

year than the previous one.

The impact of GDP on trade flows has been positive and statistically significant in all

periods, thus indicating that countries with higher GDP trade more with South Africa

than an average country. However, the coefficients have been declining although not

as rapidly as the distance coefficients. This decline means that even though a trade

partner’s Income plays an important role in determining its level of bilateral flows

with South Africa, trade patterns are changing in a way that South Africa is

increasingly trading more, even with countries that have lower Incomes.

This is a very important development because it implies that South Africa has been

diversifying markets for its products from mainly targeting high income countries to

including production targeting demand from lower income countries. Given the

volatility of markets in high income countries as has been the case in the past five

years, diversifying South Africa’s incomes seems to be a very necessary strategy that

can help stabilise public finances.

Finally, the language coefficient is also statistically significant and positive indicating

that on average, South Africa’s bilateral trade flows have been high with Countries

that speak English as their official language. This makes economic sense since any

trader is bound to find it easier to do business with partners with whom it is easy to

communicate.

Page 105: South Africa’s Trade Flows -A Gravity Model Analysis

94

6.2.2. Panel Estimations

The study runs Panel estimations of the gravity equation. Given that the variables also

included time invariant variables, estimations were done in two stages. The first

included all time invariant variables and the second estimated country specific effects.

The data exhibited severe signs of autocorrelation. As a result, generalized least

squares estimation was performed on the data. This significantly improved the

Durbin-Watson statistic but the resulting R-squared were very low. To avoid loss of

the first observation due to GLS, Prais-Winsten transformation technique was used

(Gujarati, 2003).

The study also runs independent regressions taking exports and imports as dependent

variables instead of using aggregated bilateral total flows. This is done largely

because by considering exports and imports independently, variations (relationships)

between variables that may be unnoticed when aggregate bilateral flows are

considered can be easily identified.

Page 106: South Africa’s Trade Flows -A Gravity Model Analysis

95

Table 6. 2: Stage 1 Estimation results: Exports model

Dependent Variable Exports

Method Panel Least Squares

Variable Coefficient Std. Error t-Statistic Prob.  

C 4.832259 0.501633 9.633057 0.0000

BORDER -0.436705 0.194743 -2.242468 0.0251

FTA 0.264385 0.051719 5.111978 0.0000

LANDLOCKED 0.184566 0.070077 2.633749 0.0085

EXCHANGE RATE -0.077493 0.024964 -3.104213 0.0019

GDP 0.901827 0.044488 20.27107 0.0000

LANGUAGE 0.265943 0.062996 4.221591 0.0000

DISTANCE -0.739103 0.056887 -12.99238 0.0000

POPULATION 0.117888 0.050563 2.331495 0.0199

R-squared 0.453303 F-statistic 164.7971

Adjusted R-squared 0.450553  Durbin-Watson stat 1.741994

Table 6. 3: Stage 1 estimation results: Imports model

Dependent Variable Imports

Method Panel Least Squares

Variable Coefficient Std. Error t-Statistic Prob.  

C 0.811403 0.660402 1.228649 0.2194

BORDER 0.012363 0.256380 0.048221 0.9615

FTA 0.158817 0.068088 2.332526 0.0198

LANDLOCKED 0.024135 0.092257 0.261607 0.7937

EXCH RATE 0.040958 0.032865 1.246236 0.2129

GDP 1.462627 0.058569 24.97264 0.0000

LANGUAGE 0.192128 0.082935 2.316618 0.0207

NDISTANCE -0.495463 0.074893 -6.615656 0.0000

POPULATION -0.191591 0.066567 -2.878173 0.0041

R-squared 0.442833 F-statistic 157.9654

Adjusted R-squared 0.440030  Durbin-Watson stat 2.069152

As expected the estimation results above indicate that GDP and language positively

influence bilateral trade flows between South Africa and its trade partners. The

Page 107: South Africa’s Trade Flows -A Gravity Model Analysis

96

coefficient of Distance is negative indicating that the value of exports and imports

respond negatively to increase in distance. The export coefficient is larger than that

from the imports model thus implying that South Africa exports less to distant

countries than it imports from them.

The coefficient of exchange rate is negative. Following a percentage depreciation of

the dollar, South Africa’s exports fall by around 7.4% [i.e. exp^(-0.07749)-1].

However, the imports model shows that South Africa’s imports are independent of

changes in the exchange rate.

While the Exports model shows that South Africa exports more to countries with

higher populations, the imports model reveals the opposite. The imports elasticity of

population is negative. This indicates that South Africa imports less from countries

with larger populations than it does from those with smaller populations. This finding

makes economic sense in a way that production in economies with a high number of

people is more domestic oriented since larger populations provide enough absorption

capacity for local produce. But also larger populations provide bigger markets for

South Africa’s exports.

The results however show that countries with no access to the sea possess unique

features that attract South Africa exports more than countries that have access to the

Page 108: South Africa’s Trade Flows -A Gravity Model Analysis

97

sea. This is an unexpected result and an outright contradiction to economic theory

and the predictions of Gwartney el at (2000) and the findings of Kwentua (2006).

Another unexpected result is the dummy for countries that share a common border

with South Africa whose coefficient is negative. Countries that are within each other’s

proximity are expected to be natural trade partners due to lower transport costs and

shorter mental distance (Krugman, 1991). In this case however, the result show that

countries that share a border with South Africa do exhibit special features that reduce

their bilateral trade flows with South Africa much lower (around -35%) than the rest

of the world. The most likely explanation for this is the fact that South Africa is

surrounded by countries with lower incomes compared to other countries included in

this study

Conversely, the imports model shows that the coefficient for border is statistically

insignificant; indicating that there is no evidence in the imports data that suggests that

South Africa imports more from its neighbours than from the rest of the world. It is

highly advisable to take this conclusion with caution because only three countries

(Mozambique, Zambia and Zimbabwe) were considered for the analysis due to

constraints regarding availability of data for the remaining countries.

The coefficient for the FTA dummy is positive and statistically significant. This

implies that the period since 2001 (period of trade liberalisation), South Africa has

experienced higher average exports (around 30.2% increases) to all countries

Page 109: South Africa’s Trade Flows -A Gravity Model Analysis

98

compared to the period before. Results show that exports have responded more to

liberalization than imports. The imports model shows that 17.2% of changes in South

Africa’s imports are due to liberalization. This is an indication that South Africa’s

approach to international trade in the past decade has stimulated trade to significantly

higher levels compared to the levels that existed before 2001.

The export and import model produced positive coefficients for the Landlocked

dummy. This implies that that South Africa generally exports more to landlocked

countries than it does from countries with access to the sea. This result does not make

any economic sense. The reason being that exporting to countries with access to the

sea usually involves lower cross-continental transport costs than land locked ones.

Table 6.4: Stage 2: Fixed effects estimation results: Bilateral trade flows model

Dependent Variable Bilateral trade flows

Method Fixed Effects

Estimation

Variable Coefficient Std. Error t-Statistic Prob.  

C -1.531556 0.145007 -10.56193 0.0000

EXCH RATE -0.037407 0.022491 -1.663182 0.0965

GDP 0.954829 0.038471 24.81946 0.0000

POPULATION 0.096518 0.045007 2.144526 0.0321

AGOA -0.053092 0.268667 -0.197612 0.8434

BRIC 0.043555 0.421539 0.103323 0.9177

EU 0.126404 0.062828 2.011906 0.0444

SADC 0.948116 0.088309 10.73641 0.0000

R-squared 0.502033 F-statistic 229.2857

Adjusted R-squared 0.499844  Durbin-Watson stat 1.497564

Unlike the first model, the fixed effects model excludes all time invariant variables so

as to avoid their effects being expended into the individual effects dummies. In table

Page 110: South Africa’s Trade Flows -A Gravity Model Analysis

99

6.4 above the coefficient of the EU and SADC are positive and statistically

significant. It can be concluded therefore that both trade arrangements stimulate trade

for South Africa to levels that are significantly higher than those that existed prior to

their initiation. Since 2001, 13.5% of the total increase in EU-South Africa (EU-SA)

bilateral trade flows is attributed to the free trade agreement where as the SADC free

trade agreement raised bilateral flows by over 158%.

Table 6. 5: Stage 2 estimation results: Exports modelDependent Variable Exports

Method

Fixed Effects

Estimation

Variable Coefficient Std. Error t-Statistic Prob.  

C -1.150567 0.167712 -6.860391 0.0000

EXCH RATE -0.086733 0.026013 -3.334271 0.0009

GDP 0.807538 0.044495 18.14914 0.0000

POPULATION 0.170243 0.052054 3.270532 0.0011

AGOA 0.094232 0.310733 0.303259 0.7617

BRIC 0.381463 0.487541 0.782422 0.4341

EU 0.093572 0.072665 1.287705 0.0198

SADC 1.037231 0.102135 10.15545 0.0000

R-squared 0.391136 F-statistic 146.1005

Adjusted R-squared 0.388458  Durbin-Watson stat 1.574014

The exports function in table 6.5 shows that the EU-SA trade agreement has increased

South Africa’s exports to the EU by over 9.8% and imports by 15.3%. Whereas the

SADC free trade agreement has augmented South Africa’s exports to SADC countries

by over 180% and imports by 133%. . By 2004, Holden and Mcmillan (2006) showed

that the SADC had increased South Africa’s exports to the SADC by over 50%

compared to only 33% by the EU-SA trade agreement. Given that the average

Page 111: South Africa’s Trade Flows -A Gravity Model Analysis

100

purchasing power of EU member countries is much higher than that of SADC, this

pattern can attributed to the negative impact of distance.

Table 6. 6: Stage 2 estimation results: Imports Model

Dependent Variable Imports

Method

Fixed Effects

Estimation

Variable Coefficient Std. Error t-Statistic Prob.  

C -3.268511 0.212367 -15.39089 0.0000

EXCH RATE 0.032000 0.032939 0.971496 0.3314

GDP 1.381841 0.056342 24.52607 0.0000

POPULATION -0.139526 0.065914 -2.116803 0.0344

AGOA -0.269355 0.393469 -0.684566 0.4937

BRIC -0.357469 0.617355 -0.579034 0.5626

EU 0.142805 0.092013 1.552001 0.1209

SADC 0.849995 0.129330 6.572293 0.0000

R-squared 0.426820 F-statistic 169.3552

Adjusted R-squared 0.424300  Durbin-Watson stat 2.013747

Just like the first stage estimations, the second stage bilateral trade flows and the

export models show that South Africa’s bilateral trade is dependent on changes in the

exchange rate. However the imports model in table 6.6 shows that the value of South

Africa’s imports is independent of changes in the exchange rate. On the other hand,

larger populations in partner countries have a negative effect on the value of South

Africa’s imports from them (around 13%) but raise South Africa’s exports to them

(by around 18.5%). On aggregate however, the bilateral trade flows model shows that

the overall impact of population on South Africa’s bilateral trade flows is positive;

thus indicating that South Africa trades more with countries with higher populations

than with countries that have smaller populations.

Page 112: South Africa’s Trade Flows -A Gravity Model Analysis

101

On the other hand, AGOA and BRIC trade arrangements have not stimulated trade

yet. AGOA has been in place since 2001 but it is not a free trade agreement; it only

gives selected South Africa’s commodities free market access to the USA. Given that

AGOA is due to expire in 2015, it can be concluded that it has created no winners and

no losers.

After running for less than one year, the BRIC trade arrangement has also not

produced any significant changes to bilateral trade flows between South Africa and

BRIC members. The period is just not enough to allow for adjustments that are big

enough to produce noticeable effects on bilateral trade flows.

Page 113: South Africa’s Trade Flows -A Gravity Model Analysis

102

CHAPTER VII

CONCLUSION AND RECOMMENDATIONS

7.1. Conclusion and Recommendations

This study examines the importance of South Africa’s trade agreements on its

bilateral trade flows since 1995-2010. To achieve this objective, the study controlled

for a number of variables such as GDP, population, exchange rate, distance, access to

the sea, border and language. Controlling for such variables has an added advantage

in a way that the study is able to identify the main determinants of South Africa’s

trade flows and also to capture the extra influence added by the trade arrangements

considered in this study.

This study shows that South Africa’s bilateral trade flows positively depend on the

partner’s purchasing power and economic growth. Also the study shows that

exchange rate is important and a depreciation of the dollar has a negative impact on

South Africa’s exports. When a dollar loses value, South Africa’s goods become more

expensive for foreign importers. However, the finding of this study in that South

African imports do not seem to be affected by the exchange rate value of the dollar is

puzzling.

On the other hand, the study found regarding the geographic dimension that distance

is important in explaining trade and as expected is typically negative and an

Page 114: South Africa’s Trade Flows -A Gravity Model Analysis

103

impediment to trade. Accounting for market size, by including a population variable

the results obtained in this study indicate that South Africa seems to import less from

larger countries. However, the larger countries seem to import more from South

Africa.

Additionally, the results obtained from panel data estimations of the individual

specific effects model show that only two out of the four trade agreements have had a

significant impact on South Africa’s trade flows. After controlling for SADC, AGOA

and BRIC effects, both exports and bilateral flows models results showed that the EU-

SA trade agreement has significantly stimulated trade for South Africa. However, the

agreement has not stimulated South Africa’s imports from the bloc as indicated by the

imports model. This is a good outcome for South Africa, especially given the fact

that the EU-SA FTA had not fully removed all tariffs on each other’s exports.

However, with further liberalization, the EU-SA partnership promises larger bilateral

trade expansion prospects for both parties.

On the other hand, the SADC FTA has also created trade for South Africa and SADC

in general. The SADC coefficient is positive and statistically significant in both the

exports and imports model. This means that SADC FTA has increased the value of

both imports and exports between SADC countries and South Africa. However, it is

important to highlight that due to data limitations not all the SADC member countries

Page 115: South Africa’s Trade Flows -A Gravity Model Analysis

104

were included in the study. Researchers such as Lattimore and Bottini (2009) have

hinted on the same limitation.

However, the BRIC and AGOA trade agreements have not had any noticeable effects

on their respective trade flows with South Africa. It is important to note that the BRIC

trade arrangement was put in place in 2010 and since this study covers the period of

1995-2010, it is possible that the effects of this trade arrangement have not been felt

yet since it takes time for traders to locate new imports and exports markets in partner

countries. Also, as far as BRIC is concerned, the adjustment period is just too short

for significant changes in trade patterns to emerge.

AGOA on the other hand, has had a fair amount of time for meaningful results to be

felt. Given that the overall objective of AGOA is to increase trade with South Africa

in particular and Africa in general using trade as a vehicle to create opportunities for

sustained business enterprise and Investments, the expected pattern should be one that

sees an increment in exports for South Africa. It can be concluded therefore that since

AGOA is not a free trade agreement per se, upgrading it to total liberalization in areas

might improve South Africa’s exports performance on the US market.

Lastly, it has been shown that where South Africa has made strong commitments to

free trade, the results have not been disappointing. Both the SADC and EU trade

agreements have tremendously improved South Africa’s trade performance. Since

Page 116: South Africa’s Trade Flows -A Gravity Model Analysis

105

multilateral trade negotiations have been moving at a snail’s pace, efforts to negotiate

free trade agreements should be pursued with renewed enthusiasm. The importance of

border, distance and GDP on bilateral trade flows have been decreasing in the past

decade. Therefore, considering that trade arrangements and other variables matter for

trade, South Africa should strengthen its trade ties and suggest targeting developing

markets particularly emerging southern countries. Additionally, South Africa should

try to tailor its export sector to the changing structure of imports demand in these

countries. Thus, ensuring a steadier and diversified source of income that will

contribute towards South Africa’s economic growth and development

Page 117: South Africa’s Trade Flows -A Gravity Model Analysis

106

Bibliography

Alesina, A., Spolaore, E. and Wacziarg, R. (2005). Trade, Growth and the Size of

Countries. In: Philippe Aghion & Steven Durlauf (ed.), Hand book of Economic

Growth, North Holland, Amsterdam: Elsevier B.V Publishers, pp. 1499-1542.

Anderson J.E. (1979). A Theoretical Foundation of the Gravity equation. American

Economics, Vol. 69, March, pp.106-116.

Anderson, J.E. and Van Wincoop, E. (2003). Gravity with gravitas: a solution to the

border puzzle. American Economic Review. Vol. 93, no. 1, March, pp.171-192.

Augier, P., Gasiorek, M. and Laitong, C. (2004). The impact of rules of origin on

trade flows. Economic Policy, CEPR and CES and MSH, Vol. 20, no. 43, July, pp.

567-624

Awokuse, T. (2003). Is the export-led growth hypothesis valid for Canada?. Canadian

Journal of Economics/Revue canadienne d'économique. Vol. 36, no. 1, March, pp.

126-136.

Babatunde, O. (2007). African regional trade agreements: Conditions, requirements

and implications for regional integration. In: African Economic Conference, 2007,

Washington DC, USA. pp. 85-107.

Bagwell, K. and Staiger, R.W. (2002). Economic Theory and the interpretation of

GATT/WTO. In: Szenberg, M., Ramrattan, L. and Samuelson, P.A., (ed). New

Frontiers in Economics, New York: Cambridge University Press, 2004, pp. 205-234.

Page 118: South Africa’s Trade Flows -A Gravity Model Analysis

107

Baharumshah, A.Z and Bergstrand, J.H (2003). On the sustainability of Current

Account Deficits: Evidence from Four ASEAN countries. Journal of Asian

Economics, Vol. 14, no. 3, June, pp. 465–487.

Baier, S.L. and Bergstrand, J.H. (2004). Economic determinants of free trade

agreements. Journal of International Economics, Vol. 64, no. 1, October, pp. 29–63.

Baldwin, R. (1971). Determinants of the commodity structure of US trade. American

Economic Review, vol. 61, no.1, March 1971, pp. 126-46.

Baldwin, R. (1993). A Domino Theory of Regionalism. (Working paper no. 4465,

September). National Bureau of Economic Research, Cambridge, MA. 02138.

Available from: <http://www.nber.org/papers/w4465>. Accessed: 3rd July 2012.

Baldwin, R. and Taglioni, D. (2006). Gravity for Dummies and Dummies for Gravity

Equations. (Working paper no. 12516, September). National Bureau of Economic

Research, Massachusetts Avenue, Cambridge, MA 02138. Available from:

<http://www.nber.org/papers/w12516.pdf>. Accessed: 23rd July 2012.

Baldwin, R. (2004). Stepping stones or building blocs? Regional and Multilateral

Integration. In: G-20 Workshop, 23rd September 2004, Mimeo, Graduate Institute of

International Studies, Geneva. pp. 1-27.

Bergeijk, A.G. and Brakman, S. (2010). The Gravity Model in International Trade:

Advances and Application. 2010 ed. New York: Cambridge University Press.

Bergstrand, J.H. (1985). The Gravity Equation in International Trade: Some

Microeconomic Foundations and Empirical Evidence. The Review of Economics and

Statistics, Vol. 67 no. 3, August, pp. 474–481.

Page 119: South Africa’s Trade Flows -A Gravity Model Analysis

108

Bhagwati, J. (1991) The World trading system at risk. 1991 ed. Princeton NJ:

Princeton University Press.

Bhagwati, J. and Panagariya, A. (1996). Preferential Trading Areas and Multilateralism:

Strangers, Friends or Foes? 1996. ed. New York: Columbia University Department of

Economics.

Blouet, O. and Blouet, W. (2005) Latin America and the Caribbean: A Systematic and

Regional Survey. 5th edition: New York: John Wiley.

Brecx, P. (1967). Leontief’s Paradox. The Review of Economics and Statistics, Vol.

44, Jan, pp. 603-607.

Breuss, F. and Egger, P. (1999). How reliable are estimations of East-West trade

potentials based on Cross-section Gravity analyses? Empirica, Vol.26, no. 2, 1999,

pp. 81-95.

Cooper, C. and Massell, B. (1965). Toward a General Theory of Customs Unions for

Developing Countries. Journal of Political Economy, Vol. 73, no. 5, September, pp.

461-76.

Deardorffs' Glossary of International Economics [online]. (2012). Available from:

<http://www-personal.umich.edu/~alandear/glossary/m.html>. [Accessed 28th june

2012].

Dhawan, U. and Biswal, B. (1999). Re-Examining the Export-Led-Growth

Hypothesis: A Multivariate Cointegration Analysis for India. Applied Economics,

Vol. 31, no. 4, April, pp. 525-530.

Page 120: South Africa’s Trade Flows -A Gravity Model Analysis

109

Dixit, A.K. and Norman, V. (1980). Theory of international trade: A Dual, General

Equilibrium Approach. 1980 ed. Cambridge: Cambridge University Press

Dominick, S. (2007). International economics. 8th ed. California: John Wiley & Sons,

Inc.

Donald R. and Weinstein, E. (1998). Economic geography and regional production

structure: an empirical investigation. European Economic Review, Vol. 43, 1999, pp.

379-407.

DTI.(2012) Department of Trade and Industry [online]. (2012). Available from:

<http://www.dti.gov.za/about_dti.jsp>. [Accessed 15th May 2012].

Du Plessis, S.P.J. (1987). International Economics. United Kingdom: Butterworth-

Heinemann limited.

Egger, P. and Larch, M. (2008). Interdependent preferential trade agreement

memberships: An empirical analysis. Journal of International Economics, Vol. 76

Issue 2, Dec, pp. 384-399.

Egger, P. (2002). An Econometric view on the estimation of gravity models and the

calculation of trade potentials. World Economy, Vol. 25 issue 2, February, pp. 297-

312.

Evenett, S.J. and Keller, W. (2002). On theories explaining the success of the gravity

equation. Journal of Political Economics.Vol. 110 no. 2, pp. 281-316.

Feenstra, R.C. (2002). Border Effects and the Gravity Equation: Consistent Methods

for Estimation. Scottish Journal of Political Economy, Vol. 49, pp. 491-506.

Page 121: South Africa’s Trade Flows -A Gravity Model Analysis

110

Feenstra, R.C and Taylor, M. (2008). International economics. 1st ed. California:

Worth Publishers.

Feenstra, R.C., Markusen, J.R and Rose, A.K. (2001). Using the gravity equation to

differentiate among alternative theories of trade. Canadian journal of economics, Vol.

34 no. 2, August, pp. 430-47.

Filippini, C. and Molini, V. (2003). The determinants of East Asian trade flows: a

gravity equation approach. Journal of Asian Economics, Vol. 14, no. 5, pp. 695-711.

Frankel, J.A., Stein, E. and Wei, S. (1993). Continental trading blocs: are they

natural or supernatural. (Working paper no. 4588, December). National Bureau of

Economic Review, 1050 Massachusetts Avenue, Cambridge, MA 02138. Available

from: <http://www.nber.org/papers/w4588.pdf?new_window=1>. [Accessed: 1st

September 2012].

Giorgia, A. (2008). Regionalism or multilateralism? A political economy choice.

(Working Paper no. 08/65, March.). International Monetary Fund working paper, IMF

institute. Available from:

<http://www.imf.org/external/pubs/ft/wp/2008/wp0865.pdf>. [Accessed: 9th May

2012].

Glick, R. and Rose, K. (2002). Does a Currency Union Affect Trade: The time series

evidence. European Economic Review, Vol. 46, no. 6, pp. 1125-1151.

Gopinath, M. and Echeverria, R. (2004). Does economic development impact the

foreign direct investment trade relationship: A gravity model approach. American

Journal of Agricultural Economics, Vol. 86, no. 3, pp. 782-787.

Page 122: South Africa’s Trade Flows -A Gravity Model Analysis

111

Gray,H., (1998). Intra - Industry Trade: An untidy phenomenon. Weltwirtschaftliches

Archive, Vol. 124, pp. 211-229.

Grossman, G.M. and Helpman, E. (1991). Innovation and Growth in the global

economy. Cambridge: MIT Press.

Grossman, G.M. and Helpman, E. (1994). Technology and trade: Handbook of

international economics volume 3. Amsterdam, Elsevier: Grossman, G.M. and K.

Rogoff.

Gujarati D.N. (2003). Basic Econometrics. 4th ed. 1221 Avenue of the Americas,

New York: MCGraw-Hill/Irwin.

Gwartney, J., Skipton, C.D. and Lawson, R.A (2000). Trade openness and long run

economic growth. November 2000. Avaliable from: <

https://docs.google.com/viewer?a=v >. [Accessed: 30th August 2012].

Hachicha, N. (2003). Exports, Export composition and growth: A simultaneous Error-

correction model for Tunisia. International economic Journal, Vol. 17 no.1, pp. 101-

120.

Helpman, E. and Krugman, P. (1985). Market Structure and Foreign Trade. US 55

Hayward Street; Cambridge, MA 02142-1315 USA: The MIT Press.

Henderson, D.J. and Millimet, D.L. (2008). Is gravity linear? Journal of applied

economics, Vol. 23, pp. 137-172.

Holden, M. and McMillan, L. (2006). Do free trade agreements create trade for South

Africa? Trade and Industrial Policy Strategies (TIPS). Trade and Industrial Monitor,

No. 37, pp. 111-122.

Page 123: South Africa’s Trade Flows -A Gravity Model Analysis

112

IMF. (2011) World Economic Outlook Database, 2011. Available from:

http://www.imf.org/external/pubs/ft/weo/2011/01/weodata/weoselgr.aspx

[Accessed: 10th February 2012].

Jayasinghe, S. and Sarker, R. (2007). Regional trade agreements and trade in agri-

food products: Evidence from the EU from gravity modeling using disaggregated

data. Journal of the international Association of Agricultural Economics, Volume 37,

no. 1, July, pp. 93–104.

Jordaan, C. and Hinaunye, E. (2007). Export8 and Economic Growth in Namibia: A

Granger causality analysis. South African Journal of Economics, volume 75, no. 3,

September, pp. 540–547.

Kanda, T.P. and Jordaan, C. (2010). Trade Diversion and Trade Creation: An

Augmented Gravity Model Study for South Africa. TIPS Small Grant Scheme

Research, Paper Series, 2010. Available from:

<http://www.tips.org.za/files/Patrick_tips_final_Jan2010.pdf>. [Accessed: 10th

August 2012].

Kangas, K. and Niskanen, A. (2003). Trade in forest products between European

Union and Central Eastern European access candidates. Forest Policy Economics,

Vol. 5 no. 3 pp. 297-304.

Kelvin, L. (1980). Intra-industry trade under perfect monopolistic competition.

Journal of International Economics, Vol. 10, pp. 151-175.

Kono, D. (2002). Are Free Trade Areas Good for Multilateralism? Evidence from the

European Free Trade Association. International Studies Quarterly, Vol. 46, pp.507-

527.

Page 124: South Africa’s Trade Flows -A Gravity Model Analysis

113

Krishna, P. (1998). Regionalism and Multilateralism: A political economy approach.

The Quarterly Journal of Economics, Vol. 113, no. 1, pp. 227-51.

Krugman, P. (1987). Is Free Trade Passe? The Journal of Economic Perspectives,

Vol. 1, no. 2, pp.456-498.

Krugma, P. (1991). The move toward Free Trade Zones: Policy implications of trade

and currency zones. Federal Reserve Bank of Kansas City: Jackson Hole, Wyoming.

Krugman, P., Obstfeld, M. and Melitz M.J. (2011). International Economics: Theory

and Policy, student value edition. 9th ed. New York: Prentice Hall publishers.

Krugman, P. and Venables, A. (1997) How Robust is the Home Market Effect? 1997.

ed. London: Mimeo, MIT and London School of Economics.

Kucera, D. and Sarna R. (2006). Trade Union rights, democracy and exports: A

gravity model approach. Review of international economics, Vol. 14, no. 5, pp. 859-

82.

Kwentua, G. (2006). Trade creation and trade diversion effects in the EU-South

Africa Free Trade Agreement. M.Sc. thesis, University of Nigeria.

Laszlo, M. (1997). Proper econometric specification of the gravity model. 1997. ed.

Oxford, UK: Blackwell Publishers Ltd.

Leontief, W. (1953). Domestic Production and Foreign Trade: the American Capital

Position Re-examined. The Proceedings of the American Philosophical Society, Vol.

97, September, pp. 331-49.

Page 125: South Africa’s Trade Flows -A Gravity Model Analysis

114

Leontief, W. (1956). Factor Proportions and the Structure of American Trade; Further

Theoretical and Empirical Analysis. Review of Economic Statistics, Vol. 38, no. 4,

November, pp. 386-407.

Lester, S. and Mercurio, B. (2009). Bilateral and regional trade agreements

commentary analysis. 2009. ed. USA: Cambridge University Press.

Linder, S. (1961). An essay on trade and transformation. 1961. ed. New York: John

Wiley and Sons.

Lipsey, R., (1957). The theory of Customs Unions: Trade Diversion and Welfare.

Economica, Vol. 24, no. 93, pp. 40-46.

Martin, R and Sunley, P. (1996). Paul Krugman’s Geographical economics and its

implications for regional development theory. Economic Geography, Vol. 72, no. 3,

July, pp. 259-292.

Martinez, Z. and Nowak L. (2003). The impact of a customs union between turkey

and the EU on Turkey’s exports to the EU. Journal of Common Markets, Vol. 45, no.

3: pp 719-43.

Meade, J. (1955). The theory of Customs Unions. 1955. ed. Amsterdam, North

Holland. North-Holland Publishing Company.

Mensbrugghe, D., Newfarmer, R. and Pierola, M.D. (2005). Regionalism vs.

Multilateralism. In: Richard Newfarmer, (ed). Trade, Doha, and Development,

Washington, DC: The World Bank, pp309-318.

Merrett, S.R. (1965). The Leontief paradox. The economic Journal, Vol. 75, No. 299,

September, pp. 641.

Page 126: South Africa’s Trade Flows -A Gravity Model Analysis

115

Mina, M., Puri, L. and Taisuke, I. (2005). Multilateralism and regionalism: the new

interface. New York and Geneva: United Nations Conference on Trade and

Development (UNCTAD).

Mthembu, S.G. (2008). The Cost of Non-tariff Barriers to Business along the North–

South Corridor (South Africa–Zimbabwe) via Beit Bridge: A Preliminary Study.

South African Institute of International Affair, Trade Report, No 20, May 2008.

Nandasiri, K. (2008). FTA influence on trade creation and diversion by regional

trading blocs. The Journal of the Korean Economy, Vol. 9, no. 2, August. pp 293-333.

Nistch, V. (2000). National borders and international trade: Evidence from the EU.

Canadian Journal of Economics, Vol. 33, no. 4, August, pp. 1091-1105.

Nuno, C.L. (2010). The Gravity Model and United States Trade. European Journal of

Economics, Finance and Administrative Sciences, no. 21, June, pp. 92.

Panagariya A. (2000). Preferential trade liberalization: the traditional theory and new

developments. Journal of economic literature, Vol. 38, July, pp.287-331.

Panagariya, A. (1996). The Free Trade Area of the Americas: Good for Latin

America? The World Economy, Vol. 19, no. 5, March, pp. 485-515.

Pelletiere, D. and Reinert, K.A. (2003). Used automobile protection and trade: gravity

and ordered probit analysis. Empirical Economics, Vol. 29, no. 4, pp. 737-51.

Perry, B. (2000). Rhetoric or reality? EU policy towards South Africa 1977-2000.

European Development Policy Study Group Discussion Paper no.19, Bradford:

Bradford University, October.

Page 127: South Africa’s Trade Flows -A Gravity Model Analysis

116

Pomfret, R. (1992). International trade policy with imperfect competition.. Special

papers in international economics, no. 17, New Jersey: International Finance Section,

Department of Economics Princeton University, August.

Robles, A. (2008). EU FTA Negotiations with SADC and Mercosur: integration into

the world economy or market access for EU firms? Routledge Taylor Francis Group.

Third World Quarterly, Vol. 29, no. 1, pp. 181 – 197.

Rodrik, D. (2001A). Trading in Illusions. Washington DC: Washington post

Newsweek Interactive publishers. In Foreign Policy no.123, March-April, pp. 54-62.

Available from: <http://www.jstor.org/stable/3183155>. Accessed: July 2012.

Rodrik, D. (2001B). The Global Governance of Trade as if Development Really

Mattered. UNDP: New York.

Rodrik, D. (2008). Is Export Led Growth Passé? Daily News Egypt, 12 September.

Available from <http://dailystaregypt.com/printerfriendly.aspx?ArticleID=16429>.

Accessed: 13th May 2012.

Rosson, P., Ford, R.C. and Moulton, K.S. (2003). Preferential Trading Arrangements:

Gainers and losers from regional trading blocs. Canada: Social research and

demonstration corporation (SRDC), no. 198-8. Available from <

http://www.ces.ncsu.edu/depts/agecon/trade/eight.html> [Accessed 26th Nov 2012.]

SADC today [online]. (2010). Available from:

<http://www.sardc.net/editorial/sadctoday/documents/v12n5.pdf>. [Accessed 22nd

March 2012].

Page 128: South Africa’s Trade Flows -A Gravity Model Analysis

117

Samad, A. (2011). Exploring Exports and economic growth causality in Algeria.

Journal of economics and behavioural studies, Vol. 2, no. 3, March, pp. 92-96.

Shujiro, U. (2002). Globalization and Growth in Free Trade Agreements. Asia Pacific

review, Vol. 9, no. 1 pp. 20-32.

Suranovic, S. (1997). Trade theories and Realities: Why Economists should study

fairness. Challenge, Vol. 40, no. 5, pp. 109-124.

Suranovic, S. (2010). International Trade: Theory and Policy. Nyack, New York: Flat

World Knowledge.

Susanto, D. Rosson, P. and Adcock, F.J. (2007). Trade Creation and Trade Diversion

in the North American Free Trade Agreement: The Case of the Agricultural Sector.

Southern Agricultural Economics Association; Journal of Agricultural and Applied

Economics, Vol. 39, no. 1, April, pp. 121–134.

Thirwall, A.P. (1995). The Terms of Trade, Debt and Development: with Special

Reference to Africa. African Development Review, Vol.7, no. 1, June, pp. 1-34.

Tiiu,P. (2000). Gravity approach for modeling trade flows between Estonia and the

main trading partners. Tartu: Tartu University Press. Faculty of Economics and

Business Administration Working Paper Series, no. 4, ISSN 1406-5967. Available

from < http://www.mtk.ut.ee/doc/febawb4.pdf>.

Time and Date (2012) [online]. Available from:

<http://www.timeanddate.com/worldclock/distance.html>. [Accessed 28th May

2012].

Page 129: South Africa’s Trade Flows -A Gravity Model Analysis

118

UNCOMTRADE. (2012) United Nations Commodity Trade Statistics Database,

2012.Available from: http://comtrade.un.org/db/ [Accessed: 10th June 2012].

UNCTAD (2012A). United Nations Conference on Trade and Development [online].

Available from: <http://unctadstat.unctad.org/TableViewer/summary.aspx?

ReportId=247sa41>. [Accessed 22nd May 2012].

UNCTAD (2012B). United Nations Conference on Trade and Development [online].

Available from: <http://unctadstat.unctad.org/TableViewer/summary.aspx?

ReportId=117>. [Accessed 22nd May 2012].

UNECA (2011). Study on the establishment of inter-REC’s Free Trade Areas in

Africa drawing lessons from the COMESA-SADC-EAC FTA experience. United

Nation Economic Commission for Africa: Addis Ababa, Ethiopia.

Viner, J. (1950). The Customs Union Issue. New York: Carnegie Endowment for

International Peace.

Washington, S., Karlaftis, M.G. and Mannering, L.F. (2003). Statistical and

econometric methods for transportation data analysis. 1st. ed. USA: Chapman &

Hall/CRC Press.

World Bank: World Development Indicators (2012), World Development Indicators

[online]. (2012). Available from: <http://data.worldbank.org/data-catalog/world-

development-indicators/wdi-2010>. [Accessed 1st May 2012].

Wei S.J. (1996). Intra-National Versus International Trade: How Stubborn are

Nations in Global Integration? (NBER working paper, no. 5531, April). National

Page 130: South Africa’s Trade Flows -A Gravity Model Analysis

119

Bureau of Economic Research, 1050 Massachusetts Avenue. Available from:

<http://www.nber.org/papers/w5531.pdf>. Accessed: 9th August 2012.

Workman, D. (2010). suite101 [online]. Available from: <http://daniel-

workman.suite101.com/us-versus-south-africa--top-export-and-import-products-

a246472>. [Accessed 5th March 2012]

Yannikaya, H. (2002). Trade openness and economic growth: a cross-country

empirical investigation. Journal of Development Economics, Vol. 72 October, pp. 57-

58.


Recommended