Working Paper Series 2009-03

Export Market Selection Methods and the

Identification of Realistic Export Opportunities for
South Africa Using a Decision Support Model

ERMIE STEENKAMP, RIAAN ROSSOUW, WILMA VIVIERS, AND

LUDO CUYVERS

November 2009

indigenous growth
 

EXPORT MARKET SELECTION METHODS AND THE IDENTIFICATION

OF REALISTIC EXPORT OPPORTUNITIES FOR SOUTH AFRICA USING A
DECISION SUPPORT MODEL

ERMIE STEENKAMP, RIAAN ROSSOUW, WILMA VIVIERS, AND LUDO CUYVERS

NOVEMBER 2009

 
 

TABLE OF CONTENTS

Page
1. Introduction 1

2. Literature overview: Market selection methods for international expansion 2

2.1 International market selection methods 3

2.1.1 Qualitative approaches 3
2.1.2 Quantitative approaches 4
2.1.2.1 Market grouping methods 4
2.1.2.2 Market estimation methods 4
2.1.2.2.1 Firm level market estimation methods 5
2.1.2.2.2 Country level market estimation methods 5
2.1.2.2.2.1 Green and Allaway’s shift-share model 6
2.1.2.2.2.2 Russow and Okoroafo’s global screening model 7
2.1.2.2.2.3 Papadopoulos et al.’s trade-off model 7
2.1.2.2.2.4 The ITC’s multiple criteria method 10
2.1.2.2.2.5 Assessment of export opportunities in emerging markets 11
2.1.2.2.2.6 Cuyvers’ decision support model 12
2.2 Summary and method selected 13
3. The South African approach 14
3.1 A decision support model for planning export promotion activities 14
3.2 Filter one: Which countries show preliminary export opportunities for South African
products 14
3.3 Filter two: Detecting possible export opportunities for South Africa 16
3.4 Filter three: The selection of realistic export opportunities for South Africa 19
3.4.1 Variable 1: Distance 20
3.4.2 Variable 2: Cost 21
3.4.3 Variable 3: Logistics Performance Index 20
3.4.4 Variable 4: Average applied tariffs 21
3.4.5 Variable 5: Frequency Coverage ratio of non-tariff barriers 21
3.4.6 Construction of an index for market accessibility 21
3.5 Filter four and results: An analysis of South Africa’s realistic export opportunities 24
4. Policy implications and Conclusion 28

ii
 

LIST OF TABLES
Page
Table 1: Papadopoulos et al.’s (2002) trade-off model – variables and measure 8
Table 2: Country X’s risk ratings 15
Table 3: Country X’s transformed risk ratings 15
Table 4: Categorisation of country-product combinations 18
Table 5: Distribution of country-product combinations according to short-term import market
growth, long-term import market growth and relative import market size, 2004 18
Table 6: Most accessible countries based on degree of restriction – “green” countries 22
Table 7: Lesser accessible countries based on degree of restriction – “orange” countries 23
Table 8: Least accessible countries based on degree of restriction – “red” countries 23
Table 9: Selection of realistic export opportunities for South Africa, 2004 24
Table 10: Realistic export opportunities per country, 2004 25
Table 11: South Africa: distribution of realistic export opportunities according to regional clusters 26
Table 12: South Africa’s realistic export opportunities according to relative market position and
market characteristics, 2004 26

iii
 

LIST OF FIGURES
Page
Figure 1: Categorisation of the international market selection literature 3
Figure 2: Papadopoulos et al.’s (2002) trade-off model 8
Figure 3: Two-dimensional matrix for plotting countries in Papadopoulos et al.’s (2002) trade-off
model. 9
Figure 4: Global clusters according to overall share of realistic export opportunities in 2004 28

iv
 

LIST OF ABBREVIATIONS / ACRONYMS

ASEAN: Association of Southeast Asian Nations
CGIC: Credit Guarantee Insurance Corporation
DTI: Department of Trade and Industry
DSM: Decision Support Model
EU: European Union
ITC: International Trade Centre
GDP: Gross Domestic Production
HS: Harmonised System
LPI: Logistics Performance Index
MFN: Most Favoured Nation
OECD: Organisation for Economic Co-operation and Development
ONDD: Office National du Ducroire
SITC: Standard International Trade Classification
TPO: Trade Promotion Organisation
TRAINS: Trade Analysis and Information System
UNCTAD: United Nations Conference on Trade and Development
WTO: World Trade Organisation
UK: United Kingdom
USA: United States of America

v
 

EXECUTIVE SUMMARY
Export promotion activities in South Africa are, to a large extent, driven by historical trends and trading
partners (the dti, 2005). However, the limited resources of government should be allocated in such a manner
that they contribute towards successful exports and increased export growth in the future. The dti has
indicated that further research on international market selection for export promotion in South Africa would
greatly assist senior management in ensuring that government resources are used with maximum return on
investment by determining priority products and markets (Erero, 2004).
Market selection methods, of which a vast number exist, are a critical tool in firms’ and government’s
policy, planning and budgeting processes. To this end, the primary aim of this paper is to determine the
international market selection method best-suited to the identification of potential export opportunities for
South Africa. The secondary aim is to apply the chosen method for South Africa in order to determine
realistic export opportunities (country-product combinations).
The decision support model (DSM) of Cuyvers et al. (1995:173-186) and Cuyvers (2004:255-278) was
chosen as the most appropriate international market selection method for the purposes of this study. It was
henceforth applied to South Africa in order to provide the dti with a tool to justify export promotion
activities more scientifically.
The methodology of the DSM developed is discussed. The DSM consists of a sequential filtering process
using four filters to eliminate countries with lower export potential. The first filter considers the macro-
economic environment of the trading partner. Indicators such as country risk ratings; GDP (GDP per capita);
and GDP growth (GDP per capita growth) play a role in the selection process.
In filter two, import market growth in the short and long term, and relative market size, were considered
for each country-product combination. A table has been constructed to show the categories that will be used
for further analysis. In filter three, the Herfindahl-Hirschmann Index gives an indication of the market
concentration of the importing countries and barriers to entry. In filter four, realistic export opportunities
identified in the previous filters are classified. This classification was done by calculating South Africa’s
relative market importance for each country-product combination and combining this with the categorisation
in filter two.
After the application and adaptation of the DSM for South Africa, 12,695 country-product combinations
were identified as realistic export opportunities. After the identification of the 12,695 country-product
combinations, a clustering process was undertaken to enable the dti to focus on specific regions, if needed,
when developing their export promotion strategies. This clustering is reported and graphically represented in
this paper.
It is recommended that the DSM results should form part of an overall strategy towards increasing
exports through utilising government resources in order to contribute to the effectiveness of export
promotion in South Africa.

vi
 

EXPORT MARKET SELECTION METHODS AND THE IDENTIFICATION

OF REALISTIC EXPORT OPPORTUNITIES FOR SOUTH AFRICA USING A
DECISION SUPPORT MODEL

ERMIE STEENKAMP1,2, RIAAN ROSSOUW1, WILMA VIVIERS1, AND LUDO CUYVERS3

Abstract
Market selection methods, of which a vast number exist, are a critical tool in firms’ and government’s policy,
planning and budgeting processes. To this end, the primary aim of this paper is to determine the international
market selection method best-suited to the identification of potential export opportunities for South Africa.
The secondary aim is to apply the chosen method to South Africa in order to determine realistic export
opportunities (country-product combinations). The decision support model chosen for application in this
study consists of a screening process of four consecutive filters, through which relevant information on
markets (such as country risk indicators, macroeconomic data, imports per product group, etc.) is fed, and
which allows the identification realistic export opportunities. Results are reported on the application of this
decision support model to the case of South Africa, adapted for an analysis of foreign trade data at the SITC
four-digit level up to 2004. In this way, South Africa’s export opportunities in individual countries are listed
and categorised according to criteria such as import market characteristics and South Africa’s market share in
the various markets.

Keywords: Exports, market research, decision support model, South Africa

JEL classification: F13, F14, D81

1. INTRODUCTION

The Department of Trade and Industry (the dti), as the primary trade promotion organisation (TPO) in
South Africa, finances or co-finances export activities on a regular basis. Several hundreds of projects
submitted annually to the TPO for funds must be evaluated for their suitability for funding from the state
budget. Until recently, however, export promotion activities in South Africa have been based on historical
export performance trends. Very little attention has been devoted to new export opportunities in unexploited
markets or opportunities for new products in existing markets. Still, the South African government wants to
fulfil its obligation, not only to assist potential exporters, but also to prioritise export promotion activities in a

1 School of Economics, North-West University, Potchefstroom Campus, Private Bag X6001, Potchefstroom, 2520, South Africa.

2 Corresponding author. E-mail addresses: ermie.steenkamp@nwu.ac.za (E.A. Steenkamp), riaan.rossouw@nwu.ac.za (R. Rossouw),
wilma.viviers@nwu.ac.za (W. Viviers), ludo.cuyvers@ua.ac.be (L. Cuyvers).

3 Department of International Management and Centre for ASEAN Studies, Faculty of Applied Economics, University of Antwerp,
Antwerp, Belgium.

1
 

manner that will yield a high return on investment of scarce resources, while increasing the success rate of
South African exporters (the dti, 2005:47).
For this to be realised, the South African government must distinguish between the vast numbers of
export opportunities that exist. Because of scarce resources, only a limited number of these can be explored.
The challenge faced by government thus lies in the selection of specific sectors for export promotion and the
allocation of its limited resources among these sectors. Market selection methods, of which a vast number
exist (e.g. Papadopoulous and Denis, 1988; Green and Allaway, 1985; Russow and Okoroafo, 1996;
Papadopoulos et al., 2002; Freudenberg and Paulmier, 2005a;b; Shankarmahesh et al., 2005; Sakarya et al.,
2007), are therefore a critical tool in government policy, planning and budgeting processes.
To this end, the primary aim of this paper is to determine the international market selection method best-
suited to the identification of potential export opportunities for South Africa. The secondary aim is to apply
the chosen method to South Africa in order to determine realistic export opportunities (country-product
combinations) for the country. With the starting point as all the countries of the world and all possible
products, the decision support model (DSM) of Cuyvers et al. (1995) and Cuyvers (2004) seemed to be the
best suited to the task. The model consists of four consecutive steps or ‘filters’ (adapted for the circumstances
of South African trade), leading to a list of realistic export opportunities in countries with sufficient
macroeconomic strength and performance. This facilitates transparent identification and evaluation of
realistic export opportunities. It is thus possible, in a simple and effective manner, to obtain an answer to the
question: how can government prioritise export assistance for potentially successful exporters?
This paper presents a modified decision support method for the identification of realistic export
opportunities for South Africa. First, various market selection methods for international expansion are
presented and discussed and this is followed by a more detailed description of the decision support model
selected. The paper concludes with a presentation of the results and policy implications for South Africa of
the application of the model.

2. LITERATURE OVERVIEW: MARKET SELECTION METHODS FOR INTERNATIONAL

EXPANSION

Governments and individual firms that want to stimulate growth through export development must
distinguish between the vast number of export combinations due to the fact that, in most circumstances, a
large number of export opportunities exist, and only a limited number of these can be explored because of
scarce resources (Papadopoulos and Denis, 1988:38). The challenge that governments and individual firms
therefore face is in choosing specific sectors for export promotion (Shankarmahesh et al., 2005:204). In order
to yield a higher return on investment and make sure that resources are not wasted on less attractive export
markets, they should focus their efforts and resources on a limited set of dominant export markets
(Shankarmahesh et al., 2005:204). Furthermore, selecting the “right” market is important as a first step in
expanding exports to ensure export success, determining foreign marketing strategies and determining where
to establish bases to establish a favourable competitive position in those markets (Papadopoulos and Denis,
1988:38). Rahman (2003:119) stated that there exists a well-developed literature of market failures
encountered by international marketers and that the biggest reason for these failures is poor market selection
resulting from inappropriate evaluation of markets.
The process of evaluating worldwide export opportunities is complicated for a number of reasons. These
include the difficulty of examining all possible export opportunities to all the countries of the world and the
availability of data for specific consumers, businesses or governments that limits the screening process to

2
 

using only published data (Jeannet and Hennessey, 1988:137; Brewer, 2001:155). Numerous attempts to
formulate appropriate international market selection processes have been made in the literature.
The purpose of this section is to find the international market selection method best-suited to the
identification of potential export opportunities for a given exporting country (in this case, South Africa). The
focus is therefore on country-level (macro-level) rather than firm-level (micro-level) market selection (see
section 2.1.2.2.1 and 2.1.2.2.2). This implies that all possible country-product combinations worldwide must
be screened in order to identify a list of priority export opportunities for the exporting country. Different
international market selection models or processes will be discussed and evaluated subsequently.

2.1 International market selection methods

The literature on international market selection methods will be discussed in this section under different
categories. In figure 1 this categorisation is illustrated.

Figure 1: Categorisation of the international market selection literature

INTERNATIONAL MARKET SELECTION METHODS

QUALITATIVE APPROACHES QUANTITATIVE APPROACHES

Market Grouping Methods Market Estimation Methods

Firm-level Country-
level

Source: Own figure constructed from Papadopoulos and Denis (1988:38-51)

Papadopoulos and Denis (1988:38-51) summarised the literature on international market selection methods
up until the late 1980s. They classified over 40 proposed international market selection models into two broad
types of approaches – qualitative approaches (rigorous and systematic gathering and analysis of qualitative
information about one or a handful of potential country markets) and quantitative approaches (analysing large
amounts of secondary statistical data about many or all foreign markets).

2.1.1 Qualitative approaches

According to Papadopoulos and Denis (1988:39) most qualitative approaches typically start with identifying a
short list of countries for further consideration by establishing objectives and constraints for exporting a
specific product to each country under consideration. Other studies focus more on the nature,
appropriateness and sources of qualitative information that could be used in the international market
selection process. These sources include government agencies, chambers of commerce, banks, distributors,
customers, international experts and foreign market visits (Pezeshkpur, 1979). Papadopoulos and Denis
(1988:39) suggest that pure qualitative approaches to international market selection could be seen as biased as
they are based on perceptions and are largely inaccurate.

3
 

Douglas et al. (1982:27) stated that the biggest challenge in international market selection is the large
number of countries throughout the world that need to be analysed. They suggest that a screening procedure
of secondary data be used to determine which countries to investigate in depth. Quantitative approaches to
international market selection do exactly this by analysing and comparing secondary data of a large number of
countries and will be discussed subsequently.

2.1.2 Quantitative approaches

Papadopoulos and Denis (1988:39) further divided quantitative approaches into two categories, namely market
grouping methods and market estimation methods. Market grouping methods cluster countries on the basis of
similarity while market estimation models evaluate market potential on firm or country level (see figure 1).

2.1.2.1 Market grouping methods

Studies undertaken to attempt market grouping have been summarised by Papadopoulos and Denis (1988: 39-
41), Steenkamp and Ter Hofstede (2002:185-213) and Shankarmahesh et al. (2005:204-206). These methods
are based on the assumption that the most attractive markets for a firm are the ones that most closely
resemble the markets it has already penetrated successfully (Papadopoulos and Denis, 1988:41). By providing
insight into structural similarities, these methods enable firms to standardise their offerings and marketing
strategies across markets (Sakarya et al., 2007:213). Countries are clustered based on similarities in social,
economic and political indicators while demand levels are, for the most part, not taken into account (Sakarya
et al., 2007:212). Market grouping methods are mostly criticised for relying exclusively on general country
indicators, rather than on product-specific market indicators, as macro indicators may not reflect market
development for a product (Sakarya et al., 2007:212; Kumar et al., 1994:31; Papadopoulos and Denis, 1988:41).
Studies that attempted to include more product-specific information faced the problem of insufficient data,
and are limited to the product ranges of a particular firm. Thus, they cannot be applied to all possible product
groups (Papadopoulos and Denis, 1988:41, 47). Sakarya et al. (2007:212) also argued that grouping methods
fail to take into account similarities among groups of consumers across national boundaries. Furthermore,
focusing only on countries with similar characteristics to markets already penetrated may hold the risk of
overlooking lucrative opportunities in countries with other characteristics (Kumar et al., 1994:32).
Referring to the abovementioned limitations, market-grouping methods will not be suitable to identify
export opportunities for a country if the trade promotion organisation or researcher needs to consider all
possible country-product combinations worldwide.
Market estimation methods will therefore be subsequently investigated in order to establish if the
international market selection method best suited to the identification of potential export opportunities for
South Africa can be found within this classification of international market selection methods.

2.1.2.2 Market estimation methods

Market estimation models evaluate foreign markets on the basis of several criteria that measure aggregate market
potential and attractiveness (Sakarya et al., 2007:212; Papadopoulos and Denis, 1988:41). The criteria vary
across methods and often include wealth, size, growth, competition and access indicators (Sakarya et al.,
2007:212). Papadopoulos and Denis (1988:40-47) summarised the different methods of measuring market
potential that were introduced up until the late 1980s and included multiple factor indices, regression analyses
and multiple criteria import demand estimations. Papadopoulos and Denis (1988:40-47) found that common

4
 

shortcomings of these methods include the lack of product specificity, the assumption of a static
environment and methodological problems due to data availability.
Henceforth, the more recent literature on market estimation methods will be discussed in detail. Most of
these methods are based on, and address, the methodological shortcomings of earlier studies (see
Papadopoulos and Denis (1988:40-47) for a discussion of these earlier studies).
For the purposes of this study, the literature on market estimation methods will be categorised into firm-
level and country-level methods (see figure 1). Firm-level methods can be applied by firms to identify markets for
their limited product ranges. These methods usually include an analysis of the firm’s objectives, profitability,
managerial experience and knowledge, customer standards and attitudes and product adaptation requirements
when identifying potential export markets. Country-level methods, on the other hand, can be applied by a
country’s export promotion agency to identify the most promising country-product combinations to focus
their export promotion efforts on. Criteria and data used in these methods should be product-specific,
applicable to many country-product combinations and generally available. These criteria might include
product-specific market growth, market size, level of competition and barriers to trade.

2.1.2.2.1 Firm level market estimation methods

Firm level market estimation methods include the studies of Davidson (1983)4, Cavusgil (1985)4, Ayal and Zif
(1978)4, Kumar et al. (1993), Hoffman (1997), Andersen and Strandskov (1998), Brewer (2001), Andersen and
Buvik (2002), Rahman (2003), Alon (2004), Ozorhon et al. (2006), and others.
Given that the purpose of this study is to identify the most appropriate country-level international market
selection method to apply to South Africa (see section 1), firm-level market estimation methods will not be
discussed in much detail. It is, however, important to note that Cavusgil (1985:30-31) and Kumar et al.
(1994:33-34) suggest that the process of evaluating the export potential of a foreign market involves the
following three stages: (1) a preliminary screening stage to select more attractive countries to investigate in
detail, based on countries’ demographic, political, economic and social environments; (2) an in-depth
screening stage in which product potential (market size and growth), competitors, market access, and other
market factors are analysed; and (3) a final selection stage that involves the analysis of company sales
potential, profitability and product adaptation to the firm’s existing portfolio. This process forms the basis of
many firm-level market estimation models and cannot be used in this exact form when identifying export
opportunities for a country due to the final selection stage that includes subjective, firm-specific variables.
With the purpose of this study being partly to find the best suited country-level international market
selection method, the focus will subsequently be on studies in this category.

2.1.2.2.2 Country-level market estimation methods

On first review, the methods of Green and Allaway (1985), Russow and Okoroafo (1996) and Papadopoulos
et al. (2002), although applied to a limited number of countries and products, seemed to be applicable for
screening a wide range of country-product combinations and are therefore categorised under country-level
market estimation methods. These methods are discussed in sections 2.1.2.2.2.1 to 2.1.12.2.2.3.
Papadopoulos and Denis (1988:43) mentioned a multiple criteria method proposed by the International
Trade Centre (ITC) to assist developing countries that want to extend exports in identifying potential export

4 Although these three studies were conducted before 1988, it was not included in Papadopoulos and Denis’ (1988:40-47) summary of
the international market selection literature, and is therefore included here.

5
 

markets. On a research visit to the ITC in Geneva in September 2008, the researchers found that the ITC is
still using a similar method to assist developing countries in identifying potential export markets. This method
can also be classified as country-level market estimation model and is discussed in section 2.1.2.2.2.4.
Three other studies that can be classified under country-level market estimation models are the studies of
Arnold and Quelsh (1998:7-20), Cavusgil (1997:87-91) and Sakarya et al. (2007:208-238). They all attempted to
assess export opportunities in emerging markets specifically, as discussed in further detail in section
2.1.2.2.2.5.
Another method that was specifically designed to be applied on a country-level is the decision support
model proposed by Cuyvers et al. (1995). This model was designed to screen all possible worldwide country-
product combinations to identify potential export opportunities for Belgium5. Cuyvers (2004) adapted and
applied this model to Thailand as the exporting country. This method is discussed in section 2.1.2.2.2.6.

2.1.2.2.2.1. Green and Allaway’s shift-share model

Green and Allaway’s (1985) shift-share approach to identify export opportunities were described by Douglas
and Craig (1992) as the only new approach to international market selection that had been proposed up until
the early 1990s. They used 20 OECD countries and 51 high-technology products6 (at the SITC four-digit
level) and the period 1974 to 1979 in their analysis.
Shift-share analysis identifies growth differentials based upon the changes that have occurred in market
shares over time. It requires import data of the countries under investigation for the products in question at
the beginning and end of the period of analysis. An expected growth figure is calculated for each country-
product combination based on the average growth of all combinations included in the analysis. The
difference between each market’s actual and expected growth is called the net shift and will be positive for
markets that gained market share over the period of analysis and negative for those that lost market share.
The net shift is therefore the difference between a market’s actual performance and the performance it would
have had if its growth rate had been equal to the average growth of the entire group of markets included in
the analysis (Green and Allaway, 1985:84).
Furthermore, the percentage net shift is calculated by dividing the net shift of each market under
investigation by the total net shift of all the markets included in the analysis and multiplying it by 100 (Green
and Allaway, 1985:85). The figure thus obtained provides the total gain or loss of market share accounted for
by each member of the group7.
Green and Allaway (1985:87) identified a few shortcomings to their analysis. These include that the
timeframe of the analysis is based on only two points in time. Moreover, shift-share analyses identify only
relative opportunities.

5 Although Shankarmahesh et al. (2005:205) classified Cuyvers et al.’s (1995) decision support model as a market segmentation /
grouping method, the authors use mostly market estimation techniques in their analysis. Market estimation is used in filters 1 to 3
and, in filter four, the identified export opportunities (country-product combinations) are classified according to market size and
growth, and the exporting countries’ current position in the different markets. No geographical or demographical grouping based
on similar country characteristics has been done. This method therefore falls under market estimation methods.

6 These products are defined by Green and Allaway (1985:85) as individual product categories possessing a high level of technological
input. Specific products are not specified in the article.

7 For a step-wise mathematical description of the shift-share methodology, see Papadopoulos et al. (2002:186-190) and Huff and
Scherr (1967).

6
 

Papadopoulos et al.’s (2002:168-169) specifically reviewed Green and Allaway’s (1985) shift-share model as
it seemed to address all the shortcomings of the international market selection models that they have
reviewed in their study. According to Papadopoulos et al. (2002:168) the core strength of the shift-share
approach is that it is simple and industry-specific. Its main weakness, on first review, is that it is limited to
import-only measures. When Papadopoulos et al. (2002:168) investigated the theoretical foundations of the
shift-share approach, they noted that other authors who applied the shift-share approach in the field of
marketing found the results to be biased, depending on the base years chosen, and to fluctuate greatly due to
outliers. Papadopoulos et al. (2002:168-169) subsequently tested the shift-share approach themselves by
performing the shift-share approach for three products and 50 importing countries. They found that one
country might perform very promisingly at one point in time but very poorly in subsequent years. They also
found that the rankings identified by the model are volatile and that the simple growth model rankings were
highly correlated to the shift-share rankings. Papadopoulos et al. (2002:169) concluded that the shift-share
approach lacked predictive power and that it was redundant, given the high correlation between the results
and those that would be obtained from the simple growth model.
Based on these findings and the fact that no other indicators except import growth are considered, the
shift-share approach does not seem appropriate for application in this study.

2.1.2.2.2.2 Russow and Okoroafo’s global screening model

Russow and Okoroafo (1996:52) used six randomly selected products and 192 countries around the world in
their analysis. From the international business theory and market screening and assessment literature, Russow
and Okoroafo (1996:50) identified three screening criteria, namely product-specific market size and growth,
factors of production and economic development. The variables used to measure market size and growth
include domestic production, imports, exports, shift-share of domestic production, shift-share of imports and
shift-share of exports of a specific product. The cost and availability of factors of production was captured by
gross fixed capital formation, money supply, total international reserves, total population, unemployment rate,
average hourly wages in manufacturing, country area and population density. The level of economic
development was measured by gross domestic product, gross domestic product per capita, agriculture as a
percentage of GDP, the contribution of manufacturing industries as a percentage of GDP, construction as a
percentage of GDP, wholesale and retail trade as a percentage of GDP and transportation and
communication as a percentage of GDP (Russow and Okoroafo, 1996:52).
A principal components analysis8 was used for every product included separately in the analysis to
determine whether the 21 variables mentioned above are interrelated. After performing the principal
components analysis for “calculators” (as an example product), seven factors were identified to be used in the
screening model. A cluster analysis was consequently conducted to group countries with similar9 market

8 Principle components analysis is a technique for identifying groups or clusters of variables. The technique is used to understand the
structure of a set of variables or to reduce a data set to a more manageable size without compromising on the original information
in the dataset (Field, 2005). Principle components analysis involves measuring the correlation between variables in order to
transform a number of correlated variables into a smaller amount of uncorrelated variables called principle components (Fields,
2005).

9 For instance, in the case of “calculators” the seven factors that were defined by means of the principle components analysis include
market size, economic development, market size growth, trade, population density, capital spending and infrastructure maintenance
and development. Countries which are similar based on these seven factors are therefore considered to have similar market
potential for “calculators”.

7
 

potential for a specific product. Each country group was then classified as having a high, medium or low
market potential10 for the product in question (Russow and Okoroafo, 1996:55-58).
Due to the fact that, in this method, a principle components analysis is performed for each product
category separately, the application thereof to over 900 four-digit level, or over 5000 HS six-digit level,
product categories would be very extensive and time consuming for research purposes. The availability of
data on, inter alia, factors of domestic production for a large array of country-product combinations would
also be problematic. This method can be used more realistically when a limited number of products have
been identified for further analysis. This method will therefore not be considered for application in this study,
although elements of the model – such as the criteria identified to measure market potential – can be very
useful.

2.1.2.2.2.3 Papadopoulos et al.’s trade-off model

According to Papadopoulos et al. (2002:169) the international market selection theory suggests that both the
pluses and minuses of the countries under review must be considered in order to make effective decisions.
They expressed these trade-offs as the demand potential (plus) and trade barriers (minus) in the countries
under review. They state that many researchers identify trade barriers as the most important deterrent to
exports, but most have not accounted for it in their international market selection models. This was probably
due to the difficulty in quantifying non-tariff barriers and most authors assumed that non-tariff barriers would
be dealt with in later stages of the internationalisation process where in-depth market analyses are conducted
(Papadopoulos et al., 2002:170). Papadopoulos et al. (2002:170) also accounted for the firm’s strategic
orientation to guide the weighting of the constructs in their model. Papadopoulos et al.’s (2002) model is
illustrated in figure 2.

Figure 2: Papadopoulos et al.’s (2002) trade-off model

Demand Potential Trade barriers
• Apparent consumption • Tariff barriers
• Import penetration • Non-tariff barriers
• Origin advantage • Geographic distance
• Market similarity • Exchange rate

Strategy
Defensive vs. Offensive

International Market Selection

Source: Papadopoulos et al. (2002 :170, figure 1).

Four variables were used for each of the two main constructs (demand potential and trade barriers). These
variables were chosen based on relevance, frequency of use in past research, evidence of satisfactory
performance in various settings, data availability, reliability, comparability and ability to express qualitative

10 The determination of the thresholds for the classification of countries into high, medium or low market potential is not specified by
Russow and Okoroafo (1996).

8
  

factors where necessary (Papadopoulos et al., 2002:170-171). The variables and their measures are
summarized in table 1.

Table 1: Papadopoulos et al.’s (2002) trade-off model – variables and measures*

Demand Potential Trade barriers
Variable 1: Apparent Consumption = Domestic production plus imports Variable 1: Tariff Barriers = Weighted mean annual tariff rate over the study
minus exports period.
Import data do not portray the total available market. This measure Tariffs have a direct effect on the exporter’s prices and pricing strategy
for apparent consumption is considered to be the appropriate discretion.
reflection of true market size in a given industry.
Variable 2: Import Penetration = Imports as % of apparent consumption. Variable 2: Non-tariff barriers = Composite quantitative index of 20 barrier
items
This measure is widely used in industry-specific analyses. A high ratio Non-tariff restrictions are often a more important obstacle to exporting than
means import market openness and low domestic producer tariffs. Papadopoulos et al. (2002:172) developed an index consisting of all
competitiveness, signalling an attractive market. 20 barrier items in the World Trade Organisation’s Trade Policy Review.
Each item was weighted based on its frequency of occurrence in the target
countries. WTO data was used.
Variable 3: Origin Advantage = Exporting country’s share in target Variable 3: Geographic Distance = Mileage distance between exporting and
market’s total imports target countries
A high overall share indicates that the exporting country has the According to Papadopoulos et al. (2002:171) distance is directly related to
benefits of critical mass, favourable image in the importing market and transport costs and affects export price. Distance between countries’ main
strong trade relations between the importing and exporting countries. ports was used (if no port, the capital or next closest major city was used).
Variable 4: Market Similarity = Overall score of four indicators, namely Variable 4: Exchange Rate = Percent change in official exchange rate vs. previous
health and education, personal consumption, production and transportation and year
trade.
According to Papadopoulos et al. (2002:171) demand tends to be According to Papadopoulos et al. (2002:171) volatile exchange rates between
higher in markets similar to the market in which a product was initially the exporting and importing countries’ currencies is a major risk element in
developed. exporting and can have a big impact on pricing and strategy.
Sethi (1971) proposed 29 indicators of market similarity that were
grouped in the above-mentioned four categories. Papadopoulos et al.
(2002:171) used the indicator in each group with the highest
correlation to the others in the group to measure the four indicators in
their market similarity score. These were:
• for health and education: life expectancy.
• for personal consumption: GNP per capita
• for production & transportation: electricity production and
• for trade: imports-to-GDP ratio.
Note : *) the variables used are all at one point in time
Source: Papadopoulos et al. (2002:170-171, Exhibit 1).

Seventeen OECD countries were chosen as the target (importing) countries due to data availability and
similarities among these developed countries. Two different countries were chosen to be the exporting
countries, namely Canada, which is a highly-developed country and an experienced exporter, and China,
which has the world’s largest population and is in its earlier stages of internationalisation. Papadopoulos et al.
(2002:173) argued that a major weakness in earlier international market selection models was that, when
screening markets, they focused on the importing countries only, without considering the identity of the
exporting country. They thus chose two greatly different exporting countries in their analysis to test the
effects of taking into account who the exporting country is.
Three products were chosen, namely aircraft (representing industrial goods), furniture (representing
consumer durables) and beverages (representing consumer non-durables)11. Two- and three-digit SITC
(Standard International Trade Classification) data were used.
Papadopoulos et al. (2002:173) stated that there is no clear guidance in the literature as to the length of the
period between when an observation is made about a country and when it is reacted upon. Papadopoulos et
al. (2002:173) chose the six-year period 1989 to 1994 with 1988 as the base year.

11 The basis for choosing aircraft, furniture and beverages to represent industrial goods, consumer durables and consumer non-
durables respectively, are not specified by Papadopoulos et al. (2002).

9
 

The data for each variable was scaled by subtracting the lowest country value from the highest and dividing
the difference by 10. Therefore 10 equal scale intervals were formed and each country could be assigned a
score from 0 to 10. Each country’s scores for each variable were averaged to get a total score for each of the
demand potential and trade barrier dimensions. High scores represented high demand potential and low trade
barriers. Countries were subsequently plotted in a two-dimensional matrix illustrated in figure 3.

Figure 3: Two-dimensional matrix for plotting countries in Papadopoulos et al.’s (2002) trade-off model
High demand potential / High demand potential /
High trade barriers Low trade barriers

Low demand potential / Low demand potential /

High trade barriers Low trade barriers

Source: Papadopoulos et al. (2002 : 174, figure 2).

Target markets in the upper right quadrant (high demand potential / low trade barriers) would offer the best
export opportunities.
As many users would prefer to rank countries on a single overall score, Papadopoulos et al. (2002:174-175)
assigned weights based on firm strategy to develop total score country attractiveness scales that combine the
two dimensions. If a firm has a defensive strategy it will focus more on markets that are easier to penetrate
and high trade barriers would carry a bigger weight. On the other hand, if a firm has an offensive strategy it
will focus on markets with high demand potential, even though it may take more effort to penetrate those
markets. Weighted scores for each of the two dimensions were then added to generate an overall score for
each country.
Papadopoulos et al. (2002:183) identified a few limitations of their model. These include deficiencies of
secondary data; the lack of direct conversion schemes between the trade coding systems12; unavailability,
unreliability and aging of data for some countries (particularly less-developed countries) and the lack of
greater product-specificity13.
Papadopoulos et al. (2002:184) attempted to address as many of the limitations identified in previous
studies as possible. They stated that their model provided a significant improvement over earlier ones by
capturing total rather than import-only demand, because it is industry-specific and was tested using three
products, 17 importing countries and two very different exporting countries (Papadopoulos et al., 2002:184).
When considering the application of this model for the purposes of this study where all countries are
included as possible export markets, a few possible problems can be raised. When dealing with a large array of
possible country-product combinations (over 200,00014 when four-digit level product data is used and over
one million15 on a six-digit level) domestic production figures for all of these products would be difficult to

12 Papdopoulos et al. (2002:183) stated that the SITC codes alone have been revised three times since 1965 and some countries still use
earlier versions to report trade data.

13 Papadopoulos et al. (2002) used the SITC two- and three-digit level product classification in their analysis. This is rather aggregated
product classifications. For example, the three-digit level SITC code 001 represents “live animals”, while the four-digit SITC codes
0011 represents “animals of the bovine species, including buffalo“; 0012 represents “live sheep and goats”; 0013 represents “live
swine”; 0014 represents “live poultry”; and 0015 represents “live horses, asses, mules and hinnies”. It is clear that the four-digit
product classifications are more specific.

14 240 countries in the world x 986 SITC four-digit product groups = 236 640

15 240 countries in the world x 5407 HS six-digit product groups = 1 297 680

10
 

collect, especially in the least-developed countries. The same applies for data availability of non-tariff barriers
and consumption figures per country-product combination. Papadopoulos et al.’s (2002) analysis is also
conducted at the product level. In other words, a list of priority countries can be identified per product (as in
Russow and Okoroafo (1996) (see section 2.1.2.2.2.2)). Again, the vast number of products dealt with poses
the problem of the application of this method being a very extensive and time-consuming exercise. This
method would, however, be very useful when a limited number of products have already been identified and
a list of priority countries has to be identified for each of these.
Papadopoulos et al.’s (2002) trade-off model will therefore not be used for the purposes of this study,
although the proposed criteria and rationale of their market potential analysis can be very useful.

2.1.2.2.2.4 The ITC’s multiple criteria method

One of the aims of the International Trade Centre (UNCTAD / WTO) is to assist developing countries that
want to extend and diversify their exports of products which are critical for their future growth and
development, as well as to effectively focus their trade promotion efforts (Freudenberg, 2006). They do this
by using a multiple criteria method to assess the export potential of these countries, thereby identifying those
sectors with the highest potential for future exports (Freudenberg, 2006).
The ITC measures the export potential of a specific product group as (Freudenberg and Paulmier, 2005a:
10-11; Freudenberg and Paulmier, 2005b: 8, Freudenberg et al., 2007:2; Freudenberg et al., 2008:11-12):
• The current export performance of the exporting country (index 1), evaluated by indicators such as
its exports of the product in value, the world market share, the growth rates of exports of the
product, net exports to the world and the relative trade balance;
• The domestic supply capacity (index 2), evaluated by a survey of companies questioning the quality
of products and the efficiency of supporting industries; and
• The characteristics of the international environment (index 3), evaluated by indicators such as size
and growth of world demand and the exporting country’s access conditions to international markets.

An export potential index is ultimately calculated for each product group under investigation, using the
abovementioned variables. The different variables are first standardised (due to the fact that the different
variables are in different units – dollar, value, % per year etc.) before they are aggregated into the composite
index. To standardise the variables, the following formula is used:

100 x (Value - Lower limit) / (Upper limit - Lower limit)

This provides a score ranging from 0 (weak performance) to 100 (best performance) for each variable. The
best performing 5% of products define the upper limit and the weakest performing 5% of products define
the lower limit for each variable. For reasons of simplicity, all variables are given the same weights within the
indexes (index 1 to index 3) and the three indexes are again weighted equally when deriving the overall index
(Freudenberg and Paulier, 2005a: 34).
According to (Freudenberg and Paulmier, 2005a: 36) the limitations of the ITC’s multiple criteria method
include that composite indexes measure only that which can be quantified and for which there are data
available; the selected variables give only a snapshot at one moment in time; growth variables are backward-
looking; weighting of the different variables is difficult to establish; and rankings should be interpreted with
caution, especially when differences between the respective indices for products are small.
To reach the objectives of this study, all possible world-wide country-product combinations must be
considered and, ultimately, a limited list of the country-product combinations with the highest export

11
 

potential for the exporting country should be produced in order for an export promotion agency to use its
resources optimally. The ITC’s export potential assessment studies discussed in this section unfortunately
concentrate only on identifying product groups/industries in which the country under investigation have
potential for future exports. The focus is not on demand for the product group in different countries
respectively, but rather on total world demand. Therefore only a ranking of product groups is done, rather
than a ranking or list of country-product combinations that hold potential for future exports.
It does, however, seem possible to apply the ITC’s method on a country-product level, but it would be a
time-consuming exercise to conduct it for all possible country-product combinations around the world. As is
the case with Russow and Okoroafo’s (1996) and Papadopoulos et al.’s (2002) models, the analysis is extensive
because no elimination process is in place and each of the more than 200,000 (when four-digit level product
data is used) and over one million (when six-digit level data is used) possible country-product combinations
should be analysed individually. When a limited list of country-product combinations is considered, the ITC’s
methodology would be very useful to rank these in terms of export potential.

2.1.2.2.2.5 Assessment of export opportunities in emerging markets

As mentioned earlier, Cavusgil (1997:87-91), Arnold and Quelsh (1998:7-20) and Sakarya et al. (2007:208-238)
all attempted to assess export opportunities in emerging markets specifically. They argue that traditional
market selection analyses fail to account for the dynamism and future potential of emerging markets (Sakarya
et al., 2007:208). Cavusgil (1997:87-91) attempted to rank the total market potential of only 25 emerging
countries. Only country-level (macro) indicators were used and no product specificity was introduced.
Arnold and Quelsh (1998:7-20) proposed a foreign market assessment framework that included three
elements, namely assessing long-term market potential (using population and GDP, thus country-level
measures), identifying business prospects (product-level assessment, companies must identify their own
indicators for assessing demand for their product) and predicting potential profits (assessing concentration of
population in urban centres versus rural villages, the distribution of wealth, telecommunications
infrastructure, penetration of key consumer durables such as telephones, televisions or cars, etc.). Because
Arnold and Quelsh’s model uses only macro-level indicators to assess market potential and then concentrates
on firm-level assessments (which are mostly situation specific, qualitative and not applicable in a model that
assesses a large array of country-product combinations), this method is not discussed in further detail and is
not applicable for the purposes of this study.
Sakarya et al. (2007:209) introduces long-term market potential (from Arnold and Quelch’s model), cultural
distance, competitive strength of the industry, and customer receptiveness as criteria for assessing emerging
markets as candidates for international expansion. Their proposed model was applied to the United States of
America as the exporting country, Turkey as the importing country and apparel as a product/industry.
Because Sakarya et al.’s (2007) model includes an in-depth, situation-specific assessment of a particular
country-product combination that requires information not readily available for a large array of country-
product combinations (such as social and moral values of consumers, wages in the industry, consumer choice
opportunities, product quality, appeal of sales promotions, level of customer service, etc.), this model cannot
be considered for the purposes of this study either.

2.1.2.2.2.6 Cuyvers’ decision support model

The basic ideas of Walvoord (1983) were used by Cuyvers et al. (1995:173-186) to construct a decision
support model (DSM) for a Belgian government export promotion institution to provide them with a limited

12
 

list of realistic export opportunities to which they could devote their limited financial resources. This DSM
was then refined and applied for Thailand in 2004 (Cuyvers, 2004:255-278).
The basic decision support model used to identify realistic export opportunities for a particular country,
starts from the assumption that all world markets hold potential export opportunities for a particular country
and, therefore, all possible country-product combinations enter the filtering process (Cuyvers, 2004:256).
After every filter, a number of opportunities are rendered uninteresting and are not considered in subsequent
filters.
In filter one, countries that hold too high a political and/or commercial risk and do not show adequate
macro-economic size or growth are eliminated. The rationale for this is that, with the 240 countries of the
world as a starting point, filter one enables the researchers to eliminate uninteresting markets in order to
concentrate in detail on a more limited set of preliminary opportunities.
In filter two, a more specific assessment of the various product groups for the remaining countries is done
to identify the market potential of each possible country-product combination (market). The main purpose
of this filter is therefore to eliminate markets that do not show sufficient demand potential. The main criteria
that are used in this filter are the growth rate of imports of a given product group by a given country (import
growth) and the value of imports of a given product group by a given country (import market size).
Three variables are calculated for each market, namely, short-term import growth, long-term import
growth and import market size. Short-term import growth is considered to be the most recent available
simple annual growth rate in imports, while long-term growth is calculated as the average annual percentage
growth in imports over a period of five years. Finally, the relative import market size is calculated as the ratio
of imports of country i for product group j and the total imports of all countries that entered filter two of
product group j (Cuyvers et al., 1995:178; Cuyvers, 2004:259-260).
In filter three, trade restrictions and other barriers to entry are considered to further screen the remaining
possible export opportunities. Two categories of barriers are considered in this filter, namely, the degree of
market concentration (competitor analysis) and trade restrictions (market accessibility).
In the last stage of the analysis (filter four), the export opportunities (country/product combinations)
identified in filters one to three, are categorised according to relative market importance and relative market
size and growth (Cuyvers, 2004:267).
One of the main benefits of the DSM is that it provides a tool to assist export promotion authorities to
decide how to allocate their scarce resources to export promotion activities in various markets. It also
provides information on export markets and export promotion efficiency to derive appropriate actions in
relevant export markets (Cuyvers et al., 1995:174). The DSM further provides export promotion agencies with
a limited list of export promotion priorities, based on measurable and objective economic data, and draws
attention to markets that have not previously been recognised as potential export markets (Cuyvers et al.,
1995:174).
Despite the abovementioned benefits of using the DSM to identify realistic export opportunities in a
country, Cuyvers et al. (1995:174) warn that it would be unwise to rest all export promotion decisions upon
the model alone. Other considerations, such as feedback from foreign trade offices (on the demand side of
exports) and export councils (on the supply side), should also be taken into consideration. Diplomatic and
political issues would also lead to government supporting exports to a particular country, even though it
might be identified by the DSM as an economically promising market (Cuyvers et al., 1995:175).
Export promotion is, furthermore, an activity that is very often only effective in the long run. Since the
DSM’s scope is more short term and based on historical data, some export opportunities that are considered
by the model as sub-optimal, might be good opportunities in the long run (Cuyvers et al., 1995:174).
Therefore, basing export promotion decisions only on the results of the DSM could also lead to missed

13
 

opportunities. Cuyvers et al. (1995:174) also state that it is important to keep in mind that the purpose of the
model is not to provide a ranking of export opportunities, but rather a list of choices of interesting markets
grouped into categories reflecting market size, market growth and market importance.
When considering the application of this model for the purposes of this study, it seems that the DSM
conforms to the prerequisites that all possible world-wide country-product combinations must be considered
and that a limited list of the country-product combinations with the highest export potential for the exporting
country should be produced in order for an export promotion agency to use their resources optimally. It also
seems to be capable of handling a large array of possible country-product combinations due to the filtering
process used. The DSM can also provide a list of priority products in each country and, vice versa, a list of
priority countries for each product. For an export promotion agency, these lists of priority products for each
specific market would be very useful.
As it seems that the DSM is the best-suited model for identifying export opportunities for South Africa,
and that most of its limitations could be overcome, the DSM will be refined and applied in this study.

2.2 Summary and method selected

The aim of this section was to evaluate the international market selection literature to find the method best-
suited to the identification of potential export opportunities for a given exporting country (in this case, South
Africa).
The literature was classified into various categories of studies (see figure 1) and the focus of this study was
established to be on country-level market estimation models. Six country-level market estimation models
could be found in the literature and are discussed in sections 2.2.1 to 2.2.6. The benefits and limitations of
each of these methods or models were discussed and each method was evaluated for application in this study.
The decision support model (DSM) of Cuyvers et al. (1995:173-186) and Cuyvers (2004:255-278) was
found to be the model best-suited to the purposes of this study. The basic methodology of the DSM and the
results of the application of the DSM to South Africa are discussed in Section 3.

14
 

3. THE SOUTH AFRICAN APPROACH

This section demonstrates the ways in which a decision support model, incorporating various adaptations for
the South African trade environment, can be used to more successfully identify realistic export opportunities.
The following explanations are therefore an extension of the work of Cuyvers et al. (1995) and Cuyvers (2004)
who developed and applied a decision support model for Belgium and Thailand.

3.1 A decision support model for planning export promotion activities

The decision support model adapted for South Africa also starts from the assumption that, in principle, all
world markets (i.e., the markets for all products in all countries) are potential markets for the exporters of the
given exporting country, and all markets should therefore enter a screening procedure. The unit of analysis is
the country-product combination. The analytical framework of the model is based on the model of
international market research proposed by Walvoord (1983), in which relevant information on markets is fed
through a screening process of four consecutive filters, with the result that less interesting market
opportunities are identified and deleted from the list (see section 2.1.2.2.2.6).
In the following sections the method used and results for each filter of the South African application and
refinement of the DSM is discussed.

3.2 Filter one: Which countries show preliminary export opportunities for South African products?

As pointed out in section 2.1.2.2.2.6, the aim of the first step in our analysis is to determine which countries
merit closer investigation as potential markets. The criteria used here are relatively low commercial and
political risks, together with total market potential as measured by macro-economic growth and/or the size of
the economy.
The commercial and political risks involved in doing business with foreign countries can be assessed using
parameters such as the current-account deficit as a percentage of GDP, the external debt service as a
percentage of export earnings, the stock of foreign debts of a country in proportion to its GDP, etc., as well
as past and future changes in these parameters (Cuyvers, 2004:258). This information is available through the
International Monetary Fund and other international organisations. In addition, some academic and private
organisations publish such information, based on commercial and political risk assessments by foreign
business people16. The assessment of the political risk of a country usually involves analysing inter alia the
system of government (e.g. amount of state control over international trade and investment activities), history
of political instability, evidence of corruption in political and financial circles, economic policies, rule of law
and the legal system (ITRISA, 2009:16-17). Commercial risk assessments generally involve analysing the
financial strength of buyers in a particular country (therefore assessing their ability and willingness to pay). It
is often easier to acquire information for countries rather than individual buyers. Therefore, foreign buyers’
general ability and willingness to pay are assessed by using the levels of insolvencies or bankruptcies in a
country, the degree to which an economy is dependent on foreign aid, the level of unemployment and social
unrest, the average per capita income, the external debt repayment record, history of balance of payments
deficits and exchange controls, history of imposing sanctions and foreign exchange reserves (ITRISA,
2009:16-17).

16 See http://www.countryrisk.com

15
  

In the application of the DSM for both Belgium (Cuyvers et al., 1995) and Thailand (Cuyvers, 2004), the
country risk ratings of the Belgian public credit insurance agency, Office National du Ducroire (ONDD) were
used in this part of filter 1. The ONDD’s ratings conform to the OECD’s Arrangement on Guidelines for
Officially Supported Export Credits17 and are not conducted from the point of view of a specific exporting
country. It can therefore be used by any exporter that wants to establish the degree of risk involved in
dealing with a specific country. Therefore the country risk ratings of the ONDD were used in this study.
The ONDD provides risk assessment on export transactions in terms of political risk in the short-,
medium-, and long-term, as well as the commercial risk of the country. From these country risk ratings a
country risk score is calculated. The country risk score is used to determine whether or not a country should
be further investigated as a potential export market.
The ONDD political risk rating rates countries on a scale of 1 to 7, where 1 indicates a low political risk in
a specific category and 7 indicates a high political risk in a specific category (short-, medium-, and long-term)
for the particular country. The commercial risk rating differs from that of the political risk rating. The
commercial risk rating is presented either as an A, a B, or a C, where A indicates that the country is
experiencing low commercial risk and C indicates that the country is experiencing a high commercial risk. The
three political risk ratings are transformed from a 1 to 7 scale to a 1 to 10 scale, whereas the commercial risk
country rating is transformed in such a manner that A represents 3.33, B represents 6.67 and C represents 10
(Cuyvers, 2004:256). This transformation is necessary to construct a country risk score. A compounded
country risk score is calculated from the risk ratings, namely short-, medium- and long-term political as well
as commercial risks. The country risk score is used to determine a critical value to eliminate less interesting
export markets from the model.
To illustrate the process, consider Country X with the following political and commercial risk ratings as an
example.

Table 2: Country X’s risk ratings

Political Risk: Political Risk: Political Risk: Commercial Risk
short term medium term long term
Country X 3 4 5 B
Source of data: excerpt from the DSM for South Africa.

In order to construct the country risk score, the country risk ratings should be transformed as discussed in
the previous paragraph. The transformed country risk rating for country X is given as:

Table 3: Country X’s transformed risk ratings

Political Risk: Political Risk: Political Risk: Commercial Risk
short term medium term long term
Country X 4.29 5.71 7.14 6.67
Source of data: excerpt from the DSM for South Africa.

In order to obtain a country risk score for a particular country an equally weighted index is constructed from
the country risk ratings in terms of political risk and commercial risk of the specific country under
investigation. In terms of the example of Country X, the country risk score is 6.19.
When a particular country’s risk score exceeds the critical value of 9.286, this country should not be
included in the further analysis of potential export markets for South Africa. Country X considered in the
example would be included in the further analysis of potential export markets because its risk score of 6.19 is
below 9.286.

17 For more information see Cutts and West, 1998:12-14; Moravcsik, 1989:173-205

16
 

Twenty-one countries (out of an original 240), belonging to the two highest credit risk groups of the
ONDD, were excluded from the analysis, leaving 219 countries18. These countries were excluded due to their
relatively high political and commercial risk ratings that exceeded the critical value of 9.28619.
To identify potential export markets, indicators that give an indication of whether the particular markets
are large enough or show relative growth should be employed. GNP and GNP per capita were chosen in the
Belgian and Thailand studies as a starting point for the filtering process in terms of macro-economic
indicators (Cuyvers et al., 1995:177; Cuyvers, 2004:256). In this study, real GDP and GDP per capita values
were used in filter 1, as well as GDP growth and GDP per capita growth to extend the model to include
countries that show general potential due to economic growth and development.
Data on GDP and per capita GDP between 2002 and 2004 could be collected for 193 of the remaining
219 countries. No, or incomplete, data was available for 26, mostly small, countries such as Andorra, the
Cayman Islands, the Cook Islands, the Faroe Islands, Liechtenstein, Vatican City, etc. In order to select the
more interesting markets from these 193 countries, a cut-off point x is calculated for the GDP and per capita
GDP values, such that:

, (1)

where is the average of X (GDP or per capita GDP), δx is the standard deviation of X, and α is a factor
which is determined in such a way that small changes in its value only marginally affect the number of
countries screened out. The same α-value is chosen for both GDP and GDP per capita. When choosing an α-
value, it is therefore also considered that a comparable number of countries should be selected for both GDP
and GDP per capita within a small range of values for α.

Countries are selected when the following condition applies:

(2) 

(GDP and per capita GDP, respectively, are larger than or equal to the cut-off value) for at least two
consecutive years of the most recent three-year period for which data are available.
Starting from α = 0.001, α is increased consecutively by 0.001 until the number of countries rejected
stabilises. An alpha value of α = 0.056 was chosen to calculate the cut-off value in this filter, in which case a
total of 95 countries meet condition (2) for GDP and/or per capita GDP. This number is the union of the
two sets of 62 and 52 countries selected on the basis of GDP and per capita GDP20. Among the countries
which do not fulfil condition (2) are some central and Eastern European countries (such as Belarus,
Macedonia, Croatia and Serbia and Montenegro), some less-developed Asian countries (such as Papua New

18 The countries eliminated include Afghanistan, Belarus, Burundi, Cambodia, Côte d’Ivoire, the Democratic Republic of the Congo,
Cuba, Eritrea, Guinea, Haiti, Iraq, North Korea, the Lao People's Democratic Republic (Laos), Lebanon, Liberia, Malawi,
Myanmar, Palestine, Rwanda, São Tomé and Príncipe, the Seychelles, Sierra Leone, Somalia, Sudan, Suriname, Tajikistan, and
Zimbabwe.

19 The ONDD’s risk ratings for 2008 were used. 241 countries are rated yearly by the ONDD.

20 The reader is reminded that countries can meet both cut-off values or either value. For example, both China and India are above

the GDP cut-off value but below the per capita GDP threshold. Both are considered for further investigation in filter two because
they are large markets, although their per capita income levels are low.

17
 

Guinea and Jordan), one North-African country (Tunisia), a number of Latin-American countries (Costa
Rica, El Salvador, Paraguay and Uruguay) and some Caribbean islands (such as the Aruba, the Dominican
Republic, Jamaica, and Grenada). Among those countries which pass the test of this first stage are the OECD
countries, the ASEAN countries China, Taiwan, Hong Kong, South Korea, India, New Zealand and
Australia, Argentina, Indonesia, Vietnam, etc.

3.3 Filter two: Detecting possible export opportunities for South Africa

In the next stage of the assessment of South Africa’s export opportunities, data on imports are analysed for
each country selected in the previous section. The data used are at the SITC (revision 2) four-digit level over
the period 2002-200421. However, for Antigua and Barbuda, Aruba, Botswana, Iran, the Isle of Man, the Lao
People’s Democratic Republic (Laos), Lebanon, Namibia, Puerto Rico, San Marino, St. Lucia, and Suriname,
no trade data were available. Therefore, only 85 countries remain for the detection of possible export
opportunities. Based on these data, 83,810 trade figures in total – henceforth called country-product
combinations – will be analysed for the remaining 85 countries, using the growth of imports and the import
market size as criteria, with the purpose of eliminating non-interesting country-product combinations.
As in the previous step (see section 3.2), cut-off points are calculated for each product group at the SITC
four-digit level. The cut-off points used also take into account whether or not South Africa is relatively
specialised in the respective products, as the so-called “revealed comparative advantage” (RCA) index
indicates22. The rationale is that if South Africa is relatively specialised in a particular product i (RCAi > 1),
one may allow the selection of interesting markets to be less restrictive than if South Africa is not specialised
in it (RCAi ≤ 1).
Evidently, the time period considered for the growth of imports of each product group is important. As
far as short-term growth is concerned, the simple percentage growth rate of the imports of each product
group j in country i is calculated for 2004. Long-term growth, however, stretches over a longer period and is
calculated as the compounded annual growth rate of imports of product group j in country i between 2002
and 2004.
In order to take into account the degree of specialisation in the exports of South Africa of a product
group j, we define a scaling factor sj, following Willeme and Van Steerteghem (1993), such that:23

, (3)

21 The data are from the World Trade Analyzer. http://data.library.ubc.ca/java/jsp/database/production/detail.jsp?id=161

22The RCA index is defined as follows: , where is the exports of the country/world of product group j, and is
the total exports of the country/world of all product groups (see Balassa, 1965).

The properties of the scaling factor are as follows: sj = 2 for RCA = 0, sj = 1 for RCA = 1, sj = 0.848 for RCA = 2 and sj = 0.8 for
23

RCA = ∞.

18
 

Denoting the rate of growth of imports of product group j by country i by gij and total world imports of the
same product group by gworld,j, the cut-off point for import growth of product group j is then

if gworld,j > 0, and

if gworld,j < 0 (see Willeme and Van Steerteghem, 1993:6-7).

Hence, the market in a particular country i for product group j will be deemed sufficiently promising if:

, (4)

This procedure is applied to calculate both short-term and long-term cut-off growth rates. In Table 4, “1”
is reflected in columns 2 and 3 if the condition in equation 4 is fulfilled for both the short and long term.
For market size of country i for product group j the import value of j in i is obviously not taken as a
proxy, but rather the share of this market in the world imports of that product group. This criterion enables
selection of markets that do not show growth, but that are interesting because of their size. Taking into
account the degree of specialisation of South Africa in a particular product group j, the cut-off point for
relative import market size Sj is determined as follows:

,
if RCAi > 1, and:

,
if RCAi ≤ 1 where MWorld,j is the aggregate imports in the world of product group j. As can be seen from the
above equations, the cut-off points for the relative import market size will vary between two and three
percent according to the RCA24.
Therefore, the relative import market size of country i for product group j will be considered as
sufficiently large, and consequently the country-product combination will be selected as a possible export
opportunity for South Africa, if:

Mi,j ≥ Sj, (5)

Each country-product combination is now assigned a 1 or a 0 in column 4 of Table 4, according to

whether condition (5) is fulfilled or not.

24 The criterion of a 2 % share in the world imports, with South Africa being relatively specialised in a given product j, is for the sake
of comparability with the previous applications of the DSM in Belgium and Thailand, when the same criterion was applied. When
South Africa is relatively less specialized (RCAi ≤ 1), the market size criterion becomes more binding.

19
  

Table 4: Categorising of country-product combinations

Category Short-term Long-term Relative
market growth market growth market size
0 0 0 0
1 1 0 0
2 0 1 0
3 0 0 1
4 1 1 0
5 1 0 1
6 0 1 1
7 1 1 1
Source of data: excerpt from the DSM for South Africa.

Table 4 represents the basis for the selection procedure within filter two. The number 1 is assigned to either
column 2, 3, or 4, if condition 4 in the short and long term and condition 5 are fulfilled respectively,
otherwise a 0 applies. Category 0 in Table 4 thus indicates those markets that do not show growth in the
short or long term, and that also have a relatively small import market size for a specific product group. The
result in Table 4 is indicated as a 0 in column 2, 3, and 4. The markets in category 7 show growth in the short
and long term and have a relatively large market size. Categories 3 to 7 will be considered as interesting
markets and will be chosen for further analyses. These categories will be further grouped into 3 groups.
Group 1 will consist of category 3, group 2 will consist of categories 4, 5, and 6 and, lastly, group 3 will
contain category 7. In filter three, trade restrictions play a part in further eliminating the number of country-
product combinations for analysis.
The distribution of the 83,810 country-product combinations according to the various combinations of
fulfilment or non-fulfilment of condition (4) for short-term and long-term market growth, and of condition
(5), is shown in Table 5.

Table 5: Distribution of country-product combinations according to short-term import market growth, long term import market
growth and relative import market size, 2004
Category Short-term market Long-term market Relative market Number of country-
growth growth size product combinations,
2004
0 0 0 0 59,323
1 1 0 0 2,296
2 0 1 0 7,441
3 0 0 1 1,084
4 1 1 0 10,331
5 1 0 1 136
6 0 1 1 1,674
7 1 1 1 1,525
14,750
Source of data: authors’ own calculations based on DSM results.

Following Cuyvers et al. (1995:179), we will consider further only the country-product combinations which
show either sufficient relative import market size or sufficiently high import market growth in the short and
long terms. This implies that the country-product combinations in categories 0, 1 and 2 of Table 4 are not
selected. This stage of the selection process ends up with 14,750 possible export opportunities for South
Africa.

3.4 Filter three: The selection of realistic export opportunities for South Africa

The purpose of the third stage of the decision support model used is to analyse further the 14,750 country-
product combinations selected in the previous stage according to their “accessibility” for South African
exporters. This “accessibility” depends on trade restrictions and other barriers to entry, which can prevent

20
 

South African exporters of product group j acquiring a significant market position in country i. The decision
support model considers two such barriers: the degree of market concentration, and import restrictions.
Market concentration is measured using the well-known Herfindahl-Hirschmann index (Hirschmann,
1964):

,
where Xk,i,j is country k’s exports of product group j to country i, and Mtot,i,j is country i’s total imports of
product group j.
It is assumed that, if an import market is relatively highly concentrated (i.e. supplied by a few countries), it
will be more difficult for South African exporters of the product group in question to penetrate that market
than if an import market shows a relatively low HHI.
In order to determine whether HHIi,j is sufficiently low, cut-off points are calculated analogously to the
procedure outlined in section 3.1, using an average standard deviation σ and a parameter α to be determined.
Therefore, the cut-off point for HHI is defined as:
hk = xh − 0.1ασ h ,
for country-product combinations of category 3 (see Table 5),
hk = xh + 0.1ασ h ,
for country-product combinations of category 4, 5 or 6 (see Table 5), and
hk = xh + 0.3ασ h ,
for country-product combinations of category 7 (see Table 5):
hk ≥ HHI ij , (6)

Subsequently an alpha value of α = 16.6 was chosen to calculate the cut-off value in this part of filter
three. Using condition (6) we then calculated the cut-off points, which are hk = -0.083 (k = category 3), 0.928
(k = category 4, 5 or 6), and 1.939 (k = category 7). Hence, in relatively large markets, a cut-off point below
zero (i.e. negative cut-off point) implies that none of the markets are selected based on concentration. . In
relatively large and growing markets the degree of concentration is allowed to be higher, and in the most
interesting markets (relatively large and growing in both the short and long terms) the cut-off point of larger
than 1 implies that all markets in that category are selected. In earlier applications of the DSM to Belgium a
more subjective approach was taken in the choice of an appropriate alpha value (Cuyvers et al., 1995:180).
However, in Cuyvers (2004:277) a more scientific method, as suggested by Glenn Rayp, was applied to
calculate an alpha value that would result in the smallest possible impact on the number of country-product
combinations selected. For this reason the same method was applied in this study25.
At this stage of the DSM, the current paper does not include the second part of filter three’s methodology
followed by Cuyvers, et al. (1995) and Cuyvers (1996; 2004) in terms of import restrictions. An alternative
methodology was applied. The first reason for not including this specific filter stems from the proxy used
within Cuyvers et al. (1995) and Cuyvers (1996) to account for import restrictions. The concept of “revealed
absence of barriers to trade” was used as an indicator of the ability for a country to export to another market.
In the Thailand DSM (Cuyvers, 2004) four ASEAN countries were used as a proxy for “revealed absence of
barriers to trade”. The four ASEAN countries were Indonesia, Malaysia, the Philippines and Singapore. The

25 For more detail on the method applied to calculate the alpha value, please refer to Cuyvers (2004:277).

21
 

notion is that, if at least one of the four ASEAN countries is able to penetrate an export market, Thailand
would also be able to penetrate that specific market because of four ASEAN countries’ geographic proximity
and similar economic structures and level of development (Cuyvers, 2004)26.
These criteria however cannot be used for the South African DSM because such proxy countries with
similar characteristics as those surrounding Belgium (OECD countries) and in the Thailand case, ASEAN
countries, could not be found for South Africa. As an alternative, an index for market accessibility was
constructed using the following five variables: distance, transport cost, the World Bank Logistics Performance
Index (LPI), average applied tariffs, and the frequency coverage ratio of non-tariff barriers.

3.4.1 Variable 1: Distance

Countries were divided into coastal countries and landlocked countries. All distances were calculated from
Durban harbour to the main port of the coastal countries or the nearest port to the landlocked countries.
For coastal countries, data on the nautical miles from Durban harbour to a port in each of the countries was
collected from SeaRates.com and Netpas Distance. Nautical miles were converted into kilometres (1 nautical
mile = 1.852 kilometres).
For landlocked countries, the distance was calculated based on data on the nautical miles from Durban
harbour to the port nearest to the landlocked country (this port is situated in another country) was obtained
from the same sources as for the coastal countries. Nautical miles were also converted into kilometres. As-
the-crow-flies distances (in kilometres) were then obtained from the port closest to the capital city in the
landlocked country by using MapCrow Travel Distance Calculator. Distances to landlocked countries were
then calculated as the sum of the kilometres from Durban harbour to the closest port and from the closest
port to the capital of the particular landlocked country.

3.4.2 Variable 2: Transport Cost

Quotes for the shipment of a 20-foot container from Durban harbour to as many countries around the world
as possible were obtained from three main shipping lines. Based on these quotes, the average shipping cost
for each country was calculated. In the case of landlocked countries, the cost of shipment to the nearest port
for which a quote was available was used. Only sea transport fees were considered due to time and cost
limitations of obtaining road and rail transportation cost per country. Sea transport cost will, however, still be
a good indicator of transport cost. Furthermore, domestic logistics cost is also considered in the construction
of the logistics performance index of each country (see section 3.4.3).

3.4.3 Variable 3: Logistics Performance Index (World Bank)

The Logistics Performance Index (LPI) that was constructed by the World Bank and built on information
from a web-based questionnaire completed by more than 800 logistics professionals worldwide was used. The
logistics professionals included the operators or agents of the world’s largest logistics service providers. Each
respondent was asked to rate the performance in seven logistics areas for eight countries with which they
conducted business. For each respondent, the eight countries were automatically generated by the survey
engine based on trade flows, income level, geographical position of respondent countries (coastal or

26 In this respect, these four ASEAN countries differ from the other ASEAN members like Laos, Cambodia, Myanmar, etc.

22
 

landlocked) and random selection. The performance was evaluated using a 5-point scale (1 being the lowest
score and 5 being the highest score).
The seven areas of performance are:
• Efficiency of the clearance process by customs and other border agencies;
• Quality of transport and information technology infrastructure for logistics;
• Ease and affordability of arranging international shipments;
• Competence of the local logistics industry;
• Ability to track and trace international shipments;
• Domestic logistics costs;
• Timeliness of shipments in reaching their destinations;

More than 5000 individual country evaluations (which cover 150 countries) were used to prepare the
Logistics Performance Index (LPI). The LPI was aggregated as a weighted average of the seven areas of
logistics performance, and constructed using the principal component analysis method in order to improve
the confidence intervals. The questionnaire and more information are available at www.worldbank.org/lpi.

3.4.4 Variable 4: Average Applied Tariffs

The latest available simple average of each county’s Most Favoured Nation (MFN) applied tariffs for all
product categories available in the latest year (mostly 2007) was used. Sources include the UNCTAD
TRAINS database and the WTO. Only the average tariff for the Netherlands Antilles was not available. This
value was left as a missing value in the dataset and was not taken in consideration in the calculation of the
average index of the Netherlands Antilles.

3.4.5 Variable 5: Frequency Coverage Ratio of Non-Tariff Barriers

The unweighted average percentages of non-tariff barriers by country were used in this category. The latest
data available varied between countries. The data are based on core non-tariff barriers which are defined as
including quantity and price restrictions as well as monopolistic trading channels. The main source for this
data is UNCTAD TRAINS. For 19 of the countries that entered filter three, non-tariff barrier percentages
were not available. These were also left as missing values in the dataset and were not taken into consideration
in the calculation of the average index of those 19 countries.

3.4.6 Construction of an index for market accessibility

Incorporating these five variables, an index for market accessibility was subsequently constructed. Given the
fact that all these variables are in different units of measurement (distance – km, price – US$, LPI – score out
of 5, tariffs – percentage and non-tariff barriers – percentage), z-scores for each variable were calculated to
standardise the data. The z-score for each variable from each country was calculated as follows:

In this manner, one can easily see if a country scores above or below the world average for the variable.

23
 

The z-scores for the different variables from each country were used to arrive at a final index. Because of
the fact that the further away from South Africa a country is, the more difficult it would be to export to, the
distance z-scores were subtracted when calculating the index. Also, higher cost to export to a particular
country would adversely affect its attractiveness. Z-scores for cost were also therefore also subtracted when
calculating the index. The higher a country’s LPI score, the better. LPI z-scores were therefore added when
calculating the index. Also, the higher the tariff/non-tariff barrier the less attractive a country would be, the
tariff and non-tariff barrier z-scores were therefore also subtracted when calculating the index. Therefore
distance, cost and tariff and non-tariff barriers are negative factors and LPI a positive factor in the calculation
of the index.
The researchers first set out to determine how to assign weights to the different variables. There is
however no clear indication in the literature hereof. It was therefore decided that no elimination of countries
would be done in this part of filter 3. The countries were ranked as most accessible (green), lesser accessible
(orange) and least accessible (red) in order for the dti to take market accessibility to the different countries
under investigation into account in their export promotion activities. Four options of weighing the different
variables, together with its results, were provided to the dti (Rossouw, Steenkamp and Viviers, 2009). For 60
per cent of the countries under review there was no difference in the classification between the different
options. It was argued that the dti’s practical experience and encounters with the different countries in
question might shed light on which weighting to choose.
After reviewing the different options, Mr. R. le Roux (2009), Chief director of the dti’s Export promotion
and development, indicated in a meeting that the best option for weighing the different variables would be as
follows: tariff and non-tariff barriers (converted into one variable): 40%, transport cost: 30%, LPI: 20%,
distance: 10%. The dti indicated that these weights cannot be considered perfect, but the ranking would at
least give some indication of the market accessibility of the different countries included in the DSM. It needs
to be stressed again that no elimination of country-product combinations was done in this part of filter 3 and
that future studies should be conducted to improve on this part of filter 3 for the application of the DSM for
South Africa. In tables 6, 7 and 8 the accessibility ranking, green, orange and red, respectively, are indicated.

24
  

Table 6: Most accessible countries based on degree of restriction – “green countries”

Average z-
Shipping score
Distance Z-score Cost Z-score Z-score Tariffs & Weighted
Country (km) Distance (US$) Cost LPI LPI NTB Index
Singapore 9052.576 -0.30917 298.00 -1.67520 4.19 2.274869 -0.86829 1.33577
Hong Kong (China) 11558.33 0.38030 285.00 -1.69938 4.00 1.975583 -1.11987 1.31485
Macau 11526.85 0.37163 614.37 -1.08681 3.32 0.904455 -1.73703 1.16458
United Arab Emirates 7502.452 -0.73569 743.33 -0.84697 3.73 1.550282 -0.98461 1.03156
Brunei 9867.456 -0.08495 577.60 -1.15520 3.48 1.156485 -0.7277 0.87743
Iceland 14001.12 1.05244 1072.50 -0.23478 3.99 1.959831 -1.28916 0.87282
Malaysia 8715.512 -0.40192 577.60 -1.15520 3.48 1.156485 -0.46131 0.80257
Israel 8765.516 -0.38816 908.00 -0.54072 3.21 0.731185 -0.99797 0.74646
Bahrain 7906.188 -0.62460 960.00 -0.44401 3.15 0.636674 -0.938 0.69820
Indonesia 8450.676 -0.47479 504.00 -1.29208 3.01 0.416147 -0.44593 0.69670
Georgia 11010.14 0.22946 1335.00 0.25342 3.15 0.636674 -1.64745 0.68735
Bulgaria 10204.52 0.00779 1335.00 0.25342 2.87 0.195621 -1.73703 0.65713
Kuwait 8248.808 -0.53033 960.00 -0.44401 2.99 0.384643 -0.93087 0.63551
Japan 13973.34 1.04479 715.86 -0.89807 4.02 2.007087 -0.1408 0.62268
Australia 10815.68 0.17595 853.33 -0.64239 3.79 1.644794 -0.24011 0.60012
Switzerland 11379.36 0.33105 997.50 -0.37427 4.02 2.007087 -0.2736 0.59003
Mauritius 2909.492 -1.99946 510.00 -1.28092 2.13 -0.97002 -0.43678 0.56493
Austria 11305.3 0.31067 997.50 -0.37427 4.06 2.070094 -0.16884 0.56277
Philippines 11189.78 0.27889 703.00 -0.92198 2.69 -0.08791 -0.79079 0.54744
Netherlands 12991.78 0.77472 1072.50 -0.23478 4.18 2.259117 -0.16884 0.51232
Oman 5793.056 -1.20604 1060.00 -0.25803 2.92 0.27438 -0.64708 0.51172
Qatar 7833.96 -0.64448 1310.00 0.20692 2.98 0.368892 -1.05627 0.49866
Germany 13178.83 0.82618 1072.50 -0.23478 4.10 2.133102 -0.16884 0.48197
Saudi Arabia 7152.424 -0.83200 1060.00 -0.25803 3.02 0.431899 -0.58282 0.48012
Thailand 10563.81 0.10665 665.00 -0.99265 3.31 0.888704 -0.02552 0.47508
United Kingdom 12671.38 0.68656 1072.50 -0.23478 3.99 1.959831 -0.16884 0.46128
China 12845.47 0.73446 614.37 -1.08681 3.32 0.904455 -0.03934 0.44923
Taiwan 12199.12 0.55661 467.00 -1.36090 3.64 1.408515 0.463823 0.44878
Norway 13736.28 0.97957 1472.50 0.50914 3.81 1.676297 -0.90392 0.44613

Table 7: Lesser accessible countries based on degree of restriction – “orange countries”

Average
Shipping z-score
Distance Z-score Cost Z-score Z-score Tariffs & Weighted
Country (km) Distance (US$) Cost LPI LPI NTB Index
Italy 10219.34 0.01187 997.50 -0.37427 3.58 1.314004 -0.16884 0.44143
Malta 10278.6 0.02817 997.50 -0.37427 3.58 1.314004 -0.16884 0.43980
Belgium 12806.58 0.72376 1072.50 -0.23478 3.89 1.802312 -0.16884 0.42606
France 12447.29 0.62490 1032.50 -0.30917 3.76 1.597538 -0.16884 0.41731
Sri Lanka 6728.316 -0.94870 810.00 -0.72298 2.40 -0.54472 -0.45936 0.38657
New Zealand 12273.2 0.57700 876.67 -0.59899 3.75 1.581786 0.171542 0.36974
Mozambique 2044.608 -2.23744 926.67 -0.50600 2.29 -0.71799 -0.31799 0.35915
Spain 10837.9 0.18207 1122.50 -0.14179 3.52 1.219493 -0.16884 0.33576
Turkey 9928.572 -0.06814 1335.00 0.25342 3.15 0.636674 -0.68486 0.33207
Mongolia 15071.52 1.34696 614.37 -1.08682 2.08 -1.04878 -0.85921 0.32528
Tanzania 3313.228 -1.88837 685.00 -0.95546 2.08 -1.04878 -0.14586 0.32406
Slovenia 10845.31 0.18411 997.50 -0.37427 3.14 0.620922 -0.16884 0.28559
Hungary 11024.13 0.23331 997.50 -0.37427 3.15 0.636674 -0.16884 0.28382
Zambia 2139.592 -2.21130 926.67 -0.50600 2.37 -0.59197 -0.04032 0.27066
Portugal 11169.41 0.27328 1147.50 -0.09530 3.38 0.998967 -0.16884 0.26859
Czech Republic 11411.46 0.33988 997.50 -0.37427 3.13 0.60517 -0.16884 0.26686
Ireland 12580.64 0.66159 1397.50 0.36966 3.91 1.833816 -0.16884 0.25724
Albania 10178.59 0.00066 997.50 -0.37427 2.08 -1.04878 -0.86751 0.24946
The Republic of Korea 14627.1 1.22468 614.37 -1.08681 3.52 1.219493 0.514063 0.24185
Sweden 13619.61 0.94746 1472.50 0.50914 4.08 2.101598 -0.16884 0.24037
Denmark 13551.08 0.92861 1372.50 0.32316 3.86 1.755057 -0.16884 0.22874
Slovakia 11080.34 0.24878 997.50 -0.37427 2.92 0.27438 -0.16884 0.20981
Greece 9850.788 -0.08954 1335.00 0.25342 3.36 0.967463 -0.16884 0.19396
Armenia 11241.92 0.29324 1335.00 0.25342 2.14 -0.95427 -1.2175 0.19080
Bosnia and Herzegovina 11725.01 0.42616 997.50 -0.37427 2.46 -0.45021 -0.51883 0.18715
Moldova 10593.86 0.11492 1335.00 0.25342 2.31 -0.68648 -0.89126 0.13169
Poland 13912.22 1.02798 1072.50 -0.23478 3.04 0.463403 -0.16884 0.12785
Finland 14912.3 1.30315 1472.50 0.50914 3.82 1.692049 -0.16884 0.12289

25
  

Table 8: Least accessible countries based on degree of restriction – “red countries”

Average
Shipping z-score
Distance Z-score Cost Z-score Z-score Tariffs & Weighted
Country (km) Distance (US$) Cost LPI LPI NTB Index
Azerbaijan 8985.166 -0.32772 960.00 -0.44401 2.29 -0.71799 -0.23219 0.11525
Viet Nam 11460.18 0.35329 757.50 -0.82062 2.89 0.227125 0.358991 0.11269
Cyprus 8880.34 -0.35656 1335.00 0.25342 2.92 0.27438 -0.16884 0.08204
Romania 10297.12 0.03327 1335.00 0.25342 2.91 0.258628 -0.16884 0.03991
Ukraine 10565.66 0.10716 1335.00 0.25342 2.55 -0.30844 -0.4582 0.03485
Trinidad and Tobago 11758.35 0.43533 1410.00 0.39290 2.75 0.006598 -0.40952 0.00372
Latvia 14312.26 1.13805 1472.50 0.50914 3.02 0.431899 -0.16884 -0.11263
Uzbekistan 9089.67 -0.29896 998.00 -0.37334 2.16 -0.92276 0.197759 -0.12176
Estonia 14397.45 1.16149 1472.50 0.50914 2.95 0.321636 -0.16884 -0.13703
Bangladesh 8730.328 -0.39784 1210.00 0.02094 2.47 -0.43445 0.221003 -0.14179
Netherlands Antilles 12099.12 0.52910 1410.00 0.39290 2.75 0.006598 -0.16946
Lithuania 14023.34 1.05855 1472.50 0.50914 2.78 0.053854 -0.16884 -0.18029
Canada 14641.91 1.22875 2429.25 2.28851 3.92 1.849568 -0.63189 -0.18676
Bhutan 9130.488 -0.28773 1210 0.02094 2.16 -0.92276 0.111374 -0.20661
India 7233.912 -0.80958 1138.25 -0.11250 3.07 0.510659 1.226561 -0.27378
United States 15364.19 1.42749 2393.90 2.22277 3.84 1.723553 -0.31527 -0.33876
Chad 7049.846 -0.86023 1560.00 0.67187 1.98 -1.2063 0.095991 -0.39519
Ecuador 14453.01 1.17678 1410.00 0.39290 2.60 -0.22968 0.420011 -0.44949
Brazil 7543.196 -0.72448 1410.00 0.39290 2.75 0.006598 1.158991 -0.50770
Barbados 11412.02 0.34004 1410.00 0.39290 2.75 0.006598 0.914347 -0.51629
Maldives 5980.108 -1.15457 810.00 -0.72298 2.4 -0.54472 1.88174 -0.52929
St.Kitts and Nevis 12065.78 0.51992 2393.90 2.22277 3.84 1.723553 0.412735 -0.53921
Peru 12404.7 0.61318 2210.00 1.88075 2.77 0.038102 0.13805 -0.67314
Russia 16645.78 1.78013 1472.50 0.50914 2.37 -0.59197 0.58358 -0.68258
Kazakhstan 10198.24 0.00606 2177.06 1.81949 2.12 -0.98577 0.055486 -0.76580
Mexico 15253.07 1.39692 2393.90 2.22277 2.87 0.195621 1.251874 -1.26815
Equatorial Guinea 5615.264 -1.25496 3010.00 3.36860 2.49 -0.40295 1.415958 -1.53206
Bahamas 13877.04 1.01830 2393.90 2.22277 3.84 1.723553 3.619465 -1.87174

In filter four a country-product combination will be considered for further investigation if condition (6) is
satisfied. This is the case for 12,695 country-product combinations.

3.5 Filter four and results: An analysis of South Africa’s realistic export opportunities

In this section we analyse the 12,695 realistic export opportunities detected according to their product group,
the geographical markets involved, and some major characteristics of these markets.

The process up to this stage of the model is summarised in table 9.

Table 9: Selection of realistic export opportunities for South Africa, 2004

Filter one: Country risk rating Filter two: Detecting possible Filter three: Realistic export Filter four: Final analysis of
and macroeconomic analysis export opportunities opportunities export opportunities
• Country risk rating (critical value • 83,810 country-product • Herfindahl-Hirschmann Index • Criteria of filters 2, 3 & 4.
9.29, 193 countries selected). combinations. (Degree of Market • 12,695 country-product
• Selection on the basis of • Category 3 – 0 Concentration). combinations analysed and
macroeconomic indicators. • Category 4, 5, 6 – 11,272 • 12,695 country-product classified.
• GDP and GDP per capita – 95 • Category 7 – 1,483. combinations selected.
countries selected. • 14,750 country-product • Market accessibility index:
• GDP growth and GDP per combinations selected. ranking of 85 countries.
capita growth – 98 countries
selected.
• 85 countries selected to enter
filter two.
Source: Decision support model analyses conducted by the authors.

26
 

When analysing the 12,695 country-product combinations identified in filters one to three, the following
product groups were identified as having the highest export potential for South Africa (shows potential in the
highest number of countries):
• Wine of fresh grapes (including fortified wine); grape must in fermentation or with fermentation
arrested (SITC 1121): 60 country-product combinations.
• Non-alcoholic beverages, n.e.s. (SITC 1110): 57 country-product combinations.
• Paper and paperboard, uncoated, n.e.s., in rolls or sheets (SITC 6415): 57 country-product
combinations.
• Aluminium and aluminium alloys, worked (SITC 6842): 56 country-product combinations.
• Iron or non-alloy steel flat-rolled products, not clad, plated or coated, hot-rolled only, specified yield
points based on thickness (SITC 6731): 55 country-product combinations.
• Iron and non-alloy steel flat-rolled products, not clad, plated, etc., cold-rolled (cold-reduced) only,
specified yield points based on thickness (SITC 6733): 55 country-product combinations.
• Iron and steel railway and tramway track construction material (SITC 6770): 55 country-product
combinations.
• Overcoats, car coats, capes, anoraks (including ski-jackets), etc. (except suit-type jackets), of woven
textile fabrics, women’s or girls’ (SITC 8421): 52 country-product combinations.
• Juices; fruit and vegetable (including grape must) unfermented (SITC 0585): 50 country-product
combinations.
• Fertilisers, n.e.s. (imports only) (SITC 5629): 50 country-product combinations.
• Millstones, grinding wheels and the like, without frameworks, for grinding, sharpening, trueing, etc., of
natural or artificial abrasives or ceramics (SITC 6631): 50 country-product combinations.
• Crustaceans and molluscs, fresh, chilled, frozen etc. (SITC 0360): 49 country-product combinations.
• Colouring matter n.e.s.; preparations based on colouring matter, n.e.s.; inorganic products used as
luminophores (SITC 5331): 49 country-product combinations.
• Passenger motorcars, for transport of passengers and goods (SITC 7810): 49 country-product
combinations.
• Minerals, crude, n.e.s. (SITC 2789): 47 country-product combinations.
• Bulbs, tubers, and rhizomes of flowering or of foliage plants; cuttings, slips, live trees and other plants
(SITC 2926): 47 country-product combinations.
• Aldehyde-, ketone- and quinone-function compounds (SITC 5162): 47 country-product combinations.
• Inorganic acids and oxygen compounds of non-metals (SITC 5222): 47 country-product combinations.
• Metallic oxides of zinc, chromium, manganese, iron, etc. (SITC 5224): 47 country-product
combinations.
• Organic surface-active agents other than soap; surface-active, washing and cleaning preparations,
whether or not containing soap, n.e.s. (SITC 5542): 47 country-product combinations.
• Sacks and bags of textile materials used for packing goods (SITC 6581): 47 country-product
combinations.
• Monumental or building stone and articles thereof (SITC 6613): 47 country-product combinations.
• Statuettes and other ornaments, and articles of adornment (SITC 6666): 47 country-product
combinations.
• Trousers, breeches etc. of textile fabrics (SITC 8423): 47 country-product combinations.
• Fluorides; fluorosilicates, fluoroaluminates and other complex fluorine salts (SITC 5231): 46 country-
product combinations.

27
 

The 85 countries under investigation were also ranked in table 10 according to export potential for South
African products. As can be seen in table 10, these export opportunities are geographically diverse, with some
European countries showing the largest numbers (e.g. Germany, France, Italy, the UK and the Netherlands),
with countries of the Pacific Rim (the US, Japan, Hong Kong, Korea, China and the other ASEAN-5
countries) coming close.

Table 10: Realistic export opportunities per country, 2004

Number of Number of
Rank in 2004 Country
Rank in 2004 Country opportunities in opportunities in
(continued) (continued)
2004 2004 (continued)
1 Germany 354 44 Albania 136
2 France 315 45 Iceland 136
3 Italy 288 46 Sweden 136
4 China 274 47 Thailand 136
5 Belgium-Luxembourg 273 48 Oman 135
6 UK 273 49 Singapore 135
7 Turkey 261 50 Switzerland 128
8 Netherlands 250 51 Sri Lanka 127
9 Spain 249 52 Azerbaijan 125
10 Poland 243 53 Finland 125
11 Russia 234 54 Estonia 122
12 Ukraine 232 55 Philippines 118
13 Romania 231 56 Mongolia 113
14 USA 230 57 Bangladesh 109
15 India 213 58 Norway 109
16 Japan 213 59 Peru 108
17 Kazakhstan 209 60 Tanzania 106
18 Czech Republic 207 61 Ireland 105
19 Bulgaria 200 62 Saudi Arabia 103
20 Indonesia 198 63 Ecuador 101
21 Malaysia 192 64 Malta 100
22 Bosnia Herzegovina 191 65 Portugal 98
23 Lithuania 189 66 Mauritius 96
24 Austria 186 67 Kuwait 95
25 Slovakia 180 68 Mozambique 94
26 Moldova 178 69 Vietnam 93
27 Georgia 176 70 Qatar 92
28 Republic of Korea 175 71 Israel 88
29 Hungary 172 72 Uzbekistan 87
30 Latvia 159 73 Armenia 81
31 Australia 155 74 Equatorial Guinea 78
32 Canada 153 75 Trinidad Tobago 78
33 Greece 151 76 Netherlands Antilles 77
34 Hong Kong 150 77 St. Kitts Nevis 70
35 Maldives 149 78 Barbados 69
36 Slovenia 149 79 Zambia 69
37 Brazil 148 80 Bahrain 68
38 Denmark 146 81 Bahamas 60
39 Macau 145 82 Brunei 48
40 New Zealand 143 83 Chad 32
41 Cyprus 142 84 Mexico 32
42 United Arab Emirates 142 85 Bhutan 21
43 Taiwan 138
Source of data: authors’ own calculations based on DSM results.

The distribution of realistic export opportunities according to the 12 identified regional clusters is shown in
Table 11. From Table 11, it can be seen that the EU and Asia have the highest shares in the number of
realistic export opportunities found.

28
  

Table 11: South Africa: distribution of realistic export opportunities according to regional clusters, 2004
Number of
Regional cluster opportunities 2004 Percentage
Africa 475 3.74
Asia 2,872 22.62
Western Europe 2,961 23.32
Middle Europe 1,660 13.08
Eastern Europe 1,110 8.74
Scandinavia 652 5.14
Baltic States 470 3.70
Middle East 1,071 8.44
Australasia 298 2.35
North America 415 3.27
South America 357 2.81
Caribbean 354 2.79
Other 0 0.00
12,695 100.00
Source of data: authors’ own calculations based on DSM results.

Realistic export opportunities can further be grouped according to their relative market importance for South
Africa and according to their relative size and growth rate. Table 12 shows this tentative grouping.

Table 12: South Africa’s realistic export opportunities according to relative market position and market characteristics, 2004
Market share of Market share of
Market share of South Africa South Africa Market share of
South Africa intermediately intermediately South Africa
relatively small small high relatively high Total
(Cell 1) (Cell 6) (Cell 11) (Cell 16)
Large product/market 0
0 0 0 0
Growing (long- and short term) (Cell 2) (Cell 7) (Cell 12) (Cell 17)
9478
product/market 7277 290 107 1804
Large product/market short (Cell 3) (Cell 8) (Cell 13) (Cell 18)
118
term growth 106 5 1 6
Large product/market long (Cell 4) (Cell 9) (Cell 14) (Cell 19)
1609
term growth 1267 288 33 21
Large product/market short- (Cell 5) (Cell 10) (Cell 15) (Cell 20)
1490
and long term growth 1307 129 21 33
Total 9957 712 162 1864 12695
Source of data: authors’ own calculations based on DSM results

The categorisation in table 12 was done by calculating a relative market importance for South Africa’s exports
in the different country-product combinations (vertical classification – small market share to high market
share) and using the categories identified in filter two (horizontal classification - category 3: large product
market to category 7: large market with short- and long-term growth).
The calculation of South Africa’s relative market share for each country-product combination is
subsequently discussed. For each chosen exporting country n (South Africa in this case), the degree of market
importance of country n’s exports of product group j to country i is defined as:

where Xn,i,j is country n’s exports of product group j to country i, XWorld,i,j is the world’s exports of product
group j to country i, Xn,j is country n’s total exports of product group j, and XWorld,j is the world’s total exports
of product group j.

29
 

A comparison can now be made for any particular country-product combination selected in the previous
section of South Africa’s μSA,i,j with μSix,i,j, the combined degree of market importance of the six exporting
countries with the largest exports of the product category to the country in question. By calculating the
difference between South Africa’s degree of market importance and that of the six dominant exporting
countries of product group j to country i, we can now determine whether South Africa’s relative market share
is large or small. We are therefore using the following rules as suggested by Cuyvers (2004:267):
• : the relative market share of South Africa is relatively small;
• : the relative market share of South Africa is intermediately small;
• : the relative market share of South Africa is intermediately high; and
• : the relative market share of South Africa is relatively high.

Table 12 shows the distribution of realistic export opportunities for 2004 according to the characteristics
of the foreign markets and South Africa’s relative market share.
As stated in section 2.1.2.2.2.6, these results could be used to derive appropriate export promotion
strategies in relevant export markets (Cuyvers et al., 1995:174). From Table 12, it can be seen that about 78
percent of the realistic export opportunities found for 2004 correspond to country-product combinations
with a relatively small market share for South Africa (cells 1 to 5). Cuyvers (2004:270) recommended that, in
the case of Thailand, country-product combinations in cells 1 to 5 should not be actively promoted due to the
relatively limited resources of Thailand’s Department for Export Promotion. It can be argued that the same
applies for South Africa and that a policy option could be not to promote actively the export opportunities in
these cells of Table 12, but rather to gather and disseminate market information regarding these opportunities
to South African exporters. It could also be argued, as in the Thai case, that export opportunities in cells 6 to
10 (representing some 6 per cent of all realistic export opportunities in 2004) should also not be actively
promoted, although these opportunities could be explored in greater depth using the official trade counsellors
at South African embassies abroad and the services of the dti.
As recommended by Cuyvers (2004:270) in the Thai case, an active and offensive export promotion
strategy of “market expansion” can be implemented by South Africa’s TPO for the country-product
combinations in cells 11 to 15. This strategy is recommended because South African exporters have already
gained some market share in these markets and this experience can be utilised for future expansion of South
Africa’s market share.
For the country-product combinations in cells 16 to 20 (which account for 15 per cent of all realistic
export opportunities in 2004) in which South Africa already has a relatively large market share, a defensive
export promotion strategy of “market maintenance” would be appropriate.
Another presentation of the results was done after consultation with the dti’s Chief Director of Export
Promotion and Development. The 12,695 realistic export opportunities were clustered according to regions
in the world as deemed important by the dti. The 13 regional clusters identified in collaboration with the dti
include: Africa, Asia, Western Europe, Middle Europe, Eastern Europe, Scandinavia, the Baltic States, the
Middle East, Australasia, North America, South America, and the Caribbean. The 13 clusters are illustrated in
Figure 4.

30
 

Figure 4: Global clusters according to overall share of realistic export opportunities in 2004

•Austria (DTI)
•Belgium‐Luxembourg (DTI)
•Bosnia Herzegovina
•France (DTI)
13 Clusters •Albania
•Germany (DTI) Realistic Export Opportunities •Bulgaria
(DTI) – existing offices
•Greece •Cyprus
21.12% •Ireland •Czech Republic
•Denmark
•Italy (DTI) Western  •Finland •Hungary
•Netherlands (DTI) Europe •Iceland •Poland
•Portugal •Norway 3.81% •Romania
13.45%
•Spain (DTI) 5.28% •Slovenia
•Sweden (DTI)
•Switzerland (DTI) •Slovakia
Baltic States
•United Kingdom (DTI) Scandinavia
•Estonia Middle Europe
•Malta 
•Latvia
•Lithuania
•Armenia
•Georgia
•Kazakhstan
•Moldova
Eastern Europe •Russia
•Bangladesh
•Turkmenistan
•Bhutan
•Ukraine
North  •Brunei
•Uzbekistan
•China (DTI)
America •Hong Kong (DTI)
8.99%
•Canada (DTI) •India (DTI)
•Mexico •Indonesia (DTI)
3.36%
•USA (DTI) Asia •Japan (DTI)
Caribbean 2.87% •Korea, Republic of
•Bahamas •Macau
•Barbados •Malaysia
•Netherlands Antilles •Maldives
Clusters •Mongolia
•Saint Kitts and Nevis Middle 
Rest of World 23.27% •Philippines
•Trinidad and Tobago East
•Singapore (DTI)
Caribbean Africa •Azerbaijan
•Sri Lanka
North America •Bahrain
•Taiwan (DTI)
•Chad •Israel (DTI) •Thailand (DTI)
Scandinavia •Equatorial Guinea •Kuwait
•Vietnam
Africa •Mauritius •Oman 8.68%
•Mozambique •Qatar
South America •Turkey
•Tanzania
Western Europe •Zambia •Saudi Arabia (DTI)
South America •United Arab Emirates (DTI)
Eastern Europe
2.89% •Brazil (DTI)
Australasia •Ecuador 3.85% Australasia
Middle Europe •Peru •Australia (DTI)
Middle East •New Zealand
2.41%
Baltic States
Asia

Source: Compiled by the authors.

From figure 4 it is clear that 23.27% of the realistic export opportunities identified for South Africa are in
Asia and 21.12% in Western Europe. Clusters with the least opportunities are Australia (2.41%), the
Caribbean (2.87%), South America (2.89%), North America (3.36%) and Africa (3.85%). This is a more
visual presentation of the results in table 11.
Although one of the main benefits of the DSM is that it provides a tool to assist export promotion
authorities to decide how to allocate their scarce resources to export promotion activities in various markets,
the list of realistic export opportunities should not be considered as guaranteeing future success. It would
therefore be unwise to rest all export promotion decisions upon the model alone. Other considerations, such
as feedback from foreign trade offices (on the demand side of exports) and export councils (on the supply
side), should also be taken into consideration.

4. POLICY IMPLICATIONS AND CONCLUSION

Export promotion activities by governments are, to a large extent, driven by historical trends and trading
partners (DTI, 2005). However, the limited resources of government should be allocated in such a manner
that they contribute towards successful exports and increase export growth in the future. The dti indicated
that further research on export promotion in South Africa would greatly assist senior management in
ensuring that government resources are used with maximum return on investment by determining priority
products and markets (Erero, 2004). In order to determine these priority country-product combinations, a
decision support model (DSM) was applied to the South African export market and will provide the dti with
a tool to justify export promotion activities more scientifically.

31
 

The methodology of the DSM developed by Cuyvers et al. (1995:173-186), Cuyvers (1996:71-96) and
Cuyvers (2004:255-278) was discussed. The DSM consists of a sequential filtering process eliminating
countries with lower export potential with the use of four filters. The first filter considers the macro-
economic environment of the trading partner. Indicators such as country risk ratings; GDP (GDP per capita)
and GDP growth (GDP per capita growth) play a role in the selection process.
In filter two, import market growth in the short and long term, and relative market size were considered
for each country-product combination. Table 4 is constructed to show the categories that will be used for
further analysis. In filter three, the Herfindahl-Hirschmann Index gives an indication of the market
concentration of the importing countries and barriers to entry. In filter four, realistic export opportunities
identified in the previous filters were classified. This classification was done by calculating South Africa’s
relative market importance for each country-product combination and combining this with the categorisation
in filter two (see Table 12).
After the application and adaptation of the DSM, 12,695 country-product combinations were identified as
realistic export opportunities. After the identification of the 12,695 country-product combinations, the dti
indicated that a clustering process would enable them to deploy their limited resources better. The clustering
process will enable the dti to focus on specific regions and products when developing their export promotion
strategies.
In conclusion, future research should aim to develop a more robust index to account for import
restrictions. Furthermore, an attempt could be made to increase the product-specificity of the DSM by using
a less aggregated product classification (e.g. six-digit instead of four-digit trade data). Future studies may also
include specific country-product studies to account for the qualitative issues that cannot be captured in a
model. In-depth analyses of the results of specific continents could also be conducted. The robustness of the
model could also be tested by varying cut-off values and comparing the results. Furthermore, it is
recommended that the DSM results should form part of an overall strategy towards increasing exports
through utilising government resources; and the dti’s current export promotion programmes should be
correlated with the realistic export opportunities identified with the DSM. The DSM results can also be used
to evaluate the appropriateness of current export promotion offerings. Ultimately the results of this study can
contribute to more efficient export assistance in South Africa.

32
 

REFERENCES

Alon, I. (2004) “International market selection for a small enterprise: A case study in international entrepreneurship”, SAM Advanced
Management Journal, pp. 25-33, Winter.
Andersen, O. and Buvik, A. (2002) “Firms’ internationalisation and alternative approaches to the international customer/market
selection”, International Business Review, Vol. 11, pp. 347-363.
Andersen, P. H., and Strandskov, J. (1998) “International market selection: A cognitive mapping perspective”, Journal of Global
Marketing, Vol. 11(3), pp. 65-84.
Arnold, D. J. and Quelsh, J. A. (1998) “New Strategies in Emerging Markets”, Sloan Management Review, pp. 7-20, Fall.
Arvis, J.F., Mustra, M.A., Ojala, L. and Naula, T (2007) “Trade Logistics in the Global Economy”, The World Bank, pp. 1-38.
Ayal, I. and Zif, J. (1978) “Competitive market choice strategies in multinational marketing”, Columbia Journal of World Business, pp. 72-
81, Fall.
Brewer, P. (2001) “International market selection: developing a model from Australian case studies”, International Business Review, Vol.
10, pp. 155-174.
Cavusgil, S. T. (1985) “Guidelines for Export Market Research”, Business Horizons, pp. 27-33, November-December.
Cavusgil, S. T. (1997) “Measuring the Potential of Emerging Markets: An Indexing Approach”, Business Horizons, pp. 87-91, January-
February.
Coughlin, C. C. and Cartwright, P. A. (1987) “An examination of state foreign export promotion and manufacturing exports”, Journal
of Regional Science, Vol. 27, pp. 439-449.
Cutts, S. & West, J. (1998) “The Arrangement on Export Credits”, The OECD OBSERVER, Vol. 211, pp. 12-14, April/May.
Cuyvers, L. (1996) “Export opportunities of Thailand: a decision support model approach”, Journal of Euro-Asian management, Vol.
l2(2), pp. 71-96.
Cuyvers, L. (2004) “Identifying export opportunities: the case of Thailand”, International Marketing Review, Vol. 21(3), pp. 255-278.
Cuyvers, L., De Pelsmacker, P., Rayp, G. and Roozen, I. T. M. (1995) “A decision support model for the planning and assessment of
export promotion activities by government export promotion institutions: the Belgian case”, International Journal of Research in
Marketing, Vol. 12, pp. 173-186.
Davidson, W. H. (1983) “Market similarity and market selection: Implications for international marketing strategy”, Journal of Business
Research, Vol. 11(4), pp. 439-456.
Douglas, S. P. and Craig, C. S. (1992) “Advances in international marketing”, International Journal of Research in Marketing, Vol. 9, pp.
291-318, December.
Douglas, S. P., Craig, C. S. and Keegan, W. J. (1982) “Approaches to Assessing International Marketing Opportunities for Small and
Medium-Sized Companies”, Columbia Journal of World Business, pp. 26-32, Fall.
Department of Trade and Industry. (2005) “National export strategy”, Pretoria: Department of Trade and Industry, DTI-campus.
No. 4, June.
DTI, see The Department of Trade and Industry.
Dunning, J. H. (1997) “How should national governments respond to globalization?” International Executive, Vol. 39(1), pp. 55-66,
Jan/Feb.
Erero, J. (2004) Letter sent by the Department of Trade and Industry to the NRF, dated 06/08/2004 - Director: Statistical Analysis
and Modelling.
Field, A. (2005) “Discovering statistics using SPSS”, London, SAGE Publications, pp. 619-620.
Freudenberg, M. (2006) “Export potential assessment: Identifying priority sectors for export promotion”, International Trade Centre
UNCTAD/WTO, Market Analysis Section.
Freudenberg, M. and Paulmier, T. (2005a) “Export potential assessment in Vietnam”, International Trade Centre, Market analysis
section.
Freudenberg, M. and Paulmier, T. (2005b) “Export potential assessment in Lao PDR”, International Trade Centre, Market analysis
section, November.
Freudenberg, M., Paulmier, T., Bijl, B. and Ikezuki, T. (2007) “Export Potential Assessment in Nepal. Implemented by the
International Trade Centre (ITC)”, Market Analysis Section In collaboration with the Trade and Export Promotion Centre
(TEPC) of Nepal.
Freudenberg, M., Paulmier, T., Ikezuki, T. and Conte, K. (2008) „Export opportunity scan for Jordan“, Second draft. International
Trade Centre UNCTAD/ WTO (ITC), Market analysis and research section. April.
Gencturk, E. F. and Kotabe, M. (2001) “The effect of export assistance program usage on export performance: a contingency
explanation”, Journal of International Marketing, Vol. 9(2), pp. 51-72.
Green, R. T. and Allaway, A. W. (1985) “Identification of export opportunities: A shift-share approach”, Journal of Marketing, Vol. 49,
pp. 83-88, Winter.
Hirshmann, A. O. (1964) “The paternity of an index”, American Economic Review, Vol. 54, pp. 761, September.
Hoffman, J. J. (1997) “A two stage model for the introduction of products into international markets”, Journal of Global Marketing, Vol.
11(1), pp. 67-86.
Huff, D. L. and Scherr, L. A. (1967) “Measure for Determining Differential Growth Rates of Markets”, Journal of Marketing Research,
Vol. 4, pp. 391-395, November.
IMF (International Monetary Fund). 2006. World Economic Outlook.
http://www.imf.org/exteranl/pubs/ft/weo/2004/02/data/index.htm. Date of access: 12 September 2005.

33
 

International Trade Institute of Southern Africa. (2009) “International finance and payments”, Study material compiled for a
certificate in export practice.
ITRISA, see International Trade Institute of Southern Africa.
Jeannet, J. P. and Hennessey, H. D. (1988) “International Marketing Management: Strategies and Cases”, Boston: Houghton Mifflin,
pp.137-168.
Kumar, V., Stam, A. and Joachimsthaler, E. A. (1993) “An interactive multi-criteria approach to identifying potential foreign markets”,
Journal of International Marketing, Vol. 2(1), pp. 29-52.
Le Roux, R. (2009) Minutes of a Export Promotion Research Group (EPRG) meeting, Zebula, 24 April. (Minutes of the meeting in
possession of the authors).
MapCrow Travel Distance Calculator (2009) http://www.mapcrow.info. Date of access: 20 January 2009.
Moravcsik, A. M. (1989) “Disciplining trade finance: the OECD Export Credit Arrangement”, International Organization, Vol. 43(1),
pp. 173-205, Winter.
Netpas Distance (2009) http://www.netpas.net. Date of access: 30 January 2009.
ONDD (Office National Du Ducroire). Nationale Delcrederedienst. (2005) “Country risk summary table”
http://www.ondd.be/wenondd/Website.nsf/ Date of access: 15 November 2005.
Ozorhon, B., Dikmen, I. and Birgonul, M. T. (2006) “Case-Based Reasoning Model for International Market Selection”, Journal of
Construction Engineering and Management, pp. 940-947, September.
Papadopoulos, N. and Denis, J. E. (1988), “Inventory, taxonomy and assessment of methods for international market selection”,
International Marketing Review, pp. 38-51, Autumn.
Papadopoulos, N., Chen, H. and Thomans, D. R. (2002) “Toward a tradeoff model for international market selection” International
Business Review, Vol. 11, pp. 165-192.
Pezeshkpur, C. (1979) “Systematic Approach to Finding Export Opportunities” Harvard Business Review, pp. 182-196, September-
October.
Rahman, S. H. (2003) “Modelling of international market selection process: a qualitative study of successful Australian international
businesses”, Qualitative Market Research: An International Journal, Vol. 6(2), pp. 119-132.
Russow, L. C. and Okoroafo S. C. (1996) “On the way towards developing a global screening model”, International Marketing Review,
Vol. 13(1), pp.46-64.
Sakarya, S., Eckman, M. and Hyllegard, K. H. (2007) “Market selection for international expansion: Assessing opportunities in
emerging markets“, International Marketing Review, Vol. 24(2), pp. 208-238.
Shankarmahesh, M. N., Olsen, H. W. and Honeycutt, E. D. (2005) “A dominant product-dominant country framework of industrial
export segmentation”, Industrial Marketing Management, Vol. 34, pp. 203-210.
Searates.Com (2009) http://www.searates.com/referebce/portdistance/. Date of access: 30 January 2009.
Steenkamp, J. E. M. and Ter Hofstede, F. (2002) “International market segmentations: issues and perspectives“, International Journal of
Research in Marketing, Vol. 19, pp. 185-213.
Rossouw, R., Steenkamp, E.A. and Viviers, W. (2009), “Identifying export opportunities: the construction of a decision support
model for evaluating and identifying realistic export promotion opportunities for South Africa”, Report delivered as part of a
research project commissioned by the Department of Trade and Industry for South Africa, Pretoria, pp. 67-81, Unpublished.
Walvoord, R. W. (1983) “Export market research”, Global Trade Magazine, p. 83, May.
Wilkinson, T. and Brouthers, L. E. (2006) “Trade promotion and SME export performance”, International Business Review, Vol. 15, pp.
233-252.
Willeme, P. and Van Steerteghem, D. (1993) “Een normatief model voor de planning van exportbevorderende activiteiten van de
Vlaamse Dienst voor de Buitelandse Handel (update)”, unpublished report, Vlaamse Dienst voor de Buitelandse Handel, Brussels.
World Bank (2005) “World development indicators”, Sept 2005. http://devdata.worldbank.org/data-query/. Date of access: 12
September 2005.

34