Comparing International Market Selection Methods Using Export Potential Values for South Africa

ABSTRACT

Resources for export promotion are scarce, and incorrect market selection can be costly. In the literature, the gravity model, International Trade Center’s Export Potential Map, and decision support model all estimate export potential values. Although the same concept is measured, different methodologies are used. This study aimed to compare the rankings assigned to the export potential values by each approach. The results indicated that 45% of the rankings differed by more than five places, while one-third were ranked in the top 10 of only one approach. Given these inconsistencies, alternative approaches for prioritization after identifying export opportunities are recommended.

KEYWORDS:

• Decision support model
• export potential map
• export potential values
• export promotion
• gravity model
• international market selection

Acknowledgements

The authors would like to thank Professor Suria Ellis for assisting with the statistical analysis of this study, as well as Ms. Ali Parry for her invaluable technical inputs.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08853908.2023.2166628.

Notes

¹ A selection of 15 of South Africa’s main manufactured exports in 2015 were used in this study, based on data availability and comparability, and number of significant variables. All export potential values of the three approaches were, thus, estimated for 2015. A list of the products used can be found in Appendix 1 online. Furthermore, the top 50 South African export destinations for each product were included in the gravity analysis. In total, 580 product-country combinations were used for the analysis. A limited number of countries with missing values, and those subject to collinearity in the gravity model, were omitted in the calculation of export potential values. Also see Footnote 9.

² In the South African context, the double log specification of the gravity model has been primarily used specifically for the estimation of export potential values (see, for example, Breytenbach and Jordaan (2010), Jordaan (2011), Jordaan and Eita (2011, 2012), and Muronda (2018)). For this reason, this study also makes use of the double-log specification.

³ The 15 gravity datasets each contained data from 2007 to 2015. This is because the Logistics Performance Index (LPI) was available only from 2007 onwards, while the method of collection of the cost data from the World Bank Doing Business report changed after 2015. Therefore, if data outside of these years were used, it would have led to inconsistent and incomparable results. The LPI, as a proxy for trade facilitation, was, however, significant for only two products and was, therefore, included in HS 330499 and HS 842139 only. Instead of trade facilitation, only the costs to import were taken into consideration for the other products, since weak trade facilitation will ultimately increase the costs associated with trading goods and services (see Table 1).

⁴ Five products’ import and export values were, however, not significant, and the importer’s and exporter’s GDPs (sourced from the World Bank) were used instead. These products are: HS 290129 (unsaturated acyclic hydrocarbons), HS 720241 (ferro-silico-manganese in granular/powder form), HS 732690 (articles of iron/steel), HS 870410 (dumpers designed for off-highway use), and HS 880330 (parts of aeroplanes/helicopters).

⁵ This section relies heavily on Decreux and Spies (2016), who wrote the methodology for the Export Potential Map on behalf of the ITC.

⁶ Any product–country combination is considered to have export potential if it either has a total export potential value (to the world) of US$200,000 or more, or if it is partially covered by the 95% cumulative export potential of products that are ranked by the exporting country in descending order. Furthermore, products should have been exported over the past three years and also imported over the past five years to be included in the EPI (Decreux and Spies 2016).

⁷ The ITC’s Export Potential Map does have a few limitations. Of most relevance to this study is the fact that export potential assessments abstain from reporting any potential dollar value that may be related to identified diversification opportunities; therefore, only within-market rankings of products are provided (Decreux and Spies 2016). Furthermore, the costs related to export promotion activities are not considered in the Export Potential Map, which may influence the feasibility of exporting certain products.

⁸ This section relies heavily on the methodology explained by Viviers et al. (2014).

⁹ For this study, export potential values were only estimated using the gravity model and the Export Potential Map of the ITC for those countries that the DSM identified as having realistic export potential.

¹⁰ Later applications of the DSM for South Africa included an alternative analysis of trade barriers in Filter 3.2 (Steenkamp 2011). Trade costs were collected for each potential market worldwide. As cutoff values are based on world averages and South Africa’s transportation cost data are not readily available in an updated dataset, this would have involved an extensive and time-consuming data collection exercise. For the purposes of this study, where the focus is on the potential value calculation for a selection of 15 products, an update of this filter was, therefore, not included. Moreover, the potential value calculation of the DSM did not include market access variables.