![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
ABSTRACT
The North African countries’ (NACs) production and export structure suffers from double constraints: insufficient diversification and weak sophistication. This study assesses the impact of diversification/sophistication variables on growth in the NACs and verifies that the current export structure is indeed a constraint to economic development. We estimate a growth model as a Barro's regression (conditional-convergence model) using panel data. The factors that determine export diversification and sophistication suggest which levers and actions would firstly allow NACs to diversify their exports to higher added-value products and secondly take the existing products to a higher level of sophistication. Our recommendations highlight the role of various policies and stakeholders.
RÉSUMÉ
La production et la structure d'exportation des pays de l'Afrique du Nord souffre d'une double contrainte: diversification insuffisante et faible sophistication. Cette étude évalue l'impact des variables de diversification/sophistication sur la croissance des pays nord africains et valide que la structure d'exportation est effectivement une contrainte au développement économique. Nous estimons un modèle de croissance comme une régression Barro (modèle de convergence conditionnelle) en utilisant des données de panel. Les facteurs qui déterminent la diversification et la sophistication des exportations suggèrent quels leviers et quelles actions pourraient tout d'abord permettre aux pays de l'Afrique du Nord de diversifier leurs exportations vers des productions à plus grande valeur ajoutée et ensuite conduire les produits actuels vers un plus grand niveau de sophistication. Nos recommandations soulignent le rôle de différentes politiques et intervenants.
Acknowledgements
The authors wish to thank the participants of the expert group meeting on diversification and sophistication in North Africa, held in Rabat, 26–27 February 2013. The views expressed in this article are those of the authors and do not necessarily represent those of their institutions.
Notes on contributors
Nizar Jouini is on the faculty of the University of Tunis, Tunisia and has served as a consultant for the African Development Bank.
Nassim Oulane is a senior economic affairs officer and chief of the Sub-Regional Data Centre at the North Africa Sub-Regional Office of United Nations Economic Commission for Africa (UNECA), located in Rabat, Morocco. Previously he worked in France's Forecasting Directorate.
Nicolas Péridy is a professor of economics in the Faculty of Economics and Business of the University of Toulon, France, and a member of Laboratoire d'Économie Appliquée au Développement (Laboratory of Applied Economics and Development) at the university.
Notes
1. Among the indicators used, beside the standard measurements of concentration and diversification (Herfindahl–Hirschman), we shall use the recent approaches developed particularly by Cadot, Carrère, and Strauss-Kahn (Citation2011), who propose a decomposition of Theil's concentration index, making it possible to distinguish the intensive and extensive margins of export diversification. With regard to measurements of sophistication, we shall use the indicator of Hausmann, Hwang, and Rodrik (Citation2007) and Hidalgo and Hausmann's (Citation2009) economic complexity index.
2. As with many empirical studies of this kind, the model specified above is not a log–log type on account of the presence of variables with a negative sign. However, a sensitivity analysis including a log–log model without the negative variables gives us results very close to those presented below in terms of sign and significance of the parameters.
3. This index is defined as follows:with hij = share of product i in total exports (or imports) of the country or group of countries j and hi = share of product i in total world exports (or imports). This index, between 0 and 1, shows if the structure by product of a country's or group of countries’ exports diverges little or greatly from the structure by product of total exports in the world. The closer the index is to 1, the greater the divergence.
4. This variable will be tested so as to verify in particular the resource curse hypothesis.
5. The estimate of the model that is strictly limited to the seven NACs (Algeria, Egypt, Libya, Mauritania, Morocco, Sudan and Tunisia) was made in a preliminary approach. The results obtained are close to those presented, particularly with regard to the sign of the parameters. However, the results are made less robust because we have only seven in cross-sectional observations, which considerably lowers the quality of the panel estimates. Thus, so as to ensure more robust results, the econometric analysis has also been to the Mashrek countries (Iraq, Jordan, Lebanon and Syria) and Turkey, a total of 12 countries in all. This enlarged sample makes it possible to improve the quality of the estimates firstly on account of a larger number of observations, and secondly as it makes it possible to increase the variance between the countries of the variables used in the panel. The results do not differ from those obtained with the NACs alone on account of the great economic weight of these countries, but the parameters are more significant on account of a larger number of observations in particular cross-sectional observations.
6. The results are also controlled in relation to multicollinearity using the vif test. This generally is in the neighbourhood of 5, a level below the generally tolerated threshold of 10.
7. The non-significance of the aggregated Theil index is explained by the fact that it is for the most part made up of the (non-significant) intra index.
8. So as to test the specificity of the NACs in relation to other countries among those 12 ultimately chosen, the model was estimated with, on the one hand, a diversification and sophistication variable applied to all 12 countries and, on the other hand, an interaction variable consisting in the same variable multiplied by a dummy variable taking 1 as the value for the NACs and 0 for the other countries. To the extent that the interaction variable is not significant, it may be concluded that the relation between diversification/sophistication and growth is not significantly different for North African countries from that of the 12 countries.
9. For example, the positive role of infrastructures in the NACs has been identified in Péridy and Bagoulla (Citation2012).
10. For the same reasons as previously mentioned, the sample of countries will be extended to the Mashrek countries and Turkey.
11. HOS = Heckscher Ohlin and Samuelson.
12. Taking account of problems of data availability with regard to loans from financial institutions, the variable tested here is limited to loans from the European Investment Bank.
13. With regard to financial liberalisation, the variable used (share of domestic credit to the private sector as a percentage of GDP) is very general and does not directly address the finer sub-variables such as guarantees for export, the rate of non-performing loans or rationing. However, the variable can be used to explain the positive effects of financial liberalisation on the diversification of the economies of the North African countries. These results are also consistent with the Mélitz-type framework, which explains this positive relationship with the entry of new exporters due to the easing of liquidity constraints related to financial liberalisation.
14. As in Ben Hammouda, Oulmane, and Sadni-Jallab (Citation2009), we have tested the existence of non-linearities relating to investment. We confirm some results relating to public investment, which raises diversification up to a certain threshold before lowering it. However, the results highlighted here suffer from problems of multicollinearity, making them less robust.
16. There was a specific estimate done for these variables on account of the serious problems of multicollinearity.
17. The indicator of legal rules gives slightly better results than that concerning corruption but is barely significant at a level below 20 per cent.