Trade in intermediate goods and Euro-Med production networks

Abstract

This paper examines the involvement of North African (NA) countries in regional production networks using a gravity model of trade augmented with imports of intermediate goods. The model is estimated using disaggregated bilateral exports of final and intermediate goods from four NA countries (Algeria, Egypt, Morocco and Tunisia) to OECD countries over the period 1995–2008. A first hypothesis argues that the Euro-Mediterranean process has led to a greater integration of NA countries in regional production networks. A second hypothesis states that manufacturing companies may transfer part of their production process to countries with lower labour costs, thereby generating increasing trade links between intermediate goods. Our results indicate that NA countries have indeed become more integrated in Euro-Mediterranean production networks and that this has a positive impact on trade flows between the two sides of the Mediterranean Sea. We conclude that the increase in exports from NA countries has been mainly channelled through changes in the rules of origin and the increase in imports of intermediate goods.

Keywords:

• production networks
• North African countries
• Euro-Mediterranean agreements
• rules of origin
• panel data

JEL Classification:

• F10
• F14

Notes

1. Jones and Kierzkowski (2005), page 7.

2. Due to political problems, the agreement concluded with Syria in 2004 was suspended a number of times.

3. Table A1 (Appendix A) shows the years in which each agreement came into force (Column 1) and the new RoO that have to be applied (Column 2).

4. Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Israel, Italy, Japan, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Turkey, UK and the USA.

5. The BEC is a three digit classification which groups (transportable) goods according their main end use. Other authors that have used it are Baldwin and Taglioni (2011) and Florensa et al. (2011). See Table A2.1 in Appendix A.

6. Fuels and lubricants (BEC-3) are excluded from the analysis.

7. Note that this is not necessary in this set-up as the right-hand side has importer-year and exporter-year fixed effects which are controlling for any variation in the GDP of the importer and the exporter.

8. The European Free Trade Association (EFTA) has also agreements with NA countries; therefore, both EuroMed and Pan_EuroMed_RoO dummies also include EFTA countries. See European Commission (2013) and Parra-Robles et al. (2012) for a list of FTA agreements, years and country members.

9. See Table A1 in the Appendix A.

10. See Appendix B. Note that this variable is constructed with exports or imports and then is potentially endogenous if it is included in the year t.

11. These two variables record a correlation of approximately 70%.

12. Results obtained by random effects including estimated coefficients for distance, contiguity, language and colony variables are available upon request.

13. Note that the variable of (lagged) imports of intermediate goods from the RoW is positive and statistically significant in regressions by random effects. As the coefficient of (lagged) imports of intermediate goods from the EU is greater than that of (lagged) imports of intermediate goods from the RoW, the rest of regressions rely on the use of that variable. In addition, results obtained by using a random effects specification show that distance and contiguity display a negative sign, whereas common language and colonial ties show positive signs.

14. These results help to explain the transmission channel of why imports of intermediate goods increased in the first place. They are available upon request.