The determinants of Turkey's exports to Islamic countries: The impact of political risks

ABSTRACT

Using the traditional gravity model, this paper aims to analyze the determinants of Turkish exports to 43 Islamic Development Bank member countries for the period from 1996 to 2015. The paper specifically investigates the effects of 12 political risk measures (bureaucracy quality, corruption, democratic accountability, government stability, internal and external conflict, investment profile, law and order, military in politics, religious and ethnic tensions, and socioeconomic conditions) in the importing countries on the total volume of exports of Turkey. After implementing various robustness checks, the paper finds that the government instability in the importing countries is negatively associated with the Turkish exports.

KEYWORDS:

• Gravity model estimation
• south–south trade
• export market diversification
• political risks
• Turkish economy
• panel data estimation techniques

JEL CLASSIFICATIONS:

• F14
• F50
• C33

Acknowledgments

This research (project number S–BAG–2016–834) was supported by the Scientific Research Projects Coordination Unit of Istanbul Medeniyet University. We would like to acknowledge them for their financial support. We would also like to thank anonymous reviewers for their comments and suggestions, which significantly improve the paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. For example, the effects of the level of democracy (Nunn 2007; Yu 2010), the economic freedom (Depken and Sonora 2005), the rule of law (Levchenko 2007; Nunn 2007; Ranjan and Lee 2007) on international trade have been examined.

2. According to the data from the foreign trade statistics database by the Turkish Statistical Institute, the share of the EU countries in the total Turkish exports is 48.5% for the period from January 1996 to May 2017.

3. We restrict the analysis on the IDB countries and assess the impact of political instability within Islamic countries, whereas these countries are more unstable than others. This is the main limitation of the paper. Moreover, the European countries are also important markets for Turkish exports, but their political risks seem to be quite homogeneous. Therefore, we neglect the European countries in the empirical analysis.

4. According to the International Monetary Fund data for the GDP with the international dollar in the purchasing power parities; China, India, Russia, Brazil, Indonesia, Mexico, and Turkey are respectively the 1st, the 3rd, the 6th, the 7th, the 8th, the 11th, and the 13th largest economies in the world in 2016.

5. The share of exports of Turkey to the IDB member countries was 36.2% in 2012, but it has declined since 2012, probably due to the political issues in most IDB member countries.

6. The first five variables are indexed from 0 to 12. The variables from 6 to 11 are scaled from 0 to 6. The bureaucracy quality is indexed from 0 to 4.

7. Albania, Algeria, Azerbaijan, Bahrain, Bangladesh, Brunei, Burkina Faso, Cameroon, Cote D'Ivoire, Egypt, Gabon, the Gambia, Guinea, Guinea-Bissau, Indonesia, Iran, Iraq, Jordan, Kazakhstan, Kuwait, Lebanon, Libya, Malaysia, Mali, Morocco, Mozambique, Niger, Nigeria, Oman, Pakistan, Qatar, Saudi Arabia, Senegal, Sierra Leone, Somalia, Sudan, Suriname, Syria, Togo, Tunisia, the United Arab Emirates (UAE), Uganda, and Yemen.

8. It is important to note that y_it   is country pair invariant as it is the evolution of Turkey's GDP that is constant on the cross-dimension of the panel. Therefore, this variable is perfectly collinear with time dummies. lnD_ij, border_ij, colony_ij, which are also constant over time; and therefore, they are perfectly collinear with the country dummy (recall that a country dummy would be d_j = 1 or 0, but the ‘j’ country is always Turkey for these dummies to replicate the fixed-effects). As a consequence, it is not possible to estimate the fixed-effect model in our model framework.

9. The empirical models also consider the dummy variables for the free trade agreements between Turkey and the IDB member countries (Albania in 2008, Egypt in 2007, Jordan in 2011, Malaysia in 2015, Morocco in 2006, Syria in 2007, and Tunisia in 2005). However, the coefficient of free trade agreements has been automatically dropped due to the multicollinearity with other control variables.

10. Although the gravity model of Anderson and Van Wincoop (2003) is estimated by the fixed-effects estimation, the fixed-effects estimation does not allow for the effects of exporting- and importing-invariant variables (e.g. border country and colonial history). Indeed, the fixed-effects allow controlling for multilateral resistance that makes the gravity equation theory-founded. However, implementing both the country and the year fixed-effects does not allow to estimate the impact of political instability that has the same dimension. In addition, as we discussed in Footnote 8, it is not possible to estimate the fixed-effect model for our case. Therefore, we consider the random-effects estimations, which are also a common approach in the literature (e.g. Prehn, Brümmer, and Glauben 2016).

11. For details, see the export-led growth hypothesis and Yu (2010).

12. Note that 8.65% = 100 × (e^0.083 − 1).

Additional information

Funding

Scientific Research Projects Coordination Unit of Istanbul Medeniyet University [grant number S-BAG-2016-834].