SEPARATIST TERRORISM AND THE ECONOMIC CONDITIONS IN SOUTH‐EASTERN TURKEY

Abstract

This paper investigates the economic roots of separatist terrorism in Turkey. The political conventional wisdom is that poverty in highly Kurdish‐populated, south‐eastern Turkey is one of the most important causes of separatist terrorism and Turkish–Kurdish conflict in Turkey. Therefore, many economic policies have been implemented to improve the economic conditions in the south‐eastern part of the country. Using the Global Terrorism Database and Vector Autoregression (VAR) methodology, I find that there is no causal relationship between economic conditions in south‐eastern Turkey and separatist terrorism. Therefore policy‐makers should be cautious in using economic measures to prevent separatist terrorism in Turkey.

Keywords:

• Economics of terrorism
• Count data
• Separatist terrorism

JEL Classification:

• D71
• D74
• H56

ACKNOWLEDGEMENTS

I would like to thank Daniele Paserman for his excellent guidance and help throughout the project. I also thank Larry Kotlikoff, Marc Rysman, Zhongjun Qu, Yohei Yamamoto and Francesco Russo.

Notes

¹South‐eastern Anatolia Project Regional Development Administration (http://www.gap.gov.tr)

²State Planning Organization (http://www.dpt.gov.tr)

³ New York Times, March 12, 2008

⁴The authors merge data from different sources like MIPT, Rodoplu, Arnold and Ersoy (2004), Sebasteanski (2005), Turkish Daily News and Turkish Press.

⁵DHKP/C (Revolutionary People's Liberation Front), TKP‐ML (Turkish Communist Party‐Marxist‐Leninist Organization), TIKKO (Turkish Workers and Peasants Army, IBDA/C (Islamic Great Eastern Raider's Front), TIJ (Turkish Islamic Jihad).

⁶In 1993 there are no terrorist incidents in the GTD but seven separatist incidents in Feridun and Sezgin (2008). Also, in 1992, the number of incidents in GTD data set is very high, which might be a concern about the data set. The estimations using GTD are also done by using year dummies for years 1992 and 1993 for robustness checks and the results do not change significantly.

⁷Same estimations have been performed by using data from 1984 to 2004. The main results in the paper do not change significantly.

⁸For example, incidents involving soldiers or military personnel and terrorists are counted as terrorist incidents in newspapers.

⁹Domestic terrorism includes the terrorist incidents where the target and the terrorist's country of origin is the same.

¹⁰Separatist terrorism is committed by domestic nationals engaged in separatist causes.

¹¹ Hürriyet, 2 February 2010.

¹²GDP data is not available in regions like south‐eastern Turkey; therefore, I use province‐level GDP data to generate the GDP level in south‐eastern Turkey.

¹³In 1991 Siirt was dividend into three provinces: Siirt; Batman; and Şırnak. The data for these three different provinces was combined after 1991.

¹⁴Another comprehensive data set on terrorism is the MIPT (Memorial Institute of Prevention of Terrorism) database. In this paper I cannot use the MIPT data as well because MIPT has information on separatist incidents only after 1998. The MIPT includes data on international terrorism since 1968.

¹⁵The Denton Stata module has been used in calculations.

¹⁶Yearly estimation results are available upon request.