Geographic concentrations in Turkey's manufacturing industry: identifying regional highpoint clusters1

1. Previous versions of the paper was presented at the METU International Conference in Economics, September, Ankara, Turkey, 2002 and 2003.

Abstract

The theory of economic development based on industry clusters emphasizes the importance of co-location of firms or industries with potential to share technical information and knowledge transfer. The paper identifies industry cluster templates in Turkey using the 1996 Turkish input/output table. Engineering and textile clusters are the largest ones with respect to the number of establishments and employment. Majority of manufacturing clusters are located in İstanbul, İzmir, Ankara/Kırıkkale and Çukurova districts. The study points out the importance of newly developing centres near the periphery of Ankara and several production centres in Anatolia.

Notes

1. Previous versions of the paper was presented at the METU International Conference in Economics, September, Ankara, Turkey, 2002 and 2003.

2. For further information, see, http://www.competitiveturkey.org

3. The differences in sector classification between the two input/output tables (1990 and 1996) and industrial break-up of the sectors makes it difficult to make objective comparisons between Akgüngör, Kumral and Lenger (2003) and this study. The 1990 input/output table includes 64 sectors while the 1996 input/output table includes 98 sectors.

4. For a detailed description of the method, see, Feser and Bergman (2000). The following paragraphs that describe Feser and Bergman's methodology draws extensively from Akgüngör et al. (2003).

5. Through rotation, the factor matrix is transformed into a simpler and easily understood matrix. A rotation, which requires the factors remain uncorrelated, is an orthogonal rotation while others are oblique rotation. The correlation coefficients of the factors identified in the rotated component matrix are low, implying that the factors are not correlated. We therefore use an orthogonal rotation and do not impose a restriction to the analysis to assume that the resulting clusters are correlated.

6. For further information, see, http://www.hhh.umn.edu/centres/slp/projects/edweb/indcook.htm.

7. Location quotient (LQ) is a measure of the industry's concentration in an area relative to the rest of the nation. LQ = [(Industry's local employment)/(Total local employment)]/[(Industry's national employment/Total national employment)]. A location quotient greater than 1 means that the cluster employs a greater share of the local workforce than it does nationally. LQ value greater than 1.25 is considered to be an initial evidence of regional specialization (for further information, see, “Business Clusters in the UK” a report for the Department of Trade and Industry by a consortium led by Trends Business Research and Hubert H. Humphrey Institute of Public Affairs, 2000).

8. For an industry break-up of the clusters, see Appendix.

9. Criteria used in determining the primary and secondary sectors are explained in the method secton earlier.