The traditional way of industrial production has transformed into high added-value products and services sector while the geography of manufacturing has been changing accordingly. On the other hand, both determinants and impacts of industrial location decisions and agglomeration economies have been crucial topics for researchers since Alfred Marshall (1920) (Krugman, 1991; Mccann, 2001; Fujita and Thisse, 2002; Capello, 2007). According to Isard (1956), who is the pioneer of implementing the location theory?s extended version in regional economy, location choice is not only caused by inter-regional differences, but also, it has a significant effect on the occurance of inter-regional disparities. The existence of industry in a region is fundamental for that region?s economical development by its stimulating effects on manufacturing, employment and other sectors. Even though the traditional location choice factors have been changing in time, initial advantages, reducing transportation costs, accessibility to market and accessibility to skilled labor pool are still significant determinants. (Fuijita and Thisse, 1996; McCann, 2001; Parr, 2002; Capello, 2007). In addition, the companies? competitiveness increases the region?s competitiveness. Therefore, industrial clusters are expected to be positively effective on the region?s economical development (Porter, 1990; Stimson et al., 2006; Keise, 2008; Capello, 2009). The purpose of this paper is to analyze the changes in the location of the industrial activities in a few decades to explore the determinants of agglomeration economies in Turkey. The previous studies on the related topic (Eraydin, 1992; Eraydin, 1999; Dincer et al., 2003; Dogruel, 2006) indicate that the industrial activities located at new industrial centers rather in addition to the traditional ones. This paper will initially present the changing and concentration pattern of industrial employment in different regions. Furthermore, the study would point out that the determinants of agglomeration may differ based on the location parameter by comparing the results of global and local regression analyses.
