ANALYZING THE SPATIAL STRUCTURE OF TURKEY?S SECTORAL AND REGIONAL EMPLOYMENT

This paper employs a beta convergence analysis for the regional employment rates in the agriculture, industry and services sectors in Turkey during 2004-2011. After performing a standard seemingly unrelated regression (SUR) model for the sectoral equations, spatial variations are also taken into account in the second part of the paper. For this purpose, first a random effects panel data model with spatial error components is utilized for the convergence problem of each sector. Then, considering possible contemporaneous correlation between these sectoral equations, a panel SUR model with spatial error components and random effects is estimated. Feasible generalized spatial three stage least squares estimations give sector-specific outcomes indicating that employment convergence analysis should be carried out by considering not only the spatial but also the sectoral interactions. In particular, the dynamics of employment rates in services sector in comparison with that of the agriculture sector has important characteristics in explaining the development path of the regions in Turkey.