Spatial Econometric Analysis of R&D Spillovers in Turkey

Abstract

Research and development (R&D) activities are of great significance in the long term development of firms and R&D expenditures have been studied by economists including Romer. This paper sets out to investigate the effects of R&D spillovers at provincial level with specific reference to Turkey. To serve this purpose, data published by Turkish Statistical Institute have been employed. The paper covers the period from 2003 to 2007. The number of data points is 342 subsequent to the aggregation of data at the firm level. Of the data aggregated, it has been found out that 162 of them are zero. While running the spatial econometric estimations, inverse distance and neighboring matrices were employed as the weighting matrix. The determinants of R&D activities employed in the study are size, represented by the number of employees; qualified labor, represented by the number of R&D employees; technology transfer represented by expenses for licensing; foreign ownership, Pavitt sector dummies, location dummies by the sea, border or airport. The results of the analyses suggest that when both weighting matrices are used, spatial lag and error yielded significant results. There is R&D knowledge spillovers at provincial level in Turkey, shown by the spatial spillover effects in nearly one third of the total effects.