On the Probabilities and Determinants of Global Sourcing in Korean Industry

Based on a recently constructed firm-level panel dataset, Survey of Business Activities, 2006~08, from the National Statistics Office (NSO), we empirically analyzed probabilities and determinants of global sourcing in Korean industry by setting up a panel logit model and a panel tobit model.This paper is likely the first attempt to utilize firm-level panel dataset in the study of global sourcing, which can be regarded as an important step forward in this field in the sense that the dataset could correct the chronic distortion that conventional Input-Output data causes, which has been taken to substantially overestimate global sourcing since, besides global sourcing, it also includes ordinary trade not classified to global sourcing.The results show that a larger firm that invests more in R&D, exports more, and is more informatized appeared more likely to engage in and expand global sourcing significantly. According to industry, the effect of lagging returns on invest, capital-labor ratio, and number of patents derived from global sourcing appears mixed. We might arguably obtain a more rigorous result than existing studies that rely only on cross-sectional or IO data by taking the endogeneity problem and industry/firm characteristics into consideration. A set of research topics for subsequent study/analysis was presented with the conclusion