공간패널모형을 이용한 산업집적의 고용 효과 분석

This paper analyzes industrial agglomeration effects. Unlike most previous studies, here the focus is on spatial interaction. Using establishments data covering the years 1998 to 2006, low-level industry-region cells are defined and analyzed through spatial panel models. Empirical results show that, for a wide variety of parameters and spatial weighting matrices, coefficients associated with spatial interactions are highly statistically significant. Direct and indirect effects derived from inter-model testing suggest that the coefficients(sign and absolute value) and statistical significance of non-spatial panel model estimates show substantially different values. In terms of industrial agglomeration, evidence of MAR externalities were found for hiring by incumbent firms while Porter externalities were found for new entrants. However, in contrast with usual results in foreign studies, Jacobs externalities were found to be statistically insignificant in the Korean data. These results imply that the issue of spatial dependence should be considered when analyzing local-level industry and employment structures. Also, the paper’s evidence supports the need for industrial policies geared toward stimulating Jacobs externalities.