Abstract
We investigate whether the convergence hypothesis in 13 Korean regional economies for the period 1985 to 2011 exists allowing technological spillover and spatial interdependence. We use the spatial dynamic panel data approach including the spatial Durbin model. When we use the spatial lag model (SAM) with spatial fixed, we find that there is a positive spillover effect of growth among Korean regions and that convergence speed rate is faster than the model without spatial interdependence. However, when we use the SAM with time period fixed, or spatial fixed and time-period fixed effects, these results do not support convergence hypothesis of regional incomes. Therefore, we have to be careful to approach conclusion about convergence hypothesis in Korean regional incomes.
Notes
1 Under decentralization policy, Gwangju was created from Jeonnam after 1987, Daejeon from Chungnam after 1989 and Ulsan from Gyeongnam after 1998. It is necessary to make some adjustments in the raw data since we consider the income level combining original identical community area and smooth the data series with balance panel for test as possible. Therefore, three regions, Gwangju, Daejeon and Ulsan, were incorporated into the original region.
2 The extended Solow model adds the enrollment rate as suggested by Mankiw et al. (Citation1992) and local tax revenue.
3 We do not show the results due to space constraints.