A dynamic panel data approach and HCW's method: Assessing the effect of China (Shanghai) Free Trade Zone on local GDP

Abstract

We propose a simple-to-implement dynamic panel data method to evaluate the impacts of place-based policies. The idea is to exploit both the cross sectional dependence and the serial correlation within a panel and implement a semi difference-in-difference decomposition. Different from the method by Hsiao et al. (2012), the proposed model is easy to implement statistical inferences and can incorporate multiple treated groups at different time thresholds. We use the proposed method to explore the effect of the Free Trade Zone (FTZ) policy implemented in Shanghai at the end of 2013 on local GDP growth rate and compare it with that from Hsiao et al. (2012). The empirical results show that the FTZ does have a positive effect on the local GDP growth rate and it adds 1.2−1.8 percent point to the growth rate of the GDP per capita for Shanghai in 2014 which is about one fourth to one third of the total GDP growth rate per capita. This effect largely comes from the growth of the tertiary sector. The results lend support to the duplication of FTZ policy to other cities and provinces in China.

Keywords

• Dynamic panel model
• Difference-in-Difference
• Free trade zone
• GDP growth
• Treatment effect