抄録
This paper explores the location patterns of China’s automobile sector including 11 four-digit industries, using employment weighted K-density function proposed by Duranton and Overman (2005). We construct a detailed micro-geographic dataset based on the firm-level dataset of the 2nd China Economic Census. We classify the firms into SHE (state-holding enterprise), NSE (nonstate-holding enterprise), FFE (foreign-funded enterprise) and DE (domestic enterprise) by their ownership. We find the following mainly. First, in many industries, the NSEs, the FFEs and the DEs are localized at short distances, respectively, while SHEs are not localized in any industry. Second, in almost all industries, the enterprises show random distribution between the SHEs and the NSEs, and also between the FFEs and the DEs. Third, firms of 3720 (vehicle manufacturing) are not localized, and they are also not co-localized with the firms of other industries at short distances. Fourth, the NSEs and the FFEs in 3725 (vehicle parts and accessories manufacturing) show localization pattern at short distances. But they are not co-localized at short distances with the NSEs and the FFEs in the other industries, respectively. Fifth, the NSEs of almost all the industries show co-dispersion patterns with the SHEs of 3725 at short distances. Sixth, the DEs of almost all the industries show co-dispersion patterns with the FFEs of 3725 at short distances, while they show co-localization patterns at long distances, too. Seventh, the SHEs are not colocalized with the NSEs of 3725 at short distances, while the FFEs are not co-localized with the DEs of 3725 at short distances, too.
