CCER讨论稿:Highways and Firms’ Exports: Evidence from China

发布日期:2018-01-21 20:03:00    来源:北京大学国家发展研究院

E2018003                       2018-01-21

Dan Liu

Shanghai University of Finance and Economics

Liugang Sheng

Chinese University of Hong Kong

Miaojie Yu

Peking University



This paper investigates the impact of highway networks on firms’ exports. The paper constructs a firm-level measure of highway access using GIS information on firm locations and highways, and adopting an instrumental variable approach. The analysis finds that improvement in highway access significantly promotes firms’ exports, after controlling for firm characteristics in addition to unobservable city and industry time-varying factors. Highway access also helps firms to expand the scope of their exports and imports. More interestingly, the positive effects of highway access on exports are stronger for less productive firms. With greater highway access, less productive firms lower their markup, but more productive firms increase their markup.     

Key words: Exports, Highways, Product Scope, Markup, GIS
JEL codes: F14, H54, L25, L92


* Dan Liu, School of Economics, Shanghai University of Finance and Economics (SUFE), email: Liugang Sheng, Economics Department, the Chinese University of Hong Kong, email: Miaojie Yu, National School of Development, Peking University, email: Dan is thankful for financial support from the Program for Innovative Research Team of Shanghai University of Finance and Economics, and the National Natural Science Foundation of China, project no. 71703187. Liugang is thankful for financial support from the Research Grants Council of the Hong Kong SAR (project no., CUHK 15056516). Miaojie thanks financial support from National Natural Science Foundation of China (No. 71625007, 71573006) and National Social Science Foundation of China(16AZD003). We thank all participants in seminars, the NSD summer institute in Yantai, and the Workshop on International Economics in Honor of Professor Robert Feenstra at Peking University. We thank Feng Chen for his assistance in GIS data processing. Special thanks to Dave Donaldson, Rob Feenstra, and Keller Wolfgang for comments and suggestions at the early stage. All remaining errors are our own.