Transport infrastructure, urban growth and market access in China

This paper quantifies the causal effects of various types of investments in the road and railroad networks on economic growth in Chinese cities and regions. We separate out the influences of changes in access to markets that have come through better inter-regional and international transport links from the more direct effects of transport infrastructure on city level productivity, which may operate through various channels. We find strong evidence that improved integration with nearby markets significantly promoted local growth in China since 1990. In particular, expansions of road infrastructure leading to a 10 percent increase in economic activity within a six hours' travel time led to an estimated 1.4 percent more rapid prefecture GDP growth and 1.1 percent more rapid prefecture city GDP growth. Expansions of road infrastructure leading to a 10 percent increase in population within six hours caused an estimated 1.6 percent increase in prefecture GDP growth and an imprecisely estimated 1.3 percent increase in prefecture city GDP growth. Estimated causal effects of more theoretically grounded measures of market access on local growth are consistent with these effects of more reduced form market potential measures. While we find that improved regional integration promoted local GDP growth in China, we find no significant effects of regional integration on prefecture or city population growth. Instead, we find evidence that improved access to international ports promoted population growth. A 10 percent decline in travel time to an international port caused a 0.6 to 0.7 percent increase in prefecture and city population growth. The context of severe migration restrictions in many cities and policies promoting foreign investment in other cities is important for interpreting these results. Our investigation of the effects of highways and railroads serving prefecture cities is less conclusive. While point estimates tend to be positive, they are generally imprecise. This study innovates on the existing literature about the effects of reduced domestic trade costs on local growth in several ways. First, we consider both highways and railroads. Second we primarily examine cities, rather than rural counties or small towns. Third we examine the effects of transport infrastructure on the growth and redistribution output and population simultaneously, rather than on inferred income or the output of specific commodities. Finally, we examine the responses to various measures of the composition of output in the regions surrounding cities in various distance and travel time bands. Critical to this evaluation is the use of pseudo-random variation in the allocation of transport networks to cities and their surrounding regions.