Funding from SCID will be used primarily for building and analyzing the spatial dataset of China’s banking system. The data will cover information on the location, ownership and operation status of every single bank unit in PRC China (N> 200,000). The basic idea of this data extraction process is that we first collect registration information (e.g. name, address, open date) of every single bank branch that has ever been built in China, and take it from there to identify the geographic information (coordinates) of each bank branch by taking advantage of Baidu map (China’s Google map). The dataset will allow researchers to answer such basic yet crucial descriptive questions as “How many banks are there in locality X in China?” “Who controls them?” None of the existing datasets can do this. A variety of measures of banking development and competition can be developed using this unique dataset. One such measure can be built on the idea of the Herfindahl index. To test proposed hypotheses, this spatial dataset will be merged with existing ones, such as the China Industry and Enterprise Database. The key identification strategy is using arbitrarily created grid squares as the unit of analysis, which can largely solve the endogeneity problem—the location of these squares are not possibly endogenous to any socioeconomic factors that might determine the location of the banks.