China's 9.6 million square kilometres hide enormous economic disparities. Over 100 million poor people live in unproductive environments, mainly in western China. But even there pockets of wealth are disguised by the province-level data typically used for poverty research in China. Similarly disguised are the pockets of poverty in more prosperous provinces. Although China has well-regarded sample surveys, these are too small for measuring poverty below the province level. The Census can be disaggregated down to either county or township level but it does not the information on expenditures and incomes needed for measuring poverty.
In this project we use small-area estimation techniques (aka "poverty mapping") to combine survey and census data to measure poverty at fine scale in China. A feature of our project is that we emphasize geographical (location) and environmental factors and allow for "spatial dependencies". China has a unique record of local environmental change coming from satellite imagery with each one kilometre square parcel of land mapped three times since 1988.
Another key aspect of our project is that we operate as a virtual research organization, spread between three countries: China, the United States, and New Zealand. Our collaboration is enabled by KAREN, which provides high-speed, high-capacity internet to the research community, allowing us to connect geographically distributed data, tools and expertise.