Statistical and economic applications of Chinese regional input-output tables

China has a long tradition in compiling Input-Output (I0) tables at the regional level. For 30 of its 31 provinces, survey-based 10 tables have been constructed by regional statistical bureaus. These bureaus follow a uniform compilation scheme prescribed by the National Statistical Bureau, which yields a unique set of harmonized tables. This thesis uses the regional 10 tables for 1997 d 2002, in two ways. In the statistical applications, the wealth of data is used evaluate existing methods for compiling 10 tables, if ta cannot be collected by means of full surveys among firms. Furthermore, some novel methods are proposed and tested against popular alternatives. Some of the results are particularly relevant for data construction efforts in China (a dynamic economy characterized by large disparities between regions), whereas other results have a more general validity. In the economic applications, China's huge regional disparities in labor productivity levels and in the presence of ICT-production activities are analyzed. Existing shift-share techniques and structural decomposition analyses are adapted (using insights from index number theory) to quantify the contributions to these disparities of several structural differences between regional economies. The results suggest that the different positions Chinese provinces have in global value chains play an important role in the dispersion of labor productivity levels. Being close to regional markets for ICT products appears relatively unimportant in explaining the strong geographical concentration of these industries.