Using a nonparametric panel data model, this paper estimates the degree of time‐varying and province‐specific capital mobility in China during 1970–2006. We estimate the savings–investment association, that is, the savings retention rate à la Feldstein and Horioka, as a measure of capital mobility. We also split the total savings (investment) into private and government savings (investment) to explore the role of government in improving capital mobility. Over time, we find an improvement in capital mobility after the mid‐1990s. Across provinces, we observe higher capital mobility in eastern/coastal regions. From the 1990s, the government is found to play a less important role in promoting capital mobility. Across provinces, the government is found to be more important in either the municipalities (Shanghai and Beijing) or the less developed inland provinces.