This paper uses linear and nonlinear Granger causality tests to study the lead-lag relations among China's segmented stock markets. In contrast to the weak lead-lag relation among A- and B-share markets disclosed by its linear counterpart, a nonlinear causality test provides evidence of strong bi-directional causal relations between two A-share markets as well as between two B-share markets. In addition, the evidence shows that since the implementation of a new policy allowing domestic citizens to invest in B-share markets, A-share markets tend to lead their B-share counterparts in the same stock exchange and B-share markets continue to lead the H-share market. © 2007 Elsevier B.V. All rights reserved.