Following recent studies on evolving stock markets in some of the European transition economies, we revisit the weak-form efficiency of China's stock markets by examining its changing behavior over the entire history of the Shanghai and Shenzhen Stock Exchanges for which data are available. The Kalman Filter estimation technique is applied to the system consisting of a time-varying AR(2) model and an asymmetric ARCH equation. The estimates of predictability combined with other non-quantifiable, evolutionary characteristics of the markets are used to infer on their efficiency. Our research shows that, at their initial development stages, both the Shanghai and Shenzhen markets were inefficient, but the latter is less so than the former. However, the past decade clearly saw a steady convergence of the two markets towards efficiency, owing to the improvements in the market infrastructure, market liquidity, regulation enforcement, and so on
