Capital flow to China and the issue of hot money: an empirical investigation

This paper tries to model the time series characteristics of capital flows to China over the period 1999-2008, namely bond flows (BF), equity flows (EF), bank credit (BC), and foreign direct investment (FDI). By utilizing the state space model and using Kalman filtering algorithm with maximum likelihood estimation, we try to gauge the relative importance of permanent and temporary components of each series. And by incorporating intervention and explanatory variables, we also try to detect if capital control measure imposed by the Chinese government and market sentiment of RMB foreign exchange rate appreciation expectation have any effect upon those flows. The empirical result shows that: all four flows are dominated by transitory component, among which BC flows have a relatively large permanent component and are the only series that are sensitive to market sentiment measure. In addition, capital control measures successfully skewed flows to come in through FDI and bond flow channels instead of equity flows. And our extended model with intervention and explanatory variables for those flows have better prediction performance compared to Sarno and Taylor (1999a) and the benchmark models