Determining the RMB Exchange Regime Using Linear and Nonlinear ANN Approaches

摘要

Since 2005 China has claimed that it will move towards a more market-oriented system of managing its foreign exchange. China, however, has remained in part a managed economic system. This article examines the relative importance of fundamentalist, chartist and currency arrangements in determining the RMB exchange regime using both traditional linear models and artificial neural networks (ANN). We find that the emphasis on the US Dollar as a reference currency has declined. Fundamentalist forces are strong determinants of the currency exchange. The RBF ANN model is among the best performing in minimizing forecasting error.