基于STAR模型的中国货币需求函数研究

A STAR-model-based Functional Research of China's Monetary Demand

本文对我国货币需求的短期误差修正模型进行非线性检验并发现指数平滑转换自回归模型更加符合我国短期的货币需求函数。通过检验,我们发现非线性模型的转换变量为滞后一期的真实国民收入,该结论与我国经济发展状况相吻合。另外,通过检验发现非线性模型具有较强的预测能力。因此,非线性模型将会给货币政策的正确制定和有效实施提供更具参考价值的作用。

Firstly, through non-nested tests, this paper chooses interest rate as the proxy of opportunity cost of holding money and on this basis, a long-run equilibrium equation is established. After that, by application of ESTAR Model, this paper makes a nonlinear empirical test for the monetary demand in China. The result shows that the model is of significant nonlinear characteristics and the transition variable is a real reflection of national income for the next period. The conclusions above are consistent with China＇s eeonomie development. Finally, by two indicators： RMSE and MAE, this paper compares the forecasting ability between ESTAR model and the traditional linear model, and the result indicates that the ESTAR model is better. This article suggests the nonlinear model has better statistical properties than traditional linear model.