摘要
本文构建了符合人民币外汇市场微观结构的噪声交易模型,并利用马尔可夫机制转换模型对其进行估计,同时采用自2014年3月扩大汇率浮动幅度以来至2015年3月的日数据实证研究了人民币汇率的动态决定过程。估计结果显著支持我国人民币汇率两机制转换的MS-AR(2)模型。实证结果表明,研究样本区间内,人民币汇率大多处于基本面机制之中。基本面机制具有突出的持续性,而噪声机制具有剧烈的波动性。在基本面机制时,人民币汇率主要受中美隔夜利率差的基本面因素影响。在噪声交易机制时,人民币汇率主要受汇率的过去趋势和即期汇率与中间价汇率的边界效应的噪声因素影响。
This paper investigates the dynamic determinants of RMB exchange rate by developing a noise trading model estimated with Markov Regime switching model and employing the daily data from March 2014 to March 2015. The estimation result reveals that MS-AR (2) captures the properties of two mechanisms: fundamental and noise trading mechanisms. The empirical results suggest that the fundamental mechanism is relatively persistent while the noise trading mechanism is relatively volatile. The results also show that the RMB exchange rate is determined by the interest rate difference between the money markets in China and the US under the fundamental mechanism, while the RMB exchange rate is determined by the past trend, and the noise factor of difference between spot rate and mid exchange rate under the noise trading mechanism.
出处
《国际金融研究》
CSSCI
北大核心
2016年第7期74-82,共9页
Studies of International Finance
基金
教育部规划基金项目(14YJA7900)
国家社科基金重大课题(14ZDA020)的资助