基于泰勒法则的人民币汇率预测

发布时间：2018-07-04 16:08

本文选题：汇率 + 泰勒法则　；参考：《复旦大学》2014年硕士论文

【摘要】：本文主要研究汇改后人民币汇率的预测模型,通过构建开放经济条件下中美两国的泰勒法则公式,结合无抛补的利率平价公式,推导出人民币兑美元汇率与宏观经济变量(产出缺口、通货膨胀率、实际汇率、名义利率)之间的关系。本文选取了2005年8月至2013年12月总共101个月的数据,分别用HP滤波法和结构化VAR模型估计了产出缺口,建立汇率波动的预测模型,以最后3年为预测期,分别使用了固定样本回归和滚动样本回归对模型的预测能力进行分析。同时,本文选择随机游走模型作为基准模型,通过考察MSPE、CW统计量、TU统计量和DMW统计量,对泰勒法则模型与无抛补利率平价模型、弹性货币模型和购买力平价模型的预测能力进行对比。本文的研究结果表明,泰勒法则模型在中短期对汇率波动的预测能力明显优于无抛补利率平价模型、购买力平价模型和弹性货币模型,其预测的MSPE在所有模型中最小。作为以经济理论为基础的预测模型,泰勒法则模型考虑了决定汇率中短期波动的货币政策方面的因素,在一定程度上弥补了传统模型的缺陷。
[Abstract]:This paper mainly studies the prediction model of RMB exchange rate after remittance. By constructing the Taylor rule formula of China and the United States under the open economic conditions and combining the formula of the non tossed interest rate parity, the relationship between the exchange rate of RMB and the macroeconomic variables (output gap, inflation rate, real exchange rate, nominal interest rate) is deduced. The data of 101 months from August 2005 to December 2013 are taken. The output gap is estimated by HP filtering and structured VAR model, and the prediction model of exchange rate fluctuation is established. The prediction ability of the model is analyzed using the fixed sample regression and rolling sample regression in the last 3 years. Taking the model as a reference model, by examining MSPE, CW statistics, TU statistics and DMW statistics, the prediction ability of the Taylor's rule model and the non tossed interest rate parity model, the elastic money model and the purchasing power parity model is compared. The results of this paper show that the prediction ability of the Taylor method is clear in the middle and short term on the exchange rate fluctuation. The MSPE model, which is based on the economic theory, considers the factors that determine the monetary policy of the short-term volatility in the exchange rate, to a certain extent, to make up for the deficiency of the traditional model. Depression.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.6

【相似文献】