基於GARCH類模型的人民幣匯率波動性研究

Research on the volatility of RMB exchange rate based on GARCH model

王晟(Sheng WANG) ； 趙一瑋(Yiwei ZHAO) ； 趙鈺(Yu ZHAO)

《IETI Transactions on Social Sciences and Humanities》 7卷 (2020/05) Pp. 220-227

https://doi.org/10.6896/IETITSSH.202005_7.0029

摘要

匯率作為開放經濟體系最為基本的宏觀經濟要素之一，其變化對國際經濟金融有著不可替代的影響。匯率的波動存在波動率聚集、尖峰厚尾和非對稱性等特點。本文搜集了2007至2017年的美元兌人民幣的日平均中間價作為樣本數據並進行處理，取對數差分建立波動率序列。基於GARCH模型探究其波動的特徵和原因，並通過預測驗證了模型的可行性。本文分析了人民幣匯率在三個階段波動的原因及影響，包括08年次貸危機、15年央行調整人民幣匯率報價制度、17年外匯儲備調整。發現匯率對數差分序列隨著匯率市場化程度的不斷深化，波動也隨之明顯。在處理數據後，進行了GARCH模型的適用性檢驗，包括平穩性核對總和ARCH效應檢驗，發現匯率波動的對數差分序列適用於GARCH（1,1）模型。在建立GARCH（1,1）模型後本文擇選了2015年1月21日至2017年10月31日共677個樣本數據進行模型檢驗，對2017年11月的人民幣匯率進行預測並與真實值比較。發現誤差在可接受範圍內，證明了模型真實可靠。利用2007年至2017年的數據建立GARCH（1,1）模型，對匯率波動性進行分析後，本文針對人民幣匯率波動的規律性和特徵對現階段下的貨幣政策，提出了完善浮動匯率制度、重視和規範數位貨幣及建立健全衍生品市場的看法和建議。

關鍵字

匯率制度 ； GARCH模型 ； 波動性聚集 ； 匯率預測 ； 外匯儲備

並列摘要

As one of the most basic macroeconomic factors of open economy, exchange rate has an irreplaceable influence on international economic and finance. The fluctuation of exchange rate is characterized by fluctuation rate, thick tail and asymmetry. In this paper, the average middle price of USD/RMB per day from 2007 to 2017 is collected and processed as sample data, and the fluctuation rate sequence is established by taking logarithmic difference. Based on the GARCH model, the characteristics and reasons of its fluctuation are explored and the feasibility of the model is verified by prediction. In descriptive statistics for 2431 data, this paper analyzes the reasons of the RMB exchange rate in three phase fluctuation and its influence, including 08 the subprime crisis, 15 years, the central bank adjusted the quotation of RMB exchange rate system adjustment, 17 years of foreign exchange reserves. It is found that the fluctuation is obvious with the deepening of the exchange rate marketization degree. The analysis indexes mainly include mean, standard deviation, skewness, kurtosis and j-b statistics. After processing the data, the applicability of the GARCH model was tested, including the stationarity test and the ARCH effect test, and the logarithmic difference sequence of the exchange rate fluctuations was found to be applicable to the GARCH(1,1) model. After the establishment of GARCH(1,1), this paper selects 677 sample data from January 21, 2015 and October 31, 2017 to conduct model test, and compares the RMB exchange rate in November 2017 with the real value. The error is found within acceptable range and the model is true and reliable. Using data from 2007 to 2017 to establish a GARCH (1, 1) model, after the analysis of exchange rate volatility, in this paper, according to the regularity and characteristics of RMB exchange rate fluctuations on the present stage of monetary policy put forward to perfect the floating exchange rate system, value and standard digital currency and establish and improve the derivatives market views and Suggestions.

並列關鍵字

Exchange rate regime ； GARCH model ； Volatility gathering ； Exchange rate forecast ； Foreign exchange reserves