基于ARIMA模型的人民币兑美元汇率进化研究

发布时间：2018-09-19 18:50

【摘要】：本文通过对2000年1月～2009年12月月平均汇率对数进行分析,建立1阶差分、2阶自回归、8阶滞后移动平均模型(ARIMA(2,1,8)),得出如下结论:汇率对数波动暨汇率进化源自新息冲击与汇率对数波动本身等两类因素,分别占83.3%和17.7%;其中当期和近2期新息冲击和汇率对数波动引起的汇率进化占55.2%;汇率对数波动对汇率进化影响具有时滞性(2期)和叠进交叉传递性(多期),新息冲击对汇率进化影响具有记忆性(1～8期)、相对长久性(0～8期)和稳定性(3～8期)。这些结论启示政府和外贸企业等市场主体,汇率改革和监管以及重大经济事项的出台,分三个阶段(即0～2期、3～8期和9期以后)影响汇率的波动,三个阶段的工作重心不应相同。模型对2010年汇率变动的预测(交替波动、整体升值趋势)符合1～3月份的实际情况,精度很高。
[Abstract]:This paper analyzes the logarithm of the monthly average exchange rate from January 2000 to December 2009. The first order difference 2 order autoregressive and 8 order lag moving average model (ARIMA (2 / 1 / 8), is established. The conclusions are as follows: exchange rate logarithmic fluctuation and exchange rate evolution are derived from two kinds of factors, such as innovation shock and exchange rate logarithmic fluctuation itself. Account for 83.3% and 17.7% respectively, of which the current and recent two periods of innovation shocks and exchange rate logarithmic fluctuations caused by the evolution of the exchange rate accounted for 55.2; the impact of logarithmic exchange rate fluctuations on the evolution of exchange rates has time-delay (2) and overlapping cross-transitivity (multi-period), innovation impulse. The impact of stroke on the evolution of exchange rate was of memory (1 / 8), relatively permanent (0 / 8) and stability (3 / 8). These conclusions enlighten the market entities such as government and foreign trade enterprises, exchange rate reform and supervision, and the introduction of major economic matters, which are divided into three stages (i.e., period 0, period 2, period 3, period 8 and period 9), which affect the fluctuation of exchange rate. The focus of the three phases should not be the same. The forecast of exchange rate change in 2010 (alternating fluctuation, overall appreciation trend) is in line with the actual situation in January and March, and the accuracy is very high.
【作者单位】：湖南科技大学;
【分类号】：F224;F832.52

【参考文献】