尹力博,李勍.投资者关注对人民币汇率价差波动的影响研究——基于GARCH-MIDAS模型[J].管理科学,2017,30(5):147-159
投资者关注对人民币汇率价差波动的影响研究——基于GARCH-MIDAS模型
A Study on the Impact of Investor Attention on RMB Exchange Rate Spread——Based on the GARCH-MIDAS Model
投稿时间:2017-03-19  修订日期:2017-07-21
DOI:
中文关键词:  投资者关注  人民币汇率价差  波动分解  央行干预  利率平价  流动性  预期
英文关键词:investor attention  RMB exchange rate spread  volatility decomposition  central bank intervention  interest rate parity  liquidity  forecast
基金项目:国家自然科学基金(71671193,71401193)
作者单位E-mail
尹力博 中央财经大学 金融学院北京 100081 yinlibowsxbb@126.com 
李勍 中央财经大学 金融学院北京 100081 liqing_cufe@126.com 
摘要点击次数: 397
全文下载次数: 
中文摘要:
      源自心理学的有限关注理论近年来被广泛应用于金融研究,其基本思路是:金融市场产品种类繁多,由于存在信息成本,个人投资者的关注便成为一种稀缺资源,于是更倾向于交易关注到的产品,进而影响价格。该理论在证券市场已得到大量的实证支持,因此对外汇市场的研究显得非常必要。
基于谷歌搜索量指数构建的人民币关注指数,以2011年6月27日至2016年7月29日为研究区间,利用最小二乘回归实证检验投资者关注对人民币在岸和离岸汇率价差波动率的影响。特别地,利用基于已实现波动率的GARCH-MIDAS模型将波动率分解为长期成分和短期成分,并分别进行回归,考虑了价差波动的结构性特征。
研究结果表明,投资者关注能够显著增大人民币汇率价差波动,货币当局可适当引导进而达到收窄价差、抑制套利的目的。与波动长期成分相比,投资者关注对波动短期成分的作用更强。加入央行干预、利率平价、流动性、预期等控制变量后,投资者关注对汇率价差波动率的预测能力依旧显著,表明其所包含的信息并未被已有变量覆盖。稳健性检验中,将4个控制变量同时加入基准模型并考虑关注与控制变量之间的耦合效应,结果表明投资者关注仍然具有独立的解释能力,且在长期内会通过影响控制变量的作用间接影响价差波动,作用机制与短期有所区别;直接利用基于月度投资者关注的GARCH-MIDAS模型对波动进行分解,通过系数的显著性进一步验证投资者关注对波动长期成分的作用;分别将在岸和离岸汇率波动率直接分解后进行最小二乘回归,投资者关注效果同样显著,且综合效果与原检验基本一致。
通过探讨投资者关注对人民币汇率价差波动的解释能力,拓展了基于谷歌搜索量指数测量的投资者关注在外汇市场的研究,为人民币汇率走势的预测提供了新思路,具有积极的现实意义,货币当局和外汇投资者可以构建人民币关注指数,以此作为参考来实现有效汇率管理或风险控制等多重目的。
英文摘要:
      The theory of limited attention arising from psychology has been widely used in financial research in recent years. The basic idea is that there are a wide variety of products in financial market and at the same time there are information costs. As a result, the attention of individual investors is becoming a scarce resource. Therefore, investors are more inclined to deal with the products they have already noticed and thus affect the price of the assets. A large number of empirical studies in the security market support the effectiveness of the theory, so the study in foreign exchange market is also necessary.
By employing the search volume index provided on Google Trend to construct a direct measure of investor attention towards RMB, with the study interval covering from June 27, 2011 to July 29, 2016, this paper applies the least squares regression to determine empirically whether the investor attention accounts for the volatility of spread between CNY and CNH exchange rates on a daily frequency. In particular, considering the structural properties, we use the GARCH-MIDAS model based on realized volatility to decompose volatility into the long-term and short-term components and then regress respectively.
The study results indicate the investors′ attention will increase the volatility of the spread, so monetary authority′s appropriate guidance may help narrow spread and limit arbitrage activities. Compared to the long horizon, the impact is stronger in short run. Furthermore, by adding central bank intervention index, interest rate parity, liquidity and forecast as control variables into the model respectively, the effect of attention remains significant, which demonstrates the information contained in investor attention are not covered by existing variables. A series of robustness tests were also conducted. First of all, we add four variables into the bench model while also consider the inter effects between attention and control variables. The results display attention still can explain the spread volatility independently and in the long run, it indirectly influences the spread fluctuation by affecting the effect of control variables, which demonstrates the mechanism is a little different from that of the short term. Secondly, the GARCH-MIDAS model based on the monthly investor′s attention is used to decompose the volatility, and the effect of the investor attention on long-term component is further verified by the significance of the coefficient. Finally, the volatility of the onshore and offshore exchange rate is directly decomposed and then apply the ordinary least squares regression. The results show investor attention is still significant and the overall effect is consistent with the original test.
By exploring the investor attention′s ability to explain the exchange rate spread fluctuation of RMB, this paper extends the research of investor attention based on Google search volume index in foreign exchange market and develops a new idea for the forecast of the RMB exchange rate. According to the conclusions, monetary authority and the participants in foreign exchange rate can also construct an investor attention index and offer a reference for multiple applications of effective exchange management or risk control and so on.
查看全文  查看/发表评论  下载PDF阅读器
关闭