基于異質(zhì)市場(chǎng)假說(shuō)的中國(guó)股市已實(shí)現(xiàn)波動(dòng)率研究
發(fā)布時(shí)間:2018-11-02 09:37
【摘要】:隨著全球經(jīng)濟(jì)一體化的發(fā)展與中國(guó)股票市場(chǎng)運(yùn)作機(jī)制的日趨完善,我國(guó)股票市場(chǎng)已成為全球金融市場(chǎng)中一個(gè)重要組成部分,在金融市場(chǎng)研究中,波動(dòng)率是一直為熱點(diǎn)課題,學(xué)者們對(duì)波動(dòng)率領(lǐng)域的大量研究得出了求解波動(dòng)率的三種主流方法:一是由Black-Scholes方程求解到的隱含波動(dòng)率;二是以(?)RCH類(lèi)和SV類(lèi)模型為代表的歷史波動(dòng)率;三是基于高頻數(shù)據(jù)研究的已實(shí)現(xiàn)波動(dòng)率,也就是本文的研究對(duì)象。隨著高頻數(shù)據(jù)可以越來(lái)越方便的獲得,以前的模型波動(dòng)率方法已不再適合高頻數(shù)據(jù)研究需要。相比于模型波動(dòng)率,已實(shí)現(xiàn)波動(dòng)率能夠更直接、準(zhǔn)確地描述股市波動(dòng)率的特征。金融研究文獻(xiàn)表明,以往的GARCH類(lèi)和SV類(lèi)模型的日收益低頻數(shù)據(jù)會(huì)損失很多對(duì)投資者波動(dòng)預(yù)期很有用的信息,而日內(nèi)高頻數(shù)據(jù)的可得性則彌補(bǔ)了這一缺陷,高頻率的采樣數(shù)據(jù)包含了盡可能全面的信息。另外,已實(shí)現(xiàn)波動(dòng)率方法是一種非參數(shù)方法,沒(méi)有模型波動(dòng)率方法帶來(lái)的參數(shù)估計(jì)問(wèn)題,更沒(méi)有因?yàn)閺?fù)雜參數(shù)估計(jì)帶來(lái)的“維數(shù)災(zāi)難”,是近年來(lái)比較新穎的波動(dòng)率研究方法。 對(duì)已實(shí)現(xiàn)波動(dòng)率的研究理論中,Muller(1993)等人的異質(zhì)市場(chǎng)假說(shuō)的理論,Peters(1994)的分形市場(chǎng)假說(shuō)理論,Lux and Marchesi(1999)的混合市場(chǎng)理論,市場(chǎng)交易的異質(zhì)性特征在這三種理論中都有所體現(xiàn),與同質(zhì)市場(chǎng)相反,異質(zhì)市場(chǎng)假說(shuō)認(rèn)為,市場(chǎng)波動(dòng)率與市場(chǎng)活躍程度成正比,即參與者越多,交易行為的多樣化會(huì)導(dǎo)致波動(dòng)的正向變化。本文選擇基于異質(zhì)市場(chǎng)假說(shuō),在考慮了市場(chǎng)微觀結(jié)構(gòu)噪聲和跳躍基礎(chǔ)上,采用HAR-RV以及相關(guān)模型(HAR-RV-GARCH, HAR-RV-J,HAR-RV-CJ)對(duì)我國(guó)股票市場(chǎng)中的已實(shí)現(xiàn)波動(dòng)率進(jìn)行了有效的估計(jì),分析研究了短期,中期,長(zhǎng)期交易類(lèi)型對(duì)股市已實(shí)現(xiàn)波動(dòng)率的影響。 本文選取從2005年4月8日至2010年4月22日上證綜指共計(jì)59816個(gè)5分鐘高頻收盤(pán)價(jià)數(shù)據(jù),計(jì)算得到1227個(gè)已實(shí)現(xiàn)波動(dòng)率數(shù)據(jù),2005年1月4日至2011年8月23日深證綜指共計(jì)77424個(gè)5分鐘高頻收盤(pán)價(jià)數(shù)據(jù),計(jì)算獲得1613個(gè)已實(shí)現(xiàn)波動(dòng)率作為研究數(shù)據(jù),基于二次冪變差理論,將波動(dòng)率中的跳躍成分進(jìn)行有效的分離,通過(guò)HAR-RV-J(?)HAR-RV-CJ模型對(duì)短中長(zhǎng)期已實(shí)現(xiàn)波動(dòng)率進(jìn)行了有效的估計(jì),本論文實(shí)證表明,對(duì)于日已實(shí)現(xiàn)波動(dòng)率,短期投資類(lèi)型對(duì)其影響最大,而中長(zhǎng)期交易類(lèi)型的邊際貢獻(xiàn)率偏低,但兩者相差不大,而三個(gè)解釋變量系數(shù)均顯著,表明我國(guó)股票市場(chǎng)異質(zhì)性特征的存在,同時(shí)也驗(yàn)證了異質(zhì)市場(chǎng)假說(shuō)中的理論,即短期已實(shí)現(xiàn)波動(dòng)率受自身前期值的影響,同時(shí)也受中長(zhǎng)期交易的影響,而對(duì)于長(zhǎng)期交易者而言,主要受其自身前期值的影響。這與于小蕾在《基于HAR模型對(duì)中國(guó)股票市場(chǎng)已實(shí)現(xiàn)波動(dòng)率研究》中的結(jié)論類(lèi)似。 而后對(duì)HAR-RV模型進(jìn)行最小二乘回歸后的殘差進(jìn)行分析,發(fā)現(xiàn)殘差序列具有群聚現(xiàn)象,因此我們檢驗(yàn)了誤差項(xiàng)的ARCH效應(yīng),并添加了GARCH項(xiàng),通過(guò)HAR-RV-GARCH模型進(jìn)一步研究了三種交易類(lèi)型對(duì)已實(shí)現(xiàn)波動(dòng)率的影響,消除了殘差項(xiàng)的ARCH效應(yīng),HAR-RV-GARCH模型實(shí)證表明,添加GARCH項(xiàng)后的模型與原來(lái)的HAR-RV模型幾乎沒(méi)有什么差異。在研究跳躍影響方面,根據(jù)二次冪變差理論,將已實(shí)現(xiàn)波動(dòng)率分解為連續(xù)樣本路徑方差和離散跳躍方差兩部分,考慮到市場(chǎng)微觀結(jié)構(gòu)噪聲對(duì)已實(shí)現(xiàn)波動(dòng)率的影響,用錯(cuò)列的收益率絕對(duì)值乘積(原始的二次冪變差公式是相鄰收益率絕對(duì)值的乘積)修正了二次冪變差的計(jì)算方式,以減小市場(chǎng)微觀結(jié)構(gòu)的影響,這樣通過(guò)修正的Z統(tǒng)計(jì)量重新得到考慮市場(chǎng)微觀結(jié)構(gòu)噪聲影響后的跳躍方差的序列,在顯著水平為0.01條件下,由修正的Z統(tǒng)計(jì)量檢驗(yàn)得到的顯著性到達(dá)的跳躍共308次,比之前的276次多了32次,捕捉率也由修正前的22.49%增強(qiáng)為25.1%,因此,修正的Z統(tǒng)計(jì)量檢驗(yàn)的跳躍變差序列更為精確一些,我們根據(jù)Huang和Tanchen(2005)的研究?jī)?nèi)容,先是對(duì)HAR-RV-J模型實(shí)證,結(jié)果表明,跳躍項(xiàng)對(duì)已實(shí)現(xiàn)波動(dòng)率具有顯著的負(fù)影響,通過(guò)HAR-RV-J和HAR-RV-CJ模型對(duì)波動(dòng)率的估計(jì)結(jié)果,對(duì)比HAR-RV模型,我們可以知道,考慮跳躍項(xiàng)的HAR-RV-J和HAR-RV-CJ模型有更好的擬合效果和預(yù)測(cè)性能。 本文創(chuàng)新點(diǎn)即充分考慮市場(chǎng)微觀結(jié)構(gòu)噪聲的條件下量化了跳躍成分對(duì)不同已實(shí)現(xiàn)波動(dòng)率的影響,采用修正的Z統(tǒng)計(jì)量有效的檢測(cè)出顯著性跳躍次數(shù),對(duì)波動(dòng)率的捕捉程度有明顯的提高,從而更有效的分析影響已實(shí)現(xiàn)波動(dòng)率的各個(gè)因子。
[Abstract]:With the development of global economic integration and the perfection of the market operation mechanism of China's stock market, our stock market has become an important part of the global financial market. Scholars have obtained three main methods for solving the fluctuation rate: one is the implicit fluctuation rate obtained by Black-Scholes equation; 2 is (?) The RCH class and SV class model represent the historical fluctuation rate; three are the realized fluctuation rate based on high frequency data research, which is the research object of this paper. As high frequency data can be obtained more and more conveniently, previous model wave rate methods are no longer suitable for high frequency data research. Compared with the model fluctuation rate, the fluctuation rate can be more directly and accurately described. The literature of financial research shows that the daily gains and low frequency data of GARCH and SV models have lost many useful information about investor's fluctuation, while the availability of high frequency data in Japan makes up for this defect, and high frequency sampling data contains as much information as possible. In addition, the method of fluctuation rate is a kind of non-parametric method, there is no parameter estimation problem brought by model fluctuation rate method, and there is no 鈥渄imension disaster鈥,
本文編號(hào):2305654
[Abstract]:With the development of global economic integration and the perfection of the market operation mechanism of China's stock market, our stock market has become an important part of the global financial market. Scholars have obtained three main methods for solving the fluctuation rate: one is the implicit fluctuation rate obtained by Black-Scholes equation; 2 is (?) The RCH class and SV class model represent the historical fluctuation rate; three are the realized fluctuation rate based on high frequency data research, which is the research object of this paper. As high frequency data can be obtained more and more conveniently, previous model wave rate methods are no longer suitable for high frequency data research. Compared with the model fluctuation rate, the fluctuation rate can be more directly and accurately described. The literature of financial research shows that the daily gains and low frequency data of GARCH and SV models have lost many useful information about investor's fluctuation, while the availability of high frequency data in Japan makes up for this defect, and high frequency sampling data contains as much information as possible. In addition, the method of fluctuation rate is a kind of non-parametric method, there is no parameter estimation problem brought by model fluctuation rate method, and there is no 鈥渄imension disaster鈥,
本文編號(hào):2305654
本文鏈接:http://sikaile.net/guanlilunwen/zhqtouz/2305654.html
最近更新
教材專(zhuān)著