本文關(guān)鍵詞： 協(xié)整 GARCH模型 主成分分析 突變理論　出處：《華東師范大學(xué)》2012年碩士論文　論文類(lèi)型：學(xué)位論文

【摘要】：統(tǒng)計(jì)套利是一種交易策略,它的目的是建這樣一個(gè)投資組合：尋找具有協(xié)整關(guān)系的一對(duì)股票,當(dāng)這對(duì)股票的股價(jià)相互偏離值達(dá)到一定程度時(shí),做空較高價(jià)位的股票,同時(shí)做多較低價(jià)位的股票。當(dāng)這對(duì)股票的股價(jià)相互偏離值縮小到一定程度時(shí),反方向平倉(cāng)掉投資組合內(nèi)股票,從而獲得投資收益。統(tǒng)計(jì)套利的優(yōu)點(diǎn)是市場(chǎng)中性,即不管市場(chǎng)如何波動(dòng),該策略所構(gòu)建的投資組合都可以對(duì)沖掉市場(chǎng)的系統(tǒng)性風(fēng)險(xiǎn)。 目前國(guó)內(nèi)外的學(xué)者對(duì)統(tǒng)計(jì)套利做了不少研究,主要集中于配對(duì)股票的選取方式和交易信號(hào)的選擇。本文將在已有研究的基礎(chǔ)上,運(yùn)用最新的數(shù)據(jù),對(duì)統(tǒng)計(jì)套利的各種常用模型在我國(guó)的證券市場(chǎng)上進(jìn)行實(shí)證檢驗(yàn),包括正態(tài)分布模型,GARCH模型等。研究結(jié)果顯示：運(yùn)用GARCH模型會(huì)產(chǎn)生過(guò)多的交易信號(hào),導(dǎo)致套利的收益無(wú)法抵償交易的費(fèi)用,而正態(tài)分布模型更加適用于我國(guó)的證券市場(chǎng)。以民生銀行和招商銀行為例,在2007年到2011年的時(shí)間段內(nèi),投資組合可以獲得平均17%的年化收益率。 早期的研究大多集中于一對(duì)一的股票配對(duì)交易,本文將對(duì)此進(jìn)行擴(kuò)展,研究一對(duì)多的配對(duì)方式。研究結(jié)果顯示一對(duì)一的配對(duì)和一對(duì)多的配對(duì)各有優(yōu)點(diǎn),兩者都可以提供很多統(tǒng)計(jì)套利的機(jī)會(huì),一對(duì)一的配對(duì)對(duì)兩只股票的相關(guān)性要求更高,一對(duì)多的配對(duì)更加適用于流通市值較小、交投活躍、波動(dòng)幅度大的股票。本文最后將突變理論引進(jìn)到統(tǒng)計(jì)套利中,突變理論是指在某個(gè)節(jié)點(diǎn)的前后時(shí)間序列的統(tǒng)計(jì)特征發(fā)生了變化。本文運(yùn)用似然估計(jì)法找出上證指數(shù)和配對(duì)個(gè)股的突變點(diǎn),結(jié)合突變點(diǎn)改進(jìn)交易信號(hào)。研究結(jié)果顯示這一方法可以獲得更好的收益,同時(shí)承受更低的風(fēng)險(xiǎn)。
[Abstract]:Statistical arbitrage is a trading strategy that aims to build a portfolio that looks for a pair of stocks with a cointegration relationship when the stock price deviates from each other to a certain extent. Short stocks at a higher price and long stocks at a lower price. When the price deviations from each other are reduced to a certain extent, sell the shares in the portfolio in the opposite direction. The advantage of statistical arbitrage is that it is market neutral, that is, no matter how volatile the market, the portfolio constructed by the strategy can hedge the systemic risk of the market. At present, scholars at home and abroad have done a lot of research on statistical arbitrage, mainly focused on the selection of matching stocks and trading signals. This paper will use the latest data on the basis of existing research. The common models of statistical arbitrage are tested in the stock market of China, including the normal distribution model. GARCH model and so on. The results show that: the use of GARCH model will produce too many transaction signals, resulting in arbitrage income can not cover the transaction costs. The normal distribution model is more suitable for China's securities market. Taking Minsheng Bank and China Merchants Bank for example, the period from 2007 to 2011. The portfolio yields an average of 17% annualized returns. Earlier studies focused on one-to-one stock pairing, which is expanded to study one-to-many pairing. The results show that one-to-one pairing and one-to-many pairing have their respective advantages. Both can provide a lot of statistical arbitrage opportunities, one-to-one pairs of two stocks of higher correlation requirements, one-to-many pairs more suitable for circulation market value small, active trading. In the end, the catastrophe theory is introduced into statistical arbitrage. Catastrophe theory refers to the change of statistical characteristics of time series before and after a certain node. In this paper, we use the likelihood estimation method to find out the mutation point of Shanghai stock index and matched stock. Improved trading signals with mutation points. Research results show that this approach can achieve better returns while at the same time taking lower risks.
【學(xué)位授予單位】：華東師范大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2012
【分類(lèi)號(hào)】：F224;F832.51

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 宿成建,陳潔;應(yīng)用變點(diǎn)模型來(lái)研究滬深股股市波動(dòng)性突變行為[J];重慶大學(xué)學(xué)報(bào)(自然科學(xué)版);2003年10期

，

本文編號(hào)：1476760