The Empirical Study about Introduction of Stock Index Futures on the Volatility of Spot Market

Transcription

1 ibusiness, 013, 5, hp://dx.doi.org/10.436/ib b04 Published Online Sepember 013 (hp:// 113 The Empirical Sudy abou Inroducion of Sock Index Fuures on he Volailiy of Spo Marke Guiliang Tian, Huixiangzi Zheng School of Business, Hohai Universiy, China. Received July,013 ABSTRACT April 16, 010, China s firs four Index Fuure conracs have been lised for rading in he sock exchange. Sock index fuures are he world's fases growing financial derivaive producs currenly, and he research of hem is of significance o he developmen of China's financial marke. Therefore, i is paricularly imporan o concern he marke operaion saus of sock index fuure afer is official lised. As we all know ha here are linkage effecs of he fuures and spo, sock index fuures marke will have some impac on he spo marke. Based on he above poin, his sudy idenified he Shanghai and Shenzhen 300 sock index fuures marke as he research objec, focusing on wha volailiy affecs he sock index fuures have made on he spo marke over he pas 3 years. Through empirical findings, we can evaluae he operaion condiions of sock index fuures marke objecively. This paper applies GARCH model mainly, and inroduce dummy variables, collecing daily rading daa beween 16 April 007 and 16 April 013 on he spo marke, and sudies he impac of volailiy on he spo marke in-deph by empirical es. The resuls showed ha over he pas 3 years, inroducion of sock index fuures has reduced he volailiy of he spo marke which has brough a posiive impac. Keywords: Index Fuure; Spo Marke; Volailiy; GARCH Model 1. Inroducion Sock index fuures, which are based on prior agreemen beween buyers and sellers, rade a a paricular ime in he fuure, in accordance wih he prior agreemen of boh paries o engage in sock index rading shares of a sandardized proocol agreemen. April 16, 010, our counry's firs financial derivaives - officially lised in Shanghai and Shenzhen 300 sock index fuures rading. However, soon afer is lising launched a vicious slump appears unilaeral marke condiions, driving he sock marke also followed he collapse, and hus some scholars have suggesed ha he sock index fuures marke crash blame. Therefore, i is necessary o explore he inroducion of sock index fuures on he volailiy of he spo marke produce wha kind of impac, in order o clarify wheher he sock index fuures is he "culpri." According o he impacion of sock index fuures on he spo marke, scholars hold differen views, and each view has a lo of heoreical research and empirical research suppor. Chan[1] on Nikkei Sock Average daily rading daa, wih is day yield research objec, hrough he esablishmen of GARCH model, he resuls found ha he inroducion of sock index fuures on he spo marke is no significanly affeced. Anoninus[] based on FTSE 100 index esed a series of economeric model, including uni roo es, vecor error correcion model, he esablishmen of such GARCH model and found ha he inroducion of sock index fuures, increased he volailiy of he spo prices sock, ha is on he spo markes had a negaive impac. Domesic scholars Tian-cai Xing and Ge Zhang[3] based on FTSE Xinhua A50 index fuures, hrough he esablishmen of GARCH model, analyzes is launch on he CSI 300 sock index, he resuls found ha he inroducion of sock index fuures slighly increases he volailiy of he spo marke, bu his impac is very small. In 010, hey seleced he CSI 300 index fuures rading simulaion daa, hrough he esablishmen of GARCH model o es residual ARCH effecs, ec., confirmed he launch of sock index fuures on he volailiy of he spo marke is no greaer impac. Bu because of he daa from he simulaion rading sysem, herefore he reliabiliy of he conclusions sill needs o be fasidious. From he exising research, he curren domesic and foreign scholars are mainly based on maure markes abroad o sudy he inroducion of sock index fuures on he spo marke volailiy impac on emerging markes such like our lack of research, especially in China he Copyrigh 013 SciRes.

2 114 The Empirical Sudy abou Inroducion of Sock Index Fuures on he Volailiy of Spo Marke CSI 300 index fuures launch dae jus over hree years, he research scholars are mosly simulaion-based ransacion daa, and conclusions vary. Therefore, his paper by choosing real ransacion daa were esablished before and afer he inroducion of index fuures GARCH model, and he inroducion of dummy variables, a series of empirical ess o explore CSI 300 sock index fuures on he impac of volailiy in he spo marke, wih cerain pracical significance.. Research Mehods.1. Sample Descripion This daa is seleced April 16, 007 o April 16, 013 in CSI 300 sock index fuures conracs and he CSI 300 sock index daily closing price, ha is lised on he CSI 300 sock index fuures before and afer he hree years of he corresponding sock he daily closing price of he sock index of 1460 daa. Among hem, he fuures daa for he selecion of he main conrac daa, namely he neares monh fuures conracs of he daily closing price daa. In he mos recen monh before he conrac selemen dae wo or hree days, monh conrac will be convered ino he main conrac, and hen urned o he nex neares monh fuures conrac closing daa, and so on, can be a coninuous se of sock index fuures price ime series. Daa mainly comes from China Financial Fuures Exchange, and he News Finance websie. In accordance wih he sock index fuures launch ime he daa is divided ino wo sages, he firs sage before he inroducion of sock index fuures, is from April 16, 007 o April 16, 010; second sage afer he inroducion of sock index fuures, from April 16, 010 o April 16, Daa Processing Firsly, in order o eliminae heeroscedasiciy, respecively fuures price and spo price series sequence logarihm. Assuming ha he rading day closing price of a fuures conrac and he spo index closing price are denoed as P and Q, hen he naural logarihm of fuures conrac price series expressed as F ln P, is logarihmic rae of reurn is F ln P ln P 1, he naural logarihm of sock price series expressed as S ln Q, is logarihm yields expressed as S ln Q ln Q 1. The sudy objec of his paper is he log reurn of sock index fuures and spo price rae..3. Model Selecion Exisence of financial ime series daa aggregaion and volailiy heeroscedasiciy, OLS regression model is he assumpion ha homoscedasiciy, herefore, in he sudy when he price volailiy of financial producs, widely used heeroscedasiciy model is he mos widely used in he 198 Engle proposed auoregressive condiional heeroskedasiciy model (ARCH model), Bollerslev deepened he ARCH model, he generalized auoregressive condiional heeroskedasiciy model (GARCH model), GARCH model can well reflec he aggregaion of price volailiy, in he previous lieraure, he widespread use of GARCH (1, 1) model reflecs he aggregaion of financial producs. This paper uses he GARCH (1, 1) model sock index fuures on he volailiy of he spo marke, and he inroducion of dummy variables in he condiional variance o reflec he CSI 300 sock index fuures on he impac of he sock marke volailiy, Before he inroducion of sock index fuures, D = 0, afer he inroducion of sock index fuures D = 1. The GARCH (1, 1) model wih he following: The mean equaion: y x The condiional variance equaion: 1 1 The GARCH (1, 1) model was respecively esablished for he differen sages, he model does no add dummy variables. 3. Empirical Tess We need o know he basic saisical characerisics of he reurn series of CSI 300 index before he empirical es. From he Figure 1, we can clearly see ha his series has he cluser effec which mos economic sequence possesses. I means ha he curren price effeced by he prices of he previous period, which we call he ARCH effec The Volailiy Analysis of he Spo Price Three Years before he Launch of CSI 300 Index Fuures 1) Do he ADF es of reurn series of he sock index Figure 1. The reurn s volailiy of sock index in whole sample period. Copyrigh 013 SciRes.

3 The Empirical Sudy abou Inroducion of Sock Index Fuures on he Volailiy of Spo Marke 115 Firs we do he ADF es of reurn series of he sock index, he es resul is as follows: From he Table 1, we can see ha he saionary es resuls are significan, hus rejecing he null hypohesis ha he reurn series is saionary. Then we esablish ARMA model. ) Esablish ARMA model a) Deermine he value of p and q Before he esablishmen of ARMA model, we should see auocorrelaion and parial auocorrelaion char of he daily reurn series of Shanghai and Shenzhen 300 index hree years before he inroducion of fuures. The figure appears in several places which auocorrelaion and parial auocorrelaion funcions are significan, herefore we canno deermine he value of p and q direcly. Then we draw suppor wih he AIC and SC principle by imiaing ARMA model wihin a 5-order. By comparing he value of AIC and SC, we can build ARMA(,3) model. b) Build ARMA(,3) model The following able is he oupu of ARMA (,3). As we can see from Table, he resuls of model esimaion are good. A he 1% confidence level, in addiion o he consan erm, he P value of all he coefficiens are less han 1%, passing he es, R ^ = 50.0%. Value of DW is around which has no auocorrelaion. A his poin, we remove he consan erm. The expression of ARMA (,3) model is as followings: Significance level ADF es Table 1. The ADF es resuls. -Saisic Prob.* % level % level % level Table. Oupu of ARMA (,3). Variable Coefficien Sd. Error -Saisic Prob. C -5.7E E AR(1) AR() MA(1) MA() MA(3) R R R u u u 1 3 c) Do he ARCH Tess of ARMA (,3) model Before he esablishmen of GARCH model, we need o es he ARCH effec of residuals series. We use he mehod of ARCH-LM es. The resul is as follows: The resul of ARCH-LM es is significan, indicaing ha we can esablish GARCH (1,1) model. 3) Build GARCH model a) Build ARMA(,3)-GARCH(1,1) model The specific form of ARMA (,3)-GARCH (1,1) model is as follows: Mean equaion: R R R 0.034u u 0.984u Variance equaion: (1) () (3) As we can see from Figure : he curren yield of he spo index will be effeced by yield of is lagged one period a a srengh of , by yield of is lagged wo period a a srengh of The curren volailiy will be affeced by he volailiy of previous period a a srengh of b) Do he ARCH-LM es of residuals series. Table 3. ARCH-LM es of residuals series. F-saisic Probabiliy Obs*R-squared Probabiliy Table 4. Oupu of GARCH model. Coefficien Sd. Error z-saisic Prob. Mean equaion AR(1) AR() MA(1) MA() MA(3) Variance equaion C 5.38E-06.74E RESID(-1)^ GARCH(-1) Copyrigh 013 SciRes.

4 116 The Empirical Sudy abou Inroducion of Sock Index Fuures on he Volailiy of Spo Marke Table 5. Oupu of ARCH-LM es. F-saisic Probabiliy Obs*R-squared Probabiliy The value of P is , hus he ARCH effec of residuals no longer exis, which indicaes ha ARMA (,3)-GARCH (1,1) model is fully fied a he reurn series of CSI 300 sock index. 3.. The Volailiy Analysis of he Spo Price Three Years afer he Launch of CSI 300 Index Fuures This secion of he concree seps is as above, so we will no repea. Esablish GARCH model of reurns on Spo Index hree years afer he inroducion of index fuures like follows: Mean equaion: R R 1 (4) Variance equaion: As we can see: he curren yield of he spo index will be effeced by yield of is lagged one period a a srengh of The curren volailiy will be effeced by he volailiy of previous period a a srengh of Differences Analysis Firsly, we draw he volailiy char of spo price in he whole sample period. From he figure, we can easily see ha he volailiy of sock price has weakened hree years afer he inroducion of sock index fuures. So we can preliminary judge ha he inroducion of sock index fuures reduces he volailiy of he spo marke. Nex, we will build a GARCH model based on he whole sample inerval, and inroduce dummy variables. Do posiive and negaive es of dummy variable coefficiens, and see if i passed he examinaion o verify wheher he inroducion of sock index fuures reduces he volailiy of he spo marke. Esablishing revise GARCH model likes follows: Mean equaion: R R R u u Variance equaion: u 1 (5) (6) (7) Figure. Volailiy of whole sample period. As we can see: he curren yield of he spo index will be effeced by yield of is lagged one period a a srengh of , by yield of is lagged wo period a a srengh of The curren volailiy will be affeced by he volailiy of previous period a a srengh of The mos imporan, he coefficien of dummy variable is negaive, and passed he es, indicaing ha he inroducion of sock index fuures reduces he volailiy of he spo marke, which made a posiive impac on he spo marke. 4. Conclusions and Recommendaions This aricle sudies he impac of he inroducion of sock index fuures on he volailiy in he spo marke by esablishing GARCH (1,1) model, and he inroducion of dummy variables, and found he sock index fuures has a sligh decrease on he volailiy of spo marke volailiy. In order o promoe he developmen of he sock index fuures marke, reduce he negaive impac o a minimum of he fuure fuures on he spo marke may, we propose he following commens and suggesions, mainly discussed from he wo aspecs include coninue o improve he spo marke and srenghen he risk supervision and managemen of he fuures marke Improve he Spo Marke Coninually The developmen of sock index fuures markes akes he developed spo marke as a precondiion. Alhough China's sock marke has been so close o he inernaional advanced level in he echnical faciliies, forming he aspec of marke size, markeing supervision sysem, informaion disclosure is also far behind, herefore, we should srenghen he sandardizaion of he spo marke, and promoe he sharing and coordinaion managemen of he sock marke and sock index fuures marke. 4.. Srenghen Risk Oversigh and Managemen of he Fuures Marke Volailiy is an imporan indicaor o measure marke Copyrigh 013 SciRes.

5 The Empirical Sudy abou Inroducion of Sock Index Fuures on he Volailiy of Spo Marke 117 risk, he size of volailiy deermines he size of he marke, alhough he empirical es he inroducion of sock index fuures in our counry did no increase he volailiy of he spo marke, because of he sock index fuures iself has a larger marke risk, in order o coninue o promoe he developmen of he fuures marke, we sill need o srenghen risk managemen of he sock index fuures. Combined wih he acual siuaion of China's marke, we should mainly make he following recommendaions: A. Perfec marke invesors, srenghen insiuional invesors, leading he reasonable and healh invesmen philosophy in he marke. For now, mainly ake he insurance companies, fund companies, securiies companies as he main body of hedging, and encourage he arbirage behavior of he marke, reduce sysemic risk along wih promoe he price discovery funcion of fuures markes yield well and effecively. B. Improve he managemen of various accouns. Since he fuures implemen he policy of margin, i is necessary o srenghen he supervision of he rading margin and clearing margin, implemen he daily free deb selemen sysem. A he same ime, we also need o coninue o srenghen he managemen of he large posiions sysem and forced open sysem as well as ake effecively care of risk reserves, ec., and adop inernaional advanced real-ime monioring echnology o effecively circumven he sock index fuures marke risks. C. Timely and argeed inroduce he insiuional arrangemens, improve relevan laws and regulaions, make he regulaory abiliies and adhere adap he speed of developmen in order o laid a good foundaion for he enire marke s healhy and sable developmen. 5. Acknowledgemens This sudy was funded by he Naional Naural Science Foundaion of China (Gran No , ) REFERENCES [1] Chan and Karolyi, Inroducion Volailiy in he Sock Index and Sock Index Fuures Review of Financial Sudies, Vol. 4, No. 3, 1991, pp [] Anoninus, Reurn and Volailiy Dynamics in he FT-SE 100 Sock Index and Sock Index Fuures Markes, Journal of Fuures Markes, Vol. 15, No. 4, 1995, pp doi:10.100/fu [3] Xing Tiancai and Zhang Ge, The Empirical Sudy abou Impac of sock index fuures on he spo marke - Based on he analysis of he FTSE Xinhua A50, Research on Financial and Economic Issues, Vol. 4, No. 7, 009, pp Copyrigh 013 SciRes.