金融實(shí)證分析方法

寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：金融實(shí)證分析方法—寧波市生產(chǎn)總值因素分析AalysisofNingbo'sGDPfactor課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：Abstract:Intheareaofeconomicdevelopment,differentindustries'scontributiontoeconomicgrowthisdifferent,findoutthelargercontributiontoeconomicgrowthoftheindustry,istofurtheraccelerateregionaleconomicdevelopmentandenhancethecorecompetitivenessofregionaleconomykey.ThispaperselectedfortheempiricalstudyofNingboCity,Ningboregionbyidentifyingfactorsthataffecteconomicgrowth,econometricmodels,andtimeseriesdatastationarityandcointegrationtest,themodelmulticollinearityanalysis,autocorrelationinspectionandtestingheteroscedasticityinqualitativeanalysisandquantitativeanalysis,basedonthefinalestablishmentofNingboGDPeconometricmodeltoarriveattheindustrialaddedvalue,totalfixedassetinvestment,totalretailsalesofsocialconsumergoods,financegeneralbudgetrevenueofGDPinNingbosignificantfactors,andthemodelissimpleeconomics.Thispaperselectedfixedassetinvestmentasanimportantfactorofeconomicgrowth,economicgrowthhasadirectstimulus.Inthispaper,1978-2010economicdataunitrootandcointegrationanalysisNingbofixedassetinvestmentandeconomicgrowthanempiricalstudy.Keywords:GDP;influencingfactors;timeseries;multicollinearity;fixedassetinvestment寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：寧波市生產(chǎn)總值因素分析摘要：在區(qū)域經(jīng)濟(jì)發(fā)展中，不同的產(chǎn)業(yè)對經(jīng)濟(jì)增長的貢獻(xiàn)是不同的，尋找出對經(jīng)濟(jì)增長貢獻(xiàn)較大的產(chǎn)業(yè)，是進(jìn)一步加快區(qū)域經(jīng)濟(jì)發(fā)展，提升區(qū)域經(jīng)濟(jì)核心競爭力的關(guān)鍵。本文選取寧波市為實(shí)證研究對象，通過找出對寧波地區(qū)經(jīng)濟(jì)增長產(chǎn)生影響的因素，建立計(jì)量經(jīng)濟(jì)模型，并對時(shí)間序列數(shù)據(jù)進(jìn)行平穩(wěn)性及協(xié)整檢驗(yàn)，對模型進(jìn)行多重共線性分析、自相關(guān)檢驗(yàn)和異方差性檢驗(yàn)，在定性分析與定量分析相結(jié)合的基礎(chǔ)上，最終確立寧波地區(qū)生產(chǎn)總值的計(jì)量經(jīng)濟(jì)模型，得出工業(yè)增加值、全社會固定資產(chǎn)投資，社會消費(fèi)品零售總額，財(cái)政一般預(yù)算收入是寧波地區(qū)生產(chǎn)總值的顯著影響因素，并對模型進(jìn)行簡單的經(jīng)濟(jì)學(xué)分析。本文主要選取了固定資產(chǎn)投資作為經(jīng)濟(jì)增長的一個(gè)重要因素,對經(jīng)濟(jì)增長具有直接的拉動(dòng)作用。本文利用1978-2010年的經(jīng)濟(jì)數(shù)據(jù)，進(jìn)行單位根檢驗(yàn)和協(xié)整分析對寧波市固定資產(chǎn)投資與經(jīng)濟(jì)增長關(guān)系進(jìn)行了實(shí)證研究。關(guān)鍵詞：地區(qū)生產(chǎn)總值；影響因素；時(shí)間序列；多重共線性；固定資產(chǎn)投資1引言研究區(qū)域的投資環(huán)境是研究一個(gè)地方經(jīng)濟(jì)發(fā)展的基礎(chǔ),也是揭示一個(gè)區(qū)域經(jīng)濟(jì)發(fā)展?jié)摿Φ挠行緩街粡膶幉ǖ慕?jīng)濟(jì)發(fā)展看,投資驅(qū)動(dòng)是寧波經(jīng)濟(jì)發(fā)展的主要?jiǎng)恿χ?保持投資規(guī)模的適度增長,努力提高投資質(zhì)量,是現(xiàn)階段寧波經(jīng)濟(jì)景氣周期延長的必要保障。本文以固定資產(chǎn)投資對經(jīng)濟(jì)的促進(jìn)作用為理論依據(jù),借助現(xiàn)代經(jīng)濟(jì)計(jì)量學(xué)的方法,對改革開發(fā)以來的寧波市宏觀經(jīng)濟(jì)統(tǒng)計(jì)資料進(jìn)行分析,客觀地評價(jià)固定資產(chǎn)投資對寧波市經(jīng)濟(jì)增長的影響。2.寧波市經(jīng)濟(jì)社會發(fā)展階段及影響因素分析2.1數(shù)據(jù)采集影響寧波市地區(qū)生產(chǎn)總值的因素比較多，在本文的分析中，我們采用社會固定資產(chǎn)投資作為影響因素。根據(jù)寧波市經(jīng)濟(jì)社會所處的發(fā)展階段和寧波地區(qū)的生產(chǎn)特點(diǎn)、各因素對生產(chǎn)總值的影響大小及資料的可比性，本文選擇全社會固定資產(chǎn)投資作為影響地區(qū)生產(chǎn)總值的主要因素（如表1所示）。年份生產(chǎn)總值（Y）社會固定資產(chǎn)投資（X）197820.175.02

197924.155.79198029.536.50198131.996.39198236.888.57198341.687.69198453.1710.95198571.0518.08198680.2222.01198795.9929.461988118.6235.811989137.2532.791990141.4039.281991169.8751.421992213.0576.251993315.11129.271994459.66184.601995602.65264.191996784.07309.971997879.10300.571998952.79309.8119991017.08318.9320001144.57360.7520011278.75470.2820021453.34601.2720031749.27835.9020042109.451103.8120052447.321336.3020062874.421502.7720073418.571597.5420083946.521728.2420094329.302004.2220105163.002193.28圖12.2.構(gòu)建模型做時(shí)間序列生產(chǎn)總值y,社會消費(fèi)品零售額xl的差分圖和單位根檢驗(yàn)，得到Y(jié)的差分寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：圖從生產(chǎn)總值差分序列圖可以看出寧波市生產(chǎn)總值在2000年有所回落，總體基本上是增長趨勢，由于生產(chǎn)總值數(shù)逐年增加，但是每年的總值增長率是逐漸下降的。從生產(chǎn)總值序列y的變化特征看，這是一個(gè)非平穩(wěn)序列。t2.3首先做lny與lnx的樣本相關(guān)圖首先得到lny的樣本相關(guān)圖

□LOG㈣Workfile:UNTITLED::Untitled'|=||回||極HView|卩「口匚|Object]Proper*亡呂|Print|Name|Freeze|Sample|Genr|目he亡11Graph|Stats|Id亡nt|C&rrelogranwfLOG(Y)Date:06/05/13Time:21:11Sample:19732010Includedobservations:33PADQ-StatProb0.9160.91630.256PADQ-StatProb0.9160.91630.2560.0000.031-0.04755.9590.0000.744-0.05477.2940.0000.656-0.06694.4120.0000.567-0.0&0107.690.0000.479-0.056117.520.0000395-0.037124.430.0000.316-0.024129.050.0000.233-0.056131.770.0000.161-0.059133.070.0000.006-0.0&6133.460.0000.011-0.067133.470.000圖4樣本的自相關(guān)函數(shù)圖可以看出，函數(shù)并沒有迅速趨向于零，并在零附近波動(dòng)，說明lny序列是非平穩(wěn)的用同樣的方法，做lnx的自相關(guān)函數(shù)圖圖5樣本的自相關(guān)函數(shù)圖可以看出，函數(shù)并沒有迅速趨向于零，并在零附近波動(dòng)，說明lnx序列是非平穩(wěn)的2.4單位跟檢驗(yàn)

課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：接下來做lny進(jìn)行單位跟檢驗(yàn)，從模型3進(jìn)行檢驗(yàn)，包括截距項(xiàng)，時(shí)間趨勢及一階滯后項(xiàng)的模型結(jié)果如下□Series^LOG[￥JWorkfile:UFSTTTrLE&::Untitled\|=||回||因|View|Pbu]Ob膽創(chuàng)Pr.o視?如]「號字司§迎取|企口「|目膽芭|_Gr日ph[貿(mào)日竺[世引日AugmentedDick&y-FulferUnitRootTestonLOG(Y}NullHypothesis:LOG[Y}hasaunitrootExogenous:Constant,LinearTrendLagLength:1(AutomaticbasedonSIC,MAXLAG=3}t-StatisticProb*AugmentedDickey-Fullerteststatistic-2.5475210.3050Testcriticalvalues:1%level5%level10%level-4.204500-3.562032-3.215267^MacKinnon(1996}one-sidedp-valuesAugmentedDickey-FullerTestEquationDependentvariable:D[LDG(Y))Method:LeastSquaresDate:06/05/13Time:21:27Sample(adjusted):19302010Includedobservations:31afteradjustments圖6從上面的伴隨概率值可以知道，在5%的顯著性水平下，不拒絕存在單位根的假設(shè)，表明lny是非平穩(wěn)的。對模型2進(jìn)行檢驗(yàn)，即不包括時(shí)間趨勢的模型，結(jié)果如下QSeries:LOGm'A'orkfiIe:UNTITLED::Untitled\耳回][View|Proc|Object]Properties|Print|Name[Freeze|Sampl巴|GenriSheet|Graph|Stats|Ident|AugmentedDickey-FullerUnitRcx>tTestonLOG(Y}NullHypothesis:LOG(Y)hasaunitrootExogenous:ConstantLagLength:1(AutomaticbasedonSIC,MAXLAG=3)?t-StatisticProb*Augmented1Dickey-Fullerteststatistic-0.6142570.8533Testcriticalvalues:1%level-3.6616615%level-2.96041110%level-2.619160^MacKinnon(1996)one-sid-ecfp-valuesAugmentedDickey-FullerTestEquationDependentvariable:D(LOG(Y})Method:LeastSquares□ate:06/05/13Time:21:36Sample(adjusted}:19802010Includedobservations:31afteradjustmentsT圖7從上面的伴隨概率值可以知道，在5%的顯著性水平下，不拒絕存在單位根的假設(shè)，表明lny是非平穩(wěn)的。對模型1進(jìn)行檢驗(yàn)，即不包括截距項(xiàng)和時(shí)間趨勢。結(jié)果如下口Series:LOGCY]Workfile:UNTrTLED::UntHed\口『回II探IViewProc|ObjectProperties|Print|NamejFreezejSample|Genr|Sheet|GraphjStats|IdentjAugmentedDrckev-FullerUnrtRootTestonLOG(Y}NullHypothesis:LQG[Y}hasaunitrootExogenous:NoneLagLength:1(Automa.ticbasedonSIC,MAXU\G=3)t-StatisticProb*AugmentedDickey-Fullerteststatistic1.8633420.982STestcriticalvalues:1%level5%level10%level-2.641672-1.952066-1.610400^MacKinnon(1996)one-sidedp-valuesAugmente總Dickey-FullerTestEquationDependentvariable:D(LQGCY)}Method:LeastSquaresDate:06/05713Time:21:37Sample^adjusted/19302010Includedobservations:31afteradjustmentsT圖8從上面的伴隨概率值可以知道，在5%的顯著性水平下，不拒絕存在單位根的假設(shè)，表明lny是非平穩(wěn)的。綜上所述，lny序列是非平穩(wěn)序列。用同樣的方法對lnx序列進(jìn)行檢驗(yàn)，模型3的結(jié)果：寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（寧波大學(xué)科學(xué)技術(shù)學(xué)院考核答題紙（20--20學(xué)年第學(xué)期）課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：AugmentedDickey-FullerTestEquationDependentvariable:D(LOG(X))Method:LeastSquares□ate:06/05/13Time:21:40Sample(adjusted):19302010Includedobservations:31afteradjustments口SedesrLOG兇Workfle;UNTTTLED:;Untitled\「=||回，|「Hi已％"Pi?□匚|Object]Prop已r?弓呂P叩tj買ame|F「e已ze|SampleGenrSheetErapE無迢Augment&dDiek&y-FuIlerUnitRcwtTestonLOG兇NullHypothesis:LOGCX)hasaunitrootExogenous:Constant,LinearTrendLagLength:1(AutomaticbasedonSIC,M.AXLAG=3)t-StatisticProb.*AugniEntEcfDiukEy-FullE「tEst￡tmtistic：了O.H斗52Testcriticalvalues:1%level-4.2845305%level-3.56200210%level-3.215267^MacKinnon0996)one-sidedp-values.圖9模型2的結(jié)果：圖10課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：模型1的結(jié)果:□Series:LOGPQWorkfile:UNTITLED::UrTtitled\口||回’||SS|.亞對Prut]Obj亡出Properties]P「int〔N日m亡|f「亡亡2亡]Smmpl亡]G亡口[]Sh亡亡Grmph]亜ts]Id亡nt]AugmentedDickey-FulterUnitRootTestonL0GQQNullHypothesis:LOG(X)hasaunitrootExogenous:NoneLagLength:1[AutomaticbasedonSIC,MAXLAG=3)t-StatisticProb/AugmentedDickey-Fullerteststatistic1.B590620.9S26Testcriticalvalues:1：%level5%level10%level-2.641672-1.952066-1.610400^MacKinnon(1996)one-sidedp-valuesAugmentedDickey-FullerTestEquation□ependentVariable:D(LOG(X))Method:LeastSquaresDate:06/05?13Time:21:41Sample[adjusted}:19302010Includedobseru'ations:31afteradjustments圖11從伴隨概率值可以看出，在5%的顯著性水平下，不拒絕存在單位根的檢驗(yàn)，加&是非平穩(wěn)的。綜上所述，lnx序列是非平穩(wěn)序列由于時(shí)間序列l(wèi)ny和lnx是非平穩(wěn)的，如果沒有進(jìn)行協(xié)整性檢驗(yàn)，直接對兩者做OLS回歸，此回歸很可能是虛假回歸3.檢驗(yàn)?zāi)Ｐ?.1單整性檢驗(yàn)單整性的檢驗(yàn)仍然通過單位根檢驗(yàn)進(jìn)行。但此時(shí)，針對的時(shí)間序列不是原序列的水平序列，而是一階差分、二階差分或更高階的差分序列為了尋找適當(dāng)?shù)哪Ｐ?，?jīng)過反復(fù)測算，發(fā)現(xiàn)lny的一階差分序列在只帶截距項(xiàng)與三階滯后項(xiàng)時(shí)，在5%的顯著性水平下可以拒絕存在單位根的假設(shè)。過程如下：

圖12結(jié)果如下r□Series:LOG^Workfile:UNTrTLEDr:Unti'tled\=回瓦一；vVie'A']P「oizObject|Prop亡「tie另Freeze]Sample|Gen「Sh亡et|GraphI.Etats|Ident]iiAugmentedDickey-FullerUnitRootTestonD(LOG(Y))NullHypothesis:D[LOG(Y))hasaunitrootExogenous:ConstantLagLength:0[AutomaticbasedonSIC,t-StatisticProb.^AugmentedDickey-Fullerteststatistic-2.9614760.0499Testcriticalvalues:1%level5%level10%level-3.661661-2.960411-2.619160^MacKinnon(1996)one-sidedp-valuesAugmentedDickey-FullerTestEquationDependentvariable:D(LQGCY),2}Method:LeastSquares□ate:06/05/13Time:21:53Sample(adjusted):19S02010Includedobservations:31afteradjustments圖13所以InGDP序列是一階單整的。即Inx?iG.

課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：用同樣的方法對加X]進(jìn)行單整性檢驗(yàn)結(jié)果如下圖14發(fā)現(xiàn)lnx的一階差分序列，只帶截距項(xiàng)與三階滯后項(xiàng)時(shí)，在5%的顯著性水平下可以拒絕存在單位根的檢驗(yàn)。所以lnx序列也是一階單整的。即Inx?I（L）.由于lny和lnx兩序列是非平穩(wěn)的，因此不宜直接建立它們的ARMA模型。但它們的一階差分序列卻是平穩(wěn)的，因此可對差分序列建立ARMA模型。記X二InX-InXy二lny-lnyttt-]ttt-]3.2做X的自相關(guān)函數(shù)與偏自相關(guān)函數(shù)圖，結(jié)果如下：

□Series:LOG?lVorkfile:UNTITLED::Untitl&d\View.Proc|Object|Properties!PrintNameFreeze|Sample]GenrSheet■GraphjStats;IdentCorr&logramofD(LOG(Y})□ate:06/05/13Time:22:02Sample:19702010Includedobservations:32AutocorrelationPartialCorrelationACFACQ-StatProbIII110.5360.53610.0710.002I□II匚120.165-0.17111.0610.004iEIiIT13-0.119-0.19011.5970.009匸II匚14-0.310-0.20515.5330.004IIi匚15-0377-0.13621.2670.001IIIl116-0.326-0.10325.7060.000I[I1□17-0.0740.13925.9460.001I□I1]100.1430.06626.3770.001I1]190.2700.056303350.000I]i匚1100.093-0.20330.8060.001I1I1J1110.0440.00930.9070.001I匚I1匚112-0.123-0.17931.7330.002圖15從上面可以看出，序列X在一階滯后后，自相關(guān)函數(shù)與偏自相關(guān)函數(shù)均迅速趨于零表明它是ARMA(1,1)的平穩(wěn)序列，因此原序列l(wèi)nx為ARIMA(1,1,1)序列。估計(jì)X序列，結(jié)果如下：DependentVariable:DLOG[Y}Method:LeastSquares□ate:06/05713Time:22:06Sample(adjusted):19S02010Ineludedobservations:31afteradjustmentsConvergenceachievedafter5iterationsMABackcast:1979VariableCoefficientStef.Errort-StatisticProb.c0.1739760.0254576.0342500.0000AR⑴0.3773100.3161031.1936300.2426MA[1}0.2303960.33&0920.6375590.4974R-squared0.300923[.leandependent^「0.173064AdjustedR-squared0.250939S.D.dependentvar0.0333023.E.ofregression0.072094Akaikeinfocriterion-2.329921Sumsquaredresid0.145532Schwarzcriterion-2.191143Loglikelihood39.11373Hmnnmn-CUJinncriter.-2.234605F-statistic6.026412Durbin-Watsonstat1.969331Prob(F-statistic)0.006650InvertedARRoots.33InvertedMARoots-23圖16

課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：然而由于MA（1）的參數(shù)不顯著，可以從模型中去掉。修正后如下圖:DependentVariable:DLQG(Y；Method:LeastSquares□ate:06/05/13Time:22:17Sample(adjusted):19302010Includedobservations:31afteradjustmentsConvergenceachievedafter3iterationsVariableCoefficientStd.Errort-StatisticProb.C0.1729160.027&746.2433340.0000AR(1)0.5356910.1567333.4167730.0019R-squared0.2S7020Meandependentvar0.173064AdjustedR-squared0.262434S.D.dependentvar0.033302S.E.ofregression0.071541Akaikeinfocriterion-2.374744Sumsquaredresid0.140426Schwarzcriterion-2.232229Loglikelihood30.00354Hannan-Quinncrite匚-2.344537F-statistic11.67434Durbin-Watsonstat1.000104ProtjfF-statistic)0.001S96InvertedARRoots.54圖17即有X二0.1729+Ut（1）其中U=0.5356U+Vtt—1tInX—InX二0.0803+0.5356(nX—InX)+Vtt—1t-1t—2t（2）于是得到：（3而0.0803表示線性趨勢的增長速度，從結(jié)果看，特征根是1/0.5356=1.87，滿足平穩(wěn)性要求3.3同樣，做Y的自相關(guān)函數(shù)與偏自相關(guān)函數(shù)圖：

□Ser[es:LOG(X)Workfile:UNTTTLED.:Un出Ied\|口|[前11玆|ViewIProc|Object|Properties|Print|Name\Freeze\SamplejGenr|Sheet|Graph|StatsjIdentjCorrelogramofD(LOG(X})Date:06/05/13Time:23:05Sample:197S2010Includedobservations:32AutocorrelationPartialCorrelationACPACQ-StatProbIZZIIZZI10.3480.3434.24210.039I□II1i20.1470.0295.022S0.031IgIIc13-0.006-0.1665.30340.151III14-0.602-0.61719.3710.001II:.I15-0.416-0.09726.3310.000I匚IIZJ16-0.1500.26127.2690.000IIIHI7-0.0030.10927.2700.000III!i80.4000.04934.5340.000I二II匚190.300-0.23939.0160.000IZlII11100.2090.05941.1730.000IaiI□1110.0740.13641.4630.000IOIiE112-0.274-0.06945.5600.000圖18從上圖可以看出，Y的自相關(guān)函數(shù)的一階滯后、4階滯后和5階滯后不為零，偏自相關(guān)函數(shù)的1階滯后與4階滯后不為零，是ARMA（4，5）的平穩(wěn)序列，所以原序列l(wèi)ny是ARIMA（4，1，5）序列。對Y序列進(jìn)行估計(jì)，輸出結(jié)果如下：tDependentVariable:DLOG(X)Method:LeastSquares□ate:06/05/13Time:23:10Sample(adjusted):19332010Includedobservations:28afteradjustmentsConvergenceachievedafter16iterationsf.-1ABackcast:197019S2VariableCoefficientStd.Errort-StatisticProb.LiJ.HI.J.YarrLiJ.VJLiJ.HI.J.YarrLiJ.VJ141450-(-rrL-fxJLrl—RRAAAAAMMM0.2020370.00316224.754140.00000.2174940.1614001.3474730.1915-0.5310310.149429-3.5537440.00130.2774370.0360203.2252490.0039-0.2112050.089S63-2.3502930.0281-0.9390720.046009-20.410710.0000R-squared0.823175Meandependentvar0.198032AdjustedR-squared0.782987S.D.dependentvar0.161513S.E.ofregression0.075240Akaikeinfocriterion-2.USS53Sumsquaredresid0.124544Schwarzcriterion-1.863301Loglikelihood36.00394Hmnnan-Quinncriter.-2.061501F-statistic20.43333Durbin-Watsonstat1.592608Prob(F-statistic)0.000000InvertedARRoots,66-.60i,66-^.60i-.55+.EQi-55-.60iInvertedMARoots.93.22-^.96i,22-.96i-84-.53i弭+.53i圖19

課號：課程名稱：金融實(shí)證分析方法閱卷教師:班級：—學(xué)號：—姓名：成績：由于AR(1)的參數(shù)不顯著，可以從模型中去掉。重新估計(jì)結(jié)果如下:DependentVariable:DLQG(X)Method:LeastSquaresDate:06/05/13Time:23:15Sample(adjusted):19332010Includedobservations:23afteradjustmentsConvergenceachievedafter30iterationsMABackcast19701982VariableCoefficientStd.Errort-StatisticProb.C0.1930410.01008010.532100.0000AR(4)-0.7291600.159754-4.5642710.0001MA(1)0.5687830.1554713.65844斗0.0013MA{4}0.4139740.2233461.3493670.0773MA(5)-0.1831430.264433-0.6924570.4956R-squared0.699236Meandependent汨r0.19S032AdjustecfR-squared0.646938S.D.dependent畑0.161513S.E.ofregression0.095963Akaikeinfocriterion-1.689283Sumsquaredresid0.211803Schwarzcriterion-1451390Loglikelihood20.64997Hannan-Quinncriter.-1.616557F-statistic13.37115Durbin-Watsonstat1.436888Protj(F-statistic)0.000009InvertedARRoots,65-.65i,65-.65i-,65+.65i-,65-n.65iInvertedMARoots.39,33-.61i,33+.61i旳+.5創(chuàng)-,01-.56i圖20MA(4)與MA(5)不顯著去掉，重新估計(jì)如下：Dependentvariable:DLOG(X)Method:LeastSquaresDate:06/05713Time:23:17Sample(adjusted):19832010Includedobservations:23afteradjustmentsConvergenceachievedafter11iterationsMABackcast1982VariableCoefficientStd.Errort-StatisticPr&b.C0.2007590.0266177.5425640.0000AR(4)-03613660.115423-3.1307070.0044MA(1)0.3973620.07235212.409630.0000R-squared0.635115Meandependentvar0.193032AdjustedR-squared0.60592&S.D.dependentvar0.161513S.E.ofregression0.101390Akaikeinfocriterion-1.638720Sumsquaredresid0.257000Schwarzcriterion-1.495903Loglikelihood25.94207Hannan-Quinncrite匚-1.595084F-statistic21.75741Durbin-Watsonstat2.274304ProbfF-stati