宏觀經(jīng)濟(jì)模型多種估計(jì)方法的EVIEWS實(shí)現(xiàn)(共10頁)

1、精選優(yōu)質(zhì)文檔-傾情為你奉上08統(tǒng)計(jì) 學(xué)號： 吳揚(yáng)一、 問題綜述建立中國宏觀經(jīng)濟(jì)模型。宏觀經(jīng)濟(jì)模型，是指以整個(gè)國民經(jīng)濟(jì)系統(tǒng)為研究對象，從總量水平和經(jīng)濟(jì)結(jié)構(gòu)方面來研究國民經(jīng)濟(jì)各變量之間的相互作用。它可用來評價(jià)宏觀經(jīng)濟(jì)政策、分析宏觀經(jīng)濟(jì)結(jié)構(gòu)和國民經(jīng)濟(jì)的發(fā)展趨勢。宏觀經(jīng)濟(jì)模型的表達(dá)可以用單一方程進(jìn)行表達(dá)，也可以用聯(lián)立方程組表達(dá)。本作業(yè)建立如下宏觀經(jīng)濟(jì)模型，完備的結(jié)構(gòu)式模型為其中，包含3個(gè)內(nèi)生變量，即國內(nèi)生產(chǎn)總值Y，居民消費(fèi)總額C和投資總額I；3個(gè)先決變量，即政府消費(fèi)G，前期居民消費(fèi)總額Ct-1和常數(shù)項(xiàng)。可以判斷，消費(fèi)方程是恰好識別的方程，投資方程是過度識別的，模型可以識別。數(shù)據(jù)來自題目提供。導(dǎo)入EVI

2、EWS二、 各種方法的EVIEWS實(shí)現(xiàn)1. 狹義的工具變量法估計(jì)消費(fèi)方程選取消費(fèi)方程中未包含的先決變量G作為內(nèi)生解釋變量Y的工具變量；在工作文件主窗口點(diǎn)擊quick/estimate equation，選擇估計(jì)方法TSLS，在equation specification對話框輸入消費(fèi)方程，在instrument list對話框輸入工具變量.點(diǎn)擊確定，得到：Dependent Variable: C01Method: Two-Stage Least SquaresDate: 06/02/11 Time: 14:08Sample (adjusted): 1979 2009Included obse

3、rvations: 31 after adjustmentsInstrument list: C G C01(-1)VariableCoefficientStd. Errort-StatisticProb.C1290.053402.73533.0.0034Y.0001C01(-1)34710.0000R-squared0.Mean dependent var34025.26Adjusted R-squared0.S.D. dependent var34218.49S.E. of regression1365.679Sum squared residF-statis

4、tic9402.761Durbin-Watson stat0.Prob(F-statistic)0.Second-Stage SSR得到結(jié)構(gòu)參數(shù)的工具變量法估計(jì)量：2. 間接最小二乘法估計(jì)消費(fèi)方程消費(fèi)方程中包含的內(nèi)生變量的簡化方程為參數(shù)關(guān)系體系為用普通最小二乘法估計(jì)第一個(gè)簡化式：Dependent Variable: C01Method: Least SquaresDate: 06/02/11 Time: 14:46Sample (adjusted): 1979 2009Included observations: 31 after adjustmentsVariableCoefficient

5、Std. Errort-StatisticProb.C1086.594386.55342.0.0089C01(-1)27720.0000G.0001R-squared0.Mean dependent var34025.26Adjusted R-squared0.S.D. dependent var34218.49S.E. of regression1380.725Akaike info criterion17.39037Sum squared residSchwarz criterion17.52914Log likelihood-266.5507Hannan-Q

6、uinn criter.17.43561F-statistic9198.948Durbin-Watson stat0.Prob(F-statistic)0.用普通最小二乘法估計(jì)第二個(gè)簡化式：Dependent Variable: YMethod: Least SquaresDate: 06/02/11 Time: 14:47Sample (adjusted): 1979 2009Included observations: 31 after adjustmentsVariableCoefficientStd. Errort-StatisticProb.C-1899.1342081.958-0.

7、0.3695C01(-1).0000G.0000R-squared0.Mean dependent var84244.67Adjusted R-squared0.S.D. dependent var95306.59S.E. of regression7436.521Akaike info criterion20.75796Sum squared resid1.55E+09Schwarz criterion20.89673Log likelihood-318.7484Hannan-Quinn criter.20.80320F-statistic2449.755Durb

8、in-Watson stat0.Prob(F-statistic)0.得到簡化式參數(shù)估計(jì)量為：由參數(shù)體系計(jì)算得到結(jié)構(gòu)參數(shù)間接最小二乘估計(jì)值為3. 二階段最小二乘法點(diǎn)擊objects/new object，選擇systemSystem: UNTITLEDEstimation Method: Two-Stage Least SquaresDate: 06/02/11 Time: 15:09Sample: 1979 2009Included observations: 31Total system (balanced) observations 62CoefficientStd. Errort-St

9、atisticProb.C(1)1290.053402.73533.0.0022C(2).0000C(3)34710.0000C(4)-2538.266948.1448-2.0.0097C(5)85760.0000Determinant residual covariance1.63E+13Equation: C01=C(1)+C(2)*Y+C(3)*C01(-1)Instruments: G C01(-1) CObservations: 31R-squared0.Mean dependent var34025.26Adjusted R-squar

10、ed0.S.D. dependent var34218.49S.E. of regression1365.679Sum squared residDurbin-Watson stat0.Equation: I=C(4)+C(5)*YInstruments: G C01(-1) CObservations: 31R-squared0.Mean dependent var34646.51Adjusted R-squared0.S.D. dependent var42513.37S.E. of regression3921.722Sum squared resid4.46E+08Durbin-Wat

11、son stat0.消費(fèi)方程的參數(shù)估計(jì)量為投資方程的參數(shù)估計(jì)量為4. 三階段最小二乘法System: UNTITLEDEstimation Method: Three-Stage Least SquaresDate: 06/02/11 Time: 15:20Sample: 1979 2009Included observations: 31Total system (balanced) observations 62Linear estimation after one-step weighting matrixCoefficientStd. Errort-StatisticProb.C(1)

12、1384.346361.67293.0.0003C(2).0000C(3)97460.0000C(4)-2538.266917.0495-2.0.0076C(5)72280.0000Determinant residual covariance1.55E+13Equation: C01=C(1)+C(2)*Y+C(3)*C01(-1)Instruments: G C01(-1) CObservations: 31R-squared0.Mean dependent var34025.26Adjusted R-squared0.S.D. depende

13、nt var34218.49S.E. of regression1390.396Sum squared residDurbin-Watson stat0.Equation: I=C(4)+C(5)*YInstruments: G C01(-1) CObservations: 31R-squared0.Mean dependent var34646.51Adjusted R-squared0.S.D. dependent var42513.37S.E. of regression3921.722Sum squared resid4.46E+08Durbin-Watson stat0.消費(fèi)方程的參

14、數(shù)估計(jì)量為投資方程的參數(shù)估計(jì)量為5. GMM（廣義矩估計(jì)）System: UNTITLEDEstimation Method: Generalized Method of MomentsDate: 06/02/11 Time: 15:27Sample: 1979 2009Included observations: 31Total system (balanced) observations 62Identity matrix estimation weights - 2SLS coefs with GMM standard errorsKernel: Bartlett, Bandwidth:

15、 Fixed (3), No prewhiteningCoefficientStd. Errort-StatisticProb.C(1)1290.053616.41172.0.0408C(2).0003C(3).0000C(4)-2538.2661067.430-2.0.0208C(5)78450.0000Determinant residual covariance1.63E+13J-statistic1.21E+13Equation: C01=C(1)+C(2)*Y+C(3)*C01(-1)Instruments: G C01(-1) CObse

16、rvations: 31R-squared0.Mean dependent var34025.26Adjusted R-squared0.S.D. dependent var34218.49S.E. of regression1365.679Sum squared residDurbin-Watson stat0.Equation: I=C(4)+C(5)*YInstruments: G C01(-1) CObservations: 31R-squared0.Mean dependent var34646.51Adjusted R-squared0.S.D. dependent var42513.37S.E. of regression3921.722Sum squared resid4.46E+08Durbin-Watson stat0.消費(fèi)方程的參數(shù)估計(jì)量為投資方程的參數(shù)估計(jì)量為三、 幾種方法的分析比較由上述各種結(jié)果可以看出，狹義的工具變量法（IV）、間接最小二乘法(ILS)、二階段最小二乘法(2SLS)與廣義矩陣法(GMM)，都得到了相同的參數(shù)估計(jì)量。前三種方法都是適用于恰好識別的結(jié)