計量經(jīng)濟學時間序列課后習題eviews解答

1、2.1980年-2013年中國全社會固定資產(chǎn)投資總額X與工業(yè)總額增加Y的統(tǒng)計資料如下:obsYXgeo1996.50091O9000力日wq2043.-00951GOOD19S22362.3001230.40019632375.600“3。J00q9B42789.OODI83.90019S5344S.TOO2543J2O09S63&&T-W>O3120eooT9674585.800370*1.TOOrsee5777.2004753.0006484.00D4410.4009906S5S.OOO45-17.000"1991OOS594.5OO-199210284.5

2、08DS0.100q99:38S,OO3073,3019941948070"04NJ190S2950.6020019.a0996之64A7：6099732?？?-4M-I-1O199S3-013,40234060199335S61.5029SS4.7020004.00336032917.7Q200-143580.6D37215口20024743T.304M99.902OD3&4$4Sc&65-5S6b,6O2口04652HO.OO7口47/一4口SOOS77230.80SS773.602006SiSdO.aO彳。99&言：之2007110534.93733.9

3、2008130260,2172S264S009135239.922459S.S201T6O72222516S3.320-111B847O2314051201219D67D.737r48自47r3210639=.-4EV1曰-Group：UNTITLEOWerkfs!e：UNTITL=ED;I1F；l=EcditOLzyVi='wProcicfccOfzrttcjProuQbjtut:PiritjNamu|Fruuxuj|Defeo!t|Boijzj試問：(1)當設定模型為lnY=B0+B1lnX+u時是否存在序列相關？(2)采用普通最小二乘法和穩(wěn)定標準誤差方法分別估計模型，比較參數(shù)估計

4、的差異和它們標準差的誤差，并說明穩(wěn)定標準誤差法克服序列相關后果的原理。(3)若原模型存在序列相關性，試用廣義最小二乘法估計模型。(1)1.圖示法DGraph:ETWcrkfile:UNTHlED：!Urtitled回|總編Vig?.：Proc,Object'PrintjManeAddTextLney-SnadeRemovelemnlaieQjtionjZoom07-0.8-0.95T.。1.1-1.2-133-.2-1.0.12.3從圖中可以殘差項的變化圖形判斷隨機干擾項的序列相關性為正相關。2 .原模型OLS估計口Uk二出口4工大科刊自UN-LED-Un*tlr-d.回,手Vfew|

5、Fnoc|otjECt，Prirrt|N3me|Fneezie|Estinate|Foreca£t5G1sResidsDepenntVariable;LOGY)Method:Lea&tSquaresDate:06/03/16Tim«:29:43Sample:19S02013IncludedobseRations.34VariableCoefficientStd.ErrortStatisticProb.C193740101280791512559Q.OOOOLOGCX)081420600127325395109O.OGOOR-squared0.99223SMeande

6、pendent'-ar9.S9O61SAdjustedR-squared0.991904S.Ddependentvar1.523353S.E.ofr&gression0136307Akaikeinfocrit&rion-1.090796Sumsquaredresid0594544Schwaizcriterion-1.001010LogliKelirood20.54354Hannan-Quinncriter.-1.D6O177F-statlStic4089,742Duriiin-V/atsonstat0.230617ProtXF-statiStC)0.000000由于DW

7、值為0.28,dL(k=2,T=34)=1.39DW值小于dL,應此可以判斷模型存在序列相關。3 .LM檢驗曲EViewc-(EqualUNTTTLEDWorkfileUNTITLED:tUrrtkEleC3FileEditOiiycrtViewProcQuickOption3WindowHelpWcf>pcObtiert|KtjnatgFbecastfStvta總重點dzBrentch-fio白心iftICcrreIartLF.4eeistF3t0titic7G.9959SPTO00.0000口yR-qusirRtfl?441日PrctiChi仔不用片1)口onnoTestEqiuti

8、onD»!PurlduntViIut>to.RESIDMMharf_Le曰31E<1ugreaDalFoa?a3/16Time2JODSample.1sso2013Includeeobservations:34PresamplemlaaingwalufrlaggedraidiualAtoz»ro.vanaDieccerrici&ntsta.Errortstatist*cftcd.c0.(1581710.009a95O.S3&2520.40S3LUG(X)-.Ubt)1/U.0(Jt>9UtiU.9i/S/U0.34MF1E3ID(-1>

9、;091SMCIT0102GG20007404c.aoooR-SQUar&OMeandiepentjentwarb.OIL16Adjusi&dR-squareciS.Dcfependentwar0134.22SS.E.ofu白gi已臺號Jon0.073623AkidIkyinfocriterion-2.298338oumsquaredresid0.175/5iscriwarzcriterion-2.163bS9Loglikelihoodl4207175Hannfln-OLiiniricritsr-2252409FWBHEIC30,49297Durbin-Watsan方tai1.

10、401173Prob(F-statistic).OOOOQ根據(jù)nRA2統(tǒng)計量對應的值得出LM=34*0.699=23.76,在5%顯著性水平下存在一階序列相關性。感EView5-Equation:UNTITLEDWorkfik：UNTUIED:UntitledPIFileEditObjectViewProcQuickOptionsWrnciowHelpVieMProc|ObjectPriit|Pame|Freeze|Estindte|Forecast|Stats|Kesds|Breuscfi-GocJEre.SerialCorrelatronLMTestF-statistic4281526Pr

11、obF(2.30)0.0D000bs*R-squ3ne<i25.173&0ProbCh!-3quare(2:0.0000TestEquation:Dependentvarisble:RESIDf.lettiod:LeastSquaresDate:06/03/16Time;2159Sampic19802013Includedobservations:34PresamplermsHngvaluelaggedresidualssettozerc.Variabi&CoefficientStC.Errort-StatislicProbC00379E4S068B350.5513740

12、.5855LOGCX)T0043560Q063S2-0.6328910316RESIDE1.15217601794656.4193370.0000RESID(-2-,3044110.189117-16096410.1130R-sqjared0.740553Meandependentvar-5.O1E-16AdjustedR-squy&dQfDBs.DJependentvar0.134220S.Eofregressiori0.071706上Maikeinfocrikrtcn23陽52Surnsquaredrasid0.1542t3Scriwatzcriterion-2.142780Log

13、likelihoodHannan-Quinnenter小pfinnF-statistic28$4390Durbin-V.,atsonstat1.794S44ProbtFtatistic)0.0口。阿2階序列相再繼續(xù)LM檢驗的2階序列相關性檢驗，發(fā)現(xiàn)在5%顯著性水平下，原模型存在關性，但在t-test中，RESID(-2)的參數(shù)為0的假設。故不存在2階序列相關性(2)序列相關穩(wěn)健標準誤差法Equation:NEWEYWorkFle:UNHTLED:UntWed歸ViewPE|Objectftinft|NamE|FreeFe|Estimate|boreca/H5括tsgesidsDependent

14、Variable.LOGtY)Metliod:LeastSquaresDate:06/03/46Time:£1:04Sanpie:I9602013IneludedQbsnatianf：34hleweyVestHACStandardErrors&Covari<mce(Faigtruncatcn)VariableCoefficientStd.Errort-StatistrcProb.C19374010202。船95866730.0000LOGCX)0.8U2CSQ.C2044S39.3184400000R-squarect0992235IVteandependentvar9

15、.990519AdjustedR-squarea0.991994S.O.dependentvar1.523353S.E.ofregressian0136307AKaikeinfocriterion-1.09079S5umsquaredresid0,5945"Schwajz.criterian-1.001010LoglikfrHhood20.543&4Hannan-Quinnenter.4.060177F-statistic4089742urbin-Wstsonstat0230517Ffob(F-statistic)0.000000與原模型OLS估計對比Q.S7/CMUN-;T

16、LED：Unttl-d.U回羽View:ProciObject'Pdnt|FreezieFEijrngg|£口那匚町與？.3GtsResidsDependentVariabk:LOGCY)Method:LeastSquaresDate：06/03/16Time：20:43Sample:19S02C13IncludedoDser<ation£：34VariableCoefficientStdErrorStatisticProb.C1937401012807915125590.0000LOGCX)081420500127320395109O.OGOOR-squar

17、edD.99223BMeandependentvar&.sgos-isAdjustedR-squaretf0.991904SDdependentvar1523353S.E.ofregression0136307Aksikeinfocriterion-1090796Sumsquaredresid0594544Schwaizcriterion-1.001010Logiikelirood20.54354Hannan-Quinncrtier.-1.060177F-statistic4089.742urtjin-?/atsonslat0.260517ProbfF-statistic；0.0000

18、00發(fā)現(xiàn)變量X的對應參數(shù)修正后的標準差OLS結果有所增大，表明原模型OLS估計結果低估了X的標準差。(3)廣義最小二乘法估計由于LM檢驗只要一階序列相關性，故：軍EVlews-(Equstton;ARIWorkfile;UNTIT_EUntitldFileEditObjectViewProcQuickOptionsWindowHelp陛依jproc|cibct|Print|Piamc|F3H&|Estiffimte，F(xiàn)g其器t|StatsDependentVariable:LOG(Y)Methoa.LeastSquaresDate:6Time:21:33Sample(adjusted):-1Q812013Includedobservations;33afteradjustmentsConvergenceachievedafter13iterationsVariableCoefficientStdErrorStatisticProc.C94344306.0fl6531556412口1301LOG兇046106S0.10571243315610.0001AR(1)0.986440Q.015S4358565960.0000R'Squared0998356Meanaependentvar10,053