eviews面板数据模型详解.docx

1、eviews面板数据模型详解1.已知 19962002年中国东北、 华北、华东 15 个省级地区的居民家庭人均消费（cp，不变价格）和人均收入（ip，不变价格）居民，利用数据（1）建立面板 数据（ panel data ）工作文件；（ 2）定义序列名并输入数据； （ 3）估计选择面板 模型；（ 4）面板单位根检验。年人均消费（con sume）和人均收入（in come）数据以及消费者价格指数（p ）分别见表 9.1， 9.2 和 9.3。表 9.2 1996 2002 年中国东北、华北、华东表 9.1 1996 2002 年中国东北、华北、华东 15 个省级地区的居民家庭人均消费（元）数据人

2、均消费1996199719981999200020012002CONSUMEAH3607.433693.553777.413901.814232.984517.654736.52CONSUMEBJ5729.526531.816970.837498.488493.498922.7210284.6CONSUMEFJ4248.474935.955181.455266.695638.746015.116631.68CONSUMEHB3424.354003.713834.434026.34348.474479.755069.28CONSUMEHLJ3110.923213.423303.153481.74

3、3824.444192.364462.08CONSUMEJL3037.323408.033449.743661.684020.874337.224973.88CONSUMEJS4057.54533.574889.435010.915323.185532.746042.6CONSUMEJX2942.113199.613266.813482.333623.563894.514549.32CONSUMELN3493.023719.913890.743989.934356.064654.425342.64CONSUMENMG2767.843032.33105.743468.993927.754195.

4、624859.88CONSUMESD3770.994040.634143.964515.0550225252.415596.32CONSUMESH6763.126819.946866.418247.698868.199336.110464CONSUMESX3035.593228.713267.73492.983941.874123.014710.96CONSUMETJ4679.615204.155471.015851.536121.046987.227191.96CONSUMEZJ5764.276170.146217.936521.547020.227952.398713.08人均收入1996

5、199719981999200020012002INCOMEAH4512.774599.274770.475064.65293.555668.86032.4INCOMEBJ7332.017813.168471.989182.7610349.6911577.7812463.92INCOMEFJ5172.936143.646485.636859.817432.268313.089189.36INCOMEHB4442.814958.675084.645365.035661.165984.826679.68INCOMEHLJ3768.314090.724268.54595.144912.885425.

6、876100.56INCOMEJL3805.534190.584206.644480.0148105340.466260.16INCOMEJS5185.795765.26017.856538.26800.237375.18177.64INCOMEJX3780.24071.324251.424720.585103.585506.026335.64INCOMELN4207.234518.14617.244898.615357.795797.016524.52INCOMENMG3431.813944.674353.024770.535129.055535.896051INCOMESD4890.285

7、190.795380.085808.966489.977101.087614.36INCOMESH8178.488438.898773.110931.6411718.0112883.4613249.8INCOMESX3702.693989.924098.734342.614724.115391.056234.36INCOMETJ5967.716608.397110.547649.838140.58958.79337.56INCOMEZJ6955.797358.727836.768427.959279.1610464.6711715.615 个省级地区的居民家庭人均收入（元）数据15个省级地区的

8、消费者物价指数表9.3 1996 2002年中国东北、华北、华东物价指数1996199719981999200020012002PAH109.9101.310097.8100.7100.599PBJ111.6105.3102.4100.6103.5103.198.2PFJ105.9101.799.799.1102.198.799.5PHB107.1103.598.498.199.7100.599PHLJ107.1104.4100.496.898.3100.899.3PJL107.2103.799.29898.6101.399.5PJS109.3101.799.498.7100.1100.899

9、.2PJX108.410210198.6100.399.5100.1PLN107.9103.199.398.699.910098.9PNMG107.6104.599.399.8101.3100.6100.2PSD109.6102.899.499.3100.2101.899.3PSH109.2102.8100101.5102.5100100.5PSX107.9103.198.699.6103.999.898.4PTJ109103.199.598.999.6101.299.6PZJ107.9102.899.798.810199.899.1（1）建立面板数据工作文件 首先建立工作文件a i YH-i

10、llli+HII!fl f IDetdifc+ASihow | FetchStore,Delete)mple|Range:Ssntple:E c 回 rsidGenerate Series. BDisplay Filter *New Obj set. Vorkfile: U1TTTTLEDTetcli from DB .Update sslectei from DB.Stor* selected to DB.Copy .slectedL . selectelFrijit Selected建立面板数据库在窗口中输入15个不同省级地区的标识EVlews - Pool: COM3UME TaikTi

11、le: UHTITLED: :UntitledX -nFil Edi t Objoc t Vi -cw Px-c-c gui ck 牡 iez Windo-w Help_ &View lProcOb)ect |Prrit(Name Fretze j Estimate Detine PoolGerir jl5heetCress Sect ion Identifiers : (Ent e r idenct if iers bel ow this： line)AH BJ FJHB HLJJL JS JX LNNMG SD SH sx TJ ZJI(2)Pool: CDHSlTVE Torkf ile

12、= U9TTTLED: r Tint it ledTji11 OilsPrint Nrrik- Fteezt*Crops S&ct ion I dam iTiers：: (Hnt ar idntif iars tT-nw this lina)AHBJFJ H呂 HLJ JL JS JX L,M NMGSU&H sx TJZJQVi-iAJ Pr 匚 I Object定义序列名并输入数据产生3*15个尚未输入数据的变量名。这样可以通过键盘输入或黏贴的方法数据 数据。(3)估计、选择面板模型打开一个pool窗口，先输入变量后缀(所要使用的变量)。点击Estimate，打开 估计窗口。 Pool:

13、FOOL1 Torkfile: UKTLTLED:UntitLedCrass Section Identifiers: (Enter identifiers below this line JAHBJ FJ HB HLJJLJSJXLNNMGSDSHSXTJ ZJSpecific iti on O-pti onsDependent vr i ibl eE+irniti:n mthdF in电d. and. Rhl&teHoneVKoneVLgjltsVCrQsssactitfeigjils No zRegressors wnd. ARO terms Ccmni确定A.混合模型的估计方法左边的C

14、ommon表示相同系数，即表示不同个体有相同的斜率Pool EatiBatdonSpecific ttLoft OptionsDepend ent vari able.rpQEstination ntholRegressors and KFi0 termsCflmmonCt&ss-5tc*in spcifi cssttingsFined tid 卫11GELS - Leurs （辺奴 AH）Sample 1996 2002确走取捎得到如下输出结果:Dependent Variable: CP?Method: Pooled Least SquaresDate: O7XK/O0 Time: 13:

15、13Sample: 1996 2002Included observations: 7Cross-sections included 15Total pool (balan匚旦d) oben/ations: 106VariableCoefficientStd. Error NStatisticProbC129.631363 69259 2 0352650.04i4IP?O76072E0.009622 79.681990.0000R-squaredQ984036Mean dependent var4917.608Adjusted R-squared0.983381S.D. dependentva

16、r1707048.E. of regression216.4270Aka ike info criterion13&1125Sum squared resid4B24588.Schwarz criterion13,66180Log likelihood-712.5905F-st artistic8349.204Durbin-Wat son stat0 734107Prob(F-statistic).000000相应的表达式是:CI? 129.63 0.76IRt48245882(2.0) (79.7) R 0.98, SSEr上式表示15个省级地区的城镇人均指出平均占收入的 76%B.个体固定

17、效应回归模型的估计方法 将截距项选择区选Fixed effects （固定效应）Dependent vari blsFix*d and Eandomittthod: LSLeast SquarsE Cand A2)tiiLgsqvs AB Q 弋旨fbwAm，or. Ampl e19QS 2CO21 BslIsjic*Safnpli得到如下输出结果:Dependent Variable: CP?Method: Pooled Least SquaresDate: O7Z12/O0 Time: 13:36Sample: 1996 2002Included observations: 7Cross-

18、sections included: 15Tot al pool (balanced) enervations: 105VariableCoefficientStd. Errort-StatisticProb.C515.S14281 596656.319061o.oaooIP?0.6975B1001269254.9G0290 0300Fixed Effects (Crass)AHC*36.30503BJ-C537.5660FJM-47.64515HB-C-164 2367HLJ-C*169.7015JL-C24.50427尽（-35.19587JX-C-319.6960LN-C106.4273

19、MMG-C-209.5484SD-C-134 1145SH-C266.9859SX-C74.60901TJ-C47.22940ZJ-C193.6202Effects SpecificaiionCross-section fixed (dumnn/ variablE司R-squared0.992408Mean deperdent irar4917.600Adjusted R-squared0.991222S.D dependent var1704704S.E. of regression159.7104Akaike info criterion13.12414Sum squared re aid

20、2270386.Schwarz criterion13.52955Log likelihood-673.0173F-statistic703.0902Durtin-Watson stat1 609517Prob(F-statistic)0 oooooo相应的表达式为:(6.3) (55)R2 0.99, SSEr2270386C? 515.6 0.70IPit 36.3D, 537.6D2 . 198.6D15其中虚拟变量Di,D2,., D15的定义是:1,如果属于第个个体,i 1,2,.,150,其他15个省级地区的城镇人均指出平均占收入 70%。从上面的结果可以看出北京市 居民的自发性消

21、费明显高于其他地区。接下来用F统计量检验是应该建立混合回归模型， 还是个体固定效应回归模型。Ho: i 。模型中不同个体的截距相同(真实模型为混合回归模型) 。Hi :模型中不同个体的截距项 i不同(真实模型为个体固定效应回归模型)。F统计量定义为：F (SSE SSEJ/(NT k 1) (NT N k) (SSE SSEU)/(N 1)SS/(NT N k) SSEu/(NT N k)其中SSE表示约束模型，即混合估计模型的残差平方和， SSEU表示非约束模型，即个体固定效应回归模型的残差平方和。非约束模型比约束模型多了 N 1个被估参数。所以本例中：(4824588 227386)/(1

22、5 1) rF 8.1 f F005 (14,89) 1.82270386/(105 15 1)所以推翻原假设，建立个体固定效应回归模型更合理。C.时点固定效应回归模型的估计方法 将时间选择为固定效应。得到如下输出结果:Depended Variable: CP?Method: Pooled Laa&t SquaresDate: 07/02/08 Time- U:44Sample: 1996 2002Included observations： 7Cross-sectiors included 15Total pool (balari匚日H) obs&rvaiions: 105Variable

23、CoefficientStd. ErrorStatisticProb.C2.57723460.BO5460.0374570.9702IP?0 7783600.01043874.619110.0000Fixed Effects (Period)1996-C105 92851997-C134 05561998C54.848281999-C-37.074042000-C-7 1183592001-C-15B747430Q2-C-93.39200Effects SpecificatiorPeriod fixed (dummy variables)R-squarad0.9366G9Mean depend

24、ent var4917.6D8Adjusted R-sqjared0.985707S.D. dependent var1701.701S.E. of regrsssioti203.8001Akaike info criteriiin13.54529Sum squared resiri4028843.Echwar! criterioh13.74749Log likelihood-703.127SF-sltistic1025.643Durbiri-Watson0.795376Pro b(F-statistic)0.000000(76.6) R2 0.987, SSE4028843相应的表达式为:其

25、中虚拟变量D!,D2,., D7的定义是:1,如果属于第t个截面，t=1996,.,20020,其他D.个体随机效应回归模型估计截距项选择Random effects （个体随机效应）得到如下部分输出结果:D 即 end ent Variable: CP?Method： Pooled EGLS (Crosse set ion random effects)Date: 07/Q2/OS Time: 15:06Sample: 1996 2002Included observtio7Cross-sections included： 15Total pool (balanced) observatio

26、ns: 105Swarny nd Arora etintor of cump&ntnt variflincejsVariableCoefficientSid. Errort-StatisticProb.C345.179575.472174.5735990.0000IPf0 7245690 01057268.538140.0000Random Effects JCrcss)AH-C-2.553433BJ-C367 0439FJ-C-51.24006HB-C-104.8357HLJ-C-101 7600JL-C54.90671JS-C-32.27363心-223.9519LN-C112.1152NMG-C-133.13775D-C-100.8713SH-C126.1820X C*22,79189TJ-C100S794ZJ-C106 0939相应的表达式是:Cpt 345.2 0.72IRt 2.6D1 367.0D2 . 106.1D15