.word.zl..word.zl.Eviews面板數(shù)據(jù)之固定效應(yīng)模型在面板數(shù)據(jù)線(xiàn)性回歸模型中，如果對(duì)于不同的截面或不同的時(shí)間序列，只是模型的截距項(xiàng)是不同的，而模型的斜率系數(shù)是一樣的，那么稱(chēng)此模型為固定效應(yīng)模型。固定效應(yīng)模型分為三類(lèi)：個(gè)體固定效應(yīng)模型個(gè)體固定效應(yīng)模型是對(duì)于不同的縱剖面時(shí)間序列〔個(gè)體〕只有截距項(xiàng)不同的模型：y=九+◎卩x+uitikkitit(1)k=2從時(shí)間和個(gè)體上看，面板數(shù)據(jù)回歸模型的解釋變量對(duì)被解釋變量的邊際影響均是一樣的，而且除模型的解釋變量之外，影響被解釋變量的其他所有〔未包括在回歸模型或不可觀(guān)測(cè)的〕確定性變量的效應(yīng)只是隨個(gè)體變化而不隨時(shí)間變化時(shí)。檢驗(yàn)：采用無(wú)約束模型和有約束模型的回歸殘差平方和之比構(gòu)造F統(tǒng)計(jì)量，以檢驗(yàn)設(shè)定個(gè)體固定效應(yīng)模型的合理性。F模型的零假設(shè)：H:九二九二九二…二九二00123N-1(RRSS-URSS)F=URSSN^1F(N-1,N(T-1)-K+1)/(NT-N-K+1)RRSS是有約束模型〔即混合數(shù)據(jù)回歸模型〕的殘差平方和，URSS是無(wú)約束模型ANCOVA估計(jì)的殘差平方和或者LSDV估計(jì)的殘差平方和。實(shí)踐：一、數(shù)據(jù)：1996—2002年中國(guó)東北、華北、華東15個(gè)省級(jí)地區(qū)的居民家庭人均消費(fèi)〔cp，不變價(jià)格〕和人均收入〔ip，不變價(jià)格〕居民，利用數(shù)據(jù)〔1〕

建立面板數(shù)據(jù)〔paneldata］工作文件；〔2〕定義序列名并輸入數(shù)據(jù)；〔3〕估計(jì)選擇面板模型；〔4〕面板單位根檢驗(yàn)。年人均消費(fèi)〔consume〕和人均收入〔ine〕數(shù)據(jù)以及消費(fèi)者價(jià)格指數(shù)〔p〕分別見(jiàn)表1，2和3。表11996—2002年中國(guó)東北、華北、華東15個(gè)省級(jí)地區(qū)的居民家庭人均消費(fèi)〔元〕數(shù)據(jù)人均消費(fèi)1996199719981999200020012002CONSUMEAH3607.433693.553777.413901.814232.984517.654736.52CONSUMEBJ5729.526531.816970.837498.488493.498922.7210284.6CONSUMEFJ4248.474935.955181.455266.695638.746015.116631.68CONSUMEHB3424.354003.713834.434026.34348.474479.755069.28CONSUMEHLJ3110.923213.423303.153481.743824.444192.364462.08CONSUMEJL3037.323408.033449.743661.684020.874337.224973.88CONSUMEJS4057.54533.574889.435010.915323.185532.746042.6CONSUMEJX2942.113199.613266.813482.333623.563894.514549.32CONSUMELN3493.023719.913890.743989.934356.064654.425342.64CONSUMENMG2767.843032.33105.743468.993927.754195.624859.88CONSUMESD3770.994040.634143.964515.0550225252.415596.32CONSUMESH6763.126819.946866.418247.698868.199336.110464CONSUMESX3035.593228.713267.73492.983941.874123.014710.96CONSUMETJ4679.615204.155471.015851.536121.046987.227191.96CONSUMEZJ5764.276170.146217.936521.547020.227952.398713.08表21996—2002年中國(guó)東北、華北、華東15個(gè)省級(jí)地區(qū)的居民家庭人均收入〔元〕數(shù)據(jù)人均收入1996199719981999200020012002INEAH4512.774599.274770.475064.65293.555668.86032.4INEBJ7332.017813.168471.989182.7610349.6911577.7812463.92INEFJ5172.936143.646485.636859.817432.268313.089189.36INEHB4442.814958.675084.645365.035661.165984.826679.68INEHLJ3768.314090.724268.54595.144912.885425.876100.56INEJL3805.534190.584206.644480.0148105340.466260.16INEJS5185.795765.26017.856538.26800.237375.18177.64INEJX3780.24071.324251.424720.585103.585506.026335.64INELN4207.234518.14617.244898.615357.795797.016524.52INENMG3431.813944.674353.024770.535129.055535.896051INESD4890.285190.795380.085808.966489.977101.087614.36INESH8178.488438.898773.110931.6411718.0112883.4613249.8INESX3702.693989.924098.734342.614724.115391.056234.36INETJ5967.716608.397110.547649.838140.58958.79337.56INEZJ6955.797358.727836.768427.959279.1610464.6711715.6表31996—2002年中國(guó)東北、華北、華東15個(gè)省級(jí)地區(qū)的消費(fèi)者物價(jià)指數(shù)物價(jià)指數(shù)1996199719981999200020012002PAH109.9101.310097.8100.7100.599PBJ111.6105.3102.4100.6103.5103.198.2PFJ105.9101.799.799.1102.198.799.5PHB107.1103.598.498.199.7100.599PHLJ107.1104.4100.496.898.3100.899.3PJL107.2103.799.29898.6101.399.5PJS109.3101.799.498.7100.1100.899.2PJX108.410210198.6100.399.5100.1PLN107.9103.199.398.699.910098.9PNMG107.6104.599.399.8101.3100.6100.2PSD109.6102.899.499.3100.2101.899.3PSH109.2102.8100101.5102.5100100.5PSX107.9103.198.699.6103.999.898.4PTJ109103.199.598.999.6101.299.6PZJ107.9102.899.798.810199.899.1二、1.輸入操作：步驟：〔1〕File——New——Workfile0FileEditObjectViewPro-匚QuickOptionsAdd-insWindowHelpflewWorLcfile...匚trl+N.0_penDatabase...Ctrl+5frogramSaveAs...TertFileClose|Irriport卜步驟：〔2〕StartdateEnddateK

步驟：〔3〕ObjectNewObject步驟：〔4〕TypeofobjectPool步驟：〔5〕輸入所有序列名稱(chēng)EPool:POOLMODELWorkfile:UNTITLIED::Untitled\-nView!ProcObjectPrintNameFreezeEstimateD-efinePooLGenrSheetCrossSectidn.Identifierz：tinieridentiEiers"belowthisline)AHBJFJHBHLJJLJ-SJXL忖NMGSDSHsxTJZJ步驟：〔6〕定義各變量點(diǎn)擊shee—輸入consume?ine?p?回Pool;POOLMODELWorkfile;UNTITLED;:Untitled\-已sc步驟：〔7〕將表1、2、3中的數(shù)據(jù)復(fù)制到Eviews中2.估計(jì)操作：obsCONSUME?INCOME?P?obsCONSUME?INCOME?P?AH-199E3607.430451Z77D109.9000AH-19973693.5504599.27D101.3000AH-19983777.410477(1.470100.0000AH--iggg3901.8105064.60097.80000AH-20004232.9805291550100.7000AH-20014517.6505663.800100.5000AH-2002斗736.5206032400gg.QoaooBJ-19965729.5207332.010111.6000BJ-19976531.8107313.1BD105.3000BJ-19986970.8303471.9SO102.4000BJ-iggg7498.4809132.760100.6000BJ-20003493.4901034969103.5000步驟：〔1〕點(diǎn)擊poolmodel——Estimate對(duì)話(huà)框說(shuō)明Dependentvariable:被解釋變量；mon:系數(shù)一樣局部Cross-sectionspecific:截面系數(shù)不同局部步驟：〔2〕將截距項(xiàng)選擇區(qū)選Fixedeffects〔固定效應(yīng)〕Cross-section:Fixed得到如下輸出結(jié)果：DependentVariable:CONSUME?Method:PooledLeastSquares□ate:07/16/14Time:11:06S-ample:19962002Includedobservations:7Cross-sectionsincluded:15Totalpool(^balanced)observations:105VariableCoefficientStd.Errort-statisticProb.C596.504989.94504-6.839263o.ooaciINCOME?0.6862^20.013S5049.&48B2o.ooaaFiicedEffects(CroesiAH-C-53.23597BJ-C592.4387FJ-C-41.75834HB^C-189.6295.HU-C-192.0^54」L_U0.4-9391&-^6.60391」X—C-^41.5000LN-C.76302NMG-C-230.1840SD-C-140.3215SH-C227.1060SX-C-9&.131S0TJ-C61.43642ZJ-C230.1580EffectsSpecificationCross-sectionfixed(duinmyvariable^jR-squared0.992490Meand&pendentyar4931.017AdjustedR-squared0.99122&■S.D.dependentvar1700.985S.E.ofregression159.34^6Akaikeinfocriterion1211944Sumsquaredresid2259743.■Schwarzcriterion13.52385Loglikelihood-672.7706Hannan-Quinneriter.13.28332F-statistie754.1521□urbin-Watsonstat1.624146Prob[F-statistic)o.oaaooo接下來(lái)用F統(tǒng)計(jì)量檢驗(yàn)是應(yīng)該建立混合回歸模型，還是個(gè)體固定效應(yīng)回歸模型。H:S。模型中不同個(gè)體的截距一樣〔真實(shí)模型為混合回歸模型〕0iH:模型中不同個(gè)體的截距項(xiàng)?不同〔真實(shí)模型為個(gè)體固定效應(yīng)回歸模型〕1i對(duì)模型進(jìn)展檢驗(yàn):(RRSS-URSS)/(4965275-2259743)/'F=URSS(NT-N-KI)=—22597439?？谇?F0(14,90)=1W3所以推翻原假設(shè)，建立個(gè)體固定效應(yīng)回歸模型更合理。RRSS求法請(qǐng)參見(jiàn)Eview面板數(shù)據(jù)之混合回歸模型相應(yīng)的表達(dá)式為：Consume二596.50+0.69Income-53.23D+592.44D+...+230.16Ditit1215(6.64)(49.55)R2二0.99,SSE二2259743r其中虛擬變量D,D，…,D的定義是：1215[1,如果屬于第個(gè)個(gè)體,i=1,2,...,15i[0,其他15個(gè)省級(jí)地區(qū)的城鎮(zhèn)人均指出平均占收入68.62%。從上面的結(jié)果可以看出市居民的自發(fā)性消費(fèi)明顯高于其他地區(qū)。時(shí)點(diǎn)固定效應(yīng)模型時(shí)點(diǎn)固定效應(yīng)模型就是對(duì)于不同的截面〔時(shí)點(diǎn)〕有不同截距的模型。如果確知對(duì)于不同的截面，模型的截距顯著不同，但是對(duì)于不同的時(shí)間序列〔個(gè)體〕截距是一樣的，那么應(yīng)該建立時(shí)點(diǎn)固定效應(yīng)模型：y=丫+*卩x+uittkkitit(2)k=2時(shí)點(diǎn)固定效應(yīng)模型與個(gè)體固定效應(yīng)模型的操作區(qū)別在于步驟〔2〕，將時(shí)間項(xiàng)選擇區(qū)選Period:Fixed〔時(shí)間固定效應(yīng)〕PoolEstimation￡<?firalKr>Options,Esbmnhon理tfriRiyaMethodIS-LeaitSquares(andAR)BalancePoolEstimation￡<?firalKr>Options,Esbmnhon理tfriRiyaMethodIS-LeaitSquares(andAR)BalanceDepie*Sample:19^62DD2得到如下結(jié)果:DepensentVariable:CONSUME'?Melhad:PaaledlLE^stSquaresDaita:Q7f2i/14TirTie:n：oaSample:19962002lndudl9dob^$rvaUon￡：7Crosseecuonsmdude<F15Taialpool{balanced);□bservztiDns105VanableCoefficientSid.Errorl-StaU5ticProb.c-2.63D22560.5&332■0.D3B3520.9695INCOME?0.7300050.0102^47599695a0000FixedEiredts(Period)1986--C11402601997-C137.5006igge-c53.93619-I599-C?30.641272O0D-C-9.0450?2O01-C■16002642Q02-C?97.74908EiiedsSp^dfic^tionPenddflxed(dummyvanR-squared0.93M39制軸ndepgriMil泊r4981.017MustedR-squared0.985460S.Dctependenlvar1700905S.E.nfregression205.1087月kaikeinfacriierian13.55809sumsquaredresid4000749.Sctiwarzcritennn13.7B030Laglikelihood-7017997Hannan-Ouinn亡『廿皂「13.64003F-sla11sllc1007.9^3Durbln-Walson￡tatD.736995Pro-Stanslic)0.000000接下來(lái)用F統(tǒng)計(jì)量檢驗(yàn)是應(yīng)該建立混合回歸模型，還是個(gè)體固定效應(yīng)回歸模型。h:a=a。模型中不同個(gè)體的截距一樣〔真實(shí)模型為混合回歸模型〕。0iH:模型中不同個(gè)體的截距項(xiàng)a不同〔真實(shí)模型為時(shí)間固定效應(yīng)回歸模型〕。1t對(duì)模型進(jìn)展檢驗(yàn):(RRSS-(RRSS-URSS)F=T一1URSS/(NT-T-K+1)=3.54>F0(56‘98)=219所以推翻原假設(shè)，可以建立時(shí)點(diǎn)固定效應(yīng)回歸模型RRSS求法請(qǐng)參見(jiàn)Eview面板數(shù)據(jù)之混合回歸模型相應(yīng)的表達(dá)式為:Consume=—2.6+0.78IP+114D+137.5D+...—97.7Ditit127(76.0)R2=0.986,SSE=4080749其中虛擬變量D,D，…,D的定義是：127f1,如果屬于第t個(gè)截面，t=1996,...,2002D=2t[o,其他時(shí)點(diǎn)個(gè)體固定效應(yīng)模型時(shí)點(diǎn)個(gè)體固定效應(yīng)模型就是對(duì)于不同的截面〔時(shí)點(diǎn)〕、不同的時(shí)間序列〔個(gè)體〕都有不同截距模型。如果確知對(duì)于不同的截面、不同的時(shí)間序列〔個(gè)體〕模型的截距都顯著地不一樣，那么應(yīng)該建立時(shí)點(diǎn)個(gè)體固定效應(yīng)模型：y=九+丫+丫卩x+u(3)itttkkitit(3)k=2時(shí)點(diǎn)固定效應(yīng)模型與個(gè)體固定效應(yīng)模型的操作區(qū)別在于步驟〔2〕，將截距項(xiàng)選擇區(qū)域：Cross-section:fixed〔個(gè)體固定效應(yīng)〕，時(shí)間項(xiàng)選擇區(qū)選Period:Fixed〔時(shí)間固定效應(yīng)〕PoolEstimationSpEdficaiionPoolEstimationSpEdficaiionOptions得到結(jié)果如下:DependentVariable:CONSUME?Method:PooledLeastSquarest-StatisticProb.t-StatisticProb..word.zl..word.zl.Date:07/21/14Time:15:44Sample:19962002Includedobservations:7Cross-sectionsincluded:15Totalpool(balanced)observations:105VariableCoefficientStd.ErrorC806.6751221.2143INE?0.6533380.034541FixedEffects(Cross)AH--C-94.50854BJ--C698.0132FJ--C-18.86465HB--C-200.3997HLJ--C-246.3712JL--C-54.16421JS--C-31.26919JX--C-392