計(jì)量經(jīng)濟(jì)學(xué)作業(yè) (6)(共5頁(yè))

1、2、設(shè)模型(mxng)（C/CP）=1+2YCP+ut，（t=1，2，,n）模型(mxng)中各變量的解釋參見式（369）。（1）利用(lyng)表3-1給出的數(shù)據(jù)，運(yùn)用廣義最小二乘法，對(duì)上述模型進(jìn)行估計(jì)。做變換時(shí)，用1-DW2代替未知參數(shù)，其中DW為對(duì)模型進(jìn)行最小二乘估計(jì)時(shí)給出的DW值。（2）對(duì)比最小二乘法、廣義最小二乘法和CO方法給出的各個(gè)估計(jì)方程，計(jì)算各個(gè)方程給出的被解釋變量的估計(jì)值，并對(duì)結(jié)果進(jìn)行解釋。城鎮(zhèn)家庭實(shí)際人均消費(fèi)性支出與世紀(jì)人均可支配收入年度C/CPY/CP19801634.561892.9819811766.431935.0319821785.442029.191983188

2、0.672098.8819842024.612360.1219852229.142447.3519862484.452801.3119872562.742903.819882692.682878.5419892503.12839.8119902563.933027.5919912819.083297.6519923046.663693.4619933353.674095.0119943650.36447619953867.464682.3519963956.694884.8219974110.385067.5619984287.865370.3219994634.445877.51200049

3、986279.9820015272.116811.9220026036.567711.2820036440.748380.8520046837.498969.5420057428.119812.9620068013.0410835.2520078789.7612120.4520089333.9913101.43200910390.0214549.65解：(1)根據(jù)(gnj)以上(yshng)數(shù)據(jù)，應(yīng)用(yngyng)最小二乘法，得到基于以上數(shù)據(jù)的回歸方程（Y=CCP , X=YCPt）最小二乘法Dependent Variable: YMethod: Least SquaresDate: 11

4、/11/14 Time: 23:23Sample: 1980 2009Included observations: 30VariableCoefficientStd. Errort-StatisticProb.C541.176438.8279713.937800.0000X0.6886150.005886116.98470.0000R-squared0.997958Mean dependent var4379.806Adjusted R-squared0.997885S.D. dependent var2472.456S.E. of regression113.6985Akaike info

5、criterion12.36932Sum squared resid361965.8Schwarz criterion12.46273Log likelihood-183.5398Hannan-Quinn criter.12.39920F-statistic13685.41Durbin-Watson stat0.541618Prob(F-statistic)0.000000故回歸方程為：CCP=541.2+0.6886YCPtt （13.9） （117）R2=0.998 R2=0.998 S=114 DW=0.54取顯著性水平為0.05，k=2,，n=30，查DW表得到臨界值為L(zhǎng)=1.35 U

6、=1.49，在模型中DW=0.54，即DWL，表明存在序列正相關(guān)。由=1-DW2，得=0.729191先對(duì)數(shù)據(jù)進(jìn)行廣義差分變換，然后運(yùn)用最小二乘法得到以下估計(jì)模型：Dependent Variable: YSTARMethod: Least SquaresDate: 11/16/14 Time: 12:42Sample: 1980 2009Included observations: 30VariableCoefficientStd. Errort-StatisticProb.CSTAR541.943682.808026.5445790.0000XSTAR0.6836440.01106461.

7、791250.0000R-squared0.990869Mean dependent var1421.397Adjusted R-squared0.990543S.D. dependent var825.5422S.E. of regression80.28320Akaike info criterion11.67334Sum squared resid180471.0Schwarz criterion11.76675Log likelihood-173.1001Hannan-Quinn criter.11.70322Durbin-Watson stat2.166851得到(d do)模型(m

8、xng)為：Y*=541.9436+0.6836X* （6.54） （61.79）R2=0.9908 R2=0.9905 S=80.28 DW=2.166取顯著性水平(shupng)為0.05，k=2,，n=29，時(shí)L=1.34 U=1.48， UDW4-1.34，所以已不存在自相關(guān)性。(2)最小二乘法：CCP=541.2+0.6886YCPtt,廣義最小二乘法：Y*=594.6372+0.6812X*CO法：CCP=650.2+0.674YCPtt由上邊的計(jì)算結(jié)果可以，邊際消費(fèi)傾向在0.68，基本消費(fèi)在550元。3、設(shè)20個(gè)家庭的消費(fèi)與收入有如下數(shù)據(jù)，其中Y表示消費(fèi)，X表示收入20家庭的消費(fèi)

9、與收入Y消費(fèi)X收入19.922.331.232.331.836.612.112.140.742.36.16.238.644.725.526.110.310.338.840.288.133.134.533.53813.114.114.816.421.624.129.330.12528.317.918.219.820.1（1）設(shè)模型Yt=+Xt+ut （t=1，2，20）運(yùn)用最小二乘法，對(duì)上述模型進(jìn)行估計(jì)。以X軸為橫軸，殘差為縱軸，畫出殘差的散點(diǎn)圖，利用圖形討論異方差性的存在性，并對(duì)出現(xiàn)的結(jié)果進(jìn)行解釋。（2）對(duì)變量Xt，Yt取對(duì)數(shù)后，運(yùn)用最小二乘法，對(duì)模型進(jìn)行估計(jì)。以lnX為橫軸，殘差為縱軸，畫出

10、殘差的散點(diǎn)圖，討論異方差的存在性，并對(duì)出現(xiàn)的結(jié)果進(jìn)行解釋。解：有上述數(shù)據(jù)計(jì)算得：Dependent Variable: YMethod: Least SquaresDate: 11/15/14 Time: 14:37Sample: 1 20Included observations: 20VariableCoefficientStd. Errort-StatisticProb.C0.8470520.7033551.2043020.2441X0.8993250.02530935.533600.0000R-squared0.985944Mean dependent var23.55500Adjus

11、ted R-squared0.985164S.D. dependent var10.78691S.E. of regression1.313895Akaike info criterion3.478509Sum squared resid31.07377Schwarz criterion3.578082Log likelihood-32.78509Hannan-Quinn criter.3.497947F-statistic1262.637Durbin-Watson stat2.582686Prob(F-statistic)0.000000故回歸方程為：Yt=0.8471+0.8993Xt (

12、1.204) (35.5336)R=0.9859 R2=0.9853 S=1.31 DW=2.5826以X為橫軸，殘差為縱軸，得到(d do)殘差的散點(diǎn)圖：對(duì)圖形進(jìn)行(jnxng)分析，可以看到隨著收入(shur)X的變大，殘差的絕對(duì)值有變大的趨勢(shì)，故模型的誤差項(xiàng)之間存在異方差性。對(duì)變量X,Y取對(duì)數(shù)之后，運(yùn)用最小二乘法，得到以下估計(jì)模型：Dependent Variable: LNYMethod: Least SquaresDate: 11/15/14 Time: 16:29Sample: 1 20Included observations: 20VariableCoefficientStd.

13、 Errort-StatisticProb.C0.0756720.0573931.3184960.2039LNX0.9561860.01825552.380220.0000R-squared0.993482Mean dependent var3.033911Adjusted R-squared0.993120S.D. dependent var0.550836S.E. of regression0.045689Akaike info criterion-3.239278Sum squared resid0.037575Schwarz criterion-3.139705Log likelihood34.39278Hannan-Quinn criter.-3.219840F-statistic2743.688Durbin-Watson stat2.166013Prob(F-statistic)0.000000故回歸方程為：ln（Y)=0.0757+0.9562ln（X） (1.3184) (52.3802)R2=0.9935 R2=0.9931 S=0.0457 DW= 2.1660以ln（X）為橫軸，殘差為縱軸，得到