《第七章练习题及参考解答(第四版)计量经济学.doc》由会员分享，可在线阅读，更多相关《第七章练习题及参考解答(第四版)计量经济学.doc（11页珍藏版）》请在taowenge.com淘文阁网|工程机械CAD图纸|机械工程制图|CAD装配图下载|SolidWorks_CaTia_CAD_UG_PROE_设计图分享下载上搜索。
1、【精品文档】如有侵权，请联系网站删除，仅供学习与交流第七章练习题及参考解答(第四版)计量经济学.精品文档.第七章练习题及参考解答7.1 表7.4中给出了1981-2015年中国城镇居民人均年消费支出(PCE)和城镇居民人均可支配收入(PDI)数据。表7.4 1981-2015年中国城镇居民消费支出(PCE)和可支配收入(PDI)数据 （单位：元）年度城镇居民人均消费支出PCE城镇居民人均可支配收入PDI年度城镇居民人均消费支出PCE城镇居民人均可支配收入PDI1981456.80500.4019994615.915854.021982471.00535.3020004998.006280.00
2、1983505.90564.6020015309.016859.601984559.40652.1020026029.887702.801985673.20739.1020036510.948472.201986799.00900.9020047182.109421.601987884.401002.1020057942.8810493.0019881104.001180.2020068696.5511759.5019891211.001373.9320079997.4713785.8019901278.901510.20200811242.8515780.7619911453.801700.
3、60200912264.5517174.6519921671.702026.60201013471.4519109.4419932110.802577.40201115160.8921809.7819942851.303496.20201216674.3224564.7219953537.574283.00201318022.6426955.1019963919.474838.90201419968.0829381.0019974185.645160.30201521392.3631790.3119984331.615425.10估计下列模型：(1) 解释这两个回归模型的结果。(2) 短期和长
4、期边际消费倾向（MPC）是多少？分析该地区消费同收入的关系。(3) 建立适当的分布滞后模型，用库伊克变换转换为库伊克模型后进行估计，并对估计结果进行分析判断。【练习题7.1参考解答】(1) 解释这两个回归模型的结果。Dependent Variable: PCEMethod: Least SquaresDate: 03/10/18 Time: 09:12Sample: 1981 2005Included observations: 25VariableCoefficientStd. Errort-StatisticProb. C149.097524.567346.0689330.0000PDI
5、0.7575270.005085148.98400.0000R-squared0.998965 Mean dependent var2983.768Adjusted R-squared0.998920 S.D. dependent var2364.412S.E. of regression77.70773 Akaike info criterion11.62040Sum squared resid138885.3 Schwarz criterion11.71791Log likelihood-143.2551 F-statistic22196.24Durbin-Watson stat0.531
6、721 Prob(F-statistic)0.000000收入跟消费间有显著关系。收入每增加1元，消费增加0.76元。Dependent Variable: PCEMethod: Least SquaresDate: 03/10/18 Time: 09:13Sample(adjusted): 1982 2005Included observations: 24 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C147.688626.735795.5240010.0000PDI0.6791230.069
7、9599.7073850.0000PCE(-1)0.1110350.1001861.1082870.2803R-squared0.999012 Mean dependent var3089.059Adjusted R-squared0.998918 S.D. dependent var2354.635S.E. of regression77.44504 Akaike info criterion11.65348Sum squared resid125952.4 Schwarz criterion11.80074Log likelihood-136.8418 F-statistic10620.1
8、0Durbin-Watson stat0.688430 Prob(F-statistic)0.000000(2) 短期和长期边际消费倾向（MPC）是多少？分析该地区消费同收入的关系。短期MPC=0.68，长期MPC=0.679/(1-0.111)=0.764(3) 建立适当的分布滞后模型，用库伊克变换转换为库伊克模型后进行估计，并对估计结果进行分析判断。在滞后1-5期内，根据AIC最小，选择滞后5期，其回归结果如下:Dependent Variable: PCEMethod: Least SquaresDate: 03/10/18 Time: 09:25Sample(adjusted): 19
9、86 2005Included observations: 20 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C167.959033.277935.0471580.0002PDI0.7079330.1248785.6689810.0001PDI(-1)0.2252720.2742930.8212830.4263PDI(-2)-0.1789110.316743-0.5648470.5818PDI(-3)-0.0695250.328725-0.2114980.8358PDI(-4)0.2648740.
10、3004700.8815320.3940PDI(-5)-0.2269660.145557-1.5592920.1429R-squared0.999382 Mean dependent var3596.396Adjusted R-squared0.999096 S.D. dependent var2254.922S.E. of regression67.79561 Akaike info criterion11.54009Sum squared resid59751.18 Schwarz criterion11.88860Log likelihood-108.4009 F-statistic35
11、01.011Durbin-Watson stat1.471380 Prob(F-statistic)0.000000当期收入对消费有显著影响，但各滞后期影响并不显著。不显著可能是分布滞后模型直接估计时共线性造成的，也可能是真没显著影响。库伊克模型估计结果见上表，PCE(-1)部分回归结果t检验不显著。7.2 表7.5中给出了中国1980-2016年固定资产投资Y与社会消费品零售总额X的资料。取阿尔蒙多项式的次数m=2，运用阿尔蒙多项式变换法估计以下分布滞后模型：表7.5中国1980-2016年固定资产投资Y与社会零售总额X数据 （单位：亿元）年份固定资产投资Y社会消费品零售总额X年份固定资产投
12、资Y社会消费品零售总额X1980910.92140.0199929854.735647.91981961.02350.0200032917.739105.719821230.42570.0200137213.543055.419831430.12849.4200243499.948135.919841832.93376.4200355566.652516.319852543.24305.0200470477.459501.019863120.64950.0200588773.667176.619873791.75820.02006109998.276410.019884753.87440.02
13、007137323.989210.019894410.48101.42008172828.4114830.119904517.08300.12009224598.8132678.419915594.59415.62010251683.8156998.419928080.110993.72011311485.1183918.6199313072.314270.42012374694.7210307.0199417042.118622.92013446294.1237809.9199520019.323613.82014512020.7271896.1199622913.528360.220155
14、61999.8300930.8199724941.131252.92016606465.7332316.3199828406.233378.1【练习题7.2参考解答】直接估计结果如下：Dependent Variable: YMethod: Least SquaresDate: 03/10/18 Time: 09:32Sample(adjusted): 1984 2016Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C-23633.423701.8
15、25-6.3842600.0000X0.4619270.9181980.5030800.6190X(-1)2.0865661.6859581.2376140.2265X(-2)-0.5432541.708205-0.3180260.7529X(-3)1.1505771.8438080.6240220.5379X(-4)-1.3173211.283331-1.0264860.3138R-squared0.993755 Mean dependent var128264.7Adjusted R-squared0.992598 S.D. dependent var180131.0S.E. of reg
16、ression15497.23 Akaike info criterion22.29768Sum squared resid6.48E+09 Schwarz criterion22.56977Log likelihood-361.9117 F-statistic859.2660Durbin-Watson stat0.229807 Prob(F-statistic)0.000000使用阿尔蒙变换估计结果如下：Dependent Variable: YMethod: Least SquaresDate: 03/10/18 Time: 09:37Sample(adjusted): 1984 2016
17、Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C-23683.133619.054-6.5440100.0000Z00.8016780.6237781.2851980.2089Z10.4823171.3667070.3529050.7267Z2-0.2333220.358793-0.6502980.5206R-squared0.993572 Mean dependent var128264.7Adjusted R-squared0.992907 S
18、.D. dependent var180131.0S.E. of regression15170.17 Akaike info criterion22.20526Sum squared resid6.67E+09 Schwarz criterion22.38666Log likelihood-362.3868 F-statistic1494.254Durbin-Watson stat0.287072 Prob(F-statistic)0.000000根据可计算出0.802=1.051=0.833=0.149=-1.002直接使用软件结果：Dependent Variable: YMethod:
19、 Least SquaresDate: 03/10/18 Time: 09:39Sample(adjusted): 1984 2016Included observations: 33 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C-23683.133619.054-6.5440100.0000PDL010.8330240.7026451.1855550.2454PDL02-0.4509710.144976-3.1106620.0042PDL03-0.2333220.358793-0.650298
20、0.5206R-squared0.993572Mean dependent var128264.7Adjusted R-squared0.992907S.D. dependent var180131.0S.E. of regression15170.17Akaike info criterion22.20526Sum squared resid6.67E+09Schwarz criterion22.38666Log likelihood-362.3868F-statistic1494.254Durbin-Watson stat0.287072Prob(F-statistic)0.000000
21、Lag Distribution of XiCoefficientStd. ErrorT-Statistic . * |0 0.80168 0.62378 1.28520 . *|1 1.05067 0.42723 2.45927 . * |2 0.83302 0.70264 1.18555 .* |3 0.14873 0.31166 0.47722 * . |4-1.00221 0.92567-1.08269Sum of Lags 1.83190 0.18562 9.869017.3利用表7.5的数据，运用局部调整假定或自适应预期假定估计以下模型参数，并解释模型的经济意义，探测模型扰动项的一
22、阶自相关性：1）设定模型其中为预期最佳值。 2）设定模型其中为预期最佳值。3）设定模型其中为预期最佳值。【练习题7.3参考解答】1）设定模型 其中为预期最佳值。Dependent Variable: YMethod: Least SquaresDate: 03/10/18 Time: 10:09Sample(adjusted): 1981 2016Included observations: 36 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C-5669.5052498.919-2.2687830
23、.0299X0.6649820.1301835.1080430.0000Y(-1)0.7335440.0778119.4272690.0000R-squared0.997893 Mean dependent var117676.6Adjusted R-squared0.997765 S.D. dependent var175881.8S.E. of regression8314.081 Akaike info criterion20.96894Sum squared resid2.28E+09 Schwarz criterion21.10090Log likelihood-374.4410 F
24、-statistic7815.118Durbin-Watson stat0.925919 Prob(F-statistic)0.000000根据回归结果，可算出h统计量为3.64，明显大于2，表明5%显著水平下存在相关性。根据回归数据，可算出调整系数为1-0.734=0.266，这表示了局部调整的速度。0.665/0.266=2.5 2）设定模型 其中为预期最佳值。假设调整方程为：，则转化为一阶自回归模型后的回归结果为:Dependent Variable: LOG(Y)Method: Least SquaresDate: 03/10/18 Time: 10:11Sample(adjusted
25、): 1981 2016Included observations: 36 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C-0.5414920.692089-0.7824030.4396LOG(X)0.2996850.2623221.1424340.2615LOG(Y(-1)0.7649000.2006083.8129090.0006R-squared0.997423 Mean dependent var10.25491Adjusted R-squared0.997267 S.D. depende
26、nt var1.956096S.E. of regression0.102265 Akaike info criterion-1.642847Sum squared resid0.345117 Schwarz criterion-1.510887Log likelihood32.57124 F-statistic6386.241Durbin-Watson stat0.873321 Prob(F-statistic)0.000000根据回归结果，计算h统计量时开方部分为负，没法计算。故没法根据h统计量判断相关性。根据回归数据，可算出调整系数为1-0.765=0.235，这表示了局部调整的速度。0
27、.2997/0.235=1.2753）设定模型 其中为预期最佳值。Dependent Variable: YMethod: Least SquaresDate: 03/10/18 Time: 10:09Sample(adjusted): 1981 2016Included observations: 36 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C-5669.5052498.919-2.2687830.0299X0.6649820.1301835.1080430.0000Y(-1)0.7335
28、440.0778119.4272690.0000R-squared0.997893 Mean dependent var117676.6Adjusted R-squared0.997765 S.D. dependent var175881.8S.E. of regression8314.081 Akaike info criterion20.96894Sum squared resid2.28E+09 Schwarz criterion21.10090Log likelihood-374.4410 F-statistic7815.118Durbin-Watson stat0.925919 Pr
29、ob(F-statistic)0.000000可算出调节系数为1-0.734=0.266，这表示了预期修正的速度。0.665/0.266=2.57.4表7.6给出中国各年末货币流通量Y，社会商品零售额X1、城乡居民储蓄余额X 2的数据。表7.6中国年末货币流通量、社会商品零售额、城乡居民储蓄余额数据 （单位：亿元）年份年末货币流通量Y社会消费品零售总额X1城乡居民储蓄年底余额X219892344.08101.45184.5019902644.48300.17119.6019913177.89415.69244.9019924336.010993.711757.3019935864.714270
30、.415203.5019947288.618622.921518.8019957885.323613.829662.3019968802.028360.238520.80199710177.631252.946279.80199811204.233378.153407.47199913455.535647.959621.83200014652.739105.764332.38200115688.843055.473762.43200217278.048135.986910.65200319746.052516.3103617.65200421468.359501.0119555.3920052
31、4031.767176.6141050.99200627072.676410.0161587.30200730334.389210.0172534.19200834219.0114830.1217885.35200938246.0132678.4260771.66201044628.2156998.4303302.49201150748.5183918.6343635.89201254659.8210306.9399551.00201358574.4237809.9447601.57201460259.5271896.1485261.34利用表中数据设定模型：其中，为长期(或所需求的)货币流通
32、量。试根据局部调整假设，作模型变换，估计并检验参数，对参数经济意义做出解释。【练习题7.4参考解答】利用表中数据设定模型：其中，为长期(或所需求的)货币流通量。试根据局部调整假设，作模型变换，估计并检验参数，对参数经济意义做出解释。假设局部调整方程为：，对，可转化为回归方程：，其回归结果如下：Dependent Variable: YMethod: Least SquaresDate: 03/10/18 Time: 10:03Sample(adjusted): 1990 2014Included observations: 25 after adjusting endpointsVariabl
33、eCoefficientStd. Errort-StatisticProb. C1618.034732.14892.2099790.0383Y(-1)0.9810200.1493126.5702800.0000X1-0.1304290.041464-3.1455900.0049X20.0783990.0337062.3259720.0301R-squared0.997519 Mean dependent var23457.75Adjusted R-squared0.997164 S.D. dependent var18266.54S.E. of regression972.7612 Akaik
34、e info criterion16.74380Sum squared resid19871553 Schwarz criterion16.93882Log likelihood-205.2975 F-statistic2813.916Durbin-Watson stat1.112498 Prob(F-statistic)0.000000各回归系数在5%显著水平下均显著。可算出调整系数为1-0.981=0.019，这表示了局部调整的速度。假设局部调整方程为：，对，可转化为回归方程：，其回归结果如下：Dependent Variable: LOG(Y)Method: Least SquaresD
35、ate: 03/10/18 Time: 10:04Sample(adjusted): 1990 2014Included observations: 25 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb. C0.6577880.2771622.3732960.0273LOG(Y(-1)0.7419100.2306023.2172700.0041LOG(X1)0.0533500.1027270.5193320.6090LOG(X2)0.1211540.1785370.6785930.5048R-squar
36、ed0.996730 Mean dependent var9.716778Adjusted R-squared0.996263 S.D. dependent var0.913771S.E. of regression0.055860 Akaike info criterion-2.786285Sum squared resid0.065527 Schwarz criterion-2.591265Log likelihood38.82856 F-statistic2133.726Durbin-Watson stat1.076480 Prob(F-statistic)0.0000007.5 根据四
37、川省19782014年的消费总额Y（亿元）和收入总额X（亿元）的年度资料，估计出库伊克模型如下：试回答下列问题：1）分布滞后系数的衰减率是多少？2）模型中是否存在多重共线性问题？请说明判断的理由。3）收入对消费的即期和长期影响乘数是多少？4）某同学查表发现，在显著性水平下，DW检验临界值为，。请问该同学试图得出什么结论？你认为该同学的做法是否存在问题？请帮该同学完成后续工作。【练习题7.5参考解答】1）分布滞后系数的衰减率为0.822）模型中各斜率系数均显著，没有明显的多重共线性问题。3）收入对消费的即期和长期影响乘数分别是：即期乘数为0.28； 长期乘数为0.28/(1-0.82)=1.56
38、4）该同学试图检验是否存在自相关性问题，但是此模型为自回归模型，模型中有滞后被解释变量，此时不能使用DW检验法。而可以用德宾h检验，可计算出其h统计量为：式中：d=1.45；n=37；h=1.82，小于，表明5%显著水平下不存在自相关性问题。7.6利用某地区19802014年固定资产投资（Y）与地区生产总值GDP（X）的数据资料（单位：亿元），使用OLS法估计出如下模型：（1）上述模型是否存在自相关性问题？（2）如果将上述模型看成是局部调整模型的估计结果，试计算调节系数。【练习题7.6参考解答】(1) 式中：d=1.5321；n=35；h=1.8038，小于，表明5%显著水平下不存在自相关性问题。(2) 如果将模型看成是局部调整模型的估计结果, 则调节系数。7.7联系自己所学的专业选择一个实际问题，设定一个分布滞后模型或自回归模型，并自己去收集样本数据，用本章的方法估计和检验这个模型，你如何评价自己所做的这项研究?【练习题7.7参考解答】本题无参考解答