《EViews软件做的计量经济学实验关于能源消费的论文.doc》由会员分享，可在线阅读，更多相关《EViews软件做的计量经济学实验关于能源消费的论文.doc（12页珍藏版）》请在三一文库上搜索。
1、我国能源消费影响因素计量分析一、 问题的提出研究背景能源是经济增长的战略投入要素，在经济增长初期，能源的投入能够带动经济快速增长。18世纪第一次工业革命，煤炭的燃烧推动蒸汽机的普及，进而带动了生产率的提高，实现了工业化的起步。随着工业化进程的深入，石油的大量使用成为经济持续增长的推动力量。可见，经济增长和能源投入之间形成了一定的互动关系，能源是经济增长的动力源泉，经济增长又拉动能源消费。研究目的我国国民经济在向工业化和现代化发展的进程中，较长时间处于能源消费需求迅速增长而供给不组的紧缺状态，20世纪末的“九五”期间发生了显著变化，能源生产和消费总量均呈下降的趋势，出现了难得的能源供需基本基本平
2、衡状况，但同时也出现了新的问题，即煤炭供过于求与石油的供不应求的结构性矛盾突出。本文拟从我国的能源消费和生产入手，分析影响我国能源消费与生产的主要因素，探讨我国能源消费的趋势。1.3研究的相关理论支持及研究状况刘凤朝等于2007年9月发表了“中国经济增长和能源消费的动态特征”一文。文章运用基于向量自回归模型的广义预测误差方差分解和广义脉冲响应分析方法，在资本，劳动和能源三要素单部门新古典生产函数的框架内，以中国19882005年间的能源消费和经济增长数据为样本，考察了二者之件的动态特征。结果显示：在长期，除了资本增长外，经济增长是能源消费的重要增长因素，贡献度为14.92%。能源消费增长的冲击
3、对经济增长有正的影响作用。刘凤朝、孙玉涛于2008年3月在中国人口.资源与环境上发表了“技术创新，产业结构调整对能源消费影响的实证分析”。指出，在产业结构调整，减少能源效率的过程中，技术创新是关键因素。在现有的研究基础上引入技术创新要素，建立技术创新，产业结构调整对能源消费影响的分析框架。通过假设建立了技术创新，产业结构调整对能源消费的计量经济模型，运用中国的数据进行了实证分析。研究结果表明，专利授权量增加能够节约能源消费，产业产值增加能够减少能源消费。研究结果认为，产业结构升级，优化和经济增长方式转变，是经济增长和能源消费脱钩的重要途径。再找点能源产出的理论支持 而对于产业结构的影响因素，钱
4、纳里和赛尔奎因在发展型式，1950-1970一书中，设计了一个国民生产总值的市场占有率模型，在模型中，钱纳里和赛尔奎因以人均国民生产总值和人口数量作为外生变量，用回归方程对样本国家的数据进行计算，得到产业结构演进的“标准结构”,Xi=ln0+1lnY+2(lnY)2+3lnN其中Xi是第i产业的粗附加价值的市场占有率，Y是人均国内生产总值，N是样本国家的人口数量。二、 模型设定理论上认为影响能源消费需求总量的因素主要有经济发展水平、产业发展、能源生产总量、人口总数等，三、 数据的收集年份能源消费标准煤总量Y/万吨国内生产总值X2/亿元工业增加值X3/亿元建筑业增加值X4/亿元交通运输邮电业增加
5、值X5/亿元人均电力消费X6/千瓦时能源加工转换效率X7/%19857668290163448.7417.9406.921.368.2919868085010275.23967525.7475.623.268.3219878663212058.64585.8665.8544.926.467.4819889299715042.85777.281066131.266.5419899693416092.3648479478635.366.5119909870318667.86858859.41147.542.467.2199110378321781.58087.11015.11409.746.965
6、.9199210917026923.510284.514151681.854.666199311599335333.9141882266.52205.661.267.32199412273748197.919480.72964.72898.372.765.2199513117660793.724950.63728.83424.183.571.05199613894871176.629447.64387.44068.593.171.519971377987897332921.44621.64593101.869.23199813221484402.334018.44985.85278.4106.
7、669.44199913383189677.135861.55172.15821.8118.269.19200013855399214.64003.65522.37333.4132.469.042001143199109655.243580.65931.78406.1144.669.032002151797120332.747431.36465.593930.4156.369.042003174990135822.854945.57490.810098.4173.769.42004203227159878.3652108694.312147.6190.270.71200522331918308
8、4.876912.910133.810526.1216.771.082006246270211923.591310.911851.112481.1249.471.242007265583249529.9107367.214014.114604.1274.971.25四、 模型的估计与调整（一）参数估计1、双击“Eviews”，进入主页。输入数据：点击主菜单中的File/Open /EV WorkfileExcel多重共线性的数据.xls ；2、在EV主页界面的窗口，输入“ls y c x2 x3 x4 x5 x6 x7”，按“Enter”.出现OLS回归结果,图2：Dependent Vari
9、able: YMethod: Least SquaresDate: 11/01/10 Time: 11:34Sample: 1985 2007Included observations: 23VariableCoefficientStd. Errort-StatisticProb.C168326.2108641.01.5493810.1408X2-0.1422900.763550-0.1863530.8545X30.5031080.2485522.0241570.0600X48.29423710.431120.7951430.4382X5-0.2030370.111019-1.8288410.
10、0861X6233.9125388.51880.6020620.5556X7-1373.3761588.868-0.8643730.4002R-squared0.980436Mean dependent var139364.6Adjusted R-squared0.973099S.D. dependent var51705.05S.E. of regression8480.388Akaike info criterion21.17469Sum squared resid1.15E+09Schwarz criterion21.52028Log likelihood-236.5089F-stati
11、stic133.6365Durbin-Watson stat1.380303Prob(F-statistic)0.000000由此可见，该模型的可决系数为0.995，修正的可决系数为0.993，模型拟和很好，F统计量为701.47，模型拟和很好，回归方程整体上显著。但是当=0.05时，=2.069，不仅X4、X5、X6、X7的系数t检验不显著，而且X2、X4、X6系数的符号与预期相反，这表明很可能存在严重的多重共线性。（即除了农业增加值、工业增加值外，其他因素对财政收入的影响都不显著，且农业增加值、建筑业增加值、最终消费的回归系数还是负数，这说明很可能存在严重的多重共线性。）（二）多重共线性的
12、诊断与修正3、计算各解释变量的相关系数：在Workfile窗口，选择X2、X3、X4、X5、X6、X7数据，点击“Quick”Group StatisticsCorrelationsOK,出现相关系数矩阵，如图3：图3： 相关系数矩阵X2X3X4X5X6X7X210.9657320838866270.9986857898553320.345345026254410.9972578885234110.743816435340805X30.96573208388662710.9655924019815920.3239441524405460.9565951672026850.71949537138
13、2113X40.9986857898553320.96559240198159210.3298546529731210.9948853467152340.755788639731427X50.345345026254410.3239441524405460.32985465297312110.3663218904310660.205555717546146X60.9972578885234110.9565951672026850.9948853467152340.36632189043106610.72634236114161X70.7438164353408050.7194953713821
14、130.7557886397314270.2055557175461460.726342361141611由相关系数矩阵可以看出，各解释变量相互之间的相关系数较高，特别是农业增加值、工业增加值、建筑业增加值、最终消费之间，相关系数都在0.8以上。这表明模型存在着多重共线性。1、采用逐步回归法，去检验和解决多重共线性问题。分别作Y对X2、X3、X4、X5、X6、X7的一元回归，结果如下图4：在EV主页界面的窗口，输入“ls y c x2”，“回车键”。依次如上推出X3、X4、X5、X6、X7的一元回归。综上所述，结果如下图4：图4.一元回归估计结果变量参数估计值0.7348351.6654811
15、3.190880.737886678.005819332.30t统计量25.3151718.0256525.963631.29452922.422944.7024270.9682710.9392930.9697890.0739030.9599070.5129060.9667600.9364020.9683500.0298030.9579980.4897112、其中，加入 的最大，以 为基础，顺次加入其他变量逐步回归。结果如下图5：图5.加入新变量的回归结果（一）变量X2X3X4X5X6X7X2,X30.5326640.4817370.970921（5.092034）（2.001247）X2,X
16、40.14876510.526020.966883（0.263146）（1.038045）X2,X50.754738-0.209480.970869（26.06504）（-1.99037）X2,X60.947069-197.21250.965588（2.373029）（-0.533247）X2,X70.754414-951.4780.965709（17.10268）（-0.596733）（三）异方差的诊断与修正该模型样本回归估计式的书写形式为：Y = 11.44213599 + 0.6267829962*X (3.629253) (0.019872) t= 3.152752 31.54097
17、S.E.=9.158900 DW=1.597946 F=994.8326（一）图形法1、在“Workfile”页面：选中x,y序列，点击鼠标右键，点击Openas GroupYes2、在“Group”页面：点击ViewGraphScatterSimple Scatter, 得到X,Y的散点图（图3所示）：2、Goldfeld-Quandt法进行检验。a.将样本X按递增顺序排序，去掉中间1/4的样本，再分为两个部分的样本，即n1=n2=9。Dependent Variable: YMethod: Least SquaresDate: 11/01/10 Time: 11:07Sample: 1 9
18、Included observations: 9VariableCoefficientStd. Errort-StatisticProb.C-15.152720.772901-19.605000.0000X0.0002108.01E-0626.285140.0000R-squared0.989970Mean dependent var5.000000Adjusted R-squared0.988537S.D. dependent var2.738613S.E. of regression0.293208Akaike info criterion0.577264Sum squared resid
19、0.601798Schwarz criterion0.621092Log likelihood-0.597687F-statistic690.9084Durbin-Watson stat1.352108Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 11/01/10 Time: 11:08Sample: 1 9Included observations: 9VariableCoefficientStd. Errort-StatisticProb.C8.8510300.9919968.9224450
20、.0000X5.42E-055.14E-0610.537710.0000R-squared0.940700Mean dependent var19.00000Adjusted R-squared0.932228S.D. dependent var2.738613S.E. of regression0.712943Akaike info criterion2.354299Sum squared resid3.558014Schwarz criterion2.398127Log likelihood-8.594347F-statistic111.0434Durbin-Watson stat0.63
21、2734Prob(F-statistic)0.000015b.分别对两个部分的样本求最小二乘估计，得到两个部分的残差平方和，即= 0.601798 ， = 3.558014求F统计量为 F= = 5.912306给定，查F分布表，得临界值为 = 3.79c.比较临界值与F统计量值，有F=5.912306 =3.79 ，说明该模型的随机误差项存在异方差。修正异方差在运用加权最小二乘法估计过程中，分别选用了权数=1/，=1/，=1/。1、在“Workfile”页面：点击“Generate”，输入“w1=1/x”OK ；同样的输入“w2=1/x2”“w3=1/sqr(x)”；2、在“Equation
22、”页面：点击“Estimate Equation”,输入“y c x”,点击“weighted”,输入“w1”,出现如图6：Dependent Variable: YMethod: Least SquaresDate: 11/01/10 Time: 12:31Sample: 1985 2007Included observations: 23Weighting series: W1VariableCoefficientStd. Errort-StatisticProb.C75342.481955.93038.520020.0000X20.8614960.0873089.8673060.0000
23、Weighted StatisticsR-squared0.986045Mean dependent var102600.2Adjusted R-squared0.985380S.D. dependent var77372.86S.E. of regression9355.386Akaike info criterion21.20823Sum squared resid1.84E+09Schwarz criterion21.30697Log likelihood-241.8947F-statistic97.36373Durbin-Watson stat0.269103Prob(F-statis
24、tic)0.000000Unweighted StatisticsR-squared0.925702Mean dependent var139364.6Adjusted R-squared0.922164S.D. dependent var51705.05S.E. of regression14425.26Sum squared resid4.37E+09Durbin-Watson stat0.141803Dependent Variable: YMethod: Least SquaresDate: 11/01/10 Time: 12:33Sample: 1985 2007Included o
25、bservations: 23Weighting series: W2VariableCoefficientStd. Errort-StatisticProb.C61583.222022.13930.454490.0000X21.8677210.17376210.748730.0000Weighted StatisticsR-squared0.998194Mean dependent var89007.62Adjusted R-squared0.998108S.D. dependent var134618.1S.E. of regression5855.843Akaike info crite
26、rion20.27121Sum squared resid7.20E+08Schwarz criterion20.36995Log likelihood-231.1189F-statistic115.5353Durbin-Watson stat0.389451Prob(F-statistic)0.000000Unweighted StatisticsR-squared-3.468653Mean dependent var139364.6Adjusted R-squared-3.681446S.D. dependent var51705.05S.E. of regression111872.4S
27、um squared resid2.63E+11Durbin-Watson stat0.023839Dependent Variable: YMethod: Least SquaresDate: 11/01/10 Time: 12:34Sample: 1985 2007Included observations: 23Weighting series: W3VariableCoefficientStd. Errort-StatisticProb.C79134.392101.45237.657000.0000X20.7416510.04145717.889810.0000Weighted Sta
28、tisticsR-squared0.892994Mean dependent var117941.4Adjusted R-squared0.887898S.D. dependent var26545.81S.E. of regression8887.964Akaike info criterion21.10572Sum squared resid1.66E+09Schwarz criterion21.20446Log likelihood-240.7158F-statistic320.0453Durbin-Watson stat0.251182Prob(F-statistic)0.000000
29、Unweighted StatisticsR-squared0.968188Mean dependent var139364.6Adjusted R-squared0.966673S.D. dependent var51705.05S.E. of regression9439.115Sum squared resid1.87E+09Durbin-Watson stat0.303249用权数的估计结果为： = 75342.48 + 0.861496 （38.52002） （9.867306）= 0.986045 DW= 0.269103 F=97.36373括号中的数据为t统计量值。由上可以看出
30、，运用加权最小二乘法消除了异方差后，参数的t检验显著，可决系数提高了不少，F检验也显著，并说明销售收入每增长1元，销售利润平均增长0.861496元。这说明在其他因素不变的情况下，当国民收入每上升1%时，能源消费就平均增加0.23585%。Dependent Variable: YMethod: Least SquaresDate: 11/01/10 Time: 12:31Sample: 1985 2007Included observations: 23Weighting series: W1VariableCoefficientStd. Errort-StatisticProb.C7534
31、2.481955.93038.520020.0000X20.8614960.0873089.8673060.0000Weighted StatisticsR-squared0.986045Mean dependent var102600.2Adjusted R-squared0.985380S.D. dependent var77372.86S.E. of regression9355.386Akaike info criterion21.20823Sum squared resid1.84E+09Schwarz criterion21.30697Log likelihood-241.8947
32、F-statistic97.36373Durbin-Watson stat0.269103Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.925702Mean dependent var139364.6Adjusted R-squared0.922164S.D. dependent var51705.05S.E. of regression14425.26Sum squared resid4.37E+09Durbin-Watson stat0.141803Dependent Variable: YMethod: Least Sq
33、uaresDate: 11/01/10 Time: 12:52Sample: 1985 2007Included observations: 23VariableCoefficientStd. Errort-StatisticProb.C79687.883069.32225.962700.0000X20.7348350.02902725.315170.0000R-squared0.968271Mean dependent var139364.6Adjusted R-squared0.966760S.D. dependent var51705.05S.E. of regression9426.7
34、50Akaike info criterion21.22343Sum squared resid1.87E+09Schwarz criterion21.32217Log likelihood-242.0695F-statistic640.8577Durbin-Watson stat0.304578Prob(F-statistic)0.000000N=23 , k=1 , dL=1.257 dU=1.437 DW=0.269103 正相关Dependent Variable: ET1Method: Least SquaresDate: 11/01/10 Time: 12:57Sample (ad
35、justed): 1986 2007Included observations: 22 after adjustmentsVariableCoefficientStd. Errort-StatisticProb.C1162.3071751.6980.6635320.5150X2-0.0074520.016204-0.4598630.6508ET1(-1)0.8247850.1188416.9402200.0000R-squared0.717966Mean dependent var437.7797Adjusted R-squared0.688279S.D. dependent var9178.
36、547S.E. of regression5124.568Akaike info criterion20.04760Sum squared resid4.99E+08Schwarz criterion20.19638Log likelihood-217.5236F-statistic24.18392Durbin-Watson stat0.782175Prob(F-statistic)0.0000060.824785ls et1 c x2 et1(-1) 回归方程et1=0.824785et1(-1)Y0.824785Yt-1=(10.824785)p1（Xt0.824785Xt-1）p2+vt
37、ls lny-0.824785*lny(-1) c lnx2-0.824785*lnx2(-1)Dependent Variable: Y-0.824785*Y(-1)Method: Least SquaresDate: 11/01/10 Time: 13:03Sample (adjusted): 1986 2007Included observations: 22 after adjustmentsVariableCoefficientStd. Errort-StatisticProb.C14784.991699.5798.6992080.0000X2-0.824785*X2(-1)0.7226700.05544913.033050.0000R-squared0.894659Mean dependent var31999.94Adjusted R-squared0.889392S.D. dependent var15083.71S.E. of r