影响猪肉价格的因素分析

1、. 影响猪肉价格的因素分析1研究问题的背景在当前通货膨胀日益严重的情况下，各种物价飞涨，给人们的日常生活带来了极大地影响，生活中一些必需的物质也在涨，蔬菜肉的价格节节攀升，因此我觉得有必要研究一下究竟是什么因素在影响着这些必需品的价格，从而控制这些因素的上涨，使必需品的价格维持在一个比较稳定的水平上。2研究的主要内容这里我选取了对人们生活影响较大的猪肉的价格，从城镇居民收入，猪的供给量，饲料价格，替代品鸡蛋的价格，猪肉供给量五个方面来研究，看看它们相不相关，是正相关还是负相关，有多大的影响程度，从而调节这些变量使猪肉的价格比较稳定，对人们的生活产生较小的影响。关键词：猪肉价格 3选取数据 年份

2、猪肉价格城镇居民收入饲料价格鸡蛋价格猪肉供给量199610.54838.91.57.83158199712.25160.31.566.23596.3199810.15425.11.495.53883.719997.558541.25.24005.6200010.10262801.47255.093966200110.656859.61.3945.34051.7200210.237702.81.5225.394123.1200310.748472.21.65.254238.6200413.769421.61.696.394341200513.19104931.85256.574555.32006

3、12.1311759.51.8686.224650.5200718.8113785.52.137.764287.8200823.4915780.82.627.844620.54建立模型将以上数据导入eviews,就可以建立以下equation其中y代表猪肉价格，x1表示城镇居民收入，x2代表饲料价格，x3代表鸡蛋价格，x4表示猪肉的供给量.Dependent Variable: YMethod: Least SquaresDate: 12/18/10 Time: 14:25Sample: 1996 2008Included observations: 13VariableCoefficient

4、Std. Errort-StatisticProb.  C9.15743513.677330.6695340.5220X10.0004980.0005410.9196120.3847X210.258233.1922473.2134820.0124X3-0.4803610.861324-0.5577010.5923X4-0.0036890.002947-1.2515900.2461R-squared0.946997    Mean dependent var12.56938Adjusted R-squared0.920496 

5、   S.D. dependent var4.232883S.E. of regression1.193522    Akaike info criterion3.475417Sum squared resid11.39596    Schwarz criterion3.692706Log likelihood-17.59021    F-statistic35.73401Durbin-Watson stat2.510756  

6、;  Prob(F-statistic)0.000038表中除x2外，概率均大于0.05,说明其对y的影响不显著，必须对其进行修正，使其对y的影响显著。经修正的结果如下：Dependent Variable: YMethod: Least SquaresDate: 12/18/10 Time: 13:42Sample (adjusted): 1997 2008Included observations: 12 after adjustmentsVariableCoefficientStd. Errort-StatisticProb.  C26.35190

7、5.2969214.9749460.0016D(X1)0.0039440.0016512.3881830.0483LOG(X2)14.449493.6619583.9458360.0056D(X3)1.4456990.5840702.4752130.0425X4(-1)-0.0060370.001353-4.4634710.0029R-squared0.973597    Mean dependent var12.74183Adjusted R-squared0.958510    S.D. dependent v

8、ar4.373145S.E. of regression0.890771    Akaike info criterion2.900878Sum squared resid5.554310    Schwarz criterion3.102922Log likelihood-12.40527    F-statistic64.53091Durbin-Watson stat1.689968    Prob(F-statistic)0.00

9、0013从表中可以看出，t检验的概率均小于0.05,此时的变量对y的影响是显著的，此模型才是可以用的。还可以看出F检验的值也较大，所以拒绝原假设，总体的显著性成立。可以得到下列模型：Estimation Command:=LS Y C D(X1) LOG(X2) D(X3) X4(-1)Estimation Equation:=Y = C(1) + C(2)*D(X1) + C(3)*LOG(X2) + C(4)*D(X3) + C(5)*X4(-1)Substituted Coefficients:=Y = 26.35189582 + 0.003943975368*D(X1) + 14.44

10、948848*LOG(X2) + 1.44569861*D(X3) - 0.006037331562*X4(-1)5异方差检验（怀特检验）原假设HO:残差项不存在异方差备择假设H1：残差项存在异方差White Heteroskedasticity Test:F-statistic0.975116    Probability0.568784Obs*R-squared8.666952    Probability0.371165Test Equation:Dependent Variable: RESID2Metho

11、d: Least SquaresDate: 12/18/10 Time: 14:44Sample: 1997 2008Included observations: 12VariableCoefficientStd. Errort-StatisticProb.  C9.59299477.996970.1229920.9099D(X1)0.0023160.0021471.0788210.3597(D(X1)22.08E-078.06E-070.2584460.8128LOG(X2)1.6597335.6856470.2919160.7894(LOG(X2)2-4.5993656

12、.490699-0.7086090.5297D(X3)0.2440161.1780230.2071400.8492(D(X3)2-0.6373930.997492-0.6389950.5683X4(-1)-0.0023800.037570-0.0633530.9535X4(-1)2-5.31E-084.56E-06-0.0116440.9914R-squared0.722246    Mean dependent var0.462859Adjusted R-squared-0.018431    S.D. depe

13、ndent var0.471484S.E. of regression0.475809    Akaike info criterion1.466105Sum squared resid0.679182    Schwarz criterion1.829785Log likelihood0.203370    F-statistic0.975116Durbin-Watson stat2.734176    Prob(F-statisti

14、c)0.568784从表中可以看出怀特检验的概率均大于0.05,所以接受原假设，说明残差项不存在异方差。6自相关检验（LM检验）Breusch-Godfrey Serial Correlation LM Test:F-statistic0.263214    Probability0.778599Obs*R-squared1.143077    Probability0.564656Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 1

15、2/18/10 Time: 14:47Presample missing value lagged residuals set to zero.VariableCoefficientStd. Errort-StatisticProb.  C-0.4681376.037956-0.0775320.9412D(X1)0.0002890.0019010.1522180.8850LOG(X2)-0.7251874.286532-0.1691780.8723D(X3)-0.2234790.730846-0.3057810.7721X4(-1)0.0001390.0015390.090

16、6470.9313RESID(-1)0.1161500.5420630.2142740.8388RESID(-2)-0.4161950.575909-0.7226750.5023R-squared0.095256    Mean dependent var-2.09E-15Adjusted R-squared-0.990436    S.D. dependent var0.710589S.E. of regression1.002519    Akaike info crit

17、erion3.134108Sum squared resid5.025226    Schwarz criterion3.416970Log likelihood-11.80465    F-statistic0.087738Durbin-Watson stat1.851760    Prob(F-statistic)0.994924从表中可以看出，检验之后的概率均大于0.05,接受原假设，说明残差之间不存在二阶自相关，通过了LM检验。7正态分布检验从表中可以看出JB统计量的

18、概率为0.725370,说明残差有百分之72.5370的概率是正态分布，大于0.05,通过了正态分布检验。8白噪声检验Date: 12/18/10 Time: 15:00Sample: 1997 2008Included observations: 12AutocorrelationPartial CorrelationAC  PAC Q-Stat Prob    . |* . |    . |* . |10.1450.1450.32050.571  &#

19、160; . *| . |    . *| . |2-0.213-0.2391.08060.583    . *| . |    . *| . |3-0.243-0.1842.18230.535    . | . |    . | . |4-0.040-0.0272.21610.696    . |* . |   &#

20、160;. | . |50.1200.0442.56210.767    . | . |    . *| . |60.000-0.0892.56210.861    . | . |    . | . |70.0000.0342.56210.922    . | . |    . | . |80.0000.0122.56210.959  &

21、#160; . | . |    . | . |90.000-0.0102.56210.979    . | . |    . | . |100.0000.0002.56210.990从图中可以看出，其自相关系数和偏自相关系数均落在二倍的标注差以内，说明其波动性较小，且在几阶之后趋近于0，说明从长期来看，其是不相干的，属于白噪声。9伪回归检验Null Hypothesis: E1 has a unit rootExogenous: NoneLag Length: 0

22、 (Automatic based on SIC, MAXLAG=2)t-Statistic  Prob.*Augmented Dickey-Fuller test statistic-2.756569 0.0107Test critical values:1% level-2.7921545% level-1.97773810% level-1.602074*MacKinnon (1996) one-sided p-values.Warning: Probabilities and critical values calculated for 20 &

23、#160;      observations and may not be accurate for a sample size of 11Augmented Dickey-Fuller Test EquationDependent Variable: D(E1)Method: Least SquaresDate: 12/18/10 Time: 15:11Sample (adjusted): 1998 2008Included observations: 11 after adjustmentsVariableCoefficient

24、Std. Errort-StatisticProb.  E1(-1)-0.9714940.352429-2.7565690.0202R-squared0.417541    Mean dependent var0.144408Adjusted R-squared0.417541    S.D. dependent var0.956933S.E. of regression0.730322    Akaike info criterion2.295845Su

25、m squared resid5.333700    Schwarz criterion2.332017Log likelihood-11.62715    Durbin-Watson stat1.779706从表中可以看出，其概率为0.0107小于0.05,所以不存在伪回归，通过了检验。9模型平稳性和预测性检验 从图中可以看出，模型的稳定性一直很好，始终在红线的范围内。从图中可以看出，模型的预测能力较强，稳定性也较强，符合我们所需要的模型。从图中可以看出，该模型的一步预测能力较好，因为蓝线一直在红线内，处在预测能

26、力之内。从图中可以看出，其N步预测能力较好，蓝线一直处在红线之内。图中的红线代表预测能力，蓝线处在两条红线之内则代表稳定性较强，在预测期内结构未发生改变，说明该模型的预测能力和稳定性较好。10参数约束检验（1）约束条件：c(1)=0Wald Test:Equation: EQUATION2Test StatisticValue  df    ProbabilityF-statistic24.75008(1, 7)  0.0016Chi-square24.750081  0.0000Null Hy

27、pothesis Summary:Normalized Restriction (= 0)Value  Std. Err.C(1)26.351905.296921Restrictions are linear in coefficients.从表中可以看出，其概率小于0.05，所以拒绝原假设，说明参数c(1)=0不成立。（2）约束条件：c(2)=0Wald Test:Equation: EQUATION2Test StatisticValue  df    ProbabilityF-statistic5.70342

28、0(1, 7)  0.0483Chi-square5.7034201  0.0169Null Hypothesis Summary:Normalized Restriction (= 0)Value  Std. Err.C(2)0.0039440.001651Restrictions are linear in coefficients.从表中可以看出，其概率小于0.05，所以拒绝原假设，说明参数c(2)=0不成立。（3）约束条件：c(3)=0Wald Test:Equation: EQUATION2Test StatisticVal

29、ue  df    Probability0.F-statistic15.56962(1, 7)  0.0056Chi-square15.569621  0.0001Null Hypothesis Summary:Normalized Restriction (= 0)Value  Std. Err.C(3)14.449493.661958Restrictions are linear in coefficients.从表中可以看出，其概率小于0.05，所以拒绝原假设，说明参

30、数c(3)=0不成立。（4）约束条件：c(4)=0Wald Test:Equation: EQUATION2Test StatisticValue  df    ProbabilityF-statistic6.126678(1, 7)  0.0425Chi-square6.1266781  0.0133Null Hypothesis Summary:Normalized Restriction (= 0)Value  Std. Err.C(4)1.4456990.584070

31、Restrictions are linear in coefficients.从表中可以看出，其概率小于0.05,所以拒绝原假设，说明参数c(4)=0不成立。（5）约束条件：c(5)=0Wald Test:Equation: EQUATION2Test StatisticValue  df    ProbabilityF-statistic19.92258(1, 7)  0.0029Chi-square19.922581  0.0000Null Hypothesis Summary:Normalized Restriction (= 0)Value  Std. Err.C(5)-0.0060370.001353Restrictions are linear in coefficients.从表中可以看出，其概率小于0.05,所以拒绝原假设，说明参数c(5)=0不成立。11残差图检验从图中可以看出，回归方程拟合的较好，残差的波动性不大，模型较稳定，预测能力比较强。12预测图检验从图中可以看出，蓝线一直处在两条红线之内，说明其预测的水平较好，一直处在预测能力之内。13经济意义