您的当前位置：首页计量经济学作业

计量经济学作业

来源：乌哈旅游

20100561 统计一班黄艳萍课本P306

(1)对LX与LM序列进行单位根检验，检验它们的平稳性。

Null Hypothesis: D(LX) has a unit root Exogenous: Constant, Linear Trend

Lag Length: 0 (Automatic based on SIC, MAXLAG=7)

t-Statistic -9.686205 -4.323979 -3.580623 -3.225334 t-Statistic -9.686205 1.703469

Prob.* 0.0000 Prob. 0.0000 0.1009 0.3152 -0.003879 0.490453 0.031226 0.173962 0.074862 1.798245

H0:0

Augmented Dickey-Fuller test statistic Test critical values:

1% level 5% level

10% level *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(LX,2) Method: Least Squares

Date: 10/23/12 Time: 17:49 Sample (adjusted): 1980 2007

Included observations: 28 after adjustments Variable D(LX(-1)) C

@TREND(1978) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic

Prob(F-statistic)

Coefficient Std. Error -1.567446 0.166497 0.005618 0.789804 0.772989 0.233680 1.365154 2.562838 46.96842 0.000000

0.161823 0.097740

0.005481 1.024938 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat

由于-9.6862明显小于给定的显著性水平的临界值，所以拒绝零假设

，即0，

不存在单位根1，原LX序列并非平稳，但一阶差分后是平稳序列。

Null Hypothesis: D(LM) has a unit root Exogenous: Constant, Linear Trend

Lag Length: 1 (Automatic based on SIC, MAXLAG=7) t-Statistic Augmented Dickey-Fuller test statistic Test critical values:

1% level 5% level 10% level

-5.317055 -4.339330 -3.587527 -3.229230 t-Statistic -5.317055 2.435392

Prob.* 0.0010 Prob. 0.0000 0.0230 0.0898 0.1618 -0.002081 0.170029 -1.268497 -1.076521 -1.211412 2.083194

H0:0 *MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation Dependent Variable: D(LM,2) Method: Least Squares Date: 10/23/12 Time: 18:05 Sample (adjusted): 1981 2007

Included observations: 27 after adjustments ariable VCoefficient Std. Error D(LM(-1)) D(LM(-1),2) C

@TREND(1978) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

-1.187734 0.418377 0.103301 0.004374 0.560167 0.502797 0.119892 0.330604 21.12471 9.764190 0.000241

0.223382 0.171790

0.058331 1.770947 0.003026 1.445444 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat

由于-5.311明显小于给定的显著性水平的临界值，所以拒绝零假设

，即0，

不存在单位根1，原LM序列并非平稳，但一阶差分后是平稳序列。（2）检验LX与LM的单整性

由（1）可知，LX与LM都是经过一次差分变成平稳序列，所以LX与LM都是1阶单整序列。

（3）检验LX与LM的协整性

对LX和LM进行回归分析，获得残差Resid，并新建e序列存放Resid。对e序列进行单整性检验。由于残差序列均值为0，所以选择无截距项、无趋势项的ADF检验。

Null Hypothesis: E has a unit root Exogenous: None

Prob.* 0.0000

Prob. 0.0000 -0.004297 0.311145 -0.076430 -0.029282

t-Statistic -4.868963 -2.647120 -1.952910 -1.610011

Lag Length: 0 (Automatic based on SIC, MAXLAG=7)

Augmented Dickey-Fuller test statistic Test critical values: 1% level

5% level 10% level

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation Dependent Variable: D(E) Method: Least Squares Date: 10/23/12 Time: 18:25 Sample (adjusted): 1979 2007

Included observations: 29 after adjustments Variable Coefficient Std. Error t-Statistic E(-1) -0.911793 0.187266 -4.868963 R-squared 0.458378 Mean dependent var Adjusted R-squared 0.458378 S.D. dependent var S.E. of regression Sum squared resid

0.228987 1.468179

Akaike info criterion Schwarz criterion

Log likelihood 2.108239 Hannan-Quinn criter. -0.061664 Durbin-Watson stat 1.958349

由图可知，残差序列的t检验统计量为-4.8690，小于已有各个水平下的临界值，从而拒绝原假设，表明残差序列不存在单位根，是平稳序列。LX与LM为（1，1）阶协整。存在协整关系。

（4）估计LX关于LM的误差修正模型。

由以上分析可知，LX与LM之间存在协整关系，表明两者之间存在长期均衡关系。但从短期来看，可能会存在非均衡的情况，为了模型能够反映短期内的动态调整，可以把(3)中分析得到的残差作为均衡误差，通过建立误差修正模型把LX的短期行为和长期变化联系起来。误差修正模型为：

LXtLMtet1t

对LX、LM两个序列进行差分，两个差分序列分别为dlx和dlm。建立ecm模型。

对dlx和dlm进行回归分析，得

Dependent Variable: DLX Method: Least Squares

Date: 10/23/12 Time: 19:57 Sample (adjusted): 1980 2007

Included observations: 28 after adjustments Variable Coefficient Std. Error DLM 0.644859 0.217209 E(-1) -0.869768 0.185643 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood

0.366332 0.341960 0.224591 1.311471 3.124489

t-Statistic 2.968843 -4.685178

Prob. 0.0063 0.0001 0.160368 0.276864 -0.080321 0.014837 -0.051230

Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter.

Durbin-Watson stat 1.277480

可见所有t值都通过t检验。因此误差修正模型为

DLXt0.6449DLMt0.8698et1

（2.9688）（-4.6852） R=0.3663 DW=1.2775

Eviews P211 我国城镇居民消费结构对比分析

面板模型类型的选择和估计

（1）计算混合模型的残差平方和

Dependent Variable: CONSUME? Method: Pooled Least Squares Date: 10/23/12 Time: 17:01 Sample: 1996 2003

Included observations: 8 Cross-sections included: 29

Total pool (balanced) observations: 232

Variable C

INCOME? R-squared

Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

因此表达式为：

Coefficient 265.0866 0.747336 0.980227 0.980141 225.4217 11687434 -1585.159 11402.02 0.000000

Std. Error 46.32223

t-Statistic 5.722664

Prob. 0.0000 0.0000 4952.143 1599.622 13.68241 13.71212 13.69439 0.633044

0.006999 106.7802 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat

comsumet265.08660.7473incomet （5.7227）（106.7802） R=0.9802 残差平方和SSEr=11687434

2模型显示：29个省市的人均收入变动1个单位，人均消费变动0.7473个单位。（2）计算个体固定效应回归模型的残差平方和

Dependent Variable: CONSUME? Method: Pooled Least Squares Date: 10/23/12 Time: 17:11 Sample: 1996 2003

Included observations: 8

Cross-sections included: 29

Total pool (balanced) observations: 232 Variable C

INCOME?

Fixed Effects (Cross) BEIJ--C TIANJ--C HEB--C SHANX--C LMG--C LIAON--C JIN--C HEILJ--C SHANGH--C JIANGS--C

Coefficient 546.2405 0.702507

528.7945 -2.098388 -206.0987 -150.5853 -231.7145 104.2279 -19.82496 -219.3924 185.5096 -124.4121

Std. Error 52.79208 0.008264

t-Statistic 10.34702 85.00834

Prob. 0.0000 0.0000

ZHEJ--C ANH--C FUJ--C JIANGX--C SHAND--C HEN--C HUB--C HUN--C

GUANGD--C GUANGX--C HAIN--C SIC--C GUIZ--C YUNN--C SHANXI--C GANS--C QINH--C NINX--C XINJ--C

107.6448 -106.1617

-120.2966 -383.5205 -212.4554 -205.7808 165.8370 139.3203 433.8359 47.39929 -146.1806 181.3348 -44.58322 96.97064 168.9827 3.968558 39.49803 79.92879 -110.1475

4952.143 1599.622 13.01743 13.46313 13.19718 1.497457

Effects Specification

Cross-section fixed (dummy variables) R-squared 0.992012 Adjusted R-squared 0.990865 S.E. of regression Sum squared resid Log likelihood F-statistic

Prob(F-statistic)

152.8873 4721654. -1480.022 865.0129 0.000000

Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter. Durbin-Watson stat

因此个体固定效应模型的相应表达式是：

consumet546.2405528.7945D2...110.1475D290.7025incomet

(85.0083) R=0.9920

2个体固定效应模型的残差平方和

其中虚拟变量

DiD2SSEu4721654

,…

D29的定义是：

1或0 且i=2，3，…29

当所有虚拟变量取0时，截距项为546.2405表示安徽省的自发消费支出；当

D2取1且其它

都虚拟变量取0时，截距项为（546.2405+528.7945）=1075.035表示北京的自发消费支出。以此类推，可得出每个地区的自发消费支出。

由模型可知，29个省市的城镇居民人均消费支出占收入的70.25%，北京市居民的自发消费支出明显高于其他地区。模型的选择检验：

H0H1：i0模型中不同个体的截距项相同（混合模型）......受约束模型：模型中不同个体的截距项i不同（个体固定效应模型）......非约束模型

统计量F=((SSEr-SSEu)/(N-1))/(( SSEu/(NT-N-k))=10.276,大于5%置信水平下的分布临界值，应拒绝零假设，所以选择个体光电效应模型更为合理。

综上所述，1996~2003年中国29个省市居民家庭人均消费和收入问题的研究应该建立个体固定效应模型，即随地区不同，自发消费支出存在显著性差异，人均消费平均占收入的70.25%。

Eviews P252 中国城镇居民的生活消费支出和可支配收入的数量关系

1、对生活消费支出和可支配收入两个序列单整性检验

Null Hypothesis: D(SR) has a unit root Exogenous: Constant, Linear Trend

Lag Length: 10 (Automatic based on SIC, MAXLAG=11)

t-Statistic -15.35611 -4.090602 -3.473447 -3.163967

Prob.* 0.0001

Std. Error

t-Statistic

Prob.

Augmented Dickey-Fuller test statistic Test critical values:

1% level 5% level 10% level

*MacKinnon (1996) one-sided p-values.

Dependent Variable: D(SR,2) Method: Least Squares Date: 10/23/12 Time: 20:29

Sample (adjusted): 1993M01 1998M12 Included observations: 72 after adjustments

Variable

Coefficient

Augmented Dickey-Fuller Test Equation

D(SR(-1)) D(SR(-1),2) D(SR(-2),2) D(SR(-3),2) D(SR(-4),2) D(SR(-5),2) D(SR(-6),2) D(SR(-7),2) D(SR(-8),2) D(SR(-9),2) D(SR(-10),2)

@TREND(1992M01)

R-squared

Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

-11.75623 9.795575 8.845714 7.929526 7.006358 6.081439 5.123690 4.149816 3.148907 2.118391 0.999209 70.61308 0.079438

0.765574 0.729268 0.685529 0.627553 0.561630 0.485560 0.409138 0.324499 0.240559 0.157661 0.082962 9.653312 0.163919

-15.35611 13.43206 12.90349 12.63563 12.47505 12.52459 12.52313 12.78837 13.08995 13.43634 12.04421 7.314908 0.484616

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.6297 0.526806 104.7211 9.691469 10.10253 9.855116 1.930907

0.938954 Mean dependent var 0.926538 S.D. dependent var 28.38341 Akaike info criterion 47531.47 Schwarz criterion -335.8929 Hannan-Quinn criter. 75.62420 Durbin-Watson stat 0.000000

可见SR是1阶单整序列

Null Hypothesis: D(ZC) has a unit root Exogenous: Constant, Linear Trend

Lag Length: 10 (Automatic based on SIC, MAXLAG=11)

t-Statistic -7.983964 -4.090602 -3.473447 -3.163967

Prob.* 0.0000

Augmented Dickey-Fuller test statistic Test critical values:

1% level 5% level 10% level

*MacKinnon (1996) one-sided p-values.

Dependent Variable: D(ZC,2) Method: Least Squares Date: 10/23/12 Time: 21:01

Sample (adjusted): 1993M01 1998M12 Included observations: 72 after adjustments

Augmented Dickey-Fuller Test Equation

Variable D(ZC(-1)) D(ZC(-1),2) D(ZC(-2),2) D(ZC(-3),2) D(ZC(-4),2) D(ZC(-5),2) D(ZC(-6),2) D(ZC(-7),2) D(ZC(-8),2) D(ZC(-9),2) D(ZC(-10),2)

@TREND(1992M01) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

Coefficient -10.48184 8.569867 7.624906 6.653426 5.773019 4.890708 3.931355 3.126982 2.297599 1.507902 0.676720 60.38091 -0.133949

Std. Error 1.312862 1.261370 1.187981 1.086581 0.966364 0.828509 0.686923 0.534175 0.383790 0.246171 0.121540 13.78526 0.221219

t-Statistic -7.983964 6.794096 6.418372 6.123265 5.973956 5.903023 5.723142 5.853856 5.986603 6.125414 5.567898 4.380105 -0.605506

Prob. 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.5472 0.370000 97.66146 10.28039 10.69146 10.44404 1.969254

0.873514 Mean dependent var 0.847788 S.D. dependent var 38.10203 Akaike info criterion 85654.11 Schwarz criterion -357.0942 Hannan-Quinn criter. 33.95448 Durbin-Watson stat 0.000000

ZC的一阶差分也是平稳序列，所以ZC也是1阶单整序列。 2、对生活消费支出和可支配收入两个序列协整性检验

由于SR序列与ZC序列都是1阶单整序列，现在使用EG检验生活消费支出和可支配收入两个序列的协整关系。

第一步，使用OLS计算非均衡误差得到残差序列et；第二步，检验et的单整性。

Null Hypothesis: E has a unit root Exogenous: Constant, Linear Trend

Lag Length: 0 (Automatic based on SIC, MAXLAG=11)

t-Statistic -7.388970 -4.072415 -3.464865 -3.158974

Prob.* 0.0000

Augmented Dickey-Fuller test statistic Test critical values:

1% level 5% level 10% level

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation Dependent Variable: D(E) Method: Least Squares Date: 10/23/12 Time: 21:16

Sample (adjusted): 1992M02 1998M12 Included observations: 83 after adjustments

Variable E(-1) C

@TREND(1992M01)

R-squared

Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

Coefficient -0.811393 -3.980125 0.095922

Std. Error 0.109811 6.980138 0.144511

t-Statistic -7.388970 -0.570207 0.663770

Prob. 0.0000 0.5701 0.5087 0.051836 40.23706 9.767360 9.854788 9.802484 1.972016

0.405635 Mean dependent var 0.390776 S.D. dependent var 31.40616 Akaike info criterion 78907.76 Schwarz criterion -402.3455 Hannan-Quinn criter. 27.29869 Durbin-Watson stat 0.000000

可见，et为稳定序列I（0），所以ZC和SR为（1，1）阶协整。

3、建立误差修正模型

由以上分析可知，ZC与SR之间存在协整关系，表明两者之间存在长期均衡关系。但从短期来看，可能会存在非均衡的情况，为了模型能够反映短期内的动态调整，可以把(2)中分析得到的残差作为均衡误差，通过建立误差修正模型把ZC的短期行为和长期变化联系起来。误差修正模型为：

ZCtSRtet1t

对ZC、SR两个序列进行差分，两个差分序列分别为dzc和dsr。建立ecm模型。

对dzc和dsr进行回归分析，得

Dependent Variable: DZC Method: Least Squares Date: 10/23/12 Time: 21:28

Sample (adjusted): 1992M02 1998M12 Included observations: 83 after adjustments

Variable DSR E(-1)

Coefficient 0.769475 -0.779417

Std. Error 0.059046 0.112456

t-Statistic 13.03180 -6.930886

Prob. 0.0000 0.0000

R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat

0.691068 Mean dependent var 0.687254 S.D. dependent var 31.15883 Akaike info criterion 78640.70 Schwarz criterion -402.2048 Hannan-Quinn criter. 1.995755

4.538434 55.71666 9.739874 9.798159 9.763290

可见所有t值都通过t检验。因此误差修正模型为

DZCt0.7695DSRt0.7794et1

（13.03）（-6.93） R=0.6911 DW=1.9958

2对dzc和dsr进行协整回归

Dependent Variable: DZC Method: Least Squares Date: 10/23/12 Time: 22:15

Sample (adjusted): 1992M02 1998M12 Included observations: 83 after adjustments

Variable C DSR

R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

Coefficient 0.924934 0.660167

Std. Error 4.333580 0.072169

t-Statistic 0.213434 9.147575

Prob. 0.8315 0.0000 4.538434 55.71666 10.20496 10.26325 10.22838 2.753565

0.508131 Mean dependent var 0.502059 S.D. dependent var 39.31642 Akaike info criterion 125208.3 Schwarz criterion -421.5060 Hannan-Quinn criter. 83.67814 Durbin-Watson stat 0.000000

得协整回归模型为：dzct0.92490.6602dsrt(dzct上需加hat)

所以，中国城镇居民月人均生活费支出的变化不仅取决于可支配收入的变化，而且还取决于上一期生活费支出对均衡水平的偏离，误差项估计的系数0.7794体现了对偏离的修正，上一期如果是负偏离，本期就是正向的修正，上一期偏离越远，本期修正的量越大，即系统存在误差修正机制，体现了生活费支出与可支配收入的动态调整过程。结论：dzc关于dsr的长期弹性为0.6602，短期弹性为0.7695.

eviewsP310中国西北地区货币政策效应差异的实证研究

该题主要是要确定该面板数据的模型选择，即在混合模型、变截距模型和变系数模型中根据F检验选择最优拟合模型。

由于有三个模型，所以需要进行两次F检验。

基本面板模型为：ln(GDPit)citln(M2it)itit分别进行3次回归分析，结果如下表所示：

Dependent Variable: GDP? Method: Pooled Least Squares Date: 10/23/12 Time: 22:28 Sample: 1990 2006 Included observations: 17 Cross-sections included: 5

Total pool (balanced) observations: 85

Variable C _XJ--M2_XJ _QH--M2_QH _GS--M2_GS _XZ--M2_XZ _NX--M2_NX Fixed Effects (Cross)

_XJ--C _QH--C _GS--C _XZ--C _NX--C

Coefficient 0.620653 0.768577 0.775791 0.595151 0.780014 0.720649 0.034162 -0.155790 0.511382 -0.338224 -0.051530

Std. Error 0.037969 0.033479 0.035122 0.028597 0.031635 0.031710

t-Statistic 16.34637 22.95722 22.08878 20.81168 24.65695 22.72618

Prob. 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 2.496976 0.498509 -2.887749 -2.600378 -2.772160 0.915719

Effects Specification

Cross-section fixed (dummy variables)

R-squared

Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

0.989505 Mean dependent var 0.988245 S.D. dependent var 0.054048 Akaike info criterion 0.219091 Schwarz criterion 132.7293 Hannan-Quinn criter. 785.6654 Durbin-Watson stat 0.000000

【上图为变系数模型回归分析】

Dependent Variable: GDP?

Method: Pooled Least Squares Date: 10/23/12 Time: 22:29 Sample: 1990 2006 Included observations: 17 Cross-sections included: 5

Total pool (balanced) observations: 85

Variable C M2?

Fixed Effects (Cross)

_XJ--C _QH--C _GS--C _XZ--C _NX--C

Coefficient 0.631320 0.719292 0.174235 -0.031141 0.135411 -0.219624 -0.058881

Std. Error 0.042604 0.016223

t-Statistic 14.81825 44.33727

Prob. 0.0000 0.0000 2.496976 0.498509 -2.671580 -2.499158 -2.602227 0.863239

Effects Specification

Cross-section fixed (dummy variables) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

0.985686 Mean dependent var 0.984780 S.D. dependent var 0.061500 Akaike info criterion 0.298799 Schwarz criterion 119.5422 Hannan-Quinn criter. 1088.031 Durbin-Watson stat 0.000000

【上图为变截距模型回归分析】

Dependent Variable: GDP? Method: Pooled Least Squares Date: 10/23/12 Time: 22:29 Sample: 1990 2006 Included observations: 17 Cross-sections included: 5

Total pool (balanced) observations: 85

Variable C M2?

Fixed Effects (Cross)

_XJ--C

Coefficient 0.631320 0.719292 0.174235

Std. Error 0.042604 0.016223

t-Statistic 14.81825 44.33727

Prob. 0.0000 0.0000

_QH--C _GS--C _XZ--C _NX--C

-0.031141 0.135411 -0.219624 -0.058881

2.496976 0.498509 -2.671580 -2.499158 -2.602227 0.863239

Effects Specification

Cross-section fixed (dummy variables) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

0.985686 Mean dependent var 0.984780 S.D. dependent var 0.061500 Akaike info criterion 0.298799 Schwarz criterion 119.5422 Hannan-Quinn criter. 1088.031 Durbin-Watson stat 0.000000

【上图为混合模型回归分析】

SSR1=0.219091 SSR2=0.298799 SSR3=0.298799

第一次：

H0:为混合模型 H1:为非混合模型

F1=84.5340>临界值，拒绝原假设，模型为非混合模型；

第二次：

H0:为变截距模型 H1:为变系数模型

F2=332.65678>临界值，拒绝原假设，模型为变系数模型。面板模型估计：

Dependent Variable: GDP? Method: Pooled Least Squares Date: 10/24/12 Time: 12:52 Sample: 1990 2006 Included observations: 17 Cross-sections included: 5

Total pool (balanced) observations: 85

Variable C

Coefficient 0.620653

Std. Error 0.037969

t-Statistic 16.34637

Prob. 0.0000

_XJ--M2_XJ _QH--M2_QH _GS--M2_GS _XZ--M2_XZ _NX--M2_NX Fixed Effects (Cross)

_XJ--C _QH--C _GS--C _XZ--C _NX--C

0.768577 0.775791 0.595151 0.780014 0.720649 0.034162 -0.155790 0.511382 -0.338224 -0.051530

0.033479 0.035122 0.028597 0.031635 0.031710

22.95722 22.08878 20.81168 24.65695 22.72618

0.0000 0.0000 0.0000 0.0000 0.0000 2.496976 0.498509 -2.887749 -2.600378 -2.772160 0.915719

Effects Specification

Cross-section fixed (dummy variables) R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic)

0.989505 Mean dependent var 0.988245 S.D. dependent var 0.054048 Akaike info criterion 0.219091 Schwarz criterion 132.7293 Hannan-Quinn criter. 785.6654 Durbin-Watson stat 0.000000

it都通过了t检验，说明货币政策长期显著影响区域经济的发展水平，正向的货币政策冲

击会导致区域经济总量的扩张，按照货币政策效应应由弱到强依次为：西藏>青海>新疆>宁夏>甘肃。所以，货币政策长期效应在西北地区内部也存在差异。

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文