我国货币需求函数的相关因素分析
根据货币学派的主张，认为只要保持货币供给量满足社会经济的发展需要，就不必需要政府来干预市场。在此选取我国1991年至2007年的数据进行了简单的分析。
(一)
Dependent Variable: M
Method: Least Squares
Date: 11/29/09 Time: 18:54
Sample: 1990 2007
Included observations: 18
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
GDP
CPI
PPI
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
+09
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
F-statistic
Durbin-Watson stat
Prob(F-statistic)
Dependent Variable: LNM
Method: Least Squares
Date: 11/29/09 Time: 19:16
Sample: 1990 2007
Included observations: 18
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
LNGDP
CPI
PPI
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
F-statistic
Durbin-Watson stat
Prob(F-statistic)
由上表知，F检验通过，但是CPI的t检验未通过。又考虑到CPI与PPI都是经济发展状况的反映，可能存在较强的多重共线性。求出CPI与PPI的相关系数矩阵，见下表：
Correlation
LNGDP
CPI
PPI
LNGDP
1
CPI
1
PPI
1
由上表可知，CPI与PPI存在多重共线性，运用逐步回归法：
保留CPI,去掉PPI，对数据进行回归，如下：
Dependent Variable: LNM
Method: Least Squares
Date: 11/29/09 Time: 19:08
Sample: 1990 2007
Included observations: 18
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
LNGDP
CPI
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
F-statistic
Durbin-Watson stat
Prob(F-statistic)
由DW值可知，模型存在正的序列相关性，由图示法做---图如下：
由图可知有很明显的正序列相关性。通过广义差分法，得下表：
Dependent Variable: LNM
Method: Least Squares
Date: 11/29/09 Time: 18:59
Sample (adjusted): 1991 2007
Included observations: 17 after adjustments
Convergence achieved after 18 iterations
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
LNGDP
CPI
AR(1)
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
F-statistic
Durbin-Watson stat
Prob(F-statistic)
Inverted AR Roots
.94
保留PPI，去掉CPI
Dependent Variable: LNM
Method: Least Squares
Date: 11/29/09 Time: 19:40
Sample: 1990 2007
Included observations: 18
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
LNGDP
PPI
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
F-statistic
Durbin-Watson stat
Prob(F-statistic)
由图可知有很明显的正序列相关性。通过广义差分法，得下表：
Dependent Variable: LNM
Method: Least Squares
Date: 11/29/09 Time: 19:41
Sample (adjusted): 1991 2007
Included observations: 17 after adjustments
Convergence achieved after 9 iterations
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
LNGDP
PPI
AR(1)
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
F-statistic
Durbin-Watson stat
Prob(F-statistic)
Inverted AR Roots
.94
(二)
考虑临界指标的虚拟变量的引入：
考虑1998年亚洲金融危机的影响，考虑在此时的货币供给量是否有明显的转折：
Dependent Variable: LNM
Method: Least Squares
Date: 11/29/09 Time: 20:04
Sample: 1990 2007
Included observations: 18
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
LNGDP
()*D1
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
F-statistic
Durbin-Watson stat
Prob(F-statistic)
由上图，知货币供给量在1998年出现转折并不明显，另外，我国在此阶段确实保持了稳定的货币供给量。
(三)
Null Hypothesis: D(A1,2) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on AIC, MAXLAG=0)
t-Statistic
Prob.*
Augmented Dickey-Fuller test statistic
Test critical values:
1% level
5% level
10% level
*MacKinnon (1996) one-sided p-values.
Warning: Probabilities and critical values calculated for 20 observations
and may not be accurate for a sample size of 10
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(A1,3)
Method: Least Squares
Date: 12/23/09 Time: 20:17
Sample (adjusted): 1998 2007
Included observations: 10 after adjustments
Variable
Coefficient
Std. Error
t-Statistic
Prob.
D(A1(-1),2)
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
93746207
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
Durbin-Watson stat
Null Hypothesis: D(GDP,2) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=2)
t-Statistic
Prob.*
Augmented Dickey-Fuller test statistic
Test critical values:
1% level
5% level
10% level
*MacKinnon (1996) one-sided p-values.
Warning: Probabilities and critical values calculated for 20 observations
and may not be accurate for a sample size of 10
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(GDP,3)
Method: Least Squares
Date: 12/24/09 Time: 14:25
Sample (adjusted): 1998 2007
Included observations: 10 after adjustments
Variable
Coefficient
Std. Error
t-Statistic
Prob.
D(GDP(-1),2)
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
+08
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
Durbin-Watson stat
Null Hypothesis: D(A3,2) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on AIC, MAXLAG=2)
t-Statistic
Prob.*
Augmented Dickey-Fuller test statistic
Test critical values:
1% level
5% level
10% level
*MacKinnon (1996) one-sided p-values.
Warning: Probabilities and critical values calculated for 20 observations
and may not be accurate for a sample size of 10
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(A3,3)
Method: Least Squares
Date: 12/23/09 Time: 20:28
Sample (adjusted): 1998 2007
Included observations: 10 after adjustments
Variable
Coefficient
Std. Error
t-Statistic
Prob.
D(A3(-1),2)
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
Durbin-Watson stat
Null Hypothesis: B1 has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on AIC, MAXLAG=2)
t-Statistic
Prob.*
Augmented Dickey-Fuller test statistic
Test critical values:
1% level
5% level
10% level
*MacKinnon (1996) one-sided p-values.
Warning: Probabilities and critical values calculated for 20 observations
and may not be accurate for a sample size of 12
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(B1)
Method: Least Squares
Date: 12/23/09 Time: 20:42
Sample (adjusted): 1996 2007
Included observations: 12 after adjustments
Variable
Coefficient
Std. Error
t-Statistic
Prob.
B1(-1)
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
3493660.
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
Durbin-Watson stat
残差进行单位根检验：
Null Hypothesis: RESID01 has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=2)
t-Statistic
Prob.*
Augmented Dickey-Fuller test statistic
Test critical values:
1% level
5% level
10% level
*MacKinnon (1996) one-sided p-values.
Warning: Probabilities and critical values calculated for 20 observations
and may not be accurate for a sample size of 12
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(RESID01)
Method: Least Squares
Date: 12/24/09 Time: 14:30
Sample (adjusted): 1996 2007
Included observations: 12 after adjustments
Variable
Coefficient
Std. Error
t-Statistic
Prob.
RESID01(-1)
R-squared
Mean dependent var
Adjusted R-squared
. dependent var
. of regression
Akaike info criterion
Sum squared resid
29717903
Schwarz criterion
Log likelihood
Hannan-Quinn criter.
Durbin-Watson stat
