计量经济学论文.doc

对我国第三产业的计量分析
学院：农生院 姓名：张祥 学号：222010326032041

摘要：近年来，第三产业成为我国GDP增长的主要推动力，该文从五个方面对第三产业的发展进行了深入的计量分析，得出第三产业的发展主要跟交通运输、仓库和邮政业、批发和零售业、住宿和餐饮业、金融业、房地产业等有关。

关键词：第三产业，批发和零售业，金融业，房地产业。
1、引言
改革开放以前，在计划经济体制下，我国一直重视发展第一、二产业，忽视了第三产业的发展。近年来，第三产业加快了发展的步伐，改善了我国人民的生活水平，为我国的经济增长作出了突出贡献，成为整个国民经济发展的重要力量。为了弄清第三产业的发展，我从计量经济学的角度出发，首先要建立一个经济模型，然后用最小二乘法来判断它的拟合程度高低，最后进行判断分析和修正。在建模时，根据“经典假设”做了以下处理：经济回归模型是线性的。主要采集的样本是1978年以后的，因为改革开放以后，我国的经济运行机制有了极大的改变，第三产业快速发展，并逐渐成为我国经济的主要成分，因此用这一时期的样本更能反映这种变化。模型中将第三产业作为被解释变量，根据经验引入交通运输、仓储和邮政业、批发和零售业、住宿和餐饮业、金融业、房地产业对模型进行回归分析，以求能使模型具有更高的可操作性。

2、模型
2.1、数据（单位：万亿元）

第三产业Y

年 份
交通运输、
批发和
住宿和
金融业X4
房地产业X5

仓储和邮政业X1
零售业X2
餐饮业X3

1978
872.5
182.0
242.3
44.6
68.2
79.9
1979
878.9
193.7
200.9
44.0
66.9
86.3
1980
982.0
213.4
193.8
47.4
75.0
96.4

1981
1076.6
220.7
231.1
54.1
79.8
99.9
1982
1163.0
246.9
171.4
62.3
114.8
110.8
1983
1338.1
274.9
198.7
72.5
149.0
121.8
1984
1786.3
338.5
363.5
96.8
203.9
162.3
1985
2585.0
421.7
802.4
138.3
259.9
215.2

1986
2993.8
498.8
852.6
163.2
356.4
298.1
1987
3574.0
568.3
1059.6
187.1
450.0
382.6
1988
4590.3
685.7
1483.4
241.4
585.4
473.8
1989
5448.4
812.7
1536.2
277.4
964.3
566.2
1990
5888.4
1167.0
1268.9
301.9
1017.5
662.2

1991
7337.1
1420.3
1834.6
442.3
1056.3
763.7
1992
9357.4
1689.0
2405.0
584.6
1306.2
1101.3
1993
11915.7
2174.0
2816.6
712.1
1669.7
1379.6
1994
16179.8
2787.9
3773.4
1008.5
2234.8
1909.3
1995
19978.5
3244.3
4778.6
1200.1
2798.5
2354.0

1996
23326.2
3782.2
5599.7
1336.8
3211.7
2617.6
1997
26988.1
4148.6
6327.4
1561.3
3606.8
2921.1
1998
30580.5
4660.9
6913.2
1786.9
3697.7
3434.5
1999
33873.4
5175.2
7491.1
1941.2
3816.5
3681.8
2000
38714.0
6161.0
8158.6
2146.3
4086.7
4149.1

2001
44361.6
6870.3
9119.4
2400.1
4353.5
4715.1
2002
49898.9
7492.9
9995.4
2724.8
4612.8
5346.4
2003
56004.7
7913.2
11169.5
3126.1
4989.4
6172.7
2004
64561.3
9304.4
12453.8
3664.8
5393.0
7174.1
2005
74919.3
10666.2
13966.2
4195.7
6086.8
8516.4

2006
88554.9
12183.0
16530.7
4792.6
8099.1
10370.5
2007
111351.9
14601.0
20937.8
5548.1
12337.5
13809.7
2008
131340.0
16362.5
26182.3
6616.1
14863.3
14738.7
2009
147642.1
17057.7
28984.5
7118.2
17727.6
18654.7

2.2模型估计
2.2.1第三产业与交通运输、仓储和邮政业的线性关系
假定第三产业与运输业、仓储和邮政业存在线性关系，,其中为随机误差项，用Eviews软件得到第三产业与交通运输、仓储和邮政业的散点图如下：

由上图可以看出第三产业与交通运输、仓储和邮政业呈线性关系。
在Eviews软件下，得到第三产业与交通运输、仓储和邮政业回归结果如下：

Dependent Variable: Y

Method: Least Squares

Date: 05/19/12 Time: 16:12

Sample: 1 38

Included observations: 32

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

C
-3543.417
1101.518
-3.216850
0.0031
X1
7.897584
0.164425
48.03160
0.0000

R-squared
0.987163
    Mean dependent var
31876.95
Adjusted R-squared
0.986735
    S.D. dependent var
40189.28
S.E. of regression
4628.694
    Akaike info criterion
19.77840
Sum squared resid
6.43E+08
    Schwarz criterion
19.87001
Log likelihood
-314.4544
    Hannan-Quinn criter.
19.80876
F-statistic
2307.035
    Durbin-Watson stat
0.381884
Prob(F-statistic)
0.000000

由上表可知，第三产业总值与交通运输、仓储和邮政业变化的一元线性回归方程为：Y=-3543.417+7.897584X1,拟合系数=0.987163，从回归结果可以看出，模型拟合度很好，可决系数很高,表明交通运输、仓储和邮政业对第三产业总值有显著影响。
2.2.2第三产业与批发和零售业的线性关系
假定第三产业与批发和零售业存在线性关系，，其中为随机误差项，用Eviews软件分析它们之间的关系得到散点图如下：

可以看出它们呈线性关系。
在Eviews软件下，得到第三产业与批发和零售业回归结果如下：

Dependent Variable: Y

Method: Least Squares

Date: 05/19/12 Time: 16:37

Sample: 1 38

Included observations: 32

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

C
-1908.629
553.2654
-3.449753
0.0017
X2
5.196721
0.055324
93.93298
0.0000

R-squared
0.996611
    Mean dependent var
31876.95
Adjusted R-squared
0.996499
    S.D. dependent var
40189.28
S.E. of regression
2378.131
    Akaike info criterion
18.44648
Sum squared resid
1.70E+08
    Schwarz criterion
18.53809
Log likelihood
-293.1437
    Hannan-Quinn criter.
18.47684
F-statistic
8823.405
    Durbin-Watson stat
0.583250
Prob(F-statistic)
0.000000

由上表可得第三产业总值与批发和零售业变化的一元线性回归方程为：Y=-1908.629+5.196721X2，拟合系数=0.996611，从回归结果可以看出，模型拟合度很好，可决系数很高,表明批发和零售业变化对第三产业总值有显著影响。
2.2.3第三产业与住宿和餐饮业的线性关系
仍然假定它们之间存在线性关系，，为随机误差项，用Eviews软件分析它得到第三产业与批发与住宿和餐饮业回归结果如下：

Dependent Variable: Y

Method: Least Squares

Date: 05/19/12 Time: 16:50

Sample: 1 38

Included observations: 32

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

C
-1728.955
712.1898
-2.427661
0.0214
X3
19.68222
0.270507
72.76057
0.0000

R-squared
0.994365
    Mean dependent var
31876.95
Adjusted R-squared
0.994177
    S.D. dependent var
40189.28
S.E. of regression
3066.676
    Akaike info criterion
18.95504
Sum squared resid
2.82E+08
    Schwarz criterion
19.04665
Log likelihood
-301.2806
    Hannan-Quinn criter.
18.98540
F-statistic
5294.101
    Durbin-Watson stat
0.516394
Prob(F-statistic)
0.000000

由上表可得第三产业总值与住宿和餐饮业变化的一元线性回归方程为：Y=-1728.955+19.68222X3，拟合系数=0.994365，从回归结果可以看出，模型拟合度很好，可决系数很高,表明住宿和餐饮业变化对第三产业总值有显著影响。
2.2.4第三产业与金融业之间的线性关系
假定第三产业与金融业之间存在线性关系，，为随机误差项，用Eviews软件分析它们之间的关系得到散点图如下：

由上图可知，第三产业与金融业之间呈一定的线性关系。在Eviews软件下，得到第三产业与批发与金融业回归结果如下：
Dependent Variable: Y

Method: Least Squares

Date: 05/19/12 Time: 17:00

Sample: 1 38

Included observations: 32

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

C
649.3815
1551.738
0.418487
0.6786
X4
9.056489
0.281299
32.19519
0.0000

R-squared
0.971871
    Mean dependent var
31876.95
Adjusted R-squared
0.970934
    S.D. dependent var
40189.28
S.E. of regression
6851.795
    Akaike info criterion
20.56287
Sum squared resid
1.41E+09
    Schwarz criterion
20.65448
Log likelihood
-327.0059
    Hannan-Quinn criter.
20.59324
F-statistic
1036.531
    Durbin-Watson stat
0.376238
Prob(F-statistic)
0.000000

由上表可得第三产业总值与住宿和餐饮业变化的一元线性回归方程为：Y=649.3815+9.056489X4，拟合系数=0.971871，从回归结果可以看出，模型拟合度较好，可决系数较高,表明金融业变化对第三产业总值有显著影响。
2.2.5第三产业与房地产业之间的线性关系
假定第三产业与房地产业之间存在线性关系，，为随机误差项，用Eviews软件分析得到第三产业与房地产业回归结果如下：

Dependent Variable: Y

Method: Least Squares

Date: 05/19/12 Time: 17:07

Sample: 1 38

Included observations: 32

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

C
1411.322
645.0973
2.187766
0.0366
X5
8.320696
0.107676
77.27539
0.0000

R-squared
0.995001
    Mean dependent var
31876.95
Adjusted R-squared
0.994835
    S.D. dependent var
40189.28
S.E. of regression
2888.429
    Akaike info criterion
18.83527
Sum squared resid
2.50E+08
    Schwarz criterion
18.92688
Log likelihood
-299.3644
    Hannan-Quinn criter.
18.86564
F-statistic
5971.485
    Durbin-Watson stat
2.241768
Prob(F-statistic)
0.000000

由上表可得第三产业总值与房地产业变化的一元线性回归方程为：Y=1411.322+8.320696X5，拟合系数=0.995001，从回归结果可以看出，模型拟合度很好，可决系数很高,表明房地产业变化对第三产业总值有显著影响。

2.3用Eviews软件对影响对第三产业的单因素的单位根检验
2.3.1 用Eviews软件来检验交通运输、仓储和邮政业的单整性水平

Null Hypothesis: X1 has a unit root

Exogenous: None

Lag Length: 3 (Fixed)

t-Statistic
  Prob.*

Augmented Dickey-Fuller test statistic
-4.029971
 0.0031
Test critical values:
1% level

-3.109582

5% level

-2.043968

10% level

-1.597318

*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 5

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(X1)

Method: Least Squares

Date: 05/19/12 Time: 22:30

Sample (adjusted): 9 33

Included observations: 5 after adjustments

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

X1(-1)
-11.61761
2.882803
-4.029971
0.1548
D(X1(-1))
-29.25747
7.034068
-4.159395
0.1502
D(X1(-2))
-46.04027
11.35384
-4.055038
0.1539
D(X1(-3))
272.2184
66.44825
4.096698
0.1524

R-squared
0.997760
    Mean dependent var
648.2904
Adjusted R-squared
0.991039
    S.D. dependent var
516.1230
S.E. of regression
48.85833
    Akaike info criterion
10.60629
Sum squared resid
2387.136
    Schwarz criterion
10.29384
Log likelihood
-22.51572
    Hannan-Quinn criter.
9.767705

由上面的图表数据，-4.029771<-3.109582, 满足在小于1%条件下，不存在单位根，所以拒绝原假设，即交通运输、仓储和邮政业原水平条件下是稳定的。
2.3.2 用Eviews软件来检验批发和零售业的单整性水平

Null Hypothesis: X2 has a unit root

Exogenous: None

Lag Length: 3 (Fixed)

t-Statistic
  Prob.*

Augmented Dickey-Fuller test statistic
 2.046926
 0.9703
Test critical values:
1% level

-3.109582

5% level

-2.043968

10% level

-1.597318

*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 5

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(X2)

Method: Least Squares

Date: 05/19/12 Time: 23:25

Sample (adjusted): 9 33

Included observations: 5 after adjustments

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

X2(-1)
0.115575
0.056463
2.046926
0.2893
D(X2(-1))
1.449616
0.291504
4.972876
0.1263
D(X2(-2))
-0.931448
0.522931
-1.781206
0.3257
D(X2(-3))
-0.749556
0.540166
-1.387640
0.3975

R-squared
0.985953
    Mean dependent var
671.3495
Adjusted R-squared
0.943813
    S.D. dependent var
662.1334
S.E. of regression
156.9502
    Akaike info criterion
12.94030
Sum squared resid
24633.35
    Schwarz criterion
12.62785
Log likelihood
-28.35074
    Hannan-Quinn criter.
12.10171

由上面的图表数据， -1.597318<2.046926，满足在小于10%的条件下，存在单位根，所以接受原假设，即批发和零售业是不稳定的。
2.3.3 用Eviews软件来检验住宿和餐饮业的单整性水平

Null Hypothesis: X3 has a unit root

Exogenous: None

Lag Length: 3 (Fixed)

t-Statistic
  Prob.*

Augmented Dickey-Fuller test statistic
 1.852630
 0.9617
Test critical values:
1% level

-3.109582

5% level

-2.043968

10% level

-1.597318

*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 5

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(X3)

Method: Least Squares

Date: 05/20/12 Time: 08:37

Sample (adjusted): 9 33

Included observations: 5 after adjustments

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

X3(-1)
0.161339
0.087086
1.852630
0.3151
D(X3(-1))
0.451425
0.117160
3.853082
0.1617
D(X3(-2))
-0.589812
0.937790
-0.628938
0.6426
D(X3(-3))
-0.203204
0.431167
-0.471289
0.7196

R-squared
0.996895
    Mean dependent var
198.7134
Adjusted R-squared
0.987580
    S.D. dependent var
203.4264
S.E. of regression
22.67090
    Akaike info criterion
9.070604
Sum squared resid
513.9697
    Schwarz criterion
8.758154
Log likelihood
-18.67651
    Hannan-Quinn criter.
8.232020

由上表可知， -1.597318<1.852630, 满足在小于10%条件下，存在单位根，所以接受原假设，即住宿和餐饮业是不稳定的。
2.3.4 用Eviews软件来检验金融业的单整性水平

Null Hypothesis: X4 has a unit root

Exogenous: None

Lag Length: 3 (Automatic based on SIC, MAXLAG=3)

t-Statistic
  Prob.*

Augmented Dickey-Fuller test statistic
-22.80003
 0.0001
Test critical values:
1% level

-3.109582

5% level

-2.043968

10% level

-1.597318

*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 5

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(X4)

Method: Least Squares

Date: 05/20/12 Time: 08:44

Sample (adjusted): 9 33

Included observations: 5 after adjustments

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

X4(-1)
-0.018413
0.000808
-22.80003
0.0279
D(X4(-1))
-0.916195
0.007588
-120.7351
0.0053
D(X4(-2))
2.738914
0.013850
197.7527
0.0032
D(X4(-3))
0.507025
0.007901
64.17010
0.0099

R-squared
0.999995
    Mean dependent var
327.3802
Adjusted R-squared
0.999981
    S.D. dependent var
292.5052
S.E. of regression
1.270666
    Akaike info criterion
3.307521
Sum squared resid
1.614591
    Schwarz criterion
2.995071
Log likelihood
-4.268802
    Hannan-Quinn criter.
2.468937

由上表可知，-22.80003 <-3.109582, 满足在小于1%条件下，不存在单位根，所以拒绝原假设，即金融业是原条件水平下是很稳定的。

2.3.5用Eviews软件来检验房地产业的单整性水平

Null Hypothesis: X5 has a unit root

Exogenous: None

Lag Length: 3 (Fixed)

t-Statistic
  Prob.*

Augmented Dickey-Fuller test statistic
 1.011106
 0.8862
Test critical values:
1% level

-3.109582

5% level

-2.043968

10% level

-1.597318

*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 5

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(X5)

Method: Least Squares

Date: 05/20/12 Time: 08:55

Sample (adjusted): 9 33

Included observations: 5 after adjustments

Variable
Coefficient
Std. Error
t-Statistic
Prob.  

X5(-1)
0.194321
0.192187
1.011106
0.4965
D(X5(-1))
0.629237
0.902650
0.697100
0.6124
D(X5(-2))
-0.649425
2.220917
-0.292413
0.8189
D(X5(-3))
-0.230146
2.157418
-0.106677
0.9323

R-squared
0.999931
    Mean dependent var
480.6340
Adjusted R-squared
0.999724
    S.D. dependent var
518.3688
S.E. of regression
8.610336
    Akaike info criterion
7.134366
Sum squared resid
74.13789
    Schwarz criterion
6.821916
Log likelihood
-13.83591
    Hannan-Quinn criter.
6.295782

由上表