Annexe 3 : Test de Ramsey
Ramsey RESET Test:
|
F-statistic 0.073191 Probability 0.792248
Log likelihood ratio 0.145848 Probability 0.702535
|
|
Test Equation:
|
|
Dependent Variable: TCPIB
|
|
Method: Least Squares
|
|
Date: 08/24/12 Time: 03:56
|
|
Sample: 1991 2010
|
|
Included observations: 20
|
|
Variable Coefficient Std. Error t-Statistic Prob.
|
|
STDPIB -1.807829
|
1.752822
|
-1.031382
|
0.3267
|
|
STDEXP -0.193514
|
0.578859
|
-0.334302
|
0.7451
|
|
VTE 6.365144
|
3.327538
|
1.912869
|
0.0848
|
|
SEDEXP 22.57015
|
12.89484
|
1.750324
|
0.1106
|
|
TCDEM 1.737920
|
1.086551
|
1.599483
|
0.1408
|
|
BCPIB 22.55928
|
16.32682
|
1.381732
|
0.1971
|
|
SOBPPIB 26.99993
|
13.12033
|
2.057870
|
0.0666
|
|
EXPPIB 68.71192
|
13.58736
|
5.057046
|
0.0005
|
|
C -21.15628
|
8.195275
|
-2.581522
|
0.0273
|
|
FITTED^2 -0.005393
|
0.019936
|
-0.270537
|
0.7922
|
|
R-squared 0.934628
|
Mean dependent var -0.125000
|
|
Adjusted R-squared 0.875792
|
S.D. dependent var 6.466991
|
|
S.E. of regression 2.279169
|
Akaike info criterion 4.792352
|
|
Sum squared resid 51.94611
|
Schwarz criterion 5.290218
|
|
Log likelihood -37.92352
|
F-statistic 15.88551
|
|
Durbin-Watson stat 1.832950
|
Prob(F-statistic) 0.000084
|
Source : Nos calculs à l?aide du logiciel E-views
3.1
Le tableau ci-dessus présente le test de Ramsey, ceci
nous a permis de vérifier l?absence d?autocorélation des erreurs.
H0 (absence d?autocorélation des résidus) est acceptée si
les probabilités sont supérieures à 0,05.
Annexe 4 : Test de Breusch-Godfrey
Breusch-Godfrey Serial Correlation LM Test:
|
F-statistic 1.555129 Probability 0.262967
Obs*R-squared 5.136568 Probability 0.076667
|
|
Test Equation:
|
|
Dependent Variable: RESID
|
|
Method: Least Squares
|
|
Date: 08/24/12 Time: 03:54
|
|
Presample missing value lagged residuals set to zero.
|
|
Variable Coefficiet Std. Error t-Statistic Prob.
|
|
STDPIB 0.194235
|
1.449500
|
0.134001
|
0.8964
|
|
STDEXP -0.258860
|
0.560241
|
-0.462052
|
0.6550
|
|
VTE 0.148563
|
3.030209
|
0.049027
|
0.9620
|
|
SEDEXP 6.692941
|
12.29040
|
0.544567
|
0.5993
|
|
TCDEM 0.514651
|
1.035090
|
0.497204
|
0.6310
|
|
BCPIB 5.801345
|
15.46907
|
0.375029
|
0.7163
|
|
SOBPPIB 0.899961
|
11.58305
|
0.077696
|
0.9398
|
|
EXPPIB 16.72447
|
15.77383
|
1.060267
|
0.3166
|
|
C -4.907776
|
7.761829
|
-0.632296
|
0.5429
|
|
RESID(-1) -0.517360
|
0.458291
|
-1.128890
|
0.2881
|
|
RESID(-2) -0.775700
|
0.443346
|
-1.749650
|
0.1141
|
|
R-squared 0.256828
|
Mean dependent var -1.50E-15
|
|
Adjusted R-squared -0.568918
|
S.D. dependent var 1.659523
|
|
S.E. of regression 2.078659
|
Akaike info criterion 4.602816
|
|
Sum squared resid 38.88743
|
Schwarz criterion 5.150469
|
|
Log likelihood -35.02816
|
F-statistic 0.311026
|
|
Durbin-Watson stat 2.307947
|
Prob(F-statistic) 0.958499
|
Source : Nos calculs à l?aide du logiciel E-views
3.1
Ce tableau présente le test Breusch-Godfrey. Il a
permis de confirmer l?absence d?auto-corélation des erreurs ci-haut
vérifier. Si les probabilités sont inférieures à
0,05, on rejette H0 de non auto corrélation des résidus
(erreurs).
| 
