ANNEXES
ANNEXE I : ANALYSE DE LA STATIONNARITE DES SERIES
I.1- Stationnarité de ihpc
|
TABLE
|
1:
|
test
Test
Statistic
|
for unit root Number of obs =
Interpolated Dickey-Fuller
1% Critical 5% Critical 10%
Value Value
|
57
Critical Value
|
|
Augmented Dickey-Fuller
|
|
Z(t)
|
|
|
3.347
|
-2.617
|
|
-1.950
|
-1.610
|
|
D.ihpc
|
|
|
Coef.
|
Std. Err.
|
t
|
P>|t|
|
[95% Conf.
|
Interval]
|
|
ihpc
|
+
|
|
|
|
|
|
|
|
|
L1.
|
|
|
.0062319
|
.0018618
|
3.35
|
0.002
|
.0024975
|
.0099662
|
|
LD.
|
|
|
.4673105
|
.1327332
|
3.52
|
0.001
|
.2010813
|
.7335397
|
|
L2D.
|
|
|
-.5715382
|
.1501091
|
-3.81
|
0.000
|
-.872619
|
-.2704574
|
|
L3D.
|
|
|
.3039482
|
.0993908
|
3.06
|
0.003
|
.1045955
|
.5033009
|
|
TABLE
|
2:
|
|
|
|
|
|
|
|
|
Phillips-Perron
|
|
test for unit root
|
|
Number of obs =
|
60
|
|
|
|
|
|
|
Newey-West lags =
|
3
|
Interpolated Dickey-Fuller
Test 1% Critical 5% Critical 10% Critical
Statistic Value Value Value
Z(rho) 0.606 -12.980 -7.740 -5.520
Z(t) 3.086 -2.616 -1.950 -1.610
ihpc | Coef. Std. Err. t P>|t| [95% Conf. Interval]
+
ihpc |
L1. | 1.010241 .0024769 407.86 0.000 1.005285 1.015197
I.2- Stationnarité de infglis
TABLE 3
Augmented Dickey-Fuller test for unit root Number of obs = 55
Interpolated Dickey-Fuller
Test 1% Critical 5% Critical 10% Critical
Statistic Value Value Value
Z(t) -3.295 -3.573 -2.926 -2.598
MacKinnon approximate p-value for Z(t) = 0.0151
D.infglis | Coef. Std. Err. t P>|t| [95% Conf. Interval]
infglis |
L1. | -.1878307 .0570071 -3.29 0.002 -.3022239 -.0734375
LD. | .2642757 .0765217 3.45 0.001 .1107236 .4178277
_cons | .0056985 .0023338 2.44 0.018 .0010154 .0103815
TABLE 4
Phillips-Perron test for unit root Number of obs = 56
Newey-West lags = 1
Interpolated Dickey-Fuller
Test 1% Critical 5% Critical 10% Critical
Statistic Value Value Value
Z(rho) -23.393 -19.008 -13.348 -10.736
Z(t) -9.985 -3.572 -2.925 -2.598
MacKinnon approximate p-value for Z(t) = 0.0000
infglis | Coef. Std. Err. t P>|t| [95% Conf. Interval]
+
infglis |
L1. | .5974365 .0345129 17.31 0.000 .5282424 .6666306
_cons | .0117074 .0021668 5.40 0.000 .0073632 .0160516
I-3 : Stationnarité de outputgap
TABLE 5
dfuller outputgap, lag(3) regress
Augmented Dickey-Fuller test for unit root Number of obs = 57
Interpolated Dickey-Fuller
Test 1% Critical 5% Critical 10% Critical
Statistic Value Value Value
Z(t) -1.699 -3.570 -2.924 -2.597
MacKinnon approximate p-value for Z(t) = 0.4316
D.outputgap | Coef. Std. Err. t P>|t| [95% Conf. Interval]
+
outputgap |
L1. | -.0634455 .0373402 -1.70 0.095 -.1383741 .0114831
LD. | .1158135 .1304984 0.89 0.379 -.1460507 .3776777
L2D. | .2095736 .1332714 1.57 0.122 -.0578551 .4770023
L3D. | .0654421 .1225056 0.53 0.595 -.1803835 .3112676
cons | -.0006298 .0012486 -0.50 0.616 -.0031354 .0018757
_
TABLE 6
Dickey-Fuller test for unit root Number of obs = 59
Interpolated Dickey-Fuller
Test 1% Critical 5% Critical 10% Critical
Statistic Value Value Value
Z(t) -6.349 -2.616 -1.950 -1.610
D2.outputgap | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
+
outputgap |
LD. | -.8013101 .1262006 -6.35 0.000 -1.053928 -.5486922
TABLE 7
Phillips-Perron test for unit root Number of obs = 59
Newey-West lags = 3
Interpolated Dickey-Fuller
Test 1% Critical 5% Critical 10% Critical
Statistic Value Value Value
Z(rho) -54.816 -12.972 -7.736 -5.518
Z(t) -6.523 -2.616 -1.950 -1.610
D.outputgap | Coef. Std. Err. t P>|t| [95% Conf. Interval]
+
outputgap |
LD. | .1986899 .1262006 1.57 0.121 -.0539281 .4513078
|
TABLE 8
|
61
|
|
|
Number of obs =
|
61
|
|
ARIMA regression Sample: 1 to
|
|
|
|
|
Wald
|
chi2(2) =
|
289.20
|
|
Log likelihood
|
= 189.1048
|
|
|
Prob
|
> chi2 =
|
0.0000
|
|
|
|
|
OPG
|
|
|
|
|
|
outputgap |
|
Coef.
|
Std. Err.
|
z
|
P>|z|
|
[95% Conf.
|
Interval]
|
|
+
outputgap |
|
|
|
|
|
|
|
|
_cons |
|
.0016938
|
.0182379
|
0.09
|
0.926
|
-.0340519
|
.0374395
|
|
+
|
|
|
|
|
|
|
|
ARMA |
ar |
|
|
|
|
|
|
|
L1. |
|
1.163599
|
.1035005
|
11.24
|
0.000
|
.9607422
|
1.366457
|
L2. |
|
|
-.2388272
|
.1050883
|
-2.27
|
0.023
|
-.4447966
|
-.0328579
|
|
|
+
|
|
|
|
|
|
|
|
/sigma |
|
.0107001
|
.0007009
|
15.27
|
0.000
|
.0093263
|
.0120739
|
ANNEXE II : ALGORITHME D'INTERPOLATION
Cette méthode proposée par Goldstein et Khan
(1976) considère trois observations annuelles consécutives d'une
variable de flux x(s), soit xt-1, xt et xt+1 par lesquelles passent la fonction
quadratique définie par le système suivant :
1
? (as2 + bs + c) ds =
xt-1
0
2
? (as2 + bs + c) ds = xt
1
3
? (as2 + bs + c) ds =
xt+1
2
La résolution du système d'équation
donne les valeurs de a, b et c en fonction des xi. Soit :
a = 0.5xt-1 - 1.0x + 0.5xt+1
b = -2.0xt-1 + 3.0x -
1.0xt+1
c = 1.833xt-1 - 1.166x +
0.333xt+1
Pour une année donnée (t), les séries
trimestrielles peuvent être alors interpolées, soit :
1.25
T1 = ? (as2 + bs + c) ds = 0.0545xt-1 +
0.2346xt - 0.0392xt+1
1
1.5
T2 = ? (as2 + bs + c) ds = 0.0079x
t-1 + 0.2655xt - 0.0234xt+1
1.25
1.75
T3 = ? (as2 + bs + c) ds = -0.0234xt-1 +
0.2655xt + 0.078xt+1
1.5
1
T4 = ? (as2 + bs + c) ds = -0.039xt-1 +
0.2343xt + 0.0547xt+1
1.75
Les séries trimestrielles au rythme annuel sont
obtenues en multipliant chaque observation par quatre. L'erreur relative se
situe en moyenne autour de 2%.
| 
