CONCLUSION
A l'issu de cette réflexion et des résultats
empiriques que nous avons obtenus, il semble effectivement que le VIH/SIDA joue
un rôle très important dans l'explication de la mortalité
infanto-juvénile dans les pays de l'Afrique Subsaharienne. En effet une
augmentation de la prévalence du VIH/SIDA entraine
inéluctablement une augmentation de la mortalité
infanto-juvénile. Cet effet du VIH/SIDA sur la mortalité des
enfants de moins de cinq ans est plus fort en présence
d'instabilité politique.
Néanmoins, on s'aperçoit à travers le
résultat de nos régressions, que la prévalence VIH/SIDA
n'est significative dans l'explication de la mortalité
infanto-juvénile que si elle associée à
l'instabilité politique. De plus, ce modèle a permis de mettre en
évidence les autres déterminants de la mortalité
infanto-juvénile en occurrence la vaccination, l'éducation des
mères, le niveau de revenu et l'accouchement assisté.
Toutefois, il est important d'insister sur le fait que ce
modèle ne peut être exhaustif, étant donné que nous
n'avions pas pu mettre en évidence les effets direct du VIH/SIDA,
à travers la prise en compte de tous les aspects de la transmission
mère-enfant par manque de données. Ainsi, il est évident
que la proportion d'enfants infectée par suite de l'allaitement ou au
cours de l'accouchement a sans nul doute, un effet sur la mortalité des
enfants due au VIH/SIDA qui n'a pu être prise en compte dans notre
modèle.
Nous soulignons à ce niveau, le fait que nous n'avons
pas trouvé de bons instruments pour pouvoir vérifier
empiriquement, l'exogénéité de notre variable de VIH/SIDA.
Il est important de souligner également, que la
littérature portant sur les effets du VIH/SIDA sur la mortalité
infantile n'est pas fournie.
Dans la perspective de l'atteinte des Objectifs du
Millénaire pour Développement, la mesure des effets du VIH/SIDA
sur la mortalité des enfants nécessite des études plus
approfondies. Ceci permettra en plus des efforts déjà
réalisés dans le domaine scientifique, de mieux cerner cette
problématique dans l'optique de leur prise en compte dans les politiques
de réduction de la mortalité infanto-juvénile dans les
Pays en Développement d'une manière générale.
ANNEXES
. *Statistiques desriptives
Variable | Mean Std. Dev. Min Max |
Observations
-----------------+--------------------------------------------+----------------
im overall | 159.3929 59.77393 18 320 |
N = 168
between | 58.52353 20 291 |
n = 42
within | 14.47225 117.1429 204.8929 |
T = 4
|
|
hiv overall | 8.27646 8.608894 .0773347 38.8 |
N = 168
between | 8.295911 .0780011 34.125 |
n = 42
within | 2.554847 -5.96854 16.85146 |
T = 4
|
|
roug overall | 61.55952 21.72989 18 97 |
N = 168
between | 19.62063 29.25 96.75 |
n = 42
within | 9.702305 36.05952 97.55952 |
T = 4
|
|
motheduc overall | 42.52262 19.41235 11 96 |
N = 168
between | 18.5753 13.5 86.5 |
n = 42
within | 6.164089 24.77262 76.02262 |
T = 4
|
|
acoucsis overall | 50.14286 21.10263 6 99 |
N = 168
between | 20.8135 9 94.75 |
n = 42
within | 4.461076 36.89286 66.89286 |
T = 4
|
|
gdppc overall | 2363.624 3009.127 453.1864 19780 |
N = 168
between | 2855.201 489.5347 13217.88 |
n = 42
within | 1024.266 -7121.056 8925.748 |
T = 4
. *Test de Hausman
Fixed-effects (within) regression Number of obs
= 168
Group variable (i): code Number of groups
= 42
R-sq: within = 0.2223 Obs per group:
min = 4
between = 0.4719
avg = 4.0
overall = 0.4331
max = 4
F(6,120)
= 5.72
corr(u_i, Xb) = 0.4522 Prob > F
= 0.0000
------------------------------------------------------------------------------
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | .4582889 .483487 0.95 0.345
-.4989816 1.415559
roug | -.2190717 .1265995 -1.73 0.086
-.46973 .0315865
motheduc | -.758119 .1926906 -3.93 0.000
-1.139633 -.3766051
acoucsis | -.2700784 .2746772 -0.98 0.327
-.8139202 .2737633
gdppc | .0016068 .0011999 1.34 0.183
-.000769 .0039825
inst_hiv | .7315258 .2602651 2.81 0.006
.216219 1.246833
_cons | 214.3106 16.12429 13.29 0.000
182.3856 246.2355
-------------+----------------------------------------------------------------
sigma_u | 48.831644
sigma_e | 15.056297
rho | .91318543 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(41, 120) = 24.08 Prob
> F = 0.0000
Random-effects GLS regression Number of obs
= 168
Group variable (i): code Number of groups
= 42
R-sq: within = 0.1970 Obs per group:
min = 4
between = 0.5861
avg = 4.0
overall = 0.5534
max = 4
Random effects u_i ~ Gaussian Wald chi2(6)
= 71.39
corr(u_i, X) = 0 (assumed) Prob > chi2
= 0.0000
------------------------------------------------------------------------------
im | Coef. Std. Err. z P>|z| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | .4254233 .4223622 1.01 0.314
-.4023914 1.253238
roug | -.3152181 .1213161 -2.60 0.009
-.5529932 -.077443
motheduc | -.9124917 .1800989 -5.07 0.000
-1.265479 -.5595043
acoucsis | -.7711354 .2271328 -3.40 0.001
-1.216307 -.3259633
gdppc | .0006957 .0011248 0.62 0.536
-.0015089 .0029003
inst_hiv | .5794249 .251837 2.30 0.021
.0858335 1.073016
_cons | 253.6693 13.67986 18.54 0.000
226.8573 280.4813
-------------+----------------------------------------------------------------
sigma_u | 38.143536
sigma_e | 15.056297
rho | .86519435 (fraction of variance due to u_i)
------------------------------------------------------------------------------
. hausman eq1
---- Coefficients ----
| (b) (B) (b-B)
sqrt(diag(V_b-V_B))
| eq1 . Difference
S.E.
-------------+----------------------------------------------------------------
hiv | .4582889 .4254233 .0328656
.2353079
roug | -.2190717 -.3152181 .0961464
.0361918
motheduc | -.758119 -.9124917 .1543727
.0685132
acoucsis | -.2700784 -.7711354 .5010569
.1544612
gdppc | .0016068 .0006957 .0009111
.0004179
inst_hiv | .7315258 .5794249 .1521008
.0656966
------------------------------------------------------------------------------
b = consistent under Ho and Ha;
obtained from xtreg
B = inconsistent under Ha, efficient under Ho;
obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 68.20
Prob>chi2 = 0.0000
. *Test de Ramsey Reset
Fixed-effects (within) regression Number of obs
= 168
Group variable (i): code Number of groups
= 42
R-sq: within = 0.2223 Obs per group:
min = 4
between = 0.4719
avg = 4.0
overall = 0.4331
max = 4
F(6,120)
= 5.72
corr(u_i, Xb) = 0.4522 Prob > F
= 0.0000
------------------------------------------------------------------------------
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | .4582889 .483487 0.95 0.345
-.4989816 1.415559
roug | -.2190717 .1265995 -1.73 0.086
-.46973 .0315865
motheduc | -.758119 .1926906 -3.93 0.000
-1.139633 -.3766051
acoucsis | -.2700784 .2746772 -0.98 0.327
-.8139202 .2737633
gdppc | .0016068 .0011999 1.34 0.183
-.000769 .0039825
inst_hiv | .7315258 .2602651 2.81 0.006
.216219 1.246833
_cons | 214.3106 16.12429 13.29 0.000
182.3856 246.2355
-------------+----------------------------------------------------------------
sigma_u | 48.831644
sigma_e | 15.056297
rho | .91318543 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(41, 120) = 24.08 Prob
> F = 0.0000
Fixed-effects (within) regression Number of obs
= 168
Group variable (i): code Number of groups
= 42
R-sq: within = 0.2847 Obs per group:
min = 4
between = 0.3970
avg = 4.0
overall = 0.3581
max = 4
F(9,117)
= 5.17
corr(u_i, Xb) = 0.3909 Prob > F
= 0.0000
------------------------------------------------------------------------------
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | 59.43723 41.34726 1.44 0.153
-22.44885 141.3233
roug | -28.25424 19.74142 -1.43 0.155
-67.35109 10.84262
motheduc | -97.91025 68.25926 -1.43 0.154
-233.0941 37.27363
acoucsis | -34.77898 24.29715 -1.43 0.155
-82.89822 13.34025
gdppc | .207229 .1444896 1.43 0.154
-.0789252 .4933831
inst_hiv | 94.43629 65.86321 1.43 0.154
-36.00235 224.8749
imhat2 | -1.215096 .9065531 -1.34 0.183
-3.010477 .5802851
imhat3 | .004999 .0040124 1.25 0.215
-.0029473 .0129452
imhat4 | -7.53e-06 6.59e-06 -1.14 0.256
-.0000206 5.52e-06
_cons | 22725.4 15987.17 1.42 0.158
-8936.358 54387.16
-------------+----------------------------------------------------------------
sigma_u | 50.760469
sigma_e | 14.62366
rho | .92336383 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(41, 117) = 25.31 Prob
> F = 0.0000
. test imhat2 imhat3 imhat4
( 1) imhat2 = 0
( 2) imhat3 = 0
( 3) imhat4 = 0
F( 3, 117) = 1.59
Prob > F = 0.1952
. *test de la normalité des residu
Fixed-effects (within) regression Number of obs
= 168
Group variable (i): code Number of groups
= 42
R-sq: within = 0.2223 Obs per group:
min = 4
between = 0.4719
avg = 4.0
overall = 0.4331
max = 4
F(6,120)
= 5.72
corr(u_i, Xb) = 0.4522 Prob > F
= 0.0000
------------------------------------------------------------------------------
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | .4582889 .483487 0.95 0.345
-.4989816 1.415559
roug | -.2190717 .1265995 -1.73 0.086
-.46973 .0315865
motheduc | -.758119 .1926906 -3.93 0.000
-1.139633 -.3766051
acoucsis | -.2700784 .2746772 -0.98 0.327
-.8139202 .2737633
gdppc | .0016068 .0011999 1.34 0.183
-.000769 .0039825
pol_hiv | .7315258 .2602651 2.81 0.006
.216219 1.246833
_cons | 214.3106 16.12429 13.29 0.000
182.3856 246.2355
-------------+----------------------------------------------------------------
sigma_u | 48.831644
sigma_e | 15.056297
rho | .91318543 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(41, 120) = 24.08 Prob
> F = 0.0000
. sktest residu
Skewness/Kurtosis tests for Normality
------- joint
------
Variable | Pr(Skewness) Pr(Kurtosis) adj chi2(2)
Prob>chi2
-------------+-------------------------------------------------------
residu | 0.158 0.795 2.09
0.3518
. *test d'homoscedasticité
Source | SS df MS Number of
obs = 168
-------------+------------------------------ F( 6,
161) = 5747.05
Model | 5.0287e+09 6 838123558 Prob >
F = 0.0000
Residual | 23479518.5 161 145835.519 R-squared
= 0.9954
-------------+------------------------------ Adj
R-squared = 0.9952
Total | 5.0522e+09 167 30252819.6 Root MSE
= 381.88
------------------------------------------------------------------------------
residu2 | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | 152.3262 4.196293 36.30 0.000
144.0393 160.6131
roug | -72.04528 1.577225 -45.68 0.000
-75.16 -68.93057
motheduc | -239.285 2.374088 -100.79 0.000
-243.9733 -234.5966
acoucsis | -89.08242 2.001651 -44.50 0.000
-93.0353 -85.12954
gdppc | .5608018 .0132142 42.44 0.000
.5347062 .5868973
pol_hiv | 216.0009 3.262524 66.21 0.000
209.5581 222.4438
_cons | 43157.03 111.3058 387.73 0.000
42937.22 43376.83
------------------------------------------------------------------------------
. * correction de l'héteroscedasticité
Linear regression, absorbing indicators Number of
obs = 168
F( 6,
120) = 9.15
Prob >
F = 0.0000
R-squared
= 0.9544
Adj
R-squared = 0.9366
Root MSE
= 15.056
------------------------------------------------------------------------------
| Robust
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | .4582889 .7076847 0.65 0.518
-.9428775 1.859455
roug | -.2190717 .1132868 -1.93 0.055
-.4433716 .0052281
motheduc | -.758119 .2301168 -3.29 0.001
-1.213734 -.3025038
acoucsis | -.2700784 .3393422 -0.80 0.428
-.9419523 .4017954
gdppc | .0016068 .0005645 2.85 0.005
.0004891 .0027244
pol_hiv | .7315258 .2732748 2.68 0.008
.1904607 1.272591
_cons | 214.3106 18.94468 11.31 0.000
176.8014 251.8197
-------------+----------------------------------------------------------------
code | absorbed
(42 categories)
. *test d'endogenéité
Source | SS df MS Number of
obs = 164
-------------+------------------------------ F( 6,
157) = 23.02
Model | 5663.88005 6 943.980009 Prob >
F = 0.0000
Residual | 6437.58416 157 41.0037207 R-squared
= 0.4680
-------------+------------------------------ Adj
R-squared = 0.4477
Total | 12101.4642 163 74.2421117 Root MSE
= 6.4034
------------------------------------------------------------------------------
hiv | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
roug | .1241879 .0259414 4.79 0.000
.0729488 .1754269
motheduc | .0882306 .0403221 2.19 0.030
.0085869 .1678744
acoucsis | .0827753 .0342647 2.42 0.017
.015096 .1504546
muslim | -.0710005 .0180237 -3.94 0.000
-.1066007 -.0354004
vmr | (dropped)
gdppc | .0001004 .0002329 0.43 0.667
-.0003597 .0005605
pol_hiv | -.0177579 .0570255 -0.31 0.756
-.1303941 .0948782
_cons | -5.008685 2.142561 -2.34 0.021
-9.24065 -.7767212
------------------------------------------------------------------------------
Linear regression Number of
obs = 164
F( 7,
156) = 32.42
Prob >
F = 0.0000
R-squared
= 0.5322
Root MSE
= 39.378
------------------------------------------------------------------------------
| Robust
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | 2.838934 1.273781 2.23 0.027
.3228514 5.355017
roug | -.9051552 .2210186 -4.10 0.000
-1.341731 -.4685799
motheduc | -1.305785 .3443464 -3.79 0.000
-1.985968 -.6256019
acoucsis | -1.360461 .1929943 -7.05 0.000
-1.74168 -.9792416
vmr | (dropped)
gdppc | .0008388 .0014134 0.59 0.554
-.0019531 .0036307
pol_hiv | .3582408 .2420676 1.48 0.141
-.1199123 .8363939
r_resid | -2.246375 1.374993 -1.63 0.104
-4.962381 .4696315
_cons | 315.1202 14.83378 21.24 0.000
285.8192 344.4211
------------------------------------------------------------------------------
. *test de la validité des instruments
Instrumental variables (2SLS) regression
Source | SS df MS Number of
obs = 164
-------------+------------------------------ F( 6,
157) = 26.02
Model | 242659.399 6 40443.2332 Prob >
F = 0.0000
Residual | 274383.552 157 1747.66594 R-squared
= 0.4693
-------------+------------------------------ Adj
R-squared = 0.4490
Total | 517042.951 163 3172.04265 Root MSE
= 41.805
------------------------------------------------------------------------------
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | 2.838934 1.657293 1.71 0.089
-.4345333 6.112402
roug | -.9051552 .2777392 -3.26 0.001
-1.453743 -.3565677
motheduc | -1.305785 .3365818 -3.88 0.000
-1.970598 -.6409724
acoucsis | -1.360461 .2478415 -5.49 0.000
-1.849995 -.8709269
gdppc | .0008388 .0015474 0.54 0.589
-.0022175 .0038952
inst_hiv | .3582408 .3820488 0.94 0.350
-.3963779 1.112859
_cons | 315.1202 19.60188 16.08 0.000
276.4027 353.8376
------------------------------------------------------------------------------
Instrumented: hiv
Instruments: roug motheduc acoucsis gdppc pol_hiv muslim
------------------------------------------------------------------------------
Linear regression Number of
obs = 164
F( 7,
156) = 4.94
Prob >
F = 0.0000
R-squared
= 0.1184
Root MSE
= 39.378
------------------------------------------------------------------------------
| Robust
epsilon | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | -2.246375 .3961061 -5.67 0.000
-3.028798 -1.463951
roug | .2789725 .1824796 1.53 0.128
-.0814771 .6394221
motheduc | .1981991 .2780586 0.71 0.477
-.3510466 .7474448
acoucsis | .1859444 .2136245 0.87 0.385
-.2360254 .6079142
gdppc | .0002254 .00141 0.16 0.873
-.0025597 .0030105
inst_hiv | -.0398909 .2364988 -0.17 0.866
-.5070441 .4272622
muslim | -.1594938 .0976252 -1.63 0.104
-.3523317 .0333441
vmr | (dropped)
_cons | -11.25139 12.20875 -0.92 0.358
-35.36717 12.8644
------------------------------------------------------------------------------
. *prise en compte des variables muettes régionales
Linear regression Number of
obs = 168
F( 10,
157) = 37.38
Prob >
F = 0.0000
R-squared
= 0.6539
Root MSE
= 36.269
------------------------------------------------------------------------------
| Robust
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | 1.252978 .3862217 3.24 0.001
.4901176 2.015839
roug | -.5128926 .1726685 -2.97 0.003
-.8539455 -.1718397
motheduc | -1.077929 .2407526 -4.48 0.000
-1.553461 -.6023969
acoucsis | -1.248732 .2066431 -6.04 0.000
-1.656891 -.840573
gdppc | .0004584 .00164 0.28 0.780
-.0027809 .0036977
water | -.2164926 .1871605 -1.16 0.249
-.5861701 .1531848
inst_hiv | -.0017336 .2349717 -0.01 0.994
-.4658473 .46238
west_afr | 46.79417 7.859004 5.95 0.000
31.27116 62.31719
east_afr | (dropped)
central_afr | 36.99559 9.076654 4.08 0.000
19.06748 54.9237
southern_afr | 31.68599 9.474774 3.34 0.001
12.97152 50.40046
_cons | 269.3585 13.57728 19.84 0.000
242.5408 296.1762
------------------------------------------------------------------------------
Linear regression Number of
obs = 168
F( 7,
160) = 38.07
Prob >
F = 0.0000
R-squared
= 0.5794
Root MSE
= 39.604
------------------------------------------------------------------------------
| Robust
im | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
-------------+----------------------------------------------------------------
hiv | 1.136047 .3416004 3.33 0.001
.4614199 1.810674
roug | -.7300074 .1727496 -4.23 0.000
-1.071171 -.3888439
motheduc | -.9940563 .2556647 -3.89 0.000
-1.498969 -.4891437
acoucsis | -1.304065 .2093723 -6.23 0.000
-1.717554 -.8905749
gdppc | .0004228 .0016705 0.25 0.801
-.0028764 .0037219
inst_hiv | .2402728 .2358977 1.02 0.310
-.2256019 .7061476
war | 2.102814 10.73026 0.20 0.845
-19.08839 23.29402
_cons | 302.304 12.23743 24.70 0.000
278.1363 326.4718
------------------------------------------------------------------------------
| 
