Annexes n°3 : Test de spécification de
Hausman généralisé pour la variable KH
local id pays
. local depvar kh
. local varlist ide dpibh ddurban
. xtivreg `depvar' `varlist' (ide = dcredit dpibh ddurban dtchang
didf ouv export interet epargne), re
G2SLS random-effects IV regression Number of obs = 189
Group variable: pays Number of groups = 7
R-sq: within = 0.2649 Obs per group: min = 27
between = 0.6452 avg = 27.0
overall = 0.2044 max = 27
Wald chi2(3) = 64.84
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
kh | Coef. Std. Err. z P>|z| [95% Conf. Interval]
ide | 3.931163 .6467688 6.08 0.000 2.66352 5.198807
ide | (omitted)
dpibh | .053091 .0151452 3.51 0.000 .0234069 .0827751
ddurban | 62.03374 19.49829 3.18 0.001 23.81779 100.2497
_cons | 56.49652 6.650325 8.50 0.000 43.46212 69.53092
sigma_u | 17.147059
sigma_e | 12.172647
rho | .664914 (fraction of variance due to u_i)
Instrumented: ide
Instruments: ide dpibh ddurban dcredit dtchang didf ouv export
interet
epargne
. hausman, save
You used the old syntax of hausman. Click here to learn about the
new syntax.
(storing estimation results as _HAUSMAN)
.
. xtivreg `depvar' `varlist' (ide
export interet epargne), fe
|
= dcredit dpibh ddurban
|
dtchang didf ouv
|
Fixed-effects (within) IV
|
regression
|
|
Number of obs
|
=
|
189
|
Group variable: pays
|
|
|
Number of groups
|
=
|
7
|
R-sq: within = 0.2650
|
|
|
Obs per group: min
|
=
|
27
|
between = 0.6438
|
|
|
avg
|
=
|
27.0
|
overall = 0.2036
|
|
|
max
|
=
|
27
|
|
|
|
Wald chi2(3)
|
=
|
4927.67
|
corr(u_i, Xb) = 0.2114
|
|
|
Prob > chi2
|
=
|
0.0000
|
kh | Coef.
|
Std. Err.
|
z
|
P>|z| [95%
|
Conf.
|
Interval]
|
+
|
|
|
|
|
|
ide | 3.861672 .641076 6.02 0.000 2.605186 5.118158
ide | (omitted)
dpibh | .0533171 .0150001 3.55 0.000 .0239175 .0827166
ddurban | 61.36754 19.31573 3.18 0.001 23.50942 99.22567
_cons | 56.5773 1.170693 48.33 0.000 54.28278 58.87182
sigma_u | 21.955144
sigma_e | 12.172647
rho | .76487949 (fraction of variance due to u_i)
F test that all u_i=0: F(6,179) = 82.15 Prob > F = 0.0000
Instrumented: ide
Instruments: ide dpibh ddurban dcredit dtchang didf ouv export
interet
epargne
. hausman, less
You used the old syntax of hausman. Click here to learn about the
new syntax.
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| Consistent Efficient Difference S.E.
+
ide | 3.861672 3.931163 -.0694916 .
dpibh | .0533171 .053091 .0002261 .
ddurban | 61.36754 62.03374 -.6661989 .
R-sq: within = 0.4811 Obs per group: min
between = 0.7875 avg
overall = 0.6450 max
=
|
189
|
=
|
7
|
=
|
27
|
=
|
27.0
|
=
|
27
|
. xtreg `depvar' `newlist' , re
Random-effects GLS regression Number of obs
Group variable: pays Number of groups
b = consistent under Ho and Ha; obtained from xtivreg B =
inconsistent under Ha, efficient under Ho; obtained from xtivreg
Test: Ho: difference in coefficients not systematic
chi2(3) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= -0.72 chi2<0 ==> model fitted on these
data fails to meet the asymptotic assumptions of the Hausman
test; see suest for a generalized test
.
. tokenize `varlist'
. local i 1
. while "``i''" != "" {
2. qui by `id': gen double mean`i' = sum(``i'') / _n
3. qui by `id': replace mean`i' = mean`i'[_n]
4. qui by `id': gen double diff`i' = ``i'' - mean`i'
5. local newlist `newlist' mean`i' diff`i'
6.
. local i = `i' + 1
7. }
.
Random effects u_i ~ Gaussian Wald chi2(6) = 330.62
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
kh | Coef. Std. Err. z P>|z| [95% Conf. Interval]
mean1 |
|
38.59088
|
diff1 |
|
.0420427
|
mean2 |
|
.0425622
|
diff2 |
|
.0153781
|
mean3 |
|
255.0256
|
diff3 |
|
34.98939
|
_cons |
|
35.64351
|
sigma_u |
|
0
|
sigma_e |
|
10.086122
|
rho |
|
0
|
2.52265 15.30 0.000 33.64658 43.53519
.9487875 0.04 0.965 -1.817546 1.901632
.0607589 0.70 0.484 -.076523 .1616474
.0210159 0.73 0.464 -.0258124 .0565686
108.7919 2.34 0.019 41.79737 468.2539
26.75798 1.31 0.191 -17.45529 87.43407
2.58779 13.77 0.000 30.57154 40.71549
(fraction of variance due to u_i)
. tempname b
. matrix `b' = e(b)
.
. qui test mean1 = mean1 , notest /* clear test */
. local i 2
. while "``i''" != "" {
2. if `b'[1 , colnumb (`b', "mean`i'")] != 0 & /*
> */ `b'[ 1 , colnumb (`b', "diff`i'") ] != 0 {
3. qui test mean`i' = diff`i' , accum notest
4. }
5. local i = `i' + 1
6. }
.
. test
( 1) = 0
( 2) mean2 - diff2 = 0
( 3) mean3 - diff3 = 0 Constraint 1 dropped
chi2( 2) = 3.65
Prob > chi2 = 0.1614
20
Burkina Faso
Bénin
Côte d'Ivoire
10
-10 0
-20
Mali
20
Niger
Sénégal
10
-10 0
-20
1980 1990 2000 2010 1980 1990 2000 2010
20 Togo
-10 0
-20
10
1980 1990 2000 2010
Années
IDE Cr
Graphique par Pays
|