Saad-Ellah Berhili
Plan de la présentation
I. Commandes de base
II. Modèles linéaires
III. Modèles à variables instrumentales
IV. Modèles à variables dépendantes limitées
V. Les commandes indispensables
2
I. Commandes de base
Préparation des données
Augmenter la mémoire allouée par Stata:
set mem 1000m (parfois plus!)
Ouvrir ou importer le fichier de données:
use "Donnees college [Link]", clear
Les données doivent être organisées selon le format
« long » : chaque observation (ligne) correspond à la paire
(individu , temps).
Nécessité de changer de format avec la commande
reshape long prop_f age_moy anc_moy eleve_clas eleve_ens
pCOL_xt pCOL_rt trCOL, i(cd_etab) j(annee)
Créer un identifiant pour pouvoir utiliser les données de panel
egen id=group(cd_etab)
Déclarer les données comme étant des données de panel (afin de
pouvoir utiliser les commandes xt):
xtset id annee 4
Description des données
. describe
Contains data from J:\Cours de données de panel\Séances\Séance 3\Panel college [Link]
obs: 24,204
vars: 13 7 Jan 2015 14:02
size: 1,936,320
storage display value
variable name type format label variable label
id float %9.0g group(cd_etab)
cd_etab str6 %9s code établissement
annee int %9.0g Année
region str32 %32s libellé région
milieu str6 %9s libellé milieu
prop_f double %10.0g Proportion des enseignantes
age_moy byte %8.0g Age moyen des enseignants
anc_moy byte %8.0g Nombre moyen d'années d'ancienneté des enseignants
eleve_clas float %9.0g ratio élèves classe collégial corrigé
eleve_ens float %9.0g Ratio élèves enseignant au collégial
pCOL_xt float %9.0g
pCOL_rt float %9.0g
trCOL float %9.0g Taux de redoublement collégial total corrigé
Sorted by: cd_etab annee
Note: dataset has changed since last saved
5
Description des données de panel
. xtdescribe
. xtdescribe
id: 1, 2, ..., 2017 n = 2017
annee: 1999, 2000, ..., 2010 T = 12
Delta(annee) = 1 unit
Span(annee) = 12 periods
(id*annee uniquely identifies each observation)
Distribution of T_i: min 5% 25% 50% 75% 95% max
12 12 12 12 12 12 12
Freq. Percent Cum. Pattern
2017 100.00 100.00 111111111111
2017 100.00 XXXXXXXXXXXX
6
Variation within et between
. xtsum annee trCOL
. xtsum annee trCOL
Variable Mean Std. Dev. Min Max Observations
annee overall 2004.5 3.452124 1999 2010 N = 24204
between 0 2004.5 2004.5 n = 2017
within 3.452124 1999 2010 T = 12
trCOL overall .1630803 .0627874 0 .4 N = 15928
between .0533071 0 .4 n = 1935
within .0460275 -.1528303 .4432758 T = 8.23152
7
Courbes individuelles
. xtline trCOL if id<=10, overlay
Taux de redoublement collégial total corrigé
0 .1 .2 .3 .4
2000 2005 2010
Année
id = 1 id =2
id = 3 id =4
id = 5 id =6
id = 7 id =8
id = 9 id = 10
8
Autocorrélation d’ordre 1
. sort id annee
. correlate trCOL [Link] [Link] [Link] [Link]
. correlate trCOL [Link] [Link] [Link] [Link]
(obs=8930)
L. L2. L3. L4.
trCOL trCOL trCOL trCOL trCOL
trCOL
--. 1.0000
L1. 0.4745 1.0000
L2. 0.4356 0.5023 1.0000
L3. 0.4077 0.4551 0.4978 1.0000
L4. 0.3714 0.4194 0.4356 0.4974 1.0000
9
II. Modèles linéaires
Estimateur MCO sur des données
de panel (pooled)
Variables du modèle
Yit : taux de redoublement au collège (trCOL)
Xit : - taille moyenne de la classe (eleve_clas)
- proportion d’élèves à l’âge normal (pCOL_xt)
- nombre moyen d’années d’ancienneté
des enseignants (anc_moy)
- indicatrices des années 2000 à 2010 (t00 à t09)
La méthode des MCO suppose que
les erreurs sont indépendantes et homoscédastiques
et ne tient pas compte de l’hétérogénéité individuelle
non observée.
11
Estimateur MCO sur des données
de panel (pooled)
. regress trCOL eleve_clas pCOL_xt anc_moy t00 t01 t02 t03
t04 t05 t06 t07 t08 t09
Source SS df MS Number of obs = 15863
F( 13, 15849) = 159.24
Model 7.1766752 13 .552051938 Prob > F = 0.0000
Residual 54.9441365 15849 .003466726 R-squared = 0.1155
Adj R-squared = 0.1148
Total 62.1208117 15862 .003916329 Root MSE = .05888
trCOL Coef. Std. Err. t P>|t| [95% Conf. Interval]
eleve_clas .0013504 .0000788 17.13 0.000 .0011958 .0015049
pCOL_xt -.0751603 .0034994 -21.48 0.000 -.0820196 -.068301
anc_moy .0018575 .0000748 24.83 0.000 .0017108 .0020041
t00 .0006824 .0024761 0.28 0.783 -.0041711 .0055359
t01 .0045072 .0024436 1.84 0.065 -.0002825 .0092969
t02 -.00257 .0024054 -1.07 0.285 -.0072849 .0021449
t03 -.0082947 .0023783 -3.49 0.000 -.0129564 -.0036329
t04 -.0198441 .0023689 -8.38 0.000 -.0244875 -.0152008
t05 -.0245458 .0023527 -10.43 0.000 -.0291573 -.0199342
t06 -.0286816 .0023467 -12.22 0.000 -.0332814 -.0240818
t07 -.0318979 .0023391 -13.64 0.000 -.0364828 -.0273129
t08 -.0406548 .0023252 -17.48 0.000 -.0452125 -.0360971
t09 -.0318277 .0022896 -13.90 0.000 -.0363155 -.0273399
_cons .1524257 .0033804 45.09 0.000 .1457997 .1590517
12
Estimateur MCO avec des écarts-
types robustes
. regress trCOL eleve_clas pCOL_xt anc_moy t00 t01 t02 t03
t04 t05 t06 t07 t08 t09, vce(cluster id)
Linear regression Number of obs = 15863
F( 13, 1915) = 76.16
Prob > F = 0.0000
R-squared = 0.1155
Root MSE = .05888
(Std. Err. adjusted for 1916 clusters in id)
Robust
trCOL Coef. Std. Err. t P>|t| [95% Conf. Interval]
eleve_clas .0013504 .0001548 8.72 0.000 .0010468 .001654
pCOL_xt -.0751603 .0092392 -8.13 0.000 -.0932803 -.0570403
anc_moy .0018575 .0001369 13.57 0.000 .0015889 .002126
t00 .0006824 .0017414 0.39 0.695 -.0027329 .0040977
t01 .0045072 .002016 2.24 0.025 .0005533 .0084611
t02 -.00257 .002001 -1.28 0.199 -.0064943 .0013544
t03 -.0082947 .0020711 -4.01 0.000 -.0123564 -.0042329
t04 -.0198441 .0019835 -10.00 0.000 -.0237343 -.015954
t05 -.0245458 .0021105 -11.63 0.000 -.0286849 -.0204066
t06 -.0286816 .0022082 -12.99 0.000 -.0330124 -.0243508
t07 -.0318979 .0022995 -13.87 0.000 -.0364076 -.0273882
t08 -.0406548 .002326 -17.48 0.000 -.0452165 -.0360931
t09 -.0318277 .0022105 -14.40 0.000 -.0361629 -.0274925
_cons .1524257 .0066578 22.89 0.000 .1393683 .165483
13
Estimateur between (BE)
. xtreg trCOL eleve_clas pCOL_xt anc_moy t00 t01 t02 t03 t04
t05 t06 t07 t08 t09, be
Between regression (regression on group means) Number of obs = 15863
Group variable: id Number of groups = 1916
R-sq: within = 0.0010 Obs per group: min = 1
between = 0.1962 avg = 8.3
overall = 0.0097 max = 11
F(13,1902) = 35.71
sd(u_i + avg(e_i.))= .0467003 Prob > F = 0.0000
trCOL Coef. Std. Err. t P>|t| [95% Conf. Interval]
eleve_clas .0018417 .0002037 9.04 0.000 .0014422 .0022411
pCOL_xt -.0839307 .0097849 -8.58 0.000 -.1031211 -.0647403
anc_moy .002452 .000192 12.77 0.000 .0020755 .0028285
t00 .1269393 .0358529 3.54 0.000 .0566241 .1972545
t01 -.0098389 .0250204 -0.39 0.694 -.0589092 .0392314
t02 -.0302448 .0270856 -1.12 0.264 -.0833655 .0228758
t03 .247742 .0352775 7.02 0.000 .1785553 .3169286
t04 .0577634 .030024 1.92 0.055 -.0011199 .1166468
t05 -.0193436 .0256454 -0.75 0.451 -.0696396 .0309524
t06 .0538775 .0261423 2.06 0.039 .002607 .105148
t07 .0478783 .023445 2.04 0.041 .0018977 .0938588
t08 .0030674 .0212807 0.14 0.885 -.0386685 .0448033
t09 .042155 .0195468 2.16 0.031 .0038196 .0804903
_cons .0619752 .0206747 3.00 0.003 .0214277 .1025227
14
Estimateur à effets aléatoires (RE)
. xtreg trCOL eleve_clas pCOL_xt anc_moy t00 t01 t02 t03 t04
t05 t06 t07 t08 t09, re vce(robust) theta
Random-effects GLS regression Number of obs = 15863
Group variable: id Number of groups = 1916
R-sq: within = 0.0574 Obs per group: min = 1
between = 0.1207 avg = 8.3
overall = 0.0998 max = 11
Wald chi2(13) = 721.77
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
theta
min 5% median 95% max
0.2451 0.2451 0.6721 0.6721 0.6721
(Std. Err. adjusted for 1916 clusters in id)
Robust
trCOL Coef. Std. Err. z P>|z| [95% Conf. Interval]
eleve_clas .0005121 .0001319 3.88 0.000 .0002536 .0007705
pCOL_xt -.0249118 .0053082 -4.69 0.000 -.0353157 -.0145078
anc_moy .0010085 .0001138 8.86 0.000 .0007854 .0012315
t00 .0014406 .0016393 0.88 0.380 -.0017723 .0046535
t01 .0070158 .0018876 3.72 0.000 .0033162 .0107154
t02 .0022494 .0018984 1.18 0.236 -.0014715 .0059702
t03 -.0035937 .0019574 -1.84 0.066 -.0074301 .0002428
t04 -.0132394 .0018823 -7.03 0.000 -.0169287 -.00955
t05 -.0189963 .0019758 -9.61 0.000 -.0228688 -.0151238
t06 -.0230257 .0020404 -11.28 0.000 -.0270248 -.0190266
t07 -.0251904 .0021177 -11.90 0.000 -.0293409 -.0210399
t08 -.0348515 .0021721 -16.05 0.000 -.0391086 -.0305943
t09 -.0246836 .0020455 -12.07 0.000 -.0286928 -.0206745
_cons .1552498 .0051054 30.41 0.000 .1452434 .1652563
sigma_u .04135501
sigma_e
rho
.0476047
.43009198 (fraction of variance due to u_i) 15
Estimateur à effets fixes (FE)
. xtreg trCOL eleve_clas pCOL_xt anc_moy t00 t01 t02 t03 t04
t05 t06 t07 t08 t09, fe vce(robust)
Fixed-effects (within) regression Number of obs = 15863
Group variable: id Number of groups = 1916
R-sq: within = 0.0605 Obs per group: min = 1
between = 0.0601 avg = 8.3
overall = 0.0656 max = 11
F(13,1915) = 43.55
corr(u_i, Xb) = 0.0926 Prob > F = 0.0000
(Std. Err. adjusted for 1916 clusters in id)
Robust
trCOL Coef. Std. Err. t P>|t| [95% Conf. Interval]
eleve_clas .0000309 .0001565 0.20 0.843 -.0002761 .0003379
pCOL_xt -.0158546 .0054434 -2.91 0.004 -.0265303 -.005179
anc_moy .0002145 .0001561 1.37 0.170 -.0000917 .0005206
t00 .0017555 .0016213 1.08 0.279 -.0014243 .0049353
t01 .0084844 .0018784 4.52 0.000 .0048005 .0121684
t02 .0046731 .0019276 2.42 0.015 .0008927 .0084536
t03 -.0010599 .0019884 -0.53 0.594 -.0049595 .0028396
t04 -.0088046 .0019998 -4.40 0.000 -.0127266 -.0048826
t05 -.0151321 .0020545 -7.37 0.000 -.0191615 -.0111028
t06 -.0188649 .0021392 -8.82 0.000 -.0230604 -.0146695
t07 -.0201419 .0022524 -8.94 0.000 -.0245594 -.0157245
t08 -.0294763 .0023541 -12.52 0.000 -.0340932 -.0248593
t09 -.0188154 .0022131 -8.50 0.000 -.0231557 -.0144751
_cons .1783407 .0059316 30.07 0.000 .1667076 .1899739
sigma_u .0506961
sigma_e .0476047
rho .5314173 (fraction of variance due to u_i) 16
Estimateur en différences
premières (FD)
. reg D.(trCOL eleve_clas pCOL_xt anc_moy t00 t01 t02 t03
t04 t05 t06 t07 t08 t09), noconstant vce(cluster id)
Linear regression Number of obs = 13930
F( 13, 1776) = 31.02
Prob > F = 0.0000
R-squared = 0.0165
Root MSE = .06179
(Std. Err. adjusted for 1777 clusters in id)
Robust
[Link] Coef. Std. Err. t P>|t| [95% Conf. Interval]
eleve_clas
D1. -.000593 .0002007 -2.95 0.003 -.0009866 -.0001993
pCOL_xt
D1. .0266515 .0057948 4.60 0.000 .0152861 .0380169
anc_moy
D1. .0001294 .000173 0.75 0.455 -.00021 .0004688
t00
D1. .0026476 .0016004 1.65 0.098 -.0004913 .0057865
t01
D1. .0093361 .0018627 5.01 0.000 .0056827 .0129895
t02
D1. .006807 .0020602 3.30 0.001 .0027662 .0108477
t03
D1. .0005984 .0022323 0.27 0.789 -.0037798 .0049766
t04
D1. -.0075412 .0024314 -3.10 0.002 -.0123099 -.0027724
t05
D1. -.0151721 .0025477 -5.96 0.000 -.0201689 -.0101754
t06
D1. -.0196973 .0026587 -7.41 0.000 -.0249119 -.0144827
t07
D1. -.0213714 .0027941 -7.65 0.000 -.0268515 -.0158913
t08
D1. -.0331748 .0029741 -11.15 0.000 -.0390079 -.0273417
t09
D1. -.0228282 .0028729 -7.95 0.000 -.0284627 -.0171936 17
Comparaison des estimateurs
• global xvar eleve_clas pCOL_xt anc_moy t00 t01 t02 t03 t04
t05 t06 t07 t08 t09
• quietly regress trCOL $xvar, vce(cluster id)
• estimates store MCO
• quietly xtreg trCOL $xvar, be
• estimates store BE
• quietly xtreg trCOL $xvar, re vce(robust)
• estimates store RE
• quietly xtreg trCOL $xvar, fe vce(robust)
• estimates store FE
• estimates table MCO BE RE FE, b(%9.4f) se stats(N)
18
Comparaison des estimateurs
Variable MCO BE RE FE
eleve_clas 0.0014 0.0018 0.0005 0.0000
0.0002 0.0002 0.0001 0.0002
pCOL_xt -0.0752 -0.0839 -0.0249 -0.0159
0.0092 0.0098 0.0053 0.0054
anc_moy 0.0019 0.0025 0.0010 0.0002
0.0001 0.0002 0.0001 0.0002
t00 0.0007 0.1269 0.0014 0.0018
0.0017 0.0359 0.0016 0.0016
t01 0.0045 -0.0098 0.0070 0.0085
0.0020 0.0250 0.0019 0.0019
t02 -0.0026 -0.0302 0.0022 0.0047
0.0020 0.0271 0.0019 0.0019
t03 -0.0083 0.2477 -0.0036 -0.0011
0.0021 0.0353 0.0020 0.0020
t04 -0.0198 0.0578 -0.0132 -0.0088
0.0020 0.0300 0.0019 0.0020
t05 -0.0245 -0.0193 -0.0190 -0.0151
0.0021 0.0256 0.0020 0.0021
t06 -0.0287 0.0539 -0.0230 -0.0189
0.0022 0.0261 0.0020 0.0021
t07 -0.0319 0.0479 -0.0252 -0.0201
0.0023 0.0234 0.0021 0.0023
t08 -0.0407 0.0031 -0.0349 -0.0295
0.0023 0.0213 0.0022 0.0024
t09 -0.0318 0.0422 -0.0247 -0.0188
0.0022 0.0195 0.0020 0.0022
_cons 0.1524 0.0620 0.1552 0.1783
0.0067 0.0207 0.0051 0.0059
N 15863 15863 15863 15863
19
legend: b/se
Effets fixes ou effets aléatoires ?
Comment choisir entre les deux estimateurs RE
et FE ?
Test de Hausman
• quietly xtreg trCOL $xvar, fe
• estimates store fe
• quietly xtreg trCOL $xvar, re
• estimates store re
• hausman fe re
On rejette H0 : l’estimateur FE est recommandé (ou
FD).
20
III. Modèles à variables
instrumentales
Description des données
Nom: PSID
Se trouve sur le site web de Stata sous le nom de
fichier : « [Link] »
Source : Baltagi et Khanti-Akom (1990)
595 individus pour les années 1976-82
But : estimer l’effet de l’éducation sur les salaires
Complication: l’éducation est invariante dans le
temps dans ces données
Utiliser les méthodes de variables instrumentales
(Hausman-Taylor). 22
Modèle à effets fixes (FE)
. use [Link], clear
. xtreg lwage exp exp2 wks ed, fe
Fixed-effects (within) regression Number of obs = 4165
Group variable: id Number of groups = 595
R-sq: within = 0.6566 Obs per group: min = 7
between = 0.0276 avg = 7.0
overall = 0.0476 max = 7
F(3,3567) = 2273.74
corr(u_i, Xb) = -0.9107 Prob > F = 0.0000
lwage Coef. Std. Err. t P>|t| [95% Conf. Interval]
exp .1137879 .0024689 46.09 0.000 .1089473 .1186284
exp2 -.0004244 .0000546 -7.77 0.000 -.0005315 -.0003173
wks .0008359 .0005997 1.39 0.163 -.0003399 .0020116
ed (dropped)
_cons 4.596396 .0389061 118.14 0.000 4.520116 4.672677
sigma_u 1.0362039
sigma_e .15220316
rho .97888036 (fraction of variance due to u_i)
F test that all u_i=0: F(594, 3567) = 40.17 Prob > F = 0.0000 23
Modèle à variables instrumentales
de Hausman-Taylor (HT)
. xthtaylor lwage occ south smsa ind exp exp2 wks ms union
fem blk ed, endog(exp exp2 wks ms union ed)
Hausman-Taylor estimation Number of obs = 4165
Group variable: id Number of groups = 595
Obs per group: min = 7
avg = 7
max = 7
Random effects u_i ~ i.i.d. Wald chi2(12) = 6891.87
Prob > chi2 = 0.0000
lwage Coef. Std. Err. z P>|z| [95% Conf. Interval]
TVexogenous
occ -.0207047 .0137809 -1.50 0.133 -.0477149 .0063055
south .0074398 .031955 0.23 0.816 -.0551908 .0700705
smsa -.0418334 .0189581 -2.21 0.027 -.0789906 -.0046761
ind .0136039 .0152374 0.89 0.372 -.0162608 .0434686
TVendogenous
exp .1131328 .002471 45.79 0.000 .1082898 .1179758
exp2 -.0004189 .0000546 -7.67 0.000 -.0005259 -.0003119
wks .0008374 .0005997 1.40 0.163 -.0003381 .0020129
ms -.0298508 .01898 -1.57 0.116 -.0670508 .0073493
union .0327714 .0149084 2.20 0.028 .0035514 .0619914
TIexogenous
fem -.1309236 .126659 -1.03 0.301 -.3791707 .1173234
blk -.2857479 .1557019 -1.84 0.066 -.5909179 .0194221
TIendogenous
ed .137944 .0212485 6.49 0.000 .0962977 .1795902
_cons 2.912726 .2836522 10.27 0.000 2.356778 3.468674
sigma_u .94180304
sigma_e .15180273
rho .97467788 (fraction of variance due to u_i)
Note: TV refers to time varying; TI refers to time invariant.
24
Modèle à variables instrumentales
dynamiques de Arellano-Bond (AB)
. xtabond lwage occ south smsa ind, lags(2) maxldep(3) pre(wks,lag(1,2))
endogenous(ms,lag(0,2)) endogenous(union,lag(0,2)) twostep vce(robust) artests(3)
Arellano-Bond dynamic panel-data estimation Number of obs = 2380
Group variable: id Number of groups = 595
Time variable: t
Obs per group: min = 4
avg = 4
max = 4
Number of instruments = 40 Wald chi2(10) = 1287.77
Prob > chi2 = 0.0000
Two-step results
WC-Robust
lwage Coef. Std. Err. z P>|z| [95% Conf. Interval]
lwage
L1. .611753 .0373491 16.38 0.000 .5385501 .6849559
L2. .2409058 .0319939 7.53 0.000 .1781989 .3036127
wks
--. -.0159751 .0082523 -1.94 0.053 -.0321493 .000199
L1. .0039944 .0027425 1.46 0.145 -.0013807 .0093695
ms .1859324 .144458 1.29 0.198 -.0972 .4690649
union -.1531329 .1677842 -0.91 0.361 -.4819839 .1757181
occ -.0357509 .0347705 -1.03 0.304 -.1038999 .032398
south -.0250368 .2150806 -0.12 0.907 -.446587 .3965134
smsa -.0848223 .0525243 -1.61 0.106 -.187768 .0181235
ind .0227008 .0424207 0.54 0.593 -.0604422 .1058437
_cons 1.639999 .4981019 3.29 0.001 .6637377 2.616261
Instruments for differenced equation
GMM-type: L(2/4).lwage L(1/2).[Link] L(2/3).ms L(2/3).union
Standard: [Link] [Link] [Link] [Link]
Instruments for level equation
Standard: _cons 25
IV. Modèles à variables
dépendantes limitées
26
Description des données
Nom: Rand Health Insurance Experiment data
Source : "The Structure of Demand for Medical Care:
Latent Class versus Two-Part Models", Journal of
Health Economics, 21, 601-625
Chaque observation est une mesure d’un individu en
une année.
Les individus peuvent apparaître dans au plus 5
années.
27
Modèle logit
. use [Link], clear
. xtset id year
. xtdescribe
. xtsum age lfam child
. xtsum dmdu
. logit dmdu lcoins ndisease female age lfam child, vce(cluster
id) nolog
. xtlogit dmdu lcoins ndisease female age lfam child, pa
corr(exch) vce(robust) nolog
. xtlogit dmdu lcoins ndisease female age lfam child, re nolog
. xtlogit dmdu lcoins ndisease female age lfam child, fe nolog
28
Modèle tobit à effets aléatoires
. xtsum med
. xttobit med lcoins ndisease female age lfam child, ll(0) nolog
Random-effects tobit regression Number of obs = 20186
Group variable: id Number of groups = 5908
Random effects u_i ~ Gaussian Obs per group: min = 1
avg = 3.4
max = 5
Wald chi2(6) = 573.45
Log likelihood = -130030.45 Prob > chi2 = 0.0000
med Coef. Std. Err. z P>|z| [95% Conf. Interval]
lcoins -31.10247 3.578498 -8.69 0.000 -38.1162 -24.08875
ndisease 13.49452 1.139156 11.85 0.000 11.26182 15.72722
female 60.10112 14.95966 4.02 0.000 30.78072 89.42152
age 4.075582 .7238253 5.63 0.000 2.656911 5.494254
lfam -57.75023 14.68422 -3.93 0.000 -86.53077 -28.96968
child -52.02314 24.21619 -2.15 0.032 -99.48599 -4.560284
_cons -98.27203 36.05977 -2.73 0.006 -168.9479 -27.59618
/sigma_u 371.3134 8.64634 42.94 0.000 354.3668 388.2599
/sigma_e 715.1779 4.704581 152.02 0.000 705.9571 724.3987
rho .2123246 .0086583 .1957541 .2296872
Observation summary: 4453 left-censored observations
15733 uncensored observations
0 right-censored observations 29
V. Les commandes
indispensables
Les commandes indispensables
des données de panel
Préparation et description des données de panel : xtset, xtdescribe,
xtsum, xttab, xtline
MCO pooled : regress
Effets aléatoires : xtreg, re
Effets fixes : xtreg, fe
Différences premières : regress (sur les données différencées)
Variables instrumentales : xtivreg, xthtaylor
GMM : xtabond, xtdpdsys, xtdpd
Test de Hausman : hausman
Modèles non linéaires pooled : poisson, nbreg, logit, probit, tobit
Modèles non linéaires : xtpoisson, xtnbreg, xtlogit, xtprobit, xttobit
Tests de racine unitaire : levinlin, ipshin, hadrilm
31
Merci !!!
32
