Décomposition des inégalités salariales selon le sexe par calage : application à l’Enquête suisse sur la structure des salaires
Section 7. Conclusion

Le phénomène de discrimination présente de multiples facettes et peut être produit par de nombreux mécanismes. Néanmoins, nous n’examinons ici son estimation que d’un point de vue méthodologique. Les deux approches de calage prises en considération représentent une généralisation de deux méthodes de décomposition existantes, à savoir la méthode de Blinder (1973) et Oaxaca (1973) et la méthode semi-paramétrique de DiNardo et coll. (1996), toutes deux exprimées en utilisant les poids de sondage. Les méthodes originales peuvent aussi être obtenues si tous les poids de sondage sont considérés comme égaux à 1. Le cas linéaire donne le même résultat que la méthode BO. Cependant, puisque les poids résultants ne sont pas bornés, des valeurs négatives sont parfois observées. Tout comme la méthode DFL, l’approche de calage permet de décomposer les écarts salariaux à d’autres points que la moyenne, tels que les quantiles. Cependant, le calage par raking ratio représente une amélioration de la méthode DFL, en ce sens que l’estimation de l’effet de structure comprendra toujours un effet résiduel nul. Donc, l’effet de structure sera composé uniquement de l’effet pur. La décomposition des écarts salariaux le long des quantiles permet de conclure que, dans les emplois faiblement rémunérés, les inégalités sont dues uniquement à la discrimination. Dans le présent article, l’accent est mis sur la généralisation de deux méthodes de décomposition bien établies au moyen de l’approche de calage.

Remerciements

Les auteurs remercient l’Office fédéral de la statistique suisse de son soutien financier, et son département des salaires (LOHN) d’avoir fourni les données. Cependant, les opinions exprimées dans le présent article ne reflètent pas nécessairement celles de l’Office fédéral de la statistique suisse.

Annexe A

Preuve du résultat 1

X ^ h β ^ h = X ^ h ( k S h d k x k x k ) 1 l S h d l x l y l = j S h d j x j ( k S h d k x k x k ) 1 l S h d l x l y l = ( j S h ς d j x j x j ) ( k S h d k x k x k ) 1 l S h d l x l y l = l S h ς d l x l y l = l S h d l y l = Y ^ h . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaafaqaaeabcaaaaeaaceWHybGbaKaada WgaaWcbaGaamiAaaqabaGcceWHYoGbaKaadaWgaaWcbaGaamiAaaqa baaakeaacaaI9aGabCiwayaajaWaaSbaaSqaaiaadIgaaeqaaOWaae WaaeaadaaeqbqaaiaadsgadaWgaaWcbaGaam4AaaqabaGccaWH4bWa aSbaaSqaaiaadUgaaeqaaOGaaCiEamaaDaaaleaacaWGRbaabaWexL MBbXgBd9gzLbvyNv2CaeHbbjxAHXgiv5wAJ9gzLbsttbaceaGaa8Ne XaaaaeaacaWGRbGaeyicI4Saam4uamaaBaaameaacaWGObaabeaaaS qab0GaeyyeIuoaaOGaayjkaiaawMcaamaaCaaaleqabaGaeyOeI0Ia aGymaaaakmaaqafabaGaamizamaaBaaaleaacaWGSbaabeaakiaahI hadaWgaaWcbaGaamiBaaqabaGccaWG5bWaaSbaaSqaaiaadYgaaeqa aaqaaiaadYgacqGHiiIZcaWGtbWaaSbaaWqaaiaadIgaaeqaaaWcbe qdcqGHris5aaGcbaaabaGaaGypamaaqafabaGaamizamaaBaaaleaa caWGQbaabeaakiaahIhadaqhaaWcbaGaamOAaaqaaiaa=jrmaaaaba GaamOAaiabgIGiolaadofadaWgaaadbaGaamiAaaqabaaaleqaniab ggHiLdGcdaqadaqaamaaqafabaGaamizamaaBaaaleaacaWGRbaabe aakiaahIhadaWgaaWcbaGaam4AaaqabaGccaWH4bWaa0baaSqaaiaa dUgaaeaacaWFsedaaaqaaiaadUgacqGHiiIZcaWGtbWaaSbaaWqaai aadIgaaeqaaaWcbeqdcqGHris5aaGccaGLOaGaayzkaaWaaWbaaSqa beaacqGHsislcaaIXaaaaOWaaabuaeaacaWGKbWaaSbaaSqaaiaadY gaaeqaaOGaaCiEamaaBaaaleaacaWGSbaabeaakiaadMhadaWgaaWc baGaamiBaaqabaaabaGaamiBaiabgIGiolaadofadaWgaaadbaGaam iAaaqabaaaleqaniabggHiLdaakeaaaeaacaaI9aWaaeWaaeaadaae qbqaaiaahk8adaahaaWcbeqaaiaa=jrmaaGccaWGKbWaaSbaaSqaai aadQgaaeqaaOGaaCiEamaaBaaaleaacaWGQbaabeaakiaahIhadaqh aaWcbaGaamOAaaqaaiaa=jrmaaaabaGaamOAaiabgIGiolaadofada WgaaadbaGaamiAaaqabaaaleqaniabggHiLdaakiaawIcacaGLPaaa daqadaqaamaaqafabaGaamizamaaBaaaleaacaWGRbaabeaakiaahI hadaWgaaWcbaGaam4AaaqabaGccaWH4bWaa0baaSqaaiaadUgaaeaa caWFsedaaaqaaiaadUgacqGHiiIZcaWGtbWaaSbaaWqaaiaadIgaae qaaaWcbeqdcqGHris5aaGccaGLOaGaayzkaaWaaWbaaSqabeaacqGH sislcaaIXaaaaOWaaabuaeaacaWGKbWaaSbaaSqaaiaadYgaaeqaaO GaaCiEamaaBaaaleaacaWGSbaabeaakiaadMhadaWgaaWcbaGaamiB aaqabaaabaGaamiBaiabgIGiolaadofadaWgaaadbaGaamiAaaqaba aaleqaniabggHiLdaakeaaaeaacaaI9aWaaabuaeaacaWHcpWaaWba aSqabeaacaWFsedaaOGaamizamaaBaaaleaacaWGSbaabeaakiaahI hadaWgaaWcbaGaamiBaaqabaGccaWG5bWaaSbaaSqaaiaadYgaaeqa aaqaaiaadYgacqGHiiIZcaWGtbWaaSbaaWqaaiaadIgaaeqaaaWcbe qdcqGHris5aOGaaGypamaaqafabaGaamizamaaBaaaleaacaWGSbaa beaakiaadMhadaWgaaWcbaGaamiBaaqabaaabaGaamiBaiabgIGiol aadofadaWgaaadbaGaamiAaaqabaaaleqaniabggHiLdGccaaI9aGa bmywayaajaWaaSbaaSqaaiaadIgaaeqaaOGaaGOlaaaaaaa@DC2E@

En divisant cette équation par k S h d k , MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaadaaeqaqaaiaadsgadaWgaaWcbaGaam 4AaaqabaaabaGaam4AaiabgIGiolaadofadaWgaaadbaGaamiAaaqa baaaleqaniabggHiLdGccaaISaaaaa@3ACD@ on obtient le résultat 1.

Annexe B

B.1  Linéarisation des moyennes

Afin de calculer la variance des moyennes et des moyennes contrefactuelles, nous avons utilisé la méthode de linéarisation proposée par Graf (2011). L’auteur propose de calculer la dérivée partielle de l’estimateur par rapport à l’indicateur d’échantillon. Cette dérivée fournit la variable linéarisée qui peut être insérée dans l’estimateur de variance. Les moyennes sont définies par :

Y ¯ ^ F = k S F d k y k k S F d k , MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaaceWGzbGbaeHbaKaadaWgaaWcbaGaam OraaqabaGccqGH9aqpdaWcaaqaamaaqababaGaamizamaaBaaaleaa caWGRbaabeaakiaadMhadaWgaaWcbaGaam4AaaqabaaabaGaam4Aai abgIGiolaadofadaWgaaadbaGaamOraaqabaaaleqaniabggHiLdaa keaadaaeqaqaaiaadsgadaWgaaWcbaGaam4AaaqabaaabaGaam4Aai abgIGiolaadofadaWgaaadbaGaamOraaqabaaaleqaniabggHiLdaa aOGaaiilaaaa@481A@

et

Y ¯ ^ M = l S M d l y l l S M d l . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaaceWGzbGbaeHbaKaadaWgaaWcbaGaam ytaaqabaGccqGH9aqpdaWcaaqaamaaqababaGaamizamaaBaaaleaa caWGSbaabeaakiaadMhadaWgaaWcbaGaamiBaaqabaaabaGaamiBai abgIGiolaadofadaWgaaadbaGaamytaaqabaaaleqaniabggHiLdaa keaadaaeqaqaaiaadsgadaWgaaWcbaGaamiBaaqabaaabaGaamiBai abgIGiolaadofadaWgaaadbaGaamytaaqabaaaleqaniabggHiLdaa aOGaaiOlaaaa@4836@

Pour les deux salaires moyens, nous obtenons les variables linéarisées :

Y ¯ ^ F I j = { d j ( y j Y ¯ ^ F ) k S F d k j S F 0 j S M , MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaadaWcaaqaaiabgkGi2kqadMfagaqega qcamaaBaaaleaacaWGgbaabeaaaOqaaiabgkGi2kaadMeadaWgaaWc baGaamOAaaqabaaaaOGaeyypa0ZaaiqaaeaafaqaaeGacaaabaWaaS aaaeaacaWGKbWaaSbaaSqaaiaadQgaaeqaaOWaaeWaaeaacaWG5bWa aSbaaSqaaiaadQgaaeqaaOGaeyOeI0IabmywayaaryaajaWaaSbaaS qaaiaadAeaaeqaaaGccaGLOaGaayzkaaaabaWaaabeaeaacaWGKbWa aSbaaSqaaiaadUgaaeqaaaqaaiaadUgacqGHiiIZcaWGtbWaaSbaaW qaaiaadAeaaeqaaaWcbeqdcqGHris5aaaaaOqaaiaadQgacqGHiiIZ caWGtbWaaSbaaSqaaiaadAeaaeqaaaGcbaGaaGimaaqaaiaadQgacq GHiiIZcaWGtbWaaSbaaSqaaiaad2eaaeqaaaaaaOGaay5EaaGaaiil aaaa@55BE@

et

Y ¯ ^ M I j = { d j ( y j Y ¯ ^ M ) l S M d l j S M 0 j S F . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaadaWcaaqaaiabgkGi2kqadMfagaqega qcamaaBaaaleaacaWGnbaabeaaaOqaaiabgkGi2kaadMeadaWgaaWc baGaamOAaaqabaaaaOGaeyypa0ZaaiqaaeaafaqaaeGacaaabaWaaS aaaeaacaWGKbWaaSbaaSqaaiaadQgaaeqaaOWaaeWaaeaacaWG5bWa aSbaaSqaaiaadQgaaeqaaOGaeyOeI0IabmywayaaryaajaWaaSbaaS qaaiaad2eaaeqaaaGccaGLOaGaayzkaaaabaWaaabeaeaacaWGKbWa aSbaaSqaaiaadYgaaeqaaaqaaiaadYgacqGHiiIZcaWGtbWaaSbaaW qaaiaad2eaaeqaaaWcbeqdcqGHris5aaaaaOqaaiaadQgacqGHiiIZ caWGtbWaaSbaaSqaaiaad2eaaeqaaaGcbaGaaGimaaqaaiaadQgacq GHiiIZcaWGtbWaaSbaaSqaaiaadAeaaeqaaaaaaOGaay5EaaGaaiOl aaaa@55D7@

B.2  Linéarisation de la moyenne contrefactuelle

Afin de calculer la moyenne contrefactuelle, nous calculons les poids v k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWG2bWaaSbaaSqaaiaadUgaaeqaaa aa@33D6@ définis par le système

A = k S F v k d k x k = k S F d k l S M d l l S M d l x l = X ¯ ^ M k S F d k , MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWHbbGaeyypa0ZaaabuaeaacaWG2b WaaSbaaSqaaiaadUgaaeqaaOGaamizamaaBaaaleaacaWGRbaabeaa kiaahIhadaWgaaWcbaGaam4AaaqabaGccqGH9aqpdaWcaaqaamaaqa babaGaamizamaaBaaaleaacaWGRbaabeaaaeaacaWGRbGaeyicI4Sa am4uamaaBaaameaacaWGgbaabeaaaSqab0GaeyyeIuoaaOqaamaaqa babaGaamizamaaBaaaleaacaWGSbaabeaaaeaacaWGSbGaeyicI4Sa am4uamaaBaaameaacaWGnbaabeaaaSqab0GaeyyeIuoaaaGcdaaeqb qaaiaadsgadaWgaaWcbaGaamiBaaqabaGccaWH4bWaaSbaaSqaaiaa dYgaaeqaaaqaaiaadYgacqGHiiIZcaWGtbWaaSbaaWqaaiaad2eaae qaaaWcbeqdcqGHris5aaWcbaGaam4AaiabgIGiolaadofadaWgaaad baGaamOraaqabaaaleqaniabggHiLdGccqGH9aqpceWHybGbaeHbaK aadaWgaaWcbaGaamytaaqabaGcdaaeqbqaaiaadsgadaWgaaWcbaGa am4AaaqabaGccaGGSaaaleaacaWGRbGaeyicI4Saam4uamaaBaaame aacaWGgbaabeaaaSqab0GaeyyeIuoaaaa@68D0@

avec

v k = F ( x k λ ) . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWG2bWaaSbaaSqaaiaadUgaaeqaaO Gaeyypa0JaamOramaabmaabaGaaCiEamaaDaaaleaacaWGRbaabaWe xLMBbXgBd9gzLbvyNv2CaeHbbjxAHXgiv5wAJ9gzLbsttbaceaaeaa aaaaaaa8qacaWFsedaaOWdaiaahU7aaiaawIcacaGLPaaacaGGUaaa aa@47C5@

Pour les variables linéarisées, nous avons considéré deux cas :

-     Si j S F MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWGQbGaeyicI4Saam4uamaaBaaale aacaWGgbaabeaaaaa@3601@

A I j = v j d j x j + [ k S F F ( x k λ ) d k x k x k ] λ I j = d j X ¯ ^ M . MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaadaWcaaqaaiabgkGi2kaahgeaaeaacq GHciITcaWGjbWaaSbaaSqaaiaadQgaaeqaaaaakiabg2da9iaadAha daWgaaWcbaGaamOAaaqabaGccaWGKbWaaSbaaSqaaiaadQgaaeqaaO GaaCiEamaaBaaaleaacaWGQbaabeaakiabgUcaRmaadmaabaWaaabu aeaacaWGgbWaaWbaaSqabeaajugybiadaITHYaIOaaGcdaqadaqaai aahIhadaqhaaWcbaGaam4AaaqaamXvP5wqSX2qVrwzqf2zLnharyqq YLwySbsvUL2yVrwzG00uaGabaiaa=jrmaaGccaWH7oaacaGLOaGaay zkaaGaamizamaaBaaaleaacaWGRbaabeaakiaahIhadaWgaaWcbaGa am4AaaqabaGccaWH4bWaa0baaSqaaiaadUgaaeaacaWFsedaaaqaai aadUgacqGHiiIZcaWGtbWaaSbaaWqaaiaadAeaaeqaaaWcbeqdcqGH ris5aaGccaGLBbGaayzxaaWaaSaaaeaacqGHciITcaWH7oaabaGaey OaIyRaamysamaaBaaaleaacaWGQbaabeaaaaGccqGH9aqpcaWGKbWa aSbaaSqaaiaadQgaaeqaaOGabCiwayaaryaajaWaaSbaaSqaaiaad2 eaaeqaaOGaaiOlaaaa@70E0@

       Donc,

λ I j = T 1 d j ( v j x j X ¯ ^ M ) , MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaadaWcaaqaaiabgkGi2kaahU7aaeaacq GHciITcaWGjbWaaSbaaSqaaiaadQgaaeqaaaaakiabg2da9iabgkHi TiaahsfadaahaaWcbeqaaiabgkHiTiaaigdaaaGccaWGKbWaaSbaaS qaaiaadQgaaeqaaOWaaeWaaeaacaWG2bWaaSbaaSqaaiaadQgaaeqa aOGaaCiEamaaBaaaleaacaWGQbaabeaakiabgkHiTiqahIfagaqega qcamaaBaaaleaacaWGnbaabeaaaOGaayjkaiaawMcaaiaacYcaaaa@480D@

       où

T = k S F F ( x k λ ) d k x k x k . MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWHubGaeyypa0ZaaabuaeaacaWGgb WaaWbaaSqabeaajugybiadaITHYaIOaaaaleaacaWGRbGaeyicI4Sa am4uamaaBaaameaacaWGgbaabeaaaSqab0GaeyyeIuoakmaabmaaba GaaCiEamaaDaaaleaacaWGRbaabaWexLMBbXgBd9gzLbvyNv2CaeHb bjxAHXgiv5wAJ9gzLbsttbaceaGaa8NeXaaakiaahU7aaiaawIcaca GLPaaacaWGKbWaaSbaaSqaaiaadUgaaeqaaOGaaCiEamaaBaaaleaa caWGRbaabeaakiaahIhadaqhaaWcbaGaam4Aaaqaaiaa=jrmaaGcca GGUaaaaa@57CD@

-     Si j S M MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWGQbGaeyicI4Saam4uamaaBaaale aacaWGnbaabeaaaaa@3608@

A I j = [ k S F F ( x k λ ) d k x k x k ] λ I j = d j ( x j X ¯ ^ M ) k S F d k l S M d l . MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaadaWcaaqaaiabgkGi2kaahgeaaeaacq GHciITcaWGjbWaaSbaaSqaaiaadQgaaeqaaaaakiabg2da9maadmaa baWaaabuaeaacaWGgbWaaWbaaSqabeaajugybiadaITHYaIOaaaale aacaWGRbGaeyicI4Saam4uamaaBaaameaacaWGgbaabeaaaSqab0Ga eyyeIuoakmaabmaabaGaaCiEamaaDaaaleaacaWGRbaabaWexLMBbX gBd9gzLbvyNv2CaeHbbjxAHXgiv5wAJ9gzLbsttbaceaGaa8NeXaaa kiaahU7aaiaawIcacaGLPaaacaWGKbWaaSbaaSqaaiaadUgaaeqaaO GaaCiEamaaBaaaleaacaWGRbaabeaakiaahIhadaqhaaWcbaGaam4A aaqaaiaa=jrmaaaakiaawUfacaGLDbaadaWcaaqaaiabgkGi2kaahU 7aaeaacqGHciITcaWGjbWaaSbaaSqaaiaadQgaaeqaaaaakiabg2da 9iaadsgadaWgaaWcbaGaamOAaaqabaGcdaqadaqaaiaahIhadaWgaa WcbaGaamOAaaqabaGccqGHsislceWHybGbaeHbaKaadaWgaaWcbaGa amytaaqabaaakiaawIcacaGLPaaadaWcaaqaamaaqababaGaamizam aaBaaaleaacaWGRbaabeaaaeaacaWGRbGaeyicI4Saam4uamaaBaaa meaacaWGgbaabeaaaSqab0GaeyyeIuoaaOqaamaaqababaGaamizam aaBaaaleaacaWGSbaabeaaaeaacaWGSbGaeyicI4Saam4uamaaBaaa meaacaWGnbaabeaaaSqab0GaeyyeIuoaaaGccaGGUaaaaa@7ED6@

       Donc,

λ I j = T 1 d j ( x j X ¯ ^ M ) k S F d k l S M d l . MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaadaWcaaqaaiabgkGi2kaahU7aaeaacq GHciITcaWGjbWaaSbaaSqaaiaadQgaaeqaaaaakiabg2da9iaahsfa daahaaWcbeqaaiabgkHiTiaaigdaaaGccaWGKbWaaSbaaSqaaiaadQ gaaeqaaOWaaeWaaeaacaWH4bWaaSbaaSqaaiaadQgaaeqaaOGaeyOe I0IabCiwayaaryaajaWaaSbaaSqaaiaad2eaaeqaaaGccaGLOaGaay zkaaWaaSaaaeaadaaeqaqaaiaadsgadaWgaaWcbaGaam4Aaaqabaaa baGaam4AaiabgIGiolaadofadaWgaaadbaGaamOraaqabaaaleqani abggHiLdaakeaadaaeqaqaaiaadsgadaWgaaWcbaGaamiBaaqabaaa baGaamiBaiabgIGiolaadofadaWgaaadbaGaamytaaqabaaaleqani abggHiLdaaaOGaaiOlaaaa@5588@

Puisque nous avons supposé qu’il existe un vecteur γ MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWHZoaaaa@32FE@ tel que γ x k = 1 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWHZoWaaWbaaSqabeaatCvAUfeBSn 0BKvguHDwzZbqegeKCPfgBGuLBPn2BKvginnfaiqaacaWFsedaaOGa aCiEamaaBaaaleaacaWGRbaabeaakiabg2da9iaaigdaaaa@4359@ pour tout k U , MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWGRbGaeyicI4SaamyvaiaacYcaaa a@35BD@ alors nous avons

γ A = k S F v k d k = k S F d k . MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWHZoWaaWbaaSqabeaatCvAUfeBSn 0BKvguHDwzZbqegeKCPfgBGuLBPn2BKvginnfaiqaacaWFsedaaOGa aCyqaiabg2da9maaqafabaGaamODamaaBaaaleaacaWGRbaabeaaki aadsgadaWgaaWcbaGaam4AaaqabaaabaGaam4AaiabgIGiolaadofa daWgaaadbaGaamOraaqabaaaleqaniabggHiLdGccqGH9aqpdaaeqb qaaiaadsgadaWgaaWcbaGaam4AaaqabaaabaGaam4AaiabgIGiolaa dofadaWgaaadbaGaamOraaqabaaaleqaniabggHiLdGccaGGUaaaaa@5606@

Considérons maintenant

Y ¯ ^ F | M = k S F v k d k y k k S F v k d k = k S F v k d k y k k S F d k . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaaceWGzbGbaeHbaKaadaWgaaWcbaWaaq GaaeaacaWGgbGaaGjcVdGaayjcSdGaaGjcVlaad2eaaeqaaOGaeyyp a0ZaaSaaaeaadaaeqaqaaiaadAhadaWgaaWcbaGaam4AaaqabaGcca WGKbWaaSbaaSqaaiaadUgaaeqaaOGaamyEamaaBaaaleaacaWGRbaa beaaaeaacaWGRbGaeyicI4Saam4uamaaBaaameaacaWGgbaabeaaaS qab0GaeyyeIuoaaOqaamaaqababaGaamODamaaBaaaleaacaWGRbaa beaakiaadsgadaWgaaWcbaGaam4AaaqabaaabaGaam4AaiabgIGiol aadofadaWgaaadbaGaamOraaqabaaaleqaniabggHiLdaaaOGaeyyp a0ZaaSaaaeaadaaeqaqaaiaadAhadaWgaaWcbaGaam4AaaqabaGcca WGKbWaaSbaaSqaaiaadUgaaeqaaOGaamyEamaaBaaaleaacaWGRbaa beaaaeaacaWGRbGaeyicI4Saam4uamaaBaaameaacaWGgbaabeaaaS qab0GaeyyeIuoaaOqaamaaqababaGaamizamaaBaaaleaacaWGRbaa beaaaeaacaWGRbGaeyicI4Saam4uamaaBaaameaacaWGgbaabeaaaS qab0GaeyyeIuoaaaGccaGGUaaaaa@67AF@

De nouveau, deux cas doivent être considérés :

-     Si j S F MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWGQbGaeyicI4Saam4uamaaBaaale aacaWGgbaabeaaaaa@3601@

Y ¯ ^ F | M I j = d j ( v j y j Y ¯ ^ F | M ) + λ I j [ k S F F ( x k λ ) d k x k y k ] k S F d k = d j ( v j y j Y ¯ ^ F | M ) d j ( v j x j X ¯ ^ M ) T 1 k S F F ( x k λ ) d k x k y k k S F d k = d j [ v j y j Y ¯ ^ F | M ( v j x j X ¯ ^ M ) B F ] k S F d k , MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaafaqaaeWacaaabaWaaSaaaeaaceWGzb GbaeHbaKaadaWgaaWcbaWaaqGaaeaacaWGgbGaaGjcVdGaayjcSdGa aGjcVlaad2eaaeqaaaGcbaGaeyOaIyRaamysamaaBaaaleaacaWGQb aabeaaaaaakeaacqGH9aqpdaWcaaqaaiaadsgadaWgaaWcbaGaamOA aaqabaGcdaqadaqaaiaadAhadaWgaaWcbaGaamOAaaqabaGccaWG5b WaaSbaaSqaaiaadQgaaeqaaOGaeyOeI0IabmywayaaryaajaWaaSba aSqaamaaeiaabaGaamOraiaayIW7aiaawIa7aiaayIW7caWGnbaabe aaaOGaayjkaiaawMcaaiabgUcaRmaalaaabaGaeyOaIyRaaC4Udmaa CaaaleqabaWexLMBbXgBd9gzLbvyNv2CaeHbbjxAHXgiv5wAJ9gzLb sttbaceaGaa8NeXaaaaOqaaiabgkGi2kaadMeadaWgaaWcbaGaamOA aaqabaaaaOWaamWaaeaadaaeqaqaaiaadAeadaahaaWcbeqaaKqzGf Gamai2gkdiIcaaaSqaaiaadUgacqGHiiIZcaWGtbWaaSbaaWqaaiaa dAeaaeqaaaWcbeqdcqGHris5aOWaaeWaaeaacaWH4bWaa0baaSqaai aadUgaaeaacaWFsedaaOGaaC4UdaGaayjkaiaawMcaaiaadsgadaWg aaWcbaGaam4AaaqabaGccaWH4bWaaSbaaSqaaiaadUgaaeqaaOGaam yEamaaBaaaleaacaWGRbaabeaaaOGaay5waiaaw2faaaqaamaaqaba baGaamizamaaBaaaleaacaWGRbaabeaaaeaacaWGRbGaeyicI4Saam 4uamaaBaaameaacaWGgbaabeaaaSqab0GaeyyeIuoaaaaakeaaaeaa cqGH9aqpdaWcaaqaaiaadsgadaWgaaWcbaGaamOAaaqabaGcdaqada qaaiaadAhadaWgaaWcbaGaamOAaaqabaGccaWG5bWaaSbaaSqaaiaa dQgaaeqaaOGaeyOeI0IabmywayaaryaajaWaaSbaaSqaamaaeiaaba GaamOraiaayIW7aiaawIa7aiaayIW7caWGnbaabeaaaOGaayjkaiaa wMcaaiabgkHiTiaadsgadaWgaaWcbaGaamOAaaqabaGcdaqadaqaai aadAhadaWgaaWcbaGaamOAaaqabaGccaWH4bWaaSbaaSqaaiaadQga aeqaaOGaeyOeI0IabCiwayaaryaajaWaaSbaaSqaaiaad2eaaeqaaa GccaGLOaGaayzkaaWaaWbaaSqabeaacaWFsedaaOGaaCivamaaCaaa leqabaGaeyOeI0IaaGymaaaakmaaqababaGaamOramaaCaaaleqaba qcLbwacWaGyBOmGikaaaWcbaGaam4AaiabgIGiolaadofadaWgaaad baGaamOraaqabaaaleqaniabggHiLdGcdaqadaqaaiaahIhadaqhaa WcbaGaam4Aaaqaaiaa=jrmaaGccaWH7oaacaGLOaGaayzkaaGaamiz amaaBaaaleaacaWGRbaabeaakiaahIhadaWgaaWcbaGaam4Aaaqaba GccaWG5bWaaSbaaSqaaiaadUgaaeqaaaGcbaWaaabeaeaacaWGKbWa aSbaaSqaaiaadUgaaeqaaaqaaiaadUgacqGHiiIZcaWGtbWaaSbaaW qaaiaadAeaaeqaaaWcbeqdcqGHris5aaaaaOqaaaqaaiabg2da9maa laaabaGaamizamaaBaaaleaacaWGQbaabeaakmaadmaabaGaamODam aaBaaaleaacaWGQbaabeaakiaadMhadaWgaaWcbaGaamOAaaqabaGc cqGHsislceWGzbGbaeHbaKaadaWgaaWcbaWaaqGaaeaacaWGgbGaaG jcVdGaayjcSdGaaGjcVlaad2eaaeqaaOGaeyOeI0YaaeWaaeaacaWG 2bWaaSbaaSqaaiaadQgaaeqaaOGaaCiEamaaBaaaleaacaWGQbaabe aakiabgkHiTiqahIfagaqegaqcamaaBaaaleaacaWGnbaabeaaaOGa ayjkaiaawMcaamaaCaaaleqabaGaa8NeXaaakiaahkeadaWgaaWcba GaamOraaqabaaakiaawUfacaGLDbaaaeaadaaeqaqaaiaadsgadaWg aaWcbaGaam4AaaqabaaabaGaam4AaiabgIGiolaadofadaWgaaadba GaamOraaqabaaaleqaniabggHiLdaaaOGaaiilaaaaaaa@EA18@

       où

B F = T 1 k S F F ( x k λ ) d k x k y k . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWHcbWaaSbaaSqaaiaadAeaaeqaaO Gaeyypa0JaaCivamaaCaaaleqabaGaeyOeI0IaaGymaaaakmaaqafa baGaamOramaaCaaaleqabaqcLbwacWaGyBOmGikaaaWcbaGaam4Aai abgIGiolaadofadaWgaaadbaGaamOraaqabaaaleqaniabggHiLdGc daqadaqaaiaahIhadaqhaaWcbaGaam4AaaqaamXvP5wqSX2qVrwzqf 2zLnharyqqYLwySbsvUL2yVrwzG00uaGabaiaa=jrmaaGccaWH7oaa caGLOaGaayzkaaGaamizamaaBaaaleaacaWGRbaabeaakiaahIhada WgaaWcbaGaam4AaaqabaGccaWG5bWaaSbaaSqaaiaadUgaaeqaaOGa aiOlaaaa@5AAF@

-     Si j S M MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWGQbGaeyicI4Saam4uamaaBaaale aacaWGnbaabeaaaaa@3608@

Y ¯ ^ F | M I j = λ I j k S F F ( x k λ ) d k x k y k k S F d k = d j ( x j X ¯ ^ M ) k S F d k l S M d l T 1 k S F F ( x k λ ) d k x k y k k S F d k = d j ( x j X ¯ ^ M ) 1 l S M d l B F . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaafaqaaeWacaaabaWaaSaaaeaaceWHzb GbaeHbaKaadaWgaaWcbaWaaqGaaeaacaWGgbGaaGjcVdGaayjcSdGa aGjcVlaad2eaaeqaaaGcbaGaeyOaIyRaamysamaaBaaaleaacaWGQb aabeaaaaaakeaacqGH9aqpdaWcaaqaaiabgkGi2kaahU7adaahaaWc beqaamXvP5wqSX2qVrwzqf2zLnharyqqYLwySbsvUL2yVrwzG00uaG abaiaa=jrmaaaakeaacqGHciITcaWGjbWaaSbaaSqaaiaadQgaaeqa aaaakmaalaaabaWaaabeaeaacaWGgbWaaWbaaSqabeaajugybiadaI THYaIOaaaaleaacaWGRbGaeyicI4Saam4uamaaBaaameaacaWGgbaa beaaaSqab0GaeyyeIuoakmaabmaabaGaaCiEamaaDaaaleaacaWGRb aabaGaa8NeXaaakiaahU7aaiaawIcacaGLPaaacaWGKbWaaSbaaSqa aiaadUgaaeqaaOGaaCiEamaaBaaaleaacaWGRbaabeaakiaadMhada WgaaWcbaGaam4AaaqabaaakeaadaaeqaqaaiaadsgadaWgaaWcbaGa am4AaaqabaaabaGaam4AaiabgIGiolaadofadaWgaaadbaGaamOraa qabaaaleqaniabggHiLdaaaaGcbaaabaGaeyypa0JaamizamaaBaaa leaacaWGQbaabeaakmaabmaabaGaaCiEamaaBaaaleaacaWGQbaabe aakiabgkHiTiqahIfagaqegaqcamaaBaaaleaacaWGnbaabeaaaOGa ayjkaiaawMcaamaaCaaaleqabaGaa8NeXaaakmaalaaabaWaaabeae aacaWGKbWaaSbaaSqaaiaadUgaaeqaaaqaaiaadUgacqGHiiIZcaWG tbWaaSbaaWqaaiaadAeaaeqaaaWcbeqdcqGHris5aaGcbaWaaabeae aacaWGKbWaaSbaaSqaaiaadYgaaeqaaaqaaiaadYgacqGHiiIZcaWG tbWaaSbaaWqaaiaad2eaaeqaaaWcbeqdcqGHris5aaaakiaahsfada ahaaWcbeqaaiabgkHiTiaaigdaaaGcdaWcaaqaamaaqababaGaamOr amaaCaaaleqabaqcLbwacWaGyBOmGikaaaWcbaGaam4AaiabgIGiol aadofadaWgaaadbaGaamOraaqabaaaleqaniabggHiLdGcdaqadaqa aiaahIhadaqhaaWcbaGaam4Aaaqaaiaa=jrmaaGccaWH7oaacaGLOa GaayzkaaGaamizamaaBaaaleaacaWGRbaabeaakiaahIhadaWgaaWc baGaam4AaaqabaGccaWG5bWaaSbaaSqaaiaadUgaaeqaaaGcbaWaaa beaeaacaWGKbWaaSbaaSqaaiaadUgaaeqaaaqaaiaadUgacqGHiiIZ caWGtbWaaSbaaWqaaiaadAeaaeqaaaWcbeqdcqGHris5aaaaaOqaaa qaaiabg2da9iaadsgadaWgaaWcbaGaamOAaaqabaGcdaqadaqaaiaa hIhadaWgaaWcbaGaamOAaaqabaGccqGHsislceWHybGbaeHbaKaada WgaaWcbaGaamytaaqabaaakiaawIcacaGLPaaadaahaaWcbeqaaiaa =jrmaaGcdaWcaaqaaiaaigdaaeaadaaeqaqaaiaadsgadaWgaaWcba GaamiBaaqabaaabaGaamiBaiabgIGiolaadofadaWgaaadbaGaamyt aaqabaaaleqaniabggHiLdaaaOGaaCOqamaaBaaaleaacaWGgbaabe aakiaac6caaaaaaa@C319@

Donc, la variable linéarisée est

z k = { d j [ v j y j Y ¯ ^ F | M ( v j x j X ¯ ^ M ) B F ] k S F d k si j S F d j ( x j X ¯ ^ M ) B F l S M d l si j S M . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaebbnrfifHhDYfgasaacH8rrps0l bbf9q8WrFfeuY=Hhbbf9y8WrFr0xc9vqFj0db9qqvqFr0dXdHiVc=b YP0xH8peuj0dYddrpe0db9Wqpepic9qr=xfr=xfr=xmeaabaqaciGa caGaaeqabaqaaeaadaaakeaacaWG6bWaaSbaaSqaaiaadUgaaeqaaO Gaeyypa0ZaaiqaaeaafaqaaeGacaaabaWaaSaaaeaacaWGKbWaaSba aSqaaiaadQgaaeqaaOWaamWaaeaacaWG2bWaaSbaaSqaaiaadQgaae qaaOGaamyEamaaBaaaleaacaWGQbaabeaakiabgkHiTiqadMfagaqe gaqcamaaBaaaleaadaabcaqaaiaadAeacaaMi8oacaGLiWoacaaMi8 UaamytaaqabaGccqGHsisldaqadaqaaiaadAhadaWgaaWcbaGaamOA aaqabaGccaWH4bWaaSbaaSqaaiaadQgaaeqaaOGaeyOeI0IabCiway aaryaajaWaaSbaaSqaaiaad2eaaeqaaaGccaGLOaGaayzkaaWaaWba aSqabeaatCvAUfeBSn0BKvguHDwzZbqegeKCPfgBGuLBPn2BKvginn faiqaacaWFsedaaOGaaCOqamaaBaaaleaacaWGgbaabeaaaOGaay5w aiaaw2faaaqaamaaqababaGaamizamaaBaaaleaacaWGRbaabeaaae aacaWGRbGaeyicI4Saam4uamaaBaaameaacaWGgbaabeaaaSqab0Ga eyyeIuoaaaaakeaacaqGZbGaaeyAaiaaysW7caaMc8UaamOAaiabgI GiolaadofadaWgaaWcbaGaamOraaqabaaakeaadaWcaaqaaiaadsga daWgaaWcbaGaamOAaaqabaGcdaqadaqaaiaahIhadaWgaaWcbaGaam OAaaqabaGccqGHsislceWHybGbaeHbaKaadaWgaaWcbaGaamytaaqa baaakiaawIcacaGLPaaadaahaaWcbeqaaiaa=jrmaaGccaWHcbWaaS baaSqaaiaadAeaaeqaaaGcbaWaaabeaeaacaWGKbWaaSbaaSqaaiaa dYgaaeqaaaqaaiaadYgacqGHiiIZcaWGtbWaaSbaaWqaaiaad2eaae qaaaWcbeqdcqGHris5aaaaaOqaaiaabohacaqGPbGaaGjbVlaaykW7 caWGQbGaeyicI4Saam4uamaaBaaaleaacaWGnbaabeaakiaac6caaa aacaGL7baaaaa@8E25@

La variable linéarisée doit uniquement être insérée dans l’estimateur de variance correspondant au plan de sondage. Notons que la variance de la moyenne contrefactuelle dépend de la variance calculée pour l’échantillon d’hommes en ce qui concerne la part qui est expliquée par la régression, et de la variance calculée pour l’échantillon de femmes en ce qui concerne la part qui demeure inexpliquée.

Bibliographie

Bielby, W.T., et Baron, J.N. (1986). Men and women at work: Sex segregation and statistical discrimination. American Journal of Sociology, 759-799.

Blinder, A.S. (1973). Wage discrimination: Reduced form and structural estimates. Journal of Human Resources, 8(4), 436-455.

Bourguignon, F., Ferreira, F.H. et Leite, P.G. (2002). Beyond Oaxaca-Blinder: Accounting for differences in household income distributions across countries. Inequality and Economic Development in Brazil, 105.

Deville, J.-C., et Särndal, C.-E. (1992). Calibration estimators in survey sampling. Journal of the American Statistical Association, 87(418), 376-382.

DiNardo, J. (2002). Propensity score reweighting and changes in wage distributions. Document de discussion, University of Michigan.

DiNardo, J., Fortin, N.M. et Lemieux, T. (1996). Labor market Institutions and the distribution of wages, 1973-1992: A semiparametric approach. Econometrica, 64(5), 1001-44.

Donzé, L. (2013). Erreurs de spécification dans la décomposition de l’inégalité salariale. Dans International Conference Ars Conjectandi, 1713-2013.

Fortin, N., Lemieux, T. et Firpo, S. (2011). Decomposition methods in economics. Dans Handbook of Labor Economics, (Éds., O. Ashenfelter et D. Card), 4, 1-102. Elsevier.

Gardeazabal, J., et Ugidos, A. (2005). Gender wage discrimination at quantiles. Journal of Population Economics, 18(1), 165-179.

Graf, M. (2011). Use of survey weights for the analysis of compositional data. Dans Compositional Data Analysis: Theory and Applications, (Éds., V. Pawlowsky-Glahn et A. Buccianti), 114-127. Wiley, Chichester.

Neumark, D. (1988). Employers’ discriminatory behavior and the estimation of wage Discrimination. Journal of Human Resources, 23(3), 279-295.

Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International Economic Review, 14(3), 693-709.

Weichselbaumer, D., et Winter-Ebmer, R. (2005). A meta-analysis of the international gender wage gap. Journal of Economic Surveys, 19(3), 479-511.

Weichselbaumer, D., et Winter-Ebmer, R. (2006). Rhetoric in economic research: The case of gender wage differentials. Industrial Relations: A Journal of Economy and Society, 45(3), 416-436.

Williams, C. (2010). Bien-être économique. Femmes au Canada : rapport statistique fondé sur le sexe, Statistique Canada.


Date de modification :