Holz, Manuel; Maye l, Jochen
A icle — Published Ve sion
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h
Inequali ies? Di e ences in T ajec o ies o Subjec i e Heal h Be ween
Mig an s and Na i e Ge mans o e Time
KZ SS Kölne Zei sch i ü Soziologie und Sozialpsychologie
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Holz, Manuel; Maye l, Jochen (2025) : Can Di e ing Occupa ional Class
Posi ions Explain Mig an Heal h Inequali ies? Di e ences in T ajec o ies o Subjec i e Heal h
Be ween Mig an s and Na i e Ge mans o e Time, KZ SS Kölne Zei sch i ü Soziologie und
Sozialpsychologie, ISSN 1861-891X, Sp inge Fachmedien Wiesbaden GmbH, Wiesbaden, Vol. 77,
Iss. 1, pp. 27-52,
h ps://doi.o g/10.1007/s11577-025-00985-3
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323561
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/deed.de
ABHANDLUNG
h ps://doi.o g/10.1007/s11577-025-00985-3
Köln Z Soziol (2025) 77:27–52
Can Di e ing Occupa ional Class Posi ions Explain
Mig an Heal h Inequali ies? Di e ences in T ajec o ies
o Subjec i e Heal h Be ween Mig an s and Na i e
Ge mans o e Time
Manuel Holz · Jochen Maye l
Recei ed: 4 Decembe 2024 / Accep ed: 21 Janua y 2025 / Published online: 11 Ma ch 2025
© The Au ho (s) 2025
Abs ac Mig an s li ing in pos indus ial coun ies a e con on ed wi h a ious
socioeconomic challenges, including lowe incomes, ex ended wo king hou s, and
lowe occupa ional s a uses han na i es. Al hough heal h dispa i ies linked o oc-
cupa ional posi ions ha e equen ly been documen ed, hey emain a ela i ely
unexplo ed ac o in he explana ion o heal h gaps o e ime be ween mig an s and
na i e popula ions. To add ess his issue, we u ilized longi udinal da a spanning om
2002 o 2018 om he Ge man Socio-Economic Panel o in es iga e dispa i ies in
physical heal h– ela ed quali y o li e ac oss di e en mig an and na i e Ge man
coho s and hei associa ions wi h occupa ional class posi ion. Ou indings e eal
ha o e all lowe occupa ional class posi ions can accoun o he heal h dispa i ies
obse ed be ween mig an s and na i e Ge mans. Fu he , ou s udy un eils complex
ela ionships be ween ini ial heal h condi ions (in e cep s), changes in heal h o e
ime (slopes), egion o o igin (Eu opean mig an s, non-Eu opean mig an s, and na-
i e Ge mans), and gende . These nuanced ou comes unde sco e he impo ance o
adop ing app oaches ha conside bo h egion o o igin and gende when seeking
o enhance wo king condi ions and acili a e access o he labo ma ke o di e se
popula ions.
Keywo ds Heal h ajec o y · Mig a ion · Inequali y · Occupa ional posi ion ·
La en g ow h · Cu e analysis
M. Holz · J. Maye l
Ins i u e o Sociology, Facul y o Beha io al and Social Sciences, Chemni z Uni e si y o
Technology
Thü inge Weg 9, 09126 Chemni z, Ge many
E-Mail: manuel.holz@hsw. u-chemni z.de
J. Maye l
E-Mail: jochen.maye l@soziologie. u-chemni z.de
K
28 M. Holz, J. Maye l
Können un e schiedliche be u liche Klassenposi ionen die
gesundhei lichen Ungleichhei en on Mig an en e klä en? Un e schiede
in subjek i en Gesundhei s e läu en zwischen Mig an en und
Deu schen ohne Mig a ionshin e g und im Längsschni
Zusammen assung Mig an en, die in pos indus iellen Lände n leben, sehen sich
mi e schiedenen sozioökonomischen He aus o de ungen kon on ie . Da un e
allen nied ige Einkommen, lange A bei szei en und nied ige be u liche Posi io-
nen. Wäh end gesundhei liche Ungleichhei en in Ve bindung mi be u lichen Posi-
ionen häu ig dokumen ie wu den, bleiben sie ein ela i une o sch e Fak o bei
de E klä ung on längsschni lichen Gesundhei sun e schieden zwischen Mig an en
und einheimischen Be ölke ungen. Um dieses P oblem anzugehen, wu den Längs-
schni da en aus dem Zei aum on 2002 bis 2018 aus dem Sozio-oekonomischen
Panel (SOEP) e wende . Die Un e schiede in de gesundhei sbezogenen Lebens-
quali ä zwischen e schiedenen Mig an en-Koho en und deu schen Koho en ohne
Mig a ionshin e g und we den im Zusammenhang mi de be u lichen Klassenzuge-
hö igkei un e such . Unse e E gebnisse zeigen, dass insgesam nied ige e be u liche
Klassenzugehö igkei en die beobach e en gesundhei lichen Un e schiede zwischen
Mig an en und einheimischen Deu schen e klä en können. Da übe hinaus zeig un-
se e S udie komplexe Beziehungen zwischen an änglichen Gesundhei szus änden
(In e cep s) und Ve ände ungen de Gesundhei im Zei e lau (Slopes) in Bezug
au die He kun s egion (eu opäische Mig an en, nich eu opäische Mig an en und
Deu sche ohne Mig a ionshin e g und) und das Geschlech . Die E gebnisse un e -
s eichen die Bedeu ung in eg a i e Ansä ze, die sowohl die He kun s egion als
auch das Geschlech be ücksich igen, wenn es da um geh , die A bei sbedingungen
zu e besse n und den Zugang zum A bei sma k ü di e se Be ölke ungsg uppen
zu e leich e n.
Schlüsselwö e Gesundhei s e läu e · Mig a ion · Ungleichhei · Be u liche
Posi ion · La en e Wachs umsku en
1 In oduc ion
Mig an popula ions o en en e hos coun ies wi h a heal h ad an age compa ed
o he na i e popula ion, a phenomenon commonly known as he “heal hy mig an
e ec ” (HME; McDonald and Kennedy 2004). Howe e , o e ime, his ad an age
diminishes, leading o a ansi ion in o a s a e o heal h disad an age o con e gence
wi h na i e heal h le els. Whe eas he e is mo e consensus and es ed explana ions
o he i s pa o he HME, es ed in e p e a ions and desc ip ions o he second
pa a e sca ce (Rells ab e al. 2016;Yang2021).
To en ich ou unde s anding o heal h ajec o ies ollowing mig a ion, we d aw
on aspec s om li e-cou se pe spec i es on socioeconomic heal h dispa i ies (En-
gelha d -Woel le and Leopold 2020). Wi hin his comp ehensi e amewo k, h ee
dis inc scena ios o heal h ajec o ies may eme ge: di e gence, con e gence, and
con inui y (Leopold and Engelha d 2011;P us2007). In ou case, di e gence occu s
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 29
when heal h dispa i ies be ween mig an s and na i es inc ease o e ime, sugges ing
an inc ease in mig an heal h disad an ages o a decline in mig an heal h ad an ages
compa ed o na i e popula ions. Con e gence, on he o he hand, implies a le eling
o o heal h di e en ials, wi h mig an heal h g adually aligning wi h ha o na i e
popula ions. Con inui y indica es ha heal h dispa i ies pe sis o e ime, wi h mi-
g an heal h emaining consis en ly abo e o below na i e le els wi hou subs an ial
change. In ou s udy, ou p ima y objec i e is o explain hese po en ially di e ing
ajec o ies o pos mig a ion mig an heal h.
Al hough po e y is o en conside ed a “ isk ac o ” o heal h dispa i ies, we
a gue ha i canno be ully unde s ood wi hou conside ing i s ela ionship o social
class, which is closely ied o he wo k en i onmen (Scamble and Higgs 1999).
The wo king condi ions o mig an s ha e eme ged as a c i ical ye unde explo ed
mechanism o longi udinal heal h inequali ies ha wa an s close examina ion.
Concu en ly, mig an s ace a ious challenges in he hos coun y’s labo ma ke ,
including lowe employmen a es and occupa ional s a us, unde employmen , and
educed ea nings (Be wing 2019; Damelang e al. 2021; Kal e and G ana o 2018;
S anek and Ramos 2013). The wo king condi ions in labo ma ke posi ions occu-
pied mainly by mig an s ha e se e e de imen al e ec s on mig an heal h. Mig an
wo ke s ace disp opo iona e occupa ional heal h haza ds such as hea y li ing;
ib a ions om machine y; long pe iods o s anding; exposu e o ex eme empe -
a u es, noise, and chemical oxins; he need o wo k a a high pace; employmen
a angemen s wi hou con ac s o wi h limi ed con ol o e ee days; and psychoso-
cial s ain (Oldenbu g e al. 2010; Ronda Pé ez e al. 2012; Simmons e al. 2022).
Hence, we hypo hesize ha he de e io a ion in mig an heal h occu s a a mo e ac-
cele a ed a e compa ed o ha o na i e popula ions, and we a ibu e his a iance
o he compa a i ely lowe occupa ional class posi ions held by mig an s.
This s udy con ibu es o he exis ing body o li e a u e by examining whe he oc-
cupa ional class posi ion can explain mig an heal h dispa i ies o e ime, and he e-
o e i can deli e an empi ically es ed explana ion o he heal h decline obse ed
in he HME phenomenon.1By u ilizing da a om he Ge man Socio-Economic
Panel (SOEP) and applying la en g ow h cu e analysis wi hin a s uc u al equa ion
modeling amewo k, his s udy es ablishes occupa ional class posi ion, using he
E ikson–Gold ho pe–Po oca e o (EGP) classi ica ion o social classes (E ikson and
Gold ho pe 1993), as a media o be ween mig a ion backg ound and physical heal h
ou comes. Mo eo e , we di e en ia e be ween di e en egions o o igin (Eu opean
s. non-Eu opean mig an s) and conduc he analysis sepa a ely by gende o gain
mo e nuanced insigh s ega ding he ela ionship be ween mig a ion backg ound,
occupa ional class, and heal h o e he cou se o ime.
1Hence, ou emphasis is no on ecen mig an s ( hose wi h a sho du a ion since mig a ion) bu a he on
mig an s as a whole, as ou p ima y in e es lies in unde s anding he heal h-decline aspec o he HME.
K
30 M. Holz, J. Maye l
Fig. 1 Heal h ajec o ies in he con ex o he heal hy mig an e ec
2 Concep ual Backg ound
The con ex o his analysis is he so-called HME (McDonald and Kennedy 2004).
In his well-documen ed phenomenon, i is obse ed ha mig an s gene ally ha e
be e heal h upon a i al in he hos coun y compa ed o he na i e popula ion.
Howe e , his ad an age ends o diminish o e ime, e en ually leading o a s a e o
heal h disad an age. The p ima y elemen o his phenomenon, which is he ini ial
heal h ad an age, has ecei ed conside able a en ion and deli e ed su icien ly es ed
in e p e a ions. These explana ions p ima ily e ol e a ound heal hy sel -selec ion
p io o mig a ion (Fulle -Thomson e al. 2016), selec i e emig a ion (Di Napoli
e al. 2021), and cul u e-speci ic heal h-p o ec i e beha io s (Lee e al. 2013).
Fo he second pa o he HME phenomenon, he disp opo iona e heal h decline,
explana ions and e idence a e ei he a e o inconclusi e (Rells ab e al. 2016;Yang
2021). Ou ocus in his s udy is he e o e he subsequen pa o he HME. The
con incing, ye un es ed, easons o pos mig a ion heal h ajec o ies being ei he
simila o wo se han he na i e heal h ajec o ies ha e been a ibu ed mainly o
accul u a ion and he adop ion o unheal hy hos coun y–speci ic li es yles (Edbe g
e al. 2011; Fujishi o and Hoppe 2020; Name and Razum 2018).
Among he many ac o s con ibu ing o mig an heal h dispa i ies, he wo king
condi ions expe ienced by mig an s a e a c i ical and unde explo ed de e minan ac-
coun ing o mig an heal h decline (Co na 2013). No ably, he employmen con ex
is a e y well- esea ched ac o in explaining di e ging heal h ou comes (Lahelma
e al. 2005; Hiesinge and Topho en 2019; Landsbe gis 2010; Gonzalez-Mulé and
Cockbu n 2017; Geye and Pe e 1999). Gi en he long la ency pe iods be ween he
socioeconomic condi ions desc ibing he mig an –na i e di e ences in long- e m
heal h ajec o ies, such as he occupa ional con ex (Clumeck e al. 2009; Johnson
e al. 1996), li e-cou se pe spec i es in socioeconomic heal h dispa i ies p o ide
a use ul amewo k (Co na 2013; Edbe g e al. 2011; Fujishi o and Hoppe 2020;
Leong e al. 2014).
The e a e ypically h ee scena ios, o hypo heses, conce ning he de elopmen
o heal h di e ences be ween social g oups o e ime: con e gence, di e gence, o
con inui y/s a us main enance (Engelha d -Woel le and Leopold 2020; Leopold and
Engelha d 2011). These ajec o y ypes we e ini ially o mula ed o dispa i ies
in heal h due o di e en socioeconomic posi ions. In his con ex , we apply his
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 31
concep o mig an heal h dispa i ies. Conside ing he heal hy mig an phenomenon,
all scena ios s a wi h a mig an heal h ad an age (Fig. 1). Howe e , in scena io a,
he di e gence ajec o y, he heal h o mig an s de e io a es cons an ly a a highe
a e han he heal h o nonmig an s (o “na i es”), esul ing in an inc easing di e -
ence in he ajec o ies o e ime. In scena io bo con e gence, he a e o heal h
de e io a ion is ini ially highe o mig an s han o nonmig an s, bu he heal h
ajec o ies app oxima e each o he and become s a is ically indis inguishable o e
he cou se o ime. In he con inui y scena io c, bo h g oups expe ience a simila
a e o heal h de e io a ion o e ime, ye a pe sis ing heal h ad an age exis s o
one g oup, and he ajec o ies s ay pa allel in he long un; i.e., heal h inequali ies
s ay wi hin a cons an ange and nei he g ow no diminish when he numbe o
obse a ion pe iods inc eases.
The di e gence ajec o y is p ima ily explained by he concep o cumula i e
ad an age/disad an age. Due o hei posi ion in he socioeconomic s a a, indi id-
uals expe ience a pa h-dependen accumula ion o dis/ad an ages. P io s a es o
posi ions ampli y subsequen condi ions, and he payo s om ad an ages o conse-
quences om disad an ages inc ease o e ime (Engelha d -Woel le and Leopold
2020). In o he wo ds, highe socioeconomic posi ions can ansla e in o less haz-
a dous wo king condi ions, leading o educed pe iods o ha m ul exposu e o e
ime (P us 2007). The disad an ages in he case o mig an wo ke s a e ep esen ed
as hese wo ke s being disp opo iona ely mo e o en exposed o occupa ional heal h
haza ds, as ou lined in he in oduc ion (Oldenbu g e al. 2010; Ronda Pé ez e al.
2012; Simmons e al. 2022). The e o e, di e ences in wo king condi ions ansla e
in o g owing heal h dispa i ies be ween social g oups, i.e., be ween mig an s and
he na i e popula ion.
The con e gence ajec o y is o en explained by he “age-as-le ele ” hypo he-
sis (Co na 2013; Engelha d -Woel le and Leopold 2020). This pe spec i e posi s
ha ini ial heal h inequali ies a e “bu e ed” by biological, social class–independen
heal h de e io a ion in la e s ages o li e. Fo he mig an –na i e heal h di e ences,
i can be a gued ha in addi ion o he biological “bu e ing,” cul u al ac o s can
be a play, which lead o con e gence. A po en ial explana ion can be de i ed by
elemen s om cumula i e inequali y heo y (Fe a o e al. 2009). This pe spec i e
highligh s ha indi iduals can be exposed o bo h heal h-p omo ing and heal h-
de e io a ing ac o s a he same ime. In his case, haza dous wo king condi ions
can indeed inc ease he a e o heal h decline, bu cul u e-speci ic ac o s such as
be e die (Da mon and Khla 2001), close amily ela ions (A ends-Tó h and an
de Vij e 2008), and eligiosi y (Auks -Ma ge i´
candMa ge i
´
c2005) can mi iga e
he nega i e e ec s o low socioeconomic posi ion on heal h, e en ually leading o
simila heal h le els. Ano he eason o con e ging heal h ajec o ies s ems om
me hodological issues, namely selec i e panel mo ali y. Due o heal h pa icipa ion
bias in su eys, unheal hie indi iduals exi he panel (o die) a a disp opo iona e
a e. As a esul , he su ey popula ion in la e yea s becomes mo e simila (Shaw
and K ause 2002).
Acco ding o he con inui y pe spec i e, heal h dispa i ies emain ela i ely s able
o e ime. This pe spec i e sugges s ha once heal h dispa i ies a e es ablished due
o socioeconomic s a us, hey pe sis h oughou he li e cou se, wi h he mig an and
K
32 M. Holz, J. Maye l
na i e lines emaining pa allel (Engelha d -Woel le and Leopold 2020). Building
simila ly on cumula i e inequali y heo y, cul u al ac o s can bu e he e ec s o
wo king condi ions on heal h decline, al hough in his scena io hese ac o s canno
cancel ou his e ec en i ely.
Rega ding he e idence o he possible heal h ajec o ies, he e is some e idence
o con e gence (Becke 2000), ye di e gence seems o be he mos common pa -
e n o li e-cou se socioeconomic heal h dispa i ies, indica ing ha he cumula i e
ad an age/disad an age pe spec i e is he mos p obable one (Leopold and Engel-
ha d 2011; Chandola e al. 2007; Shaw and K ause 2002;P us2007; Sacke e al.
2011). Ye in hese s udies, he concep o socioeconomic posi ion used is mos ly in
he o m o educa ion. Job con ex o exposu e has no been applied as a mode a o
be ween age/ ime and heal h.
One consis en inding ac oss s udies ega ding heal h ajec o ies o mig an s and
e hnic mino i ies (Edbe g e al. 2011; Ja ana e al. 2018; Shuey and Willson 2008)
is he exis ence o a heal h disad an age in he o m o as e heal h decline o hese
g oups compa ed o he na i e majo i y popula ion. Ano he aspec ha eme ges is
he he e ogenei y o mig an heal h inequali y. In he wo k o Edbe g e al. (Edbe g
e al. 2011), mig an s om coun ies cha ac e ized by highe economic inequali y
end o exhibi a mo e p onounced pos mig a ion heal h de e io a ion han mig an s
om coun ies wi h lowe economic inequali y. Simila ly, in a s udy on heal h a-
jec o ies o mig an s in Aus alia, Ja ana e al. (2018) ound ha mig an s om
English-speaking coun ies showed pe sis ing heal h ad an ages in he li e cou se
compa ed o na i es, while mig an s om non–English-speaking coun ies showed
heal h disad an ages ha inc eased o e ime. I is he e o e c ucial o di e en ia e
be ween di e en g oups ega ding heal h ajec o ies. Fu he , he mi iga ing e ec
o socioeconomic posi ion on heal h decline does no show a clea end ega ding
mino i y s a us. In a s udy by Assa i (2017), educa ional a ainmen bu e ed heal h
de e io a ion only o Black men, and Shuey and Willson (2008) ound ha be-
longing o an e hnic mino i y g oup is associa ed wi h as e de e io a ion in heal h
compa ed o he majo i y popula ion. Howe e , socioeconomic posi ion se es as
a bu e agains his decline, al hough he bu e ing e ec appea s o be mo e p o-
nounced among he whi e majo i y popula ion.
2.1 Occupa ional Class as a Measu e o Wo king Condi ions
B inging he a o emen ioned opics all oge he in o one esea ch p og am is chal-
lenging in e ms o da a a ailabili y. In o de o inc ease sample sizes and obse a ion
pe iods, we conside ed da a om he Ge man SOEP o be he mos app op ia e o
ou pu pose. The exis ing job exposu e ma ices, based on In e na ional S anda d
Classi ica ion o Occupa ions (ISCO) codes, as gi en by K oll (K oll 2011), deli e
a ela i ely di ec measu e o job exposu e pe occupa ion (job exposu e in his sense
is he exposu e o ei he physical s ain, such as hea y li ing o ca cinogenic agen s,
o psychosocial s ain). Ye he numbe o missing alues c ea ed by ma ching his
exposu e measu e wi h esponden occupa ional in o ma ion (B ünge e al. 2020)
hinde s a longi udinal analysis wi h a high numbe o obse a ion pe iods and di e -
en ia ion be ween mig an and gende g oups. We p opose ha ano he ISCO-based
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 33
measu e, namely occupa ional class, as ou lined in he EGP classi ica ion o social
classes (E ikson and Gold ho pe 1993) can e ec i ely se e as a p oxy o assess-
ing job exposu e, while exhaus ing la ge pa s o he da ase . Theo e ically, wi hin
he amewo k o powe dynamics, indi iduals in highe social class posi ions, as
gi en in he EGP scheme, end o exe g ea e con ol o e hei wo king condi ions
and asks. Highly educa ed p o essionals, being equipped wi h specialized knowl-
edge, can a oid and mi iga e ins ances o exploi a ion (McCa ney e al. 2019).
Fu he mo e, he e exis s a p esumed g adien along he dimensions o social class
conce ning job secu i y and le els o unde - and unemploymen ha subsequen ly
in luence cul u al capi al, isk beha io , and o e all heal h ou comes (Scamble and
Higgs 1999). Al hough he empi ical e idence does no s ic ly adhe e o a linea
ela ionship be ween he EGP scheme and job exposu e, i is obse ed ha a e age
job exposu e ends o inc ease om EGP class I o VII. This is pa icula ly no able
o exposu e o haza dous physical condi ions, limi ed job con ol, and inadequa e
social suppo (K oks ad e al. 2002; Sch ij e s e al. 1998). Simila associa ions
ha e been demons a ed in Ge man da a, whe e hea y li ing, exposu e o ib a-
ions, and con ac wi h chemical oxins we e ound o be mo e p e alen in highe
EGP class alues (and hus lowe social class posi ions) compa ed o lowe EGP
class alues (D agano e al. 2016). Al e na i e ISCO-based me ics, such as he In-
e na ional Socio-Economic Index o Occupa ional S a us (ISEI) and he S anda d
In e na ional Occupa ional P es ige Scale (SIOPS), end o emphasize inancial ac-
o s o in angible ewa ds such as p es ige. Howe e , we con end ha he ela ional
dimensions o he EGP be e encapsula e wo k exposu e han ISEI and SIOPS do.
To ensu e he obus ness o ou indings, we ecalcula ed ou models using ISEI and
SIOPS as p ima y media ing a iables (see Online Appendix).
3 Expec a ions
O e all, we expec ed occupa ional class o ac as a media o o he e ec o mi-
g a ion backg ound on heal h. Conside ing he possible ajec o ies (as in Fig. 1)
and he e idence, we expec ed di e gence o be he dominan pa e n o heal h a-
jec o ies be ween mig an s and na i es (H1). This di e gence can be explained by
di e ing occupa ional class posi ions; i.e., he di e ence in heal h ajec o ies be-
ween mig an s and na i e Ge mans ei he anishes o dec eases signi ican ly when
occupa ional class posi ion is con olled o (H2). Fu he , mig an s a e expec ed o
ind hemsel es mo e equen ly in lowe occupa ional class posi ions (H3).
Because mig an s exhibi di e se backg ounds in e ms o selec ion, cul u e, and
skills, he impac on hei heal h ajec o ies can a y signi ican ly. Eu opean mi-
g an s in Ge many, o ins ance, o en come om coun ies wi h high p o iciency in
Ge man as a second language (Auswä iges Am [Fede al Fo eign O ice] 2015), and
hei educa ional quali ica ions end o hold mo e alue in he Ge man job ma ke
compa ed o mig an s om non-Eu opean coun ies (Kogan 2011). Con e sely, non-
Eu opean mig an s end o display ce ain heal h-p omo ing beha io s such as lowe
a es o alcohol abuse diso de s (Cook e al. 2021), highe ege able consump ion
(Kleise e al. 2010), and g ea e eligiosi y ( an Tube gen and Sind adó i 2011),
K
34 M. Holz, J. Maye l
bu hey a e also mo e p one o ha e expe ienced auma (Nes e ko e al. 2019)
compa ed o hei Eu opean coun e pa s. We he e o e in es iga ed he dis inc ion
be ween Eu opean and non-Eu opean mig an s in an explo a o y manne .
Rega ding gende dispa i ies in he wo kplace, women equen ly encoun e ha-
assmen (Bade e al.2018), a e mo e likely o ha e empo a y con ac s wi h low
wages (Campos-Se na e al. 2013), and ace addi ional s esso s such as ole o e -
load (Pea son 2008). Con e sely, men a e mo e commonly exposed o haza dous
wo king condi ions leading o wo kplace acciden s (Rommel e al. 2016)anda e
less inclined o adop heal h-p o ec i e beha io s (Hille e al. 2017). Gi en he a y-
ing s esso s and isks, p edic ing he physical heal h ajec o ies o each gende
becomes challenging. Mo eo e , conside ing ha iden ical job i les may in ol e
di e en asks and le els o exposu e based on gende (Messing e al. 1994), we
conduc ed sepa a e analyses o women and men, as ecommended in he li e a u e
on occupa ional heal h (Messing e al. 2003).
4Da a
We used indi idual da a om he Ge man SOEP, a longi udinal s udy wi h andom
sampling conduc ed by he Ge man Ins i u e o Economic Resea ch (DIW) since
1984. The SOEP in ol es o e 12,000 p i a e households in Ge many and is col-
lec ed annually. This da ase is pa icula ly sui able o ou pu poses— i s , due o
i s la ge sample size, and second, because i includes an o e sampling o mig an
esponden s om (Sou h)Eas e n Eu ope and Sou hwes Asia (B ücke e al. 2014).
The esponden s in his da ase a e aged 17 yea s and olde . The heal h a iables,
collec ed as epea ed measu es, a e a ailable biannually. In o de o inc ease sample
sizes while a he same ime including as many ime poin s as possible, we analyzed
he wa es 2002–2018 (nine ime poin s). We es ic ed esponden age o a maxi-
mum o 49 yea s in he ini ial wa e o exclude cases o esponden s who po en ially
d op ou o he sample due o e i emen . The inal sample consis s o n= 13,049
esponden s (11,386 na i e Ge mans, 1082 Eu opean mig an s, 581 non-Eu opean
mig an s).
Fo he main ou come a iables, we used he physical heal h scale o he 12-
I em Sho -Fo m Heal h-Rela ed Quali y o Li e Ques ionnai e (Wa e e al. 1996).
Physical heal h is measu ed as gene al heal h s a us, limi a ions in climbing up
s ai s and conduc ing e e yday ac i i ies, he p esence o se e e physical pain in he
p e ious 4 weeks, limi a ions in achie emen due o physical heal h, and gene al
limi a ions due o physical heal h (see Online Appendix o exac wo ding and
scales).
We classi y mig an s as indi iduals who we e no bo n in he Fede al Republic
o Ge many. Na i e Ge mans a e ca ego ized as esponden s whose bi h and he
bi h o bo h o hei pa en s occu ed in Ge many. This analysis does no ake
in o accoun indi iduals wi h an indi ec mig a ion backg ound, in which only one
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 41
Table 4 Pa ame e es ima es om condi ional la en g ow h models o la en slopes o physical heal h
Dependen a iable= la en slope o he la en ac o physical heal h
Women Men
Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Va iable
Eu opean ( e e ence: na i e
Ge man)
–0.026**
(0.009)
–0.078
–0.020* (0.009)
–0.058
–0.019* (0.009)
–0.057
–0.007
(0.010)
–0.021
–0.000 (0.010)
–0.001
0.003 (0.010)
0.007
Non-Eu opean ( e e ence:
na i e Ge man)
–0.021
(0.014)
–0.040
–0.010 (0.014)
–0.020
–0.010 (0.009)
–0.019
–0.034* (0.016)
–0.064
–0.024 (0.016)
–0.046
–0.027 (0.016)
0.007
Occupa ional class – 0.013*** (0.002)
0.135
0.009*** (0.003)
0.092
–0.010*** (0.002)
0.145
0.007*** (0.002)
0.106
Educa ion – – 0.004** (0.001)
0.087
––0.004* (0.002)
0.085
Age – – –0.123*** (0.021)
–0.155
–––0.147*** (0.023)
–0.188
Mean la en slope –0.047*** (0.002) –0.077*** (0.006) –0.066*** (0.006) –0.051*** (0.002) –0.073*** (0.004) –0.065*** (0.005)
Coe icien s a e ep esen ed as b (s anda d e o ) b*, wi h buns anda dized coe icien , b* s anda dized coe icien
***p≤0.001, **p≤0.01, *p≤0.05, (quasi-)me ic a iables mean cen e ed
K
42 M. Holz, J. Maye l
Ge mans, while he pa e n o non-Eu opean mig an s e lec s mo e con inui y. Wi h
he u iliza ion o a singula physical heal h indica o and he absence o g ow h
p ocess modeling, he e is no indica ion ha he p e ailing pa e n is di e gence.
Ye he e a e di e ences in he ajec o ies ha equi e u he analysis.
6.2 Mul i a ia e
F om Table 2we can inspec model- i measu es o each eg ession model. In he
uncondi ional g ow h model, Model 0, he CFI is a 0.959, he RMSEA a 0.028
and he SRMR a 0.067. These alues a e in line wi h he h esholds in he li e a u e
and indica e an adequa e model i (Kline 2023; Ma sh e al. 2009). We he e o e
conclude ha he linea g ow h pa e n is applicable o he la en a iable physical
heal h, i.e., es ic ing ac o loadings o he la en slope pa ame e o he alues as
epo ed in he me hods sec ion e lec s he da a s uc u e su icien ly. Subsequen
models simila ly show an adequa e da a i .
Figu e 43displays he heal h ajec o ies de i ed om he g ow h cu e mod-
els. Compa ed o he uni a ia e case depic ed in Fig. 3, he pa e n o women is
clea e , esembling a di e gence ajec o y. Addi ionally, he e is no disce nible con-
as be ween Eu opean and non-Eu opean women ega ding hei physical heal h
ajec o y, as bo h g oups expe ience a s eepe decline compa ed o na i e Ge man
women. Con e sely, among men, he pa e n is mo e a ied. The dispa i y be ween
Eu opean men and na i e Ge man men aligns mo e wi h a con inui y pa e n, while
non-Eu opean men exhibi a clea di e gence scena io. No ably, o men, he la en
in e cep s, o s a ing poin s, a e consis en ly highe han hose o women. O e all,
women consis en ly ha e lowe heal h ajec o ies han men, wi h mig an women
ha ing he lowes ajec o y.
In he uncondi ional g ow h model, Model 0, he a iances o la en slopes
(women: 0.005 [s anda d e o , SE,= 0.00], men: 0.005 [SE= 0.00]) and in e cep s
(women: 0.338 [SE= 0.008], men: 0.296 [SE= 0.008]) signi ican ly di e om 0
(p≤0.001). This sugges s ha esponden s exhibi subs an ial de ia ion in bo h he
means o la en in e cep s (women: 4.149 [SE= 0.009], men: 4.266 [SE= 0.009]) and
means o la en slopes (women: –0.049 [SE= 0.002], men: –0.052 [SE= 0.002]), in-
dica ing he p esence o a iance ha equi es explana ion.
In he mul i a ia e analysis wi h he la en in e cep o physical heal h as de-
penden a iable (Table 3), we ind a signi ican (p≤0.05) nega i e e ec o non-
Eu opean women (b= –0.086, SE= 0.040, b*= –0.032) compa ed o na i e Ge man
women (Model 1). This e ec disappea s (p> 0.05) as soon as occupa ional class is
in oduced o he model (Model 2) and s ays his way a e con olling o u he
co a ia es (Model 3). No signi ican di e ence (b= –0.102 SE= 0.060, b*= –0.025)
can be ound in he la en in e cep be ween non-Eu opean women and na i e Ge -
man women (Model 1). This esul s ays s able a e con olling o occupa ional
class (Model 2) and co a ia es (Model 3).
Pu di e en ly, when we examine he beginning o he physical heal h ajec o y,
i does no ma ch he depic ion in Fig. 1. We do no see an ini ial heal h bene i
3The es ima es o his plo can be inqui ed in Tables 3and 4.
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 43
Table 5 Pa ame e es ima es om condi ional la en g ow h models o la en in e cep s o occupa ional
class
Dependen a iable= la en in e cep o occupa ional class
Women Men
Model 2 Model 3 Model 2 Model 3
Va iable
Eu opean ( e e ence:
na i e Ge man)
–0.585***
(0.049)
–0.162
–0.577***
(0.044)
–0.160
–0.725***
(0.072)
–0.136
–0.636***
(0.059)
–0.119
Non-Eu opean ( e e -
ence: na i e Ge man)
–0.805***
(0.086)
–0.148
–0.672***
(0.078)
–0.123
–0.867***
(0.108)
–0.110
–0.720***
(0.089)
–0.091
Educa ion – 0.250***
(0.007)
0.483
–0.449***
(0.009) 0.631
Age – –1.486***
(0.112)
–0.174
––1.330***
(0.150)
–0.113
Coe icien s a e ep esen ed as b (s anda d e o ) b*, wi h buns anda dized coe icien , b* s anda dized
coe icien
***p≤0.001, **p≤0.01, *p≤0.05, (quasi)me ic a iables mean cen e ed
o mig an women compa ed o na i e Ge man women. Ins ead, we no ice a heal h
dispa i y a he ou se be ween Eu opean women and na i e Ge man women. This
could be a ibu ed o he inclusion o all mig an s a he han jus ecen ones.
In e ms o men, bo h Eu opean (b= –0.116, SE= 0.041, b*= –0.044) and non-
Eu opean (b= –0.153, SE= 0.061, b*= –0.039) mig a ion backg ounds show signi -
ican nega i e e ec s on he la en in e cep o physical heal h. This sugges s ha
mig an men ha e lowe ini ial s a ing poin s in hei physical heal h ajec o ies.
Howe e , bo h e ec s anish upon con olling o occupa ional class (p> 0.05) in
Model 2 and Model 3. This sugges s ha he ini ial heal h dispa i ies obse ed in
mig an men can be a ibu ed o hei occupa ional class.
Wi h he slope pa ame e o physical heal h as dependen a iable (Table 4),
he ollowing obse a ions can be made: Wi hou co a ia es (Model 1), Eu opean
mig an women (b= –0.026, SE= 0.009, b*= –0.078) ha e, on a e age, a lowe la en
slope o physical heal h han na i e Ge man women. The signi icance o he e ec
disappea s a e con olling o occupa ional class (Model 2) and u he co a ia es
(Model 3). Fo non-Eu opean women he e is no signi ican di e ence in he la en
slope o physical heal h in any model; i.e., na i e Ge man women and non-Eu opean
women expe ience a simila a e o heal h de e io a ion o e ime.
In he analysis o men, he e is no la en slope di e ence in any model o
Eu opean men compa ed o na i e Ge man men. Ye non-Eu opean men ha e, on
a e age, a lowe la en slope in physical heal h han Ge man men (b= –0.034, SE=
0.016, b*= –0.064) in he i s model, bu when occupa ional class is con olled
o , his e ec becomes insigni ican (Model 2 and Model 3). The a e o heal h
de e io a ion is he e o e simila o Eu opean men and na i e Ge man men, whe eas
he heal h o non-Eu opeans declines a an accele a ed a e.
K
44 M. Holz, J. Maye l
In Table 5, he es ima es o la en in e cep o he EGP class scheme as a de-
penden a iable a e shown. In bo h gende g oups, Eu opean (women: b= –0.585,
SE= 0.049, b*– 0.162; men: b= –0.725, SE= 0.072, b*= –0.136) and non-Eu opean
mig an s (women: b= –0.805, SE= 0.086, b*= –0.148, men: b= –0.867, SE= 0.108,
b*= –0.110) ha e lowe le els in he la en in e cep o occupa ional class (Model 1).
This e ec is s able a e con olling o co a ia es (Model 2).
7 Discussion and Conclusion
In his s udy, we examined heal h dispa i ies be ween mig an s and na i e Ge mans
and assessed he ole o occupa ional class in explaining hese dispa i ies. We used
longi udinal da a om a Ge man household panel s udy spanning om 2002 o 2018
and employed la en g ow h cu e analysis o in es iga e how mig a ion backg ound
in luences he ajec o y o physical heal h when occupa ional class is in oduced as
a media o . We in ended o desc ibe heal h ajec o ies o Eu opean mig an s, non-
Eu opean mig an s, and na i e Ge mans, while accoun ing o di e ences be ween
women and men. On a subs an ial le el, his s udy deli e s a es ed explana ion o
pos mig a ion heal h decline in he con ex o he HME phenomenon, which has
been absen in p io esea ch (Holz 2022).
We hypo hesized ha di e gence would eme ge as he p ima y pa e n o heal h
dispa i ies be ween mig an s and na i e Ge mans o e ime. Di e gence, cha ac e -
ized by widening heal h gaps, was con i med in ou indings, aligning wi h exis ing
li e a u e in which i is he p edominan pa e n in socioeconomic heal h gaps (Chan-
dola e al. 2007; Leopold and Engelha d 2011;P us2007; Sacke e al. 2011;Shaw
and K ause 2002). Ou s udy ex ends his unde s anding by demons a ing ha mi-
g an s ha e a highe p opensi y o expe ience he cumula i e disad an age o lowe
socioeconomic posi ions.
Howe e , his asse ion holds ue only o some ex en . Al hough such an e ec is
indeed e iden among women, he si ua ion among men is mo e in ica e. No ably,
he di e ging end is compa able o bo h Eu opean and non-Eu opean women. I
is plausible ha ac o s such as disc imina ion (Scholaske e al. 2019), ole o e load
(Pea son 2008), o wo kplace ha assmen (Bade e al. 2018) could con ibu e o
he exace ba ed decline in heal h. Fu he in es iga ion is wa an ed o unde s and
how hese ac o s in e ac wi h p eca ious wo king condi ions, pa icula ly among
Eu opean women, since occupa ional class does no en i ely explain hei di e ing
heal h ajec o ies. Eu opean women a e po en ially less selec ed o heal h due o
lowe mig a ion cos s (Bo jas 1987) and migh he e o e exhibi a g ea e a iance o
heal h le els, inc easing he likelihood o obse ing lowe heal h s a uses. Mo eo e ,
compa ed o non-Eu opeans, Eu opeans end o ha e less a o able die a y habi s
(Cook e al. 2021; Kleise e al. 2010), which could also in luence hese indings.
The pa h o Eu opean men appea s o ollow a end o con inui y (Engelha d -
Woel le and Leopold 2020). Al hough a heal h gap pe sis s be ween Eu opean men
and na i e Ge mans, i emains s eady o e ime, indica ing a a ia ion in ini ial
heal h s a uses a he han changes in heal h o e ime. This con as diminishes when
occupa ional s a us is conside ed, sugges ing ha occupa ional class may accoun
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 45
o his con inui y. I is concei able ha lowe le els o wo kplace ha assmen and
ewe ca egi ing esponsibili ies (Nu z e al. 2023) alle ia e he s ain associa ed
wi h lowe occupa ional posi ions, he eby sus aining he obse ed s abili y.
On he con a y, he ajec o y o non-Eu opean men exhibi s di e gence, as
an icipa ed. O e ime, he challenges encoun e ed by non-Eu opean mig an s ac-
cumula e, esul ing in an inc easing heal h gap. This phenomenon appea s o be
p ima ily d i en by occupa ional class, which is unsu p ising gi en ha non-Eu o-
pean mig an men a e p edominan ly employed in he lowes occupa ional s a a.
I is plausible ha heigh ened disc imina ion (Hönekopp e al. 2002)and auma
(Nes e ko e al. 2019) add o he e ec o lowe occupa ional s a us on heal h de e-
io a ion. Mo e esea ch is needed o explain and unde s and he in e play o wo k
s ain and psychosocial s esso s o mig an s h oughou he li e cou se.
Ano he no ewo hy obse a ion is ha in he uni a ia e analysis o heal h a-
jec o ies, whe e a simple eg ession line is applied, he obse ed pa e ns di e , po-
en ially leading o al e na i e conclusions. While he ajec o y o women appea s
qui e simila in bo h Fig. 3and Fig. 4, o men he less elabo a ed analysis migh
sugges ha Eu opean men would e en ually con e ge wi h Ge man men in e ms o
physical heal h. Howe e , a close examina ion—conside ing ac o s such as mea-
su emen e o s, g ow h modeling, and longi udinal and mul ig oup measu emen
in a iance, and employing a mul idimensional app oach o heal h (SF-12)— e eals
dis inc ajec o ies, indica ing a pa e n o con inui y o s a us main enance. This
inding is consis en wi h p e ious esea ch, unde sco ing he impo ance o accoun -
ing o longi udinal and mul ig oup measu emen in a iance o a oid signi ican bias
in heal h ajec o y analyses (Holz and Maye l 2024).
Ou s udy bene i s om a longi udinal design, spanning om 2002 o 2018,
allowing o he assessmen o heal h ajec o ies o e ime, which p o ides a mo e
comp ehensi e unde s anding o heal h dispa i ies among mig an s. Fu he , he use
o la en g ow h cu e analysis o model mul iple heal h indica o s as la en a iables
is a obus app oach ha can be e cap u e he complex in e play o a ious heal h-
ela ed ac o s o e ime. Fu he , he inclusion o occupa ional class as a g ow h
cu e allowed us o con ol o po en ial ime- a ian dynamics.
The inco po a ion o mul ig oup longi udinal measu emen in a iance ensu es
ha he obse ed heal h dispa i ies a e no con ounded by measu emen bias ac oss
di e en g oups o ime poin s, enhancing he alidi y o he indings. Con olling o
di e en coun ies o o igin allows o a nuanced examina ion o heal h dispa i ies
among mig an s, acknowledging ha expe iences and heal h ou comes may a y
ac oss di e en na ional backg ounds.
None heless, ou s udy does e eal ce ain limi a ions. Al hough we managed o
accoun o he ime-in a ian s uc u e o he EGP a iable by modeling i as a la en
g ow h cu e, we ound only minimal a iance o e ime, which was s a is ically
insigni ican . Consequen ly, we we e unable o inco po a e he la en slope o EGP
in o ou analysis, es ic ing i s ole o co a ying wi h he co a ia es o in e es . I
would ha e been in iguing o in es iga e how social mobili y, speci ically in e ms
o occupa ional class ajec o ies, in e ac s wi h heal h ajec o ies, shedding ligh on
he ques ion o selec ion—whe he heal hie indi iduals end o secu e be e jobs
K
46 M. Holz, J. Maye l
and consequen ly expe ience mo e g adual declines in heal h. Howe e , due o ou
app oach, his in es iga ion was no easible.
Addi ionally, we we e unable o ully cap u e he heal hy mig an phenomenon.
Despi e ou ini ial heo e ical conside a ions inco po a ing he heal hy mig an pe -
spec i e, logis ical cons ain s, pa icula ly ela ed o cell size issues, p e en ed us
om na owing he sample exclusi ely o ecen mig an s. Gi en ha he pape
p ima ily ocuses on pos mig a ion heal h ajec o ies, we conside his limi a ion
ela i ely mino . Howe e , o comp ehensi ely unde s and heal h selec ion p io o
mig a ion, he impac o mig a ing in o lowe social s a a, and he subsequen long-
e m heal h implica ions, i may be necessa y o conside he du a ion o s ay as
well. Ye ob aining a da ase wi h ex ensi e obse a ion pe iods, mul idimensional
heal h measu es, and dis inc ions be ween coun ies o o igin and gende g oups
p esen s conside able challenges.
While he s udy iden i ies he in luence o occupa ional class on heal h dispa i ies,
i does no del e deeply in o he exac mechanisms h ough which his in luence
ope a es. I emains unclea whe he ac o s such as hose men ioned abo e add
di e ing heal h de imen s be ween mig an g oups. Fu he , we did no accoun o
he e ec o unemploymen and unde employmen , which ha e been demons a ed
o be signi ican ac o s in explaining heal h dispa i ies among mig an s (Espinoza-
Cas o e al. 2019; Rells ab e al. 2016;Yang2021).
I is no ewo hy ha his s udy ocuses on a his o ical mig an popula ion, which
may no comple ely e lec he cu en composi ion o mig an s in Ge many. The
bigges g oup wi hin he Eu opean mig an sample in his s udy consis s o Eas e n
Eu opeans om Poland, Russia, he o me So ie Union, and Romania. The second
la ges g oup is Sou he n Eu opeans, mainly om I aly and G eece. The non-Eu o-
pean g oup in ou sample p edominan ly comp ises Tu kish mig an s, whose p ima y
mig a ion in lux occu ed in he pas and in he con ex o gues wo king con ac s
and amily euni ica ion (Özel and Nauck 1987). To assess he heal h ajec o ies
and hei in e ac ion wi h occupa ional class among he mo e ecen mig an g oups,
se e al decades mus elapse be o e a compa able analysis wi h an equal numbe o
obse a ion pe iods can be ca ied ou .
P edic ing heal h ajec o ies is challenging, bu gi en he simila con ex s o o -
me and ecen mig a ion pa e ns, i is easonable o assume ha pa e ns o heal h
di e ences will pe sis . Namely, based on he con ex o Ge many’s mig a ion poli-
cies, mig an s om bo h o me and ecen wa es can be g ouped ega ding he ease
o mig a ion, which mani es s i sel in s uc u al ba ie s, mig a ion cos s, and em-
ig a ion pa e ns: Be o e 2000, Sou he n Eu opean mig an s in eg a ed mo e easily
and had highe emig a ion a es, while Tu kish wo ke s aced highe mig a ion
cos s and in eg a ion challenges and showed lowe emig a ion a es (Gundel and
Pe e s 2008; Smolny and Riebe 2016). These pa e ns pe sis oday, wi h Eas e n
Eu opean mig an s, being Eu opean Union ci izens, expe iencing lowe mig a ion
cos s and easie in eg a ion, while non-Eu opean mig an s such as e ugees om
Sy ia, A ghanis an, and I aq a e acing highe mig a ion cos s and mo e ba ie s.
Consequen ly, heal h disad an ages a e s ill mo e p onounced o mig an s om
non-Eu opean coun ies han om Eu opean coun ies.
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 47
Fu he mo e, ecen mig a ion pa e ns also esemble hose o he pas among
ce ain e ugee g oups. O e 80% o Uk ainian e ugees in Ge many plan o e-
u n a e he wa (B ücke e al. 2023), simila o o me Yugosla ian e ugees in
he 2000s (Baha e al. 2024). In con as , only abou 38% o Sy ian e ugees in
Ge many in end o e u n (Al Husein and Wagne 2023), esembling Tu kish gues
wo ke s who s ayed longe han ini ially planned. Thus, he simila i ies be ween o -
me and ecen mig a ion pa e ns indica e a con inua ion o heal h dispa i ies based
on coun y o o igin and mig a ion con ex . We assume ha mig a ion policies ha
selec i ely acili a e in eg a ion and na iga ion in he hos coun y will consis en ly
c ea e a g oup wi h mo e ad an ages and a g oup wi h mo e disad an ages, ega d-
less o he coun y. The e o e, we belie e ha he esul s o his s udy ex end beyond
he Ge man case. Howe e , hese assump ions ha e o be es ed empi ically.
This s udy aimed o unco e he ajec o y o mig an heal h, examining whe he
heal h dispa i ies be ween Eu opean mig an s, non-Eu opean mig an s, and na i e
Ge mans e ol e o e ime. Addi ionally, we sough o de e mine whe he a ying
heal h ajec o ies could be elucida ed by occupa ional class, se ing as a p oxy o
job exposu e. Unde s anding he complexi ies unde lying hese inequali ies is chal-
lenging, as hey a e shaped by a combina ion o ac o s including he indi idual’s
coun y o o igin and gende . Ou indings indica e ha di e gence is he p e ailing
pa e n in mig an heal h dispa i ies o e ime, wi h he heal h gap be ween mig an s
and na i e Ge mans widening p og essi ely. This di e gence can be p ima ily ex-
plained by he lowe occupa ional class posi ions inhabi ed by mig an s. Howe e ,
o Eu opean mig an men, he ajec o y mo e closely esembles a pa e n o con-
inui y, esul ing in a sus ained heal h disad an age ha does no exace ba e o e
ime. Fu he mo e, ou analysis highligh s ha women, pa icula ly mig an women,
expe ience mul iple heal h disad an ages compa ed o bo h hei na i e coun e pa s
and men in gene al, leading o a no ably p onounced decline in heal h o e ime.
These esul s emphasize he c i ical impo ance o conside ing he unique expe i-
ences o mig an s om di e en coun ies and o di e en gende s. Conside ing
hese aspec s in bo h policy de elopmen and e o s o p omo e unioniza ion ep e-
sen s essen ial s a egies o add essing labo ma ke disc imina ion and imp o ing
wo king condi ions and hus heal h, especially o mig an women.
Au ho Con ibu ion M. Holz concei ed o he p esen ed idea. M. Holz de eloped he heo y and pe -
o med he compu a ions. M. Holz and J. Maye l e i ied he analy ical me hods. J. Maye l supe ised he
indings o his wo k. M. Holz and J. Maye l discussed he esul s. M. Holz w o e he manusc ip .
Funding Open Access unding enabled and o ganized by P ojek DEAL.
Da a A ailabili y S a emen The da a used in his s udy we e made a ailable by he Ge man Socio-Eco-
nomic Panel S udy a he Ge man Ins i u e o Economic Resea ch (DIW), Be lin. The da a a e no publicly
a ailable due o con aining in o ma ion ha could comp omise he p i acy o esea ch pa icipan s. The
code used du ing he cu en s udy is a ailable om he co esponding au ho on easonable eques o all
in e es ed esea che s.
Decla a ions
Con lic o in e es M. Holz and J. Maye l decla e ha hey ha e no compe ing in e es s.
K
48 M. Holz, J. Maye l
E hical s anda ds All da a analysis p ocedu es we e ca ied ou s ic ly in acco dance wi h he guidelines
o he da a p o ide s and he s anda ds o good scien i ic p ac ice.
Open Access Diese A ikel wi d un e de C ea i e Commons Namensnennung 4.0 In e na ional Lizenz
e ö en lich , welche die Nu zung, Ve iel äl igung, Bea bei ung, Ve b ei ung und Wiede gabe in
jeglichem Medium und Fo ma e laub , so e n Sie den/die u sp ünglichen Au o (en) und die Quelle
o dnungsgemäß nennen, einen Link zu C ea i e Commons Lizenz bei ügen und angeben, ob Ände ungen
o genommen wu den. Die in diesem A ikel en hal enen Bilde und sons iges D i ma e ial un e liegen
eben alls de genann en C ea i e Commons Lizenz, so e n sich aus de Abbildungslegende nich s ande es
e gib . So e n das be e ende Ma e ial nich un e de genann en C ea i e Commons Lizenz s eh und
die be e ende Handlung nich nach gese zlichen Vo sch i en e laub is , is ü die oben au ge üh en
Wei e e wendungen des Ma e ials die Einwilligung des jeweiligen Rech einhabe s einzuholen. Wei e e
De ails zu Lizenz en nehmen Sie bi e de Lizenzin o ma ion au h p://c ea i ecommons.o g/licenses/by/
4.0/deed.de.
Re e ences
Al Husein, Naw as, and Na ascha Wagne . 2023. De e minan s o In ended Re u n Mig a ion among
Re ugees: A Compa ison o Sy ian Re ugees in Ge many and Tu key. In e na ional Mig a ion Re-
iew 57(4):1771–1805.
Allison, Paul D. 2003. Missing da a echniques o s uc u al equa ion modeling. Jou nal o Abno mal
Psychology 112(4):545–557.
A ends-Tó h, Judi , and Fons J. R. an de Vij e . 2008. Family Rela ionships among Immig an s and Ma-
jo i y Membe s in he Ne he lands: The Role o Accul u a ion. Applied Psychology 57(3):466–487.
Assa i, She in. 2017. Combined Racial and Gende Di e ences in he Long-Te m P edic i e Role o
Educa ion on Dep essi e Symp oms and Ch onic Medical Condi ions. Jou nal o Racial and E hnic
Heal h Dispa i ies 4(3):385–396.
Auks -Ma ge i´
c, B anka, and B animi Ma ge i´
c. 2005. Religiosi y and Heal h Ou comes: Re iew o Li -
e a u e. Collegium an opologicum 29(1):365–371.
Auswä iges Am (Fede al Fo eign O ice). 2015. Deu sch als F emdsp ache wel wei . Da ene hebung
2015. Be lin.
Bade , Benjamin, Sebas ian S oe me , Anna Ka ha ina Bade and Tassilo Schus e . 2018. Ins i u ional
disc imina ion o women and wo kplace ha assmen o emale expa ia es. Jou nal o Global Mobili y
6(1):40–58.
Baha , Dany, And eas Haup mann, Cem Özgüzel and Hillel Rapopo . 2024. Mig a ion and Knowledge
Di usion: The E ec o Re u ning Re ugees on Expo Pe o mance in he Fo me Yugosla ia. Re-
iew o Economics and S a is ics 106(2):287–304.
Baue , Geo g F., Ca ola A. Hube , G ego J. Jenny, F i hjo Mülle and Oli e Hämmig. 2009. Socioeco-
nomic s a us, wo king condi ions and sel - a ed heal h in Swi ze land: explaining he g adien in men
and women. In e na ional Jou nal o Public Heal h 54(1):23–30.
Becke , Megan. 2000. Con e ging Heal h Inequali ies in La e Li e-An A i ac o Mo ali y Selec ion?
Jou nal o Heal h and Social Beha io 41(1):106.
Be wing, S e an. 2019. Mig an sel -employmen in Ge many: on he isks, cha ac e is ics and de e mi-
nan s o p eca ious wo k. In Sel -Employmen as P eca ious Wo k, ed. Wie eke Conen and Joop
Schippe s, 186–214. Edwa d Elga Publishing.
Bollen, Kenne h A., and Pa ick J. Cu an. 2006. La en cu e models: a s uc u al equa ion pe spec i e.
Hoboken, NJ: John Wiley & Sons.
Bo jas, Geo ge. 1987. Sel -selec ion and he ea nings o immig an s. NBER Wo king Pape Se ies No.
2248. Na ional Bu eau o Economic Resea ch.
B ücke , He be , And eas E e, Ma kus M. G abka, Yuliya Kosyako a, Wenke Niehues, Nina Ro he ,
C. Ka ha ina Spieß, Sabine Zinn, Ma in Buja d, Ad iana R. Ca dozo Sil a, Jean Philippe Décieux,
Am ei Maddox, Nadja Milewski, Leno e Saue , Sophia Schmi z, Sil ia Schwanhäuse , Manuel
Siege , Hans S einhaue and Ke s in Tanis. 2023. Uk ainian Re ugees in Ge many: E idence F om
a La ge Rep esen a i e Su ey. Compa a i e Popula ion S udies – Zei sch i ü Be ölke ungswis-
senscha 48:395–424.
K
Can Di e ing Occupa ional Class Posi ions Explain Mig an Heal h Inequali ies? Di e ences... 49
B ücke , He be , Simone Ba sch, Jan Goebel, Simon Kühne, Elisabe h Liebau, Pa a i T übswe e , In-
g id Tucci and Jü gen Schupp. 2014. The new IAB-SOEP Mig a ion Sample: an in oduc ion in o he
me hodology and he con en s. SOEP Su ey Pape s 216:Se ies C.
B ünge , Ma in, Sebas ian Be ne and Ka la Spy a. 2020. Be u s ä igkei als P oxy ü A bei sbelas un-
gen? Rou ineda enanalyse am Beispiel de Rehabili a ion. Das Gesundhei swesen 82(S 01):S41–S51.
Campos-Se na, Ja ie , Elena Ronda-Pé ez, Lucia A azcoz, Ben e E. Moen and Fe nando G. Bena ides.
2013. Gende inequali ies in occupa ional heal h ela ed o he unequal dis ibu ion o wo king and
employmen condi ions: a sys ema ic e iew. In e na ional Jou nal o Equi y in Heal h 12:57.
Chandola, Ta ani, Jane Fe ie, Amanda Sacke and Michael Ma mo . 2007. Social inequali ies in sel
epo ed heal h in ea ly old age: ollow-up o p ospec i e coho s udy. BMJ (Clinical esea ch ed.)
334(7601):990.
Clumeck, Nicolas, Chan al Kempenae s, Isabelle Godin, Michèle D amaix, Ma cel Ko ni ze , Paul
Linkowski and F ance Ki el. 2009. Wo king condi ions p edic incidence o long- e m spells o sick
lea e due o dep ession: esul s om he Bels ess I p ospec i e s udy. Jou nal o Epidemiology and
Communi y Heal h 63(4):286–292.
Cook, Won Kim, Xinjun Li, K is ina Sundquis , Kenne h S. Kendle , Jan Sundquis and Ka he ine J. Ka -
ike -Ja e. 2021. D inking cul u es and socioeconomic isk ac o s o alcohol and d ug use diso de s
among i s - and second-gene a ion immig an s: A longi udinal analysis o Swedish popula ion da a.
D ug and Alcohol Dependence 226:108804.
Co na, Lau ie M. 2013. A li e cou se pe spec i e on socioeconomic inequali ies in heal h: a c i ical e iew
o concep ual amewo ks. Ad ances in Li e Cou se Resea ch 18(2):150–159.
Damelang, And eas, Sabine Ebenspe ge and Felix S ump . 2021. Immig an s’ Labou Ma ke Disad-
an ages Ac oss Wes e n Eu ope: he Role o Composi ion and Con ex . Jou nal o In e na ional
Mig a ion and In eg a ion 22 (4):1525–1550.
Da mon, Nicole, and My iam Khla . 2001. An o e iew o he heal h s a us o mig an s in F ance, in
ela ion o hei die a y p ac ices. Public Heal h Nu i ion 4(2):163–172.
Da ido , Eldad, Ba Meuleman, Jan Cieciuch, Pe e Schmid and Jaak Billie . 2014. Measu emen Equi -
alence in C oss-Na ional Resea ch. Annual Re iew o Sociology 40(1):55–75.
Di Napoli, An eo, Alessand a Rossi, Gian anco Alicand o, Ma ina Ven u a, Luisa F o a and Alessio
Pe elli. 2021. Salmon bias e ec as hypo hesis o he lowe mo ali y a es among immig an s in
I aly. Scien i ic Repo s 11(1):8033.
D agano, Nico, Mo en Wah endo , Ka h in Mülle and Tho s en Lunau. 2016. A bei und gesundhei liche
Ungleichhei : Die ungleiche Ve eilung on A bei sbelas ungen in Deu schland und Eu opa. Bundes-
gesundhei sbla , Gesundhei s o schung, Gesundhei sschu z 59(2):217–227.
Edbe g, Ma k, Sean Clea y and Ami a Vyas. 2011. A ajec o y model o unde s anding and assess-
ing heal h dispa i ies in immig an / e ugee communi ies. Jou nal o Immig an and Mino i y Heal h
13(3):576–584.
Engelha d -Woel le , Hen ie e, and Liliya Leopold. 2020. Gesundhei liche Ungleichhei im Lebenslau .
In Handbuch Gesundhei ssoziologie, eds. Pe e K iwy and Monika Jungbaue -Gans, 359–372. Wies-
baden: Sp inge Fachmedien Wiesbaden.
E ikson, Robe , and John H. Gold ho pe. 1993. The cons an lux: A s udy o class mobili y in indus ial
socie ies. Ox o d: Cla endon.
Espinoza-Cas o, Be na da, Luis E. Vásquez Rueda, Rossana V. Mendoza Lopez and Ka ja Radon. 2019.
Wo king Below Skill Le el as Risk Fac o o Dis ess Among La in Ame ican Mig an s Li ing in
Ge many: A C oss-Sec ional S udy. Jou nal o Immig an and Mino i y Heal h 21(5):1012–1018.
Fe a o, Kenne h F., Te yana P. Shippee and Ma kus H. Scha e . 2009. Cumula i e inequali y heo y o
esea ch on aging and he li e cou se.InHandbook o heo ies o aging, 2nd edi ion, eds.Ve n
L. Bengs on, Daphna Gans, No ella M. Pulney and Me il Sil e s ein, 413–433. New Yo k: Sp inge
Publishing Company.
Fujishi o, Kao i, and Anneka in Hoppe. 2020. Towa d a Li e-Cou se Pe spec i e o Mig an Wo ke
Heal h and Well-Being. In Heal h, Sa e y and Well-being o Mig an Wo ke s: New Haza ds, New
Wo ke s, eds. Angeli San os and F ancisco Díaz B e ones, 23–42. Cham: Sp inge .
Fulle -Thomson, Esme, Sa ah B ennens uhl, Rachel Coope and Diana Kuh. 2016. An in es iga ion o he
heal hy mig an hypo hesis: P e-emig a ion cha ac e is ics o hose in he B i ish 1946 bi h coho
s udy. Canadian Jou nal o Public Heal h 106(8):e502–8.
Geye , Sieg id, and Richa d Pe e . 1999. Occupa ional heal h. Occupa ional s a us and all-cause mo ali y.
A s udy wi h heal h insu ance da a om No d hein-Wes alen, Ge many. Eu opean Jou nal o Public
Heal h 9(2):114–118.
K
50 M. Holz, J. Maye l
Gonzalez-Mulé, E ik, and Be hany Cockbu n. 2017. Wo ked o Dea h: The Rela ionships o Job Demands
and Job Con ol wi h Mo ali y. Pe sonnel Psychology 70(1):73–112.
Gundel, Sebas ian, and Heiko Pe e s. 2008. Wha de e mines he du a ion o s ay o immig an s in Ge -
many? In e na ional Jou nal o Social Economics 35(11):769–782.
Hiesinge , Ka olin, and Silke Topho en. 2019. Job equi emen le el, wo k demands, and heal h: a p ospec-
i e s udy among olde wo ke s. In e na ional A chi es o Occupa ional and En i onmen al Heal h
92(8):1139–1149.
Hille , Julia, Ka h ina Scha z and Hans D exle . 2017. Gende in luence on heal h and isk beha io in
p ima y p e en ion: a sys ema ic e iew. Jou nal o Public Heal h 25(4):339–349.
Holz, Manuel. 2022. Heal h inequali ies in Ge many: Di e ences in he ‘Heal hy mig an e ec ’o Eu-
opean, non-Eu opean and in e nal mig an s. Jou nal o E hnic and Mig a ion S udies 48(11):2620-
2641.
Holz, Manuel, and Jochen Maye l. 2024. Mig an Heal h Inequali ies o Unequal Measu emen s? Tes -
ing o C oss-cul u al and Longi udinal Measu emen In a iance o Subjec i e Physical and Men al
Heal h. me hods, da a, analyses 18.1:20.
Ho en, Hanno, Mo en Wah endo and Johannes Sieg is . 2015. Occupa ional posi ion, wo k s ess and
dep essi e symp oms: a pa hway analysis o longi udinal SHARE da a. Jou nal o Epidemiology and
Communi y Heal h 69(5):447–452.
Hönekopp, Elma , Gisela Will and S e an Rühl. 2002. Mig an s, mino i ies and employmen in Ge many:
Exclusion, disc imina ion and an i-disc imina ion. (RAXEN Repo / Eu opean Moni o ing Cen e
on Racism and Xenophobia, Na ional Focal Poin Ge many). Bambe g: eu opäisches o um ü mi-
g a ionss udien (e ms) Ins i u an de Uni e si ä Bambe g.
Ja ana, San osh, Ken Richa dson and Samba Si a Rao Pasupule i. 2018. In es iga ing he Dynamics o
Mig a ion and Heal h in Aus alia: A Longi udinal S udy. Eu opean jou nal o popula ion = Re ue
eu opeenne de demog aphie 34(4):519–565.
Johnson, Je ey V., Wal e S ewa , Ellen M. Hall, Pee e F edlund and To es Theo ell. 1996. Long- e m
psychosocial wo k en i onmen and ca dio ascula mo ali y among Swedish men. Ame ican Jou nal
o Public Heal h 86(3):324–331.
Kal e , F ank, and Nadia G ana o. 2018. Mig a ion und e hnische Ungleichhei au dem A bei sma k .
In A bei sma k soziologie, eds. Ma in Ab aham and Thomas Hinz, 355–387. Wiesbaden: Sp inge
Fachmedien Wiesbaden.
Kleise , Ch is ina, Ge B. M. Mensink, Hannelo e Neuhause , Liane Schenk and Bä bel-Ma ia Ku h.
2010. Food in ake o young people wi h a mig a ion backg ound li ing in Ge many. Public Heal h
Nu i ion 13(3):324–330.
Kline, Rex B. 2023. P inciples and p ac ice o s uc u al equa ion modeling. New Yo k: The Guil o d
P ess.
Kogan, I ena. 2011. New Immig an s—Old Disad an age Pa e ns? Labou Ma ke In eg a ion o Recen
Immig an s in o Ge many. In e na ional Mig a ion 49(1):91–117.
K oks ad, S eina , K is en Ringdal and S eina Wes in. 2002. Classi ying people by social class in popula-
ion based heal h su eys. No sk epidemiologi 12(1):19–25.
K oll, La s E ic. 2011. Kons uk ion und Validie ung eines allgemeinen Index ü die A bei sbelas ung
in be u lichen Tä igkei en au Basis on ISCO-88 und KldB-92. Me hoden, Da en, Analysen (mda)
5(1):63–90.
Lahelma, Ee o, Pekka Ma ikainen, Ossi Rahkonen, E a Roos and Peppiina Saas amoinen. 2005. Oc-
cupa ional class inequali ies ac oss key domains o heal h: esul s om he Helsinki Heal h S udy.
Eu opean Jou nal o Public Heal h 15(5):504–510.
Landsbe gis, Paul A. 2010. Assessing he con ibu ion o wo king condi ions o socioeconomic dispa i ies
in heal h: a commen a y. Ame ican Jou nal o Indus ial Medicine 53(2):95–103.
Lee, Sunmin, Allison H. O’Neill, Emily S. Iha a and Da id H. Chae. 2013. Change in sel - epo ed heal h
s a us among immig an s in he Uni ed S a es: associa ions wi h measu es o accul u a ion. PloS one
8 (10):e76494.
Leong, F ede ick T. L., Donald E. Egge h and Michael A. Flynn. 2014. A Li e Cou se Pe spec i e on
Immig an Occupa ional Heal h and Well Being. In Con empo a y Occupa ional Heal h Psychology,
97–113. Hoboken, NJ: John Wiley & Sons.
Leopold, Liliya, and Hen ie e Engelha d . 2011. Bildung und Gesundhei sungleichhei im Al e : Di-
e genz, Kon e genz ode Kon inui ä ? Kölne Zei sch i ü Soziologie und Sozialpsychologie
63(2):207–236.
Ma sh, He be W., Beng Mu hén, Tihomi Aspa ouho , Oli e Lüd ke, Alexande Robi zsch, Alexand e
J. S. Mo in and Ul ich T au wein. 2009. Explo a o y S uc u al Equa ion Modeling, In eg a ing CFA
K
