Ga cía-Mo án, E a; Jiang, Ming-Jin; Rachinge , Heiko
A icle
Minimum wage e ec s on job a achmen : A gende
pe spec i e
Jou nal o Economics and S a is ics
P o ided in Coope a ion wi h:
De G uy e B ill
Sugges ed Ci a ion: Ga cía-Mo án, E a; Jiang, Ming-Jin; Rachinge , Heiko (2024) : Minimum wage
e ec s on job a achmen : A gende pe spec i e, Jou nal o Economics and S a is ics, ISSN
2366-049X, De G uy e Oldenbou g, Be lin, Vol. 244, Iss. 1/2, pp. 83-112,
h ps://doi.o g/10.1515/jbns -2022-0012
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/333275
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 ps://c ea i ecommons.o g/licenses/by/4.0/
E a Ga cía-Mo án, Ming-Jin Jiang and Heiko Rachinge *
Minimum Wage Effec s on Job A achmen :
A Gende Pe spec i e
h ps://doi.o g/10.1515/jbns -2022-0012
Recei ed Ma ch 8, 2022; accep ed Ap il 25, 2023
Abs ac : We examine whe he he effec s o he in oduc ion o a minimum wage
on low-pay employmen du a ion in Ge many in 2015 a e he e ogeneous by gende .
In o de o disen angle he effec s on women and men, we es ima e a du a ion
model wi h unobse ed he e ogenei y in which we allow gende diffe ences and
diffe ences be o e and a e he in oduc ion o he minimum wage. We find ha
he e o m does affec women and men diffe en ly, in pa icula , i mainly
inc eases men’s job a achmen . These gende diffe ences in job a achmen a e he
s onges o ull- ime employmen . In consequence, al hough he minimum wage
may ha e been se up as a gende -neu al ins umen , in an indi ec way, i affec s
women and men diffe en ly. We discuss diffe en mechanisms ha could accoun
o ou esul and ca y ou se e al obus ness checks.
Keywo ds: employmen , minimum wage, du a ion, gende diffe ences
JEL Classifica ion: J31, J2, J38, J16
1 Mo i a ion
We aim a examining gende diffe ences in he effec s o a minimum wage on
employmen du a ion o low paid wo ke s. Gende p o es o be an impo an
de e minan o labo ma ke ou comes especially in he low wage sec o . In
pa icula , in Eu ope, among all employees, he incidence o low pay is wice as
A icle No e: This a icle is pa o he special issue “Gende Economics“published in he Jou nal o
Economics and S a is ics. Access o u he a icles o his special issue can be ob ained a www.deg uy e .
com/jou nals/jbns .
*Co esponding au ho : Heiko Rachinge , Depa men o Applied Economics, Uni e si a de les Illes
Balea s, C a, de Valldemossa, km 7.5, 07122 Palma de Mallo ca, Spain, E-mail: [email p o ec ed].
h ps://o cid.o g/0000-0002-2782-6545
E a Ga cía-Mo án, Ins i u e o Ad anced S udies (IHS), Jose s äd e S aße 39, 1080 Vienna, Aus ia,
E-mail: [email p o ec ed]m
Ming-Jin Jiang, Ma bella In e na ional Uni e si y Cen e (MIUC), A enida Don Jaime de Mo a y A agón,
sn Finca El Pinillo, 29601 Ma bella, Spain, E-mail: [email p o ec ed]
Jou nal o Economics and S a is ics 2024; 244(1–2): 83–112
Open Access. © 2023 he au ho (s), published by De G uy e . This wo k is licensed unde he
C ea i e Commons A ibu ion 4.0 In e na ional License.
high o emale wo ke s han o male wo ke s, whe e low pay is defined as less han
wo- hi ds o he na ional median o he na ional mean o g oss hou ly wages (Schnabel
2016). Fo Ge many, his is also ue. Da a om he Fede al S a is ical Office o Ge many
shows ha 26.4 % o employed women a e on low pay employmen , while o men his
numbe only is 16.3 % (Fede al S a is ical Office 2018).
This is pa ially a consequence o almos 50 % o employed women being
wo king pa - ime and 66 % o wo ke s wi h mini-jobs being women (Fede al
Employmen Agency 2019).
The ac ha women a e o e - ep esen ed in low paying jobs is a ac o con ibu ing
o he obse ed gende wage gap as well as ha ing consequences o li e- ime income,
leading o lowe pensions and diffe ences in po e y s a us. In pa icula , women ha e
on a e age hal he amoun o pension han men ha e, making hem mo e likely han
men o be a isk o being poo when old. Indeed, in 2018, 20 % o o e 65 yea old women
we e a isk o being poo , while his pe cen age was 17 % o men (Fede al S a is ical
Office 2019). The e o e, unde s anding whe he a minimum wage policy can help o
educe gende diffe ences is a ele an issue om a public policy pe spec i e.
The s anda d model o compe i i e labo ma ke s p edic s ha he in oduc ion o a
minimum wage ha is se highe han he compe i i e equilib ium wage educes
employmen among low-skilled wo ke s. Howe e , empi ical esea ch has dispu ed his
heo y. Fo he US, he seminal pape by Ca d and K uege (1994) ound no effec s on he
employmen in he US as ood indus y a e he 1992 inc ease in New Je sey minimum
wage. Since hen, o he s udies ha e ound no effec s o small effec s going in bo h
di ec ions. A me a-analysis o 23 minimum wage s udies by Belman and Wol son (2014)
finds ha hese s udies we e as likely o find posi i e as nega i e effec s o he minimum
wage and he es ima e ypically was close o ze o. Fo he UK, S ewa (2004) finds ha
he in oduc ion o he na ional minimum wage in Ap il 1999 did no ha m he
employmen p ospec s o he low-paid. Howe e , Dickens e al. (2015) find ha he
in oduc ion o a minimum wage educes employmen e en ion o pa - ime wo king
women. They a gue ha S ewa (2004) finds no effec ashedidno conside heeffec s
on he pa - ime employmen o women who a e one o he mos ulne able g oup o
changes in he minimum wage.
Thus, he esul s in he li e a u e o no effec s on employmen migh hide he -
e ogeneous effec s by gende .
The e a e se e al empi ical s udies o he in oduc ion o he minimum wage
in 2015 in Ge many. This minimum wage o 8.50 eu o conside ably exceeds he
a e age wage o he affec ed wo ke s be o e he e o m o 6.01 eu o (Bossle and
Schank 2017) and, hus po en ially ma e s o labo ma ke ou comes. These
s udies exploi a ia ion in he minimum-wage a he egional o es ablishmen -
le el o ea men assignmen in a diff-in-diffse ing. In pa icula , Caliendo
e al. (2018) and Bossle and Ge ne (2016) find ha in he sho un he
84 E. Ga cía-Mo án e al.
in oduc ion o he minimum wage had a mode a e nega i e effec on o e all
employmen (see Caliendo e al. 2019 o a li e a u e su ey on he causal effec s o
he minimum wage in oduc ion in Ge many). Ga loff(2019) and Schmi z (2019)
also find small nega i e (o insignifican ) effec s on o e all employmen d i en
mainly by less ma ginal employmen and Hol emölle and Pohle (2020) find
nega i e effec s on ma ginal employmen while finding small posi i e effec s on
egula employmen .
1
As o he effec s on wages and he gende wage gap, Boll e al. (2015) ob ain ia
simula ion s udies ha he minimum wage inc eases women’s wages and hus
dec eases he gende gap. Bossle and Schank (2017), on he o he hand, es ima e
ha he in oduc ion o he minimum wage can accoun o abou hal o he ecen
dec ease in wage inequali y. They findaposi i ewageeffec on all affec ed em-
ployees (up o he median), bu also a posi i e wage effec on he exis ing jobs.
Finally, Caliendo and Wi b od (2020) find ha he minimum wage educed gende
wage dispa i ies, in pa icula , he egional gende wage gap.
We ake ano he app oach by analyzing gende diffe ences in he 2015 minimum
wage e o m’seffec on low-wage wo ke s’employmen du a ion. In pa icula , we
use da a o Ge many o he pe iod 2013–2016 and es ima e a du a ion model wi h
unobse ed he e ogenei y. We allow o gende diffe ences and diffe ences be o e
and a e he in oduc ion o he minimum wage.
The mos ela ed pape o ou s is Phimis e and Theodossiou (2009). They
examine gende diffe ences in he du a ion o low pay employmen spells p io
o and a e he in oduc ion o a na ional minimum wage in 1999 in he UK.
Specifically, hey conside how gende diffe en ials in he du a ion o low-paid
jobs and he p obabili y o p og essing o a highe -paid job change due o he
in oduc ion o a minimum wage. While hey in es iga e gende diffe ences by
es ima ing sepa a e specifica ions o men and women, we ca y ou a join es i-
ma ion which allows us o join ly es o gende diffe ences and o e o m effec s.
While sepa a e es ima ions a e mainly a heu is ic de ice, ou join es ima ion
allows o p ope ly disen angle effec s om gende and om he e o m. In
pa icula , i enables us o s a is ically es whe he men and women’sjob
a achmen diffe s be o e and a e he e o m. In effec , we find ha he e o m
affec s mainly men’s sepa a ions. This leads he job sepa a ion a es o women and
men o con e ge.
1Also, he e a e a ew pape s s uc u ally es ima ing he effec s o he in oduc ion o a minimum
wage in Ge man employmen , see Blöme e al. (2018) and Mülle and S eine (2010).
Minimum Wage and Job A achmen 85
2 Me hodology
We exploi he quasi expe imen al na u e o an in oduc ion o a minimum wage
in Ge many a he ede al le el in 2015. In pa icula , we analyze he du a ion o
low pay employmen o men and women, as well as he p obabili y o becoming
unemployed o inac i e. Fi s , we es ima e a single spell disc e e ime haza d
model wi h one exi ype o unemploymen o inac i i y. A a second s age,
unobse ed he e ogenei y is inco po a ed by assuming a p opo ional haza d
model wi h Gaussian Mixing (And ews e al. 2002; Lancas e 1990).
Each pe son in low pay employmen po en ially exi s low pay in an in e al
j−1,j
[]
o one s a e. The la en a iable, T
un
ep esen s he po en ial ime in low pay
wi h exi o ype un. The andom a iable T= min Tun
()
ep esen s he low pay
du a ion. The haza d h
j
co esponds o he p obabili y o an exi o s a e un du ing
pe iod jgi en ha he low pay du a ion las ed o j−1. Finally, he su i al unc ion is
defined as Sj=Πj
s=11−hs
()
.
Specifically, we model he exi ype using a p opo ional haza d model wi h
Gaussian mixing, i.e.
hjx′
i, i
()
=hm,u
j01−d
[]
1−d
[]
+hm,
j01−d
[]
d +h ,u
j0d 1−d
[]
+h ,
j0d d
()
i
⋅exp 1−d
[]
1−d
[]
x′
iβm,u+1−d
[]
d x′
iβm, +d 1−d
[]
x′
iβ ,u+d d x′
iβ ,
()
,
(1)
whe e hp,q
j0,p∈m,
{}
;q∈u,
{}
a e baseline haza ds o any combina ion o men
(p=m) and women (p= ) as well as un ea ed (q=u) and ea ed (q= ) (i.e. be o e and
a e he in oduc ion o he minimum wage).
i
is a andom a iable cap u ing
unobse ed he e ogenei y so ha u=log
()
is no mally dis ibu ed. x
i
is ec o o
co a ia es including indi idual and job cha ac e is ics (specified below).
The p obabili y o exi ing employmen is gi en by,
Π=∑
∞
j=1
hjSj−1,
he expec ed low pay du a ion gi en an exi , by
E=1
Π∑
∞
j=1
jhjSj−1,
and finally he expec ed low du a ion by ΠE.
While Phimis e and Theodossiou (2009) sepa a ely es ima e he model –
allowing o gende diffe ences –be o e and a e he in oduc ion o he minimum
wage, we join ly model bo h dimensions. The e o e, because o he wo dummies –
86 E. Ga cía-Mo án e al.
gende and in oduc ion o minimum wage –and hei in e ac ions bo h baseline
haza ds and he impac o each cha ac e is ic a e allowed o a y ac oss male and
emale, as well as ac oss he samples be o e and a e he in oduc ion o he min-
imum wage. Mo e impo an ly, his enables us o do in e ence along bo h dimensions
a he han me ely desc ip i ely compa e hem. In pa icula , i enables us o s a-
is ically es whe he he e a e gende diffe ences in he job a achmen and in he
ole o o he co a ia es be o e and a e he e o m.
3 Da a
We use da a om he Ge man Socio Economic Panel, SOEP( 34) (2019) o he yea s
2013–2016. We use spell mon hly employmen da a –diffe en om mos p e ious
s udies –allowing us o ha e a mo e comple e pic u e o how employmen dy-
namics change due o he in oduc ion o a minimum wage ac oss gende . E e
since i s s a in 1984, he SOEP con ained a calenda sec ion asking abou
employmen s a us as o Janua y h ough Decembe o he p e ious yea . An
employmen s a us spell can ake se e al alues such as ull- ime employed, pa -
ime employed, unemployed, housewi e/househusband, sho -wo k hou s, ma e -
ni y lea e. We define employmen in a b oad sense i he spell ype is ei he ull (1),
pa (3), sho -wo k hou s (2) o mini-job (15). We define ou o he labo o ce
(unemployed o inac i e) i he spell ype is unemployed (5), housewo k (10) o
o he (12). Mo eo e , he e a e spells whe e indi iduals epo o be in diffe en
s a uses a he same ime (o e lapping spells). Fo example hey epo o be
wo king pa - ime and o he same spell ime hey epo being a housewi e. We
ea hose spells as employed i one o he s a uses epo ed co esponds o ou
defini ion o employmen while he o he spell is epo ed o be diffe en o
egis e ed unemployed.
We define low-wage ea ne s as hose whose hou ly wage belongs o he
bo om 25 h pe cen ile o he hou ly wage dis ibu ion o each yea . Using ela i e
measu es is s anda d in he li e a u e (see Phimis e and Theodossiou 2009, Cap-
pella i and Jenkins 2008). To cons uc ou h eshold o low-wage ea ning we use
con ac ual weekly wo king hou s and g oss mon hly wages adjus ed by a e age
weeks pe mon h o calcula e he hou ly wage.
F om he cons uc ed da a se , wo inflow samples o low-pay spells we e
d awn co e ing wo wo-yea pe iods —p io o and pos he in oduc ion o he
minimum wage. Specifically, low pay spells s a ing be ween Janua y 2013 and
Decembe 2014 a e included in he p e-minimum wage sample, and spells s a ing
be ween Janua y 2015 and Decembe 2016 a e included in he pos -minimum wage
Minimum Wage and Job A achmen 87
sample. All con inuing spells a e ea ed as censo ed i hey con inued beyond he
end cu -offpoin s, i.e. Decembe 2014 and Decembe 2016 espec i ely.
3.1 Co a ia es
Va ious a iables a he household and indi idual le el –such as sec o , wo k
expe ience, fi m enu e, educa ion, age, ma i al s a us –a e included as co a ia es.
Table 1 p o ides desc ip i e s a is ics o ou a iables o in e es . P e ious
esea ch shows ha no only gende bu also human capi al, job enu e
and expe ience and pa - ime wo king, a e all impo an in low pay mobili y
(Phimis e and Theodossiou 2009).
In ou sample, women accoun o mo e han 60 % o low paid jobs. We can
obse e ha women and men a e simila be o e and a e he e o m in mos
Table :Desc ip i e s a is ics by gende and yea .
– –
Women Men Women Men
Age . . . .
Seconda y educa ion . . . .
Te ia y educa ion . . . .
Ma ied . . . .
Single . . . .
Child en unde in household . . . .
Residence in Eas Ge many = . . . .
Hou ly wages . . . .
Ac ual weekly wo king hou s . . . .
Full ime (> h) = . . . .
Mini-job = . . . .
Manu ac u ing . . . .
Se ices . . . .
Full- ime expe ience, yea s . . . .
Tempo a y con ac = . . . .
Sec o wi h minimum wage = . . . .
Fi m size < . . . .
Fi m size – . . . .
Fi m size > . . . .
Public sec o . . . .
N, , , ,
Sou ce: SOEP: o . Means o he diffe en cha ac e is ics and numbe o obse a ions sepa a ed by he pe iod
be o e and a e he e o m and by gende .
88 E. Ga cía-Mo án e al.
co a ia es. Men ha e sligh ly less educa ion (less wi h seconda y educa ion) and a e
less likely ma ied han women. A s a k diffe ence is he sha e o indi iduals who a e
in a ull- ime employmen o in a mini-job. The sha e o men in a ull- ime low paid
job is double he sha e o women in ull- ime while he opposi e is ue o mini-jobs.
The a e age hou ly wage in he da a-se was, be o e he in oduc ion o he mini-
mum wage, 6.86 eu o o women and 6.88 eu o o men, and a e , 7.57 eu o o
women and 7.27 eu o o men. This eflec s he wage inc ease due o he in oduc ion
o he minimum wage. Su p isingly, a e he e o m he e is an uncondi ional
nega i e gende wage gap, i.e. in his wage segmen , he a e age wage o women is
la ge han he one o men.
3.2 Associa ion Analysis Be ween Gende , Job Sepa a ion and
Re o m
We use con ingency ables o analyze he associa ion be ween job sepa a ions and
e o m o women and o men. Table 2 shows he absolu e equency oge he wi h
he p opo ion condi ional on he e o m s a us. Howe e , he educ ion is la ge
o men ( om 1.32 % o 0.87 %) han o women ( om 1.01 % o 0.88 %), sugges ing
ha job sepa a ion a es o women and men migh con e ge a e he e o m.
Pea son χ2
1 es s confi m ha while o women he e is no associa ion be ween
e o m and sepa a ions ( es s a is ic o 1.81 wi h a p- alue o 0.178), o men he e is
one ( es s a is ic o 10.17 wi h a p- alue o 0.00).
Table :Con ingency ables: Associa ion be ween e o m and job sepa a ion by gende .
(a) Women (b) Men
Job sepa a ion Job sepa a ion
Re o m No Yes To al Re o m No Yes To al
Be o e , , Be o e , ,
. . . .
A e , , A e , ,
. . . .
To al , , To al , ,
. . . .
Sou ce: SOEP, o . Absolu e equencies o be o e and a e e o m (column) and sepa a ion ( ow) oge he
wi h he p opo ion condi ional on he e o m s a us by gende .
Minimum Wage and Job A achmen 89
3.3 Uncondi ional Su i al Func ions
Figu e 1 displays he Kaplan-Meie uncondi ional su i al unc ions by gende and
e o m s a us. The su i al unc ion es ima es he p obabili y o a low-paid job o
su i e pe iod jcondi ional on ha ing su i ed un il hen. Thus, we can obse e
om he figu e ha he su i al o men a e he e o m ( he line “male a e ”)is
he highes . Thus, he p obabili y o emaining in a low paid job is highes o men
who en e a job in 2015–2016. Compa ed o he su i al unc ion be o e he e o m
o men (“male be o e”), hey seem o be diffe en . I is less clea i women’s
su i al unc ions a e diffe en be o e (“ emale be o e”)anda e (“ emale a e ”)
he e o m and whe he a e he e o m, men’s and women’ssu i al unc ions
a e diffe en . We can es o mally whe he hese su i als a e diffe en by ca -
ying ou log- ank es s o equali y. We do pai wise compa isons o he su i al
unc ions. The null hypo hesis is ha he su i al unc ions a e equal. Fo he p e-
e o m pe iod, 2013–2014, we ejec he equali y o he su i al unc ions o
women and men (p- alue = 0.00). Howe e , o he pos - e o m pe iod, 2015–2016,
he equali y o he su i al unc ions o women and men canno be ejec ed
(p- alue = 0.98). Be o e he e o m, men we e less likely o emain in a job han
women, while a e he e o m, he e a e no diffe ences in he uncondi ional
su i al o a job. On he o he hand, he uncondi ional su i als a e equal be o e
and a e he e o m o women (p- alue = 0.30), while o men we ejec he
equali y o hei su i al be o e and a e he e o m a he 1 % le el
(p- alue = 0.00).
.96 .97 .98 .99 1
P opo ion Su i ing
0 5 10 15 20 25
Du a ion
male be o e emale be o e
male a e emale a e
Su i al unc ion
Figu e 1: Kaplan Meie su i al unc ions. Sou ce: SOEP, 2013 o 2016. Kaplan Meie su i al unc ions
calcula ed o du a ions up o 24 mon hs sepa a ed by gende and by be o e and a e he e o m.
90 E. Ga cía-Mo án e al.
In pa icula , he exi a e o men dec eases while he one o women emains
cons an .
Fi s we conside a supply side mechanism whe e wo ke s eac o he in o-
duc ion o a minimum wage. Recall ha he in oduc ion o he minimum wage
esul s in a posi i e wage effec (Bossle and Schank 2017). We would expec ha an
inc ease in wages would lead o a highe job alue o wo ke s so ha hey would
be mo e likely o choose o keep hei jobs. While we do obse e his o men, his is
no he case o women. A possible eason is ha i women alued no only wages in
a job, an inc ease in wages would no necessa ily ansla e in o a highe alue o a
job. Indeed, he e is e idence ha women alue wages less han men do and p e e
diffe en a ibu es, see Fo in (2005), Blau and Kahn (2017). The e o e, women
migh ade in lowe wages o some o he a ibu e, such as flexibili y. Indeed,
Goldin (2014) finds ha p e e ence o flexibili y has a nega i e effec on wages.
Thus, i women in low-wage jobs alued flexibili y, an inc ease in wages c ea ed by
he in oduc ion o a minimum wage migh no necessa ily inc ease he alue o
hei job and consequen ly migh no lead o highe a achmen . Rela edly, his
could imply ha o men he subs i u ion effec is dominan , while o women i is
no , inso a as he inc eased wage makes wo king ela i ely mo e a ac i e o
men, bu no women.
In p inciple, he e could also be a demand side mechanism which is howe e
no suppo ed by he da a. Fi ms migh alue a job ma ch wi h men highe han
wi h women as hey expec women o ha e mo e ca ee in e up ions. Be o e he
e o m, fi ms could hen offe lowe wages o women o accoun o he expec a ion
o mo e in e up ions. Wi h a minimum wage, he e would be es ic ions on hese,
and fi ms hen would maximize he expec ed alue o he ma ch by hi ing men.
Fi mscouldno adjus o women’s p e e ence o flexibili y wi h lowe wages, hus,
hey would be mo e likely o b eak he ma ch.
Bo h he associa ion analysis and he uncondi ional su i al unc ions sugges
ha he women’s likelihood o job sepa a ion emains simila a e he e o m.
The e o e, we a gue ha he supply side explana ion migh be mo e ele an o
Table :F- es o in e ac ed e ms, model wi h ail y.
Null hypo hesis Gende Re o m Re o m and gende
F es . . .
P obabili y . . .
Sou ce: SOEP, o .F- es s a is ics o (join ) significance oge he wi h he co esponding p- alues o he
hypo heses o he coefficien s o gende , e o m, and e o m and gende , espec i ely, in he model wi h ail y.
Minimum Wage and Job A achmen 97
ou esul s since we would expec fi ms o inc ease sepa a ions o women i i was
mainly demand side d i en.
6 Robus ness Checks
We ca y ou wo obus ness checks o confi m ou esul s: fi s , we allow o a mo e
flexible baseline haza d and, second, we conside ac ual hou s wo ked ins ead o he
dummy o ull- ime employmen . In Table 7 we epo ou esul s o he obus ness
checks. In he fi s column, we allow o a mo e flexible baseline haza d and, in he
second column, we conside ac ual hou s wo ked ins ead o a ull- ime dummy. We
only epo he a iables o in e es we emphasized in he main analysis (see he
de ailed es ima ion in Tables A.10 and A.11 in he Appendix A). We ename he
a iable “ ull- ime” o “wo k ime” o e e o bo h o he wo diffe en a iables: a
dummy ha akes he alue o one i he wo ke wo ks mo e han 30 h pe week in (1)
and ac ual hou s pe week in (2). We s a by allowing o a mo e flexible baseline
haza d. In ou final es ima ion, column (3) in Table 5 he fi s yea is di ided in o
h ee s eps, wo h ee-mon hs s eps a he beginning, ollowed by a six-mon hs s ep
while he second yea is conside ed as one s ep. This baseline haza d choice could be
oo b oad, so we allow o a baseline haza d wi h h ee-mon hs s eps o e he whole
wo yea pe iod and es ima e ou final model wi h his al e na i e baseline haza d.
In he fi s column o Table 7, we can obse e ha allowing o a mo e flexible
baseline haza d does no affec ou esul s, we s ill find ha he e o m had diffe en
effec s on women and men (F- es o join significance o he iple in e ac ions is
6.47 wi h p- alue = 0.039) and he size o he effec emains s able, when compa ed o
Table 5.
In helas column,wechange hedefini ion o ull- ime. We con ol o ac ual
hou s wo ked pe week ins ead o he dummy o ull- ime employmen . Thus
we p o ide a mo e flexible a iable ha can con ol o ime a wo k. We
find again, ha ou esul s o he e o m ha ing diffe en effec s by gende hold
(F- es o significance o he iple in e ac ion is 10.36 wi h p- alue = 0.0056). As
o he coefficien o he iple in e ac ions, wo k ime- emale- e o m,bo h
coefficien s (1.141 o he specifica ion wi h he mo e gene al haza d and 0.049 o
he specifica ion wi h he ac ual wo k hou s) a e posi i e and significan . No e
ha o he ac ual hou s, he scale is diffe en , and hus he coefficien s diffe as
well. The disco e ed gende diffe ence is hen e en mo e gene al, inso a as no
only ull- ime wo king bu e en each addi ional wo ked hou inc eases his
diffe ence.
98 E. Ga cía-Mo án e al.
7 Conclusions
Women’s economic labo ma ke ou comes s ill lag behind he ones o men. Se e al
gende diffe ences a e well documen ed. Among o he s, women a e o e -
ep esen ed in he low wage sec o .
Using da a o Ge many, we examine whe he he effec s o he in oduc ion o a
minimum wage on low-pay employmen du a ion in 2015 diffe by gende . To do so,
Table :Robus ness checks: baseline haza d and hou s wo ked. Model wi h ail y.
()()
Re o m . .
(.)(.)
Women . .
(.)(.)
Female- e o m −. −.
(.)(.)a
Minijob −. −.
(.)(.)
Minijob- emale . .
(.)(.)
Minijob- e o m . .
(.)(.)
Wo k ime . .
(.)(.)
Wo k ime- emale −. −.
(.)b(.)b
Wo k ime- e o m −. −.
(.)b(.)
Wo k ime- e o m- emale . .
(.)b(.)c
Full ime expe ience (exp ) −. −.
(.)(.)
Exp - e o m . −.
(.)(.)
Exp - emale −. .
(.)(.)
LR es ( ail y) . .
(.)c(.)c
Obs. , ,
Sou ce: SOEP, o . The dependen a iable is “exi ed employmen a mon h ”. The coefficien s
a e ma ked wi h ai hey a e significan a he % le el, bi hey a e significan a he %andci hey a e
significan a he % le el. LR es ( ail y) is a es o ail y (null hypo hesis o no ail y) oge he wi h he
p- alue.
Minimum Wage and Job A achmen 99
we es ima e a join du a ion model o men and women be o e and a e he e o m.
We show ha a minimum wage aiming a imp o ing he si ua ion o low wage
wo ke s affec s women’s and men’s employmen du a ion diffe en ly. In pa icula ,
we find ha he e o m had an effec on employmen du a ion and ha his effec is
highe o men han o women. These gende diffe ences in employmen a ach-
men a e he s onges o ull- ime employmen . This p esen s an impo an
ou come since an inc ease in pe sis ence in ull- ime employmen seems mo e
desi able han one in pa - ime o in mini-jobs.
We a gue ha men’s highe job a achmen could be in line wi h a supply side
mechanism acco ding o which women, unlike men, migh place a highe alue
on o he job cha ac e is ics such as flexibili y a he han wages. The e o e, an
inc ease in wages due o he in oduc ion o a minimum wage would no inc ease
he alue o a job o women bu would inc ease he one o men. This would lead o
an obse ed highe job a achmen o men han o women a e he e o m. This
dominance o he supply side effec (o e an al e na i e demand side effec ) would
be in line wi h Bossle and Ge ne (2016) and Bellmann e al. (2016) who sugges ha
labo demand is mainly adjus ed h ough a educ ion in hi ing (which would no be
eflec ed in job sepa a ions) a he han fi ing.Inanycase,amo e ho ough
analysis o disen angle supply om demand side effec s could be in e es ing. Fo
doing so, we would need o sepa a e employee-ini ia ed and employe -ini ia ed job
sepa a ions.
Un o una ely, he a iable inqui ing he ype o job sepa a ion in ou da a-
se has oo many missing alues, making a meaning ul analysis una ainable. This
dis inc ion would be u he ele an o concluding whe he he esul ing highe
job a achmen as consequence o he minimum wage e o m indeed is a desi -
able ea u e. In pa icula , o employe -ini ia ed job sepa a ions, a highe
a achmen should be desi able. Employee-ini ia ed job sepa a ions, on he o he
hand, a e no necessa ily undesi able since a ce ain amoun o job mobili y
migh help o climb he job ladde o e en ually escape he low wage segmen o
he labo ma ke .
Ou esul s highligh he impo ance o unde s anding ha e o ms ha e
he e ogeneous effec s, women’sandmen’s employmen du a ions a e affec ed
diffe en ly by an in oduc ion o a minimum wage. In pa icula , while men’sjob
du a ion inc eases, he one o women emains unchanged. The minimum wage
could hen ha e misleading effec s on he gende wage gap inso a as i (pa ially)
dec eases due o a selec ion effec in which men inc easingly emain in he low-
wage sec o . I could be in e es ing o quan i y he effec on he gende wage gap in
iew o Caliendo and Wi b od (2020) finding o a dec eased gende wage gap. In
any case, Bossle and Schank (2017) a o emen ioned finding o a posi i e wage
100 E. Ga cía-Mo án e al.
effec on he exis ing jobs sugges s ha he desc ibed misleading effec is a mos
seconda y.
Since he p eceding analysis co e s su i als o only up o 24 mon hs, he
disco e ed employmen effec s could be empo a y. I could be in e es ing o u he
analyze o which deg ee he ound effec s a e pe manen changes in he employmen
a achmen . Finally, a model o job sea ch explici ly allowing o gende diffe ences
could help examine he empi ical esul s o he pape and conside coun e - ac ual
policy exe cises and is le o u he esea ch.
Resea ch unding: E a Ga cía-Mo án would like o hank he DIW o suppo du ing
a esea ch s ay unde he esea ch p ojec ’E alua ing he Minimum Wage
In oduc ion in Ge many (EVA-MIN)- Inno a i e Knowledge T ans e and
E idence-Based E alua ion’.
Conflic o in e es : The au ho s decla e ha hey ha e no conflic o in e es .
Appendix A
Table A.:Du a ion model o exi ing employmen , de ailed.
()()()()
Re o m −. −. . .
(.)c(.)c(.)(.)
Women −. −. . .
(.)c(.)a(.)(.)
Female- e o m . . −. −.
(.)a(.)(.)(.)
d . . . .
(.)c(.)c(.)(.)
d . . . .
(.)b(.)(.)(.)
d −. −. −. −.
(.)a(.)a(.)(.)b
Re o m-d −. −.
(.)(.)
Re o m-d −. −.
(.)(.)
Re o m-d −. −.
(.)(.)
Minimum Wage and Job A achmen 101
Table A.:(con inued)
()()()()
Female-d . .
(.)(.)
Female-d −. −.
(.)(.)
Female-d . .
(.)(.)
- -d .
(.)
- -d −.
(.)
- -d .
(.)
Minijob −. −. −.
(.)a(.)(.)
Minijob- emale −. .
(.)(.)
Minijob- e o m . .
(.)(.)
Minijob- e o m- emale .
(.)
Full ime (> h) = −. . .
(.)(.)(.)
Full ime- emale −. −.
(.)b(.)b
Full ime- e o m −. −.
(.)b(.)b
Full ime- e o m- emale . .
(.)b(.)b
Full ime expe ience (exp ) −. −. −.
(.)(.)(.)
Exp - e o m −. .
(.)(.)
Exp - emale −. −.
(.)(.)
Exp - e o m- emale .
(.)
Residence in Eas Ge many = −. −. −.
(.)(.)(.)
102 E. Ga cía-Mo án e al.
Table A.:(con inued)
()()()()
Age −. −. −.
(.)(.)(.)
P ima y-educa ion . . .
(.)(.)(.)
Seconda y-educa ion . . .
(.)(.)(.)
Te ia y-educa ion
Ma ied −. −. −.
(.)(.)(.)
Di o ced −. −. −.
(.)(.)(.)
Single −. −. −.
(.)(.)(.)
Widow
Kids-u-hh . . .
(.)(.)(.)
Temp-con ac . . .
(.)c(.)c(.)c
Ag icul u e . . .
(.)(.)(.)
Manu ac u ing . . .
(.)b(.)b(.)b
Se ices
Sec o -MW . . .
(.)(.)(.)
Fsize- . . .
(.)c(.)c(.)c
Fsize-- . . .
(.)(.)(.)
Fsize-
Public-sec o −. −. −.
(.)(.)(.)
Obs. , , , ,
Sou ce: SOEP, o , The dependen a iable is “exi ed employmen a mon h ”. The coefficien s a e ma ked wi h
ai he le el o significance is be ween %and %, bi he le el o significance is be ween % and %andci he le el o
significance is less han %.
Minimum Wage and Job A achmen 103
Table A.:Du a ion model o exi ing employmen wi h ail y, de ailed.
()()()
Re o m −. −. .
(.)c(.)c(.)
Women −. −. .
(.)b(.)a(.)
Female- e o m . . −.
(.)(.)(.)
d . . .
(.)c(.)c(.)
d . . .
(.)c(.)c(.)b
d . . −.
(.)(.)(.)
Re o m-d −.
(.)
Re o m-d −.
(.)
Re o m-d −.
(.)
Female-d .
(.)
Female-d −.
(.)
Female-d .
(.)
Minijob −. −.
(.)a(.)
Mini- emale .
(.)
Mini- e .
(.)
Full ime (> h) = −. .
(.)(.)
Full ime- emale −.
(.)b
Full ime- e o m −.
(.)b
Full ime- e o m- emale .
(.)b
Full ime expe ience −. −.
(.)(.)
104 E. Ga cía-Mo án e al.
Table A.:(con inued)
()()()
Exp - e o m .
(.)
Exp - emale −.
(.)
Residence in Eas Ge many = −. −.
(.)(.)
Age −. −.
(.)(.)
P ima y-educa ion . .
(.)(.)
Seconda y-educa ion . .
(.)(.)
Te ia y-educa ion
Ma ied −. −.
(.)(.)
Di o ced −. −.
(.)(.)
Single −. −.
(.)(.)
Widow
Kids-u-hh . .
(.)(.)
Temp-con ac . .
(.)c(.)c
Ag icul u e . .
(.)(.)
Manu ac u ing . .
(.)b(.)a
Se ices
Sec o -MW . .
(.)(.)
Fsize- . .
(.)b(.)b
Fsize-- . .
(.)(.)
Fsize-
Public-sec o −. −.
(.)(.)
Minimum Wage and Job A achmen 105
Table A.:Robus ness check: -mon h s ep baseline haza d, de ailed.
Re o m .
(.)
Women .
(.)
Female- e o m −.
(.)
d .
(.)
d .
(.)b
d .
(.)
d −.
(.)
d .
(.)
d .
(.)
d −.
(.)
Re o m-d −.
(.)
Re o m-d −.
(.)
Re o m-d −.
(.)
Re o m-d .
(.)
Re o m-d .
(.)
Table A.:(con inued)
()()()
LR es ( ail y) . . .
(.)c(.)c(.)c
Obs. , , ,
Sou ce: SOEP, o . The dependen a iable is “exi ed employmen a mon h ”. The coefficien s a e ma ked wi h
ai he le el o significance is be ween %and %, bi he le el o significance is be ween %and% and ci he le el o
significance is less han %. LR es ( ail y) is a es o ail y (null hypo hesis o no ail y) oge he wi h he p- alue.
106 E. Ga cía-Mo án e al.
