Baue , Anja
A icle
Accoun ing o quali ica ion in misma ch unemploymen
Jou nal o Labou Ma ke Resea ch
P o ided in Coope a ion wi h:
Ins i u e o Employmen Resea ch (IAB)
Sugges ed Ci a ion: Baue , Anja (2024) : Accoun ing o quali ica ion in misma ch unemploymen ,
Jou nal o Labou Ma ke Resea ch, ISSN 2510-5027, Sp inge , Heidelbe g, Vol. 58, Iss. 1, pp. 1-23,
h ps://doi.o g/10.1186/s12651-024-00386-7
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/308509
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/
Baue J Labou Ma ke Res (2024) 58:27
h ps://doi.o g/10.1186/s12651-024-00386-7
ORIGINAL ARTICLE Open Access
© The Au ho (s) 2024. Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License, which
pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as you gi e app op ia e c edi o he
o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons licence, and indica e i changes we e made. The images o
o he hi d pa y ma e ial in his a icle a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line
o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended use is no pe mi ed by s a u o y
egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om he copy igh holde . To iew a copy o his
licence, isi h p:// c ea i eco mmons. o g/ licen ses/ by/4. 0/.
Jou nal o Labou Ma ke Resea ch
Accoun ing o quali ica ion inmisma ch
unemploymen
Anja Baue 1*
Abs ac
The pape shows how impo an i is o conside he equi emen le el when measu ing misma ch unemploy-
men . While he misma ch be ween occupa ions dec eases o e ime, he imbalances in he dis ibu ion o unem-
ployed wo ke s and acan jobs ac oss equi emen le els inc eased, which, in sum, led o a s agna ion o misma ch
on he combined occupa ion- equi emen le el. Fu he mo e he pape shows ha misma ch unemploymen
eme ges especially a he le el o un- and semi-skilled ac i i ies, as he e is excess supply ega dless o he occupa-
ions. And, mo e impo an ly, he excess supply is ising ecen ly.
Keywo ds Alloca ion, Job inding a e, Misma ch, Occupa ion, Quali ica ion
JEL Classi ica ion J6, E24
1 In oduc ion
Labo ma ke igh ness has isen sha ply in Ge many in
he las decade, and many occupa ions a e a ec ed by
sho ages o skilled wo ke s due o demog aphic sh ink-
age (Bossle e al. 2022). In addi ion, employmen g ow h
has slowed since he COVID-c isis and he job- inding
a e has no isen o p e-c isis le els. Mo eo e , long-
e m unemploymen is on he ise. Agains his back-
g ound, he ques ion a ises o wha ex en he cu en
obse a ions can be a ibu ed o misma ch unemploy-
men , i.e. a misalloca ion o unemployed pe sons and job
acancies. P incipally, misma ch occu s whene e he
“quali ica ions o skills o wo ke s, indi idually o in he
agg ega e, a e di e en om he quali ica ions o skills
equi ed o he jobs” (Sa inge 2012,p. 3). The li e a u e
dis inguishes be ween wo ypes o misma ch: (1) mis-
ma ch wi hin a ma ch, whe e wo ke s do no i well o
he job hey a e ope a ing and (2) misma ch be o e he
ma ch, o en labeled as “misma ch unemploymen ”, whe e
unemployed wo ke s and open job acancies would no
o m a ma ch e en i sea ch ic ions o impe ec in o -
ma ion did no exis . This pape looks a he la e ype,
which s ems om s uc u al shi s in labo demand and/
o supply. Resea ch on misma ch unemploymen da es
back o he 1980s (see Jackman and Rope 1987; Jackman
e al. 1991; Schioppa 1991) bu new ad ances we e made
in he seminal pape by Şahin e al. (2014). In he a e -
ma h o he Global Financial C ises (GFC), he au ho s
*Co espondence:
Anja Baue
[email p o ec ed]
1 Ins i u e o Employmen Resea ch, Regensbu ge S . 104,
90478 Nu embe g, Ge many
27 Page 2 o 23
A.Baue
showed ha misma ch unemploymen explains up o one
hi do he ise in unemploymen a e he G ea Reces-
sion (Şahin e al. 2014). A e he COVID-c isis (CC)
he opic gained in e es again, bu misma ch appea ed
o play a mino ole in he a e ma h o he COVID-19
shock in he US (Fo sy he e al. 2022). Recen s udies
con i m his esul o di e en coun ies, like he UK
(Pizzinelli and Shiba a 2023; Tu ell e al. 2021; Pa e son
e al. 2016) o Japan (Shiba a 2020).1
The con ibu ion o my pape o his li e a u e is wo-
old. Fi s , no much is known abou misma ch unem-
ploymen hus a o Ge many (see o example, Baue
2013; Hu e and Webe 2017). Howe e , Ge many is an
in e es ing case o s udy because as a Eu opean coun y
wi h ypically qui e signi ican s uc u al unemploymen ,
misma ch unemploymen migh beha e di e en ly han
in coun ies such as in he US o in he UK. So I con ib-
u e o he ques ion o how he Ge man expe ience is di -
e en . Second, I highligh he impo ance o conside ing
he equi emen le el in analysing misma ch unemploy-
men o Ge many. Şahin e al. (2014)’s app oach, and
he esul ing misma ch indica o quickly eme ged as a
benchma k, as i is easily implemen able, o en used in
he li e a u e and, unde ce ain condi ions, also com-
pa able ac oss coun ies. Howe e , i is no clea which
dimensions (e.g., occupa ions, skills, sec o s, egions, o
e en in e ac ions among hese dimensions) adequa ely
measu e misma ch unemploymen . Fo example, Piz-
zinelli and Shiba a (2023) look a sec o s and occupa-
ions, bu conclude ha he e migh exis “mo e sub le
dimensions o e which misma ch may play a ole”. How-
e e , i is o ele ance o know in which dimension mis-
ma ch unemploymen exis s, as i has an in luence on
which policy measu e could be e ec i e in educing mis-
ma ch unemploymen . In his espec , he pape con ib-
u es o he discussion on misma ch unemploymen by
looking a ano he dimension, namely he equi emen
le el which se es as a p oxy o quali ica ion. I mis-
ma ch unemploymen occu s a he occupa ional le el,
he unemployed would ha e o be mo i a ed o change
he occupa ion in o de o educe misma ch unemploy-
men . Howe e , i misma ch unemploymen occu s a he
quali ica ion le el, i would be mo e expedien o p o-
ide he unemployed wo ke s wi h u he quali ica ions
wi hin an occupa ion (o o encou age i ms o make
concessions ega ding he en y skill se ). The Ge man
Occupa ional Classi ica ion 2010 (Kldb2010) is designed
such ha i combines occupa ional expe ise wi h he
equi emen le el wi hin he job (simila as in ISCO). In
ha espec i allows o analyse misma ch unemploymen
ac oss occupa ions and he equi emen le el.
I ind ha be ween 5 and 18% o hi es a e los due o
a misalloca ion looking a he ime ho izon om 2007
o 2022. Ac oss his pe iod, misma ch unemploymen
came down om o e 40% o 12 o 22% on he occupa-
ion-le el and o 22 o 32% when aking he equi emen
le el in o accoun . Mos o his e olu ion happened un il
2012, a e wa ds he e is only li le mo emen o e ime
excep in he CC. Du ing he CC, misma ch unemploy-
men spiked empo a ily. Compa ing he occupa ion o
he occupa ion- equi emen le el, he esul s show ha
misma ch unemploymen is ising s onge on he occu-
pa ion- equi emen le el since 2021. Mo e impo an ,
I can show ha a subs an ial pa o misma ch unem-
ploymen esul s om a misalloca ion on he quali ica-
ion le el a he han on he occupa ional le el. Fi s ,
ac oss he equi emen le el, he sha e o hi es los due
o imbalances in demand and supply almos doubled
ac oss he pe iod. Second, in he ma ke o unskilled
and semi-skilled wo ke s, an o e -supply exis s in almos
all occupa ions. Re e sely, on highe equi emen le els,
mo e occupa ions exhibi sho ages. Looking a he cycli-
cal pa e ns du ing he GFC and CC e eals, ha he CC
hi occupa ions di e en ly ac oss he equi emen le -
els han he GFC. Du ing he GFC, a dec ease in unde -
supply a he (complex) specialis ac i i ies le el is isible,
while he CC a he inc eased he o e -supply in occupa-
ions a he unskilled and semi-skilled equi emen le el.
2 The Ge man labo ma ke
Fi s , I b ie ly discuss he Be e idge Cu e. While mo e-
men s along he cu e a e associa ed wi h ups and downs
in he business cycle, mo emen s o he cu e a e a he
associa ed wi h s uc u al changes ha a ec he o e all
unc ioning o he labo ma ke . In Ge many, he Be e-
idge cu e shi ed inwa d in he mid 2000s a e a se o
labo ma ke e o ms ( he so-called Ha z e o ms) we e
in oduced (see Fig.1). The e is a as li e a u e on he
unde lying sou ces o his shi anging om in ensi ied
job sea ch due o lowe unemploymen bene i s, a be -
e placemen h ough he es uc u ing o he Employ-
men Agency, and wage mode a ion (Launo and Wälde
2013; K ebs and Sche el 2013; Launo and Wälde 2016;
B adley and Kügle 2019; Hochmu h e al. 2021). Du ing
he GFC and he CC he Ge man labo ma ke ho e ed
a ound a s able Be e idge cu e. While he pe iod o he
GFC was loca ed in he lowe igh end o he cu e, an
upwa d mo emen is isible in he 2010s. The pe iod
o he CC is loca ed a he uppe le end o he cu e.
1 Wha hese pape s ha e in common, is ha hey ex end he app oach o
Şahin e al. (2014) by explo ing di e en da a sou ces, di e en g oupings
o labo ma ke s, o da a o di e en coun ies. O he pape s, like He z and
Van Rens (2020) o Ba nichon and Figu a (2015) ins ead ely on di e en
app oaches o measu e misma ch unemploymen . None heless, also hese
o he app oaches ely on some kind o segmen a ion o he labou ma ke
and hus he same issue a ises as he choice o he le el o disagg ega ion is
c ucial.
Page 3 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
Rega ding ha pic u e, one would expec ha in Ge -
many misma ch unemploymen also did no play a majo
ole du ing he CC.
Howe e , looking a job- inding and sepa a ion a es
o e ime, he e seems o be a pe sis en shock o he job-
inding a e (see le panel o Fig.2). The job- inding a e
was inc easing in he mid 2010’s. The inc ease s opped
a ound 2018 and hen he COVID-19 shock hi , which led
o a d op o he job- inding a e, om which i eco e ed
o some ex en , bu did no each he p e-pandemic le el
un il he end o 2022. While he sepa a ion a e spiked
du ing he i s phase o he CC, i quickly eco e ed and
ollowed again i s downwa d end (Baue and Webe
2021b). I a all, he sepa a ion a e d opped e en mo e in
he a e ma h o he CC. A he same ime, he numbe o
acan jobs was ising s ongly since he second qua e
2020 (see IAB Job Vacan cy Su ey) and labou sho ages
a e also ising (see Labou Sho age Index). A ques ions
Fig. 1 Be e idge Cu e 2007–2022. Vacancies and Unemploymen a e no malised as a es by employmen . Sou ce: S a is ical O ice o he Fede al
Employmen Agency, own calcula ions. ©IAB
Fig. 2 Flow a es, 2007–2022. The job- inding a e is calcula ed as he mo emen om unemploymen o employmen o e he s ock
o las pe iods unemploymen . The sepa a ion a e is calcula ed as mo emen om employmen o unemploymen o e las pe iods s ock
o employmen . Sou ce: S a is ical O ice o he Fede al Employmen Agency, own calcula ions. ©IAB
27 Page 4 o 23
A.Baue
ha na u ally a ises, is why he unemployed pe sons do
no ma ch wi h hese acan posi ions. I migh be due
o ic ions, o s uc u al imbalances such as a bad i
be ween he unemployed and he acan posi ions, which
would be conside ed as misma ch unemploymen .
3 Me hod
I am applying a me hod p oposed by Şahin e al. (2014)
which de i es an indica o ha is widely used as bench-
ma k indica o . The main idea is ha he labo ma ke
comp ises se e al subma ke s. While ic ional unem-
ploymen is c ea ed by ic ions wi hin each subma -
ke ha p e en unemployed wo ke s o ma ch wi h
un illed acancies, misma ch unemploymen a ises
because o an subop imal alloca ion o acancies and
unemployed wo ke s ac oss subma ke s. In my analy-
sis I de ine subma ke s o be di e en occupa ions and
occupa ion- equi emen le el combina ions as ou lined
abo e. This subop imal alloca ion (compa ed o a plan-
ne ’s solu ion) can be cap u ed by an index.
3.1 Theo y
The index M measu es hi es ha a e los due o a mis-
ma ch by compa ing he ac ual (obse able) numbe hi es
h o an ideal numbe o hi es
h∗
. The numbe o hi es
(ac ual and ideal) depends on he dis ibu ion o unem-
ployed wo ke s and job acancies o e a de ined ange
o subma ke s (e.g., occupa ions). In each dis inc ma -
ke , hi es a e go e ned by a Cobb-Douglas ype ma ching
unc ion wi h cons an e u ns o scale (
hi
=φ
i
α
i
u
1−α)
i
).
Hence sea ch ic ions exis in e e y ma ke and a e cap-
u ed by ma ching e iciency.2 The ideal numbe o hi es
comes ou o he model’s social planne solu ion, in which
he planne can mo e unemployed pe sons cos lessly
ac oss ma ke s. This leads o an equaliza ion o ma ke -
speci ic labo ma ke igh ness ( acancy o unemploymen
a io) ac oss ma ke s (weigh ed by ma ching e iciency).
This ensu es “ ha he planne alloca es mo e job seeke s
o hose labo ma ke s wi h mo e acancies and highe
ma ching e iciency” (Şahin e al. 2014,p. 3534).
whe e i e e s o a subma ke , deno es ime- a ian a -
iables a mon hly in e als, is he acancy s ock, u is he
unemploymen s ock.
(1)
M =1−
h
h∗
=1−
I
i=1
φi
¯
φ i
αui
u
1
−
α
The i s e m o he exp ession ela es o ma ching e i-
ciency, i.e., ¯
φ
=
I
i=1φ
1
α
i
i
α
, whe e
φ
deno es
ma ching e iciency.
3.2 Es ima ions
The ma ching elas ici y
α
is ex ac ed by a simple OLS
eg ession o a educed o m ma ching unc ion. The
eg ession equa ions ead as ollows:
The job- inding- a e
j
is de ined as changes om
unemploymen o employmen o e las pe iod’s s ock o
unemploymen (
j
=
EU
/
U
−
1
) and
θ
deno es labou
ma ke igh ness as he a io o acancies o unemploy-
men
θ =V /U
. Fo he ma ching elas ici y I ecei e
a coe icien o
α≈0.5
. I e i ied his esul in se e al
di e en speci ica ions in which he coe icien anges
be ween 0.41 and 0.54, depending on he ime pe iod
co e ed, and whe he and which ime ends a e used.
The subma ke -speci ic ma ching e iciency
φi
is
deduc ed by panel es ima ions. I eg ess he subma ke -
speci ic labo ma ke igh ness
θi
on he subma ke ’s
job- inding a e
j i
, whe e, again,
θi
is de ined as he
a io o un illed acancies o unemployed wo ke s and
he job- inding a e is measu ed as ansi ions be ween
unemploymen and employmen o e las pe iod’s s ock
o unemploymen .
βi
is cap u ing he ixed e ec o e e y
subma ke and
ei
is he e o e m.
whe e
I use he subma ke -speci ic cons an ou o his ixed
e ec s eg ession and expona e i o ecei e he subma -
ke -speci ic ma ching e iciency. This gi es me a ma ch-
ing e iciency ha anges be ween 0.7 and 2.2 ac oss
occupa ions a he 3-digi le el.
4 Da a
The da a ha he me hod is applied o is adminis a i e
da a o he Fede al Employmen Agency, which co e s
he uni e se o unemployed wo ke s and acancies ha
a e egis e ed a he Fede al Employmen Agency. I ely
on e y de ailed occupa ion in o ma ion, which is col-
lec ed a he 5-digi le el in he Ge man Classi ica ion
o Occupa ions (Kldb2010). I can dis inguish unem-
ployed wo ke s by he occupa ion o o igin (occupa ion
o las employmen o app en iceship) and des ina ion o
sea ch. The acancy da a ela es o jobs, which a e eg-
is e ed a he Fede al Employmen Agency, which is no
(2)
ln
(
j
)=β0+α
ln
(θ
)+
e
(3)
ln
(
j i
)=β
i
+β1
ln
(θ
i
)+
ei
(4)
φi=exp(βi)
2 Though ma ching e iciency could be a guably ime- a ying, I assume i
o be ime-cons an . Robus ness checks show ha he di e ences be ween a
ime-cons an and ime- a ying a ian a e small.
Page 5 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
manda o y. In a obus ness check, I combine his admin-
is a i e da a wi h su ey da a, which p ojec s acancies
a a na ional le el (i.e., IAB Vacan cy Su ey, see Appendix
Fig.12). The job- indings ela e o mo emen s om he
s a us “unemployed and sea ching o wo k” o “employ-
men subjec o social secu i y (wi hou any subsidised
employmen )” and he sepa a ions a e calcula ed ice
e sa. Fu he mo e I exclude he mili a y occupa ions as
he e is no acancy in o ma ion a ailable o hem.
The panel includes s ocks o unemployed by occupa-
ions and s ocks o acancies by occupa ions, and occu-
pa ion-speci ic job- indings and sepa a ions, a a mon hly
in e al unning om Janua y 2007 o Decembe 2022.
Based on his in o ma ion, I agg ega e unemploymen ,
acancies, job- indings and sepa a ions o he na ional
le el by summing o e he occupa ions.3
A ema kable ea u e o he Ge man Occupa ion Clas-
si ica ion is i s ho izon al and e ical dimension. Tha
means, i allows o dis inguish occupa ions by he occu-
pa ional expe ise (assessed by equi ed skills, abili ies and
knowledge) and equi emen le els (complexi y wi hin
an occupa ion; Paulus and Ma hes 2013). A a 5-digi
le el he classi ica ion comp ises 1300 di e en occupa-
ions. The 5 h digi , which holds he equi emen le el,
has ou ca ego ies: (1) Unskilled o semi-skilled ac i i ies,
(2) Specialis ac i i ies, (3) Complex specialis ac i i ies,
(4) Highly complex ac i i ies. Those ca ego ies e lec he
o mal oca ional quali ica ions ypically demanded o a
ce ain occupa ional ac i i y. I is independen o a pe -
son’s o mal quali ica ion as no jus o mal quali ica ions
equi ed o pe o m he occupa ion a e used o classi ica-
ion bu also in o mal educa ion and/o p o essional expe-
ience a e impo an .4 A he 4-digi le el, he classi ica ion
has a ound 700 occupa ions, which implies ha no all
equi emen le els a e p esen in all 4-digi occupa ions.5
Fo he main analysis, I use an agg ega ion which maps
he 5-digi classi ica ion in o 14 occupa ion segmen s
(Ma hes e al. 2015). These occupa ional segmen s di ide
he labou ma ke such ha he swi ching be ween di -
e en occupa ions is minimised. This is impo an as he
app oach o Şahin e al. (2014) is mos eliable in such a
se -up, as i minimizes po en ial bias. Basically, i ensu es
ha unemployed wo ke s in an occupa ion sea ch o
acancies in ha pa icula occupa ion, o simply speak-
ing he chosen classi ica ion ep esen s he ele an ma -
ke .6 Fo he explo a ion o he quali ica ion dimension, I
ely on he 5 h digi o he scale and also in e ac i wi h
he occupa ion segmen s.
5 Resul s
Figu e 3 shows he indices o e ime conduc ed as in
Eq.1. They ha e e y simila pa e ns o e ime, bu di -
e en le els anging om 0.05 (occupa ion segmen s, le
panel) o almos 0.2 (3-digi occupa ion
×
equi emen
le el, igh panel), which means ha , less han 20% o
hi es a e los due o misma ch. Be ween 2008 and 2010
misma ch dec eased, hen i inc eased again un il 2012.
A e wa ds he e is a dis inc mo emen be ween he le
and igh panel o Fig.3. While he e is slow downwa d
mo emen in he occupa ion misma ch measu e (le
panel), he e is a he a s agna ion o he occupa ion-
equi emen le el ( igh panel). In he second hal o 2016
he indices s a o inc ease sligh ly, which migh be an
e ec o he e ugee in low in 2015 and 2016. As B ücke
e al. (2020) and Bundesagen u ü A bei (2020) poin
ou , unemploymen o asylum seeke s inc eased s eadily
om 2016 o 2020, al hough on a e age, 35% o he e u-
gees who en e ed Ge many be ween 2013 and 2016 ound
a job a e he second hal o 2018. This migh explain he
dec ease in he misma ch index un il he COVID-19-c i-
sis un olded. The in e up ions in he beginning o 2020
due o he CC a e isible in bo h se ies.
To explo e he di e ences be ween he le and he
igh panel o Fig.3, I plo he misma ch index exclusi ely
o he equi emen le el (see Fig.4). Though he equi e-
men le el canno be e alua ed s and-alone because he
in e p e a ion di e s sligh ly wi hin occupa ions,7 i is
3 This p ocedu e has he ad an age ha i excludes mo emen s beyond
he occupa ion dimension as, e.g., no all unemployed wo ke s ha e a alid
occupa ion in o ma ion. I use he s ocks o unemploymen o he occu-
pa ion o des ina ion, hence hese missing da a is negligible. The se ies o
unemploymen epo ed by he Fede al Employmen Agency and he se ies
I gene a e by agg ega ing he occupa ion panel da a is ai ly simila , he gap
is small. Fo he o he se ies, he same holds.
4 See h ps:// s a i s ik. a bei sage n u . de/ DE/ S a i sche - Con e n / G und lagen/
Me ho dik- Quali ae / Me ho dische- Hinwe ise/ uebe g ei end- Me hH inwei se/
An o de un gsni eau- Be u e. h ml and o a deepe unde s anding: h ps:// www.
a bei sage n u . de/ da ei/ Klass i ika ion- de - Be u e_ ba017 989. pd , p.27 e seqq..
5 The da a limi a ions a e he ollowing: The Ge man Occupa ion Classi-
ica ion was enewed in 2010, and a con e sion o he old classi ica ion is
possible bu leads o coding e o s. Hence, he mo emen s o e ime be o e
2011 a e augh wi h highe insecu i ies. On op, quali y issues in unskilled
and low skilled occupa ions be ween Sep embe 2009 and June 2010 a e
p esen and no esol able.
6 The index used o measu e misma ch p oposed by Şahin e al. (2014) has
he p ope y ha i is ising in he numbe o subma ke s: The mo e ma -
ke s a e obse ed, he highe he misalloca ion be ween he ma ke s, which
inc eases he sha e o misma ch unemploymen . Hence he e is a ce ain
bias: I he ma ke s a e simila , wo ke s would no jus sea ch in hei s a ed
a ge occupa ion bu also in e y simila occupa ions.
7 The Fede al Employmen Agency gi es he ollowing in e p e a ion o
he equi emen le el: “The equi emen le el desc ibes he complexi y o
an occupa ional ac i i y. I is always ypical o a speci ic occupa ion and is
also independen o a pe son’s o mal quali ica ions. Al hough he o mal
quali ica ions equi ed o pe o m he occupa ion a e used o classi ica-
ion, in o mal educa ion and/o p o essional expe ience a e also impo an .
See h ps:// s a i s ik. a bei sage n u . de/ DE/ S a i sche - Con e n / G und lagen/
Me ho dik- Quali ae / Me ho dische- Hinwe ise/ uebe g ei end- Me hH inwei se/
An o de un gsni eau- Be u e. h ml.
27 Page 6 o 23
A.Baue
in o ma i e on whe he he quali ica ion imbalances
ose. And indeed, he index is ising ac oss ime, and
shows an upwa d end ha is especially p onounced
a e 2018. Concluding, while misma ch ac oss occupa-
ions dec eased o e ime, he imbalances wi h espec o
he quali ica ion mix in he economy ose, such ha mis-
ma ch ac oss occupa ion segmen s in combina ion wi h
he equi emen le el inc eased. Fu he mo e, a coun e -
ac ual simula ion in which I hold he dis ibu ion o he
equi emen le el cons an ac oss ime, shows, ha his
would lead o a pa allel in he misma ch index o occu-
pa ion segmen s (see Fig.14 in he Appendix.)
5.1 Implica ions o unemploymen
A ea u e o Şahin e al. (2014)’s app oach is, ha i allows
o cons uc a coun e ac ual unemploymen a e (which
is a e e ence o how unemploymen would beha e
wi hou misma ch). Fo his pu pose, a coun e ac ual
job inding a e which measu es job indings ela i e o
unemploymen in he absence o misma ch, is conduc ed.
By he assump ion o a s anda d law o mo ion o he
unemploymen a e, a coun e ac ual unemploymen a e
can be backed ou . The coun e ac ual unemploymen
a e is as ollows:
Gi en an ini ial alue o
u∗
, a sequence o coun e ac ual
unemploymen a es wi h he s anda d law o mo ion (
s
deno es he sepa a ion a e) can be calcula ed. Neces-
sa y o calcula e he sequence o coun e ac ual unem-
ploymen a es is he coun e ac ual job inding a e
∗
,
which is de ined as ollows:
As s a ing alue, I choose
u∗
=u
. As he unemploy-
men a e in my da a is downwa d biased,8 i is no
help ul o depic he coun e ac ual and ac ual unem-
ploymen a e in le els. The e o e I p oceed wi h a cal-
cula ion whe e I use he di e ence be ween ac ual and
coun e ac ual unemploymen (i.e., misma ch unemploy-
men ) o e he ac ual unemploymen a e. This measu e
can be in e p e ed as he pe cen age sha e o misma ch
unemploymen on ac ual unemploymen .
(5)
u∗
+1=s +
1−s −
∗
u
∗
(6)
∗
=¯
φ �
u
∗
α
= ·
1
1−
M u
u
∗
α
Fig. 3 Misma ch indices, 2007–2022, occupa ions (le ) and occupa ions
×
equi emen le el ( igh ). All se ies a e seasonal adjus ed using
X-12-ARIMA. Da a be o e 2011 is con e ged om KldB88 o KldB2010 and migh be augh wi h coding e o s. Sou ce: S a is ical O ice
o he Fede al Employmen Agency, own calcula ions. ©IAB
8 No e ha I gene a e he se ies o unemploymen by summing ac oss occu-
pa ions o e e y poin in ime. Tha implies, ha he unemploymen a e is
somewha lowe han he o icial unemploymen a e gi en ha he e a e
missing occupa ion in o ma ion. Fu he mo e, he mili a y sec o is excluded.
Page 7 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
Figu e5 plo s his sha e o e ime. Misma ch unem-
ploymen anges, on a e age, be ween 13 and 22% o
o al unemploymen on he occupa ion le el in he end
o 2022, and be ween 23 and 33% on he occupa ion
×
equi emen le el. The sha e dec eased om high al-
ues in he beginning o abou 40 o 50% o hese lowe
le els. Du ing he CC, he misma ch sha e ose quickly
bu only empo a ily. Compa ed o o he coun ies, he
shown pa e ns o e all beha e simila ly. Fo sy he e al.
(2022) shows ha du ing he CC misma ch only occu ed
empo a ily in he US. Pizzinelli and Shiba a (2023) con-
i ms his pa e n o he US and he UK. In hei da a,
as in mine, he COVID-19-shock induces a spike, which
is no he case in Fo sy he e al. (2022). A e he CC,
he e is an upwa d mo emen . This upwa d mo emen
b ings he occupa ion se ies back o he p e-c isis le el,
bu o he in e ac ion o occupa ions and equi emen
le el, misma ch unemploymen ises abo e i s p e-c isis
le el. In absolu e numbe s, he sha e o misma ch unem-
ploymen in o al unemploymen ose abou 2 pe cen -
age poin s om 20.5% (mean be ween Janua y 2019 and
Janua y 2020) o 22.4% (in Augus 2022) o he occu-
pa ion segmen s x equi emen le el. Hence, he ise
in he imbalances in he quali ica ion mix spills o e o
unemploymen .
When I plo he sha e o misma ch unemploymen
o occupa ion segmen s and o occupa ion segmen s
×
equi emen le el o e lapping ac oss ime (see Fig.6),
his becomes e en clea e . While he e is a simila mo e-
men be o e he COVID-c ises in he se ies, a s ong
inc ease a e ha can be seen, which is pa icula ly
p onounced o he in e ac ion o occupa ions and he
equi emen le el.9 When COVID-19 hi he Ge man
economy, he go e nmen en o ced wo s ic lockdowns,
one in sp ing 2020 and one in win e 2020/2021 (Baue
and Webe 2021a). In he i s lockdown, he unemploy-
men a e inc eased om 5.1 o 6.3% (i.e., by mo e han
600,000 pe sons), he pool o egis e ed acan posi ions
d opped by 20% (see Key Figu es o he Labou Ma ke
- Ge ma ny (Mon h ly Repo )).10 Howe e , compa ed o
o he coun ies, he Ge man “sho - ime-wo k-scheme”
se ed as a s abilize and sa ed jobs (Geh ke and Webe
2020; Ch is l e al. 2022). This implies ha he e ec o
COVID-19 on unemploymen in Ge many is a he mod-
e a e compa ed o he US. None heless, his is sugges i e
e idence ha he igh labo ma ke in Ge many may be
explained pa ially by a misalloca ion on he quali ica ion
le el.
6 Occupa ions and he equi emen le el inde ail
To s eng hen my esul s, I exploi he quali ica-
ion dimension in mo e dep h. Fi s , Fig.7 depic s he
acancy and unemploymen sha es ac oss he equi e-
men le el o e ime. Appa en ly, he dis ibu ion o he
equi emen le el o he acancy sha es is di e en han
o he unemploymen sha es. Fu he mo e he de elop-
men o e ime is di e en : he ela i e demand o high
complex jobs is ai ly s able, while he ela i e demand
o complex jobs, bu also o jobs a he unskilled and
semi-skilled le el inc eased o e ime. Con a y, he ela-
i e demand o specialis s sank. Ha ing a look a he
unemploymen sha es, shows: he ela i e supply a he
unskilled and semi-skilled le el is also inc easing, bu
much s onge han he demand. The ela i e supply a
he specialis le el is dec easing much s onge han on
he demand side. O e all, he igu e indica es ha he
inc ease in misma ch unemploymen is closely ela ed o
sho ages o skilled labou .
Tigh ly connec ed o his esul is he ques ion whe he
he di e ences ac oss he equi emen le el eme ge om
Fig. 4 Misma ch index o equi emen le el. All se ies a e seasonal
adjus ed using X-12-ARIMA. Da a be o e 2011 is con e ged
om KldB88 o KldB2010 and migh be augh wi h coding e o s.
Sou ce: S a is ical O ice o he Fede al Employmen Agency, own
calcula ions. ©IAB
9 In June 2022 bo h se ies inc eased again when he e ugees om he
Uk aine en e ed he unemploymen pool in Ge many.
10 Acco ding o he IAB-Job Vacancy su ey, he pool o all acan posi-
ions d opped by e en 40% om 4 h qua e 2019 o second qua e 2020.
27 Page 8 o 23
A.Baue
ce ain occupa ions.11 Especially, because sho ages o
skilled labou a e no obse ed in all occupa ions. The e-
o e, I decompose he con ibu ion e e y occupa ion has
had in e ms o unemploymen by looking a de ia ions
be ween ac ual unemploymen and op imal unemploy-
men o each occupa ion ac oss equi emen le els. The
decomposi ion is calcula ed unde he condi ion ha he
social planne is no able o dis ibu e mo e unemployed
wo ke s han ac ual unemployed wo ke s exis . Pe con-
s uc ion, alues below ze o imply ha he social planne
would like o dis ibu e mo e unemployed o he sec o
han ac ually a e p esen , i.e., (
u−u∗<0
). Con e sely,
alues abo e ze o would indica e ha he occupa ion
exhibi s mo e unemployed han he social planne would
choose. This exe cise is simila o he decomposi ion
exe cise o Pizzinelli and Shiba a (2023) and akes he
demand side, hence he acancy e olu ion, as gi en. Pu
di e en ly, he exe cise shows o each equi emen le el,
in which occupa ion segmen he social planne would
lo e o dis ibu e unemployed wo ke s in compa ison o
he ac ual dis ibu ion o minimize misma ch. Pe con-
s uc ion, alues below ze o imply ha he social planne
would choose o dis ibu e mo e unemployed wo ke s
han he e ac ually a e p esen (labou sho ages) while
alues abo e ze o would indica e ha he e a e mo e
unemployed wo ke s han op imal (excess supply). Fig-
u e16 shows he esul s o occupa ions independen ly
o he equi emen le el. O e all, 6 ou o 14 occupa ions
show excess unemploymen , 5 ou o 14 occupa ions
show sho ages and 3 occupa ion segmen s swi ch o e
ime.
Fig. 5 Misma ch Unemploymen , 2008–2022. All se ies a e seasonal adjus ed using X-12-ARIMA. The se ies s a s in 2008 o accoun o he e ec
o he s a ing alue on he se ies in he beginning. Sou ce: S a is ical O ice o he Fede al Employmen Agency, own calcula ions. ©IAB
Fig. 6 Misma ch Unemploymen , 2017–2022. All se ies a e seasonal
adjus ed using X-12-ARIMA. Sou ce: S a is ical O ice o he Fede al
Employmen Agency, own calcula ions. ©IAB
11 Figu e15 in he appendix shows he occupa ion sha es wi hin Fig.7.
Page 15 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
is used in he main analyses. In a compa ison o he base-
line and he occupa ion o o igin, he la e has a lowe
le el bu a simila mo emen o e ime. When I al e na e
he job indings o be he one o ac ual ake-up, he e is
almos no di e ence in he indices.
Second, I explo ed he in luence o ma ching e iciency
on he index. I calcula ed wo al e na i e e sions o he
index, (1) an index wi h ime- a ying ma ching e iciency,
and (2) and index wi hou he e m ela ed o ma ching
e iciency in Eq. (1). While he i s e sion appea s o be
mo e ola ile han he baseline, he la e has a e y simi-
la mo emen o e ime, howe e a a di e en le el.
In he main analysis, I used egis e ed acancies o he
Fede al Employmen Agency. In Ge many o i ms i is
no manda o y o egis e he acancies, hence his se ies
is p one o unde epo ing and also bias wi h ega ds o
he equi emen le els. To ci cum en his issue, I use
he ela ion o egis e ed acancies o all acancies by
equi emen le el o he IAB Job Vacancy Su ey. The
IAB Job Vacancy Su ey is a ep esen a i e su ey, ha
allows o p ojec ions o he o e all s ock o he acan-
cies in he economy. Howe e , he e a e some sho -
comings. Fi s , he da a is a ailable only in a qua e ly
equency. Second, he da a is no ep esen a i e ac oss
he occupa ion dimension. Luckily, in o ma ion on he
equi emen le el o a acan posi ion is a ailable, which
is especially aluable o my indings. I p ojec he acan-
cies sepa a ely a e e y equi emen le el and agg ega e
hem up a e wa ds. This gi es me a new measu e o he
acancy sha es. O e all, he ela ion o egis e ed o all
acancies a ies be ween 33 and 49% du ing he obse a-
ion pe iod. Wi h espec o he equi emen le el, he
highe he expe ise equi ed he less likely i is ha he
acancy is egis e ed a he Fede al Employmen Agency.
Figu e12 shows ha he index beha es simila o e ime,
howe e he index cons uc ed wi h IAB Job Vacancy
da a is highe in he beginning and he end o he obse -
a ion pe iod. Howe e he e is no sys ema ic bias in he
ime se ies.
Coun e ac uals
In Fig.13 I compa e he baseline index o an index
whe e I ei he hold he acancy sha e ixed a he
beginning o he pe iod o he unemploymen sha e.
This is a simila exe cise as in Hu e and Webe
(2017). This allows o ecei e a hin o which ex en he
Fig. 12 Misma ch Index egis e ed s. IAB Job Vacancy Su ey, 2008–2022. All se ies a e seasonal adjus ed using X-12-ARIMA. The se ies s a s
in 2008 o accoun o he e ec o he s a ing alue on he se ies in he beginning. Sou ce: S a is ical O ice o he Fede al Employmen Agency,
own calcula ions. ©IAB
27 Page 16 o 23
A.Baue
Fig. 13 Misma ch indices holding ei he he unemploymen sha es o he acancy sha es cons an . All se ies a e seasonal adjus ed using
X-12-ARIMA. Sou ce: S a is ical O ice o he Fede al Employmen Agency, own calcula ions. ©IAB
Fig. 14 Misma ch indices holding he equi emen le el dis ibu ion cons an (a he mean le el o 2011) ac oss ime. All se ies a e seasonal
adjus ed using X-12-ARIMA. Sou ce: S a is ical O ice o he Fede al Employmen Agency, own calcula ions. ©IAB
Page 17 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
Fig. 15 Dis ibu ion o acancies and unemploymen ac oss he equi emen le el, 2007–2022. All se ies a e seasonal adjus ed using X-12-ARIMA.
Sou ce: S a is ical O ice o he Fede al Employmen Agency, own calcula ions. ©IAB
27 Page 18 o 23
A.Baue
mo emen o e ime is induced by mo emen s in he
dis ibu ion o acancies o unemployed. The igu e
shows he mo emen s om Janua y 2019 o he end
o 2022. While he pa e ns a e simila un il mid 2021,
a e he second hal o 2021 he e appea s some di e -
gence. The index holding he acancy sha es cons an ,
ends sideways. Hence, i i we e jus he dis ibu ion
o unemploymen ha would ha e mo ed misma ch
unemploymen s ayed ele a ed (compa e ed line wi h
blue line). The o ces ha pulled misma ch unemploy-
men down a e he COVID19-shock is likely o be
he ebound o acancies (see g een line in compa ison
o blue line). Howe e , he sligh upwa d mo emen
in mid 2022 appea s o be a less a o able e olu ion o
acancies.
In Fig.14 I compa e he index o occupa ion seg-
men s o an index a he occupa ion segmen s
×
Fig. 16 Dis ance be ween ac ual and op imal unemploymen ac oss occupa ion segmen s. S a is ic is calcula ed by assuming ha
i
u∗=
iu
.
Values below ze o indica e sho ages, abo e ze o excesses in e ms o unemployed pe sons. Sou ce: S a is ical O ice o he Fede al Employmen
Agency, own calcula ions. ©IAB
Table 1 Pe cen age change be ween business cycle da es in he dis ance be ween ac ual and op imal unemploymen ac oss
occupa ion segmen s
Peak and ough da es a e gi en by he Ge man Economic Council. The GFC pe iod e e s o da a om Janua y 2008 (peak) o Ap il 2009 ( ough) and o he CC o
da a be ween Feb ua y 2020 (peak) and Ap il 2020 ( ough)
Occupa ion segmen GFC CC
21: Occupa ions in he ood indus y, in gas onomy and in ou ism −3.70 1.07
22: Medical and non-medical heal h ca e occupa ions −1.80 −0.16
23: Se ice occupa ions in social sec o and cul u al wo k −0.83 −0.63
41: Se ice occupa ions in he IT-sec o and he na u al sciences −0.66 −0.49
14: Occupa ions in building and in e io cons uc ion −0.43 −0.27
33: Business ela ed se ice occupa ions −0.33 −0.20
11: Occupa ions in ag icul u e, o es y and ho icul u e −0.22 −0.01
31: Occupa ions in comme ce and ade −0.20 0.23
53: Occupa ions in cleaning se ices −0.14 0.10
Page 19 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
equi emen le el, o which calcula ion I used he
dis ibu ion o he acancy and unemploymen sha es
in 2011 and hold i cons an un il 2022. The igu e
shows, ha his coun e ac ual leads o a pa allel shi
o he misma ch index in compa ison o he index o
occupa ion segmen s. The index inc eases due o he
inc ease in he numbe o subma ke . Mo e impo -
an , he coun e ac ual index shows ha he de ia ion
be ween occupa ion segmen s and occupa ion seg-
men s
×
equi emen in he end mus s ems om he
e olu ion o he dis ibu ion o he equi emen le el
ac oss he acancy and unemploymen sha es.
Requi emen le els andoccupa ions
Fo he sake o eadabili y, I ocus an he le el o un- and
semi-skilled and specialis ac i i ies and plo he e olu-
ion o he acancy and unemploymen sha es ac oss
he di e en occupa ion segmen s. In bo h equi emen
le els, he occupa ional sha es a e di e en ly dis ib-
u ed, bu he e is no pa icula occupa ion, which d i es
he inc eases and dec eases. The igu e shows, ha he
unemploymen sha es o un-skilled and semi-skilled
wo ke s inc eased s onge han he acancy sha es, and
o he le el o specialis s he e is a s onge dec ease
in he unemploymen sha es han in he acancy sha es
(Fig.15).
Occupa ion segmen s
I s ands ou ha especially occupa ions wi h a high sha e
o unskilled and semi-skilled unemployed wo ke s end
o show excesses like occupa ions in cleaning se ices,
occupa ions in business managemen and o ganisa ion,
occupa ions in comme ce and ade. Occupa ions which
exhibi sho ages like manu ac u ing occupa ions, occu-
pa ions conce ned wi h p oduc ion echnology, medi-
cal and non-medical heal h ca e occupa ions o se ice
occupa ions in he IT-sec o and he na u al sciences
end o ha e a highe sha e o unemployed ha sea ch
o (highly) complex specialis ac i i ies. Two occupa ion
segmen s, namely occupa ions in building and in e io
cons uc ion and occupa ions in social sec o and cul-
u al wo k, swi ch om excesses o sho ages o e ime.
Occupa ions in he ood indus y, in gas onomy and in
ou ism change om sho ages o excesses. Gi en he
di e en na u e o he GFC and he CC, Fig.16 shows
ha di e en occupa ions played a ole du ing he c i-
ses. While du ing he GFC sho ages in he occupa ions
conce ned wi h p oduc ion echnology and manu ac u -
ing occupa ions we e educed, excess in occupa ions in
a ic and logis ics inc eased. Du ing he CC, i has been
occupa ions in he ood indus y, in gas onomy and in
ou ism and somewha occupa ions in cleaning se ices
whe e excesses inc eased. A special case a e medical and
non-medical heal h ca e occupa ions, whe e in bo h c i-
ses sho ages inc eased (Fig.16).
To analyse he cyclical pa e n, ied o di e en occupa-
ions as highligh ed abo e, Table1 shows he changes in
he de ia ion be ween ac ual and op imal unemploymen
ac oss occupa ion segmen s among peak and ough o
he GFC and he CC. I shows ha hese do no ollow
a clea pa e n. While in some occupa ions dis ances
dec eased in bo h c ises, in o he occupa ions hey
almos did no eac , and o some occupa ions he di e -
ences e ol ed di e en ly in he wo c ises.
Wi hin‑segmen e olu ion
See Fig.17.
27 Page 20 o 23
A.Baue
-200000
-100000
0
100000
200000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i ies Specialis ac i i ies
Complex specialis ac i i ies Highly complex ac i i ies
All equi emen le els
21: Occupa ions in he ood indus y, in gas onomy and in ou ism
-400000
-300000
-200000
-100000
0
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i ies Specialis ac i i ies
Complex specialis ac i i ies Highly complex ac i i ies
All equi emen le els
22: Medical and non-medical heal h ca e occupa ions
-50000
0
50000
100000
150000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i ies Specialis ac i i ies
Complex specialis ac i i iesHighly complex ac i i ies
All equi emen le els
23: Se ice occupa ions in social sec o and cul u al wo k
-50000
0
50000
100000
150000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i ies Specialis ac i i ies
Complex specialis ac i i ies Highly complex ac i i ies
All equi emen le els
31: Occupa ions in comme ce and ade
Fig. 17 Dis ance be ween ac ual and op imal unemploymen ac oss occupa ion segmen s. S a is ic is calcula ed by assuming ha
iu∗
=
iu
.
No e ha he e a e changes in he assignmen s o occupa ions wi h espec o he equi emen le el in he segmen s o sa e y and secu i y
occupa ions and he se ice occupa ions in social sec o and cul u al wo k. Sou ce: S a is ical O ice o he Fede al Employmen Agency, own
calcula ions. ©IAB
Page 21 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
Mobili y
This sec ion shows he mobili y upon hi ing ac oss
equi emen le els (see Table2) and occupa ion segmen s
(see Table3) in 2017 and 2022. The chances o mo ing
o a highe equi emen le el as sea ched o upon hi -
ing be ween 2017 and 2022 o e all dec eased. Especially
o he un- and semi-skilled wo ke s he e ec is la ge.
The chances o swi ching occupa ion segmen s inc eased
o e all sligh ly. In occupa ions such as Sa e y andsecu i y
occupa ions, Medical and non-medical heal h ca e occupa-
ions o Manu ac u ing occupa ions he chances o mo ing
o ano he segmen s inc eased clea ly, while he chances
dec eased o se ice occupa ions in social sec o and cul-
u al wo k.
-30000
-20000
-10000
0
10000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i ies Specialis ac i i ies
Complex specialis ac i i ies Highly complex ac i i ies
All equi emen le els
41: Se ice occupa ions in he IT-sec o and he na u al sciences
0
20000
40000
60000
80000
100000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Specialis ac i i ies Unskilled o semi-skilled ac i i ies
Complex specialis ac i i ies Highly complex ac i i ies
All equi emen le els
51: Sa e y and secu i y occupa ions
-100000
0
100000
200000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i ies Specialis ac i i ies
Complex specialis ac i i iesHighly complex ac i i ies
All equi emen le els
52: Occupa ions in a ic and logis ics
0
50000
100000
150000
200000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i iesSpecialis ac i i ies
Complex specialis ac i i ies All equi emen le els
53: Occupa ions in cleaning se ices
0
50000
100000
150000
200000
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Unskilled o semi-skilled ac i i ies Specialis ac i i ies
Complex specialis ac i i ies Highly complex ac i i ies
All equi emen le els
32: Occupa ions in business managemen and o ganisa ion
-200000
-150000
-100000
-50000
0
2007m1
2008m1
2009m1
2010m1
2011m1
2012m1
2013m1
2014m1
2015m1
2016m1
2017m1
2018m1
2019m1
2020m1
2021m1
2022m1
2023m1
Specialis ac i i ies Complex specialis ac i i ies
Highly complex ac i i ies All equi emen le els
33: Business ela ed se ice occupa ions
Fig. 17 con inued
27 Page 22 o 23
A.Baue
Table 3 Mobili y a es ac oss occupa ion segmen s in 2017 and 2022 o UE ansi ions. Sou ce: S a is ical O ice o he Fede al
Employmen Agency, own calcula ions. A ailable a : h ps:// s a i s ik. a bei sage n u . de/ Si eG lobals/ Fo ms/ Suche/ Einze lhe suche_
Fo mu la . h ml? opic_ = besch ae i gung- sozbe- bs - be u l- mobi ©IAB
*Mobili y desc ibes he sha e o mo es be ween unemploymen and employmen o a di e en occupa ion segmen han sea ched o
Ta ge occupa ion 2017 2022
Ra e o mobili y* Ra e o mobili y
Occupa ions in ag icul u e, o es y and ho icul u e 51.46 54.05
Manu ac u ing occupa ions 63.05 68.45
Occupa ions conce ned wi h p oduc ion echnology 71.11 69.46
Occupa ions in building and in e io cons uc ion 49.19 49.29
Occupa ions in he ood indus y, in gas onomy and in ou ism 52.86 55.65
Medical and non-medical heal h ca e occupa ions 35.43 41.45
Se ice occupa ions in social sec o and cul u al wo k 57.48 45.88
Occupa ions in comme ce and ade 62.35 58.82
Occupa ions in business managemen and o ganisa ion 76.37 75.70
Business ela ed se ice occupa ions x x
Se ice occupa ions in he IT-sec o and he na u al sciences 84.17 82.44
Sa e y and secu i y occupa ions 52.73 59.01
Occupa ions in a ic and logis ics 47.38 47.44
Occupa ions in cleaning se ices 49.53 52.02
Ac oss all occupa ion segmen s 53.87 54.63
Table 2 Mobili y ac oss equi emen le els in 2017 and 2022. Sou ce: S a is ical O ice o he Fede al Employmen Agency, own
calcula ions. ©IAB
Mobili y (up) desc ibes he sha e o mo es be ween unemploymen and employmen which a e abo e he equi emen le el sea ched o
2017 Hi ed a
Sea ched o Unskilled Specialis Complex Highly complex Mobili y
Unskilled 0.60 0.37 0.02 0.01 0.40
Specialis 0.19 0.70 0.06 0.04 0.11
Complex specialis 0.06 0.33 0.40 0.21 0.21
Highly complex 0.03 0.13 0.13 0.71 0.00
2022 Hi ed a
Job sea ched o Unskilled Specialis Complex Highly complex Mobili y (up)
Unskilled 0.67 0.29 0.02 0.02 0.33
Specialis 0.29 0.59 0.08 0.05 0.13
Complex specialis 0.09 0.29 0.42 0.21 0.21
Highly complex 0.05 0.11 0.15 0.69 0.00
Acknowledgemen s
I hank Enzo Webe , Ben Lochne and Alex Pa ke o hei aluable commen s
and sugges ions. Fu he mo e I wan o hank pa icipan s o he ASSLE
Taiwan 2023, 19 h IWH/IAB Wo kshop on labou ma ke policy, and o he 16 h
Wo kshop on Labou Economics o he Ins i u e o Labou Law and Indus ial
Rela ions in he Eu opean Union (IAAEU) and T ie Uni e si y o hei aluable
inpu .
Funding
I ecei ed no inancial suppo o he esea ch, au ho ship, and/o publica ion
o his a icle. The au ho con i ms sole esponsibili y o he ollowing: s udy
concep ion, da a collec ion, econome ic analysis and discussion o esul s.
Da a a ailabili y
The da ase s analysed in he cu en s udy a e pa ially publicly a ailable a
he Fede al Employmen Agency (h ps:// s a i s ik. a bei sage n u . de/). The job-
inding and sepa a ion a es a he occupa ional le el a e no publicly a ail-
able due o da a p o ec ion easons bu a e a ailable om he co esponding
au ho on easonable eques .
Decla a ions
Compe ing in e es s
I decla e ha I ha e no known compe ing inancial in e es s o pe sonal
ela ionships ha could ha e appea ed o in luence he wo k epo ed in his
pape .
Page 23 o 23 27 Accoun ing o quali ica ion inmisma ch unemploymen
Recei ed: 9 Feb ua y 2024 Accep ed: 30 Oc obe 2024
Re e ences
Ba nichon, R., Figu a, A.: Labo ma ke he e ogenei y and he agg ega e
ma ching unc ion. Am. Econ. J. Mac oecon. 7(4), 222–249 (2015)
Baue , A.: Misma ch unemploymen : E idence om ge many, 2000–2010.
Technical epo , IAB-Discussion Pape (2013)
Baue , A., Webe , E.: Co id-19: how much unemploymen was caused by he
shu down in Ge many? Appl. Econ. Le . 28(12), 1053–1058 (2021a)
Baue , A., Webe , E.: Lockdown leng h and s eng h: labou -ma ke e ec s in
Ge many du ing he co id-19 pandemic. Technical epo , IAB-Discussion
Pape (2021b)
Böheim, R., Ch is l, M.: Misma ch unemploymen in aus ia: The ole o egional
labou ma ke s o skills. Reg. S ud. Reg. Sci. 9(1), 208–222 (2022)
Bossle , M., Popp, M.: Labo demand on a igh leash. a Xi p ep in a Xi : 2203.
05593 (2022)
Bossle , M., Popp, M.: Labo demand on a igh leash. Technical epo , IZA
Discussion Pape s (2024)
B adley, J., Kügle , A.: Labo ma ke e o ms: an e alua ion o he ha z policies
in ge many. Eu . Econ. Re . 113, 108–135 (2019)
B ücke , H., Kosyako a, Y., Schuss, E.: low-skilled mig an s and jobs, p. 51.
In eg a ing low-skilled mig an s in he digi al age: Eu opean and us
expe ience (2020)
Bundesagen u ü A bei , S.: Be ich e. In A bei sma k kompak - Fluch mig a-
ion, Nu embe g (2020)
Ch is l, M., De Poli, S., Hu kens, T., Peichl, A., and Ricci, M.: The ole o sho - ime
wo k and disc e iona y policy measu es in mi iga ing he e ec s o he
co id-19 c isis in ge many. In e na ional Tax and Public Finance, p. 1–30
(2022)
Fo sy he, E., Kahn, L.B., Lange, F., Wicze , D.: Whe e ha e all he wo ke s gone?
ecalls, e i emen s, and ealloca ion in he co id eco e y. Labou Econ.
78, 102251 (2022)
Geh ke, B., Webe , E.: Sho - ime wo k, layo s, and new hi es in ge many: how
he co ona c isis di e s om he inancial c isis o 2009. In IAB-Fo um, ol.
16, p. 2020 (2020)
He z, B., Van Rens, T.: Accoun ing o misma ch unemploymen . J. Eu . Econ.
Assoc. 18(4), 1619–1654 (2020)
Hochmu h, B., Kohlb eche , B., Me kl, C., Ga ne , H.: Ha z i and he decline
o ge man unemploymen : a mac oeconomic e alua ion. J. Econ. Dyn.
Con ol 127, 104114 (2021)
Hu e , C., Webe , E.: Misma ch and he o ecas ing pe o mance o ma ching
unc ions. Ox o d Bull. Econ. S a . 79(1), 101–123 (2017)
Jackman, R., Laya d, R., Sa ou i, S.: Misma ch, a amewo k o hough . In
Padoa-Schioppa, F. (ed.) Misma ch and labou mobili y. Camb idge
Uni e si y P ess, Camb idge (1991)
Jackman, R., Rope , S., e al.: S uc u al unemploymen . Ox o d Bull. Econ. S a .
49(1), 9–36 (1987)
K ebs, T., Sche el, M.: Mac oeconomic e alua ion o labo ma ke e o m in
ge many. IMF Econ. Re . 61(4), 664–701 (2013)
Launo , A., Wälde, K.: Es ima ing incen i e and wel a e e ec s o nons a iona y
unemploymen bene i s. In . Econ. Re . 54(4), 1159–1198 (2013)
Launo , A., Wälde, K.: The employmen e ec o e o ming a public employ-
men agency. Eu . Econ. Re . 84, 140–164 (2016)
Ma hes, B., Meinken, H., Neuhause , P.: Be u ssek o en und be u ssegmen e
au g undlage de kldb 2010, p. 8. Me hodenbe ich de S a is ik de BA,
Nü nbe g (2015)
Pa e son, C., Şahin, A., Topa, G., Violan e, G.L.: Wo king ha d in he w ong place:
a misma ch-based explana ion o he uk p oduc i i y puzzle. Eu . Econ.
Re . 84, 42–56 (2016)
Paulus, W., Ma hes, B., e al.: The Ge man classi ica ion o occupa ions
2010-s uc u e, coding and con e sion able. FDZ-Me hoden epo 8,
2013 (2013)
Pizzinelli, C., Shiba a, I.: Has co id-19 induced labo ma ke misma ch? e i-
dence om he US and he UK. Labou Economics, p. 102329 (2023)
Rolle , C., Rul , C., Tamminga, M.M.: I ’s a misma ch! o e educa ion and ca ee
mobili y in Ge many. Ge man Econ. Re . 21(4), 493–514 (2020)
Şahin, A., Song, J., Topa, G., Violan e, G.L.: Misma ch unemploymen . Am. Econ.
Re . 104(11), 3529–3564 (2014)
Sa inge , M. e al.: Quali a i e misma ches. Found. T ends Mic oecon.
8(1–2):1–168 (2012)
Schioppa, F. P. e al.: Misma ch and labou mobili y. Technical epo , Cam-
b idge Uni e si y P ess (1991)
Shiba a, I.: Is labo ma ke misma ch a big deal in Japan? BE J. Mac oecon.
20(2):20160179 (2020)
Tu ell, A., Speigne , B., Copple, D., Djumalie a, J., Thu good, J.: Is he UK’s
p oduc i i y puzzle mos ly d i en by occupa ional misma ch? an analysis
using big da a on job acancies. Labou Econ. 71, 102013 (2021)
Publishe ’s No e
Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in pub-
lished maps and ins i u ional a ilia ions.
