Is training apprentices worthwhile? Company perspectives on costs, benefits and training quality 2025

Alexande Geh e , And eas Kuhn, Jü g Schwe i
IS TRAINING APPRENTICES WORTHWHILE?
COMPANY PERSPECTIVES ON COSTS,
BENEFITS AND TRAINING QUALITY 2025
SFUVET
Swiss excellence in oca ional
educa ion and aining
SWISS FEDERAL UNIVERSITY
FOR VOCATIONAL EDUCATION
AND TRAINING
EXECUTIVE SUMMARY 2
INTRODUCTION 3
1 WHY CONSIDER COSTS AND BENEFITS IN VOCATIONAL EDUCATION
AND TRAINING? 4
1.1 S a egic conside a ions o aining companies 4
1.2 Cos –bene i s udies gene a e insigh s o s ee ing he sys em 5
2 THE COST–BENEFIT MODEL 7
2.1 G oss cos s 7
2.2 P oduc i e ou pu 9
2.3 Ne bene i 10
2.4 Bene i o con inued employmen ( ec ui men oppo uni y e u ns) 11
3 SURVEY PROCESS 13
3.1 Re ision o he su ey amewo k 13
3.2 P e es 15
3.3 Sampling 15
3.4 Field phase 17
3.5 Responses in he wo subsamples 17
3.6 Composi ion o he wo ne samples 18
3.7 Plausibili y checks 21
3.8 App oach o missing a iable alues 22
3.9 Sample weigh ing 22
4 RESULTS 23
4.1 A e age cos s and bene i s o ini ial oca ional educa ion and aining 23
4.2 Cos s and bene i s o he wo-yea ini ial oca ional educa ion and aining
leading o an FVETC 25
4.3 Cos s and bene i s o he h ee-yea ini ial oca ional educa ion and aining
leading o an FVETD 28
4.4 Cos s and bene i s o he ou -yea ini ial oca ional educa ion and aining
leading o an FVETD 30
4.5 Resul s by company size 33
4.6 Resul s by sec o 34
4.7 Resul s by majo egion 35
4.8 Resul s by occupa ion 36
4.9 Findings on he bene i s o con inued employmen 39
4.10 Ex apola ion o aining cos s, p oduc i e ou pu and ne bene i 39
4.11 Compa ison wi h Ge many 40
4.12 Conclusion 41
5 MULTIVARIATE ANALYSES 42
5.1 Analysis o he 2022/23 su ey 42
5.2 Mul i a ia e compa ison wi h he su ey o he 2016/17 app en iceship yea 44
CONTENTS
Imp in
Edi o : Swiss Fede al Uni e si y o Voca ional Educa ion and T aining SFUVET
Sugges ed ci a ion: Geh e , A., Kuhn, A. & Schwe i, J. (2025). Is aining app en ices wo hwhile? Company
pe spec i es on cos s, bene i s and aining quali y 2025. Zolliko en: Swiss Fede al Uni e si y o Voca ional Educa ion
and T aining SFUVET.
No embe 2025. All igh s ese ed. Rep oduc ion o use o con en only wi h pe mission om SFUVET
Ve sion: 1.0
Con ac : Swiss Fede al Uni e si y o Voca ional Educa ion and T aining SFUVET, Jü g Schwe i and Alexande Geh e ,
Ki chlindachs asse 79 CH-3052 Zolliko en, +41 58 458 27 82 and +41 58 458 27 88, cos -bene i @s u e .swiss,
www.s u e .swiss/cos -bene i
6 TRAINING PLANS AND THE VOCATIONAL BACCALAUREATE 46
6.1 VET o dinances and aining plans 46
6.2 Fede al Voca ional Baccalau ea e 1 om a company pe spec i e 48
7 IN-HOUSE TRAINING QUALITY 50
7.1 Wha is he de ini ion o in-house aining quali y? 50
7.2 How do aining companies assess hei aining quali y? 50
7.3 Fi e aining ypes o aining companies 53
7.4 How do he companies in he di e en aining ypes di e ? 55
7.5 Why is aining quali y impo an ? 56
8 REASONS TO TRAIN (OR NOT) 57
8.1 Companies ha do no p o ide app en iceship aining 58
8.2 Compa ison o easons o and agains aining 58
REFERENCES 62
2 Company pe spec i es on cos s, bene i s and aining quali y 2025
EXECUTIVE SUMMARY
comp ehensi e occupa ional skills among app en ic-
es ha go beyond he needs o indi idual companies.
Con e sely, 19 pe cen o companies epo ha hey
p o ide addi ional quali ica ions no included in he
aining plan.
– The Fede al Voca ional Baccalau ea e Type 1 aken
alongside he app en iceship (BM 1) esul s in sligh -
ly lowe p oduc i e ou pu o companies due o ap-
p en ices’ inc eased absences. This educes he ne
bene i o companies by a ound CHF 3,000, bu i e-
mains posi i e.
– T aining companies we e asked eigh ques ions abou
he quali y o hei aining and, based on hei e-
sponses, we e g ouped in o i e aining ypes. O e -
all, mo e han h ee qua e s o aining companies
belong o he h ee ypes ha p o ide sa is ac o y
o e y good aining quali y. A qua e o companies
p o ide excellen aining quali y. A hi d o compa-
nies o e a e age quali y aining. Nine een pe cen
o e aining on a mo e ad hoc basis wi h li le plan-
ning, bu s ill achie e a e age o good quali y esul s.
The ou h g oup, desc ibed as ‘au ho i a ian’ in hei
aining app oach, makes up 8 pe cen : hese gi e
app en ices li le scope o ind au onomous solu ions
and pay less a en ion o a oiding po en ial o e load.
Almos 15 pe cen o companies a e hei aining
quali y as below a e age ac oss all dimensions.
– Fo aining companies, he s onges mo i a ion o
o e ing aining is he sense ha i is a collec i e e-
sponsibili y o he business communi y and he e o e
a con ibu ion o socie y. A he same ime, econom-
ic mo i es ha e gained in impo ance compa ed wi h
he s udy conduc ed six yea s ea lie : he second mos
common eason ci ed is he wish o ain skilled wo k-
e s who can be e ained in he company in he long
e m. Appa en ly, he sho age o skilled labou is
s eng hening his mo i e o o e ing aining. Com-
panies also place g ea e impo ance on being able
o make use o app en ices as pa o hei wo k o ce
du ing he aining pe iod.
– The online su ey p oduced a da ase wi h in o -
ma ion om 6,655 aining companies and 3,655
non- aining companies. The esponse a e among
companies andomly selec ed om he Fede al S a-
is ical O ice’s business egis e was 39 pe cen o
aining companies and 31 pe cen o non- aining
companies.
– Ac oss all app en iceships, he a e age ne bene i in
he aining yea 2022/23 was jus o e CHF 4,500 pe
app en iceship yea and place. A majo i y o app en-
iceships (a ound 71 pe cen ) he e o e achie ed a
ne bene i o he aining company. The ne bene-
i is de ined as he alue o app en ices’ p oduc i e
wo k du ing he aining pe iod (‘p oduc i e ou pu ’)
minus all aining cos s (‘g oss cos s’).
– This posi i e a e age esul is consis en wi h com-
panies’ own assessmen s, wi h 80 pe cen epo ing
ha hey a e ai ly o e y sa is ied wi h he cos –ben-
e i balance o aining hei app en ices.
– The esul s o he wo- and h ee-yea app en ice-
ships do no di e signi ican ly om hose o he
p e ious cos –bene i su ey in he 2016/17 aining
yea . The a e age ne bene i pe app en iceship is
CHF 9,630 o he Fede al Voca ional Educa ion and
T aining Ce i ica e (FVETC) and CHF 13,940 o he
h ee-yea Fede al Voca ional Educa ion and T aining
Diploma (FVETD). The ne bene i o CHF 17,510 o
ou -yea FVETD app en iceships is signi ican ly high-
e han in 2016/17. This is pa ly due o app en ice-
ships ha swi ched om a h ee- o a ou -yea du a-
ion. In addi ion, he su ey ques ionnai e has been
simpli ied and imp o ed since he las su ey, which
may also ha e con ibu ed o he change in esul s.
– Companies gene ally ind ha he VET o dinances
(BiVo) and aining plans (BiPla) co espond well o
hei needs. On a e age, companies conside 82 pe
cen o he aining con en de ined in he BiPla o be
ele an o hei ope a ions. This is a high p opo ion
gi en ha hese documen s a e in ended o ensu e
Company pe spec i es on cos s, bene i s and aining quali y 2025 3
Ini ial oca ional educa ion and aining (IVET) is based
on companies’ willingness o ain young people – and
also adul s – in an app en iceship. Fo his willingness o
be main ained, companies mus see IVET as wo hwhile.
The S a e Sec e a ia o Educa ion, Resea ch and Inno-
a ion (SERI) commissioned his s udy o p o ide an up-
o-da e pic u e o company cos s and bene i s and he
ac o s in luencing hem.
The Swiss Fede al Uni e si y o Voca ional Educa ion
and T aining (SFUVET) was awa ded he con ac o con-
duc he su ey h ough a WTO public p ocu emen p o-
cedu e. This epo ou lines he concep , planning and
ield phase o he su ey, as well as he da a p ocessing
and main indings.
The i h su ey on he cos s and bene i s o ini ial oca-
ional educa ion and aining builds on a se ies o ea lie
su eys. The basic me hodology o he p e ious cos –
bene i su eys has emained la gely unchanged since
he i s su ey in he 2000/2001 aining yea (Schwe i e
al., 2003). To educe he bu den on pa icipa ing compa-
nies and imp o e da a quali y, he online ques ionnai e
o he cu en su ey was comple ely o e hauled. Ne -
e heless, he esul s o he a ious su eys desc ibed in
Chap e 1 emain la gely compa able. Wi h an inc eased
esponse a e, his s udy is based on da a om mo e
han 6,600 aining companies. To make he esul s mo e
accessible, indings on indi idual app en iceship occu-
pa ions a e p o ided in dedica ed chap e s on he SFU-
VET websi e.
h ps://s u e .swiss/cos -bene i
Since he epo eplica es many o he co e esul s o he
p e ious su ey om he 2016/17 aining yea (Geh e e
al., 2019) using newly collec ed da a o 2022/23, i does
no a emp o e o mula e any a gumen s ha ha e e-
mained unchanged. This makes i easie o eade s o
compa e he esul s o he wo epo s. A he same ime,
he cu en epo also p esen s many en i ely new ind-
ings, including Chap e 7 on in-house aining quali y.
The au ho s would like o hank all pa icipa ing compa-
nies and ins i u ions in ol ed in he su ey, in pa icula
Jü g Bie i and Philipp Theile om SERI, and Alain Chas-
so , Na alia Do on so a, Lau en In e sin and F édé ic
on Kessel om he Fede al S a is ical O ice (FSO). They
a e also g a e ul o Samuel Mühlemann, Ha ald P ei e ,
Gud un Schöne eld and Felix Wenzelmann o hei al-
uable inpu in e ising he ques ionnai e. Las bu no
leas , he au ho s would like o hank he SFUVET s a
in ol ed in he s udy, especially Yolanda Hübsch, I ana
Lo ic, Tania Pichie i, Filippo Pus e la, Lea Ra hle and
Ma ilde Wenge , as well as assis an s Anacle Deguenon,
Espe anza Ma x, Lo enzo Peduzzi, Lea Raas and Julia
S adle .
INTRODUCTION

4 Company pe spec i es on cos s, bene i s and aining quali y 2025
In o ma ion on he cos s and bene i s o IVET om he
pe spec i e o companies is ele an o di e en a ge
g oups: i s ly, o he companies hemsel es, and sec-
ondly o all s akeholde s in he VET sys em who depend
on companies’ olun a y pa icipa ion in aining. The
ollowing sec ion he e o e s a s by ou lining he ela-
ionship be ween aining s a egies and cos s and ben-
e i s a company le el. I hen desc ibes how cos –bene-
i da a can con ibu e o s ee ing he VET sys em.
The a ailable li e a u e on labou ma ke and educa ion
economics analysing wha mo i a es companies o ain
app en ices is ex ensi e. This chap e p o ides only a
b ie in oduc ion and does no a emp a comp ehen-
si e e iew. Fo Ge man-language o e iews, see Wol e
(2008) and S uple & Wol e (2012); o English-language
o e iews, see Mühlemann (2019) and Mühlemann &
Wol e (2014).
1.1 S a egic conside a ions o
aining companies
F om a business pe spec i e, app en iceship aining
should be wo hwhile o companies. A ange o cos and
bene i ac o s come in o play, bu hese a e no always
easy o iden i y o quan i y, since aining is a ely p e-
sen ed in ull in company accoun s. Wi h he help o he
cos model de eloped o his su ey (see Chap e 2),
companies can ca y ou hei own cos –bene i calcu-
la ion. The su ey esul s also show companies he a e -
age cos s and bene i s obse ed among o he companies
in Swi ze land, including in compa ison wi h companies
in he same sec o o o e ing he same app en iceship.
This allows hem o benchma k hei own aining ig-
u es agains he su ey a e ages and be e assess he
cos -e iciency o hei aining.
Bu he issue is no simply whe he a company does
‘be e ’ o ‘wo se’ by showing lowe o highe ne cos s.
Cos –bene i esul s e lec di e ences in he aining
s a egies and o ganisa ion chosen by each company.
Depending on he objec i es a company pu sues wi h
aining, he cos –bene i balance will a y. The c ucial
ques ion is whe he he way in which aining is o gan-
ised enables companies o mee hei business goals
while a he same ime p o iding app en ices wi h
high-quali y aining and sound p epa a ion o hei
ca ee s. Cos –bene i analyses gi e aining manage s
an oppo uni y o e lec on and imp o e hei aining
p ac ices acco dingly.
Companies mus cons an ly decide which employees
wi h which quali ica ions hey need o p oduce hei
goods o se ices e icien ly. In Swi ze land, mos com-
panies employ skilled wo ke s wi h ei he a Fede al Vo-
ca ional Educa ion and T aining Diploma (FVETD) o a
Fede al Voca ional Educa ion and T aining Ce i ica e
(FVETC). They mus decide whe he o ec ui skilled
wo ke s wi h IVET om he ex e nal labou ma ke o
(also) o ain hem in-house. This decision is complex
because many ele an ac o s a e unce ain. App en-
iceships las wo o ou yea s, and app en iceship
con ac s canno easily be e mina ed. Hi ing app en-
ices, he e o e, ep esen s a mul i-yea commi men o
companies – one ha ca ies isks, since u u e p oduc
demand, s a ing equi emen s and app en ice pe o -
mance canno be p edic ed wi h ce ain y. Companies
mus he e o e weigh he expec ed cos s and bene i s o
aining agains hese isks.
Cos s include he s a cos s a ising om aining and
supe ising app en ices, app en ice wages, and all
aining- ela ed ma e ial cos s and ees. These cos s a e
o se by a ious bene i s. Du ing hei aining, app en-
ices wo k p oduc i ely in he company and can pe o m
asks ha would o he wise ha e o be ca ied ou by
o he employees. These may be skilled wo ke s bu also
semi-skilled o unskilled employees. Since app en ice
wages a e lowe han hose o hese ca ego ies o em-
ployees, he p oduc i e ou pu o app en ices alone can
gene a e su icien bene i s o co e all aining cos s.
Ano he bene i is ha app en ices can be employed as
skilled wo ke s a e comple ing hei app en iceship.
This bene i s he company because he app en ices
a e al eady well in eg a ed in o he company, whe eas
skilled wo ke s om he ex e nal labou ma ke mus
i s be ec ui ed, hi ed and subsequen ly ained. Keep-
ing app en ices on allows companies o a oid hese
cos s. In addi ion, aining companies a e able o assess
he wo k quali y and p oduc i i y o hei app en ices
1 WHY CONSIDER COSTS AND BENEFITS IN
VOCATIONAL EDUCATION AND TRAINING?
Company pe spec i es on cos s, bene i s and aining quali y 2025 5
and can selec i ely e ain hose who ha e pe o med
pa icula ly well. This in ol es less isk han ec ui ing
skilled wo ke s om he labou ma ke . Skilled wo ke s
ained in-house also o en show g ea e loyal y o he
company, which educes u no e cos s (Euwals & Win-
kelmann 2004). F om an economic heo y pe spec i e,
he bene i o e aining app en ices lies in he ac ha
hose ained in-house p o ide companies wi h highe
p oduc i i y, bu compa ed wi h skilled wo ke s om
he ex e nal labou ma ke , his is no – o only pa ly –
e lec ed in highe wages. T aining companies he e o e
ea n a ‘ en ’ on hei o me app en ices. The ele an
li e a u e (e.g. Acemoglu & Pischke, 1999) examines he
condi ions unde which companies ha e mo e o ewe
oppo uni ies o ea n such en s on he labou ma ke .
App en iceship aining also b ings o he ad an ages
ha bene i companies di ec ly. No only can i boos
hei epu a ion wi h cus ome s and (po en ial) employ-
ees, bu i can e en con ibu e o knowledge ans e
and inno a ion (Backes-Gellne & Lehne , 2023). Ap-
p en ices may pass on ideas hey ha e picked up in in-
e -company cou ses o a oca ional school, whe he
om eache s o classma es. T aining also keeps compa-
nies mo e closely connec ed o hei ade associa ion
o p o essional o ganisa ion, s eng hening communi-
ca ion ac oss he sec o and encou aging knowledge ex-
change (Rupie a & Backes-Gellne , 2018).
A undamen al dis inc ion be ween aining s a egies
is based on he p ima y economic mo i e o aining.
When a company ains app en ices p ima ily o secu e
i s u u e skilled wo k o ce, his is e e ed o as an ‘in-
es men -o ien ed’ aining s a egy. The company
seeks o ain i s u u e skilled wo ke s as closely as pos-
sible o i s own needs and is he e o e o en p epa ed
o bea ne cos s du ing he app en iceship. These a e
ecouped by keeping app en ices on and bene i ing om
hei subsequen p oduc i i y. In con as , unde a ‘p o-
duc ion-o ien ed’ aining s a egy (Lindley, 1975), he
company seeks o co e all aining cos s du ing he ap-
p en iceship h ough he p oduc i e wo k o app en ic-
es. Companies ollowing his s a egy a e no necessa ily
dependen on e aining app en ices: e en i app en ices
a e ec ui ed by o he companies a e comple ing hei
aining, he app en iceship has s ill been wo hwhile.
The si ua ion is di e en o companies pu suing an
in es men -o ien ed s a egy, as hese depend on e-
aining enough app en ices o mee hei skilled labou
needs and a oid he ec ui men cos s o hi ing om he
ex e nal labou ma ke . I many app en ices do no e-
main wi h he company because hey a e ec ui ed else-
whe e – o because hey con inue hei educa ion wi h a
Fede al Voca ional Baccalau ea e o a uni e si y o ap-
plied sciences deg ee – he company canno ecoup i s
in es men . This is likely o be an impo an eason why
la ge companies equen ly es ablish egional aining
alliances (Wal he e al., 2005) o in ol e o he compa-
nies in app en iceship aining in hei a ea.
Companies can use a a ie y o le e s o os e he de-
elopmen o app en ices’ skills du ing hei app en-
iceship, and hese also shape he cos –bene i balance.
Pa icula ly impo an and cos ly a e he aining hou s
p o ided by aining manage s in he company. In a-
s uc u e o aining also plays a ole. Bo h aspec s e-
qui e good planning and a ge ed use. Ye e en du ing
he app en iceship, hese should no be seen pu ely as
cos ac o s: i app en ices build hei skills as a esul ,
hey can also be deployed mo e p oduc i ely – bo h
du ing he app en iceship and, whe e applicable, a e -
wa ds. A he same ime, he ype o p oduc i e labou
i sel con ibu es o hei skills de elopmen . By he end
o he app en iceship a he la es , app en ices should
be compe en in he asks hey will la e pe o m as
skilled wo ke s. I he e o e makes sense o i ms ha
app en ices no only become amilia wi h hese p oduc-
i e asks du ing he app en iceship, bu also gain some
p ac ical expe ience o hem. Fo companies, deploying
app en ices in his way also helps educe he wo kload
o skilled wo ke s.
The cos –bene i su eys allow hese le e s in aining
companies o be analysed in de ail. In his way, compa-
nies can lea n how he di e en elemen s a ec cos s
and bene i s in aining companies (Chap e 2). The a e -
age alues o he main indica o s a e also p esen ed o
each app en iceship occupa ion in dedica ed chap e s
(a ailable online a h ps://s u e .swiss/cos -bene i ).
6 Company pe spec i es on cos s, bene i s and aining quali y 2025
1.2 Cos –bene i s udies gene a e
insigh s o s ee ing he sys em
Fo VET pa ne s, policymake s and esea che s, cos –
bene i su eys make i easie o unde s and and ake
in o accoun companies’ aining mo i es. When VET
o dinances o IVET p og ammes a e e ised, he num-
be o lessons and days in oca ional school o e en he
o al du a ion o he app en iceship may, o example,
be changed. Such changes a ec bo h he aining e o
equi ed o companies and he bene i hey de i e om
he p oduc i e wo k o app en ices. The ac ual changes
in companies’ cos s and bene i s he e o e depend on
how aining is o ganised in he espec i e app en ice-
ship occupa ion, o ins ance on how aining e o and
p oduc i e wo k a e dis ibu ed ac oss he app en ice-
ship yea s.
Resea ch using cos –bene i da a om Ge many and
Swi ze land has undamen ally changed he unde -
s anding o how he dual app en iceship sys em wo ks
in hese wo coun ies. In Ge many, companies eco ded
subs an ial ne cos s a he end o he app en iceship
pe iod in he cos –bene i su eys conduc ed by he
Fede al Ins i u e o Voca ional Educa ion and T aining
(BIBB) om he ou se (e.g. on Ba deleben e al., 1995).
In e na ional labou ma ke economics esea ch he e-
o e compa ed Ge many wi h he USA and emphasised
he impo ance o labou ma ke egula ions in Ge many
o he dual sys em. Va ious in e en ions in he labou
ma ke (dismissal p o ec ion, collec i e ag eemen s,
s ong ade unions, wo ks councils, e c.) limi he mo-
bili y o Ge man skilled wo ke s compa ed wi h he el-
a i ely ligh ly egula ed US labou ma ke . This allows
aining companies o con inue employing hei o me
app en ices a wages somewha below hei p oduc i -
i y, gene a ing a en ha o se s he ne cos s o hei
aining. F om his pe spec i e, labou ma ke egula ion
p o ided he basis o he exis ence o he Ge man dual
sys em: Ge many is cha ac e ised by a ‘low qui s/high
aining’ equilib ium – a si ua ion in which app en ices
seldom lea e he company a e aining (low qui s) and
in which a g ea deal o aining akes place (high ain-
ing). The USA, on he o he hand, has a ‘high qui s/low
aining’ equilib ium, wi h high labou mobili y bu li le
company-based aining (Acemoglu & Pischke, 1998).
When he esul s o he 2000 su ey o Swi ze land we e
published (Schwe i e al., 2003), hey showed ha com-
panies achie ed a ne bene i on a e age. Compa ed
wi h he Ge man esul s, his came as a majo su p ise
o bo h he s udy au ho s and hei colleagues a BIBB.
Howe e , i explains how Swi ze land is able o sus ain
a ‘high qui s/high aining’ equilib ium. Since he Swiss
labou ma ke is a less egula ed han he Ge man one,
app en ice mobili y a e aining is highe han in Ge -
many. Companies can he e o e ely less han in Ge ma-
ny on co e ing hei aining cos s by e aining app en-
ices. Ins ead, mos companies ensu e ha hey co e
hei g oss cos s du ing he app en iceship h ough ap-
p en ices’ p oduc i e wo k (Mühlemann e al., 2010).
Swi ze land he e o e ep esen s a ‘ hi d way’ o oca-
ional educa ion and aining compa ed wi h he USA,
which has no egula ed VET sys em, and Ge many,
which combines VET wi h a egula ed labou ma ke .
These indings we e con i med in ano he Ge man s udy
(Jansen e al., 2015), which showed ha de egula ion o
he labou ma ke since 2003 had led Ge man companies
o make g ea e p oduc i e use o app en ices, he eby
educing he ne cos s o aining.
Cos –bene i s udies o e ecen decades ha e he e o e
sha pened ou unde s anding o how he dual sys ems
in Ge many and Swi ze land unc ion, as well as o hei
simila i ies and di e ences.
Company pe spec i es on cos s, bene i s and aining quali y 2025 7
The cos –bene i model is used o calcula e he cos s
and bene i s o aining companies based on he in o -
ma ion hey p o ide. The aim is o compa e he si ua-
ion o aining companies wi h a hypo he ical si ua ion
in which he same companies had no app en ices. How
much ex a ime do companies spend supe ising ap-
p en ices? Wha asks do app en ices pe o m ha add
alue o he company?
Fo ce ain cos elemen s, companies can p o ide ig-
u es di ec ly in mone a y e ms. These eed s aigh in o
he cos –bene i model – o ins ance, app en ice wages
o ma e ial expenses. O he da a, such as he ime spen
by aine s and he p oduc i e hou s con ibu ed by ap-
p en ices, mus i s be con e ed in o mone a y cos s
and bene i s.
The ne bene i o aining app en ices o a compa-
ny co esponds o o al p oduc i e ou pu minus g oss
cos s. The esul is a posi i e ne bene i when p oduc-
i e ou pu exceeds g oss cos s, and ne cos s when he
opposi e is he case.
In p inciple, cos s and bene i s o a aining company
can a y be ween indi idual app en ices. Some app en-
ices equi e less guidance and become p oduc i e mo e
quickly. To educe he e o equi ed by companies,
howe e , his su ey did no ask abou e e y indi idual
app en ice, bu abou he a e age o all app en ices in
he same app en iceship yea . The ne bene i calcula ed
in he cos –bene i model he e o e always e e s o he
a e age app en ice in a gi en yea and may a y ac oss
he cou se o he app en iceship wi hin a company.
2 THE COST–BENEFIT MODEL
The cos s o aining app en ices na u ally a ise o a
company only du ing he app en iceship pe iod. By con-
as , companies also bene i a e he app en iceship i
hey e ain ained app en ices and he eby educe he
need o ec ui s a om he labou ma ke . This bene i
om e en ion – known as he ‘ ec ui men oppo uni-
y e u ns’ – is also included in he cos –bene i model.
O he bene i elemen s, such as epu a ional gains om
aining ac i i y, a e no included.
The cos –bene i model desc ibed below essen ially
co esponds o he one used in p e ious Swiss s udies
(Schwe i e al., 2003; Mühlemann e al., 2007b; S uple
& Wol e , 2012; Geh e e al., 2019) and Ge man s udies
(Beich e al., 2004; Schön eld e al., 2016; Wenzelmann
& Schön eld, 2022) on he cos s and bene i s o app en-
iceship aining. Since he ques ionnai e was comp e-
hensi ely e ised o his su ey, he cos –bene i model
had o be sligh ly adjus ed in a ew elemen s. O e all,
howe e , compa abili y wi h ea lie Swiss and wi h ea li-
e and mo e ecen Ge man s udies (see Chap e 4.11) is
la gely ensu ed. The ex explici ly highligh s any esul s
ha may ha e been in luenced by he e ised su ey
amewo k.
2.1 G oss cos s
App en ice wage cos s
One o he la ges cos i ems o companies is app en ice
wages. These consis mainly o he app en ices’ mon h-
ly g oss wage, bu also include non-wage labou cos s,
any egula and i egula special paymen s, and he hi -
een h-mon h sala y.
14 Company pe spec i es on cos s, bene i s and aining quali y 2025
ime spen a oca ional school…’), since he con ex o
each ques ion was clea e and esponden s could ollow
a logical h ead. The issue wi h his, howe e , is ha his
kind o s uc u e equi es he ques ionnai e o be com-
ple ed in sequence. Fo his eason, unlike in 2016/17, i
was no longe possible o skip en i e chap e s.
Simple language
The 2016/17 su ey ques ionnai e was based on hose
used in ea lie su eys in Ge many and Swi ze land. Fo
easons o compa abili y, much o he e minology and
sen ence s uc u e was e ained; howe e , he analy-
sis o p e ious su eys desc ibed abo e indica ed ha
simpli ying he language would make he ques ionnai e
easie o comple e and imp o e da a quali y. Whe e e
possible, echnical ja gon was eplaced by clea e , mo e
sel -explana o y e ms. Fo ins ance, a he han asking
abou ‘ aining acili ies’ – a e m o en con used wi h
he company i sel – and dis inguishing hese om ‘ded-
ica ed aining ooms o company-based ins uc ion’,
companies a e now asked whe he hey ha e ‘speci ic
aining ooms’ ha a e ‘used exclusi ely o aining
and con inuing educa ion’. This app oach makes i pos-
sible o cap u e he in o ma ion needed o he calcula-
ions wi hou ha ing o ely on a common unde s and-
ing o e minology o leng hy explana ions.
An impo an example o he linguis ic e ision is he
e minology o he wo ca ego ies o asks ha app en-
ices pe o m in he wo kplace (see Chap e 2). Ins ead
o asking abou ‘p oduc i e ou pu I: simple asks’ and
‘p oduc i e ou pu II: di icul asks’, he ques ionnai e
now consis en ly e e s o ‘unskilled asks’ and ‘skilled
asks’. The e ms ‘simple’ and ‘di icul ’ depend on he
pe spec i e o he esponden : someone wi h ex en-
si e wo k expe ience is likely o ega d a la ge pa o
app en ices’ asks as ‘simple’ and only a small pa as
‘di icul ’, whe eas someone wi h li le expe ience would
p obably see hings di e en ly. By con as , ‘skilled’ and
‘unskilled asks’ a e a less dependen on esponden s’
pe spec i e. ‘Unskilled asks’ a e ac i i ies ha can also
be ca ied ou by unskilled pe sons, while ‘skilled asks’,
whe he simple o di icul , a e ypically pe o med by
quali ied skilled wo ke s. These e ms also align wi h he
economic concep ha he con ibu ion o app en ices
co esponds o he cos s a company would ace i hese
asks we e pe o med by o he ca ego ies o employees.
Simpli ying he language and e minology educes e-
sponden s’ cogni i e load. This in u n should lessen
measu emen e o s – meaning he gap be ween wha
happens in p ac ice and wha is epo ed in he ques-
ionnai e – and imp o e da a quali y. The downside is
ha educing measu emen e o ine i ably comes a
he expense o s ic compa abili y wi h ea lie ques-
ion wo ding. As a esul , di e ences in a e age alues
be ween su eys may no e lec eal changes in com-
panies, bu a he he ou come o lowe measu emen
e o . E en so, his ade-o is conside ed wo hwhile,
as i yields a mo e accu a e pic u e o companies’ cos s
and bene i s.
G ea e ques ionnai e cus omisa ion
By mo ing ully o an online o ma , he ques ionnai e
can now be ailo ed mo e closely o each company and
adap ed in eal ime.
Ce ain ques ions a e asked only o companies o occu-
pa ions wi h speci ic cha ac e is ics, he eby educing
eading e o and he po en ial o misunde s andings.
Fo example, ques ions on he basic aining yea and
seasonal block cou ses a e only asked o occupa ions in
which such cou ses ac ually exis . O he ques ions (e.g.
on ‘ ull- ime aine s’) a e only displayed once a com-
pany exceeds a minimum numbe o app en ices. In o -
ma ion al eady p o ided is some imes epea ed la e in
he ques ionnai e o emind esponden s and help hem
gi e consis en answe s.
Adjus men s ha we e in oduced in 2016/17 and p o ed
e ec i e ha e been e ained in he new ques ionnai e.
Fo example, pa icipan s can co ec conspicuous al-
ues o con i m hem and lea e explana o y commen s in
eal ime. Companies ha we e mis akenly con ac ed as
non- aining companies a e edi ec ed o he ques ion-
nai e o aining companies (and ice e sa).
All o hese adjus men s ha e also been ca ied o e o
he non- aining company ques ionnai e, whe e e he
same in o ma ion is collec ed.

Company pe spec i es on cos s, bene i s and aining quali y 2025 15
3.2 P e es
The ques ionnai e was ialled in a se ies o cogni i e
p e es s conduc ed on si e in selec ed companies and
e ined s ep by s ep. Pa icipan s comple ed i on hei
compu e a wo k while a membe o he su ey eam
obse ed. They we e asked o oice hei hough s as
hey wen along, and bo h sides had he oppo uni y o
ask ques ions.
The es s esul ed in he ollowing main adjus men s:
– Less is mo e: many addi ional explana ions we e no
ead and led o unce ain y o misunde s andings.
Some o hese ex s we e he e o e emo ed, while
o he s we e mo ed o oo no es a he bo om o he
ele an ques ion ins ead o appea ing in he main
ex .
– Abili y o ‘pa k’ ce ain ques ions: in some cases, e-
sponden s did no ha e he equi ed in o ma ion o
hand and needed o consul colleagues. Fo hese
ques ions, an op ion was in oduced o pa k hem,
con inue wi h he es o he ques ionnai e, and e-
u n o hem la e .
– Some pa icipan s we e unable o answe he ques-
ions on skilled wo ke ec ui men because hey had
no insigh in o his p ocess. This pa o he ques ion-
nai e can now be o wa ded o ano he pe son ia an
inpu mask.
– Fu he linguis ic adjus men s.
The ou comes o he e ision and he p e es we e dis-
cussed in hal -day expe wo kshops wi h Ha ald P ei e ,
Gud un Schöne eld and Felix Wenzelmann om BIBB, as
well as Samuel Mühlemann om Ludwig Maximilian Uni-
e si y o Munich, and we e hen submi ed o he con-
ac ing au ho i y.
Once he e ision had been app o ed, M.I.S. T end c ea -
ed an online e sion o he ques ionnai e, which he p o-
jec eam hen es ed ex ensi ely in e ms o bo h unc-
ionali y and con en . A he same ime, he ex s we e
ansla ed in o F ench and I alian and inco po a ed in o
he online ques ionnai e by M.I.S. T end. These e sions
oo we e es ed in a inal s ep by he p ojec eam.
3.3 Sampling
Essen ially, sepa a e s a i ied andom samples we e
d awn o aining and non- aining companies, and he
companies selec ed we e in i ed by pos o ake pa in
he su ey (see Chap e 3.4). The sampling app oach ol-
lowed he p ocedu e used in he p e ious su ey (Geh e
e al., 2019; Kuhn & Schwe i, 2019). The key pa ame e s
o he sampling om he wo sub-popula ions a e ou -
lined below, wi h u he de ails on he p ocedu e p o-
ided in he accompanying me hods epo (Kuhn e al.,
2025).
De ining he a ge popula ion
The i s s ep was o apply a small se o common e-
s ic ions o bo h sub-popula ions o companies. Speci -
ically, businesses om he p ima y sec o (ag icul u e),
p i a e households and ex a e i o ial o ganisa ions o
bodies we e excluded om he a ge popula ion. In ad-
di ion, he a ge popula ion was limi ed o companies
wi h a leas wo employees. The subsequen p ocedu e
and he da a basis o d awing he sample di e be ween
he wo sub-popula ions and a e, he e o e, discussed
sepa a ely below.
T aining companies
The sample o aining companies was d awn om an
exis ing da ase linking he Voca ional Educa ion and
T aining S a is ics (SBG) wi h he Business and En e -
p ise Regis e (BUR). Company in o ma ion om he
BUR was combined wi h he popula ion o all ac i e ap-
p en iceship con ac s in a gi en yea . Fo he pu pose o
d awing he sample, he Fede al S a is ical O ice (FSO)
g an ed he p ojec eam access o an anonymised e -
sion o his da a.
The linked da ase – e e ed o he ea e as SBG-BUR –
con ained a ound 186,200 dual app en iceship con ac s
o he 2022/23 aining yea . A e applying he es ic-
ions on he popula ion (e.g. excluding ag icul u al com-
panies and companies wi h ewe han wo employees),
he ele an numbe o app en iceship con ac s was e-
duced o a ound 176,200.
16 Company pe spec i es on cos s, bene i s and aining quali y 2025
I is impo an o no e ha 'company' always e e s o
he local uni o an en e p ise, since he aining o ap-
p en ices o en akes place locally, i.e. in he local uni
(b anch).
The speci ic linkage o SBG and BUR da a has he majo
ad an age o add essing wo s uc u al ea u es in he
da a ha would o he wise make sampling much mo e
complica ed. Fi s ly, only a ela i ely small p opo ion
o all companies in Swi ze land ain app en ices in any
gi en app en iceship yea (FSO, 2024). A he same ime,
he e a e la ge di e ences in how equen ly he a ious
app en iceship occupa ions a e o e ed (Kuhn, 2024).
This c ea es a challenge, since a simple andom d aw
would mainly esul in non- aining companies and he
mo e common occupa ions being included in he sample.
In he SBG–BUR da a, howe e , aining companies can
al eady be iden i ied be o e he ac ual sampling be-
gins. This makes i possible o d aw a sample in which
aining companies a e o e - ep esen ed compa ed
wi h non- aining companies. Secondly, he da a also
shows– again be o e sampling – which companies o -
e which app en iceship occupa ions. This in o ma ion
makes i possible o iden i y companies ha ain in less
common occupa ions and assign hem a highe p oba-
bili y o selec ion. This in u n ensu es ha less common
occupa ions a e also ep esen ed in he su ey in su i-
cien numbe s.
In a p epa a o y s ep, each company ha ained in
mo e han one app en iceship occupa ion was assigned
a single occupa ion. This was necessa y because compa-
nies in i ed o ake pa in he su ey we e gene ally only
able o p o ide in o ma ion o one. The assignmen o
a unique occupa ion was ca ied ou using p obabili ies,
based on he equency o he app en iceships o e ed
wi hin he company. In o he wo ds, he mo e common
an app en iceship occupa ion was wi hin a company,
he mo e likely i was o be chosen; less common occu-
pa ions we e selec ed wi h p opo iona ely lowe p ob-
abili y.
The nex s ep was o d aw a s a i ied andom sample by
app en iceship occupa ion and company ype, meaning
samples we e aken om g oups o companies (s a a).
This had he ad an age o allowing ce ain occupa ions
o be delibe a ely o e - ep esen ed in he sample. In his
s ep, an a emp was made o mee se e al sampling ob-
jec i es a once. Fi s ly, sepa a e e alua ions we e o be
possible o a leas 25 di e en app en iceship occupa-
ions, including i e FVETC p og ammes, which equi ed
a su icien numbe o companies p o iding in o ma ion
o each o hese occupa ions. Secondly, companies o -
e ing app en iceships wi h a longe aining du a ion
we e gi en a highe p obabili y o selec ion, since in o -
ma ion mus be a ailable o all app en iceship yea s.
Thi dly, s a i ica ion was used o gi e mo e weigh o
companies wi h a la ge numbe o app en iceship con-
ac s, in o de o co e as many ac i e con ac s as pos-
sible h ough he su ey. Finally, companies belonging o
en e p ises wi h a e y la ge numbe o es ablishmen s
(e.g. majo e ail chains wi h many b anches) we e un-
de - ep esen ed in he sample, as hey a e o he wise un-
likely o pa icipa e.
Based on he p e ious su ey, we assumed a esponse
a e o a ound 27.4 pe cen o aining companies
(Geh e e al., 2019; Kuhn & Schwe i, 2019). To ensu e
a ne sample o a leas 4,500 aining companies, he
g oss sample was he e o e se a 17,400 companies,
wi h an addi ional sa e y ma gin. This co esponds o
a ound 32 pe cen o all aining companies in he ele-
an ini ial popula ion.
Non- aining companies
A sepa a e sample was also d awn om he popula ion
o non- aining companies. Unlike he aining compa-
ny sample, his d aw was ca ied ou by he FSO. The
i s s ep was o ma k he aining companies in he BUR
using he SBG–BUR linkage. The linked companies ep-
esen ed he popula ion o aining companies, while
he complemen a y se cons i u ed he popula ion o
non- aining companies. F om his popula ion, a s a -
i ied andom sample was hen d awn by company size
class, wi h la ge companies gi en a highe p obabili y
o selec ion.
Based on he p e ious su ey, he g oss sample o
non- aining companies was se a 12,600, wi h he aim
o achie ing a ne sample o a leas 3,500 companies.
In he 2016/17 app en iceship yea , he esponse a e
Company pe spec i es on cos s, bene i s and aining quali y 2025 17
among non- aining companies was a ound 28.8 pe
cen ; he e oo, an addi ional ese e o companies was
added. The g oss sample size co esponds o a ound
4.6pe cen o all non- aining companies in he a ge
popula ion.
T aining a e in he a ge popula ion
Using he popula ion da ase s om which he wo sub-
samples we e d awn, i is also possible o es ima e he
agg ega ed aining a e – ha is, he pe cen age o com-
panies ha ained a leas one app en ice du ing he
2022/23 app en iceship yea .
In he de ined a ge popula ion, a o al o 53,828 ain-
ing companies and 273,481 non- aining companies
we e iden i ied, gi ing 327,309 companies in o al. This
esul s in an agg ega ed aining a e o a ound 16.5 pe
cen o he 2022/23 app en iceship yea . In o he wo ds,
a ound one in six companies in Swi ze land ained a
leas one app en ice du ing ha yea .
In he g oss sample as a whole, aining companies a e
signi ican ly o e - ep esen ed, accoun ing o 58 pe
cen o he companies con ac ed. This e lec s he cen-
al aim o he s udy, which is o p o ide in o ma ion p i-
ma ily on he cos s and p oduc i e ou pu s o app en-
ice aining om he pe spec i e o aining companies.
3.4 Field phase
To ensu e he su ey co e ed an app en iceship yea no
longe signi ican ly a ec ed by COVID, he ield phase
was pos poned om 2022 o 2023. Companies in he
g oss samples we e in i ed o ake pa om he i s
week o May onwa ds. Each company ecei ed an in i a-
ion le e in he ele an na ional language – o in Ge -
man o companies in he Romansh-speaking a ea – sen
o he business add ess eco ded in he Business and
En e p ise Regis e . The le e included he login de ails,
basic in o ma ion on he su ey and he suppo se ice
con ac de ails, oge he wi h a sho in o ma ion b o-
chu e.
The mailing was suppo ed by a ee suppo se ice,
bo h by elephone and email, o he companies. This
se ice was a ailable Monday o F iday om 8 a.m. o
6p.m. in h ee languages. Companies could use i o ask
ques ions abou he su ey in gene al and abou he
ques ionnai e in pa icula .
An FAQ websi e was also main ained and upda ed on an
ongoing basis. Bo h he ho line and he email con ac
op ion we e used ex ensi ely, pa icula ly a e in i a-
ion and eminde le e s we e dispa ched. As in ea lie
su eys, esponse a es ose sha ply a e he wo e-
minde s we e sen ou . The en i e ield phase p oceeded
smoo hly wi hou any no able echnical p oblems.
Di ec con ac s and aining cen es
Unlike in p e ious su eys (S uple & Wol e , 2012;
Geh e e al., 2019), la ge companies and he main ain-
ing cen es and associa ions we e no ea ed sepa a ely
a he in i a ion s age. A e he end o he p ima y ield
phase, hese o ganisa ions we e con ac ed la e whe e
necessa y. Depending on he alloca ion o esponsi-
bili ies wi hin he o ganisa ion and a e a p elimina y
phone call, his ook he o m o a mul i-s age w i en
su ey.
As shown in Table 4 in Chap e 3.6, la ge companies
a e well ep esen ed in he ne sample and accoun o
a highe p opo ion han in he p e ious su ey (see
Geh e e al., 2019, p. 22, Table 4).
3.5 Responses in he wo subsamples
In he end, only a po ion o he companies con ac ed
could be included in he analysis. The main easons we e
incomple e o inco ec add esses, companies ha had
al eady ceased ading by he ime o he su ey, a lack
o willingness o pa icipa e in he su ey, and oo many
missing o unclea answe s despi e companies’ ac i e
pa icipa ion.
18 Company pe spec i es on cos s, bene i s and aining quali y 2025
Table 3 documen s he esponses in he wo subsamples
o ou su ey. The g oss sample comp ised 17,400 ain-
ing companies and 12,600 non- aining companies. As
in he p e ious su ey, he add ess da a om he SBG–
BUR p o ed ela i ely eliable, since only a small num-
be o companies could no be con ac ed by pos due
o inco ec add esses. O he aining companies in he
g oss sample, in i a ions could no be deli e ed o only
301 companies, co esponding o a ound 1.7 pe cen .
Among non- aining companies (d awn di ec ly om he
BUR), on he o he hand, a signi ican ly la ge sha e o
a ound 5.4 pe cen could no be eached by pos .
Mo e signi ican a e d op-ou s due o e usal o pa ic-
ipa e and cases whe e companies le a compa a i ely
la ge o c ucial pa o he ques ionnai e blank o illed
i in inco ec ly, making hese esponses unusable o
da a analysis. A ound 61 pe cen o aining companies
a leas opened he ques ionnai e, and 39.1 pe cen
o esponses we e usable. Among non- aining compa-
nies, a ound 35 pe cen opened he ques ionnai e and
a ound 31 pe cen o esponses we e usable. In addi-
ion, 999 non- aining companies – a ound 27 pe cen
o he usable esponses – epo ed ha hey could no
ain app en ices because hey we e no economically
independen en i ies o we e e y small, highly special-
ised b anch o ices. These companies a e he e o e no
included in he subsequen analyses.
In absolu e e ms, his lea es in o ma ion om
6,686 aining companies and 2,656 non- aining compa-
nies o he analyses ha ollow.
3.6 Composi ion o he wo ne
samples
In he nex s ep, he composi ion o he wo ne samples
is desc ibed wi h e e ence o se e al key cha ac e is-
ics. As he ollowing wo ables ocus on unweigh ed
companies, hey also e lec he pa ly disp opo iona e
s a i ica ion o he sampling plan. Fo example, he e
a e ma ked di e ences be ween language egions in he
impo ance o ini ial oca ional educa ion and aining,
esul ing in a highe sha e o aining companies in Ge -
man-speaking Swi ze land (Aepli e al., 2021a; Kuhn e
al., 2022). Since aining companies we e o e - ep esen -
ed in he g oss sample, his also leads o a disp opo -
iona ely la ge sha e o companies om Ge man-speak-
ing Swi ze land.
T aining companies
Non- aining
companies
Numbe Pe cen age Numbe Pe cen age
Numbe o in i a ions 17,400 100.00 12,600 100.00
Undeli e able in i a ions 301 1.73 679 5.39
Numbe o in i a ions (minus pos al e u ns)
17,099 100.00 11,921 100.00
Ques ionnai e accessed 10,465 61.20 4,155 34.85
Ques ionnai e comple ed 7,379 43.15 3,655 30.66
Usable esponses 6,686 39.10 3,655 30.66
Too specialised o aining   999 27.27
Table 3: Responses in he wo g oss samples
Company pe spec i es on cos s, bene i s and aining quali y 2025 19
Table 4 i s ocuses on he ne sample o 6,686 aining
companies. A ound 29.6 pe cen o hese companies
ha e up o 9 employees, 40.4 pe cen be ween 10 and
49, 10.8 pe cen be ween 50 and 99, and he emain-
ing 19.1 pe cen ha e 100 o mo e. B oken down by
sec o , abou 14 pe cen o companies belong o he
cons uc ion and 14 pe cen o he manu ac u ing sec-
o , a majo i y o 46 pe cen a e in he se ices sec o
and a ound 26 pe cen in public adminis a ion and
non-p o i o ganisa ions. Finally, he as majo i y o
aining companies – a ound 76.6 pe cen – a e loca -
ed in Ge man-speaking Swi ze land (including egions
whe e Romansh is he p edominan language), 19.7 pe
cen a e om F ench-speaking Swi ze land, and 3.8 pe
cen om Ticino and o he p edominan ly I alian-speak-
ing egions.
The lowe pa o he able shows how many ques ion-
nai es we e comple ed o each app en iceship occupa-
ion. Fo occupa ions wi h a leas 40 obse a ions pe
app en iceship yea (14 o FVETC app en iceships) om
a minimum o 65 companies (25 o FVETC), sepa a e
e alua ions a e a ailable on he SFUVET websi e (see
h ps://s u e .swiss/cos -bene i ). Ac oss all app en ice-
ships, abou 7 pe cen o companies p o ided in o ma-
ion on a wo-yea FVETC app en iceship, 61 pe cen on
a h ee-yea FVETD app en iceship, and 32 pe cen on a
ou -yea FVETD app en iceship. The dis ibu ion ac oss
speci ic app en iceship occupa ions u he shows ha
da a om a leas 30 di e en aining companies is
a ailable o a o al o 30 occupa ions (including i e
FVETC app en iceships). The smalles g oup comp is-
es 30 companies (Logis ics Assis an , FVETC), while he
la ges comp ises 750 companies (Comme cial Employ-
ee, FVETD).
Cha ac e is ic Ca ego y Numbe Pe cen age
Company size class 1–9 employees 1,982 29.64
10–49 employees 2,704 40.44
50–99 employees 723 10.81
100+ employees 1,277 19.10
Sec o Manu ac u ing 928 13.88
Cons uc ion 935 13.98
Se ices 3,054 45.68
Adminis a ion, non-p o i o ganisa ions 1,769 26.46
Language egion Ge man 5,121 76.59
F ench 1,314 19.65
I alian 251 3.75
Du a ion 2 yea s 456 6.82
3 yea s 4,082 61.05
4 yea s 2,148 32.13
Table 4: Composi ion o he ne sample o aining companies

20 Company pe spec i es on cos s, bene i s and aining quali y 2025
Cha ac e is ic Ca ego y Numbe Pe cen age
App en iceship

Comme cial Employee, FVETD 750 11.22
D augh sman/D augh swoman, FVETD 444 6.64
Re ail P o essional, FVETD 422 6.31
Ce i ied Social Ca e Wo ke , FVETD 397 5.94
Joine , FVETD 202 3.02
Au omo i e Technician, FVETD 193 2.89
Den al Assis an , FVETD 185 2.77
Heal hca e Assis an , FVETD 178 2.66
Medical Sec e a y and Assis an , FVETD 169 2.53
Elec ician, FVETD 166 2.48
Ca pen e , FVETD 140 2.09
Pha macy Assis an , FVETD 140 2.09
Che , FVETD 131 1.96
Au omo i e Mecha onics Technician, FVETD 117 1.75
Pain e , FVETD 116 1.73
Building and G ounds Cus odian, FVETD 114 1.71
Logis ics P o essional, FVETD 114 1.71
Hai d esse , FVETD 110 1.65
In o ma ion Technologis , FVETD 107 1.60
Ga dene , FVETD 104 1.56
Polymechanic, FVETD 98 1.47
Plumbe , FVETD 95 1.42
B icklaye , FVETD 93 1.39
Au oma ion Technician, FVETD 85 1.27
Building Se ices Assis an , FVETC 78 1.17
Cons uc ion Elec ician, FVETD 67 1.00
Re ail Assis an , FVETC 64 0.96
O ice Assis an , FVETC 52 0.78
Heal h and Social Ca e Wo ke , FVETC 48 0.72
Logis ics Assis an , FVETC 30 0.45
O he ou -yea FVETD 789 11.80
O he h ee-yea FVETD 704 10.53
O he FVETC 184 2.75
(6,686 companies, unweigh ed)
Company pe spec i es on cos s, bene i s and aining quali y 2025 21
In addi ion, Table 5 shows he composi ion o he ne
sample o 2,656 non- aining companies. A la ge sha e
o hese companies – 67.6 pe cen – ha e ewe han
10employees, 20.3 pe cen ha e be ween 10 and 49 em-
ployees, 6.8 pe cen ha e be ween 50 and 99 employees,
and 5.3 pe cen ha e 100 o mo e employees. By sec o ,
he majo i y – 63.5 pe cen – a e in se ices, ollowed by
public adminis a ion and non-p o i o ganisa ions wi h
19.7 pe cen , manu ac u ing wi h 8.8 pe cen , and con-
s uc ion wi h a ound 8.0 pe cen . In e ms o language
egions, 66.4 pe cen a e based in Ge man-speaking
Swi ze land, 26.4 pe cen in F ench-speaking Swi ze -
land, and 7.2 pe cen in mainly I alian-speaking egions.
3.7 Plausibili y checks
To gua an ee he quali y o he da a, he su ey espons-
es wen h ough se e al s ages o plausibili y checking.
Fi s , du ing da a en y in he b owse , esponses we e
checked agains p ede ined h esholds. I hese h esh-
olds we e exceeded o no me , esponden s we e ale -
ed o he unusual alues. They could hen co ec he
alues di ec ly in he b owse o con i m and commen
on hem. These commen s we e used in subsequen
plausibili y checks o be e unde s and anomalies o , in
some cases, o adjus alues e ospec i ely.
Compa ed wi h he p e ious su ey, he quali y o e-
sponses was gene ally be e , wi h ewe implausible an-
swe s. E en so, many cases s ill equi ed a ollow-up call.
To iden i y hese cases and he companies conce ned,
he da a was analysed once again agains h eshold
alues. I esponses con ained implausible o missing
en ies, he ele an companies we e con ac ed by ele-
phone, whe e possible, abou he a ec ed pa s o he
ques ionnai e.
Cha ac e is ic Ca ego y Numbe Pe cen age
Company size class 1–9 employees 1,796 67.62
10–49 employees 539 20.29
50–99 employees 180 6.78
100+ employees 141 5.31
Sec o Manu ac u ing 233 8.77
Cons uc ion 213 8.02
Se ices 1,687 63.52
Adminis a ion, non-p o i o ganisa ions 523 19.69
Language egion Ge man 1,764 66.42
F ench 700 26.36
I alian 192 7.23
(2,656 companies, unweigh ed)
Table 5: Composi ion o he ne sample o non- aining companies
22 Company pe spec i es on cos s, bene i s and aining quali y 2025
3.8 App oach o missing a iable
alues
Consequences o missing a iable alues
Calcula ing he cos s and bene i s o company-based
aining (see Chap e 2) equi es a company o answe
all ques ions ela ing o he a iables included in he
model. In p ac ice, howe e , many companies le one o
mo e o hese ques ions unanswe ed.
The di icul y he e is ha a missing alue o one a i-
able leads o missing alues a all subsequen le els o
calcula ion o analysis. This applies e en in he ex eme
case whe e only a single a iable equi ed o a calcula-
ion s ep is missing.
Impu a ion o missing alues
Fo his eason, missing alues o a iables equi ed in
he cos –bene i calcula ion we e impu ed using s a is-
ical me hods (Dong & Peng, 2013; Li le & Rubin, 2019).
Impu a ion means ha missing alues in indi idual a -
iables a e eplaced wi h plausible alid alues. Pu sim-
ply, he basic p inciple o impu a ion is ha a missing
en y o one company is subs i u ed wi h a alid en y
om o he , as a as possible compa able, companies.
In p ac ice, his p ocess is o en highly complex, since
missing alues may occu simul aneously in se e al
a iables wi hin he same company, and he pa e ns
o missing alues a y be ween companies. Mode n im-
pu a ion me hods ope a e i e a i ely – ha is, missing
alues a e g adually eplaced o e se e al epea ed uns
o a complex algo i hm. This makes i possible o co ec
p ac ically any pa e n o missing da a (see an Buu en,
2007, 2018).
Building on he impu a ion p ocess used in he p e ious
su ey, his ime a mul iple impu a ion me hod was cho-
sen. This means ha missing alues we e eplaced no
wi h a single alue, bu wi h 35 di e en alues. The da a
analyses we e hen essen ially ca ied ou 35 imes, and
he esul s o hese es ima es a e aged. This app oach
o e s wo ad an ages: i s , i accoun s o he addi ion-
al unce ain y caused by impu a ion in he calcula ion o
a iances and s anda d e o s; second, mul iple impu a-
ion is less in luenced by ou lie s, which can occu when
impu ing single alues. The ade-o , howe e , is ha
his p ocess signi ican ly inc eases compu ing equi e-
men s and makes subsequen analyses mo e complex.
3.9 Sample weigh ing
As explained in Chap e 3.3, he sampling p ocess delib-
e a ely selec ed a disp opo iona e numbe o aining
companies compa ed wi h non- aining companies, as
well as less equen ly chosen app en iceship occupa-
ions compa ed wi h mo e popula occupa ions. This
al e ed he composi ion o he g oss sample compa ed
wi h he o e all popula ion o companies. Sample
weigh s se e i s and o emos o co ec o his in he
analysis. They a e also used o adjus o di e ences in
ealised esponse a es ac oss he a ious s a a o he
sample.
The sample weigh s we e de e mined in wo successi e
s eps, each based on he s a a used in he sampling
design – namely app en iceship occupa ion and compa-
ny ype o aining companies, and company size class
o non- aining companies. In he i s s ep, he ini ial
o design weigh is de e mined. This is de ined as he
in e se o he p obabili y o a company om he pop-
ula ion being included in he g oss sample. This weigh
hus enables in e ence om he g oss sample o he
popula ion, compensa ing o he pa ly disp opo ion-
a e selec ion p obabili ies in he sampling p ocess. In
he second s ep, he esponse a e wi hin each s a um
is de e mined. This gene a es a second weigh ha co -
ec s o an incomple e esponse. This second weigh is
used o d aw in e ences om he ne sample o he g oss
sample, he eby balancing di e ences in esponse a es
ac oss s a a. The inal sample weigh is he p oduc o
hese wo weigh s and hus co ec s simul aneously o
bo h he disp opo iona e sampling design and he di -
e ences in esponse a es.
Fu he in o ma ion on he sample weigh s is p o ided
in he accompanying me hods epo (Kuhn, Geh e &
Schwe i, 2025).
Company pe spec i es on cos s, bene i s and aining quali y 2025 23
This chap e p esen s he esul s o he cos –bene i su -
ey. The i s s ep is o examine he g oss cos s, p oduc-
i e ou pu and ne bene i o ini ial oca ional educa-
ion and aining om a company pe spec i e. These a e
conside ed i s in agg ega e, and hen di e en ia ed by
aining du a ion ( wo-yea app en iceships leading o
an FVETC and h ee- and ou -yea app en iceships lead-
ing o an FVETD), as well as by company size, sec o and
majo egion. The chap e hen summa ises he ne ben-
e i o he mos equen app en iceship occupa ions. I
also illus a es how companies can gene a e addi ional
bene i s by employing app en ices once hei aining is
comple e. The chap e concludes by ex apola ing he
esul s o Swi ze land as a whole and compa ing hem
wi h hose o he la es Ge man cos –bene i s udy.
4.1 A e age cos s and bene i s o
ini ial oca ional educa ion and
aining
As ou lined in Chap e 2 in he discussion o he cos -ben-
e i model, companies incu g oss cos s as a esul o
aining. These consis o app en ice wages, s a cos s,
ma e ial and equipmen cos s, and addi ional cos s. In
he 2022/23 aining yea , he a e age g oss cos ac oss
all app en iceships was CHF 27,990 pe app en iceship
yea . A he same ime, companies bene i om he wo k
p o ided by app en ices, who can be assigned o ei he
unskilled o skilled asks. The mone a y alue o his
is cap u ed in he p oduc i e ou pu , which a e aged
CHF32,530 pe app en iceship yea ac oss all app en-
iceships.
Since p oduc i e ou pu exceeded g oss cos s, compa-
nies achie ed an a e age ne bene i o CHF 4,540 pe
app en iceship pe yea . This undamen al inding – ha
aining app en ices gene a es a ne bene i on a e -
age ac oss all app en iceships – con i ms he esul s o
p e ious Swiss su eys in 2016/17 (Geh e e al., 2019),
2009/10 (S uple & Wol e , 2012) and 2008/09 (Fuh e &
Schwe i, 2010).
The sligh ly highe ne bene i compa ed wi h 2016/17 is
due o a small dec ease in g oss cos s, while p oduc i e
ou pu emained cons an . Pe sonnel cos s ose sligh -
ly, bu companies on a e age es ima ed ma e ial and
equipmen cos s somewha lowe .
Al hough he o al p oduc i e ou pu has changed li le
compa ed wi h 2016/17, i s composi ion has shi ed un-
damen ally. Whe eas unskilled asks p e iously accoun -
ed o he la ge sha e, i is now skilled asks ha b ing
g ea e bene i s o companies. This change was e y
likely caused by he e ision o he su ey amewo k.
In assessing app en ices’ ac i i ies in he wo kplace, he
e minology was delibe a ely e ised and s anda dised.
As explained in Chap e 2, he ca ego ies ‘simple asks’
and ‘di icul asks’ we e no longe used, as his dis inc-
ion le oo much oom o in e p e a ion. Ins ead, he
e ms ‘unskilled asks’ and ‘skilled asks’ we e adop ed
consis en ly, bo h o which a e di ec ly ela ed o he
company’s wo k and p oduc ion p ocesses. F om a ped-
agogical pe spec i e, oo, he key ques ion is whe he
app en ices a e able o acqui e and p ac ise he skills
needed o pe o m he asks expec ed o hem as skilled
wo ke s. I is he e o e encou aging ha skilled asks
now ou weigh unskilled asks.
When in e p e ing he indings, i is impo an o bea in
mind ha he ne bene i e lec s he cos –bene i bal-
ance a he end o he app en iceship (o o a speci ic
yea o aining) and does no include ce ain bene i el-
emen s, such as he gains om employing app en ices
a e hei aining.
An a e age posi i e ne bene i does no mean, howe e ,
ha all companies achie e a ne bene i om aining
app en ices by he end o he app en iceship. Figu e 1
illus a es his by showing he dis ibu ion o ne bene-
i pe app en iceship place. App en iceships abo e he
ze o- anc h eshold ep esen cases wi h a ne bene-
i , while hose below i ep esen cases wi h ne cos s.
In ewe han 30 pe cen o app en iceships, ne cos s
a e incu ed. By con as , he majo i y – a ound 71 pe
cen – esul in a smalle o la ge ne bene i .
4 RESULTS
30 Company pe spec i es on cos s, bene i s and aining quali y 2025
in he way app en ices’ ime is dis ibu ed in he wo k-
place. While companies assessed he sha e o unp oduc-
i e ime simila ly o 2009/10 and 2016/17, he di ision
be ween unskilled and skilled asks di e s in all aining
yea s. In con as o 2016/17 and also 2008/09, compa-
nies now es ima e he sha e o skilled asks signi ican ly
highe in e e y app en iceship yea . As wi h he wo-yea
FVETC p og ammes (see Chap e 4.2), his change is like-
ly due o adjus men s in he su ey amewo k. Fo ex-
ample, he es ima e ha app en ices spend 66 pe cen
o hei ime on skilled asks in he inal yea , compa ed
wi h 45 pe cen in 2016/17, also seems ealis ic o he
h ee-yea FVETD p og ammes.
0
20
40
60
80
100
Pe cen age
App en iceship yea 1 App en iceship yea 2 App en iceship yea 3
Skilled asks Unskilled asks
O he /exe cises
4.4 Cos s and bene i s o he ou -
yea ini ial oca ional educa ion
and aining leading o an FVETD
In he 2022/23 aining yea , acco ding o FSO da a
(and he popula ion de ini ion in Chap e 3.3), almos
51,300app en ices we e ained o e he cou se o he
ou -yea IVET in 79 occupa ions leading o a Fede al VET
Diploma (FVETD). Among hese app en ices, 17 pe cen
we e women and 4.5 pe cen we e adul app en ices
(aged o e 25). Fewe han 7 pe cen o he app en ice-
ship places we e sho ened aining p og ammes.
The ollowing analysis is based on 58 di e en occu-
pa ions in 2,148 aining companies, wi h a o al o
4,773app en ices ac oss di e en yea s. In his sample,
he Fede al Voca ional Baccalau ea e (BM 1) a e was al-
mos 14pe cen .
Figu e 6: Time spen by app en ices a he company wo kplace in h ee-yea FVETD app en iceships

Company pe spec i es on cos s, bene i s and aining quali y 2025 31
The o al ne bene i ha a aining company achie es
on a e age pe app en iceship in ou -yea IVET p o-
g ammes (see Table 10) is a ound CHF 17,510 – signi i-
can ly highe han in he 2016/17 su ey. The 95 pe cen
con idence in e al (‘+/–’) is plus/minus CHF 4,060, i.e.
be ween CHF 13,450 and CHF 21,570.
0
10
20
30
40
In CHF housands
App en ice-
ship yea 1
App en ice-
ship yea 2
App en ice-
ship yea 3
App en ice-
ship yea 4
App en ice wage cos s Pe sonnel cos s
Ma e ial/equipmen cos s Addi ional cos s
Figu e 7: B eakdown o g oss cos s in ou -yea FVETD app en iceships
App en iceship yea 1234To al
G oss cos s 26,300 27,090 29,850 32,670 115,910
+/- 700 670 660 720 1,960
P oduc i e ou pu 21,650 27,160 37,600 46,990 133,410
+/- 950 940 1,130 1,390 3,590
Ne bene i -4,650 70 7,750 14,330 17,510
+/- 1,150 1,120 1,320 1,600 4,060
Table 10: G oss cos s, p oduc i e ou pu and ne bene i pe ou -yea FVETD app en iceship (in CHF)
32 Company pe spec i es on cos s, bene i s and aining quali y 2025
G oss cos s
Fo g oss cos s and hei componen s (see Figu e 7),
he ypical inc ease in cos s pe app en iceship yea is
e iden as he aining du a ion ises. The eason lies
in ising app en ice wages (see also Table 11), while
pe sonnel cos s decline sligh ly. Companies es ima e
he a e age weekly ime spen on aining app en ices
a 6.6hou s in he i s yea , 5.6 hou s in he second,
4.9hou s in he hi d and 4.3 hou s in he ou h. Ma e i-
al and equipmen cos s and addi ional cos s a e on a e -
age highe han in wo- o h ee-yea IVET p og ammes,
mainly because many o he ou -yea IVET p og ammes
a e in echnical occupa ions.
When looking a he mon hly median wage o app en ic-
es (see Table 11), i is s iking ha i is lowe in he i s
wo yea s han in wo-yea IVET p og ammes and lowe
in he i s h ee yea s han in h ee-yea p og ammes.
This pa e n, also obse ed in ea lie su eys, is like-
ly ela ed o he ac ha app en ices in he mo e de-
manding ou -yea p og ammes ake longe o each a
p oduc i i y le el compa able wi h ha o skilled wo k-
e s. In he inal yea , howe e , wages ise o CHF 1,350.
As in he sho e p og ammes, wages show some a i-
a ion: in he inal yea , 25 pe cen o app en ices ea n
CHF1,450 o mo e pe mon h, while ano he 25 pe cen
ea n CHF1,200 o less.
App en iceship yea 1234
25 h pe cen ile 550 750 950 1,200
Median wage 650 800 1,050 1,350
75 h pe cen ile 700 930 1,200 1,450
Table 11: Mon hly g oss wage o app en ices in ou -yea FVETD
app en iceships (in CHF)
0
20
40
60
80
100
Pe cen age
Skilled asks Unskilled asks O he /exe cises
App en ice-
ship yea 1
App en ice-
ship yea 2
App en ice-
ship yea 3
App en ice-
ship yea 4
Figu e 8: Time spen by app en ices a he company wo kplace in ou -yea FVETD app en iceships
Company pe spec i es on cos s, bene i s and aining quali y 2025 33
P oduc i e ou pu
Gi en ha ou -yea FVETD p og ammes end o co -
e mo e demanding occupa ions, i is ha dly su p ising
ha i s -yea app en ices spend abou a qua e o hei
ime in he company on unp oduc i e ac i i ies such as
aining exe cises (Figu e 8). This sha e dec eases as
aining p og esses. In he inal yea , app en ices spend
on a e age mo e han 87 pe cen o hei ime in he
company on p oduc i e ac i i ies, mos o which in ol e
asks ha would o he wise be pe o med by quali ied
skilled wo ke s. The sha e o such skilled asks is al eady
44 pe cen in he i s yea and ises o 74 pe cen in
he inal yea . App en ices’ pe o mance le el in hese
skilled asks inc eases om 31 pe cen o 46 pe cen ,
63pe cen and inally 78 pe cen in he las yea .
Compa ison wi h he p e ious su ey
Compa ed wi h he 2016/17 aining yea , ne bene i
ises mo e s eeply o e he ou yea s o aining and is
almos CHF 10,000 highe in o al. This signi ican di e -
ence is explained by lowe g oss cos s combined wi h
highe p oduc i e ou pu . As wi h he wo- and h ee-
yea p og ammes, he lowe g oss cos s a e la gely due
o educed ma e ial and equipmen cos s and sligh ly
lowe addi ional cos s. I is possible ha hese cos s
ha e, in ac , allen, bu i is also possible ha hey ap-
pea lowe because he e ised ques ionnai e desc ibed
hem in clea e and simple e ms.
Ano he eason o he inc ease in ne bene i since
2016/17 lies in he changed composi ion o ou -yea ap-
p en iceships. The new ou -yea aining p og ammes
(be o e: h ee yea s) in he occupa ions o Ca pen e
FVETD, Plumbe FVETD, Hea ing Ins alle FVETD and
Shee -Me al Wo ke FVETD, which ha e only been in o-
duced a e he las su ey, aise he a e age ne bene i .
Fo he h ee building-se ices occupa ions, howe e ,
his applies only o he i s h ee yea s o aining, since
he newly in oduced ou h yea ook place o he i s
ime in he yea a e he su ey.
As wi h he wo- and h ee-yea p og ammes, he b eak-
down o p oduc i e ou pu in ou -yea app en iceships
has shi ed s ongly owa ds skilled asks. This e lec s
he e ised way in which wo kplace ac i i ies we e as-
sessed, and is linked o clea e ques ion wo ding in he
ques ionnai e.
So a , he esul s ha e been p esen ed by aining du a-
ion, whe eas he ollowing chap e s desc ibe he ind-
ings by company size, sec o and majo egion. The pe -
spec i e changes he e: ins ead o analysing indi idual
yea s o aining ( i s yea , second yea , e c.), he cos s
and bene i s o an a e age yea o aining a e examined.
This allows wo-, h ee- and ou -yea app en iceships o
be conside ed as a whole.
4.5 Resul s by company size
In he 2022/23 aining yea , acco ding o FSO da a om
he SBG-BUR, mos app en ices we e ained in small
companies wi h 10 o 49 employees (38 pe cen o ap-
p en iceships) and in la ge companies wi h 100 o mo e
employees (27 pe cen ). Ve y small companies wi h up
o 9 employees also ained a subs an ial sha e a o e
20 pe cen , while medium-sized companies wi h 50 o
99 employees accoun ed o almos 13 pe cen . Only a
small numbe o aining companies in he BUR did no
epo hei company size.
Figu e 9 shows g oss cos s, p oduc i e ou pu and ne
bene i by company size. The esul s show ha , on a e -
age, companies o all sizes achie e a ne bene i . In con-
as o he 2009/10 and 2016/17 su eys, medium-sized
and la ge companies no longe eco d a ma ked d op in
ne bene i . P oduc i e ou pu is now also e y simila
ac oss he di e en company sizes.
34 Company pe spec i es on cos s, bene i s and aining quali y 2025
4.6 Resul s by sec o
In he 2022/23 aining yea , a ound 63 pe cen o ap-
p en iceships we e in he se ice sec o , jus unde
16pe cen in manu ac u ing, almos 14 pe cen in con-
s uc ion and nea ly 8 pe cen in public adminis a ion
o non-p o i o ganisa ions (SBG-BUR). As shown in Fig-
u e 10 – and in con as o he esul s by company size–
he e a e la ge di e ences be ween sec o s in a e age
ne bene i . All sec o s achie e a clea ne bene i , bu
cons uc ion s ands ou wi h a signi ican ly highe ig-
u e. Al hough g oss cos s in cons uc ion a e sligh ly
highe han in public adminis a ion and non-p o i o -
ganisa ions and also in se ices, i s p oduc i e ou pu is
clea ly g ea e han in he o he sec o s.
The highe p oduc i e ou pu in cons uc ion is mainly
d i en by he g ea e numbe o days app en ices spend
in he company. On a e age, app en ices in cons uc-
ion spend 7 o 13 mo e days pe yea in he wo kplace.
They also spend less ime on unp oduc i e ac i i ies
such as aining exe cises. This e ec is ein o ced by he
skilled-wo ke wages used o alue p oduc i e skilled
asks, which a e highe in cons uc ion han in public
adminis a ion/non-p o i o ganisa ions and in he se -
ice sec o .
Highe skilled-wo ke and unskilled-wo ke wages in he
calcula ion o p oduc i e ou pu also explain why ap-
p en iceships in manu ac u ing show highe p oduc i e
ou pu han hose in public adminis a ion/non-p o i s
and in se ices.
0
10
20
30
40
<10 10–49 50–99 >99
In CHF housands
Numbe o employees wi hin he company
G oss cos s P oduc i e ou pu Ne bene i
Figu e 9: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea , by company size
Company pe spec i es on cos s, bene i s and aining quali y 2025 35
4.7 Resul s by majo egion
In he 2022/23 aining yea , app en iceship places we e
sp ead ac oss all majo egions. Espace Mi elland ac-
coun ed o mo e han 22 pe cen o all con ac s, Zu ich
19 pe cen , Eas e n Swi ze land 16 pe cen , he Lake
Gene a egion 14 pe cen , Cen al Swi ze land 11 pe
cen and Ticino jus o e 3 pe cen .
Compa ed wi h he sec o esul s, he di e ences be-
ween he a ious majo egions a e smalle (see Figu e
11). Once again, all g oups show a posi i e ne bene i ,
bu he di e ence is s a is ically signi ican only be-
ween he Lake Gene a egion and Ticino. The highe
ne bene i in he Lake Gene a egion is explained by
highe p oduc i e ou pu , especially om skilled asks,
despi e sligh ly highe g oss cos s. P oduc i e ou pu in
he Lake Gene a egion is also highe han in Espace Mi -
elland and in No he n and Eas e n Swi ze land, owing
o sligh ly highe skilled-wo ke wages, which a e used
o calcula e he alue o skilled asks.
As in 2016/17, he cos –bene i balance in Ticino is
ma ked by lowe p oduc i e ou pu . This e lec s he
impac o lowe wages o skilled and unskilled wo ke s
on he calcula ion o ou pu : when wages a e lowe , he
alue o he wo k pe o med by app en ices is educed.
Compa ed wi h 2016/17, i is also no able ha p oduc-
i e ou pu in he Zu ich egion has con e ged wi h ha
o he o he egions (excep Ticino), leading o a mo e
a ou able cos –bene i balance.
Figu e 10: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea , by sec o
In CHF housands
Se ices Manu ac u ing Cons uc ion Adminis a i e
and non-p o i
G oss cos s P oduc i e ou pu Ne bene i
0
10
20
30
40

36 Company pe spec i es on cos s, bene i s and aining quali y 2025
4.8 Resul s by occupa ion
The ollowing igu es show he a e age ne bene i o
he mos common FVETC occupa ions (Figu e 12) and
FVETD occupa ions (Figu e 13 and Figu e 14). De ailed
analyses o hese occupa ions a e a ailable online a he
ollowing link:
h ps://s u e .swiss/cos -bene i
The occupa ional i les used co espond o hose in
o ce in he 2022/23 aining yea . Fo example, he Ge -
man i le ‘Fach au/Fachmann Apo heke EFZ’ also co -
e s app en ices in he occupa ion p e iously known as
‘Pha ma-Assis en /-in EFZ’ who we e in hei second o
hi d yea o aining a ha ime.
In he wo-yea IVET p og ammes, Logis ics Assis an
FVETC and Building Se ices Assis an FVETC can now
be epo ed sepa a ely, while he case numbe s o Au-
omo i e Assis an FVETC and Ki chen Employee FVETC
a e no longe su icien o indi idual analysis. As wi h
o he FVETC occupa ions wi h oo ew cases ( ewe han
25 pa icipa ing companies o ewe han 14 obse a-
ions pe app en iceship yea ), hese a e now g ouped
unde ‘O he FVETC’. O e all, he a e age ne bene i o
FVETC occupa ions is highe han in 2016/17 (see Chap-
e 4.2). Wha s ands ou is he ela i ely high a e age
ne bene i in he Logis ics Assis an FVETC occupa ion,
d i en by compa a i ely high p oduc i e ou pu along-
side sligh ly below-a e age g oss cos s. By con as , he
d op in ne bene i o O ice Assis an FVETC compa ed
wi h 2016/17 is s iking, as i e lec s a ma ked decline in
p oduc i e ou pu .
Figu e 11: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea , by majo egion
0
10
20
30
40
In CHF housands
Lake
Gene a
Region
Espace
Mi elland
No h-
wes e n
Swi ze land
Zu ich Eas e n
Swi ze -
land
Cen al
Swi ze -
land
Ticino
G oss cos s P oduc i e ou pu Ne bene i
Company pe spec i es on cos s, bene i s and aining quali y 2025 37
Figu e 12: A e age ne bene i o he mos common FVETC app en iceship occupa ions
010 20 30
Ne bene i in CHF housands
Heal h and Social Ca e Wo ke , FVETC
Building Se ices Assis an , FVETC
O ice Assis an , FVETC
O he FVETC
Re ail Assis an , FVETC
Logis ics Assis an , FVETC
Figu e 13: A e age ne bene i o he mos common h ee-yea FVETD app en iceship occupa ions
Ne bene i in CHF housands
B icklaye , FVETD
Che , FVETD
Heal hca e Assis an , FVETD
Plumbe , FVETD
O he h ee-yea FVETD
Hai d esse , FVETD
Au omo i e Technician, FVETD
Re ail P o essional, FVETD
Comme cial Employee, FVETD
Cons uc ion Elec ician, FVETD
Medical Sec e a y and Assis an , FVETD
Pain e , FVETD
Ce i ied Social Ca e Wo ke , FVETD
Building and G ounds Cus odian, FVETD
Pha macy Assis an , FVETD
Logis ics P o essional, FVETD
Ga dene , FVETD
Den al Assis an , FVETD
010 20 30 40
38 Company pe spec i es on cos s, bene i s and aining quali y 2025
In he h ee-yea FVETD p og ammes, ewe app en ice-
ship occupa ions a e epo ed han in 2016/17 because
case numbe s in some occupa ions we e oo low. Occu-
pa ions no longe epo ed sepa a ely include Hospi-
ali y Se ices P o essional FVETD, Res au an Se ice
Employee FVETD and Specialis in Ho el Housekeeping
FVETD. These and o he h ee-yea FVETD occupa ions
wi h low case numbe s ( ewe han 65 pa icipa ing com-
panies o ewe han 40 obse a ions pe app en iceship
yea ) ha e been g ouped unde ‘O he h ee-yea FVETD’.
Wi h he mos ecen e o m o he VET o dinance, he
Plumbe FVETD p og amme was ex ended om h ee
o ou yea s. Since he e we e no app en ices in hei
ou h yea in 2022/23, his occupa ion is s ill shown as
a h ee-yea p og amme in all analyses. The ac ual ne
bene i may change once he ou h yea is included, as
he inal app en iceship yea usually inc eases ne bene-
i . By con as , he Ca pen e FVETD occupa ion was ob-
se ed ac oss all aining yea s and is now lis ed unde
he ou -yea app en iceships.
Compa ed wi h 2016/17, ne bene i has also isen in
he h ee-yea p og ammes. None o he occupa ions
analysed now eco ds a signi ican ly nega i e ne ben-
e i . Mos show a modes o subs an ial ne bene i . The
Plumbe app en iceship appea s a i s glance o con-
adic his posi i e end, showing a ma ked decline in
ne bene i . Howe e , since only he i s h ee app en-
iceship yea s we e obse ed he e, i is no possible o
es ima e how much he ne bene i will inc ease in he
inal yea o app en iceship.
The end owa ds highe ne bene i is also e iden in
he ou -yea FVETD p og ammes. This is pa icula ly
clea in he Polymechanic FVETD app en iceship, which
in ea lie su eys (2016/17 and 2009/10) was s ill shown
as an occupa ion wi h signi ican ne cos s. G oss cos s
s ill ou weigh p oduc i e ou pu in his case, bu o a
lesse ex en han be o e. A simila pa e n can be seen
in he In o ma ion Technologis FVETD and Au omo-
i e Mecha onics Technician FVETD app en iceships.
Fo he la e , his is he i s ime since he ini ial su -
ey in 2000/01 (see Schwe i e al., 2003; Mühlemann e
al., 2007b; S uple & Wol e , 2012; Geh e e al., 2019)
ha a posi i e ne bene i has been es ima ed. Fou -
yea FVETD app en iceship occupa ions wi h low case
numbe s ( ewe han 65 pa icipa ing companies o ew-
e han 40 obse a ions pe app en iceship yea ) ha e
been g ouped unde ‘O he ou -yea FVETD’.
Figu e 14: A e age ne bene i o he mos common ou -yea FVETD app en iceship occupa ions
Ne bene i in CHF housands
Polymechanic, FVETD
Au oma ion Technician, FVETD
In o ma ion Technologis , FVETD
O he ou -yea FVETD
Au omo i e Mecha onics Tech., FVETD
D augh sman/-woman, FVETD
Joine , FVETD
Ca pen e , FVETD
Elec ician, FVETD
-20 020 40 60
Company pe spec i es on cos s, bene i s and aining quali y 2025 39
4.9 Findings on he bene i s o
con inued employmen
The p e ious chap e s ocused on he cos s and bene-
i s ha a ise du ing aining. As explained in Chap e
2, companies can de i e addi ional bene i s by keeping
app en ices on a e hei aining. This is because hey
a oid some o he ec ui men and induc ion cos s ha
would o he wise a ise when hi ing skilled wo ke s om
he ex e nal labou ma ke . In his way, aining com-
panies achie e an a e age addi ional bene i o a ound
CHF 13,490 pe app en iceship ac oss all company sizes.
Table 12 lis s he a e age ec ui men and induc ion
cos s and he a e age addi ional bene i om e en-
ion ( ec ui men oppo uni y e u ns) by company size.
The esul s show ha la ge companies achie e g ea -
e gains. This is consis en wi h he obse a ions om
2009/10 and 2016/17, al hough he di e ences a e now
once again mo e p onounced as in he 2009/10 su ey.
One eason is ha , as in 2009/10, ec ui men and induc-
ion cos s also ise clea ly wi h company size.
4.10 Ex apola ion o aining cos s,
p oduc i e ou pu and ne bene i
Table 13 shows he ex apola ion o aining cos s, p o-
duc i e ou pu and ne bene i o all ongoing app en-
iceship con ac s o he 2022/23 aining yea . This is
calcula ed by mul iplying he ele an a e age es ima es
by he numbe o places in he co esponding app en-
iceship yea .
The numbe o ongoing app en iceship con ac s can be
de e mined om he SBG-BUR da a, which o m he ba-
sis o he sample o aining companies, as ou lined in
Chap e 3.3. Fo he a ge popula ion o aining com-
panies de ined he e, a o al o 176,183 ongoing dual
app en iceship con ac s was eco ded o he 2022/23
aining yea . Since some companies we e excluded om
he a ge popula ion (e.g. ag icul u al companies), he
ex apola ed cos s and bene i s o Swi ze land a e like-
ly o be somewha unde es ima ed.
To al cos s and bene i s pe app en iceship yea ,
in CHF millions
G oss cos s 4,930
+/- 71
P oduc i e ou pu 5,731
+/- 123
Ne bene i 800
+/- 129
No e: The ex apola ions e e o a o al o 176,183
app en iceship con ac s in he 2022/23 aining yea .
The co esponding 95 pe cen con idence in e al
(‘+/–’) is shown in he second line in each case.
Table 13: Ex apola ion o cos s and bene i s o he 2022/23
app en iceship yea
Based on he a e age g oss cos s in he ne sample, he
ex apola ed o al amoun s o a ound CHF 4,930 million
in aining cos s (g oss cos s). This compa es wi h a o al
o a ound CHF 5,731 million in p oduc i e ou pu gene -
a ed by app en ices du ing hei aining, esul ing in an
ex apola ed posi i e ne bene i o a ound CHF 800 mil-
lion o he 2022/23 aining yea . The 95 pe cen con i-
dence in e al (‘+/-’) is plus/minus CHF 129 million, i.e.
Numbe o employees wi hin he company <10 10–49 50–99 >99
Rec ui men and induc ion cos s 14,130 17,880 20,500 22,130
Bene i o con inued employmen 8,550 11,190 12,400 17,860
Table 12: A e age ec ui men and induc ion cos s pe skilled wo ke and bene i o con inued employmen , by
company size
46 Company pe spec i es on cos s, bene i s and aining quali y 2025
The educa ion sys em includes egula ions ha di ec -
ly guide companies in how hey ain app en ices, and
o he s ha shape he aining con ex mo e indi ec ly.
A cen al example o p esc ip i e egula ions a e VET o -
dinances and aining plans, which a e also binding o
companies. By con as , he Fede al Voca ional Bacca-
lau ea e Type 1 (BM1), which accompanies he app en-
iceship, is an educa ional pa hway aimed p ima ily a
app en ices, hough i also a ec s companies’ aining
en i onmen s. This chap e discusses he signi icance o
aining plans and he Fede al Voca ional Baccalau ea e
o aining companies.
6.1 VET o dinances and aining plans
VET o dinances (“Bildungs e o dnung” BiVo) and ain-
ing plans (“Bildungsplan” BiPla) se ou he asks and
skills ha aining companies a e expec ed o each
app en ices in a gi en occupa ion. They ensu e ha ap-
p en ices acqui e comp ehensi e occupa ional skills ha
ex end beyond he needs o any single company.
This equi es a ce ain deg ee o s anda disa ion in he
aining con en o app en iceships. Bu his also means
ha BiVo and BiPla may no always ma ch he eali ies o
e e y company o e ing aining in he same occupa ion.
Fo example, he day- o-day wo k o an in o ma ion ech-
nologis in an indus ial company is likely o di e om
ha o an in o ma ion echnologis in a se ice company,
ye aine s in bo h a e bound by he same BiVo and BiPla.
This aises he ques ion o how well companies eel he
BiVo and BiPla ele an o hem mee hei needs.
Assessmen o BiVo and BiPla by he companies
Fo he second ime a e 2016/17, companies we e
asked abou he BiVo and BiPla ele an o hem. The
i s ques ion asked whe he he BiVo and BiPla mee he
equi emen s o hei company.2 Figu e 15 summa ises
he esponses om he 6,562 aining companies ha
ook pa .
In mos cases, he BiVo and BiPla appea o mee compa-
ny equi emen s. Fo 55 pe cen o companies his was
he case, and o a u he 13 pe cen i was e en ‘ ully
he case’. Fo 27 pe cen , he BiVo and BiPla only pa ly
me company equi emen s. Fo 6 pe cen , hey did no
o did no a all mee company equi emen s. These as-
sessmen s ha e emained la gely unchanged since he
2016/17 su ey.
In a second ques ion, companies we e asked o indica e
he sha e o aining con en ha is speci ied in he o -
icial aining plan bu no equi ed in hei company.3
Acco ding o Figu e 16, companies conside 82 pe cen
o he aining con en de ined in he BiPla o be ele-
6 TRAINING PLANS AND THE VOCATIONAL
BACCALAUREATE
0% 20 % 40% 60% 80% 100%
Yes, in all espec s Yes, o he mos pa Neu al No eally No a all
2 The ques ion was ac ually ph ased as ollows: “Wi h ega d o he
VET o dinance o SERI and he o icial aining plan o he p o es-
sional/ ade o ganisa ion o he selec ed occupa ion, does he
ollowing s a emen apply o you company? ‘The VET o dinance
and aining plan mee he equi emen s o ou company.’”
3 The ques ion was ac ually ph ased as ollows: “Wha p opo ion
o he aining con en speci ied in he o icial aining plan o
he selec ed occupa ion would you es ima e is no equi ed in
you company?”
Figu e 15: Do he VET o dinances (BiVo) and aining plans (BiPla) mee he company’s equi emen s?

Company pe spec i es on cos s, bene i s and aining quali y 2025 47
an o hei ope a ions. Al hough his igu e is one pe -
cen age poin lowe han in 2016/17, i s ill appea s e y
high gi en ha aining plans a e in ended o ensu e
comp ehensi e occupa ional skills ha ex end beyond
he needs o indi idual companies. Ne e heless, he
a e age esponses di e ac oss app en iceship occupa-
ions. The con idence in e al, shown by he ine lines,
indica es he deg ee o s a is ical unce ain y. Fo he
occupa ions shown in he lowe pa o he igu e, he
p opo ion o unnecessa y aining con en is es ima ed
o be below a e age, while o hose in he uppe pa , a
g ea e sha e o con en in he aining plan is deemed
unnecessa y o he company.
Heal h and Social Ca e Wo ke , FVETC
Heal hca e Assis an , FVETD
Pha macy Assis an , FVETD
Den al Assis an , FVETD
Plumbe , FVETD
Ce i ied Social Ca e Wo ke , FVETD
Au omo i e Technician, FVETD
Cons uc ion Elec ician, FVETD
Ga dene , FVETD
Pain e , FVETD
O he FVETC
Elec ician, FVETD
Building and G ounds Cus odian, FVETD
B icklaye , FVETD
Ca pen e , FVETD
O he h ee-yea FVETD
O ice Assis an , FVETC
Medical Sec e a y and Assis an , FVETD
Che , FVETD
D augh sman/D augh swoman, FVETD
O he ou -yea FVETD
Re ail Assis an , FVETC
Joine , FVETD
Au omo i e Mecha onics Technician, FVETD
Polymechanic, FVETD
Hai d esse , FVETD
Comme cial Employee, FVETD
Re ail P o essional, FVETD
Building Se ices Assis an , FVETC
Au oma ion Technician, FVETD
Logis ics Assis an , FVETC
In o ma ion Technologis , FVETD
Logis ics P o essional, FVETD
A e age
0% 5% 10% 15% 20% 25% 30%
Figu e 16: P opo ion o con en no equi ed in aining plan (BiPla)
48 Company pe spec i es on cos s, bene i s and aining quali y 2025
Companies also p o ide hei app en ices wi h addi ion-
al quali ica ions ha a e no co e ed by he BiPla. This
may be he case, o example, when such quali ica ions
a e ele an o in e nal company p ocesses. Nine een
pe cen o companies epo ha hey each hei ap-
p en ices addi ional quali ica ions beyond he BiPla – a
signi ican ly highe p opo ion han in he 2016/17 su -
ey (12 pe cen ).4
Using da a om he 2016/17 su ey, Schwe i e al. (2021)
and Aepli e al. (2021b) examined whe he a misma ch
be ween he skills p esc ibed in he aining plan and
hose equi ed in he wo kplace esul s in cos s o com-
panies. Acco ding o he au ho s, his is indeed he case:
bo h ypes o misma ch – skills no equi ed o addi-
ional quali ica ions augh – a e associa ed wi h highe
g oss aining cos s and a lowe p oduc i e con ibu ion
om app en ices. These e ec s a e s a is ically signi -
ican bu no e y la ge: e.g., o e e y en pe cen age
poin s o aining con en no equi ed, he ne bene i
pe app en iceship yea is a ound CHF 630 lowe .
The key inding om hese analyses is ha b oade
aining plans a e associa ed wi h cos s o companies.
Ve y b oad skills p o iles, such as hose ha would esul
om a signi ican educ ion in he numbe o app en-
iceship occupa ions, could he e o e en ail subs an ial
cos s o companies and lowe hei willingness o p o-
ide aining.
6.2 Fede al Voca ional
Baccalau ea e1 om a
company pe spec i e
The numbe o app en ices able o pu sue a Fede al Vo-
ca ional Baccalau ea e Type 1 which accompanies he
app en iceship depends in pa on whe he companies
a e willing o suppo o ac i ely p omo e his op ion. A
key ac o he e is he ex en o which educed ime spen
in he wo kplace – due o he highe numbe o school
lessons – a ec s he cos –bene i balance. As T ede e al.
(2020, pp. 17–20) show, he cos –bene i balance di e ed
in he 2016/17 aining yea be ween app en iceship
places wi h and wi hou BM1. Places wi h BM1 displayed
a lowe ne bene i du ing aining because app en ices
spen less ime in he company.
Since BM1 app en ices a e une enly dis ibu ed ac oss
occupa ions, sec o s and companies, hese di e ences
canno be measu ed di ec ly in he da a bu mus be
adjus ed o using mul i a ia e eg ession models (see
T ede e al., 2020). Figu e 17 p esen s he esul s o his
es ima ion, based on 11,388 obse a ions pe app en-
iceship yea om 5,519 companies o he 2022/23 ain-
ing yea . The cos alues we e calcula ed o an a e age
app en iceship yea in wo hypo he ical companies: one
aining no BM1 app en ices and he o he aining only
BM1 app en ices.
As in he 2016/17 aining yea , he a e age ne bene i
pe aining yea in 2022/23 is lowe o app en ices en-
olled in BM1, once di e ences in company cha ac e -
is ics a e con olled o . The educ ion in ne bene i is
jus unde CHF 3,000, somewha smalle han in 2016/17
(jus unde CHF 4,000), bu s a is ically clea ly signi i-
can . Ne e heless, companies aining BM1 app en ices
also achie e a posi i e ne bene i on a e age al eady
du ing he aining pe iod. The di e ence be ween ap-
p en iceship places wi h and wi hou BM1 is due o he
lowe p oduc i e ou pu o he app en ices – no o high-
e g oss cos s. Companies ul il hei aining esponsi-
bili y and do no compensa e o lowe a endance by
cu ing back on hei aining inpu . Ins ead, app en ices
a e less equen ly assigned o p oduc i e asks, pa icu-
la ly unskilled ac i i ies.
4 The ques ion was ac ually ph ased as ollows: “In you com-
pany, we e app en ices in he 2022/23 aining yea augh addi-
ional quali ica ions beyond he con en speci ied in he o icial
aining plan o he selec ed occupa ion?”
Company pe spec i es on cos s, bene i s and aining quali y 2025 49
These esul s do no ake in o accoun ha BM1 enables
companies o a ac pa icula ly capable app en ices.
I hey succeed in e aining hem as skilled wo ke s, he
company gains an addi ional bene i – especially whe e
ec ui ing and onboa ding skilled wo ke s h ough he
ex e nal labou ma ke is a demanding p ocess.
0
10
20
30
40
In CHF housands
No BM1 app en ices Only BM1 app en ices
G oss cos s P oduc i e ou pu Ne bene i
Figu e 17: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea wi h o wi hou BM1
app en ices
50 Company pe spec i es on cos s, bene i s and aining quali y 2025
Since app en ices spend he majo i y o he aining pe-
iod in hei hos company, hese play a key ole in hei
skills de elopmen . I is he e o e unsu p ising ha he
quali y o company aining has been he subjec o e-
sea ch o many yea s (Schwande , 1949; Hä eli e al.,
1981). App en ices (and hei pa en s) ha e a s ong in-
e es in ecei ing op imal suppo om he company,
no leas o educe he isk o e mina ing he app en-
iceship con ac p ema u ely o ailing he inal exam-
ina ions.
As pa o his la es cos –bene i su ey, companies
we e asked abou hei aining p ocesses, i.e. how hey
wo k wi h app en ices and alloca e asks o hem. This
chap e e alua es hose esponses, p o iding an o e -
iew o aining quali y in Swiss companies.
7.1 Wha is he de ini ion o in-house
aining quali y?
Despi e he ob ious impo ance o aining companies
o oca ional educa ion and aining, in-house aining
quali y is di icul o de ine and measu e (Be ge e al.,
2023). Quali y models o en dis inguish be ween inpu ,
p ocess and ou pu quali y (Gonon, 2008). F om an eco-
nomic pe spec i e, ou pu quali y is o pa icula sig-
ni icance: good aining means ha app en ices mas e
hei occupa ion and a e able o ac compe en ly in all
p o essional si ua ions. This de ini ion o quali y, how-
e e , does no speci y how companies can achie e a high
le el o ou pu quali y.
Resea ch in wo k and o ganisa ional psychology, as well
as in oca ional and business educa ion, has iden i ied
company p ocesses ha make employees in gene al and
app en ices in pa icula bo h sa is ied and p oduc i e
(Hackman & Oldham, 1975; Mo geson & Humph ey, 2006;
Böhn & Deu sche , 2021). In a company su ey, Neg ini e
al. (2016) used ques ionnai e i ems de eloped in his li -
e a u e. Using da a om 335 aining companies o he
occupa ions o cook and pain e , hey demons a ed a
signi ican co ela ion be ween hei quali y i ems and
he numbe o p ema u e app en iceship con ac e mi-
na ions.
Measu ing aining quali y
This cos –bene i su ey used eigh i ems as measu es o
p ocess quali y in aining companies, mos o hem ak-
en om Neg ini e al. (2016). Responden s we e asked o
a e he ollowing i ems:
1. Planning: I plan he aining uni s in ad ance.
2. Au onomous solu ions: I allow he app en ices o ind
au onomous solu ions when wo king on an assign-
men .
3. Va ie y: I ensu e ha he app en ices a e gi en a a i-
e y o asks.
4. Feedback: I gi e he app en ices egula eedback on
hei pe o mance.
5. Insigh : The app en ices gain insigh s in o all phases
o he wo k and p oduc ion p ocess.
6. Responsibili y: I assign he app en ices asks ha ca -
y eal esponsibili y.
7. Lis ening: I make an e o o lis en ca e ully o he ap-
p en ices.
8. Op imal wo kload: I make su e he app en ices a e
no o e whelmed.
In addi ion o hese eigh i ems, u he in o ma ion ha
in luences aining quali y was collec ed om aining
companies. This no ably includes da a on he ime e-
sou ces a ailable o company aining manage s and on
he a ailable in as uc u e. These a iables a e unde -
s ood as measu es o inpu quali y and a e la e used o
examine whe he di e ences be ween companies in he
aining p ocesses desc ibed abo e can be a ibu ed o
di e ing le els o esou ces.
7.2 How do aining companies assess
hei aining quali y?
The su ey p omp ed companies o indica e how accu-
a ely hey el he eigh s a emen s abo e applied o
hei own aining, using a scale om one (no a all) o
se en (in all espec s).
On he whole, companies a ed hemsel es e y posi-
i ely, as shown in Figu e 18: he a e age esponses o
all eigh i ems lie be ween 5.1 and 6.2. Wha ma e s o
he analyses ha ollow, howe e , is no he gene ally
7 IN-HOUSE TRAINING QUALITY
Company pe spec i es on cos s, bene i s and aining quali y 2025 51
high a e ages, bu he a ia ion be ween companies. I
can be assumed ha highe sel -assessmen s also co -
espond o a highe le el o aining quali y. In addi ion
o su eying companies, Neg ini e al. (2016) also asked
hei app en ices abou pe cei ed aining quali y and
ound ha , on a e age, app en ices ga e lowe a ings
han hei company. Ne e heless, he esponses o ap-
Neu al
Yes, o
some ex en
Yes, o he
mos pa
Yes, in all
espec s
I make an e o
o lis en ca e ully
o he app en ices.
I assign he app en ices
asks ha ca y
eal esponsibili y.
I allow he app en ices o ind
au onomous solu ions when
wo king on an assignmen .
I make su e he app en ices
a e no o e whelmed.
The app en ices gain insigh s
in o all phases o he wo k
and p oduc ion p ocess.
I gi e he app en ices
egula eedback
on hei pe o mance.
I plan he aining
con en in ad ance.
I ensu e ha he
app en ices a e gi en
a a ie y o asks.
4 5 6 7
Figu e 18: S a emen s on aining quali y, as a ed by companies
p en ices and companies co ela ed s ongly, meaning
ha in companies wi h highe sel - a ings, app en ices
also ga e highe assessmen s, and ice e sa. This sug-
ges s ha companies’ sel -assessmen s do indeed e-
lec di e ences in aining quali y – a leas as seen by
hei app en ices.

52 Company pe spec i es on cos s, bene i s and aining quali y 2025
S ong aining companies s and ou because hey main-
ain good p ocesses in all eigh a eas. I can he e o e be
assumed ha he eigh i ems a e posi i ely co ela ed.
In o he wo ds, i a company sco es abo e a e age on
one i em, i is mo e likely o sco e abo e a e age on he
o he s oo.
This assump ion is con i med in Figu e 19, which shows
he pai wise co ela ion coe icien s o all i ems. Theo-
e ically, co ela ion coe icien s can ange be ween mi-
nus one and plus one, bu in his case all co ela ions
a e abo e ze o, albei consis en ly below 0.6. The highe
he coe icien , he blue he ield. Only he wo i ems
‘Feedback’ and ‘Lis ening’ show a co ela ion o mo e
han 0.5. A clea link be ween eedback and lis ening
seems plausible, as bo h a e likely o occu in con e sa-
ions be ween aine s and app en ices. Companies ha
ake ime mo e equen ly o such con e sa ions a e
likely o sco e highe on bo h i ems.
0.32
0.32
0.33
0.41
0.33
0.47
0.21
0.26
0.42
0.41
0.21
0.38
0.44
0.41
0.47
0.31
0.33
0.46
0.53
0.40
0.46
0.34
0.35
0.34
0.38
0.29
0.36
0.49
Au onomous solu ions
Va ie y
Feedback
Insigh
Responsibili y
Lis ening
Op imal wo kload
Au onomous solu ions
Planning
Va ie y
Feedback
Responsibili y
Insigh
Lis ening
Figu e 19: Co ela ions be ween indi idual i ems on aining quali y
Company pe spec i es on cos s, bene i s and aining quali y 2025 53
7.3 Fi e aining ypes o aining
companies
The eigh i ems make i possible o g oup companies
acco ding o hei aining p ac ices. A clus e analy-
sis assigns companies o aining ypes (clus e s) so
ha companies wi hin a ype a e as simila as possible
ac oss he eigh i ems, while di e ing as much as pos-
sible om companies in o he ypes. Fi e aining ypes
can be iden i ied, each cha ac e ised below by a sho
label and a desc ip ion o hei aining p ac ices. These
labels we e in en ionally chosen o be clea and some-
wha o e s a ed so ha hei cha ac e is ics can be eas-
ily unde s ood.
1. TOP: Companies in his ype achie e consis en ly e y
high aining quali y.
2. OKAY: Companies in his ype show sa is ac o y – i.e.
a e age – aining quali y ac oss all i ems.
3. SPONTANEOUS: Companies in his ype p o ide sa is-
ac o y o good aining quali y ac oss all i ems, bu
aining is planned much less equen ly in ad ance.
4. AUTHORITARIAN: Companies in his ype a ely allow
app en ices o ind au onomous solu ions. They also
plan aining less in ad ance and pay less a en ion
o p e en ing app en ices om eeling o e whelmed.
On he emaining i ems, howe e , hey gene ally sco e
sligh ly highe han he OKAY g oup.
5. BASIC: Companies in his ype a e cha ac e ised by
consis en ly below-a e age aining quali y.
The i e ypes and he associa ed companies a e p e-
sen ed in mo e de ail in Table 17.
54 Company pe spec i es on cos s, bene i s and aining quali y 2025
TOP
24.7%
SPONTA-
NEOUS
18.6%
OKAY
34.4%
AUTHORI-
TARIAN
7.8%
BASIC
14.5%
T aining quali y
Planning 6.2 3.45.7 5.0 4.0
Au onomous solu ions
6.4 5.75.8 3.3 4.6
Va ie y 6.6 5.75.6 5.8 4.6
Feedback 6.7 5.75.8 6.0 4.5
Insigh 6.7 6.25.9 6.3 4.8
Responsibili y
6.5 6.05.7 5.6 4.5
Lis ening 6.8 6.26.1 6.3 4.9
Op imal wo kload 6.5 5.45.6 5.1 4.5
Company cha ac e is ics
T aining
equi emen s
T aine s ha e
less ime
Poo e aining
in as uc u e
Company
digi alisa ion
Mo e impac ed
by digi alisa ion
Less impac ed
by digi alisa ion
Company size Mo e e y small
companies
(<10 employees)
Fewe e y
small companies
(<10 employees)
Language egion
( e e ence: Ge man)
– F ench
+ I alian
+ F ench
+ I alian
– F ench
– I alian
– F ench
– I alian
+ F ench
T aine s ha e
mo e ime
Be e aining
in as uc u e
T aine s ha e
less ime
Table 17: Types o aining
Company pe spec i es on cos s, bene i s and aining quali y 2025 55
The uppe hal o he able shows he a e age alues o
companies o all eigh ques ionnai e i ems on aining
quali y. The main indings ha e al eady been desc ibed
abo e in he desc ip ions o he indi idual ypes.
The colou ed ba s indica e he sha e o all aining com-
panies in he sample assigned o each o he i e ypes.
O e all, mo e han h ee qua e s o aining companies
belong o he h ee ypes ha p o ide sa is ac o y o
e y good aining quali y (TOP, OKAY, SPONTANEOUS).
This is an encou aging esul . I should be no ed, how-
e e , ha he alloca ion o indi idual companies o he
i e ypes also depends on he exac speci ica ions o he
clus e analysis. The epo ed sha es o he i e aining
ypes should he e o e be ega ded as app oxima e al-
ues.
The une en dis ibu ion o oughly h ee qua e s o one
qua e is due o he ac ha many companies epo ed
high alues, esul ing in a concen a ion o esponses a
he uppe end o he scale. T aining quali y is he e o e
no symme ically dis ibu ed, as in a Gaussian no mal
dis ibu ion, bu skewed o he igh . This asymme y
sugges s ha mos aining companies s i e o p o ide
good aining, wi h companies o e ing poo e aining
being less common. A he same ime, he pa e n may
also e lec o e es ima ion by companies hemsel es–
as sugges ed by Neg ini e al. (2016), who ound app en-
ices ended o gi e lowe a ings o aining quali y.
Fu u e s udies on aining quali y should he e o e aim
o e i y he dis ibu ion o aining quali y using ep-
esen a i e da a no based solely on company sel -as-
sessmen s. As Chap e 7.5 will show, howe e , company
sel -assessmen s a e closely linked o he subsequen
aining success o hei app en ices and he e o e, de-
spi e some open ques ions, cons i u e a alid measu e
o aining quali y.
7.4 How do he companies in he
di e en aining ypes di e ?
The aining companies in he sample we e assigned
o i e aining ypes solely based on hei assessmen
o aining p ocesses. The lowe pa o Table 17 shows
which o he cha ac e is ics dis inguish he companies
in he di e en ypes – in o he wo ds, which ypes o
companies a e o e - o unde - ep esen ed in each
g oup. These igu es a e based on a mul i a ia e eg es-
sion analysis. This also means ha he epo ed e ec s
– such as he e ec o ime esou ces o aining o o
digi alisa ion – canno be a ibu ed o he sepa a e-
ly con olled e ec s o company size, language egion,
economic sec o and occupa ion.
A decisi e ac o is whe he companies p o ide aine s
wi h su icien ime o aining ac i i ies and good in-
as uc u e. Companies ha do so a e mo e equen ly
ound in he TOP g oup. Whe e ime is sca ce and/o in-
as uc u e is subop imal, a company is mo e likely o
all in o he g oup wi h he lowes aining quali y. Such
companies a e also mo e o en ound in he g oup wi h
spon aneous, less planned aining p ac ices. Appa en -
ly, aine s in his g oup manage o achie e a sa is ac-
o y o good le el o aining (p ocess) quali y despi e
un a ou able condi ions and limi ed planning.
The signi icance o ime esou ces is no en i ely su -
p ising, as hey a e ega ded in he li e a u e as pa o
inpu quali y. The s ong associa ion wi h p ocess qual-
i y is empi ically con i med in he p esen analysis and
also suppo s he indings o Lamam a e al. (2019) and
Wenge and Lamam a (2023), which emphasise he im-
po ance o su icien ime o in-company aine s.
Companies mo e a ec ed by digi alisa ion a e mo e
likely o be in he TOP g oup and less likely o be in he
BASIC g oup. I may be assumed ha companies mo e
exposed o digi alisa ion gene ally ha e be e p ocess-
es, which also ansla e in o highe aining quali y. Ve y
small companies wi h ewe han en employees end o
be among hose ha plan aining less, while being less
o en ep esen ed in he OKAY g oup.
A s ong e ec is also e iden in he language egions:
companies om F ench-speaking Swi ze land a e less
o en ep esen ed in he i s h ee ypes and mo e e-
quen ly in he ype wi h au ho i a ian aining as well as
in he ype wi h below-a e age aining quali y (BASIC).
I emains unclea o wha ex en he lowe popula i-
y and weake ancho ing o dual oca ional educa ion
and aining in F ench-speaking Swi ze land play a ole
62 Company pe spec i es on cos s, bene i s and aining quali y 2025
Acemoglu, D., & Pischke, J. (1998). Why do i ms ain?
Theo y and e idence. Qua e ly Jou nal o Econom-
ics, 113, 79–119.
Acemoglu, D., & Pischke, J. (1999). The S uc u e o
Wages and In es men in Gene al T aining. Jou nal
o Poli ical Economy, 107(3), 539–572. h ps://doi.
o g/10.1086/250071
Aepli, M., Kuhn, A., & Schwe i, J. (2021a). Cul u e, no ms,
and he p o ision o aining by employe s: E idence
om he Swiss language bo de . Labou Economics,
73, 1–21. h ps://doi.o g/10.5281/zenodo.5749913
Aepli, M., Schwe i, J., & Kuhn, A. (2021b). Bildungspläne
siche n b ei e Ausbildung – e u sachen abe auch
Kos en ü die Leh be iebe. EHB-S udie zu Passung
on Bildungsplänen und be ieblichen Bedü nissen.
T ans e . Be u sbildung in Fo schung und P axis,
6(3).
A Bu én, P., & Schwe i, J. (2024). Fi ms’ aining p ocess-
es and hei app en ices’ educa ion success. Swiss
Leading House – Economics o Educa ion, Fi m Be-
ha iou , T aining Policies, Wo king Pape No. 225.
Backes-Gellne , U., & Lehne , P. (2023). Be u liche Bil-
dung als Inno a ions eibe : Ein lange e nachläs-
sig es Fo schungs eld. Pe spek i en de Wi scha -
spoli ik, 24(1), 85–97.
Ba deleben, R. on, Beich , U., & Fehé , K. (1995). Be-
iebliche Kos en und Nu zen de Ausbildung:
Rep äsen a i e E gebnisse aus Indus ie, Handel
und Handwe k. In Ge man Fede al Ins i u e o Voca-
ional Educa ion and T aining BIBB (Eds.), Be ich e
zu be u lichen Bildung, issue 187. Biele eld: Be els-
mann.
Beich , U., Walden, G., & He ge , H. (2004). Kos en und
Nu zen de be ieblichen Be u sausbildung in
Deu schland. Biele eld: Be elsmann.
Be ge , J.-L., Wenge , M., & Sauli, F. (2023). Engagemen
en o ma ion p o essionnelle ini iale duale e pe -
cep ions de la quali é de la o ma ion. Didac ique,
4(2), 33–64.
Böhn, S., & Deu sche , V. (2021). De elopmen and ali-
da ion o a lea ning quali y in en o y o in-compa-
ny aining in VET (VET-LQI). Voca ions and Lea ning,
14, 23–53.
Bonin, H., F ies, J., Hille ich-Sigg, A., Maie , M.F., & Wal e ,
T. (2014). Beglei o schung "Auswi kungen des Aus-
bildungsbonus au den Ausbildungsma k und die
ö en lichen Haushal e". Abschlussbe ich . Be lin:
Ge man Fede al Minis y o Labou and Social A ai s
(BMAS).
Dionisius, R., Muehlemann, S., P ei e , H., Walden, G.,
Wenzelmann, F., & Wol e , S. C. (2009). Cos s and
bene i s o app en iceship aining. A compa ison o
Ge many and Swi ze land. Applied Economics Qua -
e ly, 1, 7–37.
Dong, Y., & Peng, CY.J. (2013). P incipled missing da a
me hods o esea che s. Sp inge Plus, 2, 222.
h ps://doi.o g/10.1186/2193-1801-2-222
Do nmay , H., G a , L., K oll, S., Ma kowi sch, J., Ne ,
D., & Neumann, J. (2025). Indika o en zu dualen
Be u sbildung 2024. Ge many – Aus ia – Swi ze -
land. Resea ch epo . Bonn – Vienna – Zolliko en:
Aus ian Fede al Minis y o Labou and Economy.
Euwals, R., & Winkelmann, R. (2004). T aining in ensi-
y and i s labo ma ke ou comes o app en ice-
ship g adua es. In e na ional Jou nal o Manpowe ,
25(5), 447–462.
Fede al S a is ical O ice (FSO) (2024). Ak uell. Be u liche
G undbildung: Le no e de Le nenden. Neuchâ el:
Fede al S a is ical O ice (FSO).
Fuh e , M., & Schwe i, J. (2010). Kos en und Nu zen
on zweijäh igen be u lichen G undbildungen.
Schlussbe ich . Zolliko en: Swiss Fede al Ins i u e
o Voca ional Educa ion and T aining.
REFERENCES

Company pe spec i es on cos s, bene i s and aining quali y 2025 63
Geh e , A., Aepli, M., Kuhn, A., & Schwe i, J. (2019). Lohn
sich die Leh lingsausbildung ü die Be iebe? Re-
sul a e de ie en Kos en-Nu zen-E hebung. Zolli-
ko en: Swiss Fede al Ins i u e o Voca ional Educa-
ion and T aining.
Gonon, P. (2008). Quali ä ssiche ung in de be u lichen
Bildung. Eine Bilanz. In E. Klieme & R. Tippel (Eds.),
Quali ä ssiche ung im Bildungswesen (p. 96–107).
Bel z. h ps://www.pedocs.de/ on doo .php?-
sou ce_opus=7271
Hackman, J.R., & Oldham, G.R. (1975). De elopmen o
he Job Diagnos ic Su ey. Jou nal o Applied Psy-
chology, 60, 159–170. h ps://dx.doi.o g/10.1037/
h0076546
Hä eli, K., F ischknech , E., & S oll, F. (1981). Schwei-
ze Leh linge zwischen Ausbildung und P oduk ion.
Fo schungse gebnis eine Leh lingsun e suchung.
Mu i: Cosmos Ve lag.
Jansen, A., S uple Leise , M., Wenzelmann, F., & Wol e ,
S.C. (2015). Labou ma ke de egula ion and ap-
p en iceship aining: A compa ison o Ge man and
Swiss employe s. Eu opean Jou nal o Indus ial Re-
la ions, 21(4), 353–368.
K iesi, I., Bonoli, L., G ønning, M., Hänni, M., Neumann, J.,
& Schwe i, J. (2022). Spannungs elde in de Be u s-
bildung in e na ional und in de Schweiz – En wick-
lungen, He aus o de ungen, Po enziale. OBS SFUVET
T end Repo 5. Zolliko en: Swiss Fede al Uni e si y
o Voca ional Educa ion and T aining SFUVET.
K iesi, I., Neumann, J., Schwe i, J., G iga, D., Kuhn, A.,
Schmid, E. und Baumele , C. (2016). "Bleiben? Ge-
hen? Neu beginnen? Häu igkei , U sachen und Fol-
gen on Leh e agsau lösungen". T end Repo 1 o
he Swiss Obse a o y o Voca ional Educa ion and
T aining. Zolliko en: Swiss Fede al Ins i u e o Voca-
ional Educa ion and T aining.
Kuhn, A., Geh e , A., & Schwe i, J. (2025). Me hoden be-
ich zu ün en Kos en-Nu zen-E hebung de be u-
lichen G undbildung. Zolliko en: Swiss Fede al Uni-
e si y o Voca ional Educa ion and T aining.
Kuhn, A. (2024). Eine Flo e aus Jollen und Ba ken,
Ku e n und Tanke n. T ans e . Be u sbildung in
Fo schung und P axis, 9(1).
Kuhn, A., & Schwe i, J. (2019). Me hodenbe ich zu i-
e en Kos en-Nu zen-E hebung de be u lichen
G undbildung. Zolliko en: Swiss Fede al Ins i u e o
Voca ional Educa ion and Lea ning.
Kuhn, A., Schwe i, J., & Wol e , S. C. (2022). Local no ms
desc ibing he ole o he s a e and he p i a e
p o ision o aining. Eu opean Jou nal o Poli i-
cal Economy, 75. h ps://doi.o g/10.1016/j.ejpole-
co.2022.102226
Lamam a, N., Duc, B., & Besozzi, R. (2019). Im He zen
de dualen Be u sbildung: Be u sbildne innen und
Be u sbildne . Fo schungse gebnisse und Massnah-
men ü die Ak eu innen und Ak eu e de Be u sbil-
dung. Zolliko en: Swiss Fede al Ins i u e o Voca-
ional Educa ion and T aining.
Lindley, R. M. (1975). The demand o app en ice ec ui s
by he enginee ing indus y, 1951–71. Sco ish Jou -
nal o Poli ical Economy, 22(1), 1–24.
Li le, R.J.A., & Rubin, D.B. (2019). S a is ical Analysis
wi h Missing Da a. 3 d edi ion. Hoboken, NJ: Wiley.
Mo geson, F. P., & Humph ey, S. E. (2006). The Wo k De-
sign Ques ionnai e (WDQ): De eloping and alida ing
a comp ehensi e measu e o assessing job design
and he na u e o wo k. Jou nal o Applied Psychol-
ogy, 91(6), 1321–1339. h ps://doi.o g/10.1037/0021-
9010.91.6.1321
Mühlemann, S. (2013). De Ein luss de In e na ionalisi-
e ung au die a bei sma k o ien ie e Bildung in de
Schweiz. Economics o Educa ion Wo king Pape Se-
ies No. 92, Uni e si y o Zu ich.
64 Company pe spec i es on cos s, bene i s and aining quali y 2025
Mühlemann, S. (2019). Measu ing Pe o mance in Voca-
ional Educa ion and T aining and he Employe ’s
Decision o In es in Wo kplace T aining. In D. Guile
& L. Unwin (Eds.), The Wiley Handbook o Voca ion-
al Educa ion and T aining (p. 187–206). h ps://doi.
o g/10.1002/9781119098713.ch10
Mühlemann, S., P ei e , H., Walden, G., Wenzelmann, F.,
& Wol e , S.C. (2010). The inancing o app en ice-
ship aining in he ligh o labo ma ke egula ions.
Labou Economics, 17(5), 799–809.
Mühlemann, S., Schwe i, J., Winkelmann, R., & Wol e ,
S.C. (2007a). An empi ical analysis o he decision o
ain app en ices. LABOUR – Re iew o Labou Eco-
nomics and Indus ial Rela ions, 21(3), 419–441.
Mühlemann, S., & Wol e , S.C. (2014). Re u n on in es -
men o app en iceship sys ems o en e p ises: E -
idence om cos –bene i analyses. IZA Jou nal o
Labo Policy, 3(1), 1–22. doi.o g/10.1186/2193-9004-
3-25
Mühlemann, S., Wol e , S.C., Fuh e , M., & Wües , A.
(2007b). Leh lingsausbildung – ökonomisch be a-
ch e . Chu : Rüegge .
Mülle , B., & Schwe i, J. (2012). Die Be iebe in de dua len
Be u sbildung: En wicklungen 1985 bis 2008. Neu-
châ el: Fede al S a is ical O ice
Neg ini, L., Fo sblom, L., Gu ne , J. L., & Schumann, S.
(2016). Is he e a ela ionship be ween aining qual-
i y and p ema u e con ac e mina ions in VET? Vo-
ca ions and Lea ning, 9(3), 361–378.
Rupie a, C., & Backes-Gellne , U. (2018). How i ms’ pa -
icipa ion in app en iceship aining os e s knowl-
edge di usion and inno a ion. Jou nal o Business
Economics. doi.o g/10.1007/s11573-018-0924-6
Schön eld, G., Jansen, A., Wenzelmann, F., & P ei e ,
H. (2016). Kos en und Nu zen de dualen Ausbil-
dung aus Sich de Be iebe. E gebnisse de ün en
BIBB-Kos en-Nu zen-E hebung. Biele eld: Be els-
mann.
Schwande , A. (1949). Die Ausbildung des Leh lings
im Be ieb. Ve lag de deu schschweize ischen
Leh lingsäm e -Kon e enz.
Schwe i, J., Aepli, M., & Kuhn, A. (2021). The Cos s o
S anda dized App en iceship Cu icula o T aining
Fi ms. Empi ical Resea ch in Voca ional Educa ion
and T aining, 13(16).
Schwe i, J., & a Bu én, P. (2024). Abwechslungs eiche
Au gaben im Leh be ieb machen Le nende e olg-
eiche . T ans e . Be u sbildung in Fo schung und
P axis, 9(12).
Schwe i, J., Mühlemann, S., Pescio, Y., Wal he , B.,
Wol e , S.C., & Zü che , L. (2003). Kos en und Nu zen
de Leh lingsausbildung aus de Sich Schweize
Be iebe. Chu /Zu ich: Rüegge .
S uple , M., & Wol e , S.C. (2012). Die duale Leh e: eine E -
olgsgeschich e auch ü die Be iebe: E gebnisse de
d i en Kos en-Nu zen-E hebung de Leh lings aus-
bildung aus de Sich de Be iebe. Zu ich: Rüegge .
Swiss Na ional Bank (2025). Fo eign exchange a es –
Yea . h ps://da a.snb.ch/en/ opics/zi ede /cube/
de kua (las accessed: 27 May 2025).
T ede, I., Hänni, M., Leumann, S., Neumann, J., Geh e ,
A., Schwe i, J., & K iesi, I. (2020). Be u sma u i ä . Bil-
dungs e läu e, He aus o de ungen und Po enziale.
OBS SFIVET T end Repo 4. Zolliko en: Swiss Fede al
Ins i u e o Voca ional Educa ion and T aining.
an Buu en, S. (2007). Mul iple impu a ion o disc e e
and con inuous da a by ully condi ional speci ica-
ion. S a is ical Me hods in Medical Resea ch, 16(3),
219–242.
an Buu en, S. (2018). Flexible Impu a ion o Missing
Da a. Second Edi ion. Boca Ra on, FL: CRC P ess.
Wal he , B., Schwe i, J., & Wol e , S.C. (2005). Shall I ain
you app en ice? An empi ical in es iga ion o ou -
sou cing o app en iceship aining in Swi ze land.
Educa ion + T aining, 47(4), 251–269.
Company pe spec i es on cos s, bene i s and aining quali y 2025 65
Wenge , M., & Lamam a, N. (2023). Les besoins e
p é é ences en ma iè e de o ma ion con inue des
pe sonnes o ma ices d’app en i-e-s. Analyses de
l’enquê e en ligne – Rappo inal. Renens: Swiss
Fede al Uni e si y o Voca ional Educa ion and
T aining SFUVET.
Wenzelmann, F., Schön eld, G., Linckh, C., & P ei e , H.
(2025). Eigene Ausbildung ode ex e ne Fachk ä ege-
winnung – mi welchen Kos en müssen Be iebe
echnen? E gebnisse de BIBB-Kos en-Nu zen-E he-
bung 2022/2023. BIBB Repo 2. Bonn.
Wol e , S.C. (2008). Ausbildungskos en und -nu zen und
die be iebliche Nach age nach Leh lingen. Pe s-
pek i en de Wi scha spoli ik, 9 (Special Issue),
90–108.
66 Company pe spec i es on cos s, bene i s and aining quali y 2025
Figu e 1: Dis ibu ion o ne bene i 24
Figu e 2: Sa is ac ion wi h he cos –bene i balance o company-based aining 24
Figu e 3: B eakdown o g oss cos s in FVETC app en iceships 26
Figu e 4: Time spen by app en ices a he company wo kplace in FVETC app en iceships 27
Figu e 5: B eakdown o g oss cos s in h ee-yea FVETD app en iceships 29
Figu e 6: Time spen by app en ices a he company wo kplace in h ee-yea FVETD
app en iceships 30
Figu e 7: B eakdown o g oss cos s in ou -yea FVETD app en iceships 31
Figu e 8: Time spen by app en ices a he company wo kplace in ou -yea FVETD
app en iceships 32
Figu e 9: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea ,
by company size 34
Figu e 10: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea ,
by sec o 35
Figu e 11: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea ,
by majo egion 36
Figu e 12: A e age ne bene i o he mos common FVETC app en iceship occupa ions 37
Figu e 13: A e age ne bene i o he mos common h ee-yea FVETD app en iceship
occupa ions 37
Figu e 14: A e age ne bene i o he mos common ou -yea FVETD app en iceship
occupa ions 38
Figu e 15: Do he VET o dinances (BiVo) and aining plans (BiPla) mee he company’s
equi emen s? 46
Figu e 16: P opo ion o con en no equi ed in aining plan (BiPla) 47
Figu e 17: G oss cos s, p oduc i e ou pu and ne bene i pe a e age app en iceship yea
wi h o wi hou BM1 app en ices 49
Figu e 18: S a emen s on aining quali y, as a ed by companies 51
Figu e 19: Co ela ions be ween indi idual i ems on aining quali y 52
Figu e 20: Reasons o no aining app en ices in-house 57
Figu e 21: Reasons o aining app en ices in-house 59
Figu e 22: Assessmen o measu es o boos aining ac i i y 60
LIST OF FIGURES
Company pe spec i es on cos s, bene i s and aining quali y 2025 67
Table 1: O e iew o g oss cos s 8
Table 2: O e iew o p oduc i e ou pu 9
Table 3: Responses in he wo g oss samples 18
Table 4: Composi ion o he ne sample o aining companies 19
Table 5: Composi ion o he ne sample o non- aining companies 21
Table 6: G oss cos s, p oduc i e ou pu and ne bene i pe FVETC app en iceship (in CHF) 25
Table 7: Mon hly g oss wage o app en ices in FVETC app en iceships (in CHF) 26
Table 8: G oss cos s, p oduc i e ou pu and ne bene i pe h ee-yea FVETD app en iceship
(in CHF) 28
Table 9: Mon hly g oss wage o app en ices in h ee-yea FVETD app en iceships (in CHF) 28
Table 10: G oss cos s, p oduc i e ou pu and ne bene i pe ou -yea FVETD app en iceship
(in CHF) 31
Table 11: Mon hly g oss wage o app en ices in ou -yea FVETD app en iceships (in CHF) 32
Table 12: A e age ec ui men and induc ion cos s pe skilled wo ke and bene i
o con inued employmen , by company size 39
Table 13: Ex apola ion o cos s and bene i s o he 2022/23 app en iceship yea 39
Table 14: Compa ison o he cos s and bene i s o oca ional aining in Swi ze land
and Ge many 40
Table 15: Mul i a ia e analysis o cos s and bene i s 2022/23 43
Table 16: Mul i a ia e analysis o cos s and bene i s 2016/17 in compa ison o 2022/23 45
Table 17: Types o aining 54
Table 18: De e minan s o aining pa icipa ion 61
TABLES

Swiss Fede al Uni e si y
o Voca ional Educa ion
and T aining SFUVET
Ki chlindachs asse 79
CH-3052 Zolliko en
+41 58 458 27 00
www.s u e .swiss
in o@s u e .swiss