Vi a elli, Ma co; A enas Díaz, Guille mo
Wo king Pape
New echnologies and employmen : The s a e o he a
UNU-MERIT Wo king Pape s, No. 2025-005
P o ided in Coope a ion wi h:
Maas ich Economic and Social Resea ch Ins i u e on Inno a ion and Technology (UNU-MERIT),
Uni ed Na ions Uni e si y (UNU)
Sugges ed Ci a ion: Vi a elli, Ma co; A enas Díaz, Guille mo (2025) : New echnologies and
employmen : The s a e o he a , UNU-MERIT Wo king Pape s, No. 2025-005, Uni ed Na ions
Uni e si y (UNU), Maas ich Economic and Social Resea ch Ins i u e on Inno a ion and Technology
(UNU-MERIT), Maas ich ,
h ps://doi.o g/10.53330/KFFP5948
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/326934
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by-nc-sa/4.0/
#2025-005
New echnologies and employmen : he s a e o he a
Ma co Vi a elli and Guille mo A enas Díaz
Published 10 Feb ua y 2025
DOI: h ps://www.doi.o g/10.53330/KFFP5948
Maas ich Economic and social Resea ch ins i u e on Inno a ion and Technology (UNU-MERIT)
email: in o@me i .unu.edu | websi e: h p://www.me i .unu.edu
Boschs aa 24, 6211 AX Maas ich , The Ne he lands
Tel: (31) (43) 388 44 00
UNU-MERIT Wo king Pape s
ISSN 1871-9872
Maas ich Economic and social Resea ch Ins i u e on Inno a ion and Technology
UNU-MERIT
UNU-MERIT Wo king Pape s in end o dissemina e p elimina y esul s o esea ch ca ied
ou a UNU-MERIT o s imula e discussion on he issues aised.
h ps://c ea i ecommons.o g/licenses/by-nc-sa/4.0/
1
New echnologies and employmen : he s a e o he a
Ma co Vi a ellia and Guille mo A enas Díazb
a Depa men o Economic Policy, Ca holic Uni e si y o he Sac ed Hea h, Milano, I aly;
UNU-MERIT, Maas ich , The Ne he lands; IZA, Bonn, Ge many. Co esponding au ho :
[email p o ec ed]
b Depa men o Economic Policy, Ca holic Uni e si y o he Sac ed Hea h, Milano, I aly.
Abs ac
The ela ionship be ween echnology and employmen has long been a opic o deba e. This
issue is e en mo e pe inen oday as he global economy unde goes a echnological e olu ion
d i en by au oma ion and he widesp ead adop ion o A i icial In elligence. The p ima y
objec i e o his pape is o p o ide insigh s in o he ela ionship be ween inno a ion and
employmen by p oposing a concep ual amewo k and by discussing he s a e o he a o he
deba es and analyses su ounding his opic.
JEL classi ica ion: O33
Keywo ds: Technology, employmen , compensa ion heo y, AI, obo
Acknowledgemen s and disclaime s
The au ho s acknowledge he suppo by he I alian Minis e o dell’Is uzione, dell’Uni e si à e
della Rice ca (PRIN-2022, p ojec 2022P499ZB: “Inno a ion and labo ma ke dynamics”;
p incipal in es iga o : Ma co Vi a elli; unded by he Eu opean Union - Nex Gene a ion EU,
Mission 4, Componen 2, CUP: J53D23004830008 ). The iews and opinions exp essed a e only
hose o he au ho s and do no necessa ily e lec hose o he Eu opean Union o he Eu opean
Commission. Nei he he Eu opean Union no he Eu opean Commission can be held
esponsible o hem.
2
1. In oduc ion
The ela ionship be ween echnology and employmen has always been a “ho ” opic bo h o
social scien is s and policy make s, a leas since he i s indus ial e olu ion. Indeed, claims o
echnologically caused unemploymen end o e-eme ge a imes o adical echnological change
such as coun ies a e cu en ly expe iencing, acing he a i al o au oma ion and a i icial
in elligence (AI) echnologies.
Today, deba e ocuses on h ee main ques ions: Wha a e he oles o echnology and inno a ion
in explaining he long- e m declining end o manu ac u ing as a sha e o he mode n economy?
A e new echnologies, such as obo s and a i icial in elligence, eplacing humans? A e job losses
due o he ad en o obo s and AI s uc u al and he e o e ine i able?
Con ex ualizing, McKinsey (2017) o ecas s ha nea ly 50% o wo k ac i i ies could be
au oma ed by 2055. Speci ic sec o s, such as “Accommoda ion and Food Se ices” (66%),
“Manu ac u ing” (64%), and “T anspo a ion and Wa ehousing” (60%), a e pa icula ly
suscep ible o au oma ion (Figu e 1).
A mo e ecen epo om Goldman Sachs (2023) es ima es ha 25% o cu en jobs in he
Uni ed S a es and 24% in he Eu opean Union could be au oma ed. In he U.S., indus ies mos
exposed o AI include “O ice and Adminis a i e Suppo ” (46%), “Legal” (44%), and
“A chi ec u e and Enginee ing” (37%), while sec o s such as “Building and G ounds Cleaning
and Main enance” (1%), “Ins alla ion, Main enance, and Repai ” (4%), and “Cons uc ion and
Ex ac ion” (6%) a e among he leas exposed (Figu e 2).
This somewha pessimis ic ou look has gained signi ican a en ion among schola s o explo e
he po en ial impac s o new echnologies on he labo ma ke . New s a is ical ools and machine
lea ning me hods a e enabling esea che s o analyze mo e g anula ly which echnologies a ec
speci ic jobs and asks. Fo ins ance, Fel en e al. (2018) and Fel en e al. (2021) ound ha
whi e-colla wo ke s in he U.S. a e mo e exposed o AI-d i en au oma ion.
Con e sely, Webb (2020) shows ha obo s p ima ily a ec low-wage occupa ions, so wa e
al e s medium-wage occupa ions, and high-wage occupa ions a e mos ulne able o AI. Mo e
ecen ly, Mon obbio e al. (2023) ound ha low-wage jobs concen a ed in p oduc ion a e
pa icula ly exposed o obo ic labo -sa ing echnologies, especially in ins alla ion and
main enance oles. Thei s udy also no es ha se ice-based ac i i ies, such as hose pe o med
by logis ic and heal hca e wo ke s, a e inc easingly exposed o obo ic echnologies.
Howe e , he impac o inno a ion on employmen is no i ial, and i equi es unde s anding
he all-possible heo e ical mechanisms in ol ed in his ela ionship: labo -c ea ing (mainly om
p oduc inno a ion), labo -sa ing (mainly om p ocess inno a ion), and he so-called ma ke
compensa ion mechanisms, po en ially able o coun e balance he ini ial labo -sa ing impac o
inno a ion (see Sec ion 2.2).
3
Figu e 1. Po en ial au oma ion by 2055 in di e en sec o s
Sou ce: Ox o d Economic Fo ecas ing; US Bu eau o Labo S a is ics; McKinsey analysis (2017)
Figu e 2. Sha e o he indus y employmen exposed o au oma ion by AI: US
Sou ce: aking om Goldman Sachs’ epo (2023). Goldman Sachs Global In es men Resea ch
66
64
60
54
54
50
50
45
44
44
44
43
42
41
41
39
38
37
34
010 20 30 40 50 60 70
Accommoda ion and ood se ices
Manu ac u ing
T anspo a ion and wa ehousing
Mining
Re ail ade
Ag icul u e, o es y, ishing and hun ing
Wholesale ade
U ili ies
Finance and insu ance
Cons uc ion
Real es a e and en al and leasing
O he se ices (excep ede al, s a e and local…
A s, en e ainmen , and ec ea ion
In o ma ion
Adminis a i e and suppo and go e nmen
P o essional, scien i ic, and echnical se ices
Heal h ca e and social assis ance
Managemen o companies and en e p ises
Educa ional se ices
Pe cen age (%)
46
44
37
36
35
33
32
31
29
28
28
28
27
26
26
25
19
12
11
9
6
4
1
0 5 10 15 20 25 30 35 40 45 50
O ice and Adminis a i e suppo
Legal
A chi ec u e and Enginee ing
Li e, Physical, and Social Science
Business and Financial Ope a ions
Communi y and Social Se ice
Managemen
Sales and Rela ed
Compu e and Ma hema ical
Fa ming, Fishing, and Fo es y
P o ec i e Se ice
Heal hca e P ac i ione s and Technical
Educa ional Ins uc ion and Lib a y
Heal hca e Suppo
A s, Design, En e ainmen , Spo s and Media
All indus ies
Pe sonal Ca e and Se ice
Food P epa a ion and Se ing Rela ed
T anspo a ion and Ma e ial Mo ing
P oduc ion
Cons uc ion and Ex ac ion
Ins alla ion, Main enance, and Repai
Building and G ounds Cleaning and Main enance
Pe cen age (%)
4
In mo e de ail, while p ocess inno a ion can be job-des oying, p oduc inno a ion can imply
he eme gence o new i ms, new sec o s, and hus new jobs. Bu e en o p ocess inno a ion,
he inal impac on labo demand is shaped by ma ke mechanisms ha can compensa e o he
di ec job- des oying impac , i ma ke and ins i u ional igidi ies do no impede hem.
Fu he mo e, a Schumpe e ian ision is essen ial o be e unde s anding he ela ionship
be ween inno a ion and employmen . Schumpe e (1939, 1947) a gues ha echnical
unemploymen a ises om dispa i ies be ween he skills and abili ies o wo ke s displaced om
old sec o s and hose equi ed by eme ging ones. This iew - ocusing on “c ea i e des uc ion”
- emphasizes ha while some jobs disappea , new oles a e simul aneously c ea ed (Díaz,
Gue e o, e al., 2024).
Addi ionally, as highligh ed by Dosi e al. (2021), inno a ion should no be iewed as an
exogenous o isola ed phenomenon. Ins ead, i in luences he en i e socio-economic sys em. In
o he wo ds, inno a ion no only impac s he i ms o sec o s ha in oduces i , bu also migh
a ec ela ed i ms and indus ies. Fo ins ance, a obo migh se e as a p ocess inno a ion o
downs eam sec o s while simul aneously unc ioning as a p oduc inno a ion o ups eam
sec o s (Díaz, Ba ge-Gil, e al., 2024; Dosi e al., 2021).
D awing on he p e ious li e a u e, he main objec i e o his s udy is o analyze he ela ionship
be ween inno a ion and employmen h ough a comp ehensi e concep ual amewo k,
conside ing on he one hand he possible labo -sa ing impac o echnological change, he scope
o job c ea ion on he o he hand and ocusing in pa icula on he ma ke and ins i u ional
mechanisms ha can shape he inal labo demand ou comes. To achie e his, we i s pu
o wa d an in e p e a i e heo e ical amewo k and hen we c i ically discuss he key empi ical
s udies ha ha e explo ed his ela ionship (including s udies ha ocus on new echnologies,
such as obo ics and a i icial in elligence, o p o ide a comp ehensi e unde s anding o how
hese ecen inno a ions impac employmen ).
The emainde o he pape is s uc u ed as ollows: in Sec ion 2 we discuss he main heo e ical
mechanisms de e mining he ela ionship be ween inno a ion and employmen . Sec ion 3
discusses he empi ical li e a u e h ough a selec ion o p e ious s udies. In Sec ion 4 we
summa ize he main indings, while Sec ion 5 concludes.
2. A heo e ical amewo k
2.1 A comp ehensi e concep ualiza ion
One o he main d i e s o he long- e m deindus ializa ion end in de eloped coun ies is he
p oduc i i y gap be ween manu ac u ing and se ices. Indeed, echnological change is singled
ou as he main de e minan o he p oduc i i y imp o emen s ha en ail job losses in
manu ac u ing and ha he e o e lead o he declining sha e o indus ial employees in o al
5
employmen . Howe e , mo e ecen ly, au oma ion and AI di usion ha e made possible a simila
labo -sa ing p ospec in se ice indus ies anging om he inancial sec o s o ade and e ail.
Re e ing o he heo y pu o wa d by he economis s o inno a ion, he e a e wo basic
inno a ion inpu s: esea ch and de elopmen (R&D), which may lead o p oduc inno a ion,
and embodied echnological change, which may lead o p ocess inno a ion. R&D in es men s
a e he key inno a ion inpu in he app oach o iginally p oposed in 1979 by Z i G iliches, who
iden i ied he concep o he “knowledge p oduc ion unc ion” (G iliches, 1979).
In his unc ional ela ionship linking inno a i e inpu s o inno a i e ou pu s, i ms pu sue new
economic knowledge as an inpu in o gene a ing inno a i e ac i i ies. Indeed, a as li e a u e
has iden i ied a s ong signi ican link be ween R&D in es men , inno a ion, and p oduc i i y
gains, demons a ing ha R&D is a main d i e o echnological p og ess a mac oeconomic,
sec o al, and mic oeconomic le els (C epon e al., 1998). Meanwhile, embodied echnological
change in ol es p ocess inno a ion, o inno a ion ha is inco po a ed in in es men s in capi al
goods (machine y and equipmen , o ins ance obo s and o he au oma ion de ices) (F eeman
& Soe e, 1987).
Mo eo e , he inno a ion li e a u e sugges s ha i is mainly la ge high- ech i ms ha ely on
o mal R&D o d i e complex p oduc inno a ion, while embodied echnological change plays
a key ole in small- and medium-size i ms in mo e adi ional indus ies (Pa i , 1984).
As men ioned abo e, o he wo main d i e s o echnological change, R&D is mainly ela ed o
p oduc inno a ion, and embodied echnological change is mo e closely ela ed o p ocess
inno a ion. Howe e , in some ci cums ances, he dis inc ion be ween p oduc inno a ion and
p ocess inno a ion is ambiguous om an empi ical poin o iew (conside , o ins ance, he
di usion o ICT in he pas decades, and a i icial in elligence nowadays), and in many cases he
wo o ms o inno a ion a e in e ela ed. Mo eo e , bo h R&D and embodied echnological
change pa icipa e in mixed inno a i e ac i i ies ha en ail bo h p oduc and p ocess inno a ion.
Figu e 3 illus a es he main links be ween inno a i e inpu s, inno a i e ou pu s and hei
e en ual impac on he labo ma ke .
Ob iously enough, p ocess inno a ion and p oduc inno a ion in ol e di e en employmen
impac s (as shown in he igh panel o Figu e 3). P ocess inno a ion esul s in a di ec labo -
sa ing (job-des oying) e ec , ela ed mainly o he in oduc ion o machine y and equipmen
ha can subs i u e o labo and allow he p oduc ion o he same amoun o ou pu wi h ewe
inpu s (gene ally wo ke s). On he one hand, p oduc inno a ion can en ail a job-c ea ing e ec
h ough he eme gence o new indus ies and new ma ke s. Howe e , on he o he hand, he
same inno a ion can play he ole o p oduc inno a ion in a gi en sec o (supply side) and he
ole o p ocess inno a ion in ano he indus y (demand/adop ion side). Fo example, he design
and implemen a ion o a new AI algo i hm is a p oduc inno a ion in he supplie indus ies and
may en ail job c ea ion (e.g. an inc ease in he demand o da a scien is s). Howe e , he same
algo i hm may imply job losses when is adop ed in he use sec o s as a p ocess inno a ion (e.g.
a d op in he demand o bank cle ks).
6
Figu e 3. The wo aces o inno a ion: how p oduc and p ocess inno a ion a ec employmen
Sou ce
: Au ho ’s own illus a ion.
2.2 The labo ma ke implica ions o p ocess inno a ion and he compensa ion mechanisms
Since by de ini ion p ocess inno a ion means p oducing he same amoun o ou pu wi h less
labo (and some imes o he ) inpu s, he di ec impac o p ocess inno a ion is job des uc ion
when ou pu is ixed. Howe e , economic analysis has demons a ed he exis ence o
coun e ailing economic o ces ha can compensa e o he educ ion in employmen a ising
om echnological p og ess. Indeed, he classical economis s pu o wa d a heo y ha Ma x
la e called he “compensa ion heo y”(Pian a, 2005; Vi a elli, 1995, 2013, 2014). These
compensa ion mechanisms include new machine y, lowe p ices, new in es men s, and lowe
wages.
2.2.1 The compensa ion mechanism ia new machine y
The e ec o he in oduc ion o new machine y ( o ins ance obo s) is ambiguous. On he one
hand, p ocess inno a ions displace wo ke s in downs eam indus ies ha in oduce he
embodied echnological change inco po a ed in he new capi al goods. On he o he hand,
addi ional wo ke s a e needed in he ups eam indus ies ha p oduce he new machine y.
Howe e , he e a e a leas h ee a gumen s agains he e icacy o his compensa ion
mechanism. Fi s , o he in oduc ion o he new machine y o be p o i able, he cos o labo
associa ed wi h he cons uc ion o he new machine y has o be lowe han he cos o labo
displaced by he new capi al goods. Second, labo -sa ing echnologies sp ead o he capi al goods
sec o as well as o he p oduc sec o , so his compensa ion can be an endlessly epea ing s o y,
13
s a is ical signi icance depend on a ious ci cums ances ( o ins ance, de eloping s. de eloped
coun ies, sec o s, pe iod o c isis, di e en me hodologies). Indeed, only ew s udies ound ou
a labo -sa ing impac o p ocess inno a ion (Díaz e al., 2020; Lim & Lee, 2019). One o he
main c i ique add essable o his bunch o s udies is ha he dummy a iable “sole p ocess
inno a ion“ ails o ully cap u e i m’s p ocess inno a ion s a egy, i s ac ual size and i s
a iabili y (Díaz, Gue e o, e al., 2024).
O he s udies ha e no adop ed he wo main app oaches men ioned abo e. Fo ins ance, a
s udy - using a dynamic employmen model and a longi udinal da a se on Ge man
manu ac u ing i ms o e he pe iod 1982–2002 - has ound a signi ican ly posi i e impac o
a ious cu en and pas p oduc and p ocess inno a ion a iables on labo demand
(Lachenmaie & Ro mann, 2011). Acco ding o his wo k, inno a ion is homogeneously
employmen iendly.
Mo e ecen s udies ha e used di e en ypes o measu es o inno a ion. A s udy ha used
pa en s as a p oxy o inno a ion o 20,000 Eu opean companies om 2003 o 2012 ound a
posi i e impac o inno a ion on employmen , bu only o i ms in high- ech manu ac u ing
sec o s (Van Roy e al., 2018). Ano he s udy om Spain om 1991 o 2012 ound a posi i e
e ec o p oduc inno a ion on employmen g ow h and no signi ican impac o p ocess
inno a ion (bo h using dummy a iables as p oxies o inno a ion) (Bianchini & Pelleg ino,
2019). A mos ecen s udy used he En e p ise su eys da ase om he Wo ld Bank and ound
ha R&D expendi u e and p ocess inno a ion os e i m’s employmen g ow h (Goel &
Nelson, 2022).
3.4 Empi ical e idence o speci ic echnologies: obo s and a i icial in elligence
The eme gence o he cu en new echnological pa adigm has gene a ed a desi e o explo e he
empi ical e ec o speci ic echnologies (namely obo s and a i icial in elligence) on he labo
ma ke .
In he case o obo s, s udies a he indus ial le el ha used da a om he In e na ional
Fede a ion o Robo ics and EUKLEMS o de eloped coun ies ha e ound a nega i e e ec
on employmen (speci ically o low-skilled wo ke s and in se ices sec o s) (Acemoglu &
Res epo, 2020a; Chiacchio e al., 2018; G ae z & Michaels, 2018).
In con as , mos s udies a he i m le el ound posi i e impac s o obo s on employmen
(mainly in coun ies such as F ance, Spain, Canada, and Ge many) (Dau h e al., 2021; Dixon
e al., 2021; Domini e al., 2021; Koch e al., 2021). Howe e , op imis ic employmen esul s
ob ained a he i m le el o analysis can be en i ely due o he “business s ealing e ec ” (see
abo e) and job c ea ion a he i m le el can well coexis wi h job des uc ion a he indus y
le el (Acemoglu e al., 2020).
New empi ical me hodologies (such as na u al language p ocesses and ex analyses) allow o
explo e o he sou ces o in o ma ion (e.g., job pos s and pa en s). These ypes o s udies analyze
he exposu e and he impac o a i icial in elligence ( obo s) on he labo ma ke . Also, hese
14
ypes o s udies can assess he p oximi y be ween speci ic inno a ions, occupa ions, and asks.
Fo ins ance, one s udy ies o look a AI-exposed es ablishmen s by combining job pos s using
Bu ning Glass Technology da a and SOC occupa ional codes. The s udy ound no appa en
e ec a he indus y and occupa ional le els, bu i did ind a e-composi ion owa d AI-
in ensi e jobs (Acemoglu e al., 2022). O he s udies using pa en s (AI- ela ed in en ions) show
a mode a i e posi i e employmen impac o AI pa en ing wi hin he indus ies which pa en in
AI, ha is he ups eam sec o s which p o ide he new echnologies (see abo e) (Damioli e al.,
2024).
O he app oaches dis inguish be ween labo -sa ing inno a ions and labo -complemen a y
echnologies. Fo ins ance, one s udy ha uses he ex ual desc ip ion o asks in he ou h
edi ion o he Dic iona y o Occupa ion Ti les (DOT) and he b eak h ough inno a ions
( h ough pa en s) ound ha he mos exposed occupa ions expe ienced a dec ease in wage and
employmen le el (mainly whi e-colla wo ke s ela i e o blue-colla wo ke s) (Kogan e al.,
2021). Mo e ecen ly, ano he s udy iden i ied labo -sa ing inno a ions using ex ual analysis o
USPTO pa en applica ions in obo ics. The main esul s show ha some ac i i ies a e mo e
exposed o labo -sa ing inno a ion, such as hose ela ed o anspo , s o age, packaging, and
mo ing objec s. Along he same line, an upda e o he p e ious s udy shows ha occupa ions
mos exposed o obo ic labo -sa ing echnologies a e associa ed wi h lowe employmen and
wage a es (Mon obbio e al., 2022, 2023).
4. Main indings
Theo e ical models canno claim o ha e a clea answe on he inal employmen impac o
p ocess and p oduc inno a ion.
While he p ice and income mechanisms desc ibed he e ha e he po en ial o compensa e, ully
o in pa , he di ec labo -sa ing impac o p ocess inno a ion, he p ecise ou come is unce ain.
De e mining ac o s include such a iables as he deg ee o compe i ion, demand elas ici y,
elas ici y o subs i u ion be ween capi al and labo , and expec a ions o consume s and
employe s. O e all, depending on ma ke s uc u e and ins i u ional con ex s, compensa ion
mechanisms can be mo e o less e ec i e, and he unemploymen impac o p ocess inno a ion
can be o ally, pa ially, o no a all neu alized.
Simila ly, he indings o empi ical s udies a e no ully conclusi e abou he possible
employmen impac o inno a ion and echnological change. Howe e , mos ecen panel
in es iga ions suppo a posi i e link. This posi i e link is especially e iden when R&D o
p oduc inno a ion a e adop ed as p oxies o echnological change and when he ocus is on
high- ech sec o s and high-g ow h i ms (Vi a elli, 2013, 2014). In many sec o s, howe e ,
especially in se ices, p oduc and p ocess inno a ion a e in e mingled and di icul o
disen angle. Mo eo e , while p ocess inno a ions display clea di ec labo -sa ing e ec s, some
p oduc inno a ions may also in ol e job displacemen . The e o e, i is no always easy and
s aigh o wa d o design indus ial and inno a ion policies ha can e ec i ely maximize he
15
posi i e employmen impac o inno a ion. Addi ional mic oeconome ic s udies o he ype
add essed by he cu en esea ch li e a u e a e needed o u he disen angle he labo impac
o inno a ion ac oss di e en sec o s and di e en ypes o i ms. Indeed, new s a is ical
echniques and sou ces o in o ma ion a e being used nowadays o cons uc di e en measu es
o labo ma ke exposu e o echnological change.
Wi h speci ic ega d o he AI echnologies, he sca ce a ailable e idence (see abo e) sugges s
ha echnological leade s wi hin he eme gence o he AI pa adigm can ealize (mode a e) labo -
iendly ou comes. Howe e , o he companies (pa icula ly in manu ac u ing) may e eal o be
unable o couple p oduc inno a ion wi h job c ea ion.
Mo eo e , compa ed wi h he labo -sa ing impac implied by he adop ion o AI and au oma ion
echnologies (massi e acco ding o some s udies, see abo e), he labo - iendly ex en in he
supply indus ies appea s limi ed in magni ude and scope (jus as a na a i e example: he hi ing
o da a scien is s in ups eam se ices and AI big- ech would ha dly compensa e job losses due
o obo s in downs eam manu ac u ing).
As a gap in he cu en li e a u e, much is needed in e ms o addi ional empi ical e idence able
o compa e he ac ual magni ude o possible employmen complemen a y e ec s wi hin he
p o ide s o new AI echnologies wi h he possible job-losses due o he subs i u ion e ec s
wi hin he use s o new AI and au oma ion echnologies.
Finally, one c ucial aspec is ha mos s udies ha analyze he ela ionship be ween employmen
and inno a ion ocus on de eloped coun ies. Howe e , he la ge e ec s o au oma ion and
inno a ion migh be in de eloping coun ies whe e many ac i i ies and job asks can be easily
subs i u ed by obo s and AI algo i hms ( hink abou manu ac u ing jobs displaced by obo s o
call-cen e jobs displaced by cha bo s). Wi h ew excep ions (see abo e (Goel & Nelson, 2022)),
he e is a lack o empi ical s udies o de eloping coun ies ha can p o ide mo e e idence o
hese phenomena.
5. Conclusion and policy implica ions
The li e a u e, bo h heo e ical and empi ical, has examined he main echnological d i e s ha
can play a ole in he loss o jobs and he c ea ion o echnological unemploymen . Indeed,
inno a ion a ec s he economy h ough bo h p ocess and p oduc inno a ion, bo h o which
can ha e employmen impac s. Fo he mos pa , R&D expendi u es ha esul in p oduc
inno a ion a e gene ally labo - iendly, c ea ing new jobs, while embodied echnological change
ha esul s in p ocess inno a ion is gene ally job-des oying. A clea policy implica ion would
seem o be ha economic policy should y o os e job c ea ion by suppo ing R&D
in es men s and p oduc inno a ion. In he AI e a his means o os e eme ging indus ies and
inno a i e s a ups, ac i e in AI design, enginee ing and pa en ing.
Howe e , he pic u e is mo e complica ed han ha . P oduc and p ocess inno a ion a e o en
in e ela ed, and p ocess inno a ion does no always lead o job des uc ion. Indeed, much o
16
he heo e ical li e a u e on he employmen impac o echnology has ocused on a ious ma ke
compensa ion mechanisms ha can coun e ac mos i no all o he echnological
unemploymen impac s o p ocess inno a ion (see he classical compensa ion heo y discussed
abo e, and i s ecen e i al pu o wa d by Acemogulglu and Res epo).
Thus, a gene al heo e ical and empi ical conclusion is ha compensa ion mechanisms a e always
a wo k bu ha he ull eabso p ion o wo ke s dismissed as a esul o echnological change
canno be assumed ex an e. In pa icula , o wo k p ope ly, compensa ion mechanisms equi e
compe i ion ( o acili a e he compensa ion mechanism ha wo ks h ough lowe p ices),
op imis ic expec a ions ( o acili a e he compensa ion mechanisms ha wo k h ough lowe
p ices and new in es men s), and a high elas ici y o subs i u ion be ween capi al and labo . In
his amewo k, compe i ion policies ha lowe en y ba ie s and educe monopolis ic en s,
along wi h expansiona y policies a ge ing in e media e and inal demand o new p oduc s, can
be impo an d i e s o job c ea ion. In his espec , he concen a ion o AI esea ch and
pa en ing in he hands o he “big ech” is ex emely wo ying and should be con as ed by a
ie ce an i us policy.
Since economic heo y o e s no clea -cu answe on he employmen e ec o inno a ion,
answe s need o come om empi ical analyses. Empi ical s udies can conside di e en o ms
o echnological change, hei di ec e ec s on employmen , a ious compensa ion mechanisms
a wo k, and any possible impedimen s o hese mechanisms.
In pa icula , mic oeconome ic s udies ha e he g ea ad an age o enabling di ec and p ecise
i m-le el mapping o inpu and ou pu inno a ion a iables (Vi a elli, 2013, 2014) . O e all, he
empi ical li e a u e, pa icula ly he mos ecen mic oeconome ic panel da a analyses, ends o
suppo a posi i e link be ween echnological ad ances and employmen , especially when he
ocus is on R&D, p oduc inno a ion and ups eam high- ech i ms. These posi i e employmen
ou comes o e idence-based s udies a e consis en wi h a li ecycle iew o di e en indus ies,
wi h eme ging sec o s cha ac e ized by p oduc inno a ion (mos ly labo - iendly) and mo e
adi ional, ma u e indus ies mo e likely o expe ience p ocess inno a ion (mos ly labo -sa ing).
As a policy implica ion, policy make s should os e he eme gence and he s eng hening o
ups eam AI-in ensi e indus ies, whe e he job c ea ion impac is concen a ed (see abo e).
Howe e , while suppo ing R&D in es men s and p omo ing knowledge and AI-in ensi e
indus ies (An onelli e al., 2023), can be a mean o os e ing compe i i eness, economic g ow h
and job c ea ion, bo h indus ial policies and inno a ion policies need ca e ully o ake in o
accoun a se ies o complex in e ac ions be ween p ocess inno a ion and p oduc inno a ion,
be ween ma u e sec o s and new sec o s, and be ween job-c ea ion e ec s in he ups eam
indus ies and job-des uc ion e ec s in he downs eam indus ies (see abo e). These complex
in e ela ionships, di icul o p edic in ad ance, highligh he need o a con inuous moni o ing
o policy implemen a ion. Fo ins ance, sa e y ne s and ac i e labo ma ke policies a e necessa y
o deal wi h he employmen displacemen due o he widesp ead di usion o AI and au oma ion
echnologies in he use indus ies.
17
Finally, as discussed abo e, i is u gen o gene a e mo e empi ical analysis o know he ac ual
e ec s o new echnologies on he labo ma ke s ou side Eu ope and he Uni ed S a es. Indeed,
he mos ulne able coun ies o new echnological changes a e hose ha execu e ou ine asks
( o ins ance, assembly plan s o adi ional se ices), and hose a e usually de eloping coun ies.
Policymake s' ision should conside his global e ec o echnological change on he labo
ma ke s, pa icula ly wi hin in e na ional o ganiza ions such as he UN, Wo ld Bank, IMF, e c.
Re e ences
Acemoglu, D., Au o , D., Hazell, J., & Res epo, P. (2022). A i icial In elligence and Jobs: E idence om
Online Vacancies. Jou nal o Labo Economics, 40(S1), S293–S340. h ps://doi.o g/10.1086/718327
Acemoglu, D., LeLa ge, C., & Res epo, P. (2020). Compe ing wi h obo s: i m-le el e idence om F ance (Vol.
2011, Issue 26738). h p://www.nbe .o g/pape s/w26738
Acemoglu, D., & Res epo, P. (2018a). A i icial In elligence, Au oma ion and Wo k. In A. Ag awal, J. S.
Gans, & A. Gold a b (Eds.), The economics o a i icial in elligence: An agenda (pp. 197–236). Uni e si y
o Chicago P ess. h ps://doi.o g/10.2139/ss n.3098384
Acemoglu, D., & Res epo, P. (2018b). The Race be ween Man and Machine: Implica ions o Technology
o G ow h, Fac o Sha es, and Employmen . Ame ican Economic Re iew, 108(6), 1488–1542.
h ps://doi.o g/10.1257/ae .20160696
Acemoglu, D., & Res epo, P. (2020a). Robo s and jobs: E idence om us labo ma ke s. Jou nal o Poli ical
Economy, 128(6), 2188–2244. h ps://doi.o g/10.1086/705716
Acemoglu, D., & Res epo, P. (2020b). The w ong kind o AI? A i icial in elligence and he u u e o
labou demand. Camb idge Jou nal o Regions, Economy and Socie y, 13(1), 25–35.
h ps://doi.o g/10.1093/cj es/ sz022
An onelli, C., O sa i, G., & Pialli, G. (2023). The knowledge-in ensi e di ec ion o echnological change.
Eu asian Business Re iew, 13(1), 1–27. h ps://doi.o g/10.1007/s40821-022-00234-z
Ba bie i, L., Pi a, M., & Vi a elli, M. (2020). R&D, embodied echnological change, and employmen :
e idence om I alian mic oda a. Indus ial and Co po a e Change, 28(1), 203–218.
h ps://doi.o g/10.1093/icc/d aa001
Bianchini, S., & Pelleg ino, G. (2019). Inno a ion pe sis ence and employmen dynamics. Resea ch Policy,
48(5), 1171–1186. h ps://doi.o g/10.1016/j. espol.2018.12.008
Bogliacino, F., Pi a, M., & Vi a elli, M. (2012). R&D and employmen : An applica ion o he LSDVC
es ima o using Eu opean mic oda a. Economics Le e s, 116(1), 56–59.
h ps://doi.o g/10.1016/j.econle .2012.01.010
Bue ge , M., B oekel, T., & Coad, A. (2012). Regional Dynamics o Inno a ion: In es iga ing he Co-
e olu ion o Pa en s, Resea ch and De elopmen (R&D), and Employmen . Regional S udies, 46(5),
565–582. h ps://doi.o g/10.1080/00343404.2010.520693
18
Chiacchio, F., Pe opoulos, G., & Pichle , D. (2018). The Impac o Indus ial Robo s on EU
Employmen and Wages: A Local Labou Ma ke App oach. In McKinsey Global Ins i u e (Vol. 10,
Issue Ma ch).
Co oche , N., Moschella, D., S accioli, J., & Vi a elli, M. (2024). Inno a ion and he labo ma ke :
heo y, e idence, and challenges. Indus ial and Co po a e Change, 33(3), 519–540.
h ps://doi.o g/10.1093/icc/d ad066
C epon, B., Dugue , E., & Mai essec, J. (1998). Resea ch, Inno a ion And P oduc i i y: An Econome ic
Analysis A The Fi m Le el. Economics o Inno a ion and New Technology, 7(2), 115–158.
h ps://doi.o g/10.1080/10438599800000031
Damioli, G., Roy, V. Van, Vé esy, D., & Vi a elli, M. (2024). D i e s o employmen dynamics o AI
inno a o s. Technological Fo ecas ing and Social Change, 201.
h ps://doi.o g/10.1016/j. ech o e.2024.123249
Dau h, W., Findeisen, S., Suedekum, J., & Woessne , N. (2021). The Adjus men o Labo Ma ke s o
Robo s. Jou nal o he Eu opean Economic Associa ion, 19(6), 3104–3153.
h ps://doi.o g/10.1093/jeea/j ab012
Díaz, G. A., Ba ge-Gil, A., & Heijs, J. (2020). The e ec o inno a ion on skilled and unskilled wo ke s
du ing bad imes. S uc u al Change and Economic Dynamics, 52, 141–158.
h ps://doi.o g/10.1016/j.s ueco.2019.09.012
Díaz, G. A., Ba ge-Gil, A., Heijs, J., & Ma zucchi, A. (2024). The e ec o ex e nal inno a ion on i m
employmen . Economics o Inno a ion and New Technology.
h ps://doi.o g/10.1080/10438599.2024.2303051
Díaz, G. A., Gue e o, A. J., & Heijs, J. (2024). The e ec s o p oduc and p ocess inno a ion on
employmen : a me a- eg ession analysis. Eu asian Business Re iew, 14(1), 35–68.
h ps://doi.o g/10.1007/s40821-024-00259-6
Dixon, J., Hong, B., & Wu, L. (2021). The obo e olu ion: Manage ial and employmen consequences
o i ms. Managemen Science, 67(9), 5586–5605. h ps://doi.o g/10.1287/mnsc.2020.3812
Domini, G., G azzi, M., Moschella, D., & T eibich, T. (2021). Th ea s and oppo uni ies in he digi al
e a: Au oma ion spikes and employmen dynamics. Resea ch Policy, 50(7).
h ps://doi.o g/10.1016/j. espol.2020.104137
Dosi, G., Pi a, M., Vi gilli o, M. E., & Vi a elli, M. (2021). Embodied and disembodied echnological
change: The sec o al pa e ns o job-c ea ion and job-des uc ion. Resea ch Policy, 50(4), 104199.
h ps://doi.o g/10.1016/j. espol.2021.104199
Feldmann, H. (2013). Technological unemploymen in indus ial coun ies. Jou nal o E olu iona y
Economics, 23(5), 1099–1126. h ps://doi.o g/10.1007/s00191-013-0308-6
Fel en, E., Raj, M., & Seamans, R. (2018). A Me hod o Link Ad ances in A i icial In elligence o
Occupa ional Abili ies. AEA Pape s and P oceedings, 108, 54–57.
h ps://doi.o g/10.1257/pandp.20181021
19
Fel en, E., Raj, M., & Seamans, R. (2021). Occupa ional, indus y, and geog aphic exposu e o a i icial
in elligence: A no el da ase and i s po en ial uses. S a egic Managemen Jou nal, 42(12), 2195–2217.
h ps://doi.o g/10.1002/smj.3286
F eeman, C., & Soe e, Luc. (1987). Fac o subs i u ion and echnical change. In Technical change and ull
employmen (pp. 2183–2187). Basil Blackwell.
Goel, R. K., & Nelson, M. A. (2022). Employmen e ec s o R&D and p ocess inno a ion: e idence
om small and medium-sized i ms in eme ging ma ke s. Eu asian Business Re iew, 12(1), 97–123.
h ps://doi.o g/10.1007/s40821-022-00203-6
Goldman Sachs. (2023). The Po en ially La ge E ec s o A i icial In elligence on Economic G ow h (B iggs/Kodnani).
h ps://www.gspublishing.com/con en / esea ch/en/ epo s/2023/03/27/d64e052b-0 6e-
45d7-967b-d7be35 abd16.h ml
G ae z, G., & Michaels, G. (2018). Robo s a wo k. Re iew o Economics and S a is ics, 100(5), 753–768.
h ps://doi.o g/10.1162/ es _a_00754
G iliches, Z. (1979). Issues in Assessing he Con ibu ion o Resea ch and De elopmen o P oduc i i y
G ow h. The Bell Jou nal o Economics, 10(1), 92. h ps://doi.o g/10.2307/3003321
Ha ison, R., Jaumand eu, J., Mai esse, J., & Pe e s, B. (2014). Does inno a ion s imula e employmen ?
A i m-le el analysis using compa able mic o-da a om ou Eu opean coun ies. In e na ional
Jou nal o Indus ial O ganiza ion, 35(1), 29–43. h ps://doi.o g/10.1016/j.ijindo g.2014.06.001
Ka soulacos, Y. S. (1986). The employmen e ec o echnical change: a heo e ical s udy o new echnology and he labou
ma ke . Whea shea .
Koch, M., Manuylo , I., & Smolka, M. (2021). Robo s and Fi ms. Economic Jou nal, 131(638), 2553–2584.
h ps://doi.o g/10.1093/ej/ueab009
Kogan, L., Papanikolaou, D., D. W. Schmid , L., & Seegmille , B. (2021). Technology-Skill Complemen a i y
and Labo Displacemen : E idence om Linking Two Cen u ies o Pa en s wi h Occupa ions (Issue 29552).
h ps://doi.o g/h ps://doi.o g/10.3386/w29552
Lachenmaie , S., & Ro mann, H. (2011). E ec s o inno a ion on employmen : A dynamic panel
analysis. In e na ional Jou nal o Indus ial O ganiza ion, 29(2), 210–220.
h ps://doi.o g/10.1016/j.ijindo g.2010.05.004
Lim, J., & Lee, K. (2019). Employmen e ec o inno a ion unde di e en ma ke s uc u es: Findings
om Ko ean manu ac u ing i ms. Technological Fo ecas ing and Social Change, 146, 606–615.
h ps://doi.o g/10.1016/j. ech o e.2018.04.022
McKinsey. (2017). A u u e ha wo ks: Au oma ion, employmen , and p oduc i i y. Mckinsey Global Ins i u e.
h p://nji 2.m ooms.ne /plugin ile.php/688844/mod_ esou ce/con en /1/Execu i e
Mon obbio, F., S accioli, J., Vi gilli o, M. E., & Vi a elli, M. (2022). Robo s and he o igin o hei labou -
sa ing impac . Technological Fo ecas ing and Social Change, 174.
h ps://doi.o g/10.1016/j. ech o e.2021.121122
20
Mon obbio, F., S accioli, J., Vi gilli o, M. E., & Vi a elli, M. (2023). Labou -sa ing au oma ion: A di ec
measu e o occupa ional exposu e. Wo ld Economy, 47(1), 332–361.
h ps://doi.o g/10.1111/ wec.13522
Pa i , K. (1984). Sec o al pa e ns o echnical change: Towa ds a axonomy and a heo y. Resea ch Policy,
13(6), 343–373. h ps://doi.o g/10.1016/0048-7333(84)90018-0
Pelleg ino, G., Pi a, M., & Vi a elli, M. (2019). Beyond R&D: he ole o embodied echnological change
in a ec ing employmen . Jou nal o E olu iona y Economics, 29(4), 1151–1171.
h ps://doi.o g/10.1007/s00191-019-00635-w
Pian a, M. (2005). Inno a ion and employmen . In J. Fage be g, D. Mowe y, & R. Nelson (Eds.), The
Ox o d Handbook o Inno a ion (pp. 568–597). Ox o d Uni e si y P ess.
h ps://o a.uniu b.i /handle/11576/1889692
Pi a, M., & Vi a elli, M. (2005). Inno a ion and Employmen : E idence om I alian Mic oda a. Jou nal
o Economics, 86(1), 65–83. h ps://doi.o g/10.1007/s00712-005-0140-z
Pi a, M., & Vi a elli, M. (2018). Technological change and employmen : is Eu ope eady o he
challenge? Eu asian Business Re iew, 8(1), 13–32. h ps://doi.o g/10.1007/s40821-017-0100-x
Schumpe e , J. A. (1939). Business cycles : a heo e ical, his o ical, and s a is ical analysis o he capi alis p ocess.
McG aw-Hill Book Co.
Schumpe e , J. A. (1947). Capi alism, Socialism and Democ acy (Second Edi). Ha pe & Row Publishe s.
Ugu , M., Chu chill, S. A., & Solomon, E. (2018). Technological inno a ion and employmen in de i ed
labou demand models: a hie a chical Me a-Reg ession Analysis. Jou nal o Economic Su eys, 32(1),
50–82. h ps://doi.o g/10.1111/joes.12187
Van Reenen, J. (1997). Employmen and echnological inno a ion: E idence om U.K. manu ac u ing
i ms. Jou nal o Labo Economics, 15(2), 255–284. h ps://doi.o g/10.1086/209833
Van Roy, V., Vé esy, D., & Vi a elli, M. (2018). Technology and employmen : Mass unemploymen o
job c ea ion? Empi ical e idence om Eu opean pa en ing i ms. Resea ch Policy, 47(9), 1762–1776.
h ps://doi.o g/10.1016/j. espol.2018.06.008
Vi a elli, M. (1995). The Economics o Technology and Employmen : Theo y and Empi ical (E. Elga , Ed.).
h ps://ideas. epec.o g/b/elg/eebook/458.h ml%22%3EThe
Vi a elli, M. (2007). Inno a ion and Employmen : A Su ey. In IZA Discussion Se ies (Issue 2621).
h p://pape s.ss n.com/sol3/pape s.c m?abs ac _id=969621
Vi a elli, M. (2013). Technology, employmen , and skills: An in e p e a i e amewo k. Eu asian Business
Re iew, 3(1), 66–89. h ps://link.sp inge .com/a icle/10.14208/BF03353818#ci eas
Vi a elli, M. (2014). Inno a ion, employmen and skills in ad anced and de eloping coun ies: A su ey
o economic li e a u e. Jou nal o Economic Issues, 48(1), 123–154. h ps://doi.o g/10.2753/JEI0021-
3624480106
Webb, M. (2020). The Impac o A i icial In elligence on he Labo Ma ke .
h ps://doi.o g/10.2139/ss n.3482150
The UNU-MERIT WORKING Pape Se ies
2025-01 De elopmen s a egies o he g een hyd ogen economy in eme ging economies
by Fabianna Bacil, An hony Black, Ma ina Domingues, Jun Jin, Rasmus Lema, Glen
Robbins and Sö en Schol in
2025-02 Do global alue chains and local capabili ies ma e o economic complexi y in EU
egions? by R. Boschma, E. He nández-Rod íguez, A. Mo ison and C. Pie obelli
2025-03 Inno a ion and p oduc i i y: he ecen empi ical li e a u e and he s a e o he a
by Jacques Mai esse, Pie e Mohnen and Ad No en
2025-04 Digi al Technologies, Hi ing, T aining, and Fi m Ou comes: E idence om AI and ICT
Adop ion in Indian i mss by Sneha Ma ydas, Nandi ha Ma hew, Gio dano De
Ma zo and Ca lo Pie obelli
2025-05 New echnologies and employmen : he s a e o he a by Ma co Vi a elli and
Guille mo A enas Díaz
