What Drives Carbon Emissions in German Manufacturing: Scale, Technique or Composition?

Ro ne , Elisa; on G ae eni z, Ka h ine
A icle — Published Ve sion
Wha D i es Ca bon Emissions in Ge man Manu ac u ing:
Scale, Technique o Composi ion?
En i onmen al and Resou ce Economics
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Ro ne , Elisa; on G ae eni z, Ka h ine (2024) : Wha D i es Ca bon Emissions in
Ge man Manu ac u ing: Scale, Technique o Composi ion?, En i onmen al and Resou ce Economics,
ISSN 1573-1502, Sp inge Ne he lands, Do d ech , Vol. 87, Iss. 9, pp. 2521-2542,
h ps://doi.o g/10.1007/s10640-024-00894-7
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/315258
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
Vol.:(0123456789)
En i onmen al and Resou ce Economics (2024) 87:2521–2542
h ps://doi.o g/10.1007/s10640-024-00894-7
1 3
Wha D i es Ca bon Emissions inGe man Manu ac u ing:
Scale, Technique o Composi ion?
ElisaRo ne 1· Ka h ine onG ae eni z2
Accep ed: 20 June 2024 / Published online: 16 July 2024
© The Au ho (s) 2024
Abs ac
Emissions o local pollu an s om indus y ha e declined ac oss many de eloped coun ies
o e he las decades. Fo ca bon emissions such educ ions ha e ye o ma e ialize. Using
Ge man adminis a i e mic o-da a a he p oduc le el, we apply he wo kho se model o
decomposing emission changes in o scale, composi ion (changes in he mix o goods p o-
duced) and echnology (emission ac o s o p oduc ion) e ec s. We ind ha he p oduc ion
composi ion in Ge man manu ac u ing shi ed owa ds less CO2 in ensi e goods, while
emission in ensi ies o p oduc ion inc eased. We show ha da a agg ega ion ma e s. Bo h
e ec s a e subs an ially unde es ima ed when using da a a he sec o le el as compa ed o
he p oduc le el. A complemen a y plan le el decomposi ion e eals ha emission in en-
si ies o p oduc ion inc eased bo h wi hin plan , and due o a ealloca ion o p oduc ion o
mo e emission in ensi e plan s.
Keywo ds Ca bon emissions· Clima e policy· S a is ical decomposi ion· Manu ac u ing
We a e g a e ul o wo excellen e e ees who helped us imp o e he pape subs an ially. We hank
Sylwia Bialek, Clai e B unel, And eas Ge s e , A i G o e , Bea Hin e mann, A ik Le inson, Dominik
Schobe , Bodo S u m and Ul ich Wagne o sugges ions and insigh ul commen s. We also hank
semina pa icipan s a he ZEW–Leibniz-Cen e o Eu opean Economic Resea ch, he 9 h Mannheim
Con e ence on Ene gy and he En i onmen , he Eu opean Associa ion o En i onmen al and Resou ce
Economis s, he annual mee ing o he Associa ion o Ge manspeaking Economis s (V S), he AURÖ
junio wo kshop and he Wo ld Bank b own bag se ies G eening P i a e and Financial Sec o . We
g a e ully acknowledge he Resea ch Da a Cen e (FDZ) o he Fede al S a is ical O ice and he
S a is ical O ices o he Ge man Lände o g an ing us access o he AFiD da a and o he use o
hei esea ch acili ies, in pa icula Denise Henke , S e an Sei z, Ke s in S ockmaye and Diane Zabel
o hei ad ice and echnical suppo . We hank he Fede al Minis y o Educa ion and Resea ch
(BMBF) o he inancial suppo h ough he P ojec TRACE (G an numbe 01LA1815A). The iews
exp essed in his pape a e hose o he au ho s and do no necessa ily ep esen hose o he ins i u ions
men ioned abo e.
* Ka h ine on G ae eni z
ong ae eni [email protected]
Elisa Ro ne
o ne @zew.de
1 ZEW - Leibniz Cen e o Eu opean Economic Resea ch andUni e si y o Basel, P.O. Box10 34
43, 68034Mannheim, Ge many
2 ZEW - Leibniz Cen e o Eu opean Economic Resea ch andUni e si y o Mannheim, P.O.
Box10 34 43, 68034Mannheim, Ge many
2522
E.Ro ne , K. on G ae eni z
1 3
JEL Classi ica ion D22· L60· Q41· Q48
1 In oduc ion
En i onmen al policy o imp o e ai quali y has been ema kably e ec i e. Local pollu an
emissions ha e been egula ed in many coun ies since he 1970s and 1980s. In con as ,
we ha e ye o make subs an ial p og ess on he educ ion o CO2 emissions. Regula ion o
CO2 emissions is much mo e ecen , and no as widesp ead e en among de eloped econo-
mies. Fo an indi idual coun y, he di ec bene i s o educing ca bon emissions a e less
clea han he bene i s o educing local pollu an s, since damages depend on emissions
globally and a y ac oss coun ies. Pe haps in consequence o less s ingen egula ion,
end-o -pipe echnologies o CO2 emo al a e no ye comme cially iable, and emissions
emain closely linked o ac i i y le els inc easing he pe cei ed cos s o egula ion.
The indus ial sec o accoun s o he lion’s sha e o global g eenhouse gas emissions,
oge he wi h he powe sec o (IPCC 2014). In his sense, ca bon emissions a e simila o
he case o local pollu an s. Elec ici y gene a ion has become cleane o e ime h ough
expansion o enewable elec ici y gene a ion om wind and sola ins alla ions. Fo indus-
y, a educ ion in g eenhouse gas emissions, and, mos p ominen ly, CO2 has no ma e i-
alized, e en in coun ies whe e clima e policies a e implemen ed and a e compa a i ely
s ic om a global poin o iew: In Ge many, ca bon emissions om manu ac u ing ha e
inc eased in ecen yea s and we e abou 32 million onnes highe in 2017 han in 2003
( on G ae eni z and Ro ne 2023). The lack o p og ess in absolu e e ms, e en in such an
indus ialized economy wi h clima e policies in place, is p oblema ic—especially i coun-
ies like Ge many a e o se e as a bluep in o deca boniza ion wi hou deindus ializa-
ion in o he coun ies.
Wha d i es he de elopmen o CO2 emissions in he Ge man manu ac u ing sec o ?
Answe ing his ques ion is c ucial o de e mine app op ia e coun e measu es. A e emission
inc eases solely g ounded in a ising p oduc ion scale? Do we ha e o wo y abou he ou -
sou cing o di y p oduc s and ca bon leakage (p oduc ion composi ion)? A e he e ech-
nological imp o emen s ha migh be expo ed o o he coun ies and acili a e emission
educ ions he e as well (p oduc ion echnique)? Decomposi ion me hods a e a use ul ool
o disen angle hese emission d i e s and ha e been widely used o local pollu an s in he
en i onmen al communi y (B unel 2017; Le inson 2009, 2015; Najja and Che niwchan
2021; Shapi o and Walke 2018).
Due o limi ed da a a ailabili y, pas esea ch has conduc ed Le inson-s yle s a is ical
decomposi ions a he sec o al le el. Using ine-g ained da a a he p oduc le el dis in-
guishing be ween mo e han 4600 9-digi p oduc s, we decompose he de elopmen o ca -
bon emissions in Ge man manu ac u ing. We show ha da a agg ega ion ma e s: Gene -
ally, we ind ha he Ge man manu ac u ing sec o has been shi ing owa ds a cleane
p oduc composi ion be ween 2005 and 2017. The clean up due o composi ional change
is 9.4% in ou 9-digi baseline decomposi ion, bu only 6.8% in he 3-digi decomposi ion
(71%). Con e sely, emission in ensi ies o Ge man manu ac u ing (p oduc ion echnique)
ha e inc eased be ween 2005 and 2017. A he 9-digi le el, his inc ease amoun s o 4.5%,
bu only 2.3% a he 3-digi le el (51%). The inc ease in p oduc ion scale and he posi i e
echnique e ec combined a e he easons why ca bon emissions in Ge man indus y ha e
no dec eased.
2523
Wha D i es Ca bon Emissions inGe man Manu ac u ing: Scale,…
1 3
The de ailed mic o-da a a ailable om he Ge man Manu ac u ing Census allow us o
combine he Le inson-s yle decomposi ion wi h a plan le el decomposi ion in he s yle
o Fos e e al. (2008). In doing so we ollow Ba ows and Olli ie (2018). Conduc ing
decomposi ion analyses using di e en uni s o obse a ion (p oduc e sus plan ), we can
exploi he espec i e me i s o each me hod and ob ain a ulle pic u e o he channels
h ough which emission in ensi ies o p oduc ion ha e inc eased. Ou esul s show ha
selec ion (plan en y and exi ) had negligible e ec s on he emission in ensi y o Ge man
indus y. Emission in ensi ies o p oduc ion inc eased wi hin plan . Re-alloca ion o p o-
duc ion be ween plan s has also led o an inc ease in emission in ensi y. The c oss e ec is
la ge and nega i e indica ing ha emission in ensi y o plan s wi h ini ially high in ensi ies
has g own ela i ely less, dampening he ne impac on emission in ensi y o he manu ac-
u ing sec o .
Wi h ou pape , we con ibu e o he li e a u e using decomposi ion ools as a means o
unde s and he d i e s o emission de elopmen s (e.g. Shapi o and Walke 2018; Le inson
2009, 2015, 2021; B unel 2017; Najja and Che niwchan 2021). Fi s , we con ibu e by
cha ac e izing he agg ega ion bias. P io wo k by Lyubich e al. (2018) show ha emission
in ensi ies in he US a y subs an ially wi hin 375 na owly de ined NAICS indus ies. The
Ge man manu ac u ing sec o displays simila he e ogenei y ( on G ae eni z and Ro ne
2023). Pa o his he e ogenei y is due o di e en emission in ensi y o p oduc s wi hin
indus ies, i.e., he wi hin indus y composi ion. In decomposi ions o emissions conduc ed
a he indus y le el, echnique e ec s a e calcula ed as wi hin sec o changes in emission
in ensi y. Changing composi ion o p oduc s wi hin an indus y gi es ise o agg ega ion
bias in he echnique and composi ion e ec . The di ec ion o bias depends on he sign o
he “ ue” echnique and composi ion e ec s. Wi h nega i e composi ion e ec s and posi-
i e echnique e ec s as in ou case, bo h e ec s a e unde es ima ed. When bo h e ec s a e
nega i e, he echnique e ec would end o be o e es ima ed in absolu e e ms: Emission
educ ions a ibu ed o he echnique e ec in eali y s em om wi hin sec o composi-
ional shi s. Whene e he ue composi ion e ec is nega i e, he bias wo ks owa ds p e-
dic ing a echnique e ec ha is oo clean. Ei he , i is posi i e bu oo small, o nega i e
and oo la ge. These indings sugges ha he clean-up a ibu ed o he echnique e ec in
p io analyses a a less g anula le el is likely o e s a ed.
Second, we con ibu e o he decomposi ion li e a u e by p o iding e idence on ca -
bon emissions. Se e al pape s examine he d i e s o local pollu an emissions (e.g.
Shapi o and Walke 2018; Le inson 2009, 2015; B unel 2017; Najja and Che niwchan
2021). These pape s gene ally documen la ge and nega i e echnique e ec s sugges ing
ha echnology adop ion has been a main d i e o local pollu an emission educ ions. A
much smalle li e a u e decomposes ca bon emissions (Ba ows and Olli ie 2018; Le in-
son 2021; B unel and Le inson 2022), bu hese s udies ha e ocused on he US and India
whe e clima e policy has been compa a i ely lax. We p o ide e idence om a de eloped
and expo -o ien ed economy in which ca bon emissions ha e been egula ed since he
ea ly 2000s.
Las ly, we con ibu e o he li e a u e e alua ing he e ec o en i onmen al policies on
he manu ac u ing sec o . A g owing li e a u e s udies he ela ionship be ween clima e
policies and ene gy demand as well as i m-pe o mance in Ge many (e.g. Flues and Lu z
2015, Lu z, 2016; Ge s e and Lamp, 2020; Hin e mann e al., 2020; Leh e al., 2020; on
G ae eni z and Ro ne , 2022). This li e a u e has ocused on iden i ying causal e ec s o
speci ic policies such as he Eu opean emissions ading scheme o he exemp ion om he
Ge man enewable ene gy su cha ge. Exploi ing quasi-na u al expe imen s, his body o
wo k mos ly quan i ies e ec s on emissions o la ge and ene gy in ensi e manu ac u ing
2524
E.Ro ne , K. on G ae eni z
1 3
plan s. Ou analysis p o ides con ex o hese indi idual policy e alua ions. We a e
he i s o apply he app oaches by Le inson (2015) and Fos e e al. (2008) o Ge man
emissions.1
2 Da a
We conduc ou analysis using he o icial plan le el mic o-da a om he ede al s a is ical
o ices o he Bund and he Lände . All manu ac u ing plan s in Ge many wi h mo e han
20 employees mus epo in he su eys. We use da a om 2005 o 2017. Ou analysis
equi es in o ma ion on each plan ’s emissions and o al ou pu , as well as agg ega e emis-
sions and ou pu . Mo eo e , we need in o ma ion on he ou pu sha e and emission in en-
si y o each subsec o and plan .
The Ge man Manu ac u ing Census does no con ain any in o ma ion on ca bon emis-
sions. We calcula e plan le el emissions by combining in o ma ion on manu ac u ing
plan s’ consump ion o 14 di e en uels and elec ici y wi h app op ia e emission ac o s
e ie ed om he Ge man En i onmen al Agency (Umwel bundesam 2008, 2020a, b).
Emission ac o s a e na ional bu ime- a ying. The emission ac o o elec ici y e lec s
he Ge man elec ici y mix as well as ansmission losses.2 Agg ega e emissions a e cal-
cula ed by summing up plan le el emissions. In he analysis, we abs ac om analysing
p ocess emissions due o da a limi a ions.3 Figu e1 shows he de elopmen o agg ega e
ca bon emissions in he Ge man manu ac u ing sec o be ween 2003 and 2017. In 2017,
ca bon emissions we e oughly 32 mio. onnes highe han in 2003.
The Ge man Manu ac u ing Census con ains in o ma ion abou p oduc s a he 9-digi
le el. The 9-digi le el is e y ine-g ained: In ou base yea 2005, we dis inguish be ween
4,672 di e en 9-digi p oduc s. Fo compa ison, he e a e only 1,356 p oduc s a he
6-digi le el; 221 di e en sec o s a he 4-digi le el, and 91 sec o s a he 3-digi le el.
Figu e2 shows an example o he b eakdown o a 2-digi sec o o he 9-digi p oduc le el
o illus a i e pu poses. Agg ega e ou pu is calcula ed by summing g oss ou pu o each
p oduc p oduced by indi idual manu ac u ing plan s.4 We de la e g oss ou pu using p o-
duce p ice indices (base yea 2015) om he Fede al S a is ical O ice (DeS a is 2018).5
1 Pe ick (2013); Kube and Pe ick (2019) analyse CO2 emissions a he i m le el using he Loga i hmic
Mean Di isia Index. They do no ake he e ec s o p oduc mix in o accoun .
2 The e a e sligh egional di e ences in emissions pe o mance ac oss Ge man powe plan s. Fo example,
coal i ed powe plan s om he Rhineland emi ed 113 CO2/TJ in 2016 whe eas a coal i ed powe plan
in middle Ge many emi ed 104 CO2/TJ in he same yea due o di e ences in hea a es (Umwel bunde-
sam 2019a). Ge many cons i u es one elec ici y ma ke and in gene al he e is no way o de e mine which
egion has deli e ed elec ici y o indi idual plan s.
3 Acco ding o he EEA (2015, 2016, 2017) p ocess emissions made up abou 10-16% o o e all Ge man
emissions egula ed unde he EU Emissions T ading Scheme in he yea s 2013 o 2015. Wi hin manu ac-
u ing, he sec o s in which p ocess emissions occu end o be he same sec o s ha also ha e la ge emis-
sions om uel combus ion, i.e. pulp and pape , coke and pe oleum, chemicals, o he non-me allic mine al
p oduc ion and me al p oduc ion.
4 No e ha by using g oss ou pu as a measu e o manu ac u ing ac i i y, we canno ule ou he possibili y
ha esul s a e d i en by he manu ac u ing sec o ou sou cing/s a ing o p oduce in e media e inpu s ha
we e p oduced/impo ed be o ehand. In o ma ion on alue added is only a ailable a he i m le el o a
s a i ied sample o i ms.
5 Whe e a ailable, p oduc le el g oss ou pu is de la ed using p ice indices a he 9-digi p oduc le el. In
o al, oughly 80% o g oss ou pu a e de la ed a he 9-digi le el, 13% a he 6-digi le el and he emain-
ing 7% a he 4-digi le el.

2525
Wha D i es Ca bon Emissions inGe man Manu ac u ing: Scale,…
1 3
Ene gy use and by ex ension CO2 emissions a e a ailable a he plan le el. We calcu-
la e emission in ensi ies a he p oduc le el by alloca ing plan emissions o he di e en
p oduc s based on ou pu sha es. The p ocedu e is desc ibed in mo e de ail in Addi ional
ile1. No e ha by using ou pu sha es o alloca e plan le el emissions on o he p oduc s,
we implici ly assume ha all p oduc s wi hin a plan a e p oduced wi h he same emission
Fig. 1 The de elopmen o ca bon emissions om uel combus ion in he Ge man manu ac u ing sec o .
No es: Sou ce: DOI 10.21242/43531.2017.00.03.1.1.0. Own calcula ions
Fig. 2 Indus ies, sec o s and p oduc s. No es: Sou ce: Goods ca alogue o p oduc ion s a is ics (DeS a is
2011)
2526
E.Ro ne , K. on G ae eni z
1 3
in ensi y.6 We conduc he Le inson-s yle decomposi ion analysis a he 3-, 4-, 6- and
9-digi sec o le el o quan i y agg ega ion bias.
3 S a is ical decomposi ion me hods
Decomposi ion ools a e equen ly used o disen angle he sou ces o emission changes.
A wide a ie y o di e en app oaches exis . We ocus he e on he decomposi ion in o
scale, composi ion and echnique e ec s as in Le inson (2015), which is he "wo kho se
app oach" in en i onmen al economics. Wi h ou excep ionally g anula da a, we a e able
o analyse composi ion e ec s in mo e de ail han pas esea ch and o assess he ex en o
agg ega ion bias in quan i ying composi ion and echnique e ec s using mo e agg ega e
da a. We complemen his p oduc le el decomposi ion o agg ega e emissions wi h a plan
le el decomposi ion o emission in ensi y in he manu ac u ing sec o inspi ed by Fos e
e al. (2008). The plan le el decomposi ion does no allow us o analyse he ole o p od-
uc composi ion on emissions de elopmen in de ail, bu speaks o he e ec o p oduc ion
alloca ion ac oss plan s and selec ion (en y and exi ) among manu ac u ing plan s in d i -
ing emission in ensi ies.7 The combina ion o bo h decomposi ion app oaches allows us o
ob ain a comp ehensi e pic u e o he d i e s o emissions.
3.1 P oduc le el decomposi ion o agg ega e emissions
The s a is ical decomposi ion o agg ega e emissions ollowing Le inson (2015) can be
ca ied ou a di e en le els o sec o al disagg ega ion. I is g ounded in a ep esen a ion
o o al emissions,
P
, o , in his case, CO2 in he Ge man manu ac u ing sec o a ime as
he sum o emissions om s di e en subsec o s in manu ac u ing. In each subsec o , emis-
sions a e de e mined by he p oduc o ou pu p oduced,
s
, and he emission in ensi y o
ha subsec o ,
zs
. Hence, o al emissions om manu ac u ing depend on agg ega e ou pu
V
o he manu ac u ing sec o as a whole, he sha e o each subsec o om agg ega e ou -
pu
𝜃s
and he emission ac o s o p oduc ion in each subsec o .
In ec o no a ion, his is:
whe e
𝛉𝐭
and
𝐳𝐭
a e
s×1
ec o s con aining he ma ke sha es and emission in ensi ies o
each o he s di e en indus ies.
To al di e en ia ion and di ision by emissions yields an exp ession o emission
changes (wi h ime subsc ip s d opped o no a ional con enience):
(1)
P
=
∑
s
ps =
∑
s
s zs =V
∑
s
𝜃s zs
(2)
P
=V
𝛉
�
𝐭
𝐳
𝐭
(3)
dP
P
=dV
V
+d𝜃
𝜃
+
dz
z
6 This mos likely leads o some measu emen e o , as Ba ows and Olli ie (2018) ound subs an ial a i-
a ion in emission in ensi ies wi hin mul i-p oduc i ms ac oss p oduc lines. To assess obus ness, we also
conduc he p oduc le el analysis based on single-p oduc plan s.
7 We hank an anonymous e e ee o his sugges ion.
2527
Wha D i es Ca bon Emissions inGe man Manu ac u ing: Scale,…
1 3
The i s e m o he equa ion is he scale e ec . The scale e ec is gi en by he change in
agg ega e ou pu and summa ises how emissions would ha e e ol ed i only he p oduc-
ion olume had changed. The second e m, he composi ion e ec , desc ibes how emis-
sions would ha e e ol ed i he sec o al composi ion o manu ac u ing had ollowed i s
his o ical pa h while keeping scale and emission in ensi ies ixed. The hi d e m is he
echnique e ec and explains how emissions would ha e e ol ed i emission in ensi ies had
ollowed hei his o ical pa hs while p oduc ion scale and composi ion we e ixed.8
The decomposi ion is s aigh o wa d on a concep ual le el. The accoun ing iden i y in
Eq.3) is used o di ec ly calcula e he scale e ec (keeping composi ion and echnique con-
s an ) and ei he he echnique o he composi ion e ec (keeping scale and composi ion o
scale and echnique cons an ).
Scale, composi ion and echnique e ec add up o he ac ually obse ed emission
changes as demons a ed in Eq. (3). Based on his iden i y, he emaining componen is
de e mined as he esidual once scale and composi ion o scale and echnique ha e been
sub ac ed om he ac ual obse ed emissions. I , gi en scale and composi ion, emissions
would ha e been highe han hey ac ually we e, he echnique e ec is nega i e. Con-
e sely, i , gi en scale and composi ion, emissions would ha e been lowe han hey ac u-
ally we e, he echnique e ec is posi i e.
No e ha es ima ing scale, echnique and composi ion in his way a ibu es all in e -
ac ions ha migh a ise be ween he h ee e ec s o he e ec es ima ed as he esidual.
Di e en es ima es o he echnique e ec depending on whe he i is es ima ed di ec ly
o indi ec ly as a esidual a e no an issue o obus ness o he decomposi ion. Ra he such
di e ences a ise when in e ac ion e ec s a e la ge. Le inson (2009, 2015) lis s se e al
po en ial ypes o in e ac ions, e.g., la ge indus ies ha ing inc easing e u ns o scale
o pollu ion aba emen o sh inking indus ies closing down he di ies plan s i s . I is
no ob ious which channel hese in e ac ions should be a ibu ed o. In many s udies, he
choice o whe he o calcula e echnique o composi ion di ec ly is mo i a ed by da a a ail-
abili y. Ou da a allow us o calcula e bo h composi ion and echnique e ec di ec ly and as
a esidual, he eby allowing us o assess he impo ance o in e ac ion e ec s.
Fo he sake o in e p e a ion, he echnique e ec can also be depic ed in a Laspey es-
like index. In his case, he p edic ed emissions wi h ac ual emission in ensi y de elop-
men s bu base yea p oduc ion scale and composi ion a e di ided by base yea emissions.
This index is equal o one when emission in ensi ies emain unchanged as compa ed o he
base yea . Falling o ising emission in ensi ies in con as lead o he echnique index ak-
ing on alues smalle o la ge han one. The index o he echnique e ec is gi en by he
ollowing equa ion:9
8 Ou decomposi ion is based on he wo kho se app oach which elies on e enues. Recen wo k by Rod-
igue e al. (2022) on SO2 emissions in China has shown ha his decomposi ion may be biased when
ma kups change o e he pe iod unde s udy. De Loecke and Eeckhou (2021) es ima e ma kups o se -
e al coun ies including Ge many o examine hei de elopmen o e ime. In Ge many ma kups emained
app oxima ely cons an h oughou ou s udy pe iod. While De Loecke and Eeckhou (2021) do no discuss
ma kup he e ogenei y wi hin Ge many, hei inding o cons an ma kups sugges s ha he bias is limi ed.
9 Simila ly, he composi ion e ec in index o m can be w i en as:
C
L=
∑
s𝜃s zs0
∑
s
𝜃
s
0z
s
0
2528
E.Ro ne , K. on G ae eni z
1 3
whe e 0 indica es he base yea o which emission in ensi y changes a e compa ed.10 Re o -
mula ing he es ima ed echnique e ec in his way cla i ies some o he p ope ies o he
calcula ed e ec . Speci ically, he agg ega e index consis s in changes in indi idual subsec-
o s s ha a e weigh ed by hei ela i e impo ance in he base yea 0. Fo ou applica ion
o emission in ensi ies, his means ha he es ima ed echnique e ec is a weigh ed a e age
o emission in ensi y changes in indi idual subsec o s s whe e he weigh s a e gi en by he
subsec o s’ sha es o o al manu ac u ing emissions in he base yea
𝜇s0
.11
In each subsec o , ou measu e o he echnique e ec does no cons i u e a echnique
imp o emen wi hin plan . The change in he emission in ensi y also depends on alloca ion
o p oduc ion ac oss plan s. Equa ion (7) shows his:
The i s e m in he las exp ession is he sha e o plan j in agg ega e p oduc ion o sub-
sec o s. The second e m is he plan ’s emission in ensi y. Clea ly, he emission in ensi y
o di e en plan s is weigh ed by hei ou pu sha e, which means ha emission in ensi y
imp o emen s (i.e., he echnique e ec in he analysis) can be achie ed bo h by emis-
sion in ensi y imp o emen s wi hin plan s, and by a ealloca ion o p oduc ion owa d less
emission in ensi e plan s. To clea ly disen angle he ole o wi hin plan changes e sus
ac oss plan ealloca ion, we complemen he sec o le el decomposi ion wi h a plan le el
decomposi ion o he echnique e ec .
3.2 The agg ega ion bias ins a is ical decomposi ions
The le el o sec o al disagg ega ion has di ec implica ions o he calcula ion o he com-
posi ion and echnique e ec . Suppose, e.g., ha no sec o da a a e a ailable a all, bu
only da a on he agg ega e manu ac u ing sec o . In ha scena io, i would no be possible
o iden i y any composi ion e ec . All changes in emissions would be a ibu ed o ei he
scale o echnique e ec . Speci ically, changes in emission in ensi ies caused by p oduc-
ion composi ion changes would be a ibu ed o he echnique e ec . Mo e de ailed sec o
le el da a makes i possible o sepa a e a composi ion e ec om he echnique e ec . I
(4)
T
L=
∑
s
z
s
s0
∑
z
s0
s0
=�
s
zs
z
s
0
∗
zs0 s0
∑s
z
s0
s0
≡�
s
zs
z
s
0
∗𝜇s
0
(5)
zs =
ps
s
=
∑
jpjs
s
=
∑
jzjs ∗ js
s
=�
j
sj
s
∗z
js
10 In ou applica ion, 2005 cons i u es he base yea .
11 Al e na i ely, Le inson (2015) p oposes o use a Paasche-like measu e whe e he weigh s a e gi en by
cu en sha es:
Di e ences be ween Laspey es and Paasche indices cap u e one in e ac ion a ibu ed o he e m es ima ed
as a esidual, namely whe he o no he manu ac u ing sec o shi s owa ds o away om sec o s in which
pollu ion in ensi ies decline he mos . The compa ison does no cap u e all possible in e ac ions be ween
scale, composi ion and echnique e ec . We epo compa isons be ween Laspey es and Paasche echnique
e ec s in Addi ional ile1. Gene ally, we ind ha sec o s o which he di e ence be ween Laspey es and
Paasche indices is big also display a big di e ence be ween es ima ing he composi ion o he echnique
e ec as a esidual. This sugges s ha a la ge sha e o in e ac ion e ec s a e be ween composi ion and ech-
nique a he han due o scale.
T
P=
∑
s
z
s
s
∑
s
z
s
0
s
2535
Wha D i es Ca bon Emissions inGe man Manu ac u ing: Scale,…
1 3
swi ches. The impo ance o p oduc swi ches o he obse ed emission in ensi y imp o e-
men s is ela i ely small. This is g ounded in he ac ha only a ela i ely small sha e o
single p oduc plan s swi ched hei p oduc s be ween 2005 and 2017 (o e all oughly 6%,
om yea o yea less han 1%). S ill, on a e age, plan s seem o ha e swi ched owa d
less emission in ensi e p oduc s (i.e., hey became less emission in ensi e a e a p oduc
swi ch). This swi ching beha iou migh be one d i ing o ce o he clean composi ion
e ec iden i ied in he p oduc le el decomposi ion.
4.4 Ca bon leakage andachange in hep oduc composi ion
The clea end owa ds a less emission in ensi e p oduc ion composi ion ha we ind a
he 9-digi p oduc le el could indica e ca bon leakage. I especially ene gy in ensi e sec-
o s ou sou ce he p oduc ion o he mos emission in ensi e p oduc s, his would lead o a
less ca bon in ensi e domes ic p oduc ion composi ion. Sugges i e e idence o he occu -
ence o ca bon leakage and o he loss o compe i i eness o Ge man ene gy in ensi e
sec o s could be ound i we obse ed a simul aneous shi in Ge man expo s owa d less
emission in ensi e sec o s and in Ge man impo s owa d mo e emission in ensi e sec o s.
We decompose Ge man ade lows ollowing Eq. (3), akin o he app oach aken in Le -
inson (2009). The decomposi ion is conduc ed a he 3-digi sec o le el using impo s and
Table 2 Alloca ion and selec ion
Sou ce: Own calcula ions.
The able shows he esul s o he plan le el decomposi ion o emission in ensi ies calcula ed a he 4-digi
sec o le el. Sec o s a e agg ega ed using ou pu weigh s. The i s ow shows esul s in long di e ences
be ween 2005 and 2017. The h ee emaining ows epo a e ages o yea -on-yea changes o dis inc pe i-
ods
Time pe iod To al change Wi hin plan Be ween plan s C oss e ec En y Exi
2005–2017 4.2 19.5 10.5 -25.4 -0.7 -0.3
2005–2009 1.3 3.6 5.4 -7.6 0.4 0.4
2010–2013 -0.0 2.9 1.0 -3.9 -0.3 -0.1
2014–2017 -0.3 0.5 2.0 -3.2 0.8 0.4
Table 3 P oduc en y e ec s
wi hin single p oduc plan s
Sou ce: Own calcula ions.
The able shows he esul s o a p oduc le el decomposi ion o emis-
sion in ensi ies calcula ed a he plan le el. Plan s a e agg ega ed
using ou pu weigh s. The i s ow shows esul s in long di e ences
be ween 2005 and 2017. The second ow epo s a e ages o yea -on-
yea changes be ween 2005 and 2017
Plan s Long
di e -
ences
To al change P oduc swi ch Technique
Single p oduc Yes −11.99 −1.48 −10.52
Single p oduc No −8.47 −0.67 −7.80

2536
E.Ro ne , K. on G ae eni z
1 3
expo s om Eu os a (2023). Base yea , i.e., 2005 emission in ensi ies a he 3-digi le el
a e compu ed om he Ge man Manu ac u ing Census.20
Table 4 shows Laspey es and Paasche indices o he composi ion e ec o Ge man
manu ac u ing impo s and expo s, espec i ely. Clea ly, he composi ion shi owa d less
emission in ensi e sec o s is no limi ed o o e all Ge man p oduc ion, bu also ex ends o
bo h Ge man impo s and Ge man expo s. Impo s ha e e en shi ed as e owa d cleane
sec o s han domes ic p oduc ion. This sugges s ha ca bon leakage and s ingen Ge man
clima e policies may no be a he oo s o he Ge man p oduc ion shi owa d less emis-
sion in ensi e goods.
Table 4 Laspey es and Paasche
indices o he composi ion e ec
o Ge man ade lows
Sou ce: Own calcula ions.
The able shows he Laspey es and Paasche indices o he composi-
ion e ec calcula ed o he decomposi ion o Ge man expo s (le )
and Ge man impo s ( igh ). Compa ing he wo shows, ha he
Paasche index is consis en ly smalle han he Laspey es index indica -
ing ha ade lows in indus ies wi h as e alling/mo e slowly g ow-
ing ca bon in ensi ies g ew a a as e a e
Yea Laspey es Paasche Laspey es Paasche
Expo s Expo s Impo s Impo s
2005 1 1 1 1
2006 1.002387 1.001079 1.008533 1.003687
2007 0.9569227 0.9535437 1.009588 1.003167
2008 0.950685 0.9394742 0.9987149 0.9887199
2009 0.9790129 0.9720478 0.8985757 0.8895001
2010 0.9624456 0.9484019 0.9516796 0.9256257
2011 0.9259918 0.9122524 0.9341809 0.910885
2012 0.9189055 0.9039791 0.9045228 0.8831315
2013 0.9114583 0.8901056 0.901652 0.8795794
2014 0.907735 0.8805468 0.9000167 0.8723223
2015 0.8880442 0.8578597 0.8838269 0.8516514
2016 0.8824099 0.8520127 0.8584881 0.819602
2017 0.8790151 0.8361229 0.8638813 0.8129885
20 In he decomposi ion o expo s, we implici ly assume ha expo s a e p oduced wi h he same emission
in ensi y as o e all p oduc ion. While his is no necessa ily ue (see, e.g., Fo slid e al. 2018 o e idence
ha expo e s end o be less emission in ensi e), no e ha he decomposi ion exploi s a ia ion ela i e o
he base yea . The e will only be bias i he “expo p emium” in e ms o emission in ensi y di e s sys-
ema ically ac oss 3-digi sec o s. In he decomposi ion o impo s, o he lack o in e na ional emissions
da a a he 3-digi sec o le el, we also ely on Ge man emission in ensi ies, no he emission in ensi ies o
p oduc ion in he o igin coun y. We us ha Ge many se es as a easonable app oxima ion o he ela i e
emission in ensi y o di e en sec o s, e en i emission in ensi ies om di e en o igin coun ies migh di -
e . Al e na i ely, as in Le inson (2009), he exe cise can be iewed as analysing he emissions displaced,
i.e., he emissions ha would ha e occu ed in Ge many absen impo s.
2537
Wha D i es Ca bon Emissions inGe man Manu ac u ing: Scale,…
1 3
4.5 Sec o al he e ogenei y in hede elopmen o emissions andemission
in ensi ies
We agg ega e sec o and p oduc speci ic de elopmen s using weigh s o hei emission
sha e in he base yea 2005 (as shown in Eq. (5)). This means ha he p oduc and sec o
speci ic decomposi ion esul s a e hea ily d i en by de elopmen s in he e y ene gy and
emission in ensi e sec o s me al p oduc ion, chemicals, coke and pe oleum, o he non-
me allic mine al p oduc s and pulp and pape .21,22 In his sec ion, we show disagg ega ed
esul s by 2-digi sec o . This e eals subs an ial he e ogenei y in emission d i e s ac oss
sec o s. I also shows ha agg ega ion and selec ion bias ma e o a di e en ex en in di -
e en sec o s.
Figu e4 con as s he Laspey es indices o he composi ion and he echnique e ec a
he 2-digi sec o le el o he yea 2017 ela i e o 2005. Bo h sign and magni ude o ech-
nique and composi ion e ec s a y subs an ially.23 Table5 p o ides a lis o b ie sec o
desc ip ions wi h he sec o codes.
Despi e he posi i e echnique e ec in 2017 o he manu ac u ing sec o o e all, in he
same yea , nega i e echnique e ec s p e ail when i comes o he 2-digi sec o s (i.e., mos
sec o s lie below he ho izon al line in he igu e). Se e al sec o s expe ienced con inu-
ous imp o emen s in e ms o hei emission in ensi y (among o he s, NACE 15: manu-
ac u e o lea he and ela ed p oduc s, NACE 18: p in ing and ep oduc ion o eco ded
media, and NACE 26: manu ac u e o compu e , elec onic and op ical p oduc s). The
mo e ene gy in ensi e sec o s (highligh ed by he black boxes in he igu e), howe e , a e
he ones ha eally ma e o manu ac u ing’s emissions, and hey end o display posi i e,
i.e., emission inc easing, echnique e ec s.24
Why ha e especially hose sec o s ha ma e in e ms o manu ac u ing’s emissions
no managed o imp o e hei emission in ensi y o p oduc ion? A po en ial eason o
ene gy in ensi e sec o s gene ally expe iencing mo e o an inc ease in emission in ensi ies
could lie in hose sec o s bene i ing o a highe deg ee om exemp ion and compensa ion
schemes ha we e in oduced alongside clima e policies o p e en a loss in compe i i e-
ness and leakage e ec s. In he obse a ion pe iod, la ge and in ensi e elec ici y use s
could, e.g., ge exemp om paying he Renewable Ene gy Su cha ge which accoun ed o
oughly a hi d o elec ici y p ices. Manu ac u ing plan s abo e a ce ain elec ici y p o-
cu emen h eshold could also bene i om paying educed elec ici y ne wo k a i s, and
since 2013 a subse ecei e elec ici y cos compensa ion due o highe elec ici y p ices
owing o he EU ETS. Also, h ough he leakage lis o he EU ETS, ene gy in ensi e and
ade exposed sec o s a e eligible o a highe sha e o ee allowances o elec ici y p ice
21 The de elopmen o ene gy consump ion in Ge man manu ac u ing, spli ac oss 2-digi sec o s, is shown
in Figu e7 in he Addi ional ile1.
22 A simila a gumen applies o he plan le el decomposi ion which is agg ega ed based on ou pu sha es.
He e ogenei y in e ms o ou pu sha es is o a simila magni ude, bu only chemicals belong o he op i e
sec o s in bo h ca ego ies. The o he ou op i e sec o s in e ms o ou pu sha es a e ood, machine y,
mo o ehicles, and me al p oduc s. The ood sec o holds a 6 h place in e ms o ene gy in ensi y.
23 The same can be seen om Tables9 and 10 con aining he Laspey es indices o echnique and composi-
ion e ec s in he Addi ional ile1.
24 The ou lie s in Figu e4 (e.g. NACE 33 (Repai s) and NACE 26 (Compu e s and elec onics) and NACE
(Tobacco)) a e small in e ms o emissions. They do no con ibu e subs an ially o he main esul s. The
de elopmen o emission in ensi ies in compu ing is pa ially d i en by echnological de elopmen and
esul ing p ice dec eases a ec ing de la ed ou pu . Repai s a e gene ally an addi ional p oduc wi h lim-
i ed impo ance o manu ac u ing plan s p oducing machine y. In consequence ac i i y displays subs an ial
a ia ion o e ime.
2538
E.Ro ne , K. on G ae eni z
1 3
compensa ion in he s udy pe iod.25 Aside om di ec ly acing di e en p ices, he ene gy
in ensi e sec o s end o use a di e en ene gy mix. Speci ically, hey ely a lo mo e on
onsi e elec ici y gene a ion, and as such ha e been less exposed o he egula ion-induced
elec ici y p ice inc eases o e ou obse a ion pe iod. Consequen ly, hey migh ha e been
subjec o a lowe shadow p ice on emissions and he e o e ha e aced lowe incen i es o
dec ease hei emission in ensi y.
Composi ion e ec s also a y ac oss sec o s wi h sec o s like NACE 13: ex iles, o
NACE 20: chemicals shi ing owa ds less ca bon in ensi e goods. Sec o s like NACE
16: manu ac u e o wood and wood p oduc s, and NACE 26: manu ac u e o com-
pu e , elec onic and op ical p oduc s shi owa d mo e emission in ensi e goods. In
some sec o s, he p oduc ion composi ion emains i ually unchanged o e ime wi h
ega ds o he emission in ensi y (e.g., NACE 22: manu ac u e o ubbe and plas ic
p oduc s).
Agg ega ion bias is also p esen in he sec o -speci ic decomposi ions. Mo ing
om he 3-digi decomposi ion o he 9-digi decomposi ion, i is clea ha dispe sion
inc eases in he sca e plo . This is consis en wi h unde es ima ion due o agg ega ion
Table 5 O e iew o e wo-digi sec o s in manu ac u ing
Sec o Code Sec o name
10 Manu ac u e o ood p oduc s
11 Manu ac u e o be e ages
12 Manu ac u e o obacco p oduc s
13 Manu ac u e o ex iles
14 Manu ac u e o wea ing appa el
15 Manu ac u e o lea he and ela ed p oduc s
16 Manu ac u e o wood and o p oduc s o wood and co k, excep u ni u e
17 Manu ac u e o pulp, pape and pape p oduc s
18 P in ing and ep oduc ion o eco ded media
19 Manu ac u e o coke and e ined pe oleum p oduc s
20 Manu ac u e o chemical p oduc s
21 Manu ac u e o basic pha maceu ical p oduc s and pha maceu ical p epa a ions
22 Manu ac u e o ubbe and plas ic p oduc s
23 Manu ac u e o o he non-me allic mine al p oduc s
24 Me al p oduc ion and p ocessing
25 Manu ac u e o ab ica ed me al p oduc s
26 Manu ac u e o compu e , elec onic and op ical p oduc s
27 Manu ac u e o elec ical equipmen
28 Machine manu ac u ing
29 Manu ac u e o mo o ehicles
30 O he anspo equipmen
31 Manu ac u e o u ni u e
32 O he manu ac u ing
33 Repai and ins alla ion o machine y and equipmen
25 An o e iew o exis ing exemp ions in Ge many o ene gy and elec ici y axes and le ies wi h he el-
e an h esholds can be ound in a publica ion by he Umwel bundesam (2019b).
2539
Wha D i es Ca bon Emissions inGe man Manu ac u ing: Scale,…
1 3
bias. We also see ha he pa e n is no en i ely consis en . Some sec o s shi a ound
conside ably depending a he agg ega ion le el. Fo example, sec o 14: Clo hing goes
om ha ing a mode a e posi i e composi ion e ec and a small bu nega i e echnique
e ec in he 3-digi decomposi ion o ha ing a small posi i e composi ion e ec and
a la ge posi i e echnique e ec in he 9-digi decomposi ion. The e a e wo easons
o hese la ge shi s: Fi s , composi ion migh ha e changed subs an ially wi hin he
clo hing indus y, wi h a shi owa d less emission in ensi e p oduc s wi hin 3-digi
indus ies. Second, he e migh be many p oduc en ies and exi s such ha he decom-
posi ions o he a ious agg ega ions di e in e ms o hei plan -co e age. Despi e
hese shi s, a gene al pa e n owa ds mo e dispe sion a highe le els o disagg ega-
ion is isible.
5 Conclusion
We ha e ye o make subs an ial p og ess on he educ ion o CO2 emissions. Despi e
he in oduc ion o se e al clima e policies in Ge many in he pe iod be ween 2005 and
2017 ca bon emissions con inued o inc ease. To unde s and d i e s o his emissions
Fig. 4 Laspey es indices o composi ion and echnique e ec in 2017 ac oss sec o s. Sou ce: DOI 10.212
42/43531.2017.00.03.1.1.0. Own calcula ions. No es: The igu es display sca e plo s o Laspey es indices
o echnique and composi ion agg ega ed a he 2-digi sec o le el. The six mos ene gy in ensi e sec o s
a e ma ked by black boxes, whe eas he emaining sec o s a e ma ked wi h ci cles. Values abo e one indi-
ca e ha he sec o uses mo e emissions in ensi e echniques in 2017 as compa ed o 2005 (Y-axis) and/o
p oduces mo e emissions in ensi e p oduc s in 2017 as compa ed o 2005 (X-axis)
2540
E.Ro ne , K. on G ae eni z
1 3
de elopmen , we apply a wo kho se decomposi ion me hod om he en i onmen al eco-
nomics li e a u e o ex ao dina ily ine-g ained da a on 9-digi p oduc s.
We show ha he inc ease in ca bon emissions om he manu ac u ing sec o can
o a la ge ex en be a ibu ed o an inc ease in p oduc ion scale. We obse e a clean-
up o Ge man manu ac u ing in he o de o 9% educ ion o emissions compa ed o
wha hey would ha e been had composi ion and echnique emained unchanged since
2005. This clean-up is due o a change in p oduc composi ion. F om 2011 onwa ds, he
Ge man manu ac u ing sec o has shi ed owa ds g eene p oduc s. In con as , emis-
sion in ensi ies o p oduc ion ha e mos ly inc eased as compa ed o 2005. We show
ha da a agg ega ion ma e s in decomposi ion analyses. Bo h he composi ion and he
echnique e ec a e subs an ially unde es ima ed i conduc ing he analysis a he mo e
s anda d 3-digi sec o le el, a he han using da a on 9-digi p oduc s: The clean-up
esul ing om changing p oduc composi ion is smalle (71% o he e ec calcula ed a
he 9-digi le el) and he inc ease in emission in ensi y o p oduc ion echnique is also
a enua ed (51% o he e ec calcula ed a he 9-digi le el), hough he conclusions s ill
hold. The changing composi ion o p oduc ion aises conce n whe he hese pa e ns a e
due o ou sou cing o mo e ca bon in ensi e p oduc s and could be indica i e o ca bon
leakage. Ou p elimina y analyses decomposing impo s and expo s do no suppo his
conclusion, bu mo e esea ch is equi ed o ule ou such conce ns.
Ou measu e o he echnique e ec depends on wi hin plan emission in ensi y as
well as p oduc ion alloca ion ac oss plan s. We complemen he p oduc le el decom-
posi ion wi h a plan le el decomposi ion o be e unde s and he oles o selec ion and
alloca ion in d i ing he inc ease in emission in ensi ies. This second decomposi ion
shows ha emission in ensi y wi hin plan s gene ally inc eased o e he pe iod. The
be ween plan s e ec is also posi i e indica ing ha alloca ion o p oduc ion shi ed
owa ds mo e emission in ensi e plan s. A la ge and nega i e c oss e ec indica es,
ha plan s ha we e ini ially highly emission in ensi e imp o ed hei emission pe o -
mance mo e han less emission in ensi e plan s. E ec s o ne en y on agg ega e emis-
sion in ensi y a e negligible. The ne e ec emains an inc ease in emission in ensi y in
he manu ac u ing sec o o e he pe iod s udied.
Supplemen a y In o ma ion The online e sion con ains supplemen a y ma e ial a ailable a h ps:// doi.
o g/ 10. 1007/ s10640- 024- 00894-7.
Funding Open Access unding enabled and o ganized by P ojek DEAL.
Decla a ions
Compe ing in e es s The au ho s ha e no compe ing in e es s o decla e ha a e ele an o he con en o
his a icle.
Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License,
which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Com-
mons licence, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his a icle
a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line o he
ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended use is no
pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly
om he copy igh holde . To iew a copy o his licence, isi h p://c ea i ecommons.o g/licenses/by/4.0/.

2541
Wha D i es Ca bon Emissions inGe man Manu ac u ing: Scale,…
1 3
Re e ences
Ba ows G, Olli ie H (2018) Cleane i ms o cleane p oduc s? How p oduc mix shapes emission
in ensi y om manu ac u ing. J En i on Econ Manag 88:134–158
B unel C (2017) Pollu ion o sho ing and emission educ ions in EU and US manu ac u ing. En i on
Resou Econ 68:621–641
B unel C, Le inson A (2022) Globaliza ion and g eenhouse gas emissions: is he Uni ed S a es a pollu-
ion ha en?, Edwa d Elga Publishing
De Loecke J, Eeckhou J (2021) Global Ma ke Powe . Wo king Pape
DeS a is (2011) Gü e e zeichnis ü P oduk ionss a is iken (GP 2009). Fu he in o ma ion: h ps://
www. des a is. de/ DE/ Me ho den/ Klass i ika ionen/ Gue e - Wi s cha sklas si ik a ion en/ klass i ik a ion-
gp- 19. h ml.
DeS a is (2018) E zeuge p eisindex gewe bliche P oduk e. Fu he in o ma ion: h ps:// www. des a is.
de/ DE/ Themen/ Wi s cha / P eise/ E zeu ge p eisin dex- gewe bliche- P odu k e- inhal . h ml
Ede ing on J, Le inson A, Minie J (2004) T ade libe aliza ion and pollu ion ha ens. B.E. J Econ Anal
Policy 4(2)
EEA (2015). Applica ion o he EU Emissions T ading Di ec i e - Analysis o na ional esponses unde
A icle 21 o he EU ETS Di ec i e in 2014. EEA Technical Repo No. 3
EEA (2016) Applica ion o he EU Emissions T ading Di ec i e—analysis o na ional esponses unde
A icle 21 o he EU ETS Di ec i e in 2015. EEA Repo No. 6
EEA (2017) Applica ion o he EU Emissions T ading Di ec i e—analysis o na ional esponses unde
A icle 21 o he EU ETS Di ec i e in 2016. EEA Repo No. 4
Eu os a (2023). EU ade 1988-2022 by CPA 2008 (DS-057009). A ailable a h ps:// ec. eu opa. eu/ eu os
a / da ab owse / (accessed on 25 May 2023)
Flues F, Lu z BJ (2015) The e ec o elec ici y axa ion on he Ge man manu ac u ing sec o : a eg es-
sion discon inui y app oach. ZEW discussion pape s No. 15-013
Fo slid R, Okubo T, Ull ei -Moe KH (2018) Why a e i ms ha expo cleane ? In e na ional ade,
aba emen , and en i onmen al emissions. J En i on Econ Manag 91:166–183
Fos e L, Hal iwange J, Sy e son C (2008) Realloca ion, i m u no e , and e iciency: selec ion on p o-
duc i i y o p o i abili y? Am Econ Re 98(1):394–425
Ge s e A, Lamp S (2020) Ene gy ax exemp ions and indus ial p oduc ion. Wo king Pape
Hin e mann B, Za ko ic M, diMa ia C, Wagne U (2020) The e ec o clima e policy on p oduc i i y
and cos pass- h ough in he Ge man manu ac u ing sec o . CRC TR224 Discussion Pape No. 249
IPCC (2014) Clima e change 2014: mi iga ion o clima e change. Con ibu ion o wo king g oup III o
he i h assessmen epo o he in e go e nmen al panel on clima e change
Kube R, Pe ick S (2019) Cleane p oduc ion o s uc u al change? A b eakdown o CO2 emissions
educ ions in he Ge man manu ac u ing sec o using i m-le el da a. Unpublished wo king pape
Leh J, Ge s e A, Piepe S, Wagne U (2020) The impac o ca bon ading on indus y: e idence om
Ge man manu ac u ing i ms. mimeo
Le inson A (2009) Technology, in e na ional ade, and pollu ion om US manu ac u ing. Am Econ
Re 99(5):2177–2192
Le inson A (2015) A di ec es ima e o he echnique e ec : changes in he pollu ion in ensi y o US
manu ac u ing, 1900–2008. J Assoc En i on Resou Econ 2(1):43–56
Le inson A (2021). Ene gy in ensi y: Deindus ializa ion, composi ion, p ices, and policies in U.S.
s a es. Resou Ene gy Econ 65
Lu z BJ (2016) Emissions ading and p oduc i i y: i m-le el e idence om Ge man manu ac u ing.
ZEW Discussion Pape , 16-067
Lyubich E, Shapi o JS, Walke R (2018) Regula ing mismeasu ed pollu ion: implica ions o i m he e o-
genei y o en i onmen al policy. AEA Pape s and P oceedings 108:136–142
Najja N, Che niwchan J (2021) En i onmen al egula ions and he clean-up o manu ac u ing: plan -
le el e idence. Re Eon S a 103(3):476–491
Pe ick S (2013) Ca bon e iciency, echnology, and he ole o inno a ion pa e ns: e idence om Ge -
man plan -le el mic oda a. Kiel Wo king Pape s 1833
Rod igue J, Sheng D, Tan Y (2022) The cu ious case o he missing Chinese emissions. J Assoc En i on
Resou Econ 9(4):755–805
Shapi o JS, Walke R (2018) Why Is pollu ion om US manu ac u ing declining? The oles o en i on-
men al egula ion, p oduc i i y, and ade. Am Econ Re 108(12):3814–3854
Umwel bundesam (2008) Bes immung spezi ische T eibhausgas-Emissions ak o en ü Fe nwä me.
Clima e Change. 08/08
2542
E.Ro ne , K. on G ae eni z
1 3
Umwel bundesam (2019a) Ak ualisie ung de Emissions ak o en ü G oß eue ungsanlagen. Tex e,
140/2019, u he in o ma ion (only a ailable in Ge man): h ps:// www. umwel bund esam . de/ publi
ka io nen/ ak ua lisie ung- emiss ions ak o en- g oss eue ungsa nlagen
Umwel bundesam (2019b) Re o m und Ha monisie ung de un e nehmensbezogenen Ausnahme egelungen
im Ene giebe eich. Tex e, 23/2019, u he in o ma ion (only a ailable in Ge man): h ps:// www. umwel
bund esam . de/ publi ka io nen/ e o m- ha mo nisie ung- de - un e nehme nsbez ogenen
Umwel bundesam (2020a) Ca bon Dioxide Emissions o he Ge man A mosphe ic Emission Repo ing
1900–2018
Umwel bundesam (2020b) En wicklung de spezi ischen Kohlendioxid-Emissionen des deu schen S om-
mix in den Jah en 1990–2019. Clima e Change, 13/2020, u he in o ma ion (only a ailable in Ge -
man): h ps:// www. umwel bund esam . de/ publi ka io nen/ en wi cklung- de - spezi isch en- kohle ndiox id-2
on G ae eni z K, Ro ne E (2022) Do manu ac u ing plan s espond o exogenous changes in elec ici y
p ices? E idence om adminis a i e mic o-da a. ZEW Discussion Pape , 22-038
on G ae eni z K, Ro ne E (2023) Ene gy Use Pa e ns in Ge man Manu ac u ing om 2003 o 2017.
Jah büche ü Na ionalökonomie und S a is ik 243(3–4):319–354
on G ae eni z K, Ro ne E, Rich e P (2023) Is Ge many becoming he Eu opean Pollu ion Ha en? ZEW
Discussion Pape , 23–069
Publishe ’s No e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps and
ins i u ional a ilia ions.