Bellia, Ma io; Di Gi olamo, E ica F ancesca; Pagano, And ea; Papadopoulos,
Geo gios
Wo king Pape
Local Banks and lood isk: The case o Ge many
JRC Wo king Pape s in Economics and Finance, No. 2023/13
P o ided in Coope a ion wi h:
Join Resea ch Cen e (JRC), Eu opean Commission
Sugges ed Ci a ion: Bellia, Ma io; Di Gi olamo, E ica F ancesca; Pagano, And ea; Papadopoulos,
Geo gios (2023) : Local Banks and lood isk: The case o Ge many, JRC Wo king Pape s in
Economics and Finance, No. 2023/13, Eu opean Commission, Isp a
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/283107
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Local Banks and lood isk: he
case o Ge many
Bellia, M.
Di Gi olamo, F .
Pagano, A.
Papadopoulos, G.
2023
JRC Wo king Pape s in Economics and Finance, 2023/13
This publica ion is a Wo king Pape by he Join Resea ch Cen e (JRC), he Eu opean Commission’s science
and knowledge se ice. I aims o p o ide e idence-based scien i ic suppo o he Eu opean policymaking
p ocess. The scien i ic ou pu exp essed does no imply a policy posi ion o he Eu opean Commission. Nei he
he Eu opean Commission no any pe son ac ing on behal o he Commission is esponsible o he use ha
migh be made o his publica ion. Fo in o ma ion on he me hodology and quali y unde lying he da a used in
his publica ion o which he sou ce is nei he Eu os a no o he Commission se ices, use s should con ac
he e e enced sou ce. The designa ions employed and he p esen a ion o ma e ial on he maps do no imply
he exp ession o any opinion wha soe e on he pa o he Eu opean Union conce ning he legal s a us o any
coun y, e i o y, ci y o a ea o o i s au ho i ies, o conce ning he delimi a ion o i s on ie s o bounda ies.
Con ac in o ma ion
Name: F ancesca Di Gi olamo
Add ess: ia En ico Fe mi, 2749, Isp a (VA), I aly
Email: ancesca.di-gi olam[email p o ec ed] opa.eu
EU Science Hub
h ps://join - esea ch-cen e.ec.eu opa.eu/
JRC136411
Isp a: Eu opean Commission, 2023
© Eu opean Union, 2023
The euse policy o he Eu opean Commission is implemen ed by he Commission Decision 2011/833/EU o 12
Decembe 2011 on he euse o Commission documen s (OJ L 330, 14.12.2011, p. 39). Excep o he wise no ed,
he euse o his documen is au ho ised unde he C ea i e Commons A ibu ion 4.0 In e na ional (CC BY 4.0)
licence (h ps://c ea i ecommons.o g/licenses/by/4.0/). This means ha euse is allowed p o ided app op ia e
c edi is gi en and any changes a e indica ed. Fo any use o ep oduc ion o pho os o o he ma e ial ha is
no owned by he EU, pe mission mus be sough di ec ly om he copy igh holde s.
All con en © Eu opean Union, 2023
How o ci e his epo : Bellia M., Di Gi olamo F.E., Pagano A., Papadopoulos M. (2023), Local Banks and lood
isk: he case o Ge many, Eu opean Commission, Join Resea ch Cen e, Isp a, JRC136411
Execu i e summa y
The clima e c isis has eme ged as a signi ican h ea o i ms a ound he wo ld, impac ing
no only hei ope a ions bu also hei deb epaymen capaci y. As ex eme wea he
e en s become mo e equen and se e e, companies ace an inc eased isk o being
unable o mee hei inancial obliga ions wi h a - eaching epe cussions on lending
ins i u ions.
Unde s anding and managing clima e- ela ed inancial isks a ising om he escala ing
equency and in ensi y o ca as ophic e en s, called physical isk, is c ucial o
main aining inancial s abili y. I equi es assessing he ulne abili ies o inancial
ins i u ions, in eg a ing clima e- ela ed ac o s in o isk managemen amewo ks, and
p omo ing clima e esilience and adap a ion measu es.
Policymake s ha e aken se e al policy ini ia i es which equi e banks o assess and
disclose isks a ising om clima e- ela ed na u al disas e s, in eg a e hese isks in o hei
isk managemen amewo ks, and align hei in es men po olios and inancial p oduc s
wi h sus ainabili y c i e ia. The speci ic equi emen s a y o each ini ia i e, bu he o e all
goal is o os e a mo e clima e- esilien inancial sec o by conside ing and add essing
clima e physical isks. A he EU le el, se e al policy ac ions equi e banks o assess, epo
and in eg a e clima e- ela ed isks in o hei isk managemen amewo ks. The
Sus ainable Finance S a egy expec s banks o conduc comp ehensi e isk assessmen s
ha conside he po en ial impac s o clima e change on hei business ac i i ies. As pa
o he S a egy, Eu opean Supe iso y Au ho i ies will wo k on clima e s ess es ing o he
inancial sys em including a one-o Fi - o -55 clima e isk scena io analysis in coope a ion
wi h he Eu opean Cen al Bank (ECB) and he Eu opean Sys emic Risk Boa d. The Non-
Financial Repo ing Di ec i e equi es banks unde i s scope o disclose in o ma ion
on hei policies, isks, and ou comes ela ed o en i onmen al ma e s, including clima e
change and physical isks. The EU Taxonomy, which p o ides c i e ia o de e mining which
economic ac i i ies can be classi ied as en i onmen ally sus ainable, de ails how banks
need o iden i y and epo on in es men s o p oduc s ha con ibu e o clima e change
adap a ion o mi iga e physical isks.
Agains his backg ound, he p ima y objec i e o his pape is o e alua e he e ec o
localized clima e- ela ed e en s, unde a ious global wa ming scena ios, on egional
banks in Ge many. By doing so, i aims o shed ligh on po en ial ulne abili ies and isks
in he Ge man banking sec o , pa ing he way o mo e in o med decision-making, in
pa icula on isk mi iga ion s a egies.
Conside ing a s a ic sys emic inancial c isis scena io, we show ha clima e- ela ed isks
would inc ease o e all losses by 0.1% unde he cu en empe a u e le el and hese could
become e en mo e signi ican wi h losses up o 0.9% wi h a 3°C empe a u e inc ease.
Howe e , banks in some egions would su e much highe losses han in o he s. Mo eo e ,
we obse e ha , e en in a no mal economic si ua ion, he e a e ew banks de aul ing
because o clima e isk. Should hese de aul s igge diso de ly ma ke adjus men s,
subsequen dynamics would be negligible unde he ac ual wa ming condi ions, howe e ,
could lead o signi ican losses o he Ge man egional banking sys em up o 1% o o al
asse s i a 3°C empe a u e ise ma e ialized. In bo h cases, he a iabili y o losses ac oss
NUTS2 egions, is qui e la ge, wi h some egions eaching 5% o o al asse s in case o an
exis ing inancial c isis, and e en o 10% in a no mal economic si ua ion. This nega i e
impac can be e ec i ely ackled wi h he implemen a ion o adap a ion s a egies.
The e o e ou indings suppo he idea ha banks ope a ing in egions p one o such
geog aphic isks should ake measu es o sa egua d hemsel es. One po en ial app oach
is o implemen addi ional bank- speci ic capi al add-ons, which would enhance he
esilience o he banking sec o agains hese na u al- ela ed shocks. This s a egy would
help mi iga e he po en ial nega i e impac s on banks’ inancial s abili y and ensu e hei
abili y o wi hs and and eco e om such e en s. Al e na i ely, banks can con ibu e o
he adap a ion e o s aimed a educing he ulne abili y o he egions hey ope a e in.
By suppo ing adap a ion ini ia i es, such as in es ing in clima e- esilien in as uc u e o
p o iding inancial se ices o suppo sus ainable p ac ices, banks can play a c ucial ole
in building he esilience o he inancial sec o and educing he o e all impac on he
economy.
Fu he esea ch could in es iga e mo e in de ail he o e all impac o physical isk on he
inancial sec o , as inc easing a en ion is de o ed bo h by policymake s and academics o
he ole o clima e isk o inancial s abili y and deb sus ainabili y

Local Banks and lood isk: he case o Ge many
Ma io Bellia, E ica F ancesca Di Gi olamo, And ea Pagano, Geo gios Papadopoulos
Abs ac
This pape uses a simula ion model o e alua e he e ec s o i e looding e en s occu ing wi hin
Ge many on egional’ banks. Unde a 1.5℃inc ease in empe a u e, he impac is o e all a he small,
e en accoun ing o he de alua ion o loans exposed o loods. Howe e , unde a 3℃inc ease, bank
losses can each 1% o o al asse s. We show ha he implemen a ion o adap a ion solu ions would
be success ul in keeping isks a he cu en le el.
Keywo ds: Physical isk, i e lood e en s, dynamic balance shee , banking c isis.
J.E.L. classi ica ion: C15; G2; Q54.
1
1 In oduc ion
The clima e c isis has eme ged as a signi ican h ea o i ms a ound he wo ld, impac ing no only hei
ope a ions bu also hei deb epaymen capaci y. As ex eme wea he e en s become mo e equen and
se e e, companies ace an inc eased isk o being unable o mee hei inancial obliga ions. A s udy om
he Bank o Japan (Yamamo o e al.,2021) iden i ies a nega i e impac o lood damages on company
p o i abili y p oxied by he a io o p o i o sales, especially in he manu ac u ing indus y. In e es ingly,
he impac ends o be la ge o i ms loca ed in places which do no equen ly expe ience loods.
Fu he mo e, lood e en s can ha e a - eaching, indi ec impac s on companies’ inancial pe o mance
h ough he global supply chain ne wo k. Recen ly, he Ge man ca manu ac u e Volkswagen announced
p oduc ion losses and suspension o ope a ions o some o i s ac o ies in Po ugal due o loods in
Slo enia which a ec ed some o i s engine pa supplie s (Reu e s,2023). In an ea lie s udy, Ha aguchi
and Lall (2015) documen declines in p o i abili y o up o abou 60 % in some Japanese au omake s due
o a lood which occu ed in Thailand. E iden ly, such subs an ial impac could nega i ely a ec i ms’
deb epaymen capaci y wi h a - eaching epe cussions on lending ins i u ions.
Unde s anding and managing clima e- ela ed inancial isks a ising om he escala ing equency
and in ensi y o ca as ophic e en s, called physical isk, is c ucial o main aining inancial s abili y.
I equi es assessing he ulne abili ies o inancial ins i u ions, in eg a ing clima e- ela ed ac o s in o
isk managemen amewo ks, and p omo ing clima e esilience and adap a ion measu es. The nascen
li e a u e shows some mixed e idence on he impac o wea he - ela ed na u al disas e s on he inancial
sec o . The s udy by Blickle e al. (2021) inds ha la ge banks a e ba ely a ec ed by ex eme wea he -
induced and hei income e en inc eases due o subsequen ly highe loan p o ision. The impac is mo e
p onounced in local banks bu s ill no la ge enough o cause sol ency conce ns. Simila ly, Caloia and
Jansen (2021) ind ha he Du ch banking sec o is su icien ly capi alised o wi hs and loods in a eas
whe e he eal es a e exposu e is low. Howe e , he impac would be la ge in case mo e se e e loods hi
he densely popula ed a eas o he coun y. The analysis om Calice and Miguel (2021) also inds a small,
ye s a is ically signi ican esul o la ge-scale na u al disas e s on banks’ c edi po olios. In pa icula ,
in La in Ame ican and Ca ibbean (LAC) coun ies banks’ NFC- ela ed nonpe o ming loans inc ease by
up o 2.5 pe cen age poin s pos -disas e . Two s udies ocus on small, egional inancial ins i u ions and
empi ically examine he indi ec impac o loods on hem. They bo h documen he ad e se e ec s o
loods on banks ei he by empo a ily inc easing hei impai ed loans (Shala and Schumache ,2022) o
2
by hinde ing hei p o i abili y (Paglia i,2023).
Policymake s ha e aken se e al policy ini ia i es which equi e banks o assess and disclose isks
a ising om clima e- ela ed na u al disas e s, in eg a e hese isks in o hei isk managemen amewo ks,
and align hei in es men po olios and inancial p oduc s wi h sus ainabili y c i e ia. The speci ic
equi emen s a y o each ini ia i e, bu he o e all goal is o os e a mo e clima e- esilien inancial
sec o by conside ing and add essing clima e physical isks. A he EU le el, se e al policy ac ions equi e
banks o assess, epo and in eg a e clima e- ela ed isks in o hei isk managemen amewo ks. The
Sus ainable Finance S a egy1expec s banks o conduc comp ehensi e isk assessmen s ha conside
he po en ial impac s o clima e change on hei business ac i i ies. As pa o he S a egy, Eu opean
Supe iso y Au ho i ies will wo k on clima e s ess es ing o he inancial sys em including a one-o
Fi - o -55 clima e isk scena io analysis in coope a ion wi h he Eu opean Cen al Bank (ECB) and he
Eu opean Sys emic Risk Boa d. The la e aims o assess he esilience o he inancial sec o and gain
insigh s in o i s capaci y o suppo he ansi ion o a lowe ca bon economy e en unde condi ions o
s ess. The Non-Financial Repo ing Di ec i e2 equi es banks unde i s scope o disclose in o ma ion
on hei policies, isks, and ou comes ela ed o en i onmen al ma e s, including clima e change and
physical isks. The EU Taxonomy3, which p o ides c i e ia o de e mining which economic ac i i ies can
be classi ied as en i onmen ally sus ainable, de ails how banks need o iden i y and epo on in es men s
o p oduc s ha con ibu e o clima e change adap a ion o mi iga e physical isks.
Agains his backg ound, he p ima y objec i e o his pape is o e alua e he e ec o localized
clima e- ela ed e en s, unde a ious global wa ming scena ios, on egional banks in Ge many. By doing
so, i aims o shed ligh on po en ial ulne abili ies and isks in he Ge man banking sec o , pa ing he
way o mo e in o med decision-making, in pa icula on isk mi iga ion s a egies.
The s udy ocuses on 677 coope a i e banks, which a e geog aphically mo e concen a ed in hei
exposu es, gi en hei s ong connec ion o he local economy compa ed o o he ypes o banks. We as-
sume ha he lending ac i i ies o a coope a i e bank a e conduc ed a he NUTS2 le el, which ep esen s
1Communica ion om he Commission o he Eu opean Pa liamen , he Council, he Eu opean Economic and Social Com-
mi ee and he Commi ee o he Regions S a egy o Financing he T ansi ion o a Sus ainable Economy, COM/2021/390
inal
2Di ec i e 2014/95/EU o he Eu opean Pa liamen and o he Council o 22 Oc obe 2014 amending Di ec i e 2013/34/EU
as ega ds disclosu e o non- inancial and di e si y in o ma ion by ce ain la ge unde akings and g oups. OJ L 330,
15.11.2014, p. 1-9
3Commission Delega ed Regula ion (EU) 2021/2178 o 6 July 2021 supplemen ing Regula ion (EU) 2020/852 o he
Eu opean Pa liamen and o he Council by speci ying he con en and p esen a ion o in o ma ion o be disclosed by unde -
akings subjec o A icles 19a o 29a o Di ec i e 2013/34/EU conce ning en i onmen ally sus ainable economic ac i i ies,
and speci ying he me hodology o comply wi h ha disclosu e obliga ion (Tex wi h EEA ele ance)m C/2021/4987 OJ L
443, 10.12.2021, p. 9–67
3
ela i ely la ge egions in e ms o a ea. While mo e de ailed da a a he NUTS3 le el would p o ide
mo e accu a e es ima es o he asse s ha may be a isk om acu e physical ac o s, we a e cons ained
by da a limi a ions and hus main ain his assump ion.4
Based on 2023 ECB analy ical indica o s on physical isks,5 he size o Ge man bank asse s cu en ly
a isk due o i e loods is 0.6% o hei po olios o non- inancial co po a ions. As he exposu e
o loods a ies wi hin he coun y, we dis ibu e he amoun o asse s ac oss di e en NUTS2 egions,
using he sha e o people exposed as ob ained om he Risk Da a Hub. To accoun o he s ain on
banks’ loan po olios ha can a ise om he impac o i e lood e en s, we conside he asse s a
isk as non-pe o ming loans. This is also one assump ion o he model, since pa o he losses om
non-pe o ming loans migh be eco e ed due o colla e al o ins ance. As global empe a u es ise due
o global wa ming, we assume ha non-pe o ming loans migh inc ease in p opo ion o he economic
losses caused by loods, as p ojec ed in PESETA IV (Feyen e al.,2022;Do o i e al.,2020).
A mic o-simula ion po olio model based on indi idual bank balance shee da a is used o de i e he
loss dis ibu ion o he banking sec o , conside ing a si ua ion in which losses om physical isk come
in addi ion o ini ial inancial sec o losses. Two cases a e conside ed: (a) a no mal economic si ua ion
whe e he impac on banks is only due o he physical isk, he ma e ializa ion o which igge s a small
de alua ion o asse s exposed modelled ia a dynamic balance shee ; (b) a c isis ype si ua ion whe e he
impac o physical isk is on op o exis ing inancial/economic c isis. Fo a s able economic si ua ion,
we conside only banks whose de aul s a e due o hei exposu e o i e lood e en s, and s a ing wi h
a mild ini ial de alua ion o clima e- ela ed loans o all banks, we in es iga e i , and how many, new
banks ail. We keep inc easing he le el o loans’ de alua ion, un il a new equilib ium is eached (i.e.
no ex a de aul s). The speci ic impac is a he small and e en ually becomes signi ican unde an
inc ease in global a e age empe a u e, wi h losses o a ound 0.2% o o al asse s unde a 1.5℃inc ease
in empe a u e, and a ound 1% unde a 3℃inc ease. When clima e isk ma e ializes on op o an exis ing
inancial c isis, he magni ude o losses anges om 0.4% o 0.9% o o al asse s, depending on he speci ic
clima e wa ming scena ios, and clima e- ela ed inancial isks a e concen a ed in ew egions pa icula ly
exposed o clima e- ela ed e en s (up o 5% o o al asse s). Finally, we in es iga e he impac o physical
isks when adap a ion s a egies a e implemen ed. The le el o isk associa ed wi h a 3℃inc ease in global
wa ming can be mi iga ed o he ex en ha i becomes compa able o he cu en si ua ion bo h in c isis
4I is wo h no ing ha he business a ea o egional banks may be smalle han NUTS2 a eas bu la ge han NUTS3
a eas.
5h ps://www.ecb.eu opa.eu/pub/pd /o he /ecb.clima e_change_indica o s202301˜47c4bbbc92.en.pd
4
• ℃ e e s o he inc ease in global empe a u e o a speci ic global wa ming scena io ( = 1.5, 3.0);
•∆L ℃indica es he inc ease in economic losses unde a ℃inc ease in empe a u e wi h espec o
oday;
•NP LFi, ℃ e e s o he new lood- ela ed losses unde he wa ming scena io (1.5℃, 3.0 ℃).
Using coun y da a om he PESETA IV s udy (Feyen e al.,2022), we de e mine he expec ed
inc ease in o e all economic losses esul ing om lood e en s unde a 1.5℃and 3℃inc ease in empe -
a u e, wi h and wi hou adap a ion s a egies. Based on his in o ma ion, we p ojec he size o i e
loods- ela ed non-pe o ming loans unde ou global wa ming scena ios: cu en clima e condi ions, 1.5℃
inc ease wi hou adap a ion, 3℃inc ease wi hou adap a ion, and 3℃inc ease wi h adap a ion s a e-
gies. These alues a e inpu s o he model desc ibed in he nex sec ions and hei PESETA unde lying
alues a e epo ed in Table 1.
This assump ion is a simpli ica ion ha should be app oached wi h cau ion. While i helps o add ess
he complexi y o he issue, i has he po en ial o o e simpli y he si ua ion, in oducing unce ain y
and po en ially o e looking impo an ac o s. Speci ically, we a e assuming ha he o e all economy
emains s a ic, he s uc u e o he banking sys em emains unchanged, and he isk appe i e o inancial
ins i u ions emains he same. I is impo an o ecognize ha in eali y, hese ac o s a e subjec o
change and can signi ican ly impac he ela ionship be ween economic losses and non-pe o ming loans.
Table 1: Va ia ion in economic losses unde h ee empe a u e inc eases
Inc ease Inc ease Dec ease wi h adap a ion
unde 1.5 ℃unde 3 ℃unde 3 ℃
95% 304% 80%
3 Simula ion model
The modelling amewo k elies on he Sys emic Model o Banking O igina ed Losses (SYMBOL, see
De Lisa e al.,2011), a mic o simula ion po olio model which uses bank-le el da a o simula e banking
c isis scena ios. Speci ically, indi idual banks may de aul based on hei ac ual capi al (K) and on he
p obabili y o de aul a ached o hei po olio (P D). The bank’s po olio P D is calcula ed by in e ing
he Basel o mula, which ela es he minimum capi al equi emen s o he P D o a bank’s po olio and
i s RW A as ollow:
11

CR(P D) = LGD ·N"√RN−1(0.08·RW A
T A ) + N−1(P D)
√1−R#−P D, (5)
whe e Riis he co ela ion among exposu es in he po olio, LGD is he loss gi en de aul equal o 0.45
as pe egula ion13,Nis he no mal dis ibu ion unc ion.
Bank-speci ic P Ds a e hen used o gene a e losses o indi idual banks ia a Mon e Ca lo simula ion,
whe e andomness is in oduced by sampling he unde lying shocks. While p o isions and w i e-o s should
co e expec ed losses on an ongoing basis, unexpec ed losses e e o seldom bu po en ially signi ican
losses ha should be co e ed by capi al. The ailu e o a bank is de e mined by compa ing he size o hese
unexpec ed losses o he ac ual egula o y capi al a ailable o abso bing hem. In o he wo ds, banks
ail when simula ed unexpec ed losses exceed he o al ac ual capi al.14 Some o he model i e a ions,
pa icula ly hose whe e la ge sampled shock, will likely esul in a leas one bank de aul ing.
The ou pu o he Mon e Ca lo simula ion is an I×Jma ix o unexpec ed losses (GL)ij, ob ained
as ollows:
GLij =LGD ·N"s1
1−Ri
N−1(P Di) +sRi
1−Ri
N−1(αij)#−ELi,(6)
whe e he i s e m gene a es he o al amoun o losses and he second one (ELi) app oxima es he
expec ed losses.15 Speci ically, i= 1, . . . , I e e s o he banks in he sample, j= 1, . . . , J = 10,000
deno es he model i e a ion. N−1(αij) a e co ela ed no mal andom shocks. The αij shocks a e co ela ed
as hey a e de ined as he sum o a common shock Zjand a bank-speci ic shock Wij, as ollows:
N−1(αij) = l·Zj+p1−l2·Wij,(7)
whe e la e he loadings, Wij a e he idiosync a ic shocks, and Zjis a common shock which migh be
linked, o example, wi h o e all economic de elopmen s. The s anda d e sion o he model, which is
used in his pape , se s lso as o yield a ixed co ela ion o 0.5 ac oss he αij .16 The shocks Zjand Wij
a e d awn om a s anda dized no mal dis ibu ion as in he Vasiceck model (see Vasicek (1977); De Lisa
e al. (2011)).
In each i e a ion, he ollowing h ee componen s a e summed up o each bank: i) losses ha canno
13h ps://www. dic.go /analysis/c /wo king-pape s/2006/2006-10.pd
14See Benczu e al. (2017) o de ails.
15Expec ed loss is calcula ed as he p oduc o P D,LGD, and o al asse s.
16See Di Gi olamo e al. (2017) o a discussion on a mo e sophis ica ed co ela ion s uc u e.
12
be abso bed by capi al, and ii) ecapi aliza ion needed o b ing he banks back o a iabili y s a us, i.e.
a o al egula o y capi al a io o 8% o RW As and iii) clima e- ela ed inancial losses (LF) o assess he
impac o clima e change on he bank’s balance shee . The amewo k subsequen ly assesses whe he he
banks’ capi al is high enough o abso b losses and clima e- ela ed losses o i hese could po en ially lead
o addi ional ailu es wi hin he banking sec o .
F om he e on, we hen de ine as losses o each bank and each i e a ion, he unexpec ed losses in
excess o capi al plus ecapi aliza ion needs (ExLRji), as ollows:
ExLRij = max(GLij −Ki+ 8%RW Ai+LF
i,0),(8)
o cou se, LF
ia e aken in o conside a ion jus when physical isk ma e ializes. Excess losses plus ecapi-
aliza ion needs o indi idual banks yield he ollowing agg ega e loss dis ibu ion o each i e a ion:
Lj=
I
X
i=1
ExLRij.(9)
Each poin in he dis ibu ion (i.e. each i e a ion) is associa ed wi h a di e en le el o inan-
cial/economic dis ess. Using indi idual banks’ dis ibu ion, hey can be agg ega ed a di e en scales
(e.g. coun y, NUTS2, e c.).17
4 Physical isk as a igge o a c isis
4.1 Modelling amewo k
In his applica ion o he model, we in es iga e unde which condi ions i e lood e en s could be he
igge o a c isis o coope a i e banks in Ge many e en in he absence o a la ge economic shock.
To do so, we s a by compa ing SYMBOL esul s wi h and wi hou clima e ela ed inancial losses o
iden i y hose banks ha would ail due o a ma e ializa ion o physical isk bu would no ail o he wise.
Essen ially, based on he simula ion esul s, we iden i y i e a ions whe e banks de aul when i e - lood
17This modelling amewo k can accommoda e di e en deg ees o commonali y by allowing o a ious shock co ela ion
s uc u es, while a con agion mechanism is no explici ly conside ed. We do no explici ly model con agion e ec s h ough
he in e bank ma ke o h ee main easons. Fi s , con agion would in oduce an addi ional laye o complexi y, inc easing
he unce ain y a ound he esul s. Second, modelling con agion dynamics is qui e challenging, as public da a do no i
o he pu pose o calib a ing he ne wo k. Thi d, he comp ehensi e c isis managemen and deposi insu ance amewo k
pu o wa d by he Commission a e he las inancial c isis (no ably, he bail-in ools and he es ablishmen o esolu ions
unds), is expec ed o p e en he sp eading o con agion as dis essed banks would be esol ed, liquida ed o ecapi alised
well be o e di ec con agion e ec s could ma e ialize (see, e.g. Benczu e al. (2017).
13
non-pe o ming loans a e aken in o accoun bu do no de aul wi hou . Fo mally, hese co espond o
he uns sa is ying he ollowing condi ion:
ej:= jsuch ha

















ExLRij = 0 wi hou physical isk,
and
ExLRij >0 wi h physical isk.
.(10)
Then, we model a dynamic mechanism whe eby, owing o hese ini ial de aul s, a de alua ion o exposed
asse s akes place, leading o u he ailu es and losses in he banking sec o . In pa icula , we assume
ha ini ial bank de aul s (only due o clima e) may igge a sequence o de alua ion dynamics o asse s
exposed, which is e lec ed on he one hand, in a lowe sha e o exposed asse s on banks’ balance shee s,
and on he o he hand, in lowe ma ke p ices o hose asse s. The mechanism is modelled as ollows.
Ini ially, when physical isk om i e - lood e en s eme ges, he e is an immedia e de alua ion o asse s
ha a e exposed o such isks. This de alua ion is de e mined by he p opo ion o non-pe o ming loans
a ibu ed o i e loods in ela ion o he o e all size o he exposed asse s. The ex en o de alua ion
a each ound is di ec ly linked o a ia ions in clima e- ela ed losses. This means ha he size o he
de alua ion is p opo ional o he changes obse ed in losses associa ed wi h clima e- ela ed e en s. The
mechanism con inues un il he sys em eaches a new equilib ium, i.e. no mo e banks de aul .
Final loss (F L) due o physical isk is hen compu ed by aking he expec ed alue o agg ega e bank
losses (as de ined in Equa ion 9) ac oss he selec ed i e a ions. This co esponds o he Expec ed Sho all
concep :
F L =Ej[(L)j|jin ej],(11)
Clima e- ela ed losses, LF(see Equa ion 8), ake in o accoun i s and second- ound e ec s om he
de alua ion o exposed asse s (see Sec ion 2):
LF
=

















NP LFi = 0
F L +NP LF
T AF
e,0×T AF
e, i = 1
F L +F L −1−F L −2
T AF
e, −1×T AF
e, i >1,
(12)
14
whe e T AF
e e e s o he asse s exposed o i e lood:
T AF
e, =






29.8% ·T A i = 0 as es ima ed by he ECB
T AF
e, −1−F L −1i >0.
(13)
4.2 Resul s
Based on he SYMBOL simula ion and conside ing clima e- ela ed losses om i e - lood e en s, ou
indings sugges ha he impac on Ge man coope a i e banks is negligible oday. Howe e , his could
become a mo e signi ican conce n i he empe a u e inc eases by 1.5 ℃and e en wo se unde a 3 ℃
inc ease.
Table 2p esen s he pe cen age o banks ha de aul a each s ep, while Table 3shows he co e-
sponding pe cen age o asse s a ec ed. A s ep 0, ou esul s indica e ha physical isk ini ially leads
o a ela i ely small numbe o ailu es in he banking sys em. Speci ically, 0.2% o he o al numbe o
banks in ou sample a e in de aul , which also co esponds o 0.2% o o al Ge man asse s. Unde he
mos se e e clima e condi ions, he ini ial sha e o ailed banks inc eases o 7% o he o al numbe o
banks, a ec ing 5% o o al asse s (see Table 2and Table 3). The dynamic ea u e in oduced o simula e
he de alua ion o exposed asse s leads o mo e banks de aul ing in he sys em. While he inal numbe
o de aul s does no inc ease signi ican ly unde he cu en empe a u e le el, he wo s global wa ming
scena io could esul in he ailu e o 29% o banks, which hold app oxima ely 24% o he o al asse s.
In e ms o inancial losses, he impac is ela i ely small unde he cu en clima e condi ions bu could
become signi ican as empe a u es inc ease up o 3℃. Speci ically, agg ega ed losses amoun o 0.02%
o o al asse s a he p esen empe a u e le el, 0.2% unde a empe a u e inc ease o 1.5°C, and 1.1%
unde a 3°C inc ease (see Table 4).18 These esul s show ha should such isk ma e ialises nowadays
he ecou se o public inances would be con ained. Howe e , an inc ease in empe a u e would pose
challenges o he co po a e sec o ha could des abilize he Ge man egional banks, a leas in some
egions.
When examining speci ic NUTS2 egions, ou esul s indica e ha a ew egions could be pa icula ly
s ongly a ec ed, wi h losses eaching up o 10% o o al asse s unde a 1.5 ℃ empe a u e inc ease.
Figu e 6shows ha he e a e a leas h ee egions wi h a sha e o losses o e o al asse s la ge han 5%,
18I is impo an o no e ha in eali y, he impac would be much mo e con ained due o bank eco e y mechanisms and
oubled asse pu chase p og ams ha would be igge ed be o e a la ge numbe o banks become insol en .
15
e en unde a 1.5°C empe a u e inc ease (and e en highe unde a 3 ℃ empe a u e inc ease). Compa ing
he wo panels in Figu e 6, one can see ha he e a e jus ew NUTS2 egions con ibu ing o he o e all
losses.
Cu en 1.5 ℃inc ease 3.0 ℃inc ease
S ep 0 0.2% 3% 7%
S ep 1 3% 7% 17%
S ep 2 3% 8% 23%
S ep 3 3% 8% 25%
S ep 4 3% 8% 27%
S ep 5 3% 8% 27%
S ep 6 3% 8% 28%
S ep 7 3% 8% 28%
S ep 8 3% 8% 29%
Table 2: Numbe o banks de aul ed o e he o al numbe o banks in he sample (c isis igge ed by
physical isk)
Cu en 1.5 ℃inc ease 3.0 ℃inc ease
S ep 0 0.2% 1% 5%
S ep 1 1% 4% 15%
S ep 2 1% 5% 19%
S ep 3 1% 5% 22%
S ep 4 1% 5% 23%
S ep 5 1% 5% 24%
S ep 6 1% 5% 24%
S ep 7 1% 5% 24%
S ep 8 1% 5% 24%
Table 3: Asse s o banks de aul ed o e he o al amoun o asse s in he sample (c isis igge ed by
physical isk)
Cu en 1.5 ℃inc ease 3.0 ℃inc ease
S ep 0 0% 0% 0.1%
S ep 1 0.02% 0.1% 0.6%
S ep 2 0.02% 0.2% 0.9%
S ep 3 0.02% 0.2% 1%
S ep 4 0.02% 0.2% 1.1%
S ep 5 0.02% 0.2% 1.1%
S ep 6 0.02% 0.2% 1.1%
S ep 7 0.02% 0.2% 1.1%
S ep 8 0.02% 0.2% 1.1%
Table 4: Size o losses in he sample as sha e o TA (c isis igge ed by physical isk)
16

Figu e 6: Losses o e TA dis ibu ions ac oss NUTS2 egions unde di e en empe a u e le els (c isis
igge ed by physical isk). Le panel: ull sample, igh panel: wi h no ou lie s. No e: The middle line
is he median, he obse a ion sepa a ing he sample in wo hal es. The ex emes o he box a e he i s
and hi d qua ile, hal o he sample is included inside he box, and a qua e is ou side on each side.
The whiske s a e he highes o lowes obse a ion lying wi hin 1.5 imes he leng h o he box om he
median.
5 A c isis no igge ed by physical isk
5.1 Modelling amewo k
In his sec ion, we use he model p esen ed abo e o assess he impac o i e - lood e en s, du ing a
sys emic banking c isis which is no igge ed by clima e c ises. We do so by looking a he e ec o
including i s -o de clima e- ela ed losses (LF) in Equa ion 8, wi hou he ampli ica ion mechanism.
Hence, LFis equal o i e loods- ela ed non-pe o ming loans (NP LF).
We ocus on he ( e y) igh ail o he loss dis ibu ion, which is associa ed wi h a se e e, bu
plausible, banking c isis. Technically, his pa o he dis ibu ion co esponds o alues o Zjwhich a e
a he han 3 s anda d de ia ions om he mean.19 This co esponds o he ollowing i e a ions ej:
ej={jsuch ha Zj>E(Zj) + 3s d(Zj)}.(14)
Final losses (F L) a e hen compu ed as in Equa ion 11 by aking he expec ed alue o agg ega e bank
losses (as de ined in 9) ac oss he selec ed i e a ions. By compa ing esul s wi h and wi hou physical
19As Zjmay be seen as a nega i e economic shock, his calib a ion is ep esen a i e o a ecession compa able o hose
obse ed on he occasion o he global inancial c isis and he Co id-19 pandemic.
17
isk one can de i e he size o inal losses when i e lood e en s ma e ialize.
5.2 Resul s
In a c isis si ua ion, app oxima ely 14% o banks ( ep esen ing nea ly 15% o o al asse s) de aul , esul ing
in losses o 0.3% o o al asse s in he Ge man banking sys em. The inclusion o clima e- ela ed ac o s
exace ba es he si ua ion, as physical isk con ibu es o mo e losses and ailu es. While he con ibu ion
o addi ional physical isks o o e all losses is ela i ely modes unde he cu en clima e condi ions, i
becomes signi ican unde a clima e wa ming scena io. In e ms o numbe o banks in ol ed, Tables 5
shows ha lood e en s would cause a 4% inc ease in de aul s among coope a i e banks unde he cu en
clima e si ua ion, and a 17% inc ease unde a empe a u e ise o up o 3 ℃. In e ms o losses, o 1.5 ℃,
agg ega e losses, as a sha e o o al asse s, inc ease om 0.3% o 0.5% on a e age and up o 0.9% unde
a 3 ℃.
Figu e 7shows he inc ease in losses: he si ua ion is qui e he e ogeneous ac oss egions, wi h some
NUTS2 subjec o e y mild (o almos ze o) impac s and o he s whe e losses can inc ease subs an ially
wi h espec o a si ua ion wi h no clima e isk. This is he case in pa icula o h ee egions, ha ing a
mo e han 5% inc ease o he 1.5 ℃ empe a u e inc ease scena io, and up o 30% in he 3 ℃ empe a u e
inc ease scena io.
Wi hou physical isk Wi h physical isk
Cu en Cu en 1.5 ℃inc ease 3.0 ℃inc ease
N.Banks 14% 18% 22% 31%
Asse s 15% 18% 22% 30%
Table 5: Numbe o banks de aul ed o e he o al numbe o banks in he sample and asse s in ol ed
o e he o al amoun o asse s in he sample (sys emic banking c isis)
Wi hou physical isk Wi h physical isk
Cu en Cu en 1.5 ℃inc ease 3.0 ℃inc ease
Losses (%TA) 0.3% 0.4% 0.5% 0.9%
Table 6: Size o losses in he sample as sha e o TA (sys emic banking c isis)
18
Figu e 7: Losses’ ela i e inc ease wi h espec o Baseline dis ibu ions ac oss NUTS2 egions, unde
h ee empe a u e le els (sys emic banking c isis). Le panel: ull sample, igh panel: wi h no ou lie s
6 Adap a ion s a egies and impac on he banking sec o
PESETA IV s udy (Feyen e al.,2022) quan i ies ha , unde a 3 ℃global empe a u e inc ease by he
end o he cen u y, in he EU+UK, i e loods ela ed losses would inc ease by six imes, wi h mo e han
hal a million people exposed e e y yea .20 The implemen a ion o e ec i e adap a ion s a egies plays a
c ucial ole in mi iga ing economic losses and enhancing he esilience o i ms o clima e isks. By p oac-
i ely implemen ing and in eg a ing adap a ion measu es, businesses can minimize hei ulne abili y o
clima e- ela ed e en s, he eby educing he size o po en ial economic losses and enabling businesses o
be e an icipa e o eco e om clima e- ela ed shocks. The Eu opean Commission adop ed i s new
EU Adap a ion s a egy on 24 Feb ua y 2021. The ou main p inciples o he s a egy a e a sma e
adap a ion, a sys emic and as e se o ac ions, and in e na ional suppo o clima e change adap a ion
and common policies.
Due o he unce ain y su ounding u u e scena ios, quan i ying clima e damages and de e mining
e ec i e adap a ion s a egies emains a di icul ask. Howe e , PESETA IV has aken an ini ial s ep
by explici ly modeling adap a ion s a egies and es ima ing he po en ial educ ion in losses. Acco ding
o he epo , adap a ion measu es would be e y e ec i e in educing popula ion exposed and po en ial
economic losses, p o ec ing ci izens, p ese ing economic ac i i ies, and building a mo e esilien inancial
sys em. Among o he s, he educ ion o lood peaks ins alling e en ion ese oi s could educe by a ound
20I one compa es hese numbe s wi h he p esen si ua ion, losses mo e om 7.8 o 48 billion pe yea and people exposed
go om 172000 o 4820000
19
EUR 40 billion he losses ela ed o i e loods, and a ound 400,000 ewe people would be exposed each
yea . O he adap a ion measu es such as he es o a ion o loodplain a eas, s eng hening he exis ing
dyke sys em, and implemen ing building-based damage educ ion measu es a e e ec i e in u he educing
he o iginal impac . O e all, PESETA es ima es poin o a 80% educ ion o he impac o clima e- ela ed
e en s unde a 3 ℃(see Table 1).21.
In ou modelling amewo k, we apply his educ ion o he es ima ed amoun o i e lood- ela ed
non-pe o ming loans unde a 3℃inc ease in empe a u e and we un ou analysis o es ima e he size o
inal losses unde a no mal economic si ua ion and a banking c isis. Resul s show ha he inancial sec o
would bene i signi ican ly, as he likelihood o ei he sudden asse de alua ions and inal losses d i en
by clima e- ela ed e en s diminishes o a poin whe e hey a e compa able o he p esen ci cums ances,
du ing a sys em c isis and in no mal imes (see Figu e 8and 9).
Figu e 8: Losses’ dis ibu ion (c isis igge ed by physical isk) o Cu en and 3.0℃inc ease wi h
adap a ion. Le panel: ull sample, igh panel: wi h no ou lie s
21As a ca ea , PESETA IV highligh ha es ima es a e su ounded by ele an unce ain y and as such cos s and bene i
migh be la ge bo h in e ms o economic losses and educ ions due o adap a ion s a egies
20
The Eu opean Commission’s
science and knowledge se ice
Join Resea ch Cen e
JRC Mission
As he science and knowledge se ice
o he Eu opean Commission, he Join
Resea ch Cen e’s mission is o suppo
EU policies wi h independen e idence
h oughou he whole policy cycle.