RESEARCH AND ANALYSIS Scenario Analysis -...
Transcript of RESEARCH AND ANALYSIS Scenario Analysis -...
R E S E A R C H A N D A N A LYS I S
Scenario AnalysisExploring the Macroeconomic Impacts ofInformation and CommunicationTechnologies on Greenhouse Gas Emissions
Lorenz Erdmann and Lorenz M. Hilty
Keywords:
impact assessmentindustrial ecologyinformation and communication
technology (ICT)rebound effectsystem dynamicsuncertainty
Supplementary material is availableon the JIE Web site
Address correspondence to:Lorenz ErdmannAssociate Research ScientistInstitute for Futures Studies and
Technology AssessmentSchopenhauerstr. 26 Berlin [email protected]
c© 2010 by Yale UniversityDOI: 10.1111/j.1530-9290.2010.00277.x
Volume 14, Number 5
Summary
During the past decade, several macroeconomic studies onthe potentials of information and communication technology(ICT) to reduce greenhouse gas (GHG) emissions have beenpublished. The mitigation potentials identified in them vary toa high degree, mainly because they are not consistently definedand diverse methodologies are applied. The characteristics ofICT—exceptional dynamics of innovation and diffusion, socialembedment and cross-sector application, diverse and com-plex impact patterns—are a challenge for macroeconomicstudies that quantify ICT impacts on GHG emissions.
This article first reviews principal macroeconomic studieson ICT and GHG emissions. In the second part, we recon-sider our own study on this topic and present an in-depthscenario analysis of the future impacts of ICT applications onGHG emissions. We conclude that forthcoming macroeco-nomic studies could strengthen the state of the art in en-vironmental ICT impact modeling (1) by accounting for thedynamics of new ICT applications and their first-, second-,and third-order effects on a global scale, (2) by reflecting theerror margins resulting from data uncertainty in the final re-sults, and (3) by using scenario techniques to explore futureuncertainty and its impacts on the results.
824 Journal of Industrial Ecology www.wileyonlinelibrary.com/journal/jie
R E S E A R C H A N D A N A LYS I S
Introduction
Applications of information and communi-cation technology (ICT) are expected to re-duce transport volumes, as well as energy andmaterial demand. Such effects of ICT appli-cations are closely coupled to mass flows intothe environment, such as emissions into air,water, and soil. In particular, the potential ofICT to reduce greenhouse gas (GHG) emis-sions is increasingly acknowledged as relevantin intergovernmental policy documents (e.g.,Organization for Economic Cooperation andDevelopment 2009). The extent, however, towhich ICTs might help governments to achieveGHG emission reduction targets in the mid-to long term is discussed controversially amongexperts.
ICT components, products, and servicesare constantly being invented and reinvented(Rejeski 2003), thus providing the basis for anincreasing diversity of new applications. If inno-vation is followed by rapid diffusion, macroeco-nomic impacts of ICT applications may materi-alize fast. ICT applications pervade, transform,and link all economic sectors. In improving theefficiency of production and consumption pro-cesses and in enabling new activities and so-cietal relations, ICT is an important driver fornew lifestyles, structural change (i.e., the sharecomprised by new industries in the economy in-creases in relation to old industries), and eco-nomic growth.
This particular set of characteristics of ICT isa challenge to deal with in macroeconomic stud-ies on ICT effects on GHG emissions. There aretwo basic ways to quantify these effects: (1) in-corporate ICTs in general macroeconomic mod-els explicitly or (2) build specific macroeconomicICT impact assessment models.
Studies of the first type commonly conceptu-alize ICT as a driver of technological efficiency,economic growth, and structural change (Rommet al. 1999; Kuhndt et al. 2003). This article fo-cuses on studies of the second type. They are oftenbuilt upon a distinction among first-, second-, andthird-order effects of ICT on the environment(Fichter 2003; Kohler and Erdmann 2004; Hilty2008):
1. First-order effects: The impacts of the life-cycle of ICT hardware, for example, theenergy consumed by the ICT hardware ofan intelligent transport system (ITS).
2. Second-order effects: The impacts of theservices provided by ICT applications, forexample, the energy saved in transport byusing an ITS.
3. Third-order effects (including rebound ef-fects): The emerging effects of ICT in theeconomic system, for example, increaseddemand for transport as a long-term impli-cation of efficiency gains induced by ITSapplications.
The dynamic impacts of ICT originate fromthe feedback of third-order effects to first- andsecond-order effects. If an increase in efficiency(second-order effect) is expected to reduce theenvironmental impact of an activity, but addi-tional demand (third-order effect) compensatesfor the reduction, this is a typical rebound ef-fect (Hilty 2008). ICT has the potential to en-hance the efficiency of production and consump-tion processes with regard to cost, time, energy,material, and other factors in the entire economicsystem significantly. Because one of the core prin-ciples of industrial ecology is to view systemsin their entirety and full complexity (Allenby1999), it seems indispensable to account for third-order effects—in particular rebound effects—inany macroeconomic assessment of ICT impactson GHG emissions.
In the next section, major ICT impact assess-ment studies are reviewed, taking these particularcharacteristics of ICT into account. We will thenpresent a new interpretation of our own study,consisting of a scenario analysis for ICT applica-tions based on the system dynamics model devel-oped in that study. In the final section we willdraw conclusions for future macroeconomic ICTimpact assessment studies.
Review of Existing Studies
In a recent review of research on the envi-ronmental impacts of e-business and ICT, Yi andThomas (2007) identified a diverse body of lit-erature ranging from micro-level to macro-level
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 825
R E S E A R C H A N D A N A LYS I S
impact assessments, from sector studies to cross-sector studies, from qualitative to quantitativeassessments. The authors of these studies weremainly motivated by the environmental impli-cations of the rise of the Internet and the “neweconomy.” The references dating back to 1998belong to a period that we classify as the “firstgreen ICT wave.”
The publication of the Yi and Thomas (2007)review almost coincided with the issuance of theFourth Assessment Report of the Intergovern-mental Panel on Climate Change in February2007 (IPCC 2007). Since then we have observeda “second green ICT wave,” in which a growingbody of literature focusing on the potentials ofICTs to reduce GHG emissions has been pub-lished. Almost all of these studies cover new ap-plications of ICT.
The present review looks at the mostimportant macroeconomic cross-sector studiesthat have been published in the past decade(table 1).1 We exclude studies that did not breaktheir macro-level assessments down to ICT appli-cations (namely, Laitner and Ehrhardt-Martinez2008; Madlener 2008). However, we make threeexceptions for older studies that stimulated thescientific debate substantially (Romm et al. 1999;Laitner et al. 2001; Kuhndt et al. 2003).
Study Type
The macroeconomic studies reviewed differ intheir basic approaches (see table 1). Some stud-ies are based on a “bottom-up” approach. Theyanalyze ICT applications at the micro-level (e.g.,efficiency gains in specific processes) and aggre-gate the impacts to the meso- and macro-level.Other studies follow a “top-down” approach byanalyzing macroeconomic indicators in order toisolate an ICT impact. Two studies take a “hy-brid” approach. They are anchored in macroeco-nomic analyses, but attribute the ICT sector im-pacts to certain ICT applications (Laitner et al.2001; Matsumoto et al. 2005). Seven studies weanalyzed include prospective applications of ICT,whereas three studies did not disaggregate to thelevel of ICT applications.
Erdmann and colleagues (2004) tailoreda system dynamics model. Laitner and col-leagues (2001) implemented their calculations
in the U.S. National Energy Modeling System.Matsumoto and colleagues (2005) used acomputable general equilibrium model to recal-culate adjusted input/output tables. Two stud-ies applied statistical methods (Romm et al.1999; Kuhndt et al. 2003). All other quantita-tive macroeconomic studies we reviewed calcu-late their results—to our knowledge—with basicarithmetic.
The studies differ in other key characteristics,namely, (1) goal, scope, and system boundaries,(2) data and methodologies, and (3) results.
Goal, Scope, and System Boundaries
The goal of all the studies reviewed is to ana-lyze the macroeconomic impact of ICT on GHGemissions and/or on energy consumption. Mostof the studies put particular emphasis on ICTapplications that produce environmental gains.The geographic scope ranges from national toglobal. The two global studies (Buttazoni 2008;Webb 2008) also provide a geographic resolutionof world economic regions. The time horizon ofthe studies with disaggregated ICT applicationsspans 10 to 25 years.
There is a core of ICT applications that arecovered by almost all studies, whereas some ap-plications are considered by one or a few studiesonly.
ICT in the production sector is broadly ad-dressed by the topics process automation andsmart motors. Several studies address ICT ap-plications in the energy sector by consideringsmart grids, smart metering, and smart appliances,which comprise a basis for an optimized elec-tricity supply and demand management. Webb(2008) also examines the impact of ICT on en-ergy conversion efficiency explicitly.
Smart buildings are covered by most ofthe studies. Applications range from ICT-controlled heating, ventilation, and air condi-tioning (HVAC) systems (e.g., Labouze et al.2008) to presence-based power management ofelectronic devices (Mallon et al. 2007). OnlyButtazoni (2008) quantifies the impact of smartcity planning on transport.
Intelligent transport systems comprise ICT-supported infrastructure and in-vehicle naviga-tion systems, and are included in most of the
826 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
Tabl
e1
Syno
psis
ofth
eSt
udie
sRe
view
ed
Cha
ract
erist
icof
the
stud
yC
EC
Sst
udy
EPA
/LB
NL
stud
yD
igita
lEur
ope
stud
yIP
TS
stud
yN
EC
stud
yT
elst
rast
udy
Eco
fyss
tudy
WW
FC
anad
ast
udy
Bio
inte
llige
nce
stud
yG
eSI
stud
y
Ref
eren
ceR
omm
etal
.(1
999)
Lait
nere
tal.
(200
1)K
uhnd
teta
l.(2
003)
Erdm
ann
and
colle
ague
s(20
04)M
atsu
mot
oan
dco
lleag
ues(
2005
)Mal
lon
and
colle
ague
s(2
007)
But
tazo
ni(2
008)
Gib
son
and
colle
ague
s(2
008)
Labo
uze
and
colle
ague
s(2
008)
Web
b(2
008)
Stud
yty
peB
asic ap
proa
chT
op-d
own
(and
sect
orst
udie
s)
Hyb
rid
Top
-dow
n(a
ndca
sest
udie
s)B
otto
m-u
pH
ybri
dB
otto
m-u
pB
otto
m-u
pB
otto
m-u
pB
otto
m-u
pB
otto
m-u
p
Gra
nula
rity
Hig
h-le
vel
indi
cato
rsH
igh-
leve
lse
ctor
cate
gori
es
Hig
h-le
vel
indi
cato
rsIC
Tap
plic
atio
nIC
Tap
plic
atio
nIC
Tap
plic
atio
nIC
Tap
plic
atio
nIC
Tap
plic
atio
nIC
Tap
plic
atio
nIC
Tap
plic
atio
n
Cal
cula
tion
met
hod
Stat
isti
calt
ime
seri
esan
alys
isIn
tegr
ated
mod
elin
gsy
stem
(NEM
S)
Stat
isti
cal
deco
mpo
siti
onan
alys
is
Syst
emD
ynam
ics
mod
elC
ompu
tabl
eG
ener
alEq
uilib
rium
mod
el
Bas
icar
ithm
etic
Bas
icar
ithm
etic
Bas
icar
ithm
etic
Bas
icar
ithm
etic
Bas
icar
ithm
etic
Scop
eG
eogr
aphi
cfo
cus
USA
USA
Euro
peEu
rope
Japa
nA
ustr
alia
Glo
bal
Can
ada
Euro
peG
loba
l
Geo
grap
hic
reso
luti
onU
SAU
SAG
erm
any,
Ital
y,U
KEU
15an
dth
e10
Acc
essi
onC
ount
ries
Japa
nA
ustr
alia
Var
ying
econ
omy
grou
ping
s
Can
ada
EU27
Chi
na,E
urop
e,In
dia,
Nor
thA
mer
ica
Bas
eye
ar19
9820
0019
95–1
997
(D)
1995
–199
8(U
K)
2000
2000
2005
2005
2006
2005
2002
Tim
eho
rizo
n20
0720
10—
2020
2010
2015
2030
2020
2020
2020
Cov
erag
eof
effe
cts
quan
tita
-ti
vely
Stru
ctur
alch
ange
and
tech
nolo
gy
Tec
hnol
ogy,
stru
ctur
alch
ange
,in
tegr
atin
g/re
boun
d
Gro
wth
,st
ruct
ural
chan
ge,
tech
nolo
gy,
priv
ate
hous
ehol
ds
Firs
t,se
cond
and
thir
dor
der
Firs
tand
seco
ndor
der/
tech
nolo
gy,
stru
ctur
alch
ange
Seco
ndor
der
(firs
tord
erex
clud
edfr
omag
greg
atio
n)
Seco
ndor
der
Seco
ndor
der
(firs
tord
erex
clud
edfr
omag
greg
atio
n)
Firs
tand
seco
ndor
der
Firs
tand
seco
ndor
der
Con
tinu
ed
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 827
R E S E A R C H A N D A N A LYS I S
Tabl
e1
Con
tinue
d
Cha
ract
erist
icof
the
stud
yC
EC
Sst
udy
EPA
/LB
NL
stud
yD
igita
lEur
ope
stud
yIP
TS
stud
yN
EC
stud
yT
elst
rast
udy
Eco
fyss
tudy
WW
FC
anad
ast
udy
Bio
inte
llige
nce
stud
yG
eSI
stud
y
Dat
aan
dm
etho
dolo
gyK
eyda
taso
urce
san
das
-su
mpt
ions
EIA
EIA
,Rom
met
al.(
1999
),ow
nas
sess
men
ts
Offi
cial
NA
MEA
data
Offi
cial
EUda
ta.
Expe
rtan
dow
nas
sess
men
ts.
Inpu
t/O
utpu
ttab
le,
adju
sted
byow
nas
sess
men
ts
Offi
cial
Aus
tral
iada
ta.
Ow
nas
sess
men
ts.
WEO
,WB
CSD
.O
wn
asse
ssm
ents
.
“rea
listi
cas
sum
ptio
ns”
Offi
cial
EUda
ta.
Ow
nas
sess
men
ts.M
cKin
sey.
Ow
nas
sess
men
ts.
Tre
atm
ento
fda
taun
cer-
tain
ty
Uns
yste
mat
ic,
qual
itat
ivel
y;fin
ales
tim
atio
nas
ada
tara
nge
Uns
yste
mat
ic,
qual
itat
ivel
yU
nsys
tem
atic
,qu
alit
ativ
ely
Bes
tcas
e/w
orst
case
sim
ulat
ion
Uns
yste
mat
ic,
qual
itat
ivel
y.U
nsys
tem
atic
,qu
alit
ativ
ely.
Syst
emat
ic,
qual
itat
ivel
y.U
nsys
tem
atic
quan
tita
tive
ly.U
nsys
tem
atic
,qu
alit
ativ
ely.
Uns
yste
mat
ic,
quan
tita
tive
ly.
Uns
yste
mat
ic,
qual
itat
ivel
y.
Tre
atm
ento
ffu
ture
un-
cert
aint
y
One
esti
mat
ion
ofen
ergy
inte
nsit
ych
ange
Fore
cast
(ref
eren
ceca
se,a
djus
ted
refe
renc
eca
se)N
opr
ojec
tion
squ
anti
fied
Scen
ario
s(T
echn
ocra
cy,
Gov
ernm
ent
Firs
t,an
dSt
akeh
olde
rD
emoc
racy
)
Scen
ario
s(W
LAN
,Se
amle
ssm
obile
netw
ork,
Coo
pera
tive
devi
cene
twor
k,C
onte
xtse
nsit
ive
netw
ork)
One op
port
unit
ype
rapp
licat
ionPo
tent
ials
(low
,med
ium
,hi
gh)
Goa
ls(s
hort
term
,str
etch
)Sc
enar
ios
(Bas
elin
e,B
aUEc
o)
Scen
ario
s(B
aU,
SMA
RT
2020
)
Bas
elin
e19
98en
ergy
inte
nsit
yva
lue20
10G
HG
emis
sion
sin
refe
renc
eca
sefo
reca
st
1995
GH
Gem
issi
ons
2020
GH
Gem
issi
onsw
ith
ICT
deve
lop-
men
t“fr
ozen
”in
2000
2000
GH
Gem
issi
ons
2005
GH
Gem
issi
ons
Unc
lear
tota
l20
30G
HG
emis
sion
base
line
Can
ada’
sKyo
toob
ligat
ions
(−6%
of19
90G
HG
emis
sion
leve
ls20
08–2
012)
EU27
elec
tric
ity
cons
umpt
ion
proj
ecti
on20
20w
ith
2005
GH
Gem
issi
onfa
ctor
2020
GH
Gem
issi
onsi
nB
aUsc
enar
io
ICT
impa
ctm
easu
reag
ains
tba
selin
e
2007
ener
gyin
tens
ity
chan
ge
2010
GH
Gem
issi
onsi
nad
just
edre
fere
nce
fore
cast
1997
/199
8G
HG
emis
sion
bala
nce
ofef
fect
s
2020
GH
Gem
issi
onsf
orsc
enar
ios
2010
GH
Gem
issi
onsf
orsc
enar
ios
2015
GH
Gm
itig
atio
npo
tent
ials
2030
GH
Gm
itig
atio
npo
tent
ials
GH
Gm
itig
a-ti
ongo
als
2020
GH
Gem
issi
onsf
orsc
enar
ios
2020
GH
Gem
issi
onsf
orsc
enar
io
Con
tinu
ed
828 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
Tabl
e1
Con
tinue
d
Cha
ract
erist
icof
the
stud
yC
EC
Sst
udy
EPA
/LB
NL
stud
yD
igita
lEur
ope
stud
yIP
TS
stud
yN
EC
stud
yT
elst
rast
udy
Eco
fys
stud
yW
WF
Can
ada
stud
yB
ioin
telli
genc
est
udy
GeS
Ist
udy
Res
ults
Ove
rall
ICT
impa
ct−1
.5to
−2.0
%−1
07M
tC
O2e/
aD
:dec
reas
e;U
K:i
ncre
ase
+2%
to−1
9%[in
2020
GH
Gem
issi
onle
vels
]
+0.6
%to
−2.1
%−2
7.3
MtC
O2e
−116
8to
−871
1M
tC
O2e
−19.
1to
−36
MtC
O2e
[7%
–13.
3%of
Can
ada’
sK
yoto
oblig
atio
ns]
+0.1
%to
−5.5
%w
itho
utC
E+1
.5%
to−4
.6%
incl
udin
gC
E[in
1990
GH
Gem
issi
onle
vels
]
Up
to−1
5%
Impa
ctby
appl
icat
ion
No di
sagg
rega
tion
No di
sagg
rega
tion
No di
sagg
rega
tion
Som
epo
siti
ve,
som
ene
gati
ve,
som
eam
bigu
ous
All
posi
tive
All
posi
tive
All
posi
tive
All
posi
tive
All
posi
tive
All
posi
tive
Not
e:B
aU=
busi
ness
asus
ual;
CE
=co
nsum
erel
ectr
onic
s;EI
A=
U.S
.En
ergy
Info
rmat
ion
Adm
inis
trat
ion;
GH
G=
gree
nhou
sega
s;IC
T=
info
rmat
ion
and
com
mun
icat
ions
tech
nolo
gy;M
tCO
2e=
met
ric
meg
aton
sCO
2eq
uiva
lent
s;N
AM
EA=
Nat
iona
lAcc
ount
ing
Mat
rix
wit
hEn
viro
nmen
talA
ccou
nts;
NEM
S=
Nat
iona
lEne
rgy
Mod
elin
gSy
stem
ofth
eEI
A;W
EO=
Wor
ldEn
ergy
Out
look
;WB
CSD
=W
orld
Bus
ines
sCou
ncil
forS
usta
inab
leD
evel
opm
ent.
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 829
R E S E A R C H A N D A N A LYS I S
studies. Few studies consider personal travelassistants for public transport explicitly (Mallonet al. 2007). Buttazoni (2008) and Webb (2008)also examine ICT-enhanced fuel efficiency ofvehicles.
E-commerce is dealt with in all studies.Business-to-business (B2B) e-commerce includesICT in supply chain management and logisticsand business-to-consumer (B2C) e-commercecovers, for example, teleshopping. Virtual mobil-ity (telework, mobile work, virtual meetings, andalso teleshopping) and dematerialization effectsof ICT (by virtualization or ICT-based services)are treated by all studies.
Some studies distinguish between the growth,structural change, and technology effects of ICT,and other studies build upon the classificationinto first-, second-, and third-order effects. Afew partially account for both conceptual frame-works. A complete coverage of the growth,structural change, and technology effect is pro-vided by Kuhndt and colleagues (2003), whileMatsumoto and colleagues (2005) do not accountfor the growth effect. Laitner and colleagues(2001) account for rebound effects, but the reportdoes not disclose a value for economic growth.First-, second-, and third-order effects are cov-ered quantitatively by Erdmann and colleagues(2004) only; all other studies focus on second-order impacts, some also considering first ordereffects.2
Data and Methodology
Data sources, assumptions, and methodologiesof the studies vary broadly. The bottom-up stud-ies typically combine data for ICT uptake withestimated data for ICT impact. Uncertainty isdealt with in many ways. We distinguish be-tween data uncertainty (incomplete data aboutfacts that could be known in principle) and fu-ture uncertainty (developments that are not de-termined and can be influenced in principle).
Both Labouze and colleagues (2008) and But-tazoni (2008) do quantitative data uncertaintyanalysis at selected points only, but not for allICT applications and impacts under study sys-tematically. Erdmann and colleagues (2004) con-ceptualize data uncertainty as differences amongexpert estimates that could not be resolved by
discussion, propagating the error margins quanti-tatively to the final results.
The consideration of future uncertainty rangesfrom parameter estimations at the desk to com-plex participatory scenario building exercises.In general, scenarios are used to define a mea-sure for the future macroeconomic impact ofICT against a baseline. The diverse definitionsused for ICT impact measures, baselines, dif-ferences in geographic focus, and time hori-zons as well as the different methodologies usedall impede a direct comparison of numericalresults.
Results
Six of the studies reviewed postulate unam-biguous overall reductions of GHG emissions dueto ICT. Four studies show an ambiguous picture.In the study of Erdmann and colleagues (2004),the ICT impact can range from a slight increaseto a strong reduction, depending on data and fu-ture uncertainties. In both the Matsumoto andcolleagues (2005) and Labouze and colleagues(2008) studies, the negative first-order effectsoutweigh the positive second-order effects un-der certain conditions. Kuhndt and colleagues(2003) show that the impact of ICT on GHGemissions can be positive or negative for a coun-try, depending on the relative sizes of the three ef-fects considered (technology, structural change,and growth effect) and the contribution of pri-vate households.
All studies that disaggregate ICT effects toICT applications postulate net GHG-reducingimpacts for each single application, with the ex-ception of the study by Erdmann and colleagues(2004), which will be discussed in detail in thesection “Scenario Analysis.”
Discussion
The pervasive character of ICT requires inclu-sion of all economic sectors in macroeconomicassessments of ICT impacts. The inevitable com-plexity of such an assessment is aggravated bythe exceptional dynamics of ICT innovation anddiffusion. To cope with this problem, mid- tolong-term studies often introduce broad ICT ap-plication domains such as ITS (medium level of
830 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
abstraction) or treat ICT as a monolithic sector(high level of abstraction).
The studies reviewed all cover ICT impactson the industry, transport, and building sectorsand agree on the relevance of most ICT applica-tions concerning GHG emissions, but some ICTapplications such as smart city planning are onlytreated by one study quantitatively. Other ICTapplications such as precision farming are nottreated by any of the studies, although they arelikely to be relevant. The main reason is presum-ably the studies’ major focus on energy efficiency.
The two conceptual frameworks used,growth/structural change/technology effect andfirst-/second-/third-order effects, have some over-lap and some differences. While the first con-ceptual framework treats ICT as a driver ofeconomic growth, the studies of the second typeeither treat growth as an external variable or in-directly as a third-order effect (additional growthinduced by ICT). Second-order effects includeboth ICT-fostered dematerialization (by substitu-tion effects) and ICT-enhanced process efficiency(by optimization effects). First-order effects arenot explicitly addressed in the first conceptualframework but can be accounted for in principle.
We did not find any discussion of the implica-tions of choosing a certain metric for the ICT im-pact on GHG emissions. Even if a common met-ric were agreed upon, the different scopes wouldstill make it difficult to compare results.
We found that the five older references arerich and diverse in methodology, whereas the fivemore recent studies take the simpler approachof quantifying GHG reduction potentials at thelevel of second-order effects, ignoring first- and/orthird-order effects.
In view of the various sources of uncertaintyin this field of study, it seems indispensable todisaggregate the uncertainty as far as possible(by breaking it down to specific ICT applica-tion fields) and to make a clear difference be-tween future uncertainty and data uncertainty.Only Erdmann and colleagues (2004) make a sys-tematic attempt to treat data uncertainty quan-titatively. As contributors to this study, how-ever, we have to concede that data uncertaintyis merged with future uncertainty in the presen-tation of the results. Therefore, we will reinter-pret the results of this study in the next section
and demonstrate the benefits of separating sce-narios from data assumptions at the level of ICTapplications.
Scenario Analysis
In this section we present a new assessmentof the future impact of ICTs on GHG emissionsbased on the study by Erdmann and colleagues(2004), which was commissioned by the Institutefor Prospective Technological Studies (IPTS) ofthe European Commission. This study (in thefollowing referred to as the IPTS study) includeda system dynamics model of the causal relation-ships between relevant ICT applications and aset of given environmental indicators.
The subsequent scenario analysis did not in-volve any change to the model or the input data.Instead, now in a new way, the two types of un-certainty were separated in the calculation at thelevel of ICT applications:
• data uncertainty indicating research de-mand;
• future uncertainty indicating a need for po-litical action.
In doing this analysis, we want to show that asystematic treatment of uncertainty can improvethe practical utility of simulation results in that itmakes a contribution to setting research agendasand policy agendas.
We will first summarize the integrated assess-ment methodology of the IPTS study and thenpresent the new scenario analysis.
Integrated Assessment Methodology
The methodology of the IPTS study has beendescribed by Erdmann and colleagues (2004) andcontains a modeling focus by Hilty and colleagues(2004, 2006). Here we recapitulate the method-ological elements necessary to understand thesubsequent analysis.
Goal Definition and ScopeThe aim of the study was to explore qualita-
tively and to assess quantitatively the way thatICT can influence a set of given environmentalindicators within the time horizon of 2020.
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 831
R E S E A R C H A N D A N A LYS I S
The geographical coverage of the study was theEuropean Union at the time the study was put outto tender (EU 15). For each result, the generaliz-ability to an enlarged European Union has beendiscussed qualitatively (Hilty et al. 2004). Im-pacts caused outside the European Union (suchas emissions from ICT production in Asia) werenot covered by the study.
The basic approach was to model the effect ofICT application domains on the sectors energy,transport, production, and waste, and then to ag-gregate the effects to estimate the overall ICT im-pacts. Table 2 lists the application domains con-sidered and examples of established and newerapplications. An ICT impact was defined as a netchange—caused by applying ICT—of one of thefollowing given indicators: total freight transportvolume (tonne-kilometers per year, tkm/a), totalpassenger transport volume (passenger kilome-ters per year, pkm/a), the share of private cars inthe modal split of passenger transport (pkm-%),total energy consumption (terajoules per year,TJ/a), the share of renewables in electricity supply(TJ-%), total municipal solid waste not recycled(tonnes per year, t/a), and total GHG emissions(CO2-equivalents in metric tons per year, t/a).For the purpose of this article, we will focus onthe last indicator.
The strength of an integrated assessment ap-proach is that it is comprehensive with regardto sectors and can address the dynamic changesimplied by the long-term diffusion of a technol-ogy. In practice, an integrated assessment of ICTimpacts is so comprehensive that it cannot con-sider as much technological detail as an LCAstudy of ICT products or services does. Our studytherefore relied on abstractions of ICT devices(“client-type device,” “server-type device,” and“network-type device”; Hilty et al. 2004) andtheir application in generically defined domainsto cope with technical progress. It was assumedthat if more specific ICT devices and applicationshad been included, they would have become ob-solete during the simulation period.
Scenario BuildingThere are many drivers other than ICT that
may change the environmental indicators (suchas energy prices or governmental regulation) anddrivers that influence the diffusion of ICT ap-
plications. To reduce the number of possibili-ties spanned by these drivers (or external fac-tors), three plausible scenarios were created andvalidated in an expert workshop (Goodman andAlakeson 2003). These scenarios differed in thefactors the experts considered both most influen-tial with regard to the system under study andmost uncertain at the same time:
• Scenario A, called “Technocracy,” assumedstrong economic growth and employmentdriven by large companies in the servicesector and is enabled by low regulation.
• Scenario B, called “Government first,” as-sumed strong environmental regulation,resulting in only moderate growth, noprogress in employment, but good condi-tions for small and medium-sized enter-prises.
• Scenario C, called “Stakeholder democ-racy,” was characterized by steady economicgrowth, leading to an increase in the num-ber of households, desk workers, and totallabor force.
The scenarios were formulated qualitativelyand had to be quantified to be used as parametersettings for the ICT impact model (for detailssee tables S1-1 and S1-2 in the supplementarymaterial on the Journal Web site).
Modelling ICT Impacts Under Given Sce-nariosThe basic approach of the model was to cover
all three types of ICT effects on the environmen-tal indicators in the following way:
• First-order effects: A generic model of theenergy consumption of ICT in the use phaseand of the ICT waste produced was built(key parameters: the growth rate of the en-ergy efficiency of ICT equipment and the[negative] growth rate of the useful life ofICT equipment). The production of ICThardware was neglected because most of theimpacts of production would occur outsidethe given system boundaries.
• Second-order effects: Potentials of ICTapplications (see table 2) for optimiza-tion and/or substitution of products and
832 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
Table 2 Generic Description of Information and Communication Technology (ICT) Application Domainsand Relevant Examples of ICT Applications
ICT applicationdomain Generic description
Examples ofestablished ICT
applicationsExamples of
newer ICT applications
Production processmanagement(PPM)
ICT applications increasing theenergy and material efficiencyof production processes
Productionplanning andcontrol systems
Networked multisensorregulators forhigh-temperatureprocesses
Power generationand distributionmanagement(smart grids)
ICT-supported measures toimprove the coordination ofelectricity supply and demand
ICT-based loaddetection
Electricity consumptionfeedback (smartmetering) anddemand-sidemanagement (smartappliances)
Intelligent transportsystems (ITS)
ICT applications assisting driversand managing transportinfrastructure (all modes)
In-vehiclenavigationsystems
Real-time road trafficrouting
Supply chainmanagement(SCM)
ICT applications supporting thecoordination of the processesand the optimization of thephysical mass flows in thesupply chain
Electronic datainterchangealong the supplychain
RFID-based supply chainmanagement
Virtual goods (de-materialization)
ICT applications supporting aproduct-to-service shift andthe digitization of physicalproducts
ICT-enabledservices (e.g.,Web-based carsharing)
E-paper, e-books
Teleshopping Demand-side e-commerce,allowing the purchase ofphysical goods away fromretail
Online orderingand payment
Peer-to-peer merchandiseexchange networks
Telecommuting ICT-supported work away fromthe firm’s premises, whichreplaces commuting travel
Personal computerand Internetaccess at home
Virtual workspace athome
Teleconferencing A mode of telecommunicationwhich replaces business travel
Audioconferencing
High-definition videoconferencing
Mobile work ICT devices making it possibleto work while travelling
PDAs WLAN hotspots in trains
Buildingmanagement
All “soft measures” supported byICT to reduce the energyconsumption of buildings
Intelligent HVACmanagement
Energy consumptionfeedback (smartmetering)
Waste management ICT applications supporting theseparation of waste fractions
Sensor-basedsorting
RFID-based sorting
Note: HVAC = heating, ventilation, and air conditioning; PDA = personal digital assistant; RFID = radio frequencyidentification; WLAN = wireless local area network.Source: Erdmann et al. (2004).
processes were modeled. The speeds atwhich these potentials would be realizedwere determined by parameters dependingon the scenarios.
• Third-order effects (including rebound ef-fects): Realized efficiency potentials changemarket prices or time-use and can thereforechange demand, depending on elasticity
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 833
R E S E A R C H A N D A N A LYS I S
parameters (see table S1-3 in the supple-mentary material on the Journal Web site).
All parts of the model could interact with oneanother; for example, the substitution of virtualfor material goods could decrease freight trans-port demand and increase ICT use. The overallimpact of ICT was calculated by creating a sec-ond variant of each scenario in which all vari-ables concerning ICT were “frozen” during thesimulation run. By comparing the two variantsof each scenario, the overall net effect of furtherICT development and diffusion was calculated.
Figure 1 shows groups of variables that wereused in the model. The details of the model,
which was implemented using the simulation sys-tem PowerSim, have been published in the origi-nal interim report to IPTS (Hilty et al. 2004).
Figure 2 shows the basic causal structure usedby this model to account for the dynamic impactsof ICT on the environmental indicators. Thiscausal-loop diagram is basically an interpretationof the definitions of first-, second-, and third-order ICT effects.
The scenario differences were expressed bydifferent settings of the external variables; thesewere thus used to represent future uncertainty (fordetails see table S1-2 in the supplementary mate-rial on the Journal Web site). In order to accountfor data uncertainty, the scenarios were combined
Figure 1 Groups of variables used to represent the scenarios and the three levels of information andcommunications technology (ICT) effects (see Hilty et al. (2004) for a description of all variables).CRT = cathode ray tube; GDP = gross domestic product; LAN = local area network; LCD = liquid crystaldisplay; SMEs = small and medium-sized enterprises.
834 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
netenvironmental
impact
energy and material needed to provide one service unit
demand for the service (number of service units consumed)
price per service unit
elasticity of demand with regard to price
efficiency and substitution potentials
provided by ICT
speed of exploiting efficiency and
substitution potentials
external influence determined by scenarios (economic, political and
cultural framework)
ICT development and use
1st order effects
2nd order effects
3rd order effects
Figure 2 Abstract representation of the causal structure used in the system dynamics model of informationand communication technology (ICT) environmental effects by Hilty and colleagues (2004).Dotted arrows = influence from outside the system boundaries; thin solid arrows = first-order effects ofICT; medium solid arrows = second-order effects of ICT; thick arrows = third-order effects of ICT.
with three parameter settings representing best-,mean-, and worst-case data assumptions.3 Thebest-/worst-case variants were calculated with theparameter vector that minimized/maximized to-tal energy consumption, which means total en-ergy consumption was chosen as a lead indicator.For the mean variant, each parameter was set tothe mean of its best- and worst-case values.
Results and Discussion of the ScenarioAnalysis
We will first show the overall ICT impact onGHG emissions that have been calculated anddiscuss the associated data uncertainty and fu-ture uncertainty. Then we will look into indi-vidual ICT application domains and discuss theircontributions to the overall impact with regardto their varying extent and uncertainties.
Overall ICT ImpactsSimulating the three scenarios in three vari-
ants yields the nine value sets of total GHGemissions in 2020 shown in figure 3. All GHGemissions are expressed in percent of the year2000 level. The left bar of each value set showsthe GHG emissions in 2020 that occur if ICTis “frozen” at the 2000 level. The second bar
shows the GHG emissions in 2020 with theprojected ICT development (2000–2020). Thethird bar shows the difference between thesetwo values, representing our metrics of ICTimpact.4
As figure 3 shows, there is only one case (sce-nario B combined with worst-case data assump-tions) in which this value is positive (indicatingthat the net impact of ICT is an increase in GHGemissions). As can be seen, the variation causedby ICT is smaller than both future uncertaintyand data uncertainty.
Future uncertainty is represented by the dif-ferences among the scenarios A, B, and C. Datauncertainty is represented by the best-, mean-and worst-case variants of each scenario. An ex-ample may illustrate the origin of this type ofuncertainty: A parameter called “virtualizationpotential for information products,” expressinghow much of the total physical mass of all infor-mation products (books, newspapers, DVDs, etc.)could potentially be replaced by online services(in EU 15 by 2020), was estimated on a modelingworkshop to range from 30% to 90%. Dozens ofsuch uncertainty ranges combine into the overalldata uncertainty reflected by the difference be-tween the best- and worst-case scenarios (all datapublished by Hilty et al. 2004).
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 835
R E S E A R C H A N D A N A LYS I S
Figure 3 Simulated developments of greenhouse gas (GHG) emissions by 2020 in EU 15 (year2000 = 100%) for the three future scenarios in three data uncertainty variants. ICT = information andcommunication technology.
If we look only at the mean variant, the pos-itive impact of ICT is lowest in scenario B, withroughly a 3% decrease of GHG emissions, andhighest in scenario C, with a 10% decrease. Onepossible interpretation is that this scenario ismore conducive to “decarbonizing” ICT appli-cations than scenario B. However, it should benoted that scenario B already has low GHG emis-sions for reasons other than ICT, as can be seenin figure 3. These reasons include weaker GDPgrowth and higher fossil energy prices in scenarioB. It is therefore possible that the most favorabledevelopment path in terms of GHG reduction isnot the one in which ICT contributes most toGHG reduction.
Impacts by ICT Application DomainsIn order to show in detail how ICTs affect
GHG emissions, we took a look at the disag-gregated effects of ten ICT application domains.These results were produced by “freezing” all butone of the ten ICT application domains for eachdomain. They thus show the net effect of each ap-plication domain on the overall GHG emissions.
“ICT use,” which aggregates the total first-ordereffects of ICT in the use phase, has been addedto figures 4 and 5 for comparison.
To reduce complexity, we selected scenarioC for this analysis because it provided the high-est GHG mitigation potential by ICT. Figure 4shows the results. Except for “ICT use,” thesevalues are net impacts including first-, second-,and third-order effects. The sum of the impactsacross the ten domains does not necessarily addup to the total impact shown in figure 3 becausethe application domains are interacting in themodel.
It can be seen that two application domainsincrease the total emissions according to the IPTSmodel, namely ITS and mobile work. ITS have,among other effects, the effect of saving time inthe transport of passengers or freight. ITS there-fore make transport not only more energy effi-cient and cheaper, but also faster. This creates arebound effect based on assumed elasticities of de-mand and the time-use approach included in themodel. In a similar way, mobile work is assumedto create a rebound effect by enabling better
836 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
e gri
ds
go
od
s
op
pin
g
mm
uti
ng
nfe
ren
cin
g
wo
rk
din
g
emen
t
6%
ICT
use
PP
M
Sm
art
gri
ds
ITS
SC
M
Vir
tual
go
od
s
Tele
sho
pp
ing
Tele
com
mu
tin
g
Tele
con
fere
nci
ng
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
0 %
)
0%
2%
4%
6%
ICT
use
PP
M
Sm
art
gri
ds
ITS
SC
M
Vir
tual
go
od
s
Tele
sho
pp
ing
Tele
com
mu
tin
g
Tele
con
fere
nci
ng
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
ar 2
000
= 10
0 %
)
4%
-2%
0%
2%
4%
6%
ICT
use
PP
M
Sm
art
gri
ds
ITS
SC
M
Vir
tual
go
od
s
Tele
sho
pp
ing
Tele
com
mu
tin
g
Tele
con
fere
nci
ng
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
ssio
ns
(yea
r 20
00 =
100
%)
-8%
-6%
-4%
-2%
0%
2%
4%
6%
ICT
use
PP
M
Sm
art
gri
ds
ITS
SC
M
Vir
tual
go
od
s
Tele
sho
pp
ing
Tele
com
mu
tin
g
Tele
con
fere
nci
ng
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
on
GH
G e
mis
sio
ns
(yea
r 20
00 =
100
%)
-12%
-10%
-8%
-6%
-4%
-2%
0%
2%
4%
6%
ICT
use
PP
M
Sm
art
gri
ds
ITS
SC
M
Vir
tual
go
od
s
Tele
sho
pp
ing
Tele
com
mu
tin
g
Tele
con
fere
nci
ng
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
ICT
imp
act
on
GH
G e
mis
sio
ns
(yea
r 20
00 =
100
%)
-12%
-10%
-8%
-6%
-4%
-2%
0%
2%
4%
6%
ICT
use
PP
M
Sm
art
gri
ds
ITS
SC
M
Vir
tual
go
od
s
Tele
sho
pp
ing
Tele
com
mu
tin
g
Tele
con
fere
nci
ng
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
ICT
imp
act
on
GH
G e
mis
sio
ns
(yea
r 20
00 =
100
%)
worst mean best
Figure 4 The impacts ofinformation and communicationtechnology (ICT) use and ICTapplications on total greenhouse gas(GHG) emissions, 2020, in EU 15 forscenario C (‘Stakeholder democracy’)as a percentage of the GHGemissions 2000. The length of thevertical lines indicates the uncertaintyof the results caused by input datauncertainty. ITS = intelligenttransport system; PPM = productionprocess management; SCM = supplychain management.
utilization of travel time: When some types ofwork and travel can be combined, the cost oftravel time decreases, which leads to an increasein the demand for travel.
With respect to data uncertainty, our anal-ysis reveals three types of ICT application do-mains: first, small impacts with small degrees ofdata uncertainty (smart grids, supply chain man-agement [SCM], teleshopping, telecommuting,teleconferencing, and mobile work); second, sig-nificant impacts with low to medium data uncer-tainty (production process management [PPM],ITS, and building management); third, signifi-cant impacts with high data uncertainty (virtual
goods and ICT use). The key driver for total datauncertainty (as shown in figure 3 for scenario C)is the data uncertainty of the impacts of virtualgoods on GHG emissions (−0.1% to −10.5%).This uncertainty represents a lack of knowledgeabout the effects of virtualizing goods.
In order to take a closer look at the future un-certainty of the application domains, we selectedthe mean data values and calculated the impactsof the application domains for all three scenarios(figure 5).
For most of the domains, scenario B is out-standing as the scenario (1) with the highest ad-verse impacts (ICT use, mobile work) and (2) the
2%
-1%
0%
1%
2%
3%
ICT
use
PP
M
Sm
art g
rid
s
ITS
SC
M
Vir
tual
go
od
s
Tel
esh
op
pin
g
Tel
eco
mm
uti
ng
Tel
eco
nfe
ren
cin
g
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
issi
on
s (y
ear
2000
= 1
00 %
)
-6%
-5%
-4%
-3%
-2%
-1%
0%
1%
2%
3%
ICT
use
PP
M
Sm
art g
rid
s
ITS
SC
M
Vir
tual
go
od
s
Tel
esh
op
pin
g
Tel
eco
mm
uti
ng
Tel
eco
nfe
ren
cin
g
Mo
bile
wo
rk
Bu
ildin
g
man
agem
ent
ICT
imp
act o
n G
HG
em
issi
on
s (y
ear
2000
= 1
00 %
)
A 'Technocracy' B 'Government First' C 'Stakeholder Democracy'
Figure 5 The impacts ofinformation and communicationstechnology (ICT) use and ICTapplications on total greenhouse gas(GHG) emissions, 2020, in EU 15 formean input values for three scenarios(A, B, and C) as a percentage of theGHG emissions 2000. The length ofthe vertical lines indicates thevariation of the results caused byfuture uncertainty. ITS = intelligenttransport system; PPM = productionprocess management; SCM = supplychain management.
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 837
R E S E A R C H A N D A N A LYS I S
lowest beneficial impacts (PPM, SCM, virtualgoods, building management). One applicationis clearly outstanding in its GHG-reducing ef-fect: building management, which includes intel-ligent HVAC systems and soft measures such ascreating user awareness for energy consumptionin buildings. The impacts as well as their varia-tion among scenarios are small for teleshopping,telecommuting, and teleconferencing. These ap-plications seem to be “robust” in having almostno net impact due to rebound effects that com-pensate for their second-order impact. In gen-eral, the impacts of ICT application domains aresimilar in scenarios A and C, with a significantadvantage of scenario C in the domain of vir-tual goods. The main driver for the outstand-ing character of scenario B are the higher energyprices assumed in this scenario (see the analysis byArnfalk 2004).
Conclusions
This article has reviewed principal macroeco-nomic studies on the impact of ICT on GHGemissions and discussed the challenges encoun-tered in modeling ICT impacts at the macroeco-nomic level. When assessing impacts of ICT, it isnecessary to introduce a baseline: This can be his-torical data, a “business as usual” scenario, or—inthe case of our own study—an “ICT freeze” sce-nario in which the development and diffusion ofICT were frozen at the beginning of the simu-lation period. As our review has shown, most ofthe studies are not comparable due to differentbaselines.
We have presented an in-depth scenario anal-ysis providing a reinterpretation of the results ofour own study. It showed that the impacts of dif-ferent ICT application domains as well as dif-ferent types of impacts can compensate for eachother as a result of the system’s dynamics. ICTapplications that are assumed to mitigate GHGemissions may turn out to have an adverse ef-fect, as shown in the cases of ITS and mobilework. The impact of some applications is sensi-tive to the scenarios chosen and to the data as-sumptions made, whereas other ICT applicationsseem to reduce GHG emissions in a relatively ro-bust manner, such as building management andvirtual goods, although the extent of the reduc-
tion remains uncertain. In the latter case, bothtypes of uncertainty are remarkably high, indi-cating a great need for further research as well assome political responsibility: Whether the dema-terialization potential of ICT will unfold in theeconomic system seems to depend on frameworkconditions—otherwise, in the worst case, it willjust fizzle out.
In contrast to the other studies, we find bothpositive and negative net impacts depending onthe type of ICT application. We attribute thisto the fact that our simulation model treatsfirst-, second-, and third-order effects of ICT inan integrated manner.
Meanwhile, there is a growing body of lit-erature on the extent of direct rebound effects(e.g., Sorrell et al. 2009) that could be used toimprove the approach. Furthermore, indirect re-bound effects, which will increase the demandfor other goods and services if less money or timeis spent on one service, could be easily incor-porated into an ICT impact assessment modeltoday as marginal consumption values of expen-diture and their related impacts on GHG emis-sions (Thiesen et al. 2008; Tukker et al. 2006).However, macroeconomic rebound effects of en-ergy efficiency measures leading to a series ofeconomy-wide price and quantity adjustments—with energy-intensive goods and sectors gain-ing at the expense of less energy-intensive ones(Greening et al. 2000)—can hardly be integratedinto an ICT impact assessment study, as they af-fect all industries. Instead, it would be necessaryto incorporate ICT in a general macroeconomicmodel, as done by Laitner and colleagues (2001).
Most of the existing studies on ICT impactson GHG emissions (including the IPTS study)have a limited geographical scope, both regard-ing the applications and the impacts of ICT.This we consider a serious limitation, since trans-boundary effects at any level (first-, second-, orthird-order) are then neglected. For example,energy consumption of mining metals abroadto produce ICT hardware may compensate forsome of the domestic progress in energy effi-ciency. Conversely, materials saved due to ICT-enabled efficiency could reduce the demand forresources outside the system boundaries. Finally,resource efficiency induced by ICT can influ-ence world market prices and produce global
838 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
rebound effects. We therefore think that the nextgeneration of ICT impact models should combinethe scope and ambition of the two global studies(Buttazoni 2008; Webb 2008) with a method-ology that is able to address effects on all threelevels.
A process that adjusts the scenarios used insuch assessments periodically to real-world devel-opments and recalculates the implications couldprovide opportunities for mutual learning be-tween researchers and decision makers.
Acknowledgements
The authors wish to thank Thomas Ruddyfor many improvements of this article and threeanonymous reviewers for their in-depth evalua-tion and helpful feedback.
Notes
1. The original Global e-Sustainability Initiativestudy (Webb 2008) was complemented by the U.S.,German, and Portuguese addenda and a subsequentindustry study focusing on mobile telecommuni-cations (McVeigh et al. 2009). The Gibson andcolleagues (2008) study represents a series of ICTindustry-sponsored studies of similar methodologyinvolving the World Wildlife Fund (e.g., Pamlinand Szomolanyi 2007). The Matsumoto and col-leagues (2005) article represents a body of hybridmodeling studies that have been prepared in Japan.We could not analyze some presumably relevantJapanese studies at a meaningful level mainly due tolanguage barriers (e.g., the background of the Reportfrom Study Group on ICT Policy for Addressing GlobalWarming [Ministry of Internal Affairs and Commu-nications 2008]). Because ICTs pervade, transform,and link all economic sectors, any cross-sector def-inition to capture the impacts of ICT is somewhatarbitrary. For example, it can be argued that e-commerce and virtual mobility are cross-sector ICTapplication domains. However, the major studiesfocusing on these domains, we found, do not pro-vide additional insight on the topic of this article.
2. From a life cycle perspective, the studies addressmainly the use phase of ICT equipment. Few stud-ies also account for upstream and/or downstreamprocesses (Kuhndt et al. 2003; Labouze et al. 2008;Webb 2008). Erdmann and colleagues (2004) in-clude electronic waste.
3. We are aware that upper and lower bounds based onexpert judgment do not imply that the true valueis within these ranges. However, if all uncertaintyranges input to the model are selected amply and thesimulation still produces some unambiguous results,these results can be considered more credible thanindividual judgments.
4. These values may differ slightly from those pub-lished previously (Erdmann et al. 2004; Hilty et al.2004) because of a different GHG emission refer-ence year. The former publications took the respec-tive GHG emissions of each ICT “freeze” simula-tion run by the year 2020 as the reference, whereasin this article the ICT impact for all simulation runsis measured against the GHG emissions in the year2000 to allow for cross-comparisons of scenarios.
References
Allenby, B. 1999. Culture and Industrial Ecology. Jour-nal of Industrial Ecology 3(1): 2–4.
Arnfalk, P. 2004. The future impact of ICT on environ-mental sustainability. Evaluation and recommenda-tions. Fifth Interim Report. International Institutefor Industrial Environmental Economics at LundUniversity (IIIEE). Lund, Sweden: IIIEE.
Buttazoni, M. 2008. Potential global CO2 emission re-ductions from ICT use: Identifying and assessing theopportunities to reduce the first billion tonnes of CO2
emissions. Ecofys Italy Srl. Solna, Sweden: WorldWildlife Fund Sweden.
Erdmann, L., L. Hilty, J. Goodman, and P. Arnfalk.2004. The future impact of ICTs on environmentalsustainability. EUR 21384. Seville, Spain: Institutefor Prospective Technology Studies.
Fichter, K. 2003. E-commerce: Sorting out the environ-mental consequences. Journal of Industrial Ecology6(2): 25–41.
Gibson, L., M. Bein, and B. Plamondon. 2008. Innovat-ing toward a low-carbon Canada: Using technologyto transform tomorrow. Toronto, Canada: WorldWildlife Fund Canada.
Goodman, J. and V. Alakeson. 2003. The future impactof ICT on environmental sustainability. Scenarios.Third interim report. London: Forum for the Fu-ture.
Greening, L., D. L. Greene, and C. Difiglio. 2000. En-ergy efficiency and consumption: The reboundeffect—A survey. Energy Policy 28(6–7): 389–401.
Hilty, L. M. 2008. Information technology and sustainabil-ity. Essays on the relationship between informationtechnology and sustainable development. Norderst-edt, Germany: Books on Demand GmbH.
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 839
R E S E A R C H A N D A N A LYS I S
Hilty, L. M., P. Arnfalk, L. Erdmann, J. Goodman,M. Lehmann, and P. Wager. 2006. The relevanceof information and communication technologiesfor environmental sustainability—A prospectivesimulation study. Environmental Modelling & Soft-ware 21(11): 1618–1629.
Hilty, L. M., P. Wager, M. Lehmann, R. Hischier, T.Ruddy, and M. Binswanger. 2004. The future im-pact of ICT on environmental sustainability. Refine-ment and quantification. Fourth interim report. St.Gallen, Switzerland: Swiss Federal Laboratoriesfor Materials Testing and Research.
IPCC (Intergovernmental Panel on Climate Change).2007. Climate change 2007: Synthesis report. Con-tributions of Working Group I, II and III to the FourthAssessment Report of the Intergovernmental Panel onClimate Change. Geneva, Switzerland: Intergov-ernmental Panel on Climate Change.
Kohler, A. and L. Erdmann. 2004. Expected envi-ronmental impacts of pervasive computing. Hu-man and Ecological Risk Assessment 10(5): 831–852.
Kuhndt, M, J. Geibler, V. Turk, S. Moll, K. Schal-labock, and S. Steger. 2003. Virtual dematerializa-tion and factor X. Digital Europe. WP 3. Wupper-tal, Germany: Wuppertal-Institut.
Labouze, E., C. S. Iyama, A. Beton, L. Stobbe, S.Gallehr, and L. G. Scheidt. 2008. Impacts of in-formation and communication technologies on energyefficiency. Final report. Bio Intelligence Service.Brussels, Belgium: European Commission DG IN-FSO.
Laitner, J. A. S. and K. Ehrhardt-Martinez. 2008. Infor-mation and communication technologies: The powerof productivity. Report number E081. Washington,DC: American Council for an Energy-EfficientEconomy.
Laitner, J. A. S., J. G. Koomey, E. Worrell, and E.Gumerman. 2001. Re-estimating the annual energy2000 outlook forecasting using updated assumptionsabout the information economy. Washington, DC:Environmental Protection Agency/Berkeley, CA:Lawrence Berkeley National Laboratory.
Madlener, R. 2008. The implications of ICT for en-ergy consumption. Study report 09/2008. Bonn,Germany: empirica/Brussels, Belgium: EuropeanCommission.
Mallon, K., G. Johnston, D. Burton, and J. Cavanagh.2007. Towards a high-bandwidth, low-carbon future:Telecommunications-based opportunities to reducegreenhouse gas emissions. Version 1.0. Fairlight,Australia: Climate Risk.
Matsumoto, M., J. Hamano, T. Tamura, and H. Iguchi.2005. Impacts of ubiquitous society on the globalwarming problem in 2010. In Proceedings of the
International Symposium on Electronics and the En-vironment, Vol. 1: 183–188. New York: Instituteof Electrical and Electronics Engineers.
McVeigh, C., G. Sandieson, T. Newitt, B. Webb, J.Keeble, S. Yurisich, A. Gordon, and X. Veillard.2009. Carbon connections: Quantifying mobile’s rolein tackling climate change. Newbury Berkshire, UK:Vodaphone/London: Accenture.
Ministry of Internal Affairs and Communications.2008. Communications news [Tokyo], 1 August,1–7. In Report from Study Group on ICT Policyfor Addressing Global Warming. Tokyo: Ministryof Internal Affairs and Communications.
Organization for Economic Cooperation and Develop-ment. 2009. Declaration on green growth. Adoptedat the council meeting at ministerial level on 25June 2009. Paris: Organization for Economic Co-operation and Development.
Pamlin, D. and K. Szomolanyi. 2007. Saving the cli-mate @ the speed of light. First roadmap for reducedCO2 emissions in the EU and beyond. Solna, Swe-den: World Wildlife Fund /Brussels: EuropeanTelecommunications Network Operators’ Asso-ciation.
Rejeski, D. 2003. E-commerce, the Internet and theenvironment. Journal of Industrial Ecology 6(2):1–3.
Romm, J., A. Rosenfeld, and S. Herrmann. 1999. TheInternet economy and global warming. A scenario ofthe impact of e-commerce on energy and the envi-ronment. Arlington, VA: Center for Energy andClimate Solutions.
Sorrell, S., J. Dimitropoulos, and M. Sommerville.2009. Empirical estimates of the direct reboundeffect: A review. Energy Policy 37(4): 1356–1371.
Thiesen, J., T.S. Christensen, T.G. Kristensen, R.D.Andersen, B. Brunoe, T.K. Gregersen, M. Thrane,and B.P. Weidema. 2008. Rebound effects of pricedifferences. International Journal of Life Cycle As-sessment 13(2): 104–114.
Tukker, A., G. Huppes, J. Guinee, R. Heijungs, A. deKoning, L. van Oers, S. Suh, T. Geerken, M. VanHolderbeke, B. Jansen, and P. Nielsen. 2006. En-vironmental impact of products (EIPRO). Analysisof the life cycle impacts related to final consumptionof the EU-25. EUR 22284 EN. Seville, Spain: In-stitute for Prospective Technology Studies.
Webb, M. 2008. SMART 2020: Enabling the low car-bon economy in the information age. The ClimateGroup. Brussels: Global e-Sustainability Initia-tive.
Yi, L. and H. Thomas. 2007. A review of research onthe environmental impact of e-business and ICT.Journal of Environment International 33(6): 841–849.
840 Journal of Industrial Ecology
R E S E A R C H A N D A N A LYS I S
About the Authors
Lorenz Erdmann is an Associate ResearchScientist at the Institute for Futures Studiesand Technology Assessment in Berlin, Germany.Lorenz M. Hilty is Head of the Technology and
Society Lab (TSL) at Empa, the Swiss FederalLaboratories for Materials Testing and Research,and is also professor of Informatics and Sustain-ability at the Department of Informatics at theUniversity of Zurich.
Supplementary Material
Additional supplementary material may be found in the online version of this article:
Supplement S1: This supplement fleshes out the three scenarios presented in the article, addingbrief characterizations for each as well as the specific value assumptions used in the final model.In these scenarios, as efficiency potentials become realized, market prices and therefore demandcan change; the extent to which this happens is dependent on elasticity parameters in tableS1–3. The final table, S1–4, presents the disaggregated simulation results for all future scenariosand data uncertainty variants.
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supple-mentary material supplied by the authors. Any queries (other than missing material) should bedirected to the corresponding author for the article.
Erdmann and Hilty , Macroeconomic Impacts of ICT on GHG Emissions 841