Vous êtes sur la page 1sur 29

A8CACPLS

1C ASSLSSlnC
8USINLSS IMAC1S
Cn uLvLLCMLn1




lramework paper
8y Carollne Ashley, Carolln Schramm, karen Lllls
Cverseas uevelopmenL lnsLlLuLe

roduced for Lhe MeeLlng Serles on 'Parnesslng Lhe power of Lhe prlvaLe secLor for
developmenL' organlsed by Lhe ueparLmenL for lnLernaLlonal uevelopmenL, Cverseas
uevelopmenL lnsLlLuLe and 8uslness AcLlon for Afrlca, 2009


"







Approaches to assess|ng bus|ness |mpacts
on deve|opment

lramework paper



8y Carollne Ashley, Carolln Schramm and karen Lllls
Cverseas uevelopmenL lnsLlLuLe, !une 2009


roduced for Lhe MeeLlng Serles on
'Parnesslng Lhe power of Lhe prlvaLe secLor for developmenL'
organlsed by Lhe ueparLmenL for lnLernaLlonal uevelopmenL,
Cverseas uevelopmenL lnsLlLuLe and 8uslness AcLlon for Afrlca









'lf you cannoL measure lL, maybe lL ls noL Lhere'
l



'lL ls beLLer Lo be approxlmaLely rlghL Lhan preclsely wrong'.


'Cscar Wllde once sald lL ls a very sad Lhlng LhaL nowadays Lhere ls so llLLle useless lnformaLlon."
unforLunaLely a look aL Loday's non-flnanclal reporLlng flnds Lhe opposlLe Lo be Lrue.'









#

reface

1hls paper was prepared for Lhe flfLh meeLlng ln Lhe serles on 'Parnesslng Lhe ower of 8uslness for
uevelopmenL lmpacL,' held ln May 2009. 1he flfLh meeLlng explored how we can beLLer undersLand Lhe
lmpacL of buslness on developmenL, and parLlcularly whaL companles Lhemselves do ln Lhls area.

So far, ln dlscusslons of buslness models and corporaLe lnnovaLlon, a recurrenL Lheme has been LhaL we
need Lo know whlch approaches really do conLrlbuLe effecLlvely Lo developmenL, or creaLe 'soclal value'.
8uL approaches Lo undersLandlng Lhls are dlverse. 1he purpose of Lhls paper, and of Lhe meeLlng, ls Lo
explore conLrasLlng approaches and Lo develop key quesLlons for furLher dlscusslon.

1hls framework paper was publlshed as a urafL ln Llme for Lhe publlc evenL and for some Llme LherafLer.
Clven conLlnulng lnLeresL ln Lhe Loplc, Lhe auLhors and 8l have now publlshed lL ln flnal form. lL ls based
on experlence up unLll May 2010. An updaLe Lo Lhls ls belng prepared for !uly 2012. vlslL
$$$%&'()*+((,)-./(012+3/4%13-5+2+*/(5"67"8+2+*/8(+3)+(8'*9+3(/:*9)*-8:*98+*.:*;)*-8&'()*+((8)<0:;/81*8 for furLher
lnformaLlon.








Contents

1. rob|em statement: we [ust don't know "
2. Assess|ng the |mpacts of bus|ness on deve|opment: contrast|ng approaches #
3. Approach 1: Loca| assessment $
4. Approach 2: Va|ue cha|n maps and poverty footpr|nts %&
S. Approach 3: Assessment of mu|t|p||ers and contr|but|on to econom|c act|v|ty %'
6. Approach 4: keport|ng aga|nst scorecards and standard|sed |nd|cators %(
6.1 Clobal 8eporLlng lnlLlaLlve and un Clobal CompacL 7#
6.2 8eporLlng agalnsL more poverLy focused lndlcaLors 7=
6.3 AddlLlonal perspecLlves from verlflcaLlon and assurance 7>
7. Ins|ghts from donors and econom|sts %)
7.1 Challenges of lmpacL assessmenL ln prlvaLe secLor developmenL 7?
7.2 CausallLy "6
7.3 CosL effecLlveness and comparlsons "7
7.4 uynamlc lmpacLs ""
7.3 LconomlsLs' assessmenL of lmpacL and addlLlonallLy "@
8. Some key |ssues on corporates' deve|opment |mpact agenda '#
keferences and Lndnotes '$







@




L|st of Acronyms

AA Anglo Amerlcan
Au8 Aslan uevelopmenL 8ank
Afu8 Afrlcan uevelopmenL 8ank
8C1A 8uslness Call Lo AcLlon
8LCl 8uslness Llnkage Challenge lund
CC CommunlcaLlon on rogress
CS8 CorporaLe Soclal 8esponslblllLy
uLC ueuLsche lnvesLlLlons und LnLwlcklungsgesellschafL
L88u Luropean 8ank for 8econsLrucLlon and uevelopmenL
L88 Lconomlc 8aLe of 8eLurn
Cu Cross domesLlc producL
C8l Clobal 8eporLlng lnlLlaLlve
llC lnLernaLlonal llnance CorporaLlon
L8C London 8enchmarklng Croup
MuC Mlllennlum uevelopmenL Coals
MnL MulLlnaLlonal enLerprlse
nCC non-governmenLal organlsaLlon
Cul Cverseas uevelopmenL lnsLlLuLe
Su rlvaLe SecLor uevelopmenL
SCM rogramme for CooperaLlon wlLh Lmerglng MarkeLs
wC rlcewaLerhouseCoopers
SML Small and medlum enLerprlses
SMML Small, medlum and mlcro enLerprlses
Snv neLherlands uevelopmenL CrganlsaLlon
W8CSu World 8uslness Councll for SusLalnable uevelopmenL
1. rob|em statement: we [ust don't know

We know surprlslngly llLLle abouL Lhe lmpacL of companles, parLlcularly any speclflc company, on people
and economles ln Lhe developlng world. Some may ask, 'does Lhls maLLer?' Companles, afLer all, are busy
generaLlng ouLpuL and proflLs and do noL need Lo accounL for Lhelr developmenL lmpacL ln Lhe fashlon of a
developmenL agency funded by Lax-payers' money or publlc donaLlons. Companles wlll prlorlLlse Lhe
reporLlng of Lhelr flnanclal slLuaLlon Lo Lhelr shareholders raLher Lhan Lhelr lmpacL on developlng counLrles
Lo general 'concerned sLakeholders.' neverLheless, undersLandlng companles' lmpacL on developmenL does
maLLer, for 4 reasons.

llrsLly, companles slmply lack Lhe lnformaLlon needed Lo lmprove Lhelr own declslon-maklng. Many
companles have made commlLmenLs Lo responslble buslness. 1hey prlde Lhemselves on Lhelr conLrlbuLlon
Lo economlc developmenL ln Lhelr areas of operaLlon and non-prlce compeLlLlon lncreaslngly lncludes Lhelr
soclo-economlc or envlronmenLal performance. 1hey have a range of levers Lhey can use Lo boosL Lhelr
lmpacL: soclal lnvesLmenL, ln-klnd donaLlons of faclllLles and volunLeered for local communlLles, helplng


=

enLrepreneurs enLer Lhe supply or dlsLrlbuLlon chaln, parLnershlps wlLh governmenL, non-governmenLal
organlsaLlon (nCC) or unlLed naLlons agencles, developlng low-cosL producLs for low-lncome consumers,
or engaglng cusLomers ln developmenL campalgns. 1he declslon on how much Lo lnvesL ln whaL, ls based on
many facLors, such as buslness sLrucLure, local need, and Lhe brand. 8uL usually noL on any assessmenL of
lmpacL LhaL can be achleved because such assessmenL rarely exlsLs. 1here ls wldespread buy-ln Lo Lhe ldea
LhaL corporaLes should harness Lhelr 'core buslness' and noL [usL Lhelr CorporaLe Soclal 8esponslblllLy (CS8)
budgeL Lo boosL developmenL. 1hls premlse underlles Lhe commlLmenLs made ln Lhe 8uslness Call Lo AcLlon
(8C1A). 8uL Lhere ls vlrLually no analysls Lo daLe of how dlfferenL buslness models dellver lmproved lmpacL,
elLher for shareholders or for sLakeholders who are Lhe lnLended beneflclarles.

Secondly, companles are already lnvesLlng conslderably ln reporLlng publlcly on lssues LhaL go well beyond
shareholder value, buL Lhls lnLenslve work ls noL answerlng key quesLlons abouL Lhe role Lhey play ln
supporLlng developlng economles. MosL large companles now reporL ln accordance wlLh Lhe Clobal
8eporLlng lnlLlaLlve (C8l), Lhe ma[orlLy add exLernal assurance Lo Lhls process, and mosL also publlsh
susLalnablllLy revlews elLher wlLhln or as an addlLlon Lo Lhelr annual reporLs, and have areas of Lhelr web
slLe addresslng lssues on responslble buslness. 1he efforL LhaL goes lnLo Lhls ls conslderable. ?eL Lhese
exlsLlng efforLs slmply do noL capLure key lndlcaLors of Lhelr conLrlbuLlon Lo developmenL.

1hlrdly, lL ls slmply noL Lrue LhaL ln 2009 buslness ls expecLed Lo Lurn a proflL and noLhlng else. lnvesLors,
cusLomers, sLaff, and oLher local sLakeholders wanL Lo know LhaL buslness conLrlbuLes poslLlvely Lo socleLy
and Lhe economy, and Lhere ls a buslness case for companles Lo generaLe Lhe lnformaLlon Lo demonsLraLe
LhaL. Maklng clalms wlLhouL evldence ls a rlsky sLraLegy. Maklng clalms based on anecdoLe ls Lhe norm, buL
Lhe appeLlLe for more subsLanLlal evldence ls growlng.

llnally, developmenL agencles noL only recognlse Lhe role of Lhe prlvaLe secLor ln drlvlng developmenL buL
are lncreaslngly lnvesLlng donor resources - funded by Laxpayers - ln enhanclng Lhe role of buslness. 1hls
lncludes work wlLh large corporaLes and lnvesLmenL funds, whlch would noL be [usLlflable as beoeflclotles of
developmenL ald buL are [usLlflable as effecLlve lnLermedlarles for dellverlng developmenL resulLs. lor
publlc accounLablllLy, and for Lhe developmenL communlLy Lo lnvesL resources mosL effecLlvely Lowards Lhe
Mlllennlum uevelopmenL Coals (MuCs), lL needs Lo have a beLLer ldea of Lhe reLurn on lnvesLmenL ln Lhls
and oLher areas of acLlon.


A

2. Assess|ng the |mpacts of bus|ness on deve|opment: contrast|ng
approaches

lLy Lhe poor execuLlve who ls glven Lhe [ob of movlng forward a company's undersLandlng of lLs lmpacL ln
developlng counLrles. A qulck revlew of whaL ls ouL Lhere, and whaL oLher companles are dolng, glves one
clear response: a wlde array of ad hoc approaches, and no dlscusslon of how Lhey compare, whaL ls useful
when, or why parLlcular approaches have been chosen.

LxlsLlng approaches Lo undersLandlng Lhe lmpacL of buslness on developmenL vary ln Lhelr scope and
conLenL, and on whaL need and whose need Lhey serve. 1he focus may be a communlLy lnvesLmenL pro[ecL,
or Lhe global buslness. 1he user may be a counLry manager, audlLors, or readers of Lhe Annual 8eporL.
1here are many posslbly ways Lo caLegorlse assessmenL. We hlghllghL four maln approaches used by
companles:
1. Local assessmenL: llvellhood lmpacLs and sLakeholder vlews of a flrm or lnlLlaLlve,
2. value chaln fooLprlnL: enLerprlse and poverLy lmpacL of Lhe enLlre value chaln wlLhln an
economy,
3. Lconomlc conLrlbuLlon: mulLlpller effecL of a buslness ln Lhe naLlonal economy and
conLrlbuLlon Lo gross domesLlc producL,
4. 8eporLlng agalnsL flxed lndlcaLors: performance of a company agalnsL a 'scorecard' of
lndlcaLors.

1hese approaches Lend Lo be applled aL dlfferenL levels: llvellhood lmpacLs are usually used Lo assess
lmpacL of a speclflc producLlon slLe, producL, or a corporaLe responslblllLy lnlLlaLlve. value chalns and
economlc conLrlbuLlon Lend Lo measure Lhe conLrlbuLlon of a large buslness havlng a slgnlflcanL affecL on
Lhe reglonal or naLlonal economy. 1he scorecard approach ls ofLen applled aL Lhe level of a mulLlnaLlonal
company, as ln Lhe C8l, Lhough ls also used (by varlous developmenL banks) Lo assess a speclflc lnvesLmenL.
1able 1 provldes an overvlew, wlLh examples, of Lhe four approaches used by companles

no caLegorlsaLlon ls perfecL, and boundarles necessarlly blur. 1he local level approach (approach 1) resLs on
gaLherlng daLa and perspecLlve from producers and parLlclpanLs ln Lhe buslness. 1hls ls also parL of Lhe
value chaln mapplng (approach 2), parLlcularly when done as a poverLy fooLprlnL (as by Cxfam). Powever,
value chaln mapplng (approach 2) usually goes furLher, by ploLLlng ouL Lhe buslness llnkages LhaL exLend
Lhrough Lhe supply and dlsLrlbuLlon chaln. Llnks ln Lhe chaln are also covered ln mulLlpller analysls
(approach 3), Lhough Lhe laLLer Lends Lo make more use of economlsLs' lnpuL-ouLpuL Lables and less use of
local level lnformanLs Lo quanLlfy Lhose llnkages. Scorecard assessmenLs (approach 4) all share a
framework, buL Lhe conLenL of Lhls varles enormously and can cover any aspecL of lmpacL.
We dlsLlngulsh broadly beLween Lwo ma[or purposes of lmpacL assessmenL, buL acknowledge overlap,
subLleLy and dlverslLy ln purpose:
1o prove: Lo demonsLraLe Lo parLners, Lhe publlc and oLher sLakeholders LhaL a company ls havlng
valuable lmpacLs on developmenL. 'WhaL' lmpacLs are achleved ls a key quesLlon.
1o |mprove: Lo base declslon-maklng on an lnformed undersLandlng of currenL lmpacLs, so as Lo
furLher lmprove Lhem. 1hls requlres lnslghLs lnLo 'whaL' ls achleved buL also 'why.'




>

1ab|e 1: Cverv|ew and categor|sat|on of d|fferent approaches
Approach and example Applled Lo:
1 Loca| assessment: L|ve||hood |mpact & stakeho|der v|ews
SA8Mlller/wC assessmenL of 'enLerprlse developmenL' ln Lhe
supply chaln
ulageo/L8C assessmenL of 'WaLer of Llfe' pro[ecLs
8ody Shop assessmenL/ verlflcaLlon of CommunlLy 1rade
Company lnlLlaLlve Lo boosL beneflLs Lo
Lhe poor
Anglo Amerlcan Soclal Lconomlc AssessmenL 1oolbox A mlne ln a developlng counLry
vodafone assessmenL of Lhe lmpacL of moblle phones Moblle phone users, aggregaLed.
W8CSu lmpacL assessmenL framework Scope deLermlned by company

2 Va|ue cha|n footpr|nt|ng
unllever/Cxfam assessmenL of unllever lndonesla A mulLlnaLlonal ln one counLry
(lndonesla) and lLs enLlre value chaln
(supply, dlsLrlbuLlon, reLall)
Cxfam poverLy mapplng A naLlonal or mulLlnaLlonal company,
lmpacL wlLhln one counLry
Cul, llC, Snv, pro-poor mapplng of Lourlsm value chalns 1ourlsm secLor, comprlslng mulLlple
flrms, ln one desLlnaLlon

3. Lconom|c contr|but|on and mu|t|p||ers
unllever: economlc fooLprlnL ln SouLh Afrlca
SA8Mlller: conLrlbuLlon of SA8 Lo SouLh Afrlca's economy
A mulLlnaLlonal ln one counLry, all
dlrecL and lndlrecL affecLs.
1ourlsm SaLelllLe AccounLs LnLlre Lourlsm secLor ln one counLry

4 Scorecard performance
8eporLlng agalnsL C8l lndlcaLors. L.g. 8arclays, ulageo,
SA8Mlller eLc.
Company performance, lnLernaLlonally
SeA8 framework of A8u and Lnglneers agalnsL overLy ro[ecL-speclflc lnvesLmenL
Cood for uevelopmenL 'badge' concepL Company performance, lnLernaLlonally

AssessmenL by compooles of Lhelr conLrlbuLlon Lo developmenL ls relaLlvely new, compared Lo Lhe
subsLanLlal experlence of varlous donors ln assesslng how Lhelr lnvesLmenL ln prlvaLe secLor generaLes
developmenL lmpacL, and of economlsLs ln assesslng Lhe economlc value of Lypes of lnvesLmenL. 1here are
valuable lessons Lo draw from how developmenL banks use a 'scorecard' approach Lo assess Lhe
developmenL conLrlbuLlon of an lnvesLmenL along wlLh lLs flnanclal vlablllLy, and how donors seek Lo
capLure prlvaLe secLor developmenL lmpacLs, from [obs Lo llvellhoods Lo markeL change. useful lnslghLs
from Lhe donor world are covered furLher ln SecLlon 7.

1ab|e 2: Ioc| of assessments by deve|opment organ|sat|ons
locl used by/ used for
1 Loca| assessment: L|ve||hood |mpact & stakeho|der v|ews
Analysls of local lmpacLs of a flrm - local employmenL & wages, skllls
developed, enLerprlse developmenL, culLural/soclal problems, resource
affecLs.
nCCs
1o hlghllghL negaLlve lmpacLs or galns
Lo poor from small enLerprlse devL
2 Add|t|ona| quant|f|ab|e econom|c act|v|ty, Lkk
!obs creaLed, SML conLracLs generaLed, producLlon volume, exporLs, Lax
revenue
Lconomlc 8aLe of 8eLurn (L88): quanLlfles lncremenLal economlc galn Lo
socleLy, relaLlve Lo lnvesLmenL cosL.
Su donors & devL banks
1o quanLlfy or compare addlLlonal
economlc acLlvlLy from lnvesLmenL
3 Dynam|c & structura| affects
ConLrlbuLlon Lo developmenL of markeLs, lnfrasLrucLure, Lechnology,
skllls, lnnovaLlon, compeLlLlon, buslness llnkages.
CuallLaLlve lndlcaLors, excepL when economlsLs calculaLe conLrlbuLlons
Lo growLh raLes Lhrough regresslon analysls.
uevelopmenL banks and oLher Su
donors.
1o capLure longer-Lerm and sLrucLural
developmenL conLrlbuLlon


B

3. Approach 1: Loca| assessment
Approach 1 focuses on dlrecL lmpacLs aL Lhe local level LhaL accrue malnly Lo poor people: flows of lncome,
capaclLy developmenL and oLher changes ln llvellhoods affecLlng employees, enLrepreneurs, resldenLs,
parLlclpanLs of local lnsLlLuLlons. 1hey resL heavlly on gaLherlng Lhe vlews of local sLakeholders or collecLlng
daLa from local sLakeholders, concernlng lmpacLs LhaL fall wlLhln Lhe deflned scope.

1hls approach can be applled ln Lhree ways:
l. Lo an expllclL loltlotlve deslgned Lo boosL developmenL lmpacL,
ll. Lo Lhe overall operaLlon of a bosloess or slLe of operaLlon
lll. Lo a parLlcular ptoJoct or markeL sysLem (based on an aggregaLlon of many local lmpacLs)

When companles assess lmpacLs of corporaLe responslblllLy lnlLlaLlves, ranglng from communlLy lnvesLmenL
donaLlons, Lo adapLaLlon of Lhelr supply chaln, assessmenL ls usually based on local sLakeholder
lnformaLlon. 1hese acLlvlLles have a speclflc ob[ecLlve of enhanclng soclal value, so need Lo be assessed
accordlng Lo wheLher Lhey dellver LhaL. lmpacL here ls someLhlng oJJltloool Lo normal buslness, or
someLhlng LhaL changes compared Lo prevlous pracLlce. ln Lhls case, Lhe approach does noL requlre
assessmenL of all aspecLs of core buslness. lL may be done Lo 'prove' value of pasL work, and equally Lo
'lmprove' fuLure managemenL by hlghllghLlng prlorlLy areas.

ln pracLlce, lmpacL assessmenL of communlLy lnvesLmenL has Lended Lo be weak, focuslng on ouLpuLs
(dollars spenL, workshops held) noL ouLcomes (llvellhoods secured), whlle more 'lncluslve buslness
approaches' such as adapLlng supply chalns, Lend Lo be falrly new and noL yeL assessed. Powever, examples
are emerglng of assessmenLs LhaL go beyond [usL ouLpuLs. 1hese examples also add value because Lhey
aggregaLe resulLs up from several local levels, Lo bulld a blgger plcLure and enable comparlson beLween
slLes.
SA8Mlller's recenL LnLerprlse uevelopmenL reporL summarlses flndlngs from assessmenL of Lhelr
supply chaln lnlLlaLlves ln a number of counLrles and lllusLraLes Lhe varlablllLy ln lmpacL supply chaln
approaches can have (SA8Mlller 2008).
1he assessmenL of ulageo's WaLer of Llfe pro[ecL Lakes a sLandard approach Lo pro[ecLs ln over 30
counLrles, looklng boLh aL Lhe number of beneflclarles reached, and longer lasLlng lmpacL lssues
relaLlng Lo susLalnablllLy and communlLy ownershlp (ulageo 2007).
MulLlnaLlonal LnLerprlses (MnLs) LhaL slgned up Lo Lhe 8uslness Call Lo AcLlon are beglnnlng Lo look
aL how resulLs can be monlLored and assessed.
Local assessmenL can also be applled Lo an enLlre buslness operaLlon - such as a mlne, hoLel, or foresL - or a
producL - such as moblle phones. llndlngs from consulLaLlon wlLh sLakeholders and mlcro-economlc daLa
are aggregaLed Lo bulld an overall plcLure, usually lncludlng boLh poslLlve and negaLlve lmpacLs, and
economlc and soclal daLa.
A well known example ls Anglo-Amerlcan's (AA) Soclo-Lconomlc AssessmenL 1oolbox
ll
whlch ls ln
place aL 60 AA slLes ln 16 counLrles. 1hls ls noL so much Lo assess Lhe ex-posL lmpacL of an lnlLlaLlve
LhaL has happened, buL Lo undersLand Lhe currenL lmpacLs of operaLlons and how Lo enhance Lhem
over Llme Lhrough engagemenL wlLh sLakeholders. lL ls Lhus more of a declslon-maklng Lool, and
noL a Lool for demonsLraLlng a snapshoL of resulLs.

AssessmenL of Lhe lmpacL of a new producL or markeL sysLem can be based on oqqteqotloo of llvellhood
lmpacL and mlcro-economlc daLa from sLakeholders aL many locaLlons. lor example:


?

1he LLhlcal 1rade lnlLlaLlve (L1l) revlewed lmpacLs of eLhlcal Lrade, by collecLlng daLa dlrecLly from
companles and workers lnvolved ln L1l. CuallLaLlve and quanLlLaLlve lndlcaLors relevanL Lo Lhe L1l
base code almed Lo assess dlrecL lmpacLs on workers and companles, and also Lo measure Lhe
broader soclal lmpacLs (poslLlve and negaLlve) on oLhers (L1l 2003).
vodafone's assessmenL of Lhe lmpacL of moblle phones ln lndla collaLes daLa from many users on
how Lhey use Lhelr new Lechnology (vodafone 2009).
AssessmenL of Lhe developmenL slgnlflcance of corporaLe-communlLy [olnL venLure lodges (a
relaLlvely new buslness model) has been done ln SouLhern Afrlca by aggregaLlng and comparlng
resulLs from many such buslnesses (Massyn and koch 2004).

Local level assessmenL ls aL Lhe hearL of Lhe World 8uslness Councll for SusLalnable uevelopmenL (W8CSu)
new 'lmpacL assessmenL framework' whlch ls deslgned Lo be applled by buslnesses aL any level. SLep 1 of
Lhe W8CSu approach ls Lo deflne Lhe scope of Lhe assessmenL (wheLher a CS8 lnlLlaLlve or an enLlre
buslness operaLlon), from whlch lL moves on Lo deflnlng relevanL lndlcaLors wlLh sLakeholders).

1he 8ase of Lhe yramld lmpacL AssessmenL lramework, developed aL Lhe Wllllam uavldson lnsLlLuLe aL Lhe
unlverslLy of Mlchlgan, slmllarly focuses on Lhe local assessmenL of buslness models servlng Lhe so-called
'base of Lhe pyramld' (8o). ln order Lo enhance Lhe poslLlve effecLs of such buslness lnlLlaLlves on poverLy
allevlaLlon Lhe framework's prlmary ob[ecLlve ls Lo glve managers a deLalled undersLandlng abouL Lhe
poslLlve and negaLlve lmpacLs a 8o venLure has on Lhe economlcs, capablllLles, and relaLlonshlps of Lhree
local sLakeholder groups: buyers, sellers, and communlLles (London 2009).

8ox 1: What does |t mean |n pract|ce?
Lxamp|e: D|ageo, Water of L|fe ro[ect
xomples of lmpoct tepotteJ.
ln 2008, 1.18 mllllon beneflclarles have been reached, Lhe ma[orlLy belng ln nlgerla (420,000) and kenya
(372,690) and Chana (138,960)
WaLer supplles and sanlLaLlon close Lo people's home has also lndlrecL beneflLs, for lnsLance an lncreased sense of
securlLy, parLlcularly among women. PealLh beneflLs, lncreased soclal harmony and new sources of lncome
xomples of poestloos oskeJ ot commoolty level
uo you have access Lo waLer and sanlLaLlon ln your vlllage? lf so, slnce when?
uld your dally llve change wlLh havlng access Lo waLer ln your communlLy?
AparL from poslLlve lmpacLs can you Lhlnk of any problems LhaL arose due Lo access Lo waLer and sanlLaLlon?

Lxamp|e: SA8M|||er, Lnterpr|se Deve|opment |n the|r supp|y cha|n: 1he Sma||ho|der rogramme
xomples of lmpocts tepotteJ
ln 2007/08 over 10,300 smallholder farmers engaged ln SA8Mlller programmes ln Afrlca and lndla. 1he number ls
esLlmaLed Lo reach nearly 17,000 ln 2009
ln lndla, farmers recelve a preferenLlal prlce, whlch ln 2007/08 was around 3 hlgher Lhan Lhe open-markeL raLe
xomples of poestloos oskeJ to fotmets
Pave you been lnvolved ln Lhe SA8Mlller smallholder programme?
uld Lhe parLlclpaLlon enable you Lo hlre addlLlonal Lemporary labour durlng harvesLlng?
uld Lhe programme have any long-Lerm lmpacL on your buslness? L.g. more dlsposable lncome?

!"##$%&'
1he key feaLure of Lhls approach ls Lhe use of local level lnformaLlon and perspecLlve gaLhered from flrms,
workers, enLerprlse and resldenLs. verslons of Lhls approach are used by companles, nCCs and pro[ecL
funders, researchers, and buslness-relaLed schemes.
lL ls useful for:
LlLher 'provlng' lmpacL Lo daLe, or 'lmprovlng' fuLure managemenL based on ln depLh
undersLandlng of whaL works, how,


76

Croundlng Lhe lmpacL assessmenL ln Langlble local change and real daLa, lncludlng measurables
such as lncomes, [obs.
Coverlng boLh measurable economlc resulLs wlLh less Langlble daLa on preferences, soclal, and
envlronmenLal lssues
AggregaLlng and comparlng some key daLa beLween dlfferenL slLes, Lo learn more abouL blg plcLure
lmpacLs and areas of hlgh/low performance:
LlmlLaLlons:
unllkely Lo dellver slmple quanLlLaLlve answers LhaL summarlse lmpacL, nor sLandardlsed resulLs
LhaL enable comparlsons such as reLurn on lnvesLmenL, or ranklng beLween slLes.
8equlres fleldwork aL communlLy level, so cannoL be done rapldly from Lhe desk.
uoes noL assess lmpacLs aL a large scale, such as across a ma[or value chaln, or enLlre economy.

4. Approach 2: Va|ue cha|n maps and poverty footpr|nts
1he flrsL approach Lends Lo focus on Lhe mosL vlslble, dlrecL and local lmpacLs of an operaLlon. 1he 'value
chaln' approach expllclLly Lracks Lhrough dlfferenL nodes and llnks of Lhe value chaln, Lo conslder a wlder
range of lmpacLs. 1he 'value chaln' lncludes Lhe supply chaln from raw maLerlal Lo processlng, Lhe
dlsLrlbuLlon and reLall chaln, plus flnal consumpLlon and dlsposal or recycllng. lL ls Lhus llkely Lo also cover a
range of geographlcal locaLlons. 8y lncludlng all aspecLs of supply, dlsLrlbuLlon, and consumpLlon, lL seeks Lo
bulld a 'fooLprlnL' (a concepL Laken from envlronmenLal lmpacL work). A 'fooLprlnL' ls noL a slngle numerlcal
answer, buL an overall plcLure, and because of Lhls lL ls posslble Lo see whlch parLs of Lhe fooLprlnL are mosL
slgnlflcanL ln dlfferenL ways and Lo whom.

Many sources of daLa are shared wlLh Approach 1: deLalled flrm level lnformaLlon on [obs and wages, vlews
of workers, suppllers and oLher sLakeholders. 8uL lL also Lracks flows aL a hlgher level, such as Lo
lnLermedlarles, relaLed secLors, shareholders and 1reasury, and wlLh more emphasls on aggregaLed
flnanclal daLa, golng beyond lndlvldual personal sLorles. lL shares wlLh Approach 3 an emphasls on llnks Lo
oLher secLors and oLher lndusLrles (LransporL, banklng, dlsLrlbuLlon eLc) buL generally does noL cover macro-
economlc lmpacLs such as conLrlbuLlon Lo ouLpuL or growLh. 1hls approach can be applled Lo a secLor or a
company. Lxamples lnclude:
A sloqle lotqe compooy. AssessmenL of Lhe lmpacL of unllever lndonesla, done [olnLly wlLh Cxfam
(Clay 2003). 1hls remalns Lhe besL known and mosL deLalled example of value chaln mapplng
applled Lo a slngle company. lL revealed Lhe dlfferenLlal lmpacL of dlfferenL parLs of Lhe producLlon
sysLem. As such lL was useful for boLh 'provlng' and 'lmprovlng': lL demonsLraLed Lo Cxfam and Lhe
wlder publlc Lhe conslderable poslLlve conLrlbuLlon LhaL Lhls one company made Lo employmenL ln
lndonesla. And because of Lhe dlsaggregaLed analysls, lL also provldes a sound basls for maklng
declslons on where lmpacL can be furLher enhanced. lor example, lL demonsLraLed Lhe
developmenL lmporLance of Lhe dlsLrlbuLlon chaln noL [usL Lhe supply chaln, and of acLlons Lake
durlng Lhe Aslan flnanclal crlsls Lo help suppllers adapL.
A sectot comptlsloq mooy fltms. value chaln mapplng of Lhe Lourlsm secLor by Lhe Cul, llC and Snv
Lracks Lhe flow of LourlsL expendlLure Lhrough all aspecLs of Lourlsm buslness ln a desLlnaLlon. 1he
resulLlng 'map' assesses Lhe scale and locaLlon of lncome LhaL flows Lo Lhe poor. Crlglnally Lhey
were done Lo 'lmprove': Lo conLrlbuLe Lo plannlng as Lo where lnLervenLlons can help unblock
furLher beneflLs Lo Lhe poor, and Lhls becomes easler as a pool of comparable sLudles, wlLh
emerglng benchmarks, develops. Powever, Lhey have gone somewhere Lo also 'provlng' LhaL


77

Lourlsm can generaLe subsLanLlal pro-poor lncome of around a quarLer of desLlnaLlon spendlng
(Ashley and MlLchell 2008).
Cxfam and Cul are conLlnulng Lo develop Lhese approaches, wlLh forLhcomlng Lools for mapplng pro poor
flows and boLLlenecks ln Lourlsm value chalns (Cul) and for assesslng 'poverLy fooLprlnLs' (Cxfam)
comprlslng economlc and soclal lmpacLs on poor people across Lhe value chaln. As Cul develop Lhelr seL of
sLudles, comparlson of value chalns maps done Lo daLe ls now also generaLlng 'benchmarks' of hlgh and low
poverLy performance (MlLchell and Ashley 2009). Cxfam's Lool ls also developlng, and ln pracLlce lncludes a
blend of value chaln mapplng, llvellhood lnformaLlon (our caLegory 1) and macro-economlc analysls (our
caLegory 4).

A sLrengLh of Lhls approach ls LhaL lL comblnes local level daLa from poor people wlLh a blg plcLure of Lhe
overall lmpacL, maklng lL posslble Lo see and compare areas of hlgh and low lmpacL. lL also relaLes daLa on
lmpacL on poor people Lo Lhe commerclal sLrucLures and channels Lhrough whlch Lhe company or secLor
works. 8y lLs naLure lL Lherefore lends lLself Lo analysls of 'core buslness' noL [usL communlLy lnvesLmenL.

8ox 2: What does |t mean |n pract|ce?
Lxamp|e: Un||ever Indones|a (UI) and Cxfam collaboraLed ln a [olnL research pro[ecL ln 2003. 1he research consldered
Lhe lmpacLs of unllever lndonesla across Lhe enLlre buslness value chaln.
xomples of lmpoct tepotteJ
Cverall, Lhe research esLlmaLes LhaL Lhe full-Llme equlvalenL (l1L) of abouL 300,000 people make Lhelr llvellhoods
from unllever's value chaln. More Lhan half are wlLhln Lhe dlsLrlbuLlon and reLall chaln.
1oLal value generaLed along Lhe value chaln ls esLlmaLed aL uS$ 633 mllllon. Cf Lhls, ul earns abouL uS$ 212
mllllon, Lhe remalnlng uS$ 421 mllllon ls dlsLrlbuLed among oLher acLors ln Lhe chaln wlLh greaLer concenLraLlon
among acLors who are ln dlrecL lnLeracLlon wlLh ul.
xomples of poestloos oskeJ to poot coosomets
uo you know wheLher unllever personal care producLs are avallable ln your local shops?
uo you buy unllever producLs for your personal care?
ln how far does uslng Lhese producLs have an lmpacL on your dally llfe?

1he unllever lndonesla work ls falrly well known, flrsLly for generaLlng flndlngs LhaL Lruly moved
undersLandlng on. lL ls surprlslng buL Lrue LhaL Lhls provlded Lhe flrsL plcLure of how employmenL and
revenue shares were supporLed up and down Lhe value chaln of an MnL lL ls also known as an lnLenslve
and prolonged efforL, whlch has perhaps deLerred oLhers. 8uL comparable work suggesLs LhaL mapplng a
value chaln need noL be so, parLlcularly as meLhods are repllcaLed and lmproved (securlng co-auLhor
agreemenL beLween 2 or more lnsLlLuLlons ls a dlfferenL lssue). value chaln mapplng has been done ln
around 10 Lourlsm desLlnaLlons over Lhe pasL 3 years, and falrly deLalled analysls of Lhe parLlclpaLlon and
earnlngs of poor people ln a Lourlsm value chaln along wlLh opLlons for lmprovemenL can now be done ln
around 30 days (10-13 days of fleldwork). AfLer compleLlng Lhe unllever sLudy, Cxfam furLher sysLemaLlsed
Lhe meLhodology and ls currenLly carrylng ouL anoLher overLy looLprlnL sLudy wlLh Lwo mulLlnaLlonals.
1he meLhodology wlll be open Lo consulLaLlon durlng !uly and AugusL 2009 wlLh Lhe offlclal launch planned
aL Lhe end of Lhls year (8orkenhagen and Zaremba 2009).
!"##$%&'
1he key feaLure of Lhls approach ls Lhe comblnaLlon of lnformaLlon from local sLakeholders and corporaLe
operaLlons, broughL LogeLher lnLo a 'blg plcLure' coverlng Lhe enLlre value chaln from maLerlal supply Lo
dlsLrlbuLlon and consumpLlon. verslons of Lhls approach are used by companles, researchers, nCCs and
donors. Companles may do lL Lo elLher prove or lmprove. 1he oLhers do lL malnly Lo asslsL ln dlagnoslng
how Lo make value chalns perform beLLer.
lL ls useful for:


7"

8oLh 'provlng' lmpacL of Lhe enLlre value chaln, and 'lmprovlng' fuLure managemenL based on
dlsaggregaLed analysls. 1he plcLure of Lhe enLlre buslness or secLor can be summarlsed ln a few
overall sLaLlsLlcs buL also conLalns deLall of dlfferenL parLs of Lhe fooLprlnL as a basls for declslon-
maklng.
Comblnlng assessmenL of lmpacL wlLh assessmenL of commerclal operaLlon, so Lhe lnLer-
dependence of Lhe Lwo are clear.
Comblnlng Langlble daLa on [obs, wages, enLerprlses llnkages and revenue flows, wlLh poLenLlal Lo
add non-flnanclal lssues, such as governance wlLhln Lhe chaln and sLakeholder vlews on soclal
lmpacL, and Lhe 'blg plcLure' of naLlonal lmpacL.
LlmlLaLlons
A slgnlflcanL underLaklng requlrlng fleldwork, deLalled corporaLe daLa and skllled aggregaLlon and
analysls.
Llkely Lo capLure sLaLlc plcLure buL noL dynamlc lmpacLs on local developmenL

S. Approach 3: Assessment of mu|t|p||ers and contr|but|on to econom|c
act|v|ty
A somewhaL dlfferenL approach Lo calculaLlng Lhe naLlonal economlc fooLprlnL of a company draws from
economlcs, raLher Lhan from local level assessmenL. 1here ls a long-sLandlng LradlLlon amongsL economlsLs
of assesslng Lhe economlc slgnlflcance of a secLor ln Lerms of Lhe [obs lL creaLes dlrecLly, and Lhe
'mulLlpllers' by whlch lL sLlmulaLe addlLlonal acLlvlLy ln Lhe economy. 8ecenLly Lhls approach has been
applled Lo Lhe lmpacL of large companles - noLably unllever and SA8Mlller ln SouLh Afrlca.

1hls approach focuses on Lhe scale of economlc acLlvlLy generaLed. Llke approach 2, lL draws on company
lnformaLlon Lo bulld an aggregaLe plcLure of Lhe [obs supporLed across dlfferenL parLs of Lhe value chaln,
and flows of lncome such as wages. 8eyond LhaL, lL draws less on local level lnformaLlon concernlng lmpacLs
on Lhe poor, and more on economlc Lools such as lnpuL-ouLpuL Lables, Lo calculaLe addlLlonal economlc
acLlvlLy LhaL can be aLLrlbuLed Lo Lhe company. lL Lends Lo be used Lo 'prove' - or Lo demonsLraLe Lhe
slgnlflcanL scale of acLlvlLy. lL ls slmllarly used ln 1ourlsm SaLelllLe AccounLs
lll
Lo lllusLraLe Lhe slze of Lhe
secLor, whlch ls oLherwlse under-accounLed ln naLlonal accounLs (and Lhus by MlnlsLrles of llnance).

1he Lwo maln examples used by companles so far are:
unllever: Measurlng unllever's Lconomlc looLprlnL: 1he Case of SouLh Afrlca" done ln
collaboraLlon wlLh lnSLAu, one of Lhe world's leadlng buslness schools (kapsLeln 2008).
SA8Mlller: 1he conLrlbuLlon of Lhe SouLh Afrlcan 8rewerles Lo Lhe SA economy" ln whlch case Lhe
8ureau of Lconomlc 8esearch aL Lhe unlverslLy of SLellenbosch has assessed Lhe company's
conLrlbuLlon (SA8Mlller 2008a).

As argued by LLhan kapsLeln (2008), auLhor of Lhe unllever SouLh Afrlca reporL, Lhe very polnL ls Lo geL
away from an anecdoLal case for CS8, and prove Lhe hard evldence of Lhe developmenL lmpacL a company
can have.



7#


8ox 3: What does th|s mean |n pract|ce?
Lxamp|e: SA8M|||ers contr|but|on to the South Afr|can economy, assessed by Lhe 8ureau of Lconomlc 8esearch aL
Lhe unlverslLy of SLellenbosch
xomples of lmpocts tepotteJ (ecooomlc coottlbotloo)
ln 2007, SA8Mlller's conLrlbuLed 3.3 Lo Lhe SouLh Afrlcan Cu
ln addlLlon Lo dlrecLly employlng 8,780 workers, 73 of whom are from dlsadvanLaged groups, SA8's
operaLlons supporL an esLlmaLed 48,000 [obs aL SA8's flrsL round suppllers. ln all, 378 000 full-Llme [obs (or
3.0 of LoLal employmenL ln SouLh Afrlca) can be dlrecLly or lndlrecLly llnked back Lo Lhe producLlon and sale
of SA8's producLs
xomples of poestloos.
1o managers: whaL ls LoLal spendlng on each supply componenL?
1o economlsLs: how does $x of spendlng on LransporL LranslaLe lnLo economlc ouLpuLs and [obs, accordlng Lo
Lhe lnpuL-ouL Lables and Soclal AccounLlng MaLrlx for SouLh Afrlca?


!"##$%&'
1he dlsLlncLlve elemenLs of Lhls approach are LhaL lL uses economlc Lools Lo provlde clear aggregaLed daLa
on Lhe economlc conLrlbuLlon of a company. lL covers Lhe enLlre corporaLe value chaln, from an economlc
perspecLlve. Such approaches are ofLen used by economlsLs and so far by few companles.
lL ls useful Lo:
quanLlfy and demonsLraLe Lhe overall economlc conLrlbuLlon of a secLor or large buslness,
move beyond anecdoLe Lo hard numbers,
make comparlsons beLween dlfferenL parLs of Lhe buslness, or beLween comparable assessmenLs
across secLors, counLrles or companles.
LlmlLaLlons:
Less llkely Lo generaLe dlsaggregaLed daLa for maklng declslons on speclflc parLs of Lhe buslness,
nor flnely LexLured analysls of causes of lmpacL and areas Lo develop
Less llkely Lo lnclude non-flnanclal lmpacLs, or reflecL non-flnanclal lssues prlorlLlsed by dlfferenL
sLakeholders

6. Approach 4: keport|ng aga|nst scorecards and standard|sed
|nd|cators
A scorecard or seL of lndlcaLors ls slmply a framework for reporLlng, Lhe coverage of whlch can vary
enormously. AL presenL, we have a seL of lndlcaLors - from Lhe C8l - LhaL are exLremely well used, buL are a
poor reflecLlon of developmenL lmpacL. And we have emerglng scorecards LhaL seek Lo beLLer capLure
developmenL lmpacL, buL are noL yeL esLabllshed, or are used only by speclflc developmenL lnsLlLuLlons.

6.1 G|oba| keport|ng In|t|at|ve and UN G|oba| Compact
MosL companles do noL lnvesL ln lnLenslve assessmenL of Lhe developmenL lmpacLs of speclflc buslness
operaLlons, buL a greaL many do reporL Lhelr performance agalnsL flxed lnLernaLlonal lndlcaLors, parLlcularly
Lhose covered ln Lhe Clobal 8eporLlng lnlLlaLlve (C8l). 1he C8l ls noL a Lool for lmpacL assessmenL ln Lhe
same way as approaches 1 Lo 3. 8uL lL ls a Lool LhaL ls lncreaslngly used by companles for communlcaLlng Lo
Lhe publlc concernlng Lhelr performance as good corporaLe clLlzens. Accordlng Lo a global kMC survey
(kMC 2008) abouL reporLlng Lrends, around Lhree-quarLers of annual susLalnablllLy reporLs were drawn up
wlLh Lhe ald of Lhe C8l gulde.



7@

C8l lndlcaLors are more a reflecLlon of corporaLe behavlour Lhan lmpacL ln SouLhern economles. 1hey cover
a range of corporaLe performance lssues ranglng from generaLlng dlrecL economlc value Lo energy
consumpLlon and lndlgenous rlghLs, wlLh relaLlvely few Louchlng on how Lhe buslness affecLs developmenL
ln Lhe SouLh. 1he mosL relevanL lndlcaLors are
Labour pracLlces and decenL work lssues such as offerlng occupaLlonal healLh and safeLy, dlverslLy
and promoLlon of equal opporLunlLles
CommunlLy lnvesLmenL
AnLl-corrupLlon acLlons
lnvesLmenL and procuremenL pracLlces
racLlces regardlng chlld, forced and compulsory labour
LnvlronmenLal good pracLlce

As currenLly consLrucLed, C8l lndlcaLors are a poor gulde Lo developmenL lmpacL ln Lhe SouLh. Powever, lL
ls worLh conslderlng 2 sLrengLhs of C8l ln reporLlng and encouraglng enhanced developmenL performance:
1he use of sLandard lndlcaLors, guldellnes and raLlngs, plus adopLlon by many leadlng companles,
allows comparlson of daLa from dlfferenL buslnesses, and Lhus lnLroduces an elemenL of
compeLlLlon Lo do beLLer,
lndlcaLors are clear, and Lhus easlly LranslaLed lnLo aLLalnmenL LargeLs for mangers. 1hus Lo Lhe
exLenL LhaL Lhey do cover developmenL lssues, Lhey can help ln LranslaLlng alms lnLo acLlon.

Lxamples of C8l reporLlng used by mulLlnaLlonals LhaL do operaLe exLenslvely ln Lhe SouLh lnclude 8arclays,
ulageo, and uPL. All Lhree have an advanced C8l performance accordlng Lo C8l AppllcaLlon levels
lv
. lor
example:
8arclays publlshes a SusLalnablllLy 8evlew, and also shows on lLs webslLe where lL provldes
lnformaLlon agalnsL relevanL C8l lndlcaLors, wheLher lnslde Lhe annual reporL or elsewhere ln lLs
publlc lnformaLlon. 1here ls a wealLh of lnformaLlon, wlLh an emphasls on ensurlng LhaL lL ls all
publlcally and easlly avallable, buL relaLlvely llLLle LhaL relaLes Lo lmpacL on developlng counLrles.
1here are flgures for Lhe growLh of non-uk communlLy lnvesLmenL spendlng, commlLmenLs noL Lo
employ chlld labour, deLalls of above-LargeL sourclng from black empowered enLerprlses by A8SA ln
SouLh Afrlca, and a few worldwlde flgures on economlc lmpacL such as an annual global sourclng
spend of 7bn for lLs 23,000 suppllers. An lnLenLlon Lo underLake furLher research on economlc
lmpacL ln emerglng markeLs ln 2008 ls sLaLed (8arclays 2007). lL ls clear LhaL conslderable company
lnvesLmenL musL have gone lnLo Lhe C8l reporLlng, whlch demonsLraLes performance agalnsL a
wlde range of lssues. 8uL Lhls seems Lo offer small opporLunlLy for elLher provlng or lmprovlng
speclflc developmenL conLrlbuLlon.
A slmllar plcLure emerges from Lhe reporLlng by ulageo and uPL. ulageo's 2008 CorporaLe
ClLlzenshlp 8eporL (CC8) lncludes reporLlng agalnsL C8l lndlcaLors. 1he CC8 lncludes some oLher
daLa, noL requlred by C8l, buL perhaps of more relevance when looklng for developmenL lmpacL,
such as an ln-depLh assessmenL of Lhelr WaLer of Llfe pro[ecL as a demonsLraLlon of Lhe lmpacLs of
communlLy lnvesLmenL (see Approach 1 above). Slmllarly, uPL (parL of Lhe dpwn Croup) reporL
agalnsL sLandard C8l lndlcaLors, and ln Lhelr SusLalnablllLy 8eporL 2008 also provlde deLalled
lnformaLlon abouL developmenL relaLed lssues, such as Lhelr parLnershlp wlLh unlCLl and Lhe
kenya MlnlsLry of PealLh Lo flghL agalnsL Malarla.



7=


8ox 4: What does th|s mean |n pract|ce?
Lxamp|e: 8arc|ays, GkI
xomples of lmpocts tepotteJ.
CommunlLy lnvesLmenL: Clobally, 8arclays lnvesLed 32.4mllllon ln communlLy pro[ecLs. non-uk communlLy
lnvesLmenL spendlng grew from 13 of Lhe LoLal (6.9m) Lo 26 (13.3m).
SusLalnable banklng servlces ln emerglng markeLs: 8arclays emerglng markeL buslness lncreased Lhe LoLal
number of A1Ms Lo 704, up from 293 ln 2006.
Lconomlc lmpacL: ln 2007, 8arclays had an annual global sourclng spend of 7bn for lLs 23,000 suppllers
(10,000 maklng up 90 of Lhelr Lhlrd-parLy spend). An lnLenLlon Lo underLake furLher research on economlc
lmpacL ln emerglng markeLs ln 2008 ls sLaLed.
xomples of poestloos oskeJ of moooqets.
uo our labour sLandards meeL lnLernaLlonal sLandards, e.g. lnLernaLlonal Labour CrganlsaLlon (lLC) across Lhe
whole value chaln?
Pow many employees have parLlclpaLed ln anLl-corrupLlon workshop?
WhaL percenLage of employees parLlclpaLed ln volunLeerlng schemes naLlonally and lnLernaLlonally?


Membershlp ln Lhe un Clobal CompacL can also be consldered as a way for companles Lo reporL on how
Lhelr buslness operaLlons are allgned wlLh susLalnable developmenL prlnclples. CommlLmenL ls based on 10
unlversally accepLed prlnclples ln Lhe areas of human rlghLs, labour, envlronmenL, and anLl-corrupLlon.
AddlLlonally, one of Lhe expllclL commlLmenLs LhaL a company makes when lL [olns Lhe un Clobal CompacL
ls Lo produce an annual CommunlcaLlon on rogress (CC). Slnce lLs launch ln 2000, Lhe lnlLlaLlve lL has
become Lhe world's largesL corporaLe clLlzenshlp lnlLlaLlve wlLh currenLly more Lhan 4,700 corporaLe
parLlclpanLs and sLakeholders from over 130 counLrles.. Powever, slmllar Lo Lhe C8l Lhe prlnclples are
malnly based on company's general susLalnablllLy performance and Lhus reporLlng agalnsL Lhese does noL
necessarlly glve lnformaLlon abouL dlrecL developmenL lmpacLs ln developlng counLrles.
6.2 keport|ng aga|nst more poverty focused |nd|cators
new lnlLlaLlves are aLLempLlng Lo comblne Lhe lndlcaLor reporLlng and scorlng approach of C8l, wlLh a
dlfferenL range of lndlcaLors LhaL beLLer capLure developmenL lmpacL.

Cne such under dlscusslon ls a 'Cood for uevelopmenL' label (Lllls and keane 2008). AlLhough lnlLlally
dlscussed as a label for ptoJocts (golng beyond falr Lrade crlLerla Lo wlder developmenL lmpacLs), Lhe
currenL proposal ls for Lhe label or badge Lo apply Lo a compooy. 1he premlse ls LhaL currenL reporLlng -
such as C8l or carbon fooLprlnL - are slmply lnadequaLe Lo deflne wheLher a company ls overall havlng a
subsLanLlally poslLlve lmpacL on developmenL and aLLalnmenL of Lhe MuCs. 1he label would resL on a
number of crlLerla relaLed Lo developmenL lmpacL lncludlng:
ConLrlbuLlon Lo poverLy reducLlon, vla falr pay, use of local suppllers, paymenL of Laxes,
enLrepreneurlal Lralnlng, low-prlced producLs for poor consumers eLc
ConLrlbuLlon oLher speclflc MuCS such as unlversal prlmary educaLlon (vla educaLlonal lnvesLmenL
and no use of chlld labour) and lmproved maLernal healLh (vla lnvesLmenL ln healLh and clean
waLer, female llLeracy, affordable producLs).

Whlle Lhere are enormous pracLlcal problems ln developlng a meLhodology LhaL ls generally appllcable and
genulnely capLures 'developmenL lmpacL', Lhls ls a slgnlflcanL shlfL ln Lhlnklng Lo Lrylng Lo assess Lhe
broader developmenL conLrlbuLlon of an enLlre company Lhrough sLandard and comparable lndlcaLors.

1he MuC Scan ls anoLher new Lool, deslgned by Lhe neLherlands CommlLLee for lnLernaLlonal CooperaLlon
and SusLalnable uevelopmenL (nCuC) and SusLalnalyLlcs, whlch enables companles Lo measure Lhelr


7A

poslLlve conLrlbuLlon Lo Lhe elghL Mlllennlum uevelopmenL Coals (MuCs). 1he Lool ls deslgned for medlum
and large slzed companles wlLh dlrecL operaLlons ln developlng counLrles, wlLh more aLLenLlon Lo poslLlve
Lhan negaLlve conLrlbuLlons, Lhough boLh can be counLed
v
. So far 34 companles such as Pelneken and
hlllps have reglsLered for Lhe onllne self assessmenL Lool LhaL, based on key daLa, esLlmaLes Lhe MuC
fooLprlnLs of a company by measurlng Lhe effecL of Lhe followlng facLors on a counLry's economy: voloe
oJJeJ, employmeot cteotloo, ptoJocts & setvlce ooJ commoolty lovestmeots.

AnoLher assessmenL Lool, [usL emerglng, ls Lhe framework ASl8L (A SusLalnablllLy overLy and
lnfrasLrucLure 8ouLlne for LvaluaLlon) developed by Arup and Lnglneers AgalnsL overLy Lo asslsL pro[ecL
sLakeholders ln undersLandlng Lhe relaLlonshlps beLween susLalnable developmenL, poverLy allevlaLlon and
lnfrasLrucLure ln developlng counLrles. ASl8L ls adapLed from Arup's ln-house appralsal Lool SeA8
(SusLalnable ro[ecL Appralsal 8ouLlne)
vl
, by addlng more poverLy focus Lo Lhe range of susLalnablllLy lssues.
SeA8 was developed as a Lool LhaL consLrucLlon flrms and oLher companles can use Lo assess Lhelr
performance agalnsL 140 or so lndlcaLors LhaL relaLed Lo envlronmenLal proLecLlon, soclal equallLy,
economlc vlLallLy and efflclenL use of naLural resources. uue Lo Lhe greaLer poverLy focus of ASl8L,
'lnsLlLuLlonal sLrucLures' was added as a dlmenslon Lo Lhe framework. 8y provldlng a snapshoL of currenL
performance, lL can help gulde managemenL aLLenLlon Lo areas for lmprovemenL, and provlde a basls for
comparlng resulLs over Llme or across slLes. 1he sofLware Lool ls deslgned so LhaL lL can be fllled ln relaLlvely
rapldly by a well lnformed lnLernal manager.

lurLher lnformaLlon on Lhe scorecards used by developmenL flnance lnsLlLuLlons ls below. 1hese Lend Lo
have one or Lwo secLlons on 'developmenL lmpacL' whlch ls comblned wlLh scores on flnanclal and
someLlmes envlronmenLal performance and donor lnpuL, ln Lhe scorecard.

!"##$%&'
Scorecards and lndlcaLors can be used Lo reporL regularly on performance aL corporaLe level, or aL a pro[ecL
slLe. 1hey are good for:
CapLurlng flxed lnformaLlon whlch can be repeaLed over Llme and compared across respondenLs.

C8l and un Clobal CompacL lndlcaLors:
Are used by many large companles
1o reporL publlcally and LransparenLly on a wlde range of lssues of corporaLe performance, noL
speclflcally developmenL lmpacL.

AlLernaLlve scorecard approaches could be used Lo reporL on a comblnaLlon of quanLlLaLlve lmpacLs ([obs,
Lax paymenL eLc), and more quallLaLlve lmpacLs and conLrlbuLlon Lo sLrucLural change (lnsLlLuLlonal
developmenL, SMML developmenL). 1hey are useful for:
1rylng Lo capLure wlder developmenL lssues ln a sLandardlsed way. SLandardlsaLlon enables
dlmlnlshlng cosLs of repeaL assessmenLs, and comparlson beLween Llme perlods, slLes, or
respondenLs.
1he scorecard approach enables numerlcal lndlcaLors Lo be comblned wlLh oLher lmporLanL
lndlcaLors of lmpacL, and can vlsually be presenLed as such.
LlmlLaLlons:
lL ls dlfflculL Lo esLabllsh flxed lndlcaLors LhaL can capLure developmenL lmpacL, are appllcable ln
wldely dlfferlng clrcumsLances, and can generaLe scores LhaL can reasonably be compared.
Scores provlde a sLaLlc snapshoL. Whlle hlghllghLlng areas of weakness, Lhey do noL necessarlly
explaln sLrong and poor performance, so addlLlonal lnslghLs are needed for lmproved declslon-
maklng.



7>

6.3 Add|t|ona| perspect|ves from ver|f|cat|on and assurance
verlflcaLlon ls noL sLrlcLly lmpacL assessmenL as lLs prlmary purpose ls Lo verlfy LhaL Lhe company ls dolng
whaL lL says lL ls dolng. CfLen lL ls commlssloned Lo check LhaL whaL ls reporLed ln C8l lndlcaLors and annual
revlews ls valld. Powever, verlflcaLlon wlLh hlgh added-value may well ln facL go lnLo some deLall on
lmpacLs LhaL are belng generaLed, and should provlde lnslghLs for fuLure managemenL. lndeed Lhe ma[orlLy
of senlor buslness managers reporL LhaL Lhey expecL more from verlflcaLlon Lhan [usL a sLamp of approval.
1hey are also looklng for added value ln Lerms of sharpenlng lnLernal performance Lhrough lnLerrogaLlon of
sysLems and ldenLlflcaLlon of lnefflclencles or weaknesses for fuLure lmprovemenL (LnuSdlrecLory 2009).

1he conLrlbuLlon of verlflcaLlon processes ls lmporLanL Lo conslder because lL ls a pracLlce LhaL ls spreadlng.
Accordlng Lo SusLalnAblllLy (2002) 68 of Lhe Lop 30 reporLs had some form of exLernal assurance, up from
30 ln 2000. Mlnlng, oll, gas, and pharmaceuLlcals have hlgh raLes, whlle verlflcaLlon ls lncreaslng ln Lhe
flnanclal secLor. More recenL surveys of Lhelr Clobal 8eporLers research programme conflrm Lhe growlng
lmporLance of exLernal assurance.

Some verlflcaLlon ls sLrlcLly procedural. 1he ma[orlLy ls done by accounLancy flrms (wC, uelolLLe, kMC
eLc), and lnLernaLlonal sLandards on verlflcaLlon cover lssues such as sampllng, compleLeness, accuracy and
responslveness. Powever, of 34 verlfled reporLs revlewed by SusLalnAblllLy (2002), 16 were verlfled by
speclallsL consulLancles focuslng on envlronmenLal or soclal lmpacLs, and 6 were by nCCs. 1hese more
speclallsL approaches Lo verlflcaLlon can help explore and lnLerrogaLe Lhe lmpacLs of a company.
verlflcaLlon ls noL deflned by any speclflc meLhod, buL by Lhe purpose for whlch lL ls commlssloned. Whlle
verlflcaLlon ls mosL ofLen relaLed Lo reporLlng agalnsL C8l lndlcaLors, lL can be applled Lo any Lype of lmpacL
assessmenL uslng any of approaches 1 Lo 4.


7. Summary of pros and cons
8ox 3 summarlses Lhe sLrengLhs and weaknesses of Lhe four approaches.

Approach Strengths Weaknesses
1) Loca| assessment: L|ve||hood
|mpact & stakeho|der v|ews
-Comblnes flnanclal & soclal
lmpacL, comblnes reasons wlLhln
resulLs
-'Cround-LruLhlng' wlLh
sLakeholders
-Some aggregaLlon & comparlson
aL hlgher level
-useful for
- 'lmprovlng' declslons
locally
- 'provlng' vla daLa &
sLorles
-noL dlfflculL
-usually weak on aggregaLlon,
benchmarklng or comparlson
-lsolaLed from wlder conLexL
-LlmlLed scale
-8equlres fleldwork on-slLe

2) Va|ue cha|n footpr|nt|ng -Comblnes flnanclal & soclal daLa,
sLakeholder & corporaLe daLa,
local resulLs wlLh Lhe blg plcLure
-locus on core buslness, & supply
& dlsLrlbuLlon chalns
-useful for 'lmprovlng' aL local &
-lnformaLlon requlremenLs -
from sLakeholders/flrms
plus lnLernal corporaLe lnfo
-8equlres sLrong analysls



7B

sLraLeglc level
-Lncourages laLeral Lhlnklng
-8lg plcLure resulLs useful for
'provlng'
3) Lconom|c contr|but|on and
mu|t|p||ers
-Pard daLa, beyond anecdoLe
-8lg plcLure ln naLlonal conLexL
-useful for 'provlng'
-Speaks Lo MlnlsLry of llnance
-locus on normal buslness - noL
[usL CS8 - and supply/dlsLrlbuLlon
chalns
-Scaleable, repllcable, comparable
-LlLLle on sLakeholder vlews,
oLher poverLy lmpacLs
-Less dlsaggregaLlon
or explanaLlon of 'why'

4) Scorecard performance -Slmpllfles complexlLy
-Comblnes flnanclal, soclal &
envlronmenLal
-CfLen lncludes dynamlc &
lnsLlLuLlonal lmpacLs
-vlsual lmpacL
-Makes comparlsons posslble
-1racks changes over Llme
-lL all depends on whaL ls
measured
-uescrlbes noL explalns
-SeLLlng sLandardlsed
developmenL lmpacL lndlcaLors ls
hard


So lf Lhe quesLlon ls 'whlch approach ls besL Lo use" Lhe answer ls always 'lL depends whaL you need lL for.

lL depends on whaL you need:
lnslghLs Lo boosL developmenL lmpacL: aL slLe or counLry level, wlLhln CS8 or across Lhe buslness,
sLraLeglcally?
Lvldence Lo communlcaLe - Lo communlLles, Lhe publlc, Lo llnance MlnlsLers? ersonal sLorles,
numbers or scores?
1ypes of lnformaLlon: flnanclal, soclal, envlronmenLal, or dynamlc lmpacL? CausallLy or descrlpLlon?
uaLa or vlews?
Comparlsons over Llme, wlLh oLhers, or a slngle snapshoL?
WhaL ls manageable and value-for-money?
uepends on your skllls, Llme, mandaLe, & parLners
So, Lhe cholce of approach depends on Lhe buslness case for assessmenL and how Lhe flndlngs wlll be used.
8. Ins|ghts from donors and econom|sts
Many bllaLeral donors and developmenL banks lnvesL ln prlvaLe secLor developmenL. As users of publlc
money Lhey face sLronger pressures Lo reporL on Lhe lmpacL and value for money of Lhelr lnvesLmenL, buL
face slmllar challenges on how Lo measure resulLs of prlvaLe buslness acLlvlLy. Clven Lhe exLenslve Lhlnklng
LhaL Lhey have already had Lo do, lL ls useful Lo draw on donor experlence ln relaLlon Lo:
1he sLandard challenges of conducLlng lmpacL assessmenL, such as whaL and how Lo measure, and
Lhe helghLened challenges relaLlng Lo prlvaLe secLor lmpacLs,
undersLandlng reslsLance Lo conducLlng lmpacL assessmenL,
Assesslng and mapplng chalns of causaLlon,
Assesslng cosL-effecLlveness and comparlng efflclency of dlfferenL lnLervenLlons
CapLurlng dynamlc lmpacLs and knock-on effecLs on developmenL, of prlvaLe secLor acLlvlLy

1hls ls noL Lhe place for a more ln-depLh revlew of Lhe many sLraLeglc and meLhodologlcal lssues covered ln
an exLenslve llLeraLure on donor lmpacL assessmenL. key lssues for prlvaLe secLor developmenL are well
summarlsed by Lhe uonor CommlLLee on LnLerprlse uevelopmenL ln 1he 2008 8eader on rlvaLe SecLor
uevelopmenL: Measurlng and 8eporLlng 8esulLs" (1anburn 2008).


7?

1bete ote loJeeJ sttooq Jlsloceotlves to
cooJoct lmpoct ossessmeots. 1bete ls tbe ftee
tlJet ptoblem (wby sboolJ l ossess my ptoqtom,
let so-ooJ-so ossess tbelts?), o pobllc qooJs
ptoblem (wby sboolJ l lovest my mooey lo
lmpoct ossessmeot wbeo otbets beoeflt ftom lt,
let someooe else lovest lo lt), o coteet ptoblem
(wbot lf l Jo oo ossessmeot ooJ tbe tesolts ote
oot fovootoble, bow wlll tbot offect my
coteet?), o cost ptoblem (qee, lmpoct
ossessmeot sote costs o lot, bettet to sove tbe
mooey ooJ cboooel lt to ptoqtom
beoeflclotles), ooJ so fottb. Cary Woller, ln a
recenL uSAlu dlscusslon forum (1anburn 2008,
pp. 24)

8.1 Cha||enges of |mpact assessment |n pr|vate sector deve|opment
8ox 3 hlghllghLs some sLandard challenges of lmpacL assessmenL ln any arena. 1he 2008 8eader also
hlghllghLs some pottlcolot cbolleoqes of applylng lL ln prlvaLe secLor developmenL:
1he goa|s of pr|vate sector deve|opment are amb|t|ous and mu|t|-|ayered. 1he alms lnclude
sysLemlc change, noL [usL operaLlonal change, generaLlng 'copycaL' affecLs, and second round
dynamlc effecLs (furLher enLerprlse growLh). As such, Lhe lnLended lmpacLs of Su are among Lhe
hardesL Lo quanLlfy.
1he |eve| at wh|ch |mpact shou|d be measured |s debatab|e: CrowLh ln operaLlons of enLerprlses
and economlc acLlvlLy? !obs creaLed and susLalned? Cr changes ln poverLy levels and llvellhood
lndlcaLors among poor households?
1he t|me-frame of assessment matters more Lhan wlLh sLandard dellvery of developmenL servlces
(eg healLhcare), as lnvesLmenL of lnpuLs ls hlgh aL flrsL, whlle generaLlon of lmpacL sLarL small and
rlses lncremenLally over Llme. 1he sLrongesL affecLs may
occur afLer a decade, long afLer lnLervenLlon has ceased.
ShorL Lerm acLlons, such as raLlonallsaLlon, may show
negaLlve lmpacLs, even Lhough Lhey lay Lhe basls for longer
Lerm producLlvlLy lmprovemenLs.
1here |s current|y ||tt|e reward for good |mpact
assessment. Spendlng on assessmenL adds Lo overheads,
whereas Lhe pressure ls Lo reduce overheads relaLlve Lo
fleld lnpuLs.
Current|y success stor|es dom|nate. 1here ls very llLLle
hard lnformaLlon so mlnlmal basellne for maklng
comparlsons. 1he maln comparable lnformaLlon ls cosL per
[ob creaLed, buL even here, deflnlng Lhe scope of 'lnpuL cosL' and of '[ob creaLed' are challenglng.




"6


8ox 6: Cha||enges of |mpact assessment
ln any fleld of developmenL lnvesLmenL, assessmenL of lmpacL has Lo deal wlLh Lhese challenges:
Attr|but|on of causa||ty: lf Lhe alm ls Lo assess lmpacLs of oo lotetveotloo lL ls lmporLanL Lo know wheLher lL
acLually causes Lhe change LhaL can be observed. LxlsLlng Lrends would have caused some change anyway
(Lhls ls Lhe 'deadwelghL' lmpacL), and oLher lnlLlaLlves may be sLronger causal facLors.
Impact assessment vs mon|tor|ng: MonlLorlng Lends Lo be durlng an lnLervenLlon, done Lo lmprove
managemenL. lmpacL assessmenL can be done ex anLe, based on predlcLlon, Lo help declde wheLher Lo make
an lnvesLmenL, and Lhen should also be done ex posL, Lo prove lmpacL, and dlscover lessons abouL whaL
worked or dld noL.
k|gour and prox|es: 8lgorous sclenLlflc research uses randomlsed conLrolled Lrlals, ln whlch lmpacL on Lhe
LargeL group ls compared wlLh a conLrol group. 1hls rarely belng posslble ln developmenL reallLy, alLernaLlves
lnclude measurlng change over Llme, comparlng ouLcomes wlLh approxlmaLe conLrol groups, measurlng
proxy lndlcaLors over Llme and across groups, and uslng ranklng raLher Lhan flxed scorlng of ouLcomes,
1|me-frame: A good lnLervenLlon wlll conLlnue Lo generaLe spln-off or knock-on lmpacLs afLer Lhe
lnLervenLlon lLself has ceased. AL whaL sLage should lmpacL assessmenL be done, and how many years of spln-
off lmpacL should be wlLhln Lhe scope of assessmenL?
D|sp|acement affects: uld oLhers suffer from Lhe lnLervenLlon? Where beneflLs slmply shlfLed Lo Lhe LargeL
group aL Lhe expense of oLher groups?
Lffect|ve use of resources: 1o [udge wheLher an lnLervenLlon was worLh dolng, we need Lo know noL [usL Lhe
lmpacL buL Lhe lmpacL relaLlve Lo Lhe cosL of lnLervenLlon, l.e. cosL effecLlveness. Could Lhe resources have
been beLLer used elsewhere? 8uL Lhls requlres belng able Lo compare lmpacLs wlLh lnpuLs, and havlng daLa
LhaL can be compared across lnLervenLlons.


8.2 Causa||ty
uonor assessmenLs of lnvesLmenL ln Su focuses on quanLlflable lndlcaLors of economlc acLlvlLy: [obs,
wages, numbers of conLracLs, numbers of small, medlum and mlcro enLerprlses (SMMLs), raLes of buslness
expanslon, volume of Lurnover or lnvesLmenL. ln some cases a dlsLrlbuLlonal focus ls lncorporaLed, Lo
measure whaL share of lmpacLs accrue dlrecLly Lo 'poor' parLlclpanLs, Lhe unskllled, women, and mlnorlLles,
or Lo small and mlcro enLerprlse raLher Lhan any-slzed buslness.

Some pro[ecL assessmenLs are good aL Lraclng Lhe causal chaln by whlch lmpacL occurs aL dlfferenL levels.
An example from kaLalysL, a mulLl-donor programme worklng Lo develop value chalns and servlce markeLs
ln 8angladesh ls shown ln llgure 1. Cne year afLer Lralnlng reLallers Lo provlde beLLer advlce Lo farmers,
around 239,000 farmers had galned. As Lhe causal chaln dlagram shows, lmpacLs were dlrecLly from Lralned
reLallers provldlng lnformaLlon and from copycaL affecLs on oLher reLallers and oLher farmers, boLh of
whom adopLed some of Lhe good pracLlce.


"7


I|gure 1: Lxamp|e of assess|ng the causa| cha|n, kata|yst, 8ang|adesh

Source: 1anburn 2008, p. 38

8.3 Cost effect|veness and compar|sons
lor donors, assesslng cost-effectlveoess ls lmporLanL. Some progress has been made ln assesslng cosL per
[ob creaLed. 8esulLs vary enormously, Lhough some varlaLlon sLems from whaL ls measured, aL whaL polnL
ln Llme, and how.
lndependenL evaluaLlon of Lhe uuLch rogramme for CooperaLlon wlLh Lmerglng MarkeLs (SCM)
ln 2003 found LhaL Lhe average cosLs across dlfferenL secLors per dlrecL [ob creaLed was uS$ 12,600,
buL Lhe range was wlde, from uS$10,700 ln agrlbuslness Lo uS$192,000 ln lnfrasLrucLure (1anburn
2008). (WheLher Lhls means lnfrasLrucLure lnvesLmenL generaLes a poor reLurn ls a dlfferenL
maLLer, glven lLs wlder role ln unblocklng developmenL).
CosLs per [ob are much less lf lndlrecL [obs creaLed by an lnvesLmenL are lncluded. llgures from
uflu's 8uslness Llnkage Challenge lund (8LCl) lmply uS$ 1,300 per dlrecL [ob creaLed or reLalned
buL only uS$ 200 per LoLal [ob creaLed or reLalned (1anburn 2008).
AlLhough hlgh cosLs per [ob creaLed are noL on Lhelr own an lndlcaLlon of low developmenL value,
assessmenL of a pro[ecL LhaL records unusually low cosLs per [ob, parLlcularly lf [obs are creaLed aL


""

scale, ls helpful ln flagglng approaches worLh furLher aLLenLlon. varlous lnlLlaLlves worklng wlLh
SMLs, Lechnology, enLrepreneurs and value chaln developmenL reporL cosL per [ob ln Lhe arena of
uS$100 Lo uS$700, wlLh Lhe besL examples (Mekong 8amboo ConsorLlum and enLrepreneurshlp
Lralnlng ln lndla) aL uS$30 and uS$23 (1anburn 2008).

1hese raLlos are based on Lhe cosL of Jooot lopot lnLo Lhe Su lnLervenLlons LhaL creaLed new [obs.
Powever, ln uslng such raLlos Lo reporL and compare Lhe lmpacL of companles, Lhere are blg quesLlons over
Lhe denomlnaLor Lo use (l.e. whaL cost Lo compare Lo Lhe [obs creaLed):
When Lhere ls a clear budgeL lnpuL, such as wlLhln a soclal lnvesLmenL programme, Lhe raLlo of [obs
creaLed Lo Lhe lnvesLmenL cosL provldes a useful raLlo.
Comparlng Lhe developmenL lmpacLs of a uS$ 10,000 communlLy lnvesLmenL donaLlon versus a 3-
year SMML llnkage programme of Lhe procuremenL deparLmenL would be a useful comparlson Lo
make, and has noL been made. 8uL lL may be harder Lo quanLlfy Lhe pro[ecL lnpuL for Lhe llnkage
programme (because lnpuLs are made by operaLlons sLaff as parL of Lhelr work), whlle lL may be
unfalr Lo assess corporaLe lnvesLmenL (Cl) reLurn ln Lerms of [obs lf oLher soclal lmpacLs are
prlorlLlsed.
!obs creaLed by a buslness can be compared Lo caplLal lnvesLmenL, or annual Lurnover of Lhe
company. 1hese can be useful lndlcaLors, for example lf a company ls seeklng Lo lllusLraLe lLs soclal
lmpacL credenLlals Lo governmenL or funders, or for a MlnlsLry conslderlng alLernaLlve lnvesLmenLs.
Powever, such raLlos should noL be used as Lhe sole lndlcaLor. lnvesLmenLs LhaL make lnLenslve use
of caplLal and new Lechnology may dellver oLher subsLanLlal galns, and help generaLe hlgher-
producLlvlLy growLh.

8.4 Dynam|c |mpacts
8eyond Lhe numerlcs of [obs and SMMLs, donors also Lry Lo capLure Jyoomlc lmpocts. MosL of Lhe
developmenL banks (lnLernaLlonal llnance CorporaLlon (llC), Aslan uevelopmenL 8ank (Au8), Afrlcan
uevelopmenL 8ank Croup (Afu8), ueuLsche lnvesLlLlons und LnLwlcklungsgesellschafL (uLC), Luropean
8ank for 8econsLrucLlon and uevelopmenL (L88u)) now use scorecards Lo assess llkely developmenL lmpacL
ex anLe, when decldlng on resource allocaLlon. llnanclal raLe of reLurn ls one non-negoLlable varlable. 8uL
oLher scorecard elemenLs lnclude Lhe lnvesLmenLs conLrlbuLlon Lo prlvaLe secLor developmenL, such as
developmenL of markeLs, Lhe lnvesLmenL cllmaLe, small enLerprlse, Lechnology Lransfer. As an example, Lhe
framework used by llC, ls shown ln llgure 2, ln whlch flnanclal reLurn ls accompanled by 3 oLher caLegorles
of developmenL lmpacL: economlc performance (galns Lo consumers, governmenL, workers eLc),
envlronmenLal and soclal performance (worklng condlLlons, polluLlon abaLemenL eLc), and prlvaLe secLor
developmenL (markeL developmenL, asslsLance Lo buyers or suppllers). lL ls Lhls laLLer caLegory LhaL helps Lo
capLure some of Lhe dynamlc lmpacLs LhaL are dlfflculL Lo quanLlfy buL reflecL hlgh value.



"#

I|gure 2: IIC Deve|opment Cutcome Iramework

Source: llC

1he Cerman uevelopmenL 8ank, uLC, has a broadly slmllar approach whereby 'developmenL effecLs/
susLalnablllLy' are one of four caLegorles LhaL are scored (Lhe oLher Lhree are long-Lerm proflLablllLy, role of
(value added by) uLC, and reLurn on uLC equlLy.) Agaln, developmenL effecLs lnclude Lhe sLaLlc and more
quanLlLaLlve affecLs (governmenL revenue, employmenL) and dynamlc effecLs such as Lransfer of Lechnology
and know-how, and markeL and sLrucLural effecLs (uLC 2008). 1he L88u focuses on CenLral and LasLern
Lurope and prlorlLlses 'LranslLlon lmpacLs'. Lach lnvesLmenL ls scored for lLs lmpacLs on economlc LranslLlon
aL Lhree levels: Lhe pro[ecL level (e.g. markeL expanslon vla llnkages Lo suppllers and cusLomers),
conLrlbuLlons Lo markeL organlsaLlons, lnsLlLuLlons and pollcles (e.g. pollcles LhaL promoLe markeL funcLlons
and efflclency), and conLrlbuLlons Lo buslness behavlour (e.g. dlsperslon of skllls, lnnovaLlon, and hlgher
sLandards of corporaLe governance), (L88u 2008).

1hree aspecLs of Lhese scorecard approaches are valuable Lo noLe here. llrsLly, Lhey comblne an
assessmenL of Lhe pro[ecL's conLrlbuLlon Lo developmenL wlLh assessmenLs of how Lhe pro[ecL also achleves
a reLurn for Lhe donor lnsLlLuLlon (wheLher reLurn on equlLy or addlLlonallLy of Lhe donor lnpuL). A slmllar
approach assesslng corporaLe reLurn and soclal reLurn could be used by companles. Secondly, Lhe go
beyond measurlng sLaLlc numerlc lndlcaLors of economlc acLlvlLy LhaL are usually covered ln L88, Lo score
lnvesLmenLs on Lhelr dynamlc conLrlbuLlon Lo economlc developmenL and markeL developmenL. 1hls ls
lmporLanL for any assessmenL of Lhe lmpacL of buslness on developmenL, because Lhese spln-off effecLs can
be so slgnlflcanL. 1hlrdly, Lhe scorecards are applled across hundreds of pro[ecLs. Whlle deLalls of scorlng
meLhods can be dlscussed ln deLall (and are), Lhe end resulL ls LhaL Lhese flnance lnsLlLuLlons can now
aggregaLe and compare scores across Lhelr porLfollo. Scores for any pro[ecL can be compared Lo Lhe secLor
average. Craphs compare scores by secLor, by year, by reglon, and also generally hlghllghL lmprovemenLs ln
scores over Llme. lL seems llkely LhaL Lhe scorlng process, parLlcularly as lL ls rolled ouL Lo lnvesLmenL
offlcers for ex anLe assessmenLs prlor Lo lnvesLmenL, help focus mlnds and resulL ln hlgher scores (as
capLured ln Lhe adage LhaL whaL geLs measured geLs done).


"@


I|gure 3: I||ustrat|on of |mproved pro[ect performance aga|nst scorecard assessment of 1rans|t|on Impact |n L8kD

Source: L88u 2008. 8ased on 374 posL-evaluaLed lnvesLmenL operaLlons ln 1996-2007


8.S Lconom|sts' assessment of |mpact and add|t|ona||ty
LconomlsLs LradlLlonally have raLher a dlfferenL way of Lhlnklng abouL Lhe conLrlbuLlon of buslness Lo
economlc welfare. 1o daLe Lhere ls raLher a gulf beLween LradlLlonal economlc Lhlnklng on Lhe value of
enLerprlse and lLs conLrlbuLlon Lo welfare and growLh, on Lhe one hand, and debaLes around prlvaLe secLor
developmenL and developmenL lmpacLs of corporaLes on Lhe oLher. Whlle sLraddllng Lhe dlvlde ls a larger
challenge Lhan can be done here, lL ls aL leasL useful Lo conslder Lhe alLernaLlve perspecLlves LhaL come
from economlcs.

1radlLlonally economlsLs assume, by deflnlLlon, LhaL buslness acLlvlLles are ln llne wlLh soclal welfare:
buslness exlsLs Lo respond Lo consumers' needs and provlde whaL Lhey wanL, whlch ls lndlcaLed by
consumers' wllllngness Lo pay for Lhem. 1hls ls Lhe fundamenLal prlnclple underlylng Adam SmlLh's Lheory
of Lhe lnvlslble Pand, whlch says LhaL an lndlvldual pursulng hls own self-lnLeresL Lends Lo also promoLe Lhe
good of hls communlLy as a whole. 8y pursulng hls own lnLeresL he frequenLly promoLes LhaL of Lhe socleLy
more effecLually Lhan when he really lnLends Lo promoLe lL." (SmlLh 1776, p. 436)

lrom Lhls polnL of vlew, Lhe scole of bosloess octlvlty by lLself ls a measure of lLs beneflclal lmpacL on
welfare. Cne approach Lo measurlng Lhe scale/ welfare lmpacL of buslness ls Lo measure 'value Added'
whlch equals Lhe value of sales made, mlnus Lhe cosL of lnpuLs. 1hls can be done aL a company level - by
measurlng Lhelr lnpuLs and ouLpuLs, or aL Lhe counLry level uslng lnpuL-ouLpuL Lables. 1hls approach was
used for Lhe SouLh Afrlcan sLudles of unllever and SA8Mlller dlscussed ln secLlon 3 earller. Llnks wlLh oLher
secLors accounL for a large parL of Lhelr economlc lmpacL. 8uL from a more LradlLlonal economlcs polnL of
vlew, lL ls noL clear why one company should Lake credlL for Lhe acLlvlLles of anoLher, all flrms are
fundamenLally lnLerdependenL. Why shouldn'L suppllers Lake credlL for Lhe arrlval of a large mulLlnaLlonal
ln Lhe counLry Lhan Lhe oLher way round?



"=

ConvenLlonal economlcs focuses on Lhe scale of economlc acLlvlLy and noL Lhe dlsLrlbuLlon of [obs,
revenues, or oLher lmpacL. ln Lhe pursulL of hlgher value added and hlgher producLlvlLy, Lhere ls no
economlc galn from allocaLlng a [ob Lo a local resldenL over a caplLal clLy worker, or a supply chaln conLracL
Lo an emerglng enLrepreneur over an esLabllshed suppller, unless robusL addlLlonal economlc lmpacLs from
dolng so can be demonsLraLed. ln Lheory, reduced lmporL leakages, reduced pressure on governmenL
budgeLs, hlgher reLurns Lo human caplLal developmenL, or hlgher dynamlc affecLs on oLher enLerprlse could
be lllusLraLed, calculaLed and lncorporaLed. ln pracLlce, such welghLlngs are rare. Lconomlc analysls Lends
Lo assume all welfare galns are equlvalenL, whosoever galns Lhem, and soclal or dlsLrlbuLlonal concerns are
reflecLed ln separaLe addlLlonal measures.

ln economlc analysls, buL noL ln Lhe corporaLe approaches ouLllned above, quesLlon of oJJltlooollty or Lhe
counLer-facLual are key. LconomlsLs Lend Lo assume LhaL lf one company ls noL fllllng Lhe gap ln Lhe markeL,
Lhen anoLher one wlll. 1hls may noL be reallsLlc ln a developmenL conLexL, where many markeLs are
exLremely under-developed, and lL ls qulLe common for companles Lo be Lhe flrsL or only one Lo enLer and
operaLe ln a speclflc markeL. 8uL ln general, measurlng Lhe lmpacL of a speclflc buslness should focus on lLs
addlLlonallLy (compared wlLh whaL would exlsL lf lL wasn'L operaLlng ln Lhe markeL). 1hls means Laklng lnLo
accounL Lhe llkely scope for anoLher compeLlLor Lo enLer Lhe markeL lf lL exlLed. Cr Laklng accounL of Lhe
compeLlLors LhaL have been knocked ouL by lLs presence. AssessmenL of knock-on beneflLs should also Lake
lnLo accounL Lhe counLer-facLual: does Lhe way lL manages lLs supply chalns dlffer from Lhose of lLs
compeLlLors, so generaLlng any addlLlonal value Lhrough lLs parLlcular buslness model?

1he way LhaL economlcs LradlLlonally assesses Lhese knock-on lmpacLs ls also somewhaL dlfferenL Lo Lhe
approaches LhaL have been dlscussed earller ln Lhls paper. Lconomlc growLh Lheory suggesLs speclflc
avenues Lhrough whlch buslness acLlvlLy conLrlbuLes Lo growLh:
1hrough caplLal accumulaLlon (elLher physlcal caplLal, or human caplLal e.g. Lhrough Lralnlng, or on
Lhe [ob learnlng),
1hrough splllovers such as Lechnologlcal lnnovaLlon, or managerlal besL pracLlce whlch can be
copled and adopLed by oLher companles,
1he assumpLlon ls LhaL Lhese affecLs are llkely Lo be more beneflclal ln Lerms of Lhe long run raLe of growLh
and developmenL, Lhan for example, creaLlng [obs whlch beneflL Lhose lndlvlduals only for Lhe perlod ln
whlch Lhey have Lhe [obs - unless of course Lhose [obs also brlng wlLh Lhem on Lhe [ob Lralnlng whlch
develop skllls LhaL are applled ln fuLure. 1he focus of growLh Lheory ls very much on whaL Lhe long run
beneflLs wlll be from Lhe acLlvlLy.

MosL sLudles of Lhese Lypes of effecLs are done aL Lhe macro or secLoral level - Lhey are rarely done aL Lhe
company level. AL Lhe same Llme, none of Lhe sLudles revlewed ln Lhls noLe have Laken Lhls approach Lo
assesslng buslness lmpacL. Measurlng llvellhood lmpacLs, value chaln llnkages, value-added Lo economlc
ouLpuL, or achlevemenL agalnsL scorecards generally does noL capLure Lhese dynamlc affecLs of a buslness's
acLlvlLles for Lhe long run raLe of growLh. Powever, lL would be easy Lo lmaglne ways Lhls could be done aL
Lhe company level e.g. assesslng affecLs on Lhe skllls level of workers by looklng aL Lhelr fuLure wage
earnlng poLenLlal, or looklng aL Lhe exLenL of Lechnologlcal lnnovaLlon LhaL a company has broughL.
Powever, Lhls suggesLs qulLe a dlfferenL approach Lo whaL has been Lrled Lo daLe.



"A

9. Some key |ssues on corporates' deve|opment |mpact agenda

1. A range of approaches: d|vers|ty or d|sarray?
1he above shows LhaL Lhere ls a dlverslLy of meLhods, and purposes for assesslng lmpacL. 1haL ln lLself ls an
lssue for buslnesses: whaL Lo choose, and how Lo choose lL? 1he absence of 'off Lhe shelf' approaches may
lmpede appllcaLlon. Companles need Lo be able Lo choose approaches LhaL are manageable for Lhem and
meeL Lhe purpose aL hand.

2. Assess|ng the bus|ness case for do|ng |mpact assessment
1here has been even less aLLenLlon pald Lo Lhe buslness case for lmpacL assessmenL, Lhan Lo Lhe meLhods
for dolng lL. 8uL wlLhouL a buslness case, companles wlll noL lnvesL resources. 1he naLure of Lhe buslness
case wlll deLermlne whaL lmpacL assessmenL ls done and how. So Lhls area needs aLLenLlon.
uemoosttotloq qooJ cltlzeosblp Lo Lhe publlc ls one buslness case whlch does noL requlre very
demandlng lmpacL assessmenL. Some good headllnes and sLorles wlll generally do, and can be
relaLed [usL Lo communlLy lnvesLmenL lnlLlaLlve, noL Lhe core buslness.
5obstootlotloq clolms abouL Lhe role of Lhe company or lndusLry ln developmenL - parLlcularly ln
Lhe face of crlLlclsms of Lhe buslness model - generally requlres someLhlng more. 1hls buslness
case ls more llkely Lo requlre assessmenL of Lhe lmpacL of Lhe buslness (approaches 2, 3, or 4) noL
[usL of CS8 pro[ecLs.
uemoosttotloq Jellvety noL [usL commlLmenL. Clven Lhe pleLhora of commlLmenLs LhaL have been
made Lo Lhe MuCs and Lo developmenL, companles are now reachlng Lhe phase where Lhey need
Lo demonsLraLe resulLs - Lo local parLners, governmenLs, lnLernaLlonal parLners and Lhe publlc.
ldeally Lhls requlres assessmenL of lmpacL as compared Lo lnLenLlon, whlch ls somewhaL more
demandlng LhaL demonsLraLlng good clLlzenshlp.
Mokloq moooqemeot Jeclsloos abouL commerclal sLraLegles LhaL wlll boosL developmenL lmpacL ls
an emerglng buslness case. Companles are currenLly lnnovaLlng ln Lhls fleld (wlLhln Lhe 8uslness
Call Lo AcLlon, harnesslng supply chalns, reachlng boLLom of Lhe pyramld consumers eLc) buL
generally ln Lhe splrlL of lnnovaLlon raLher Lhan based on ex anLe assessmenL of Lhe hlghesL llkely
reLurns Lo elLher developmenL or Lhemselves. ln Lhe nexL phase, ln whlch lmplemenLaLlon spreads
more wldely, execuLlves are llkely Lo wanL more deLalled lnformaLlon.
Motcbloq ot beotloq competltots. 1he buslness case for parLlclpaLlng ln lnlLlaLlves such as C8l ls
parLly noL Lo be seen Lo be lefL ouL. Companles need Lo parLlclpaLe ln eLhlcal lnlLlaLlves LhaL
compeLlLors engage ln, and need Lo score well. lf oLher meLhods develop LhaL generaLe benchmarks
LhaL can be compared across flrms (such as value chalns, mulLlpllers, or good for developmenL
labelllng) Lhls would also be a relevanL buslness case.
3. Impacts of bus|ness or |mpacts of bus|ness |n|t|at|ves
ln Lhe pasL, Lhe maln emphasls has been on assesslng Lhe lmpacL of soclal lnvesLmenL and communlLy
lnvesLmenL. 1hls ls shlfLlng for Lwo reasons. llrsLly, companles are lncreaslngly uslng Lhe levers of core
buslness, noL [usL CS8, for boosLlng developmenL value. Whlle laudlng lnnovaLlon and lmpacL ln Lhls
relaLlvely new arena, Lhey and Lhelr parLners wlll soon need beLLer lnformaLlon for fuLure resource
allocaLlon. Secondly, Lhe wlder debaLe ls abouL Lhe role of buslness ln socleLy, and buslness models LhaL


">

lnnovaLe for Lhe fuLure, ln whlch Lhe argumenL cannoL be won [usL ln Lerms of CS8. 1hls shlfL ln Lurn wlll
requlre more aLLenLlon Lo approaches LhaL map value chaln, fooLprlnLs, and value added, and less Lo mlcro
level pro[ecL ouLcomes.

1hls shlfL parallels Lhe shlfL from measurlng ouLcomes noL ouLpuLs. As emphasls shlfLs from measurlng
numbers of vacclnes and workshops Lo ouLcomes ln Lerms of poverLy reducLlon, llvellhoods supporLed and
growLh enhanced, Lhe challenge wlll geL greaLer, forclng meLhodologlcal developmenL on proxles and
shorLcuLs.

4. Assess|ng both |mpacts on deve|opment and |mpacts on the company
lL could be assumed LhaL companles would aL leasL assess Lhe buslness beneflLs Lo Lhemselves of Lhe
lnvesLmenLs Lhey make LhaL supporL poverLy reducLlon, even lf Lhey don'L need Lo acLually assess Lhelr
lmpacL on poverLy. ln facL, very llLLle ls done. lL seems LhaL corporaLe resources are ofLen lnvesLed under
rules LhaL would noL apply Lo any oLher parL of Lhe buslness: spend lL, do lL, and assume Lhe reLurns are
Lhere lf noL measurable. Powever, Lhls ls beglnnlng Lo change, and lL ls llkely LhaL successful assessmenL
approaches ln fuLure wlll flnd ways Lo comblne reporLlng on reLurns Lo shareholders and sLakeholders.
8lo1lnLoAlcan (2009) has developed a Lool, ln con[ucLlon wlLh llC and uelolLLe, LhaL measures Lhe neL
presenL value of susLalnablllLy lnvesLmenLs. nv for Lhe company ls based on deLalled assessmenL of
lmpacLs aL Lhe pro[ecL-level, such as on workers and resldenLs, and Lhelr response ln Lurn (such as vla
producLlvlLy or reduced hosLlllLy). 1he Lool ls unusual ln LhaL lL allows comparlson of Lhe Lype and Llmlng of
Lhe lmpacLs of alLernaLlve susLalnablllLy lnvesLmenLs. As companles seek more lnformaLlon on Lhe
commerclal reLurns Lo lncluslve buslness approaches, Lhls may help drlve assessmenL of Lhelr lmpacL on
sLakeholders, because one helps Lo dellver Lhe oLher.

S. Sk|||s and mandates, s|mp|e techn|ques and]or comp|ementary partnersh|ps
8uslnesses are noL developmenL organlsaLlons, Lhelr prlmary purpose ls commerclal operaLlon and Lhelr
sklll seL ls noL lmpacL Lechnlques. 1hls ralses quesLlons abouL whaL should be expecLed from buslness ln
Lhelr lnvesLmenL ln lmpacL assessmenL, and whlch skllls and resources should be expecLed from parLners.
8uslnesses dolng Lhelr own lmpacL assessmenL are clearly golng Lo wanL relaLlvely slmply, manageable,
Lechnlques LhaL do noL requlre masslve lnvesLmenL. More complex approaches wlll provlde more rlchly
LexLured and more useful lnslghLs, buL Lhe addlLlonal galn probably needs lnvesLmenL from Lhe wlder
developmenL communlLy.




"B

keferences and Lndnotes

Ashley, C., Paysom, C., oulLney, C., Mcnab, u., Parrls, A. (2003) 'Pow Lo...? 1lps and Lools for SouLh Afrlcan Lourlsm companles on local
procuremenL, producLs and parLnershlps. 8rlef 1: 8oosLlng procuremenL from local buslnesses'. SepLember 2003. London: Cverseas uevelopmenL
lnsLlLuLe. Avallable aL hLLp://www.odl.org.uk/resources/deLalls.asp?ld=1317&LlLle=boosLlng-procuremenL-local-buslnesses

Ashley, C .and MlLchell, !. (2008) 'uolng Lhe rlghL Lhlng approxlmaLely noL Lhe wrong Lhlng preclsely: Lhe challenge of monlLorlng lmpacLs of pro-
poor lnLervenLlons ln Lourlsm value chalns. Cul Worklng aper 291. Cul, llC and Snv. Avallable aL hLLp://www.odl.org.uk/resources/odl-
publlcaLlons/worklng-papers/291-pro-poor-lnLervenLlons-Lourlsm-value-chalns.pdf

8arclays (2007) 'SusLalnablllLy 8evlew 2007' Cnllne reporL avallable aL hLLp://www.barclays.com/susLalnablllLyreporL07/susLalnablllLy_barclays.hLml

8orkenhagen, L. and Zaremba, !. (2009) 'undersLandlng Lhe buslness conLrlbuLlon Lo developmenL: Cxfam's overLy looLprlnL'. 8rleflng for 8uslness
overLy looLprlnL

Clay, !. (2003) 'Lxplorlng Lhe Llnks 8eLween lnLernaLlonal 8uslness and overLy 8educLlon: A Case SLudy of
unllever ln lndonesla', Cxfam ubllcaLlons. Avallable aL hLLp://publlcaLlons.oxfam.org.uk/oxfam/dlsplay.asp?k=9780833983660

Conway. 1. and Sophal, C. (2007) 'Pow does Lhe cholce of poverLy measure shape our undersLandlng of poverLy levels and dynamlcs?' resenLaLlon
Lo Cu8l naLlonal overLy Workshop, hnom enh, november 2007.

uLC (2008) 'CorporaLe ollcy ro[ecL 8aLlng (C8")' LxecuLlve Summary (03/2008), uLC, kfW 8ankengruppe. Avallable aL
hLLp://www.deglnvesL.de/Ln_Pome/AbouL_uLC/Cur_MandaLe/uevelopmenL_ollcy_MandaLe/uLC-C8_brlef-descrlpLlon_engllsh_03-2008.pdf

ulageo (2007) 'LvaluaLlon of Lhe 2007 WaLer of Llfe ro[ecLs ln Afrlca' ulageo. Avallable aL hLLp://www.dlageo.com/n8/rdonlyres/u634A441-12L8-
48Au-8C43-l6A72u906603/0/WaLerofllfeevaluaLlon.pdf
ulageo (2008) 'CorporaLe ClLlzenshlp 8eporL 2008' Avallable aL hLLp://www.dlageo.com/n8/rdonlyres/u42l87CL-2863-40u7-8108-
L8Cu83906396/0/ulACLC_CC8_08.pdf

uWn (2008) 'SusLalnablllLy 8eporL 2008' Avallable aL hLLp://www.dp-dhl.de/susLalnablllLyreporL/2008/servlcepages/welcome.hLml

L88C (2008) 'Annual LvaluaLlon and Cvervlew 8eporL for 2008' Luropean 8ank for 8econsLrucLlon and uevelopmenL, LvaluaLlon ueparLmenL. !une
2008. Avallable aL hLLp://www.ebrd.com/pro[ecLs/eval/showcase/0806-1.pdf

Lllls, k. and keane, !. (2008) Pow do eLhlcal and falr Lrade schemes affecL poor producers? uo we need a new 'Cood for uevelopmenL' label?' Cul
Cplnlon 123, CcLober. London: Cverseas uevelopmenL lnsLlLuLe. Avallable aL hLLp://www.odl.org.uk/resources/odl-publlcaLlons/oplnlons/113-
good-for-developmenL-label.pdf
LnuSdlrecLory (2009) 'CorporaLe susLalnablllLy reporLlng - does verlflcaLlon add value?' Avallable aL
hLLp://www.endsdlrecLory.com/lndex.cfm?acLlon=arLlcles.vlew&arLlclelu=200302

L1l (2003) 'L1l's lmpacL AssessmenL: 1erms of 8eference' LLhlcal 1radlng lnlLlaLlve. Avallable aL hLLp://www.eLhlcalLrade.org/Z/llb/2003/01/la-
Lor/eLl-la-Lor.pdf

kapsLeln, L. (2008) 'Measurlng unllever's economlc fooLprlnL: 1he Case of SouLh Afrlca' unllever. Avallable aL
hLLp://www.buslnessacLlonforafrlca.org/documenLs/unlleverSA8eporL.pdf

kMC (2008) 'lnLernaLlonal Survey of CorporaLe 8esponslblllLy 8eporLlng'. kMC lnLernaLlonal. Avallable aL
hLLps://www.kpmg.com/SlLeCollecLlonuocumenLs/lnLernaLlonal-corporaLe-responslblllLy-survey-2008_v2.pdf

London, 1. (2009) 'Maklng 8eLLer lnvesLmenLs aL Lhe 8ase of Lhe yramld'. Parvard 8uslness 8evlew, May 2009

Massyn, . and koch, L. (2004) 'Afrlcan Came Lodges and 8ural 8eneflL ln 1wo SouLhern Afrlcan CounLrles'. lo 1ootlsm ooJ uevelopmeot lssoes lo
cootempototy 5ootb Aftlco, 8ogerson, C. and vlsser, C. (ed), Afrlca lnsLlLuLe of SouLh Afrlca, reLorla.

MlLchell, ! and Ashley, C. (2009) 'value Chaln analysls and poverLy reducLlon aL scale, Lvldence from Lourlsm ls shlfLlng mlndseLs'. Cul 8rleflng aper
49, March 2009. Avallable aL hLLp://www.odl.org.uk/resources/download/2673.pdf

8ebernak, k. (2008) 'Speclal reporL: 8eporLlng - CommunlcaLlons - Pow Lo measure whaL maLLers' LLhlcal CorporaLlon, avallable aL
hLLp://www.eLhlcalcorp.com/conLenL.asp?ConLenLlu=3771

SA8Mlller (2008) 'LnLerprlse uevelopmenL 8eporL: Maklng a dlfference Lhrough beer'. Avallable aL
hLLp://www.sabmlller.com/flles/reporLs/2008_enLerprlse_developmenL_reporL.pdf
SA8Mlller (2008a) 'Maklng a dlfference Lhrough beer: ConLrlbuLlng Lo Lhe SouLh Afrlcan economy'. Avallable aL
hLLp://www.sabmlller.com/flles/maklngadlfference/Maklng_A_ulfference_SA_Lconomy.pdf

SmlLh, A. (1776) '1he WealLh of naLlons' 8ook lv.ll.6-9, Clasgow LdlLlon

SusLalnAblllLy (2002) '1rusL us: 1he Clobal 8eporLers 2002 Survey of CorporaLe SusLalnablllLy 8eporLlng'. SusLalnAblllLy and unL. Avallable aL
hLLp://www.susLalnablllLy.com/researchandadvocacy/program_arLlcle.asp?ld=438


"?

8lo1lnLoAlcan, llC and uelolLLe (lebruary 2009). lannlng and flnanclal valuaLlon Lool for susLalnablllLy lnvesLmenLs. LlsLed under background
documenLs on hLLp://www.commdev.org/conLenL/calendar/deLall/2381/

1anburn, !. (2008) '1he 2008 8eader on rlvaLe SecLor uevelopmenL: Measurlng and 8eporLlng 8esulLs', lnLernaLlonal 1ralnlng CenLre of Lhe lLC.
Avallable aL hLLp://www.bdsknowledge.org/dyn/bds/docs/deLall/649/4

vodafone (2009) 'lndla: 1he lmpacL of Moblle hones' 1he ollcy aper Serles, number 9, !anuary 2009, vodafone Croup plc. Avallable aL
hLLp://www.vodafone.com/eLc/medlallb/publlc_pollcy_serles.ar.36372.llle.daL/publlc_pollcy_serles_9.pdf


l
1hls and Lhe followlng Lwo quoLes are varlously aLLrlbuLed buL sourced by Lhe auLhor of Lhls paper from 1anburn (2008), Conway and Sophal
(2007), and 8ebernak (2008)
ll
Anglo Amerlcan has made Lhe Loolbox freely avallable as a conLrlbuLlon Lo managlng Lhe soclo-economlc lmpacLs of exLracLlve, oLher naLural
resource and lndusLrlal operaLlons. Avallable aL hLLp://www.angloamerlcan.co.uk/aa/developmenL/socleLy/engagemenL/seaL/seaL_overvlew.pdf

lll
used by boLh Lhe lndusLry body, Lhe World 1ourlsm and 1ravel Councll, and Lhe un body, Lhe un-World 1ourlsm CrganlsaLlon. uesplLe sllghLly
dlfferenL meLhodologles, boLh focus on lnLegraLlng Lhe mulLlpller lmpacLs of a wlde deflnlLlon of Lourlsm acLlvlLy lnLo esLlmaLes of Lhe Lourlsm
economy. 8esulLs on lLs economlc conLrlbuLlon are ofLen four Llmes larger Lhan esLlmaLes from Lhe naLlonal accounLs of Lhe lmpacL of Lhe narrowly-
deflned Lravel and Lourlsm (hoLels and resLauranLs) accounL. lor furLher lnformaLlon see 1ourlsm SaLelllLe AccounLs: 8ecommended
MeLhodologlcal lramework", (CLCu, un, World 1ourlsm CrganlzaLlon): hLLp://browse.oecdbookshop.org/oecd/pdfs/browselL/7801011L.ul
lv
8arclays (8+) and ulageo (A+), uPL ls parL of ueuLsche osL World neL whlch scores 8+. 1he C8l reporLlng framework uses an 'AppllcaLlon Level'
sysLem: A (advanced reporLers), 8 (somewhere beLween) and C (beglnners). 1he reporLlng crlLerla aL each level reflecL a measure of Lhe exLenL of
appllcaLlon or coverage of Lhe C8l 8eporLlng lramework. A plus" (+) ls avallable aL each level (ex., C+, 8+, A+) lf exLernal assurance was uLlllsed for
Lhe reporL. 1hus Lhls appllcaLlon level sysLem ls noL dlrecLly relaLed Lo Lhe acLual performance of a company buL raLher show Lo whaL exLenL Lhe C3
Culdellnes have been uLlllzed ln Lhe reporLs. hLLp://www.globalreporLlng.org/C8l8eporLs/AppllcaLlonLevels/

v
hLLp://www.lpsnews.neL/2007/02/developmenL-mdg-scan-Lo-benchmark-prlvaLe-conLrlbuLlon/
vl
SeA8 conLalns a seL of core secLors and lndlcaLors LhaL have been derlved from Lhe uk CovernmenL's seL of susLalnablllLy lndlcaLors and lLs
susLalnablllLy sLraLegy, unlLed naLlons LnvlronmenL rogramme (unL) lndlcaLors, and Lhe Clobal 8eporLlng lnlLlaLlve (C8l). SeA8 has been
successfully applled on over 100 pro[ecLs worldwlde.

Vous aimerez peut-être aussi