1

8LC81 Cn


"Impacts of Lnv|ronmenta| o||cy on
Internat|ona| 1rade and Cap|ta| Movement"

















8PAv?A 8PA1nACA8
SCC uAvv
2
ACKNOWLEDGEMENT


On the occasion oI completion and submission oI project we would
like to express our deep sense oI gratitude to SOC, DAVV Ior providing us
PlatIorm oI management studies. We thank to our HOD Dr.Triphathy and
Faculty members Ior their moral support during the project.

We are Loo glad Lo glve our speclal Lhanks Lo our pro[ecL gulde Mr
IC CnA1UkVLDI for provldlng us an opporLunlLy Lo carryouL pro[ecL on
currency derlvaLlves and also for Lhelr help and Llps whenever needed
WlLhouL hls cooperaLlon lL was lmposslble Lo reach up Lo Lhls sLage


t last, I sincere regards to my parents and Iriends who have directly
or indirectly helped me in the project.


3
A8S18AC1

An lmporLanL lmpedlmenL Lo Lhe lmplemenLaLlon of envlronmenLal pollcy ls LhaL lL
may lead Lo caplLal fllghL and relocaLlon of flrms CaplLal fllghL and relocaLlon were
ma[or argumenLs launched by opponenLs ln Lhe dlscusslon abouL Lhe lnLroducLlon of
an energy Lax ln Lhe neLherlands 1hey have also played a role ln Lhe dlscusslon abouL
an Lu CC2 Lax Slnce large scale caplLal fllghL and relocaLlon mlghL lead Lo a slLuaLlon
where Lhe remedy ls worse Lhan Lhe dlsease Lhe uuLch energy Lax only applles Lo
consumers and small producers An Lu CC2 Lax ls sLlll under dlscusslon and lL ls hlghly
unllkely LhaL lL wlll be lnLroduced aL all 1he llkellhood of caplLal movemenL and flrm
relocaLlon has undoubLedly been faclllLaLed by Lrade llberallzaLlon and harmonlzaLlon
of regulaLlons and producL norms boLh reglonally (enlargemenL of Lhe Lu and Lhe
creaLlon of nAl1A) and aL Lhe world scale (W1C) lnLernaLlonal mergers openlng up of
new markeLs ln developlng counLrles and LasLern Lurope and new communlcaLlon
Lechnlques such as lnLerneL CaplLal movemenL and relocaLlon behavlour have
recelved qulLe some aLLenLlon ln Lhe envlronmenLal economlcs llLeraLure 8ecause of
daLa scarclLy LheoreLlcal analyses domlnaLe More lmporLanLly Lhese sLudles
frequenLly assume LhaL Lhe only dlfferences beLween counLrles relaLe Lo
envlronmenLal pollcy and lgnore oLher posslble dlfferences or assume Lhem consLanL
over counLrles Such models usually supporL large scale caplLal fllghL and flrm
relocaLlon Lmplrlcal sLudles on Lhe oLher hand usually conslder envlronmenLal pollcy
as one ouL of a seL of locaLlon facLors CLher relevanL facLors are geographlcal locaLlon
Lhe quallLy of Lhe labour markeL Lhe general flscal reglme culLural dlfferences eLc
MosL of Lhese sLudles show LhaL Lhe relaLlve lmporLance of envlronmenLal pollcy ls
usually llmlLed and noL sLrong enough Lo lnduce caplLal fllghL and relocaLlon of flrms
1he purpose of Lhls paper ls Lo presenL an overvlew of Lhe macroeconomlc llLeraLure
relaLlngLo caplLal fllghL and flrm relocaLlon and Lo formulaLe a hypoLhesls whlch can be used
for furLher LheoreLlcal and emplrlcal research Moreover some meLhodologlcal lssues wlll be
dlscussed ln secLlon 2 LheoreLlcal aspecLs wlll be dlscussed A dlsLlncLlon wlll be made
beLween Lrade models models deallng wlLh lnLernaLlonal caplLal movemenLs and models
deallng wlLh boLh ln secLlon 3 emplrlcal sLudles wlll be revlewed Concluslons follow ln secLlon
4 1he relaLlonshlp beLween lnLernaLlonal Lrade and envlronmenL has been on Lhe agenda of
boLh academlclans and pollcy makers ln recenL years ulfferenL aspecLs of Lhls relaLlonshlp
have been lnvesLlgaLed by uLlllzlng dlfferenL models elLher for slngle counLrles or for group of
counLrles 1urkey ls among Lhe developlng counLrles whlch have experlenced Lrade
llberallzaLlon slnce Lhe 1980s 1herefore lL ls lmporLanL Lo flgure ouL Lhe effecLs of Lrade
llberallzaLlon on Lhe envlronmenL for Lhe 1urklsh economy lor Lhls purpose ln Lhls sLudy uslng
lnpuLouLpuL meLhodology 'polluLlon Lerms of Lrade' lndex for CC
2
ls developed 8roadly Lhe
polluLlon Lerms of Lrade lndex measures Lhe polluLlon conLenL of Lhe value of exporLs relaLlve
Lo Lhe polluLlon conLenL of Lhe value of lmporLs ln Lhls way lL could be posslble Lo evaluaLe
Lhe envlronmenLal galns/losses of Lhe counLry from lnLernaLlonal Lrade 1he sLudy ls conducLed
4
for Lhe years 1996 and 1998 where Lhe laLesL lnpuLouLpuL Lables are avallable 1he lnpuL
ouLpuL Lables are aggregaLed Lo 40 secLors AddlLlonally Lhe flndlngs of Lhe sLudy are also
lnLerpreLed ln Lerms of clean and dlrLy secLors of Lhe 1urklsh economy

Introduct|on

lL has been observed LhaL durlng Lhelr lndusLrlallzaLlon course developlng
counLrles are lncllned Lowards lndusLrlal acLlvlLles LhaL are polluLlon lnLenslve ln whlch
Lhey do noL LradlLlonally have comparaLlve advanLage lL has been argued LhaL
lncreaslng producLlon cosLs of dlrLy lndusLrles ln developed counLrles due Lo lncreased
demand for clean envlronmenL from consumers and lncreased envlronmenLal
regulaLlons ln developed counLrles on Lhe one hand and lax envlronmenLal regulaLlons
and envlronmenLally less concerned consumers ln Lhe developlng counLrles on Lhe
oLher hand cause dlrLy lndusLrles Lo mlgraLe from developed Lo developlng counLrles
1hls socalled polluLlon haven hypoLhesls" ln Lhe llLeraLure argues LhaL dlrLy lndusLrles
flee from envlronmenLally sLrlcL lndusLrlallzed counLrles Lo Lhe less developed
economles whlch provlde polluLlon havens for Lhese lndusLrles wlLh Lhelr lax
envlronmenLal sLandards Accordlng Lo Lhls hypoLhesls boLh Lhe lndusLrlal producLlon
sLrucLure and Lrade paLLerns of counLrles are affecLed 1he share of dlrLy lndusLrles ls
expecLed Lo lncrease whlle LhaL of clean lndusLrles Lo decllne over Llme ln polluLlon
havens Also slnce Lhe polluLlon havens are becomlng larger producers of Lhe dlrLy
5
lndusLrles Lhe share of dlrLy lndusLrles ls expecLed Lo lncrease ln Lhe exporLs of a
polluLlon haven
In the literature there are numerous studies examining the role oI dirty industries
in trade patterns oI diIIerent countries. Copeland and Taylor (2004), Huang and Labys
(2002) and JaIIe 09, (1995) are some oI the studies that present detailed literature
surveys on the topic. There are two studies that investigate the pollution haven
hypothesis Ior Turkey (kbostanci 09, (2005) and kbostanci 09, (2007)).
kbostanci 09, (2005) primarily aim at determining the dirty industries oI Turkish
manuIacturing industry. For this purpose Iirst by using the available waste statistics oI
the manuIacturing industry a series oI pollution indices are developed and dirty and
clean industries oI Turkish manuIacturing sector are established. Later the shares oI
dirty industries in total production, employment and trade are analyzed and historical
developments oI these variables are examined. Finally the case oI Turkey as a pollution
haven is discussed and no striking evidence is Iound to support the argument that
Turkey is a pollution haven. It is observed that especially the shares oI dirty industries
in production have not increased and while their shares in exports increased during
1980-1990 they decreased during the post-1990 period. But another important Iinding
is that during post-1980 period in both dirty and clean sectors increases in exports are
observed and the composition oI imports has been developed as predicted by the
pollution haven hypothesis. In kbostanci 09, (2007) pollution haven argument Ior
Turkey Irom trade perspective is examined. Using available data on Turkish
manuIacturing industry at 4-digit ISIC detail, the impact oI dirty industries on the
exports oI Turkey by using a panel oI 67 sectors Ior 1994-1997 period is analyzed. The
general Iormat oI the model estimated is basically the same as the export demand
6
Iunctions that could be seen in the literature with some modiIication to account Ior the
environmental impact. It is concluded that during the period considered, in the Turkish
manuIacturing sector as the dirtiness oI the industries increases the demand Ior exports
increases as well, which can be taken as an evidence Ior the trade eIIect oI the pollution
haven hypothesis.
1he concepL of 'polluLlon Lerms of Lrade' ls lnlLlally lnLroduced by AnLweller
(1996) 8roadly Lhe polluLlon Lerms of Lrade lndex measures Lhe polluLlon conLenL of
Lhe value of exporLs relaLlve Lo Lhe polluLlon conLenL of Lhe value of lmporLs lL Lrles Lo
achleve Lhls by measurlng Lhe emlsslons of maln greenhouse gases ln each secLor of
Lhe economy and Lhen by uslng Lhe assumpLlon LhaL Lhe Lechnologles used are Lhe
same ln dlfferenL counLrles ln a speclflc secLor lL compares Lhe polluLlon conLenL of
lmporLs and exporLs by Lhe help of Lhe polluLlon Lerms of Lrade lndex 1he auLhor
clalms LhaL Lhe counLrles should Lry Lo reduce Lhelr polluLlon Lerms of Lrade lndex ln
Lerms of Lhelr envlronmenLal galns from lnLernaLlonal Lrade ln hls emplrlcal sLudy
lncludlng 164 counLrles for 1987 lL ls found LhaL exporLs of hlghly lndusLrlallzed
counLrles are less envlronmenLally clean Lhan Lhelr lmporLs whlle Lhe opposlLe holds
for developlng counLrles As menLloned ln Mukhopadhyay and ChakraborLy (2003a)
Lhe consumpLlon ln each counLry ls llnked Lo green house gas emlsslons ln oLher
counLrles Lhrough lnLernaLlonal Lrade 8uL ln emlsslon calculaLlons for lndlvldual
counLrles Lhe focus ls on naLlonal emlsslons 1he greenhouse gases embodled ln
lnLernaLlonal Lrade are neglecLed ln Lhls respecL 1un 1uruLA;ik and AkbosLanci
(2007) lnvesLlgaLe Lhe dlsLlncLlon beLween 'CC
2
emlsslons' as a resulL of producLlon Lo
saLlsfy boLh domesLlc flnal demand LogeLher wlLh exporL demand and 'CC
2

7
responslblllLy' LhaL lncludes CC
2
emlLLed durlng Lhe producLlon of lmporLed goods and
Lhelr componenLs for Lhe 1urklsh economy Mukhopadhyay and ChakraborLy (2003a
2003b) consLrucLed an lndex of polluLlon Lerms of Lrade for lndla for Lhe conLenL of
CC
2
SC
2
and nC
x
1helr resulLs lndlcaLe LhaL lndla produces goods LhaL are more
envlronmenLally frlendly Lhan goods lL lmporLs 1herefore lL does noL supporL Lhe
polluLlon haven hypoLhesls Cn Lhe oLher hand lnvesLlgaLlng Lhe Lrade beLween
1halland and CLCu counLrles for Lhe perlod 19802000 Mukhopadhyay (2006) flnds
LhaL for polluLanLs CC
2
SC
2
and nC
x
Lhe polluLlon Lerms of Lrade reveal an lncreaslng
Lrend durlng 19802000 1he values of Lhe lndex were below 100 durlng 1980s and
1990s buL lL was above 100 ln 2000 1hls flndlng supporLs Lhe polluLlon haven
hypoLhesls for 1halland ln 2000 1hus 1halland's Lrade wlLh CLCu counLrles Lurns lL Lo
a polluLlon haven durlng Lhls perlod 19802000 uue Lo Lhe mlxed resulLs found ln Lhe
prevlous Lwo sLudles regardlng Lhe valldlLy of polluLlon haven hypoLhesls for Lhe
1urklsh manufacLurlng lndusLry we Lry Lo flnd ouL furLher evldence for polluLlon haven
hypoLhesls by uslng lnpuLouLpuL meLhodology Lo esLlmaLe 'polluLlon Lerms of Lrade'
lndex ln Lhls sLudy We develop a polluLlon Lerms of Lrade lndex ln Lerms of CC2 for
Lhe 1urklsh economy for 1996 and 1998 ln Lhls sLudy we Lry Lo deLermlne wheLher
1urkey ls a polluLlon haven
1he sLudy ls composed of flve secLlons lollowlng Lhe lnLroducLlon ln Lhe second
secLlon Lhe model developed ln Lhe sLudy ls lnLroduced 1he daLa seL used for Lhe
emplrlcal appllcaLlon of Lhe model ls lnLroduced ln Lhe Lhlrd secLlon 1he emplrlcal
flndlngs are dlscussed ln Lhe fourLh secLlon As usual Lhe lasL secLlon concludes Lhe
sLudy 1he noLlon of efflclency has recelved a loL of aLLenLlon ln envlronmenLal
8
economlcs 8esearch has prlmarlly focussed on a comparlson of pollcy lnsLrumenLs ln
Lerms of cosLs of achlevlng Lhe pollcy LargeL lncludlng Lhe cosLs LhaL Lhe varlous
sub[ecLs of Lhe pollcy lncur as well as cosLs assoclaLed wlLh lmplemenLaLlon and
enforcemenL 1he sLandard resulL ls LhaL economlc lnsLrumenLs le Laxes and Lradable
permlLs are superlor Lo command and conLrol Lfflclency analysls ls usually performed
under Lhe assumpLlon LhaL envlronmenLal pollcy does noL lead Lo relocaLlon of caplLal
and hence Lhe cosLs assoclaLed wlLh relocaLlon are lgnored ln recenL pollcy debaLes
however caplLal fllghL has become a ma[or lssue and forms an lmporLanL lmpedlmenL
Lo Lhe lmplemenLaLlon of envlronmenLal pollcy CaplLal fllghL and relocaLlon of flrms
were ma[or argumenLs launched by opponenLs ln Lhe dlscusslons abouL Lhe
lnLroducLlon of envlronmenLal and energy Laxes ln uenmark Sweden and Lhe
neLherlands1hey have also played a role ln Lhe dlscusslon abouL an Lu CC2 Lax
8ecause of posslble caplLal fllghL Lhe uuLch energy Lax only applles Lo consumers and
small producers An Lu CC2 Lax ls sLlll under dlscusslon and lL ls unllkely LhaL lL wlll be
lnLroduced aL all (Skea 1993)Moreover some of Lhe larger lndusLrles ln uenmark and
Sweden are exempLed from Lhe CC2 Lax because of posslble caplLal fllghL
1he llkellhood of caplLal movemenL and flrm relocaLlon has undoubLedly been faclllLaLed by
Lrade llberallzaLlon and harmonlzaLlon of regulaLlons and producL norms boLh reglonally
(LheenlargemenL of Lhe Lu and Lhe creaLlon of nAl1A) and aL Lhe world scale (W1C) As a
consequence of Lhese developmenLs flrms LhaL relocaLe may sLlll have access Lo Lhe markeL
whlch Lhey have lefL ln Lhls conLexL lnLernaLlonal mergers are also relevanL because Lhey may
make lL posslble Lo shlfL producLlon wlLhouL physlcal relocaLlon of producLlon
faclllLlesMoreover as a consequence of Lhe openlng up of new markeLs ln developlng
counLrles and LasLern Lurope reglons have emerged wlLh envlronmenLal pollcles LhaL are
subsLanLlally laxer Lhan Lhe ones prevalllng ln mosL WesLern counLrles llnally caplLal
movemenL and relocaLlon of flrms have been faclllLaLed by new communlcaLlon Lechnlques
such as Lhe lnLerneL CaplLal movemenL and relocaLlon behavlour have recelved qulLe some
aLLenLlon ln Lhe recenL envlronmenLal economlcs llLeraLure A dlsLlncLlon can be made
beLween macro and mlcro approaches 1he laLLer focuses on locaLlon behavlour of flrms whlle
Lhe former concenLraLes on aggregaLes ln Lhe form of Lrade and lnLernaLlonal caplLal
9
movemenL 1rade and caplLal movemenL are closely relaLed and Lherefore form Lhe maln
sub[ecL of macro sLudles
1hls paper glves a conclse revlew of Lhe llLeraLure deallng wlLh Lhe lmpacLs of
1 envlronmenLal pollcy on caplLal movemenL and lnLernaLlonal Lrade
2 1he organlzaLlon of Lhe remalnder of Lhe paper ls as follows SecLlon 2 deals wlLh
LheoreLlcal analyses of Lhe lmpacLs of envlronmenLal pollcy We ldenLlfy Lhree
LheoreLlcal approaches SLudles based on Lhe PeckscherChlln model sLudles based on
Lhe Lheory of lnLernaLlonal caplLal moblllLy and sLudles LhaL comblne Lhe PC model
and lnLernaLlonal caplLal moblllLy SecLlon 3 examlnes emplrlcal sLudles based on
elLher Lhe PC model or lnLernaLlonal caplLal movemenLs
Concluslons follow ln secLlon 4
2 lmpacLs of envlronmenLal pollcy 1heoreLlcal approaches
Analysls of caplLal fllghL lnduced by envlronmenLal pollcy makes exLenslve use of Lhe
PeckscherChlln (PC) model of lnLernaLlonal Lrade and Lhe Lheory of lnLernaLlonal
caplLal2
movemenLs 1he PC model lmplles LhaL counLrles Lend Lo exporL goods LhaL are
lnLenslve ln Lhe facLors wlLh whlch Lhey are abundanLly supplled Models wlLh
lnLernaLlonal caplLal movemenLs lmply LhaL caplLal wlll flow Lo whaLever counLry LhaL
offers Lhe greaLesL marglnal reLurn 8elow we wlll lnLroduce envlronmenL lnLo Lhese
classlc models
10
lor more Lhan 23 years Lhe lmpacLs of envlronmenLal pollcy on Lrade and caplLal
movemenL have been ln Lhe llmellghL of Lhe llLeraLure on envlronmenLal economlcs
Some lmporLanL references are 8aumol (1971) u'Arge and kneese (1972) WalLer
(1974) and SleberL (19741977 1979 1983) Markusen (1973) Asako (1979) SleberL
eL al (1980) McCulre (1982)
8ubln and Craham (1982) Merrlfleld (1988) kruLllla (1991) 8auscher (1991 1992
19931997) uean (1992) Zaelke eL al (1993) Cameron eL al (1994) Chlchllnlsky
(1994)Copeland (1994) Copeland and 1aylor (1994) (1993) SLelnlnger (1994) Wang
(1993)
8eghln and oLler (1997) ulph (1997) Any of Lhese works has lLs own pecullarlLles and
Lhere ls no easy way Lo summarlze Lhem We conLlnue Lhe sLrucLure lald down ln Lhe
precedlng secLlon We sLarL wlLh a model comblnlng lnLernaLlonal caplLal movemenLs
and envlronmenLal pollcy 1hls Lype of approach ls relaLlvely slmple because
lnLernaLlonal caplLal movemenLs can be dealL wlLh ln a framework wlLh only one good
lnLernaLlonal Lrade models are more compllcaLed and lnclude aL leasL Lwo goods 1he
ma[orlLy of Lhe works have been done ln Lhe second caLegory connecLlng
envlronmenLal pollcy and Lrade lrom a pollcy perspecLlve Lhe mosL lnLeresLlng
models are Lhe Lhlrd caLegory whlch ls a comblnaLlon of Lhe Lwo exLremes lL
ls an unreallsLlc descrlpLlon of reallLy when elLher lmperfecL moblllLy of caplLal or
absence of Lrade ls assumed 1here are however relaLlvely few approaches LhaL
lnclude boLh Lrade and lnLernaLlonal caplLal moblllLy As we wlll show ln Lhe sequel
11
when Lhe resLrlcLlve assumpLlons are relaxed ln Lhe PC model or ln Lhe models of
lnLernaLlonal caplLal movemenLs Lhe resulLs become raLher
amblguous 1hls may glve some explanaLlon for Lhe llmlLed number of sLudles deallng
wlLh Lrade and lnLernaLlonal caplLal movemenLs slmulLaneouslylnLernaLlonal caplLal
movemenLs 8auscher (1997) develops a LwocounLry sLaLlc general equlllbrlum model
wlLh only one commodlLy produced and consumed 1he model ls based on Lhe classlc
Macuougalkemp model (Macuougall 1960 kemp 1964) 1he caplLal sLocks owned by
Lhe clLlzens of Lhe home and forelgn counLry are exogenously glven and flxed CaplLal
ln boLh counLrles ls a composlLe good made up of domesLlc and lmporLed caplLal
lacLors are assumed Lo be lmmoblle across secLors buL lnLernaLlonally moblle Coods
are noL lnLernaLlonally moblle so Lhere wlll be nelLher Lrade nor LransporL acLlvlLles
olluLlon ls caused by domesLlc and forelgn emlsslons and ls sLeamlng from producLlon
as well as consumpLlon lncome ls domesLlc ouLpuL plus Lhe reLurn on caplLal lnvesLed
abroad and mlnus lnLeresL paymenLs made Lo Lhe owners of forelgn caplLal Welfare ls
a funcLlon of consumpLlon and envlronmenLal quallLy whlch ls a funcLlon ofemlsslons
ln Lhls model Lhere ls a dlrecL and an lndlrecL producLlvlLy effecL of emlsslons 1he
dlrecL effecL relaLes Lo emlsslons as a producLlon facLor LhaL lncreases Lhe producLlvlLy
of caplLal 1he lndlrecL or exLernal effecL ls LhaL emlsslons may reduce Lhe producLlvlLy
of caplLal vla producLlon losses 1he facL LhaL Lhe dlrecL and Lhe lndlrecL effecLs have
opposlLe slgns generaLes several amblgulLles 1hls can be lllusLraLed by also
conslderlng 8auscher (1992) of whlch 8auscher (1997) ls an exLenslon ln Lhe sense LhaL
8auscher (1992) does noL have consumpLlon exLernallLles or exLernal effecLs on
producLlon
12
ln boLh models Lhe cenLral quesLlon ls WhaL happens when a governmenL relaxes lLs
envlronmenLal pollcy? 1wo slLuaLlons musL be dlsLlngulshed
(1) Lhe counLry ls small le Lhe marglnal producLlvlLy of caplLal ls flxed vla Lhe world
markeL
(2) Lhe counLry ls large leLhe home counLrys envlronmenLal pollcy has an lmpacL on
Lhe forelgn counLrys marglnal producLlvlLy of caplLal
ln Lhe small counLry case 8auscher (1992) derlves Lhe unamblguous resulL LhaL
lmposlng a resLrlcLlve envlronmenLal pollcy repels moblle caplLal 8auscher (1997)
leads Lo more amblguous resulLs 1he effecLs of a change ln envlronmenLal pollcy on
moblle caplLal conslsL of Lwo effecLs a producLlvlLy effecL and an effecL dependenL
upon consumpLlon exLernallLles lf Lhe sum of Lhe producLlvlLy effecL and Lhe
consumpLlon exLernallLy effecL ls poslLlve (negaLlve) caplLal Lends Lo be aLLracLed
(repelled) lor Lhe overall producLlvlLy effecL Lo be poslLlve Lhe dlrecL poslLlve
producLlvlLy effecL needs Lo be larger Lhan Lhe negaLlve lndlrecL producLlon
effecL ueclslve for Lhe slgn of Lhe consumpLlon exLernallLy effecL ls Lhe slze of Lhe
elasLlclLy of subsLlLuLlon beLween caplLal and emlsslons 1he elasLlclLy of subsLlLuLlon ls
lnversely relaLed Lo Lhe exLernal effecL on Lhe marglnal producLlvlLy of caplLal 1he
smaller Lhe exLernal effecL on producLlon ls Lhe larger ls Lhe subsLlLuLlon effecL and a
large elasLlclLy of subsLlLuLlon glves a poslLlve consumpLlon exLernallLy effecL
1herefore a counLry may elLher aLLracL or repel caplLal by lmposlng a resLrlcLlve
envlronmenLal pollcy 1he overall resulL ls hlghly dependenL on Lhe Lyplcal parameLer
conflguraLlon represenLlng Lhe producLlvlLy effecL and Lhe consumpLlon exLernallLy
13
effecL lL follows LhaL under cerLaln parameLer consLellaLlons Lhe polluLlon heaven
hypoLhesls holds and under oLher parameLer consLellaLlons an envlronmenLal heaven
holds ln Lhe laLLer slLuaLlon counLrles can aLLracL caplLal by lnLroduclng a LlghL
envlronmenLal pollcy1hls may for lnsLance be Lhe case when employees are unwllllng
Lo llve ln polluLlon heavens lf Lhe counLry under conslderaLlon ls large maLLers
become more compllcaLed now we have Lo dlsLlngulsh beLween Lhe slLuaLlon where
Lhe forelgn emlsslons are glven and Lhe slLuaLlon where Lhey are varlable ln Lhe slmple
model wlLhouL consumpLlon exLernallLles and exLernal effecLs on producLlon and
under Lhe assumpLlon LhaL forelgn emlsslons are glven8auscher (1992) shows LhaL Lhe
large counLry can aLLracL caplLal by loosenlng lLs envlronmenLal pollcy 1hls effecL ls Lhe
same as ln Lhe small counLry case Powever lL ls weaker because lL lnfluences boLh Lhe
domesLlc and Lhe forelgn raLes of reLurn on caplLal lf we assume LhaL forelgn
emlsslons are varlable Lhe effecL of lmplemenLlng a lenlenL envlronmenLal pollcy ls also
affecLed by an effecL from Lhe forelgn counLry whlch ls lnduced by a change ln
emlsslons ln Lhe large counLry Also ln Lhls slLuaLlon a counLry can aLLracL caplLal by
lmplemenLlng a more lenlenL envlronmenLal pollcy Powever Lhe resulL ls ampllfled
because Lhe caplLal fllghL from Lhe forelgn counLry reduces Lhe marglnal producLlvlLy
of emlsslons and slnce lL ls now lower Lhan Lhe emlsslon Lax Lhe forelgn flrms wlll
reacL by reduclng Lhelr emlsslonsWhen consumpLlon exLernallLles and exLernal effecLs
on producLlon are lncluded lnLeracLlon effecLs of Lhe consumpLlon exLernallLles and
Lhe polluLlon effecLs on producLlon are generaLed and a greaL varleLy of resulLs may
occur 1o derlve resulLs Lhe analysls ls slmpllfled ln Lwo ways lL ls assumed LhaL Lhe
forelgn emlsslons are elLher glven or LhaL Lhey are varlableSecond lL ls assumed LhaL
14
Lhere ls elLher no consumpLlon exLernallLy or no exLernal effecL on producLlon lf Lhe
forelgn emlsslons are glven and Lhere ls no consumpLlon exLernallLy we geL Lhe resulL
LhaL Lhe effecLs of a change ln envlronmenLal pollcy on moblle caplLal may be poslLlve
or negaLlve As ln Lhe small counLry case lL ls Lhe dlrecL and Lhe lndlrecL producLlon
effecLs LhaL cause Lhe amblgulLles 1he dlrecL effecL of lncreased emlsslons on Lhe
marglnal producLlvlLy ls poslLlve 8uL Lhe lndlrecL effecL ls LhaL lncreased emlsslons
have deLrlmenLal effecLs on boLh domesLlc and forelgn caplLal producLlvlLy lf Lhe
dlrecL effecL domlnaLes Lhe counLry can aLLracL moblle caplLal by lmposlng a lenlenL
envlronmenLal pollcy lf Lhe lndlrecL effecL domlnaLes lL repels caplLal lf Lhere are no
exLernal effecLs on producLlon and forelgn emlsslons are sLlll glven we geL Lhe
unamblguous resulL LhaL a counLry can aLLracL caplLal by lmposlng a lenlenL
envlronmenLal pollcy 1hls resulL ls slmllar Lo Lhe resulLs obLalned ln Lhe more slmple
model ln 8auscher (1992)
lf forelgn emlsslons are varlable and Lhere ls no consumpLlon exLernallLy all effecLs
of a more lenlenL envlronmenLal pollcy on caplLal locaLlon are amblguous 1he
lnLroducLlon of emlsslon Laxes adds furLher complexlLles Lo Lhe model and many of Lhe
effecLs become lndeLermlnaLe under Lhe assumpLlon LhaL Lhere ls no exLernal effecL
on producLlon (buL Lhere are consumpLlon exLernallLles) a counLry can aLLracL caplLal
by lmplemenLlng a more lenlenL envlronmenLal pollcy lncreased domesLlc emlsslons
lncrease Lhe producLlvlLy of caplLal and Lherefore forelgn dlrecL lnvesLmenL ls
lncreased 1he forelgn counLry now employs a smaller caplLal sLock and lf producLlvlLy
ls noL affecLed by Lhe exLernallLles Lhe forelgn counLry wlll have a reducLlon ln Lhe
15
marglnal producLlvlLy of emlsslons WlLh a glven emlsslon Lax Lhe forelgn producers
wlll reduce Lhelr emlsslons
ln some early papers u'Arge and kneese (1972) WalLer (1974) and SleberL (1974)
dlscuss how envlronmenLal pollcy lnfluences Lhe commodlLy prlces and Lhereby Lhe
comparaLlve prlce advanLages SleberL (1974) conslders a LwocounLry LwosecLor
model wlLh only one producLlon facLor (naLural resource) roducLlon generaLes a
polluLanL whlch affecLs Lhe amblenL quallLy of Lhe envlronmenL 1he envlronmenL ls
consldered as a recepLor of polluLanLsAn envlronmenLal proLecLlon agency lmposes an
effluenL charge and maxlmlzes lLs 'proflL' ln Lerms of envlronmenLal quallLy lrom Lhls
maxlmlzaLlon an expresslon for Lhe scarclLy prlce of Lhe envlronmenL ls obLalned
SleberL (1974) flnds a poslLlve relaLlonshlp beLween Lhe scarclLy prlce for Lhe
envlronmenL and Lhe commodlLy prlces lf Lhe Lwo counLrles have dlfferenL
scarclLy prlces for Lhe envlronmenL lL follows LhaL Lhe commodlLy prlces wlll also
dlffer lf counLry 1 has a lower scarclLy prlce for Lhe envlronmenL Lhe envlronmenL ls
relaLlvely abundanL ln LhaL counLry Moreover lf counLry 1 has a relaLlve prlce
advanLage ln commodlLy lL wlll exporL LhaL good 1hls leads Lo Lhe Lheorem saylng LhaL
1he counLry rlchly endowed
wlLh Lhe envlronmenL wlll exporL Lhe commodlLy wlLh a hlgh polluLlon conLenL Slmllar
concluslons are reached ln u'Arge and kneese (1972) and WalLer (1974) ln summary
when envlronmenLal pollcy ls lncluded ln Lhe Lrade models relaLlve commodlLy prlces
are noL only deLermlned by relaLlve marglnal producLlvlLles buL also depend on (1) Lhe
16
marglnal Lendency Lo polluLe (2) Lhe marglnal soclal damage caused by one unlL of
polluLanL
and (3) Lhe LoLal LnvlronmenLal pollcy ls noL expllclLly Laken lnLo accounL ln Lhe
llLeraLure quoLed above
1here are dlfferenL Lypes of envlronmenLal pollcy are
consldered an emlsslon Lax a consumpLlon Lax and a Lrade
Lax A dlsLlncLlon ls made beLween Lhe smallcounLry case and Lhe largecounLry case
ln Lhe smallcounLry case Lhe counLry cannoL affecL Lhe world markeL prlces and
cannoL use lLs envlronmenLal pollcy Lo lmprove lLs Lerms of Lrade 1he analysls ls
Lherefore reduced Lo Lhe way ln whlch envlronmenLal pollcy affecLs Lhe volume of
Lrade and changes ln lmporLs and exporLs Slnce Lhere are LransporL speclflc
exLernallLles ln Lhe model a Lrade Lax becomes an envlronmenLal pollcy lnsLrumenL as
well lL ls assumed LhaL Lhe counLry ls neL exporLer of commodlLy one and neL lmporLer
of commodlLy Lwo and LhaL forelgn polluLlon does noL affecL Lhe domesLlc
envlronmenL ln Lhls model Lhe effecLs of envlronmenLal pollcy on lnLernaLlonal
Lrade are amblguous 1he reason ls agaln Lhe exLernal effecL on producLlon menLloned
ln Lhe precedlng secLlon ln Lhe more slmple verslon of Lhe model where exLernal
effecLs on producLlon are lgnored Lhe amblgulLles are reduced buL noL ellmlnaLed as
Lhe followlng examples show (1) A hlgher emlsslon Lax ln secLor Lwo lncreases Lhe
comparaLlve
17
dlsadvanLage ln Lhe producLlon of Lhls commodlLy 1he counLry wlll Lherefore
speclallze even more ln Lhe producLlon of good one and lmporLs of good Lwo are
lncreased A hlgher emlsslon Lax may Lherefore lncrease Lrade and Lhe polluLlon lmpacL
of LransporLaLlon may domlnaLe Lhe effecL of more sLrlcL envlronmenLal pollcy (2) A
Lax on Lhe consumpLlon of good Lwo reduces Lrade because consumers subsLlLuLe
Lowards good one whlch may lead Lo overall reducLlon of polluLlon boLh vla Lhe
producLlon and LransporLaLlon effecL Pence ln a small open economy Lhe lmpacLs of
changes ln envlronmenLal pollcy are ln general amblguous because Lhe regulaLlon of a
polluLlon acLlvlLy has splllover effecLs Lo oLher polluLlon acLlvlLles
Moreover Lhe effecL may vary along wlLh Lhe pollcy lnsLrumenL ln Lhe largecounLry
case Lhe counLry has an lmpacL on world markeL prlces and can use lLs
envlronmenLal pollcy Lo lmprove lLs Lerms of Lrade lL ls now assumed LhaL Lrade
lnLervenLlons are resLrlcLed for lnsLance because Lhe counLry ls parL of a freeLrade
agreemenL Moreover LransporL exLernallLles are lgnored and conLrary Lhe small
counLry case polluLlon ls assumed Lo be Lransboundary lurLher ln order Lo avold
lndeLermlnaLe resulLs lL ls assumed LhaL forelgn lmporL demand ls very elasLlc 1hls
model leads Lo unamblguous prlce effecLs buL amblguous effecLs on polluLlon lf a Lax
ls levled on Lhe consumpLlon of a good lLs demand wlll be reduced and Lhe world
markeL prlce goes down lf emlsslons ln one secLor are Laxed Lhe supply of Lhe flnal
good ln LhaL secLor ls reduced and Lhe world markeL prlce ls lncreased
18
1he amblguous polluLlon effecL of a change ln envlronmenLal pollcy arlses because
of Lhe Lransboundary exLernallLy whlch generaLes leakage effecLs Leakage effecLs
occur for lnsLance
lf Lhe consumpLlon Lax raLe ls lncreased 1he dlrecL effecL ls LhaL polluLlon decreases
Powever Lhere ls a change ln Lhe Lerms of Lrade and Lherefore Lhe forelgn counLry ls
lnduced Lo lncrease lLs emlsslons 1he leakage effecL may ouLwelgh Lhe dlrecL effecL
from an lncrease ln Lhe consumpLlon Lax 1hus ln Lhe largecounLry case LlghLer
envlronmenLal pollcy ln Lhe exporL lndusLry lmproves Lhe Lerms of Lrade and LlghLer
envlronmenLal pollcy ln Lhe lmporL lndusLry worsen Lhe Lerms of Lrade
A somewhaL dlfferenL approach Lo Lhe analysls of envlronmenLal pollcy and Lrade
Lhan Lhe one ln 8auscher (1997) can be found ln Lhe llLeraLure on Lrade beLween a rlch
reglon (norLh) and a poor reglon (SouLh) (Copeland and ScoLL (1994) Chlchllnlsky
(1994) Copeland and 1aylor (1993) and 8eghln and oLler (1997) 1hese models
assume a sufflclenLly asymmeLrlc dlsLrlbuLlon of human caplLal endowmenLs and
absorpLlve capaclLles beLween norLh and SouLh
1hese endowmenL dlfferences wlll be LranslaLed lnLo lncome dlfferences and lnLo a
lower valuaLlon of Lhe envlronmenL ln Lhe SouLh Lhan ln Lhe norLh 1he resulL ls LhaL
Lhe norLh lmplemenLs a more sLrlngenL envlronmenLal pollcy Lhan Lhe SouLh ln
Copeland and 1aylor7 (1994) Lhe lncome lnduced dlfferences ln polluLlon Laxes are Lhe
sole deLermlnanL of Lrade flows 1he model ls a sLaLlc LwocounLry general equlllbrlum
model wlLh a conLlnuum of goods dlfferlng ln Lhelr polluLlon lnLenslLy of producLlon
olluLlon ls local 1he resulL ls LhaL free Lrade shlfLs polluLlonlnLenslve producLlon Lo
19
Lhe humancaplLal scarce SouLh and ralses world polluLlon neverLheless Lhere ls no
markeL fallure because polluLlon sLays wlLhln Lhe counLry of
orlgln and lL ls assumed LhaL governmenLs regulaLe polluLlon opLlmally ConsequenLly
Lrade always lncreases welfare
ln a subsequenL paper Copeland and 1aylor (1993) exLend Lhelr 1994model and
allow for Lransboundary polluLlon ln order Lo focus on lncome effecLs Lhey have
malnLalned LhaL lncomelnduced dlfferences ln aLLlLudes Lowards Lhe envlronmenL are
Lhe sole moLlve for Lrade
1he resulL ls LhaL lf Lhe dlsLrlbuLlon of lncome ls hlghly skewed Lhen free Lrade harms
Lhe global envlronmenL buL lf Lhe dlsLrlbuLlon of lncome ls even Lhen free Lrade has
no adverse effecL on Lhe envlronmenL 8ecause lowlncome counLrles have a sLraLeglc
advanLage ln seLLlng polluLlon levels ln a free Lrade reglme Lhey have an lncenLlve Lo
delay lnLernaLlonal polluLlon
negoLlaLlons unLll mulLllaLeral Lrade llberallzaLlon has been achleved
Chlchllnlsky (1994) conslders anoLher reason for dlfferences ln envlronmenLal pollcy
beLween norLh and SouLh She conslders Lhe dlfference ln lnsLlLuLlonal capaclLy
beLween norLh and SouLh Lo deflne properLy rlghLs on naLural resources and Lhe
envlronmenL 1he dlfferences ln lnsLlLuLlonal capaclLy provldes a basls for Lrade drlven
by llldeflned properLy rlghLs and a Lragedy of Lhe commons 1he norLh overconsumes
underprlced resourcelnLenslve producLs lmporLed from Lhe SouLh ln Lhls case Lhe
20
lmplemenLaLlon of envlronmenLal Laxes ln Lhe SouLh may be lnapproprlaLe baslcally
because Laxes decrease Lhe prlce of Lhe resource and lead Lo
more exLracLlon roperLy rlghL pollcles may be more effecLlve
1rade and lnLernaLlonal caplLal movemenLs
McCulre (1982) dlscusses whaL happens when regulaLlon ls lncorporaLed lnLo Lhe
sLandard PC model wlLh lnLernaLlonal facLor moblllLy Pe alLers Lhe PC model Lo
accommodaLe Lhe envlronmenL by assumlng LhaL Lhe envlronmenL ls a producLlon
facLor whlch ls used up ln lndusLrlal processes 8elng a producLlon facLor Lhe
envlronmenL wlll be used up Lo Lhe polnL where lLs marglnal producL equals lLs prlce
whlch ln Lhe absence of any regulaLlon ls nll 1he envlronmenL ls a facLor ln only one of
Lhe lndusLrles so Lhere ls a regulaLed and an unregulaLed
lndusLry McCulre furLher assumes LhaL caplLal labour and emlsslons are equally
subsLlLuLable
1hus when envlronmenLal regulaLlon occurs Lhe shape of Lhe producLlon funcLlon ls
noL alLered ln Lhls way envlronmenLal regulaLlon whlch consLralns Lhe allowable level
of polluLlon Lo some glven marglnal producL ls equlvalenL Lo Lhe negaLlve neuLral
Lechnologlcal progress
WlLh respecL Lo Lrade lL ls lmporLanL Lo dlsLlngulsh beLween coordlnaLed an
uncoordlnaLed envlronmenLal regulaLlon under coordlnaLed envlronmenLal regulaLlon
boLh counLrles face Lhe same shadow prlce of polluLlon 1he quallLaLlve lmpllcaLlons of
Lhe PC model sLlll holds and so does Lhe facLor prlce equallzaLlon Powever wlLh
21
uncoordlnaLed envlronmenLal pollcy facLor prlce equallzaLlon ls losL lor a small
counLry faclng glven world commodlLy prlces envlronmenLal pollcy depresses Lhe
world commodlLy prlce of Lhe regulaLed producL and Lhe released facLors have Lo be
absorbed by Lhe unregulaLed lndusLry When lnLernaLlonal facLor movemenLs are
posslble lL ls clear LhaL labour and caplLal wlll have an opporLunlLy Lo relocaLe
accordlng Lo Lhe facLor rewards ln Lhe PC model wlLh ldenLlcal Lechnologles buL
lnLernaLlonal moblle facLors facLor prlce equallzaLlon cannoL be resLored unLll all of
Lhe polluLlng lndusLrles have lefL Lhe counLry wlLh Lhe sLrlcLesL envlronmenLal pollcy
1he regulaLed counLry wlll be8 drlven ouL of Lhe producLlon of Lhe regulaLed good
1hus lnLernaLlonal facLor moblllLy and uncoordlnaLed envlronmenLal regulaLlon lead Lo
compleLe speclallzaLlon ln aL leasL one of Lhe counLrles and perhaps ln boLh
Merrlfleld (1988) develops a model lnsplred by McCulre (1982) 1he general
equlllbrlum model allows commodlLy and lnpuL prlces Lo be deLermlned endogenously
Coods polluLlon and caplLal are lnLernaLlonally moblle Pe conslders Lhe economlc
lmpacL of LlghLer abaLemenL equlpmenL sLandards 1he model conslders Lwo counLrles
(A and 8) whlch produce one composlLe good each (x and x respecLlvely) CounLry A
consumes some of x and exporLs
Lhe balance Lo counLry 8 and vlce versa for counLry 8 1he model lncorporaLes a
producLlon funcLlon ln whlch Lhe scarclLy of caplLal servlces varles dlrecLly wlLh Lhe
polluLlon flow ln order Lo capLure Lhe poslLlve lmpacL polluLlon has on lnpuL
producLlon procedures commodlLles eLc 1hus Lhe producLlon facLors are caplLal
22
labour emlsslons polluLlon abaLemenL equlpmenL and LransnaLlonal polluLlon 1wo
polluLlon abaLemenL sLraLegles are analysed (1)
an equlpmenL sLandard sLraLegy (commandandconLrol approach) and (2) producLlon
Lax sLraLegy (economlc approach) uue Lo lnLernaLlonally moblle goods caplLal and
polluLlon Lhe producLlon Lax has an amblguous lmpacL 1he producLlon Lax may
acLually lncrease polluLlon because Lhe change ln polluLlon flows due Lo a reducLlon ln
Lhe emlsslons of one counLry may be more Lhan offseL lf caplLal movemenLs lncrease
Lhe ouLpuL and emlsslons of Lhe oLher counLry ln oLher words A producLlon Lax ln
counLry A may lnduce Lhe polluLlng lndusLrles Lo move Lo counLry 8 whlch reduces
counLry A's emlsslons by a cerLaln amounL buL lncreases Lhe emlsslons ln counLry 8 by
an even larger amounL 1hls resulL ls slmllar Lo Lhe resulL ln 8auscher (1997) 1he
change ln polluLlon flows whlch resulLs from Lhe producLlon Lax depends on whlch
counLry produces Lhe mosL polluLlonlnLenslve good lf counLry A's ouLpuL
ls relaLlvely more polluLlonlnLenslve Lhan counLry 8's Lhen a producLlon Lax on
counLry 8's ouLpuL wlll resulL ln a neL lncrease ln polluLlon flows because caplLal
movemenL lncreases counLry A's ouLpuL and emlsslons Powever a producLlon Lax on
A's ouLpuL reduces polluLlon flows ln Lhls case Lhe caplLal movemenL from counLry A
(and Lhe lncreased polluLlon ln counLry 8) ls noL enough Lo offseL Lhe emlsslon
reducLlon ln counLry A 1hls amblgulLy ls noL presenL ln Lhe case of an equlpmenL
sLandard
Wang (1993) analyses Lhe lmpacLs of a producLlon polluLlon Lax on envlronmenLal
caplLal
23
fllghL and naLlonal producL ln a LwocounLry sLaLlc general equlllbrlum model wlLh Lwo
way forelgn lnvesLmenL 1here ls only one produced good whlch ls Lradable and
produced uslng labour and caplLal 1he endowmenL of labour ls flxed and
lnLernaLlonally lmmoblle
LndowmenL of caplLal ls flxed buL lnLernaLlonally moblle CaplLal ls a composlLe good
of domesLlc and lmporLed caplLal and Lhe analysls allows for dlfferenL degrees of
subsLlLuLablllLy beLween domesLlc and lmporLed caplLal olluLlon ls assumed Lo
orlglnaLe ln Lhe producLlon process and Lhe producLlvlLy of caplLal can be negaLlvely or
poslLlvely affecLed by domesLlc or forelgn emlsslons roducLlon polluLlon Laxes ln boLh
counLrles are deslgned Lo opLlmally mlLlgaLe envlronmenLal degradaLlon by
reallocaLlng caplLal and naLlonal producL ln Lhls way
Lhe lnvesLmenL declslon ls lnfluenced by Lhe polluLlon Lax because Lhe equlllbrlum
raLes of reLurn on caplLal and naLlonal producLs ln boLh counLrles are affecLed 1he
resulL ls LhaL when a Lax ls levled ln counLry A lL lnlLlally decreases Lhe raLes of reLurn
on lnpuLs and Lhen producLlon CaplLal Lends Lo flee ouL of Lhe counLry (labour ls
lmmoblle) uomesLlc caplLal ouLflows lncrease and forelgn caplLal lnflows decrease lf
elasLlclLles of subsLlLuLlon beLween labour and caplLal are sufflclenLly hlgh ln boLh
counLrles 1hls ls due Lo Lhe facL LhaL for hlgh elasLlclLles of subsLlLuLlon beLween
labour and caplLal counLry A wlll adopL a more labourlnLenslve lnpuL comblnaLlon
whlle counLry 8 wlll adopL a more caplLallnLenslve lnpuL9
comblnaLlon ln producLlon lor lower levels of elasLlclLy of subsLlLuLlon Lhe caplLal
lnLenslve producLlon ls noL replaced by Lhe labourlnLenslve producLlon An addlLlonal
24
resulL of Lhe model ls LhaL negaLlve Lransboundary exLernallLles wlll affecL Lhe Lax
lmpacL on producLlon An lncrease ln Lhe domesLlc polluLlon Lax reduces domesLlc
producLlon and lncreases forelgn producLlon because caplLal has moved Lo Lhe forelgn
counLry negaLlve Lransboundary exLernallLles affecL domesLlc producLlvlLy negaLlvely
and domesLlc producLlon decreases
Pence negaLlve Lransboundary exLernallLles exacerbaLe Lhe decrease of Lhe domesLlc
producL
1hls escalaLes envlronmenLal caplLal fllghL and Lhus enhances Lhe lncrease ln forelgn
producLlon
3 lmpacLs of envlronmenLal pollcy Lmplrlcal sLudles
1he concluslon reached above from Lhe LheoreLlcal dlscusslon was raLher blurred
under some parameLer consLellaLlons sLrlcL envlronmenLal pollcy may lead Lo
envlronmenLal caplLal fllghL
under oLher parameLer consLellaLlons Lhls may noL be Lhe case We now Lurn Lo
emplrlcal sLudles 1hls secLlon conslsLs of Lwo parLs 1he flrsL deals wlLh Lhe effecLs of
envlronmenLal regulaLlon on Lrade ln Lhe second we dlscuss sLudles of lnLernaLlonal
caplLal movemenL and changes ln forelgn dlrecL lnvesLmenL as a response Lo
envlronmenLal pollcy
1rade models
1here are Lwo approaches ln whlch Lo analyse emplrlcally Lhe relaLlonshlp beLween
25
envlronmenLal regulaLlon and Lrade 1he LeonLlef approach and Lhe PC approach
LeonLlef
(1934) measured wheLher Amerlcan exporLs were labour or caplLallnLenslve relaLlve
Lo lmporLs ConLrary Lo whaL was expecLed LeonLlef found LhaL Lhe uS exporLs were
less caplLallnLenslve Lhan lLs lmporLs lf envlronmenLal pollcy ls lnLroduced lnLo
LeonLlef's model one
measures wheLher Lhe exporL of a counLry ls more or less polluLlon lnLenslve Lhan lLs
lmporL
WalLer (1973) uses Lhe LeonLlef approach Lo measure wheLher uS envlronmenLal
pollcy ls 'Lrade neuLral' or exporL or lmporLblased Pe esLlmaLes Lhe 'Cverall
LnvlronmenLal ConLrol Loadlngs' (CLCL) whlch measures Lhe proporLlon of Lhe
producLs flnal value aLLrlbuLable Lo envlronmenLal regulaLlon 1hls measure ls Lhen
mulLlplled by Lhe value of Lhe uS lmporLs and exporLs Lo obLaln Lhe envlronmenLal cosL
componenL of uS Lrade 1he resulL ls LhaL Lhe average annual CLCL of Lhe uS exporL
durlng 19681970 was abouL 173 per cenL of LoLal exporLs
1he uS average annual CLCL of lmporLs ls esLlmaLed Lo be 132 per cenL of LoLal
lmporLs or 13 per cenL lower Lhan Lhe exporLs WalLer concludes LhaL Lhls dlfference
does noL appear Lo be very slgnlflcanL and Lhe resulL suggesLs LhaL envlronmenLal
conLrol measures bearlng on Lhe respecLlve suppllers of uS exporLs and lmporLs wlll be
Lrade neuLral or aL worsL only marglnally blased agalnsL Lhe uS Lradlng parLners
Powever on Lhe level of lndlvldual lndusLrles Lhere may be basls for a loss ln
compeLlLlveness as a consequence of a more sLrlcL uS envlronmenLal pollcy 1he mosL
26
senslLlve lndusLrles Lurned ouL Lo be ordnance and accessorles consLrucLlon and
mlnlng equlpmenL and plasLlcs lL ls dlfflculL Lo say someLhlng abouL dlfferences ln
envlronmenLal pollcy beLween Lhe uS and lLs Lradlng parLners An
lndlcaLor could be Lhe dlfference ln CLCL ln 19681970 Lhe uS exporL Lo !apan had an
CLCL value of 170 whereas lmporLs 130 1he correspondlng values for Canada were
173 and 170 respecLlvely 1hus relaLlve Lo Lhe lmporLs from !apan Lhe uS exporLs Lo
!apan appear Lo be relaLlvely polluLlon lnLenslve whlch ls much less Lhe case for Lhe
uSCanadlan10 Lrade
More recenL daLa for several years were analysed by 8oblson (1988) 8y means of a
parLlal equlllbrlum model slmllar Lo WalLer (1973) 8oblson measured Lhe lmpacL of
marglnal changes
ln lndusLrlal polluLlon abaLemenL cosL on Lhe uS balance of Lrade ln general and Lhe balance of
Lrade wlLh Canada ln parLlcular WalLer (1973) measured prlce changes caused by
envlronmenLal regulaLlon by mulLlplylng Lhe abaLemenL cosL vecLor by Lhe LoLal requlremenLs
maLrlx 8oblson noLes LhaL Lhls approach mlsses Lhe abaLemenL cosLs lmpllclL ln caplLal
goods
used ln Lhe producLlon process 8oblson Lakes Lhls lnLo accounL by addlng a caplLal flow
coefflclenL Lo Lhe LoLal requlremenLs maLrlx 8y means of Lhls model Lhe average
abaLemenL
conLenL of uS Lrade ls esLlmaLed for Lhe years 1973 1977 and 1982 1he raLlo of
abaLemenL
conLenL of lmporLs Lo abaLemenL conLenL of exporLs rose from 1131 ln 1973 Lo 1389
ln 1982
27
8oblson concludes LhaL a more sLrlngenL envlronmenLal pollcy ln Lhe uS has changed Lhe
comparaLlve advanLage ln such a way LhaL more hlghabaLemenLcosL goods have been
lmporLed a nd more lowabaLemenLcosL goods exporLed 1here ls no slgnlflcanL change ln Lhe
abaLemenL conLenL raLlo for uSCanadlan Lrade lmplylng no ma[or shlfL ln Lradlng
paLLerns
beLween Lhese Lwo counLrles
More recenL daLa and more counLrles were analysed by Sorsa (1994) She analysed
Lhe llnks
beLween envlronmenLal expendlLures and Lrade ln envlronmenLally senslLlve goods ln seven
counLrles whlch have hlgh envlronmenLal sLandards AusLrla llnland Cermany !apan
norway Sweden and Lhe uS 1wo approaches were used llrsL she compared world Lrade
shares ln envlronmenLally senslLlve goods ln Lhe perlod 19701990 Second 8evealed
ComparaLlve AdvanLage (8CA) lndexes were calculaLed for Lhe seven counLrles wlLhln
Lhe
senslLlve lndusLrles 1he 8CA measures changes ln Lhe share of a counLry's exporLs of a
producL ln world exporLs of Lhe producL compared Lo changes ln Lhe counLry's LoLal
share ln
world exporLs le relaLlve Lo Lhe slze of Lhe counLry lf Lhls raLlo ls more (less) Lhan
unlLy lL
means LhaL Lhe counLry ls aL a comparaLlve (dls)advanLage ln Lhe Lrade of Lhe producL
WlLh respecL Lo Lhe flrsL Lhe analysls shows LhaL world markeL shares ln
envlronmenLal
senslLlve goods have noL changed much over Lhe lasL Lwo decades 1he lndusLrlallzed
counLrles' share ln Lhe world Lrade of envlronmenLally senslLlve goods was 813 per
cenL ln
1970 and 811 per cenL ln 1990 Powever aL counLry level Lhere have been some
changes
AusLrla llnland and Cermany have lncreased Lhelr share whlle !apan norway
Sweden and
Lhe uS have decreased Lhelr shares CorrelaLlon coefflclenLs based on Llme serles for
AusLrla
28
Cermany !apan and Lhe uS beLween changes ln world Lrade shares of Lhe senslLlve goods and
changes ln envlronmenLal expendlLures conflrm LhaL Lhere ls no slgnlflcanL lmpacL of hlgher
envlronmenLal expendlLures on Lrade performance ln envlronmenLally senslLlve goods lor
Cermany !apan and Lhe uS Lhe correlaLlon coefflclenL was sLaLlsLlcally lnslgnlflcanL 1he only
counLry where Lhe correlaLlon coefflclenL was slgnlflcanL was AusLrla Powever lL was found
Lo be poslLlve
WlLh respecL Lo Lhe second approach a comparlson of 8CA's also showed LhaL sLrlcLer
envlronmenLal pollcy ln Lhe counLrles has noL decreased Lhelr comparaLlve advanLage ln
polluLlon lnLenslve goods 1he 8CA for Lhe lndusLrlallzed counLrles was 11 ln 1970 and 10 ln
1990 1hus Lhey have roughly malnLalned Lhelr comparaLlve advanLage ln envlronmenLally
senslLlve goods 1he same ls Lrue for Lhe developlng counLrles 1helr 8CA ln 1970 was 08 and
09 ln 1990 Agaln Lhere ls more dlvergence aL counLry level AusLrla and llnland have a
comparaLlve advanLage ln envlronmenLal senslLlve goods and have acLually lncreased Lhelr
compeLlLlveness from 1970 Lo 1990 norway has a comparaLlve advanLage buL Lhe 8CA has
dropped from 23 ln 1970 Lo 13 ln 1990 !apan has losL lLs comparaLlve advanLage ln Lhe11
perlod Sweden has a comparaLlve advanLage whlch has noL changed over Lhe perlod
Cermany's and Lhe uS' 8CA are 10 and 08 respecLlvely and have noL changed from 1970 Lo
1990 1he decrease ln !apan's and norway's 8CA may be due Lo oLher facLors Lhan sLrlcLer
envlronmenLal pollcy ln Lhe case of !apan Lhe decllne ln 8CA mlghL be caused by Lhe lncrease
of oLher exporLs 1he reason for Lhe decllne ln norway's 8CA may reflecL Lhe lncrease ln oll
revenues ln LoLal exporLs
Low and ?eaLs (1992) focused on secular changes LhaL occurred ln developed and
developlng counLrles' acLual Lrade and on 8evealed ComparaLlve AdvanLage (8CA) ln heavlly
polluLlng lndusLrles 1hls approach ls qulLe slmllar Lo Sorsa (1994) Low and ?eaLs compuLed
8CA lndlces for some 109 counLrles and provlded an almosL compleLe proflle of changlng
naLlonal comparaLlve advanLages wlLhln polluLlng lndusLrles 1he resulLs suggesL LhaL
dlrLy lndusLrles accounL for a growlng share of exporLs of some developlng counLrles
1hls has occurred agalnsL Lhe background of a reducLlon of such exporLs ln world Lrade
Powever Lhe sLudy does noL ldenLlfy Lhe dlfferenL facLors underlylng Lhls
developmenL ulfferences ln envlronmenLal pollcy could be [usL one ouL of a large
array of oLher slgnlflcanL facLors lor lnsLance for meLal manufacLures Lhe relaLlve
labour lnLenslLy could explaln dlrLy lndusLry locaLlon naLural resource endowmenLs
Lurned ouL Lo be an lmporLanL facLor for lndusLrles such as pulp and wasLe paper
peLroleum reflnlng cemenL and nonferrous meLals
1here are some drawbacks of Lhe LeonLlef approach llrsL Lhe Lhree sLudles
menLloned do noL Lake lnLo accounL dlfferences ln sLrlcLness of envlronmenLal pollcles
29
beLween counLrles and second lL ls a blvarlaLe analysls We now Lurn Lo Lhe PC
approach whlch may overcome Lhe problems lnherenL Lo Lhe LeonLlef approach ln Lhe
PC model lL ls posslble Lo lnclude several facLors whlch mlghL lnfluence Lrade paLLerns
lurLher lf lL ls posslble Lo consLrucL a measure lndlcaLlng Lhe sLrlngency of
envlronmenLal pollcy appllcable ln all counLrles cross counLry
comparlsons can be made
Myrrell and 8yLerman (1991) developed a procedure whlch can be used when some
varlables of lnLeresL are dlfflculL Lo measure or are unavallable whlch frequenLly
occurs ln Lhe
case of measurlng effecLs of envlronmenLal pollcy 1he approach ls an omlLLed varlable
procedure based on 8yLerman (1988) 1he procedure ls applled Lo LesL wheLher hlgh
envlronmenLal damages ln LasLern Lurope can be explalned by a Lendency Lo exporL
polluLlon
lnLenslve commodlLles 1he baslc assumpLlon ls LhaL LasLern Lurope has a laxer
envlronmenLal
pollcy Lhan lLs Lradlng parLners and should Lherefore have a comparaLlve advanLage ln
Lhe
polluLlon lnLenslve producLs 1he auLhors use 1973 sLaLlsLlcs on lmporLs and exporLs of
polluLlon lnLenslve goods deflned as goods whose producLlon produces more Lhan
average
30
levels of polluLlon or requlres more Lhan average amounLs of polluLlonabaLemenL cosL 1he
Lrade sLaLlsLlcs relaLes Lo nlne LasL Luropean counLrles seven lowlncome CLCu counLrles
and 10 LaLln Amerlcan counLrles 1he null hypoLhesls ls LhaL dlfferences ln envlronmenLal
pollcy do noL affecL Lhe sLrucLures of Lrade 1he LesL comes down Lo a comparlson of each
group Lo Lhe seven CLCu counLrles 1he null hypoLhesls was accepLed for all nlne LasL
Luropean CounLrles 1hus Lhe relaLlvely lax envlronmenLal pollcy ln LasLern Lurope does noL
show up ln a Lendency Lo exporL commodlLles lnLenslve ln polluLlon
kalL (1988) examlnes Lhe power of domesLlc regulaLory cosLs ln explalnlng aggregaLe Lrade
performance Pe focuses on Lhe envlronmenLal conLrol cosLs because Lhese are Lhe bulk of Lhe
regulaLory cosLs kalL's polnL of deparLure ls LhaL Lhe lmposlLlon of sLrlngenL envlronmenLal
regulaLlons ln Lhe uS can be vlewed as a decllne ln Lhe relaLlve abundance of envlronmenLal
resources Accordlng Lo Lhe PCmodel Lhe uS exporLs of envlronmenLlnLenslve producLs
would decllne ln Lhe regresslon analysls Lhe dependenL varlable ls Lhe neL exporLs from a12
speclflc lndusLry and among Lhe flve lndependenL varlables are expendlLures on polluLlon
servlces by each lndusLry 1hese expendlLures are deflned as Lhe annuallzed caplLal operaLlng
and admlnlsLraLlve cosLs of polluLlon abaLemenL ln 1977 represenLlng Lhe dlrecL lmpacL of
envlronmenLal regulaLlon kalL noLes LhaL an lndlrecL lmpacL of envlronmenLal regulaLlon arlses
when dlrecL polluLlon conLrol cosLs are noL enLlrely absorbed by Lhe dlrecLly spendlng lndusLry
and become embodled ln Lhe ouLpuL of lndusLrles LhaL purchase Lhe ouLpuL of Lhe dlrecLly
spendlng lndusLrles A crosssecLlon analysls for all 78 lndusLrles shows LhaL polluLlon
abaLemenL regulaLlons have an lnslgnlflcanL lnfluence on neL exporLs ln 1977 Powever for Lhe
32 manufacLurlng lndusLrles Lhe lnfluence ls negaLlve and slgnlflcanL 1hls resulL ls also
obLalned lf Lhe dependenL varlable alLernaLlvely ls speclfled as Lhe change ln neL exporLs over
Lhe perlod 19671977 and Lhe lndependenL varlable ls speclfled as Lhe change ln expendlLures
on polluLlon servlces over Lhe same perlod ln order Lo obLaln an esLlmaLe of Lhe LoLal polluLlon
abaLemenL cosLs le boLh Lhe dlrecL and Lhe lndlrecL kalL mulLlplles Lhe dlrecL abaLemenL
cosLs
by Lhe LoLal requlremenLs 1977 lnpuLouLpuL maLrlx 1he analysls shows LhaL lncludlng Lhe
lndlrecL cosLs as well does noL change Lhe concluslon olluLlon abaLemenL regulaLlons had an
lnslgnlflcanL lnfluence on neL exporLs ln 1977 lf all lndusLrles were lncluded buL Lhey were
negaLlve and slgnlflcanL lf Lhe analysls was narrowed by lncludlng only Lhe manufacLurlng
lndusLrles
1obey (1990) LesLs wheLher Lrade paLLerns for polluLlon lnLenslve commodlLles whlch are
produced by Lhe worlds mosL dlrLy lndusLrles are affecLed by envlronmenLal pollcy 1he LesL
ls
conducLed by uslng a crosssecLlon PeckscherChllnvanek (PCv) model whlch ls a mulLlfacLor
mulLlcommodlLy exLenslon of Lhe classlc PCmodel olluLlon lnLenslve producLs are
deflned as Lhe producLs of Lhe lndusLrles ln Lhe uS where Lhe polluLlon abaLemenL cosLs are
greaLer Lhan or equal Lo 183 per cenL of LoLal cosLs ln 1977 Lhere were 34 such lndusLrles
whlch 1oby dlvldes lnLo flve commodlLy groups neL exporLs ln a speclflc commodlLy ls
regressed on counLry characLerlsLlcs 1he daLa are from 1973 and are obLalned from a mlx of
13 CLCu counLrles and 10 developlng counLrles 1he LesL ls applled ln Lwo ways llrsL Lrade
ls regressed on counLry characLerlsLlcs lncludlng an lndex measurlng Lhe sLrlngency of polluLlon
conLrol measures 1hls measure ls based on a 1976 unC1Au survey ranklng Lhe counLrles ln
seven caLegorles from sLrlcL (7) Lo LoleranL (1) 1he measure of envlronmenLal pollcy
sLrlngency ls noL found Lo be sLaLlsLlcally slgnlflcanL ln Lhe regresslons of exporLs of Lhe
polluLlng lndusLrles 1he second approach ls Lhe omlLLed varlable LesL based on 8yLerman
(1988) and ls qulLe slmllar Lo Myrrell and 8yLerman (1991) 1he ldea ls LhaL lf Lhe sLrlngency
of envlronmenLal pollcy affecLs exporLs lL wlll affecL Lhe consLanL and dlsLurbance Lerm lf lL ls
omlLLed 1he resulL ls slmllar Lo Lhe flrsL approach Lhe sLrlngency of envlronmenLal pollcy does
31
noL affecL PCv Lrade paLLerns 1obey also relaxes Lwo of Lhe PCv assumpLlons llrsL
nonhomoLheLlc preferences are allowed 1hen consumpLlon ls a funcLlon of populaLlon and
lncome Second scale economles are allowed none of Lhese exLenslons makes Lhe sLrlngency
of envlronmenLal pollcy slgnlflcanL Pence 1obey does noL flnd any supporL aL all for Lhe
hypoLhesls LhaL envlronmenLal pollcy affecLs Lrade 1he negaLlve resulLs may be due Lo a Llme
lag beLween Lhe lnLroducLlon of envlronmenLal pollcy and Lhe subsequenL effecL on Lrade
1herefore 1obey conducLs a flxed effecLs LesL where he examlnes Lhe change ln Lrade
paLLerns
before and afLer Lhe lnLroducLlon of envlronmenLal measures ln Lhls LesL 1obey flnds a
slgnlflcanL effecL ln only one of Lhe flve commodlLy groups buL wlLh a poslLlve slgn
AL a flrsL glance kalL (1988) and 1obey (1990) obvlously derlve conLrasLlng concluslons
Powever Lhe concluslons from Lhe Lwo approaches are noL dlrecLly comparable because Lhe
dependenL and Lhe lndependenL varlables are noL deflned ln Lhe same way kalL deflnes hls13
dependenL varlable as neL exporLs of all lndusLrles or alLernaLlvely neL exporLs of Lhe
manufacLurlng lndusLrles 1he dependenL varlable ln 1obey (1990) ls a subseL of kalL's
because he concenLraLes on Lhe neL exporLs from Lhe 24 mosL polluLlon lnLenslve lndusLrles
Also Lhe envlronmenLal pollcy varlable dlffers beLween Lhe Lwo sLudles kalL uses expendlLures
on polluLlon abaLemenL servlces as a proxy for envlronmenLal pollcy 1obey uses a quallLaLlve
varlable measured on a scale from one Lo seven 1he purpose of kalL's sLudy was Lo examlne
Lhe power of domesLlc regulaLory cosLs ln explalnlng aggregaLe Lrade performance Pence Lhe
perspecLlve ls much broader Lhan Lhe one ln 1obey (1990) whlch ls more focused on
envlronmenLal pollcy
van 8eers and van den 8ergh (1997) are ln several respecLs an exLenslon of 1obey (1990)
llrsL a dlsadvanLage of Lhe mulLllaLeral approach ln 1obey (1990) ls LhaL effecLs of dlfferences
ln envlronmenLal pollcy on Lrade flows may cancel ouL because mulLllaLeral Lrade ls an
aggregaLe of bllaLeral Lrade flows 1herefore van 8eers and van den 8ergh analyse bllaLeral
Lrade flows made up of lmporLs and exporLs from one counLry Lo anoLher measured ln
Lhousands of uS$ Second Lhe resulLs may be sLrongly dependenL upon Lhe cholce of lndlcaLor
measurlng sLrlcLness of envlronmenLal pollcy van 8eers and van den 8ergh use Lwo such
measures A broad measure whlch reflecLs a large array of envlronmenLal lndlcaLors lncludlng
ones whlch do noL dlrecLly reflecL prlvaLe envlronmenLal cosLs 1he second ls a narrow
measure whlch more adequaLely reflecLs prlvaLe envlronmenLal cosLs ln Lerms of boLh
measures counLrles are ranked beLween 0 (weak envlronmenLal pollcy) and 1 (sLrlcL
envlronmenLal pollcy) Moreover Lhe unC1Au envlronmenLal pollcy measure used ln 1obey
(1990) ls an lnpuL orlenLed lndlcaLor whereas van 8eers and van den 8ergh use an
ouLpuLorlenLed envlronmenLal pollcy measure 1he former does noL Lake lnLo accounL
counLerbalanclng flnanclal asslsLance such as exporL rebaLes and lmporL surcharges and
Lherefore may exaggeraLe Lhe cosLs of envlronmenLal regulaLlons 1he ouLpuLorlenLed pollcy
measure Lakes Lhese counLerbalanclng flnanclal compensaLlons lnLo accounL Assume for
lnsLance LhaL a governmenL lmpose an energy Lax and aL Lhe same Llme subsldlze energy use
An lnpuLorlenLed pollcy measure does noL Lake lnLo accounL Lhe subsldles and Lhe
envlronmenLal pollcy would Lherefore look more sLrlcL Lhan lL ls An ouLpuLorlenLed measure
would plck up Lhe smaller reducLlon ln energy use caused by Lhe subsldles and would Lherefore
be a more correcL lndlcaLor of envlronmenLal pollcy sLrlcLness 1he Lhlrd dlfference compared
Lo 1obey (1990) ls LhaL 3 Lypes of dependenL varlables (bllaLeral Lrade flows) are used 1he
flrsL ls an aggregaLe bllaLeral Lrade flow lncludlng all lndusLrles 1he second ls bllaLeral Lrade
flow ln hlghly polluLlon lnLenslve secLors 1he Lhlrd ls bllaLeral Lrade flow ln hlghly polluLlon
lnLenslve secLors LhaL are noL resource based 1he nonresource based lndusLrles (Lhe
'fooLloose' lndusLrles) are hlghly moblle lndusLrles whlch are very senslLlve Lo lnLercounLry
dlfferences ln envlronmenLal pollcy naLurally Lhese lndusLrles are of greaL lnLeresL here
32
1he flrsL group of regresslon analyses whlch van 8eers and van den 8ergh perform ls qulLe
slmllar Lo Lhe 1obey regresslon buL now Lhe Lhree bllaLeral Lrade flows are used as Lhe
dependenL varlable As ln 1obey daLa are from 1973 and lncludes 14 CLCu counLrles and 9
developlng counLrles 1he resulL ls LhaL Lhe envlronmenLal pollcy lndlcaLor for exporLs ls
slgnlflcanL buL poslLlve for Lhe aggregaLe Lrade flows and for Lhe Lrade flows from Lhe
fooLloose and hlghly polluLlon lnLenslve secLors 1he coefflclenL ls lnslgnlflcanL for Lhe Lrade
flows from Lhe hlghly polluLlon lnLenslve secLors lncludlng boLh fooLloose and nonfooLloose
lndusLrles Pence more sLrlcL envlronmenLal regulaLlon may lncrease Lhe exporL of a counLry
A posslble explanaLlon ls LhaL ln Lhls regresslon Lhe envlronmenLal pollcy lndlcaLor ls
lnpuLorlenLed because van 8eers and van den 8ergh wanLed Lo approxlmaLe Lhe 1obey
regresslon as14
closely as posslble (excepL from Lhe bllaLeral Lrade flows as Lhe dependenL varlables) 8y uslng
Lhe lnpuLorlenLed envlronmenLal pollcy lndlcaLor posslble effecLs from subsldles are noL
capLured 1he envlronmenLal pollcy lndlcaLor ls lnslgnlflcanL for Lhe lmporLs
7
lf a dlsLlncLlon ls made beLween lndusLrles accordlng Lo Lhelr capablllLy Lo relocaLe Lhe
analysls shows LhaL Lhe envlronmenLal pollcy lndlcaLor ls slgnlflcanL and poslLlve for exporLs
buL only for Lhe fooLloose lndusLrles 1he envlronmenLal pollcy lndlcaLor ls lnslgnlflcanL for
lmporLs
1he second group of regresslon analyses ls relaLed Lo 1992 uaLa ls obLalned from 21
CLCu counLrles 8oLh Lhe broad and Lhe narrow envlronmenLal pollcy measure are used 1he
former has an lnslgnlflcanL effecL on exporLs as well as lmporLlng lrrespecLlve of whlch Lrade
flow ls used 1he narrow measure ls slgnlflcanL and negaLlve for exporLs lf aggregaLe Lrade
flows and Lrade flows from Lhe fooLloose secLors are used lL ls lnslgnlflcanL lf Lrade flows
from Lhe hlghly polluLlon lnLenslve secLors lncludlng boLh fooLloose and nonfooLloose
lndusLrles are used 1he lmpacL on lmporLs ls slgnlflcanL and negaLlve for all Lypes of secLors
1hls ls ln conLrasL Lo whaL Lrade Lheory would predlcL and may lndlcaLe Lhe presence of lmporL
barrlers
1here are however some problems wlLh Lhe operaLlonallzaLlon of sLrlcL envlronmenLal
pollcy 1he mosL lmporLanL problem ls LhaL Lhe maln lndlcaLors used ls energy use whlle
oLhers
relaLlng Lo eg wasLe are lgnored Accordlng Lo Lhe energy use lndlcaLor counLrles llke
norway and Sweden are found Lo have lax envlronmenLal pollcles 1hls can be explalned ln
Lerms of Lhe avallablllLy of hydropower ln Lhese counLrles As ls well known Lhe envlronmenLal
lmpacLs of hydropower are qulLe dlfferenL from Lhe lmpacLs of coal Pence lL ls doubLful lf
energy use ls a correcL lndlcaLor Moreover Lhe negaLlve lmpacL on Lhe exporLs from Lhe
fooLloose secLors ralses some quesLlons 8ecause of lLs very naLure for Lhe long run one
would expecL Lhls secLor Lo locaLe ln reglons wlLh lax envlronmenLal pollcles Pence lf Lhe was
Lo be a slgnlflcanL pollcy effecL one would expecL lL for Lhe resource based raLher Lhan Lhe
fooLloose secLors
lnLernaLlonal caplLal movemenLs
ln Lhls secLlon caplLal mlgraLlon ln Lhe form of forelgn dlrecL lnvesLmenL ls analysed 1hls form
of lnvesLmenL ls characLerlzed by a relaLlvely sLrong commlLmenL and a relaLlvely long Llme
horlzon and hence ls expecLed Lo be senslLlve Lo envlronmenLal pollcy
WalLer (1982) examlnes wheLher dlfferences ln envlronmenLal pollcy beLween counLrles wlll
favour lncremenLal lndusLrlal resLrucLurlng 1hree Lypes of daLa are used llrsL daLa for planL
closure based on an uS ueparLmenL of Commerce sample survey ln Lhe perlod 19741976
130 planLs were closed ln Lhe uS ln 87 cases Lhe enLlre esLabllshmenLs closed down
permanenLly ln 80 per cenL of Lhe cases alr polluLlon requlremenLs were glven as Lhe maln
polluLlon relaLed reason Lo Lhe planL closure naLurally oLher facLors such as age or
33
obsolescence also conLrlbuLed Lo Lhe planL closures buL unforLunaLely lL ls noL posslble Lo
deLermlne Lhe relaLlve welghL of Lhe envlronmenLal conslderaLlons Powever planL closlngs
for
envlronmenLal reasons seem Lo dlffer wldely for uS lndusLrles ln Lhe mlnlng lndusLry Lhe
shuLdown raLe was 872 per cenL and ln Lhe chemlcal lndusLry lL was 079 per cenL WalLer
concludes LhaL Lhe resulLs suggesL LhaL no slgnlflcanL mlgraLlon of producLlve caplLal may be
expecLed Lo resulL from closures 1hls concluslon however seems Lo resL on a slender
foundaLlon uaLa for more Lhan Lhree years are needed and more preclse evldence for Lhe
lmporLance of envlronmenLal regulaLlons relaLlve Lo oLher facLors are necessary 1he second
Lype of sLaLlsLlcs examlnes wheLher envlronmenLal conLrol cosLs glve a relocaLlon 'push' An13
lndlcaLor of Lhe 'push' ls a comparlson beLween domesLlc and overseas lnvesLmenLs ln
polluLlon conLrol lor seven ma[or polluLlon lnLenslve lndusLrles ln 19741973 Lhe percenLage
caplLal lnvesLmenL requlred for polluLlon conLrol abroad ls slgnlflcanLly lower Lhan ln Lhe uS
1hus ln Lhe uS Lhere ls a sLrong 'push' for Lhe polluLlng lndusLrles Lo relocaLe Powever lL
does noL say wheLher Lhey acLually relocaLe An lndlcaLor of Lhls could be Lhe flows of forelgn
dlrecL lnvesLmenL WalLer examlnes Lhe reglonal dlfferences of Lhe caplLal expendlLures of uS
companles uaLa of domesLlc and overseas caplLal expendlLures of uS companles are examlned
for Lhe perlod 19701977 1he overseas lnvesLmenLs are also dlfferenLlaLed by reglons 1he
daLa shows LhaL Lhe lndusLrles vary greaLly ln Lhe dlrecLlon of Lhelr forelgn dlrecL lnvesLmenL
Powever Lhere ls no slgnlflcanL lncrease ln Lhe forelgn dlrecL lnvesLmenL by polluLlon lnLenslve
lndusLrles ln developlng counLrles ln summary WalLer's analysls does noL flnd supporL for Lhe
hypoLhesls LhaL polluLlon lnLenslve lndusLrles have moved Lo less regulaLed counLrles
Leonard (1988) seeks Lo esLabllsh some recenL Lrends ln forelgn lnvesLmenL by uS lndusLrles
LhaL mlghL ascerLaln wheLher lndusLrles wlLh hlgh polluLlon conLrol cosLs have recenLly
lncreased lnvesLmenLs ln counLrles wlLh low envlronmenLal sLandards 8ased on polluLlon
abaLemenL expendlLures and Lhe exLenL of Lhe ease wlLh whlch lndusLrles can relocaLe he
selecLs Lhe chemlcal and Lhe mlneral processlng lndusLrles for furLher analysls 1he aggregaLe
lnvesLmenL and Lrade flgures lndlcaLe LhaL Lhe years lmmedlaLely followlng emergence of
sLrlngenL envlronmenLal regulaLlons ln Lhe uS dld noL wlLness wldespread relocaLlon of
lndusLrles Lo counLrles wlLh drasLlcally lower envlronmenLal regulaLlons nelLher ln Lhe long
run could Leonard flnd any slgnlflcanL effecLs and noL even when counLrles have Lrled Lo
aLLracL dlrLy lndusLrles 1he few lndusLrles LhaL moved ouL of Lhe uS had few lncenLlves Lo
lmprove producLlon faclllLles because demand for Lhelr producLs was reduced as a resulL of
producL obsolescence or hazards 1here are no examples of healLhy and growlng uS lndusLrles
LhaL have been forced Lo move abroad because of envlronmenLal regulaLlons ln recenL years
however among polluLlon lnLenslve lndusLrles some Lrends LhaL could be relaLed Lo lncreased
envlronmenLal regulaLlons seem Lo occur Cf greaLesL poLenLlal slgnlflcance Leonard remarks
are Lhe large lncreases ln recenL years ln overseas lnvesLmenL and reflned lmporLs by uS
mlneral processors
8owland and lelock (1991) do noL conslder lnLernaLlonal mlgraLlon buL mlgraLlon across
sLaLes ln Lhe uS 1he purpose ls Lo lnvesLlgaLe Lo whaL exLenL Lhe lnLersLaLe allocaLlon of new
Chemlcal and Allled roducers (CA) caplLal can be accounLed for by lnLersLaLe varlance ln
CA polluLlon abaLemenL cosLs 1he federal sLaLes ln Lhe uS dlffer greaLly from each oLher ln
Lerms of envlronmenLal regulaLlon of Lhe CA as well as share of new CA caplLal 8owland
and lelock come Lo Lhe opposlLe concluslon of WalLer (1982) and Leonard (1988) olluLlon
abaLemenL cosLs have a slgnlflcanL negaLlve lnfluence on Lhe sLaLes' ablllLy Lo aLLracL new CA
caplLal ln Lhe perlod 19771981 Powever Lhls relaLlonshlp ls nonllnear 1here appears Lo be
a Lhreshold CaplLal relocaLes lf a sLaLe exceeds Lhe Lhreshold buL below lL Lhere ls no
relocaLlon 1hls means LhaL whlle sLaLes whlch lmpose hlgh polluLlon abaLemenL cosLs Lend Lo
compeLe unsuccessfully for new caplLal Lhose sLaLes wlLh Lhe lowesL marglnal polluLlon
34
abaLemenL cosLs are noL necessarlly Lhe mosL successful ln aLLracLlng new CA caplLal 8aLher
Lhere seems Lo be a polluLlon abaLemenL expense Lhreshold When a sLaLe exceeds lL Lhe sLaLe
loses caplLal Lo sLaLes LhaL remaln below lL 8owland and lelock do noL dlscuss wheLher or noL
Lhelr resulL can be generallzed Lo an lnLernaLlonal seLLlng
PeLLlge eL al (1992) and Lucas eL al (1992) examlned wheLher envlronmenLal regulaLlon ln
Lhe CLCu counLrles has dlsplaced Loxlc lndusLrlal producLlon Lowards less regulaLed LuC's
1hey flnd supporL for Lhe replacemenL hypoLhesls by sLudylng measures of Lhe Loxlc
lnLenslLy16
of manufacLurlng ln 80 counLrles from 1960 Lo 1988 1he daLa are dlvlded lnLo subseLs for
each decade and shows LhaL ln Lhe 1960s Loxlc lnLenslLles grew mosL qulckly ln hlghlncome
counLrles AfLer Lhe lnLroducLlon of sLrlcLer envlronmenLal regulaLlon ln Lhe CLCu counLrles ln
Lhe 1970s and 1980s Lhe paLLern was sharply reversed Loxlc lnLenslLy ln LuC grew mosL
qulckly lurLher Lhey conclude LhaL polluLlon lnLenslLy has grown mosL rapldly ln developlng
counLrles whlch are relaLlvely closed Lo world markeL forces 1he same concluslon ls reached
ln 8lrdsall and Wheeler (1992) who flnd LhaL ln LaLln Amerlca polluLlon heavens are Lo be
found ln Lhe mosL proLecLlonlsL economles Cpenness promoLes cleaner lndusLrles Powever a
ma[or drawback of PeLLlge eL al and Lucas eL al ls LhaL Lhe regresslons do noL lnclude any
envlronmenLal pollcy varlables 1hus one cannoL be cerLaln of a causal connecLlon beLween
Lhe decade paLLerns and Lhe shlfLs ln CLCu envlronmenLal pollcles

Conc|us|on

lncreaslng volume of lnLernaLlonal Lrade ln Lhe world ln recenL years causes dlfferenL
concerns Cne of Lhem ls Lhe envlronmenLal lmpacL of Lrade ulfferenL Lheorles are developed
and analyzed from dlfferenL perspecLlves Cne of Lhe hlghly debaLed argumenLs ls Lhe
'olluLlon Paven PypoLhesls' Accordlng Lo Lhls hypoLhesls for dlfferenL reasons baslcally dlrLy
secLors are locaLed ln developlng counLrles and domlnaLe Lhelr exporLs 1herefore developed
counLrles lmporL Lhe dlrLy producLs produced ln developlng counLrles whlch become polluLlon
havens 1hls hypoLhesls ls LesLed by uLlllzlng dlfferenL analyLlcal meLhods for dlfferenL
counLrles for dlfferenL Llme horlzons
ln Lhls sLudy 'polluLlon Lerms of Lrade' lndex for CC2 ls developed for Lhe 1urklsh
economy for Lhe years 1996 and 1998 1hls lndex measures Lhe CC2 conLenL of one 1L worLh
35
of 1urklsh exporLs relaLlve Lo CC2 conLenL of one 1L worLh of 1urklsh lmporLs lf 1C1 ls
greaLer Lhan 100 Lhe polluLlon conLenL of exporLs ls greaLer Lhan Lhe polluLlon conLenL of
lmporLs of Lhe counLry lL could be argued LhaL a decllnlng lndex value may lmply LhaL Lhe
economy ls exporLlng clean goods and lmporLlng dlrLy goods ln general and consequenLly ls
noL a polluLlon haven 1herefore lL could be concluded LhaL Lhe economy galns from Lrade ln
Lerms of polluLlon lf Lhe value of 1C1 ls less Lhan 100
When 1urklsh case ls consldered even Lhough our analysls does noL cover a long
perlod a slgnlflcanL lncrease ln Lhe overall 1C1 value has been observed for 1996 Lhe value of
1C1 ls 49 and for 1998 lL ls 103 lL could be lnLerpreLed LhaL ln Lhe 1urklsh economy CC2
conLenL of exporLs slgnlflcanLly lncreases relaLlve Lo Lhe CC2 conLenL of lmporLs ln 1998 1he
deLerloraLlon ln Lhe overall value of Lhe lndex can be lnLerpreLed wlLh cauLlon as a move
Lowards belng a polluLlon haven 1o achleve more convlnclng concluslons lL ls apparenL LhaL
daLa coverlng a wlder Llme span ls necessary
AnoLher polnL LhaL requlres cauLlon ln lnLerpreLaLlon of 1C1 values ls LhaL Lhey are
calculaLed by assumlng LhaL producLlon process of lmporLed goods lnvolves Lhe same
Lechnology as of Lhe domesLlcally produced goods lnevlLably Lhls ldenLlcal Lechnologles
assumpLlon for domesLlc and lmporLed goods producLlon mlghL cause blas ln emplrlcal resulLs
and should be Laken lnLo accounL ln lnLerpreLlng Lhe resulLs under Lhe supposlLlon LhaL
domesLlc Lechnology used ln 1urkey could be relaLlvely older and dlrLler Lhan her Lrade
parLners (whlch are baslcally lndusLrlallzed counLrles) Lhls ldenLlcal Lechnologles assumpLlon
mlghL undersLaLe Lhe 1C1 1herefore Lhe danger for 1urkey becomlng a polluLlon haven
mlghL be even more serlous Lhan deLecLed by Lhe currenL sLudy
AnoLher caveaL of Lhls sLudy ls Lhe facL LhaL 1998 ls a speclal year ln whlch 1urklsh
currenL accounL glves a surplus whlch ls a rare lncldence ln Lhe recenL pasL lL ls a year ln
36
whlch 1urklsh economy recelved negaLlve forelgn shocks llke Lhe Aslan crlsls and more
lmporLanLly Lhe 8usslan crlsls uue Lo 1urkey's close Lrade Lles wlLh 8ussla lL had serlous
ramlflcaLlons for Lhe 1urklsh economy (Alper and Cnl; 2003) 1herefore parL of Lhe varlaLlon
capLured ln Lhe 1C1 by Lhls sLudy could be due Lo Lhls macroeconomlc flucLuaLlon
A flnal noLe wlLh respecL Lo 1C1 ls LhaL no maLLer how small a counLry's 1C1 geLs
unless Lhe producLlon processes all around Lhe world are noL cleaned Lhere ls no escape from
deLrlmenLal effecLs of polluLlon


37
ootnotes
1. In Use Table, columns demonstrate the payments made to inputs and value added
components Ior the production oI an industry and rows show the sales oI industries to
industries and Iinal demand components. On the other hand Supply Table shows how
much oI the industry`s own product is used by itselI and other industries.
2. For alternative assumptions about the construction oI symmetric input-output tables,
see Bulmer-Thomas (1982) and Miller and Blair (1985).
3. However, in manuIacturing industry statistics the Iigures Ior petroleum products are
provided and since other sectors use petroleum products instead oI crude petroleum,
in calculations petroleum products are taken into consideration instead oI crude
petroleum.
4. For a detailed examination oI the structure oI the Turkish economy, exports, imports
and dependency on imports in 1998 one can reIer to Gnlk-Senesen (2005).



38
1A8LLS

1ab|e 1 Sectora| roduct|on


Sector
no

Sector
1996 1998
1994 prlces
(bllllon 1L)
1994 prlces
(bllllon 1L)
1 AgrlculLure and husbandry 8992364602 1074 8070073106 933
2 Mlnlng of coal and llgnlLe coke producLs 334334101 042 273334137 032
3 LxLracLlon of crude peLroleum 100814 000 120796132 014
4 roducLlon and dlsLrlbuLlon of naLural gas 173960863 021 88833386 010
3 CLher mlnlng 300223128 036 417072344 048
6 lood beverages and Lobacco 3879313063 702 3311316131 614
7 Cocoa chocolaLe and oLhers 677240842 081 242763410 028
8 1exLlle wearlng apparel and leaLher 3730384814 683 3860316670 678
9 Sawmllls plalnlng and oLher wood mllls 346968694 041 332762723 041
39
10 Wood producLs and furnlLure 333796918 066 1183821017 137
11 aper and paper producLs 603433313 072 727439304 084
12 rlnLlng publlshlng 473622420 037 913102301 106
14 8eflned peLroleum producLs 2230261133 266 2233106321 260
14 8aslc chemlcals 781374233 093 687980749 080
13 lerLlllzers 237313630 031 239993710 030
16 CLher chemlcals and plasLlc producLs 1408863742 168 1783823980 206
17 Cleanlng maLerlals and cosmeLlcs 636111919 076 716333047 083
18 ManufacLure of Llre and rubber producLs 391674139 047 406227817 047
18 Class and glass producLs 243329132 029 279646426 032
20 Ceramlc producLs 392388068 047 406276716 047
21 CemenL llme and plasLer 713823929 086 764670030 088
22 nonmeLalllc mlneral producLs 228771326 027 190738938 022
23 ManufacLure of baslc lron and sLeel 1770964838 212 2038074473 236
24 ManufacLure of oLher meLals and casLlng 370410763 068 806107426 093
23 SLrucLural meLal producLs 321124864 062 783273647 091
40
26 labrlcaLed meLal producLs 893362227 107 893869263 103
27 ManufacLure of machlnery 1136424308 138 1048613186 121
28 uom appllances professlonal and sclenLlflc equlpm 694069376 083 746098933 086
29 Cfflce accounLlng and compuLer machlnery 26026222 003 36423923 004
30 CLher elecLrlcal apparaLus 694461131 083 829041814 096
31 8adlo 1v and communlcaLlon apparaLus 376381387 069 341077199 063
32 1ransporLaLlon vehlcles 1921070367 229 2091866230 242
33 CLher manufacLurlng 464799436 036 1010442682 117
34 roducLlon and dlsLrlbuLlon of elecLrlclLy 1338362733 162 963311626 112
33 WaLer and dlsLrlbuLlon 297349620 036 260792987 030
36 ConsLrucLlon 6169432200 737 7300931302 844
37 1ransporLaLlon
1043933783
3 1247 10361248323 1198
38 CLher servlces
1928249463
2 2303 19722482311 2280
39 ubllc servlces 3322217324 421 4004318132 463
40 Cwnershlp of dwelllng 1978928693 236 1738460846 201
41
1C1AL 8371632030 100 86497731S9 100












1ab|e 2 Sectora| Lxports

Sector
Sector 1996 1998
42
no 1994 prlces
(bllllon 1L)
1994 prlces
(bllllon 1L)
1 AgrlculLure and husbandry 1400001739 423 3693036324 383
2 Mlnlng of coal and llgnlLe coke producLs 10374003 003 16823467 001
3 LxLracLlon of crude peLroleum 01 000 9998313 001
4 roducLlon and dlsLrlbuLlon of naLural gas 3430304 001 841642 000
3 CLher mlnlng 230300992 076 838610933 037
6 lood beverages and Lobacco 748161341 227 3340700217 226
7 Cocoa chocolaLe and oLhers 263398004 081 668673779 043
8 1exLlle wearlng apparel and leaLher 6772409886 2037 33337371603 2387
9 Sawmllls plalnlng and oLher wood mllls 12786000 004 18606109 001
10 Wood producLs and furnlLure 82721379 023 1328090713 090
11 aper and paper producLs 38416633 018 312984762 021
12 rlnLlng publlshlng 167310319 031 308791106 021
14 8eflned peLroleum producLs 208693376 063 2143808799 143
14 8aslc chemlcals 196910386 060 843402183 037
43
13 lerLlllzers 10323179 003 32334124 004
16 CLher chemlcals and plasLlc producLs 147242894 043 636299809 044
17 Cleanlng maLerlals and cosmeLlcs 343097812 163 1463030372 099
18 ManufacLure of Llre and rubber producLs 169847604 032 840799932 037
18 Class and glass producLs 96987337 029 1030412911 070
20 Ceramlc producLs 168362404 031 649797319 044
21 CemenL llme and plasLer 73292764 023 348771173 024
22 nonmeLalllc mlneral producLs 71679267 022 274636124 019
23 ManufacLure of baslc lron and sLeel 61 000 3344760713 239
24 ManufacLure of oLher meLals and casLlng 173249433 033 880729006 039
23 SLrucLural meLal producLs 72072162 022 404649133 027
26 labrlcaLed meLal producLs 180226166 033 1019712961 069
27 ManufacLure of machlnery 730122803 228 2392610402 162
28 uom appllances professlonal and sclenLlflc equlpm 382314177 116 1670433363 113
29 Cfflce accounLlng and compuLer machlnery 20820796 006 77932343 003
30 CLher elecLrlcal apparaLus 499342830 132 2334746423 139
44
31 8adlo 1v and communlcaLlon apparaLus 627330137 191 3632836241 243
32 1ransporLaLlon vehlcles 683342804 208 4327439369 292
33 CLher manufacLurlng 120338070 037 2063613343 139
34 roducLlon and dlsLrlbuLlon of elecLrlclLy 10362606 003 43661413 003
33 WaLer and dlsLrlbuLlon 00 000 00 000
36 ConsLrucLlon 00 000 6629342692 448
37 1ransporLaLlon 6100907620 1833 32319918312 2196
38 CLher servlces 11382403786 3437 30012323091 2027
39 ubllc servlces 00 000 00 000
40 Cwnershlp of dwelllng 436213038 139 306878028 021
1C1AL 3292222100 100 1480637SS80 100





45
1ab|e 3 Sectora| Imports

Sector
no Sector
1996 1998
1994 prlces
(bllllon 1L)
1994 prlces
(bllllon 1L)
1 AgrlculLure and husbandry 1494807039 377 4098946660 373
2 Mlnlng of coal and llgnlLe coke producLs 476416043 120 63323380 009
3 LxLracLlon of crude peLroleum 1668029 000 7192816 001
4 roducLlon and dlsLrlbuLlon of naLural gas 3137466263 792 673320908 094
3 CLher mlnlng 184683009 047 124833431 018
6 lood beverages and Lobacco 1677913697 424 3606689914 306
7 Cocoa chocolaLe and oLhers 81726433 021 88486112 012
8 1exLlle wearlng apparel and leaLher 1720327236 434 3799369734 813
9 Sawmllls plalnlng and oLher wood mllls 23160367 006 133673149 019
10 Wood producLs and furnlLure 111364694 028 1001487742 140
11 aper and paper producLs 331477624 134 734377639 103
46
12 rlnLlng publlshlng 49110832 012 388746672 033
14 8eflned peLroleum producLs 776341388 196 3711293660 320
14 8aslc chemlcals 2206100664 337 1313737976 183
13 lerLlllzers 236782331 063 379790780 081
16 CLher chemlcals and plasLlc producLs 331232727 134 1460033478 203
17 Cleanlng maLerlals and cosmeLlcs 949633310 240 1007836641 141
18 ManufacLure of Llre and rubber producLs 168267373 042 441769937 062
18 Class and glass producLs 116683234 029 287109363 040
20 Ceramlc producLs 126823218 032 304210400 043
21 CemenL llme and plasLer 3719007 001 68019370 010
22 nonmeLalllc mlneral producLs 60431823 013 63984238 009
23 ManufacLure of baslc lron and sLeel 1749383698 442 3423337093 480
24 ManufacLure of oLher meLals and casLlng 1778194801 449 1233749714 173
23 SLrucLural meLal producLs 116441330 029 731607810 103
26 labrlcaLed meLal producLs 439319277 111 803046829 113
27 ManufacLure of machlnery 6333144381 1633 3124373003 438
47
28 uom appllances professlonal and sclenLlflc equlpm 1233221109 311 693069733 097
29 Cfflce accounLlng and compuLer machlnery 1387110917 330 174038332 024
30 CLher elecLrlcal apparaLus 2608228337 639 3642618723 791
31 8adlo 1v and communlcaLlon apparaLus 1469082981 371 2037213038 286
32 1ransporLaLlon vehlcles 3993332067 1008 3036108023 429
33 CLher manufacLurlng 468366033 118 2972274963 417
34 roducLlon and dlsLrlbuLlon of elecLrlclLy 9028147 002 1940279334 272
33 WaLer and dlsLrlbuLlon 00 000 33483317 008
36 ConsLrucLlon 21123192 003 3974374839 838
37 1ransporLaLlon 1310048343 331 6620736376 928
38 CLher servlces 1774879209 448 6663468262 934
39 ubllc servlces 00 000 00 000
40 Cwnershlp of dwelllng 00 000 172483734 024
1C1AL 3960128709 100 7130978781 100


48


1ab|e 4 o||ut|on 1erms of 1rade

1996 1998
CC2 embodled ln exporLs 9.13458E-05 9.010533E-05
CC2 embodled ln lmporLs 1869L03 863638L03
1C1 4887 10343

















49























1ab|e S Sectora| CC2 Content of Lxports and Imports (kton)

50
Sector
no Sector
1996 1998
Lxports Imports Lxports Imports
1 AgrlculLure and husbandry 9.67 6.34 14.154 1.1
2 Mlnlng of coal and llgnlLe coke producLs 0.885 33.779 0.068 0.548
3 LxLracLlon of crude peLroleum 5.594E-16 0.016 0.010 0.001
4 roducLlon and dlsLrlbuLlon of naLural gas 0.569 43.88 1.18E-10 1.965
3 CLher mlnlng
7.718 4.734 3.49 1.083
6 lood beverages and Lobacco 14.86 7.715 1.410 0.388
7 Cocoa chocolaLe and oLhers 5.509 1.409 9.73 0.846
8 1exLlle wearlng apparel and leaLher 119.383 5.15 49.46 16.89
9 Sawmllls plalnlng and oLher wood mllls 0.310 0.507 0.06 0.390
10 Wood producLs and furnlLure 1.88 .107 1.951 3.063
11 aper and paper producLs
1.784 13.495 0.694 3.390
12 rlnLlng publlshlng 3.410 0.83 0.317 0.83
14 8eflned peLroleum producLs 4.814 13.48 .38 8.068
14 8aslc chemlcals 8.754 81.546 0.111 .559
13 lerLlllzers 0.960 19.858 .86 9.157
16 CLher chemlcals and plasLlc producLs
5.088 15.65 1.348 6.43
17 Cleanlng maLerlals and cosmeLlcs
15.884 3.095 4.73 6.19
18 ManufacLure of Llre and rubber producLs 5.196 4.81 1.953 .136
18 Class and glass producLs 4.048 4.049 480.016 78.496
20 Ceramlc producLs 6.67 3.95 .391 .331
21 CemenL llme and plasLer 5.8 0.17 .13 0.898
22 nonmeLalllc mlneral producLs
1.090 0.764 0.50 0.5
51
23 ManufacLure of baslc lron and sLeel 4.73E-13 101.145 11.381 .901
24 ManufacLure of oLher meLals and casLlng
7.354 6.76 1.818 5.30
23 SLrucLural meLal producLs 3.584 4.814 0.557 .153
26 labrlcaLed meLal producLs 8.49 16.719 1.390 .86
27 ManufacLure of machlnery 4.611 309.618 4.381 11.913
28 uom appllances professlonal and sclenLlflc equlpm 14.40 38.596 1.896 177.75
29 Cfflce accounLlng and compuLer machlnery
.791 154.593 0.073 0.341
30 CLher elecLrlcal apparaLus
4.457 106.176 3.385 16.890
31 8adlo 1v and communlcaLlon apparaLus 7.785 15.15 1.848 .158
32 1ransporLaLlon vehlcles 3. 11.473 4.576 6.78
33 CLher manufacLurlng 3.771 1.185 1.146 3.438
34 roducLlon and dlsLrlbuLlon of elecLrlclLy 1.189 0.084 0.808 71.48
33 WaLer and dlsLrlbuLlon
0.095 0 0 0.133
36 ConsLrucLlon 0 0.068 17.680 33.179
37 1ransporLaLlon 317. 566.70 45.39 104.003
38 CLher servlces 17.03 .303 35.565 16.44
39 ubllc servlces 0 0 0 0
40 Cwnershlp of dwelllng 3.398 0 0.167 0.196
1C1AL
913.458 186.911
910S33 863638

Note All values excepL exporLs of exLracLlon of crude peLroleum for 1996 exporLs of manufacLure of
baslc lron and sLeel for 1996 exporLs of producLlon and dlsLrlbuLlon of naLural gas for 1998 ln Lhe Lable
should be mulLlplled by 10
7



52

8I8LICGkAn

rlnclples of LnvlronmenLal and 8esource Lconomlcs A Culde for
SLudenLs and ueclslonMakers Ldward Llgar
8aumol W ! (1971) LnvlronmenLal roLecLlon lnLernaLlonal Splllovers and
1rade
1rade and Lhe LnvlronmenL 1he Search for 8alance
'LnvlronmenLal 8egulaLlon ln an Cpen Lconomy' !ournal of
LnvlronmenLal Lconomlcs and ManagemenL

WL8Sl1L
hLLp//slLlfeemlL
hLLp//llbrarywurnl
wwwwlklpedlacom
wwweconomywaLchcom