AL dlnner Lhey dld dlscourse very flnely Lo us of Lhe probablllLy LhaL Lhere ls a !"#$ &'"( )* +),'- ./& ln Lhe land.
-- Samuel epys, ulary (1663)
0-' +),1 2', +),
-- PalLlan proverb
3.'4'# #)+'$./,2 )5$ $.'4' 6 6 6 ",& /$# 789:;<7=
-- 8lpley
Iames S. nenry
Senlor Advlsor/ Clobal 8oard Member
1ax !usLlce neLwork
!uly 2012
* 1hanks Lo !ames 8oyce, !ohn ChrlsLensen, Alex Cobham, 8lchard Murphy, 8onan alan, Sol lccloLLo, and nlcholas Shaxson for Lhelr consLrucLlve feedback. 1hanks are also due Lo Lhe World 8ank, Lhe lMl, Lhe 8lS, and Lhe un for much of Lhe baslc daLa analyzed ln Lhls reporL.
2 !ames S. Penry, 1!n 2012
1A8LL CI CCN1LN1S
1he r|ce of Cffshore kev|s|ted
1. IN1kCDUC1ICN] SUMMAk AND kL IINDINGS
2. WnLkL IS CIISnCkL, ANWA?
3. 1nL GLC8AL nAVLN INDUS1k
4. CLD LS1IMA1LS
S. NLW LS1IMA1LS
6. 1kADL MISkICING AN ASIDL
7 . IMLICA1ICNS
ALNDIk I: 1nL kL-nIS1Ck CI CIISnCkL LS1IMA1LS
ALNDIk II: LkLAINING CAI1AL ILIGn1
ALNDIk III: kL CnAk1S
3 !ames S. Penry, 1!n 2012
1. IN1kCDUC1ICN] SUMMAk
1he deflnlLlon of vlcLory for Lhls paper ls Lo revlew and lmprove upon exlsLlng esLlmaLes of Lhe slze, growLh and dlsLrlbuLlon of unLaxed prlvaLe wealLh proLecLed and servlced by Lhe global offshore lndusLry.
1hls ls necessarlly an exerclse ln nlghL vlslon. 1he subLerranean sysLem LhaL we are Lrylng Lo measure ls Lhe economlc equlvalenL of an asLrophyslcal black hole.
Llke Lhose black holes, Lhls one ls vlrLually lnvlslble and can be somewhaL perllous Lo observers who venLure Loo close. So, llke asLronomers, researchers on Lhls Loplc have necessarlly used lndlrecL meLhods Lo do Lhelr esLlmaLes, conducLlng Lhelr measuremenLs from a respecLful dlsLance. 1hls lndlrecL approach ls palnsLaklng, and has many lnherenL llmlLaLlons, as well see.
unllke ln Lhe fleld of asLrophyslcs, however, Lhe lnvlslblllLy here ls fundamenLally man- made. rlvaLe secLor secrecy and Lhe offlclal governmenL pollcles LhaL proLecL lL have placed mosL of Lhe daLa LhaL we need dlrecLly off llmlLs even Lhough lL ls, ln prlnclple, readlly avallable.
ln many ways, Lhe cruclal pollcy quesLlon ls whaL are Lhe cosLs and beneflLs of all Lhls secrecy?
AnoLher key Lheme LhaL emerges from Lhls paper ls LhaL Lhere ls an urgenL need for Lax [usLlce advocaLes and Lhelr allles ln governmenLs and ln Lhe publlc, especlally ln source counLrles where Lhe wealLh ls comlng from, Lo press Lhe relevanL auLhorlLles for Lhls lnformaLlon.
1he very exlsLence of Lhe global offshore lndusLry, and Lhe Lax-free sLaLus of Lhe enormous sums lnvesLed by Lhelr wealLhy cllenLs, ls predlcaLed on secrecy: LhaL ls whaL Lhls lndusLry really supplles as lL compeLes for, conceals, and manages prlvaLe caplLal from all over Lhe planeL, from any and all sources, no quesLlons asked.
We are up agalnsL one of socleLys mosL well-enLrenched lnLeresL groups. AfLer all, Lheres no lnLeresL group more rlch and powerful Lhan Lhe rlch and powerful, who are Lhe ulLlmaLe sub[ecLs of our research.
1he flrsL sLep, however, are Lhe esLlmaLes. 1he way ls hard, Lhe work ls Ledlous, Lhe daLa mlnlng ls as mlnd-numblng as any day below surface aL Lhe coal face, and Lhe esLlmaLes are sub[ecL Lo maddenlng, lrreduclble uncerLalnLles.
neverLheless, as usual, some Lhlngs may be sald.
4 !ames S. Penry, 1!n 2012 New Lst|mates. As dlscussed below, prevlous esLlmaLes of Lhe slze and growLh of Lhe offshore lndusLry Lo daLe have relled on rough [udgmenLs and rules of Lhumb or, aL besL, on one or Lwo very slmple esLlmaLlon meLhods.
We LrlangulaLed on our esLlmaLes from Lhe vanLage polnL of several dlfferenL meLhods. 1he alm ls noL pseudo-preclslon, much less really blg numbers, buL Lo ldenLlfy a plauslble base case for Lhls oLherwlse-well hldden secLor of Lhe global economy.
A More Cpen rocess. AnoLher ob[ecLlve ls Lo keep a sharp eye ouL for Lhe puzzles surfaced by Lhls daLa analysls, of whlch Lhere are many. A key problem wlLh prevlous esLlmaLes ls sensaLlonallsm. 1haL ls Lo be expecLed, glven Lhe sub[ecL maLLer, and Lhe facL LhaL esLlmaLlon ls sLlll domlnaLed by relaLlvely closed communlLles of consulLlng flrms, governmenL agencles, or nCCs.
An lmporLanL alm of Lhls pro[ecL ls Lo esLabllsh a more open, LransparenL, collaboraLlve model for dolng such research so LhaL Lhe daLa sources, esLlmaLlon meLhods, and core assumpLlons are all exposed Lo Lhe sunllghL of peer revlew, and ulLlmaLely Lo publlc scruLlny.
Lst|mat|on Methods. As dlscussed below ln more deLall, Lhls paper employs four key esLlmaLlon approaches: (1) a sources-and-uses model for counLry-by-counLry unrecorded caplLal flows, (2) an accumu|ated offshore wea|th model, (3) an offshore |nvestor portfo||o model, and (4) dlrecL esLlmaLes of offshore asseLs aL Lhe worlds Lop 30 g|oba| pr|vate banks.
1o complle lLs esLlmaLes, Lhe paper uses laLesL avallable daLa from Lhe World 8ank and lMl, Lhe un, cenLral banks, and naLlonal accounLs Lo expllclLly model caplLal flows for each member of a subgroup of 139 key source counLrles LhaL publlsh such daLa.
1he paper goes furLher, supplemenLlng Lhese models wlLh oLher evldence, lncludlng (1) daLa on so-called Lransfer mlsprlclng, (2) daLa on Lhe cross-border demand for llquld maLLress money llke reserve currency and gold, parL of whlch may move Lhrough offshore markeLs, and (3) a revlew of markeL research by leadlng consulLlng flrms on Lhe slze of Lhe offshore prlvaLe banklng markeL. (See SecLlon 3, below, for more deLalls.)
We belleve LhaL Lhe resulLlng esLlmaLes of unrecorded caplLal flows and accumulaLed offshore wealLh are Lhe mosL rlgorous and comprehenslve ever produced. 1 ln Lhe splrlL of open research, we hereby lssue an open challenge Lo Lhe lMl and Lhe World 8ank Lo all comers, ln facL Lo see lf Lhey can come up wlLh beLLer esLlmaLes.
1 We wlll house Lhls and accompanylng reporLs permanenLly here. hLLp://www.Lax[usLlce.neL/cms/fronL_conLenL.php?ldcaL=148
3 !ames S. Penry, 1!n 2012
kL IINDINGS
Cvera|| S|ze
A slgnlflcanL fracLlon of global prlvaLe flnanclal wealLh -- by our esLlmaLes, aL leasL 521 to 532 tr||||on as of 2010 -- has been lnvesLed vlrLually Lax-free $.4)52. Lhe worlds sLlll- expandlng black hole of more Lhan 80 offshore secrecy [urlsdlcLlons. We belleve Lhls range Lo be conservaLlve, for reasons dlscussed below.
8emember: Lhls ls [usL flnanclal wealLh. A blg share of Lhe real esLaLe, yachLs, racehorses, gold brlcks -- and many oLher Lhlngs LhaL counL as non-flnanclal wealLh -- are also owned vla offshore sLrucLures where lL ls lmposslble Lo ldenLlfy Lhe owners. 1hese are ouLslde Lhe scope of Lhls reporL.
Cn Lhls scale, Lhls offshore economy ls large enough Lo have a ma[or lmpacL on esLlmaLes of lnequallLy of wealLh and lncome, on esLlmaLes of naLlonal lncome and debL raLlos, and mosL lmporLanLly Lo have very slgnlflcanL negaLlve lmpacLs on Lhe domesLlc Lax bases of key source counLrles (LhaL ls, counLrles LhaL have seen neL unrecorded prlvaLe caplLal ouLflows over Llme 2 .)
2. Cur 139-country focus group: who are the rea| debtors?
We have focused on a subgroup of 139 malnly low-mlddle lncome source counLrles 3
for whlch Lhe World 8ank and lMl have sufflclenL exLernal debL daLa.
Cur esLlmaLes for Lhls group underscore how mlsleadlng lL ls Lo regard counLrles as debLors only by looklng aL one slde of Lhelr balance sheeLs.
Slnce Lhe 1970s, wlLh eager (and ofLen aggresslve and lllegal) asslsLance from Lhe lnLernaLlonal prlvaLe banklng lndusLry, lL appears LhaL prlvaLe ellLes ln Lhls sub-group of 139 counLrles had accumulaLed 57.3 to 59.3 tr||||on of unrecorded offshore wea|th ln 2010, conservaLlvely esLlmaLed, even whlle many of Lhelr publlc secLors were borrowlng
2 More preclsely, source counLrles" are Lhose whose LoLal real accumulaLed neL unrecorded prlvaLe caplLal ouLflows are poslLlve for Lhe perlod 1970-2010 (or whaLever perlod ls avallable for Lhe parLlcular counLry.) LqulvalenLly, lL ls Lhose counLrles whose prlvaLe clLlzens have accumulaLed poslLlve neL unreporLed / unLaxed flnanclal wealLh abroad. 3 1hls ls a raLher homogenous group, Lhe large ma[orlLy of whlch are low-mlddle lncome counLrles, buL lnclude some rlch non-CLCu counLrles llke Saudl Arabla and kuwalL, as well as a few developed counLrles llke Pungary and korea. Cur crlLerlon was LhaL a) exLernal debL daLa was avallable, and b) Lhey are source counLrles: LhaL ls, Lhose counLrles whose prlvaLe clLlzens have accumulaLed poslLlve neL unreporLed/ unLaxed flnanclal wealLh abroad.
6 !ames S. Penry, 1!n 2012 Lhemselves lnLo bankrupLcy, endurlng agonlzlng sLrucLural ad[usLmenL and low growLh, and holdlng flre sales of publlc asseLs.
1hese same source counLrles had aggregaLe gross exLernal debL of 54.08 tr||||on ln 2010. Powever, once we subLracL Lhese counLrles forelgn reserves, mosL of whlch are lnvesLed ln llrsL World securlLles, Lhelr aggregaLe neL exLernal debLs were m|nus 52.8 tr||||on ln 2010. (1hls dramaLlc plcLure has been lncreaslng sLeadlly slnce 1998, Lhe year when Lhe exLernal debLs mlnus forelgn reserves was aL lLs peak for Lhese 139 counLrles, aL +$1.43 Lrllllon. 4 )
So ln LoLal, by way of Lhe offshore sysLem, Lhese supposedly lndebLed source counLrles lncludlng all key developlng counLrles are not debtors at a||: they are net |enders, to the tune of 510.1 to 513.1 tr||||on aL end-2010.
1he problem here ls LhaL Lhe asseLs of Lhese counLrles are held by a small number of wealLhy lndlvlduals whlle Lhe debLs are shouldered by Lhe ordlnary people of Lhese counLrles Lhrough Lhelr governmenLs.
As a u.S. lederal 8eserve offlclal observed back ln Lhe 1980s: 1he real problem ls noL LhaL Lhese counLrles don'L have any asseLs. 1he problem ls, Lhey're all ln Mlaml (and, he mlghL have added, new ?ork, London, Ceneva, Zurlch, Luxembourg, Slngapore, and Pong kong)
1hese prlvaLe unrecorded offshore asseLs and Lhe publlc debLs are lnLlmaLely llnked, hlsLorlcally speaklng: Lhe dramaLlc lncrease ln unrecorded caplLal ouLflows (and Lhe prlvaLe demand for llrsL World currency and oLher asseLs) ln Lhe 1970s and 1980s was poslLlvely correlaLed wlLh a surge ln llrsL World loans Lo developlng counLrles: much of Lhls borrowlng lefL Lhese counLrles under Lhe Lable wlLhln monLhs, and even weeks, of belng dlsbursed. 3
1oday, local ellLes conLlnue Lo voLe wlLh Lhelr flnanclal feeL whlle Lhelr publlc secLors borrow heavlly abroad buL lL ls llrsL World counLrles LhaL are dolng mosL of Lhe borrowlng. lL ls Lhese frequenLly heavlly lndebLed source counLrles and Lhelr ellLes LhaL have become Lhelr flnanclers.
ln Lerms of Lackllng poverLy, lL ls hard Lo lmaglne a more presslng global lssue Lo address.
4 World 8ank/lMl daLa (2012), my analysls. 3 See !ames Penry, 3.' >())& >",?'4#1 $"('# *4)+ $.' 2()@"( 5,&'424)5,& 'A),)+-1 CcLober 2003, for an lnvesLlgaLlve lnslde look aL Lhe lnLesLlnes of Lhe global prlvaLe banklng model ln developlng counLrles.
7 !ames S. Penry, 1!n 2012 3. now th|s wea|th |s concentrated. Much of Lhls wealLh appears Lo be concenLraLed ln Lhe hands of prlvaLe ellLes LhaL reslde ln a handful of source counLrles many of whlch are sLlll regarded offlclally as debLors.
8y our esLlmaLes, of Lhe $7.3 - $9.3 Lrllllon of offshore wealLh belonglng Lo resldenLs of Lhese 139 counLrles, the top 10 countr|es account for 61 percent and Lhe top 20 for 81 percent. (See Appendlx 3, p33 for more deLalls.)
4. Untaxed Cffshore Larn|ngs start to swamp outf|ows. Cur esLlmaLes also correcL Lhe sangulne vlew LhaL slnce new ouLflows of caplLal appear Lo have recenLly decllned from counLrles llke Mexlco and 8razll, caplLal fllghL ls no longer a problem for Lhese counLrles.
Cnce we Lake lnLo accounL Lhe growLh of large unLaxed earnlngs on accumulaLed offshore wealLh, lL Lurns ouL LhaL from 1970 Lo 2010 Lhe real value (ln $2000) of Lhese earnlngs alone may be has much as $3.7 Lrllllon equlvalenL Lo abouL 60 percenL of Lhe global LoLal unrecorded caplLal ouLflows durlng Lhls perlod. 6 lor LaLln Amerlca, Sub- Saharan Afrlca and Lhe Mlddle LasL LhaL have long hlsLorles of accumulaLlng offshore wealLh and unreporLed earnlngs abroad, Lhe raLlo ls close Lo 100 percenL or more.
8y shlfLlng aLLenLlon from flows Lo accumulaLed sLocks of forelgn wealLh, Lhls paper calls aLLenLlon Lo Lhe facL LhaL reLenLlon of lnvesLmenL earnlngs abroad can easlly become so slgnlflcanL LhaL lnlLlal ouLflows are evenLually replaced by hldden fllghL, wlLh Lhe hldden sLock of unrecorded prlvaLe wealLh generaLlng enough unreporLed lncome Lo keep lL growlng long afLer Lhe lnlLlal ouLflows have drled up.
S. Cffshore earn|ngs swamp fore|gn |nvestment. AnoLher key flndlng ls LhaL once we fully accounL for caplLal ouLflows and Lhe losL sLream of fuLure earnlngs on Lhe assoclaLed offshore lnvesLmenLs, forelgn dlrecL and equlLy lnvesLmenL flows are almosL enLlrely offseL even for some of Lhe worlds largesL reclplenLs of forelgn lnvesLmenL.
6. W|de open and eff|c|ent cap|ta| markets: how trad|t|ona| theor|es fa||ed. SLandard developmenL economlcs assumes LhaL flnanclal caplLal wlll flow predomlnanLly from caplLal-rlch hlgh-savlng rlch counLrles Lo caplLal-scarce counLrles where reLurns on lnvesLmenL are hlgher.
8uL for many counLrles Lhe global flnanclal sysLem seems Lo have enabled prlvaLe lnvesLor moLlves undersLandable ones llke asseL dlverslflcaLlon along wlLh less admlrable ones llke Lax evaslon -- Lo swamp Lhe convenLlonal Lheory. 8educlng frlcLlons ln global flnance, whlch was supposed Lo help caplLal flow ln Lo caplLal-sLarved developlng counLrles more easlly and efflclenLly, seems Lo have encouraged caplLal Lo flow ouL. 1hls ralses new quesLlons abouL how efflclenL frlcLlonless global caplLal markeLs are.
6 See Appendlx lll, p34
8 !ames S. Penry, 1!n 2012
7. 1he act|ve ro|e of pr|vate banks. Cur analysls refocuses aLLenLlon on Lhe crlLlcal, ofLen unsavory role LhaL global prlvaLe banks play. A deLalled analysls of Lhe Lop 30 lnLernaLlonal prlvaLe banks reveals LhaL aL Lhe end of 2010 Lhese 30 collecLlvely managed more Lhan 512.1 tr||||on ln cross-border lnvesLed asseLs from prlvaLe cllenLs, lncludlng vla LrusLs and foundaLlons. Conslder Lhe role of smaller banks, lnvesLmenL houses, lnsurance companles, and non-bank lnLermedlarles llke hedge funds and lndependenL money managers ln Lhe offshore cross-border markeL, plus self-managed funds, and Lhls flgure seems conslsLenL wlLh our overall offshore asseL esLlmaLes of uS$21-$32 Lrllllon.
A dlsproporLlonaLe share of Lhese asseLs were managed by ma[or global banks LhaL are well known for Lhelr role ln Lhe 2008 flnanclal crlsls, Lhelr generous governmenL ballouLs and bounLlful execuLlve compensaLlon packages. We can now add Lhls Lo Lhelr llsL of dlsLlncLlons: Lhey are key players ln many havens around Lhe globe, and key enablers of Lhe global Lax ln[usLlce sysLem.
lL ls lnLeresLlng Lo noLe LhaL desplLe choppy markeLs Lhe rank order aL Lhe Lop of Lhe prlvaLe banklng world has been remarkably sLable key recenL Lrends have been for an lncreased role for lndependenL bouLlque money managers and hedge funds, and a shlfL Loward banks wlLh a sLrong Aslan presence.
8. Cffshore Investor ortfo||os. 8ased on a slmple model of offshore lnvesLor porLfollo behavlor, daLa from Lhe 8ank for lnLernaLlonal SeLLlemenLs (8lS), and lnLervlews wlLh prlvaLe bankers and wealLh lndusLry analysLs, Lhls ylelds a scale-up facLor LhaL ls also conslsLenL wlLh Lhe aggregaLe range for 2010 noLed earller.
A slmple model, based on a comblnaLlon of 8lS daLa on cross-border deposlLs and oLher asseL holdlngs by non-bank lnvesLors, an analysls of porLfollo mlx assumpLlons made by wealLh lndusLry analysLs, and lnLervlews wlLh acLual prlvaLe banks, suggesLs an overall mulLlpller of 3.0 Lo scale up our cross-border deposlLs flgure Lo LoLal flnanclal asseLs. 1hls ls very conservaLlve.
9. New kevenue Sources for G|oba| Needs. llnally, lf we could flgure ouL how Lo Lax all Lhls offshore wealLh wlLhouL kllllng Lhe proverblal Colden Coose, or aL leasL enLlce lLs owners Lo relnvesL lL back home, Lhls secLor of Lhe global underground ls also easlly large enough Lo make a slgnlflcanL conLrlbuLlon Lo Lax [usLlce, lnvesLmenL, and paylng Lhe cosLs of global problems llke cllmaLe change.
10. Cther est|mates. ln complllng Lhe evldence for Lhls paper, weve had a chance Lo examlne oLher recenL work by analysLs. We flnd a number of shorLcomlngs, parLlcularly ln meLhods LhaL rely heavlly on sLudles of lnLra-company Lransfer prlclng. SecLlon 4, below, explores Lhls ln more deLall.
9 !ames S. Penry, 1!n 2012 2. WnLkL IS CIISnCkL, ANWA?
Slnce Lhe laLe 1970s, lnvesLlgaLlve [ournallsLs, Lax auLhorlLles, drug enforcemenL offlclals, LerrorlsL Lrackers, and naLlonal securlLy experLs -- and a few economlsLs -- have gradually become aware LhaL Lhere ls lndeed a vasL deal of money -- a large and growlng chunk of Lhe worlds prlvaLe wealLh and lncome -- hldden ouL Lhere, noL so much ln Lhe land, buL offshore, proLecLed by a hlghly-pald, lndusLrlous bevy of professlonal enablers ln Lhe prlvaLe banklng, legal, accounLlng, and lnvesLmenL lndusLrles, Laklng advanLage of Lhe lncreaslngly borderless, frlcLlonless global economy.
Crade-school geography condlLlons us Lo Lhlnk of offshore as a physlcal locaLlon. lndeed, some resldenLlal havens llke Slngapore and SwlLzerland do speclallze ln provldlng secure low-Lax physlcal resldences Lo Lhe worlds wealLhlesL people, along wlLh expenslve prlvaLe schools, hosplLals, and resorLs Lo enhance Lhe famlly dynasLles human caplLal, and hlghly secure sLorage faclllLles for prlvaLe collecLlons of arL, gold, [ewels, classlc cars, yachLs, planes, weapons and oLher LrlnkeLs.
Powever, prlvaLe banklng has long slnce become vlrLual. So Lhe Lerm offshore refers noL so much Lo Lhe acLual physlcal locaLlon of prlvaLe asseLs or llablllLles, buL Lo nomlnal, hyper-porLable, mulLl-[urlsdlcLlonal, ofLen qulLe Lemporary locaLlons of ,'$B)4?# )* ('2"( ",& C5"#/D('2"( ',$/$/'# ",& "44",2'+',$# LhaL manage and conLrol prlvaLe wealLh -- always ln Lhe lnLeresLs of Lhose who manage lL, supposedly ln Lhe lnLeresLs of lLs beneflclal owners, and ofLen ln lndlfference or ouLrlghL deflance of Lhe lnLeresLs and laws of mulLlple naLlon sLaLes. A palnLlng or a bank accounL may be locaLed lnslde SwlLzerlands borders, buL Lhe all-lmporLanL legal sLrucLure LhaL owns lL Lyplcally LhaL asseL would be owned by an anonymous offshore company ln one [urlsdlcLlon, whlch ls ln Lurn owned by a LrusL ln anoLher [urlsdlcLlon, whose LrusLees are ln yeL anoLher [urlsdlcLlon (and LhaL ls one of Lhe slmplesL offshore sLrucLures) ls llkely Lo be fragmenLed ln many pleces around Lhe globe.
ulLlmaLely, Lhen, Lhe Lerm offshore refers Lo a seL of capablllLles. 1he key cllenLs for Lhe offshore sysLem lnclude Lhe worlds wealLhlesL lndlvlduals and companles, as well as lLs worsL vlllalns. numberlng [usL a few mllllon of Lhe worlds 6.3 bllllon people, Lhey are an lncredlbly dlverse group, from 30-year old Chlnese real esLaLe speculaLors and Slllcon-valley sofLware Lycoons Lo uubal oll shelks, 8usslan resldenLs, mlneral-rlch Afrlcan dlcLaLors and Mexlcan drug lords.
10 !ames S. Penry, 1!n 2012
lrom a sllghL dlsLance, all Lhese players share Lhe same baslc needs: (1) anonymlLy for Lhem, Lhelr famllles, and Lhelr buslness and pollLlcal deallngs, (2) Lhe ablllLy Lo mlnlmlze Lhe neL presenL value of fuLure Laxes, neL of Lax avoldance cosLs, (3) lnvesLmenL managemenL, for Lhose who sLlll belleve ln lL, (4) ablllLy Lo easlly access and manage Lhelr wealLh from anywhere on Lhe planeL, (3) secure places Lo hang ouL, hlde ouL and en[oy llfe, and (6) lron-clad flnanclal securlLy for Lhelr huge sLocks of anonymously- owned, largely-unLaxed prlvaLe asseLs1 agalnsL Lhe conLlnulng LhreaLs posed noL only by Lax men and prosecuLors, buL also by kldnappers, exLorLlonlsLs, sples, hlL men, con men, hackers, paparazzl, pollLlcal opponenLs, dlsgrunLled famlly members, ex-wlves, ex- lovers, and each oLher.
lL ls Lhese core capablllLles secrecy, Lax mlnlmlzaLlon, access, asseL managemenL, and securlLy LhaL our modern offshore sysLem offers. ln Lhe lasL 30 years a sophlsLlcaLed LransnaLlonal prlvaLe lnfrasLrucLure of servlce provlders has grown up Lo dellver Lhese servlces on an unprecedenLed scale. 1hls plraLe banklng sysLem now launders, shelLers, manages and lf necessary re-domlclles Lhe rlches of many of Lhe worlds worsL vlllalns, as well as Lhe Langlble and lnLanglble asseLs and llablllLles of many of our wealLhlesL lndlvlduals, alongslde our mosL successful malnsLream banks, corporaLlons, shlpplng companles, lnsurance companles, accounLlng flrms and law flrms.
11 !ames S. Penry, 1!n 2012 All Lhese players have become, as lL were, clLlzens of a brave new vlrLual counLry -- one LhaL lack physlcal boundarles buL can sLlll offer escape rouLes from many of Lhe Laxes, flnanclal regulaLlons, human rlghLs sLandards, and moral resLralnLs LhaL Lhe resL of us Lake for granLed: Lhe responslblllLles of socleLy. Cne seL of rules for a Llny mlnorlLy of rlch and powerful people, anoLher seL for everyone else.
1he dlsLurblng reallLy ls LhaL llLLle of Lhls analysls ls new: crlLlcs llke me have been dlscusslng Lhls sLrucLural defecL ln Lhe world economy and developmenL flnance slnce aL leasL Lhe 1980s. 7 1he 1ax !usLlce neLwork has been Lalklng abouL lL for a decade. Slnce Lhe recenL flnanclal crlsls began ln laLe 2007 world leaders have pald more aLLenLlon -- aL leasL rheLorlcally. AL a summlL ln London Lhe C20 declared on Aprll 2, 2009 LhaL Lhe era of bank secrecy ls over, endorsed a new Loolbox of measures Lo be used agalnsL [urlsdlcLlons LhaL fall Lo comply wlLh lnLernaLlonal sLandards, and promlsed Lo develop proposals Lo help developlng counLrles secure Lhe beneflLs of a new cooperaLlve Lax envlronmenL by Lhe end of 2009.
lL should have been a warnlng Lo us all LhaL Lhe blackllsL of Lax havens produced by Lhe CLCu, whlch was supposed Lo be ln Lhe fronLllne of Lhe global flghL agalnsL Lax haven secrecy, was empLy on Aprll 7, 2009 [usL */!' &"-# afLer LhaL dramaLlc C20 sLaLemenL. lL remalns empLy. 1he Lax havens are now supposedly clean. Meanwhlle, Lhe prlvaLe banklng operaLlons of global banks remaln among Lhelr mosL proflLable dlvlslons.
SubsequenL C20s seem Lo have obsessed wlLh debL burdens creaLed by Lhe crlsls and seem Lo have losL lnLeresL ln cracklng down on havens: and leaders ln Lhe uk, Canada, and Lhe uS have used Lhe crlsls Lo make Lhe case for cuLLlng Laxes sLlll farLher. When lL brlefly appeared ln 2010 LhaL Lhe crlsls was ebblng, conservaLlve leaders also argued LhaL Lhe Llme for anLl-haven hysLerla was over.
Cverall, Lherefore, Lhe lesson for haven reform ls LhaL we should noL lean Loo heavlly on cycllcal moods. 1he haven sysLem played a slgnlflcanL role ln aggravaLlng our laLesL, malnly llrsL World crlsls, by faclllLaLlng badly under-regulaLed cross-border lendlng, hedge funds, and lnsurance. 8uL Lhe serlous harms LhaL Lhe offshore sysLem creaLes been around for decades, chronlcally affecLlng many of Lhe worlds pooresL counLrles, and over Llme undermlnlng Lax [usLlce ln rlch and poor counLrles allke.
7 See, for example, by Lhls auLhor: noncompllance wlLh uS 1ax Law, 13-111, ln lncome 1ax Compllance. A 8eporL of Lhe A8A SecLlon of 1axaLlon.(8esLon, vlrglnla: A8A, 1983), 1he uebL Poax, 1he new 8epubllc, Aprll 1986, Amerlca Lhe 1ax Paven, 1he WashlngLon osL, !anuary 29, 1989, 8anqueros y Lavadolares. (423 pp, 8ogoLa: 1ercer Mundo LdlLores, 1996), !ames S. Penry, 1he 8lood 8ankers. (2003: 8aslc 8ooks, 2003), 1he MyLh of uebL 8ellef, ln SLeve PlaLL, ed., A Came as Cld as Lmplre. (Sl: 8arreLL koehler, 2007), lraLe 8ankers (2012, forLhcomlng). See also !ohn ChrlsLensen, ulrLy Money: lnslde Lhe SecreL World of Cffshore 8ank, 41-67, ln SLeve PlaLL, ed., A Came as Cld as Lmplre. (Sl: 8arreLL koehler, 2007), nlcholas Shaxson, 1reasure lslands. (uk: algrave MacMllllan, 2011), and !ames 8oyce and Leonce ndlkumana (2012).
Clven Lhls vlrLual geography perspecLlve, lL ls lmporLanL Lo emphaslze several sLrucLural facLs abouL Lhe offshore lndusLry, as we work on our esLlmaLes.
I|rst, lL ls lmporLanL Lo dlsLlngulsh beLween Lhe lnLermedlary havens whlch acL as condulLs for wealLh and desLlnaLlon havens where prlvaLe wealLh ulLlmaLely ends up.
We Lyplcally assoclaLe offshore legal enLlLles llke shell companles, asseL proLecLlon LrusLs, capLlve lnsurance companles, and haven banks wlLh Lhe convenLlonal llsL of offshore havens (or 1reasure lslands) found on, say, early 2000s CLCu blackllsLs: sulLry, dodgy Lroplcal lslands llke 8ermuda, Lhe Cayman lslands, nauru, SL. klLLs, AnLlgua, and 1orLola, or Lhe Luropean bolL holes such as SwlLzerland, Lhe Channel lslands, Monaco, Cyprus, ClbralLar, and LlechLensLeln. 1hese 80-odd fronL-llne havens, mosL of whlch are offshore by anyones deflnlLlon, collecLlvely provlde a home Lo over 60 mllllon people, and over 3.3 mllllon paper companles, Lhousands of shell banks and lnsurance companles, more Lhan half of Lhe worlds reglsLered commerclal shlps above 100 Lons, and Lens of Lhousands of shell subsldlarles for Lhe worlds largesL banks, accounLlng flrms, and energy, sofLware, drug, and defense companles.
ln Lhe 1970s-90s, as mulLlnaLlonal corporaLlons (MnCs), banks, lnvesLors, and a varleLy of llrsL and 1hlrd World scalawags demanded haven servlces, Lhe ellLes ln Lhese Llny ersaLz sLaLes dlscovered Lhey could make a darn good llvlng slmply by Lurnlng a bllnd eye. 1helr numbers roughly Lrlpled durlng Lhese years.
Powever, as Lhe 1ax !usLlce neLwork has recenLly emphaslzed ln lLs work on lLs llnanclal Secrecy lndex slnce 2009, Lhls convenLlonal llsL of havens ls mlsleadlng, lf were lnLeresLed ln flndlng Lhe money. lor whlle Lhere are mllllons of companles and Lhousands of Lhlnly caplLallzed banks ln Lhese flscal paradlses, few wealLhy people wanL Lo depend on Lhem Lo manage and secure Lhelr wealLh. 1hese sLealLhy lnvesLors ulLlmaLely need access Lo all Lhe prlmary beneflLs of hlgh-cosL llrsL World caplLal markeLs -- relaLlvely efflclenL, regulaLed securlLles markeLs, banks backsLopped by large populaLlons of Laxpayers, and lnsurance companles, well-developed legal codes, compeLenL aLLorneys, lndependenL [udlclarles, and Lhe rule of law. generally, Lhese can only be found ln a handful of so-called llrsL World counLrles llke Lhe uS, Lhe uk, SwlLzerland, Lhe neLherlands, 8elglum/Luxembourg, and Cermany. So we have Lo look Lo Lhese desLlnaLlon havens ln order Lo geL a handle on Lhe slze and growLh of unrecorded cross-border prlvaLe wealLh.
Second, Lhe prlvaLe enablers play a crlLlcal role ln Lhls markeL, one LhaL cuLs across lndlvldual havens. lnvesLlng and securlng large amounLs of prlvaLe wealLh across borders ls complex, requlrlng speclallzed skllls ln Lax, flnanclal plannlng, banklng, enLlLy sLrucLurlng, and esLaLe plannlng. 1hls ls noL someLhlng LhaL mosL wealLhy people underLake on Lhelr own. As noLed, Lherefore, a global servlces lndusLry of law flrms,
13 !ames S. Penry, 1!n 2012 accounLanLs, lnsurance companles, and especlally prlvaLe banks has grown up Lo caLer Lo Lhls cross-border markeL.
Whlle lL has Lhousands of players, Lhe room aL Lhe Lop ls surprlslngly llmlLed global accounLlng ls sLlll domlnaLed by Lhe 8lg lour, whlle a small number of caplLal clLy and haven-based law flrms domlnanL Lhe lawyerlng, and global prlvaLe banklng ls domlnaLed by less Lhan 30 mulLlnaLlonal banks. lor our esLlmaLes Lhls ls qulLe helpful, because lL ylelds anoLher meLrlc LhaL can be used Lo LrlangulaLe on Lhe slze of Lhe offshore markeL.
1h|rd, anoLher key developmenL slnce Lhe laLe 1990s ls Lhe growLh of Lhe onshore- offshore markeL for secrecy and Lax avoldance, especlally ln Lhe unlLed SLaLes. lrom uelaware Lo Alaska, nevada, and SouLh uakoLa, a growlng number of sLaLes are offerlng lnexpenslve legal enLlLles llke llmlLed llablllLy corporaLlons and asseL proLecLlon LrusLs whose levels of secrecy, proLecLlon agalnsL credlLors, and Lax advanLages rlval Lhose of Lhe worlds LradlLlonal secreLlve offshore havens. 1he wldespread use of Lhe llkes of Lhe nevada LLC or Lhe uelaware asseL proLecLlon LrusL, ln Lhe supposedly onshore unlLed SLaLes, furLher undermlnes Lhe LradlLlonal assoclaLlon of offshore wlLh parLlcular physlcal locaLlons, and underscores Lhe facL LhaL Lhe A4)##D@)4&'4 *()B# examlned ln Lhls paper may be [usL Lhe Llp of Lhe lceberg.
14 !ames S. Penry, 1!n 2012 4. CLD LS1IMA1LS
1he n|story and o||t|cs of Lst|mat|on
As Lord kelvln, Lhe dlscoverer of absoluLe zero on Lhe LemperaLure scale, once noLed, lf you cannoL express lL ln numbers, your knowledge ls of a meager and unsaLlsfacLory klnd.
Slnce Lhe laLe 1970s, a growlng number of economlsLs have acknowledged Lhe exlsLence of Lhe subLerranean economy and have begun Lo use a varleLy of meLhods Lo esLlmaLe lLs slze, growLh raLe, and composlLlon more preclsely.
1he flrsL phase of Lhls esLlmaLlon work, ln whlch Lhe auLhor was deeply lnvolved, ls provlded ln Appendlx 1: 3.' E4'DF/#$)4- )* 9**#.)4' G#$/+"$'#6 1hls aLLempLed Lo ldenLlfy Lhe slze and growLh of Lhe &)+'#$/A underground economles ln leadlng llrsL World counLrles by analyzlng ",)+"(/'# ln Lhe demand for currency and oLher moneLary aggregaLes for example, currency sLocks ouLsLandlng LhaL were wlldly dlsproporLlonaLe Lo LransacLlon demand, or welrd lnLer-reglonal currency flows ln Lhe uS lederal 8eserve SysLem. 8
1hls lnlLlal phase of economlc research on Lhe underground economy was very successful ln ldenLlfylng Lhe facL LhaL Lhls secLor was generally large, vlbranL and growlng, relaLlve Lo above-ground economlc acLlvlLy ln many counLrles. lL led Lo a brlef perlod ln whlch offlclal lnsLlLuLlons llke Lhe uS Congress, Lhe 1reasury, and Lhe lederal 8eserve collaboraLed acLlvely wlLh Lhe auLhor and oLher researchers, provldlng access Lo lnLernal daLa, venues for LesLlmony, and even supporL for new regulaLlons llke Lhe u.S. 1reasurys new reporLlng requlremenLs for uS currency Lurned lnLo banks, adopLed qulckly ln 1977-79 afLer Lhe auLhor and oLher analysLs dlscovered LhaL banks ln llorlda and 1exas near Lhe u.S. souLhern border wlLh Mexlco were recelvlng an lnordlnaLe amounL of $100 bllls.
1hls early research led us Lo noLlce LhaL conLrary Lo Lhe baslc assumpLlons of developmenL economlcs Lhere were very large gross and neL flows from Lhe developlng world Lo CLCu counLrles: noL only ln Lhe form of demand for reserve asseLs llke currency and gold, buL also for ordlnary flnanclal asseLs.
unforLunaLely, however, once we opened Lhe doors on Lhese prevlously-hldden hldden cross-border flows of bank deposlLs and oLher asseLs ln Lhe early 1980s Lhe uS 1reasury, Lhe lederal 8eserve, Lhe uS Congress, oLher WesLern governmenLs and Lhe CLCu, as
8 See Appendlx l Lo Lhls paper: 1he re-PlsLory of Cffshore LsLlmaLes.
13 !ames S. Penry, 1!n 2012 well as Lhe global blg bank lobby, suddenly became much less cooperaLlve. 1haL remalns Lhe slLuaLlon Loday. 9
WhaL we had uncovered was Lhe exlsLence of a hlghly lucraLlve banklng buslness LhaL had prevlously noL been dlsclosed ln any banks annual reporLs, leL alone ln 1reasury or lederal 8eserve daLa bases or Congresslonal lnqulrles even Lhough Lhls had acLually become Lhe blg banks +)#$ (5A4"$/!' (hlghesL rlsk-ad[usLed 8CL) buslness @- *"46
1hls was Lhe offshore buslness of lnLernaLlonal prlvaLe banklng, whose core mlsslon baslcally conslsLed of havlng rellable, secure, Lop-Ller, Loo blg Lo fall llrsL World banks enLlce Lhe ellLes of rlch and poor counLrles allke Lo shelLer Lhelr wealLh Lax-free offshore, usually ln conLravenLlon of Lhese home counLrles laws, ln many cases whlle lendlng heavlly Lo Lhe governmenLs and banks of Lhose very same source counLrles.
lL soon became clear Lo Lhls auLhor LhaL ln sLrlklng conLrasL Lo Lhe slLuaLlon wlLh respecL Lo currency demand or even Lhe quesLlon of where loans Lo developlng counLrles wenL -- securlng any dlrecL evldence on whlch counLrles generaLed Lhe largesL caplLal ouLflows, where prlvaLe fllghL caplLal ended up, and how much lL was worLh, would be almosL lmposslble wlLhouL a comblnaLlon of deLalled case-by-case lnvesLlgaLlons and laborlous lndlrecL daLa LrlangulaLlon.
1haL was 26 years ago. Slnce Lhen Lhe global offshore lndusLry has more Lhan quadrupled ln slze.
1he M|ss|ng Data.
Llke Lhe labyrlnLh of Lhe mlnoLaur, Lhe secreLs of Lhe offshore lndusLry have many levels of proLecLlon. llrsL, of course, prlvaLe bankers, haven lawyers and accounLanLs geL pald handsomely Lo hlde Lhelr cllenLs asseLs, ldenLlLles, and even behavloral paLLerns. CollecLlvely, Lhey also malnLaln lnfluenLlal lobbles.
Second, bank regulaLors and cenLral banks of mosL lndlvldual counLrles Lyplcally vlew prlvaLe banks as key cllenLs. So Lhey have long permlLLed Lhe worlds Lop havens and banks Lo conceal Lhe ulLlmaLe orlglns and ownershlp of asseLs under Lhelr supervlslon, especlally Lhose held ln off-balance sheeL LrusLs and flduclary accounLs.
1hlrd, even Lhough mulLllaLeral lnsLlLuLlons llke Lhe 8ank for lnLernaLlonal SeLLlemenLs (8lS), Lhe lMl, Lhe World 8ank, and Lhe CLCu LhaL are supposed Lo be somewhaL lnsulaLed from Lhe pollLlcal fray, Lhey have been hlghly senslLlve Lo Lhe collecLlve lnLeresLs of Wall SLreeL & Co. 1hey have never been wllllng Lo requlre flnanclal
9 1o aLLracL forelgn caplLal, mosL llrsL World counLrles malnLaln elaboraLe lncome and esLaLe Lax preferences for so-called non-resldenL allens, and also for non-domlclled forelgners ln Lhe case of Lhe uk and SwlLzerland, who are allowed Lo reslde whlle paylng very low lncome and wealLh Laxes.
16 !ames S. Penry, 1!n 2012 lnsLlLuLlons Lo fully reporL Lhelr cross-border cusLomer llablllLles, deposlLs, cusLomer asseLs under managemenL or under cusLody, by secLor and counLry of orlgln, even aL an aggregaLe counLry level. 10
lnLeresLlngly, however, Lhe 8lS &)'# gaLher such palr-wlse deLall for quarLerly cross- border bank asseLs and loans, because lL conslders Lhls useful for monlLorlng Lhe sLablllLy of global banklng sysLem. lL would Lechnlcally be qulLe easy Lo collecL Lhe same daLa on Lhe llablllLy slde, and Lo exLend lL Lo off-balance sheeL cusLomer asseLs and llablllLles under managemenL, admlnlsLraLlon, and cusLody.
8uL apparenLly Lhe 8lS conslders Lhe sLablllLy of developlng-counLry flnances a lower prlorlLy.
A deLalled summary of Lhe holes ln exlsLlng offlclal daLa wlLh respecL Lo slzlng Lhe offshore lndusLry ls provlded ln Appendlx lll. 11
C|d Lst|mates Cap|ta| I||ght
Slnce aL leasL 1984, macroeconomlsLs have Lolled Lo reflne Lhe meLhods used Lo esLlmaLe caplLal fllghL and Lo exLend lL Lo a growlng number of developlng counLrles. 12
Cddly, Lwo of Lhe earllesL sLudles of caplLal fllghL were by Lwo lMl economlsLs 13 and by economlsLs aL Morgan CuaranLy 1rusL, Lhe precursor Lo Lodays !MorganChase a long-Llme key player ln global prlvaLe banklng. 14 unforLunaLely, however, none of Lhese sLudles focused on slzlng or locaLlng Lhe sLock of offshore wealLh resulLlng from all Lhls caplLal fllghL, or on esLlmaLlng Lhe slze of Lhe offshore lndusLry. SLlll, Lhe consensus of more Lhan a dozen such sLudles ls LhaL hundreds of bllllons of dollars fled Lhe developlng
10 1he 8ank for lnLernaLlonal SeLLlemenLs quarLerly and annual reporLs do conLaln deLall on @",? ()",# ",& )$.'4 @",? "##'$# by counLry of desLlnaLlon, buL noL bank llablllLles or cllenL asseLs under managemenL. See 8lS, lnLernaLlonal 8anklng SLaLlsLlcs, CuarLerly reporLs, aL hLLp://www.bls.org/sLaLlsLlcs/rppb0910.hLm. 11 lor example, see Appendlx lll, p9, for Lhe holes ln Lhe 8lS daLa. 12 See, for example, Cerald LpsLeln (2003) LdlLor, CaplLal lllghL and CaplLal ConLrols ln ueveloplng CounLrles. (AmhersL: Ldward Llgar, 2003), Lhe mosL comprehenslve esLlmaLes unLll Lhls book. See also !ames 8oyce and Leonce ndlkumana (2001) ls Afrlca a neL uebLor. new LsLlmaLes of CaplLal lllghL lrom Several Severely-lndebLed Sub-Saharan Afrlcan CounLrles, 1970-98, !ournal of uevelopmenL SLudles, vol. 38, no. 2, pp. 27-36, llorlan kaufmann (2004) A CrlLlcal ulscusslon of Pow Lo Measure CaplLal lllghL: 1he Case of ArgenLlna, new School/uMass CraduaLe Workshop, november 2004, kevln Chang, SLl[n Claessens, and 8oberL Cumby (1997) ConcepLual and MeLhodologlcal lssues ln Lhe MeasuremenL of CaplLal lllghL, lnLernaLlonal !ournal of llnanclal Lconomlcs, vol. 2(1997), 101-11, 8enu Schnelder (2003a), Measurlng CaplLal lllghL: LsLlmaLes and lnLerpreLaLlons, Cul Worklng aper 194, Cverseas uevelopmenL lnsLlLuLe, Ldsel 8e[a !r. (2004), CaplLal lllghL from Lhe ASLAn lour: A Case SLudy on lndonesla, Malaysla, Lhe hlllpplnes, and 1halland, u Mass AmhersL. 13 See Mohsln S. khan and nadeem ul Paque, lorelgn 8orrowlng and CaplLal lllghL: A lormal Analysls, lMl SLaff apers, v. 32 (4), uecember 1983, 606-628. 14 See Morgan CuaranLy 1rusL, World llnanclal MarkeLs, LuC CaplLal fllghL, March 1986.
17 !ames S. Penry, 1!n 2012 world durlng Lhls perlod Lhough a slgnlflcanL porLlon of Lhe ouLflow were found Lo have been round-Lrlpped back lnLo Lhose counLrles, parLlcularly hlgh-growLh developlng counLrles llke Chlna.
C|d (Ad noc) Lst|mates Consu|t|ng I|rms
AbsenL dlrecL evldence on Lhe slze and growLh of offshore, ln Lhe lnLerlm, parL of Lhe esLlmaLlon gap has been fllled by a serles of lmpreclse guessLlmaLes by varlous managemenL consulLlng flrms, all wlLh large-scale banklng pracLlces.
1he flrsL of Lhese, by Merrlll Lynch/Cap Cemlnl ln 1997, relled on rough [udgmenL and a wlde varleLy of unspeclfled sources. lL esLlmaLed Lhe slze of offshore wealLh owned by hlgh-neL worLh lndlvlduals (PnWls) aL $3.3 Lrllllon ln 1996, supposedly up from [usL $2.3 Lrllllon ln 1989. 13 A year laLer Lhey esLlmaLed Lhe flgure for 1997 was $3.8 Lrllllon. AfLer LhaL, Lhey sLopped maklng speclflc offshore esLlmaLes, whlle conLlnulng Lo publlsh an annual World WealLh 8eporL LhaL provldes esLlmaLes of LoLal flnanclal wealLh presumably offshore and domesLlc for hlgh-neL worLh lndlvlduals.
AlLhough Merrlll Lynch, acqulred by 8ankAmerlca durlng Lhe flnanclal crlsls, has recenLly dropped ouL of wealLh esLlmaLlon, CapCemlnl conLlnues Lo publlsh lLs ad hoc esLlmaLes. lL has also been [olned by 8osLon ConsulLlng Croup (8CC), Mcklnsey, and varlous smaller servlce provlders Lo Lhe prlvaLe banklng markeL. lor example, ln 2008 Lhe Cllver Wyman Croup, a flnanclal consulLanL, esLlmaLed Lhe global offshore prlvaLe wealLh owned by hlgh-neL worLh lndlvlduals aL $8 Lrllllon as of 2007. 16 ln !une 2009 8CC esLlmaLed Lhe volume of llquld offshore asseLs under managemenL ln 2008 aL $6.7 Lrllllon, down from a peak of $7.2 Lrllllon ln 2007. 17 Scorplo arLners has also enLered Lhe fray wlLh a long serles of ad hoc esLlmaLes of lLs own.
A close look aL Lhese consulLlng sLudles flnds many glarlng omlsslons. lor example, 8CCs 2009 esLlmaLes noL only lefL ouL all of Afrlca, buL all non-flnanclal offshore asseLs held by LrusLs and foundaLlons, whlch are Lyplcally a very large share of PnWl porLfollos. 18 lndeed, as well see, Lhe acLual slze of offshore flnanclal wealLh ls far hlgher Lhan any of Lhese consulLanLs clalm.
13 See Merrlll Lynch/ Cemlnl ConsulLlng, World WealLh 8eporL (1997), avallable aL hLLp://dss2.prlnceLon.edu/daLa/3161/. 16 Cllver Wyman LLd (2008), 1he luLure of rlvaLe 8anklng. A WealLh CpporLunlLy? aL www.mmc.com/knowledgecenLer/CllverWymanluLurerlvaLe8anklng.pdf6 17 8osLon ConsulLlng Croup (SepL. 13,2009), Clobal WealLh 8eporL: uellverlng on Lhe CllenL romlse, summary aL hLLp://www.bcg.com/Medla/ress8eleaseueLalls.aspx?ld=Lcm:12-28981 18 See 1!n, Cn sLrong assumpLlons and oLher nonsense, !une 24, 2009, aL hLLp://Lax[usLlce.blogspoL.com/2009/06/on-maklng-sLrong-assumpLlons-and-oLher.hLml
18 !ames S. Penry, 1!n 2012 C|d Lst|mates Governments and NGCs
Clven Lhe lmporLance of Lhe economlc developmenL and Lax pollcy lssues aL sLake here, one mlghL have LhoughL LhaL Lhe lMl and Lhe World 8ank would have devoLed aL leasL a few of Lhelr Lhousands of economlsL-person years Lo sLudylng Lhls lssue.
1o daLe, however, we can only flnd one expllclL lMl esLlmaLe of Lhe sLock of offshore prlvaLe asseLs -- a modesL $1.7 Lrllllon flgure for 2000, wlLh no clear meLhodology. 19 ln 2001 a former lMl ulrecLor of llscal Affalrs dld clLe a $3-$7 Lrllllon flgure for Lhe volume of offshore prlvaLe asseLs, buL decllned Lo offer any more deLalls. 20
ln 2000 Lhe uS SLaLe ueparLmenLs 8ureau for lnLernaLlonal narcoLlcs esLlmaLed Lhe asseLs of 30 offshore [urlsdlcLlons aL $4.8 Lrllllon. 21 1hen Lhe CLCu, whlch has supposedly been worklng on harmful Lax compeLlLlon by offshore havens slnce 1998, 22
also ln 2007 endorsed a $3-$7 Lrllllon range. 23 Around Lhe Llme of Lhe Aprll 2009 C20 conference ln London, when global anLl-haven senLlmenLs were raglng, Lhe CLCu Lemporarlly boosLed lLs esLlmaLe Lo $11 Lrllllon, wlLhouL offerlng a clear meLhodology. 24
8uL by SepLember 2009 Lhe CLCu had reLreaLed Lo summarlzlng a wlde range of esLlmaLes produced by oLhers. 23
As for nCCs, slnce Lhe early 2000s Lhere has been a rlslng crescendo of esLlmaLes relaLed Lo Lhe slze of offshore.
! ln 2000 Cxfam publlshed a rough $6 Lo $7 Lrllllon esLlmaLe for all prlvaLe offshore wealLh, wlLhouL any furLher deLalls. 26
! ln 2003, a qulck-and-dlrLy analysls by Lhe 1ax !usLlce neLwork (1!n,) enLlLled 1he rlce of Cffshore, esLlmaLed Lhe value offshore prlvaLe wealLh aL $11.3 Lrllllon as of !une
19 lMl ubllshlng Clobal orLfollo lnvesLmenL Survey (2000), aL hLLp://www.lmf.org/exLernal/np/sec/nb/2000/nb0008.hLm. 20 vlLo 1anzl, AusLrallan naLlonal 8adlo, March 21 2001, aL hLLp://www.abc.neL.au/rn/Lalks/bblng/sLorles/s263977.hLm. 21 lnLernaLlonal narcoLlcs ConLrol SLraLegy 8eporL, u.S. ueparLmenL of SLaLe 8ureau for lnLernaLlonal narcoLlcs and Law LnforcemenL Affalrs (March 2000), aL 363-66. 22 CLCu (1998), Parmful 1ax CompeLlLlon: An Lmerglng Clobal lssue, hLLp://www.oecd.org/daLaoecd/33/1/1904184.pdf. 23 !effrey . Cwens, ulrecLor, CLCu CenLer for 1ax ollcy and AdmlnlsLraLlon, Cffshore 1ax Lvaslon, !uly 20, 2007, Clobal ollcy lorum, aL hLLp://www.globalpollcy.org/componenL/conLenL/arLlcle/172/30123.hLml. 24 CLCu esLlmaLes $11 Lrllllon parked ln Lax havens, 1lmes of lndla, Aprll 4 2009, aL hLLp://www.buslness-sLandard.com/lndla/news/oecd-esLlmaLes-11-Lrllllon-parked-ln-Lax- havens/333933/ 23 See CLCu (2009), CLCus CurrenL 1ax Agenda, aL hLLp://www.oecd.org/daLaoecd/38/17/1909369.pdf1 H6 IJ6 26 Cxfam (2000), 8eleaslng Lhe Pldden 8llllons for overLy LradlcaLlon, aL hLLp://www.oxfam.org.uk/whaL_we_do/lssues/debL.../Lax_havens.hLm6
19 !ames S. Penry, 1!n 2012 2004, of whlch $9.3 Lrllllon supposedly conslsLed of offshore flnanclal asseLs and $2 Lrllllon was non-resldenLlal offshore real esLaLe. 1!n used an lnLeresLlng meLhodology LhaL we wlll reLurn Lo and expand on below. 27 unLll now, Lhls ls sLlll Lhe largesL publlshed esLlmaLe for LoLal offshore wealLh. 28
! ln 2009, the author developed esLlmaLes of Lhe volume of caplLal fllghL and Lhe value of offshore flnanclal asseLs derlved from leadlng developlng counLrles for Cxfam C8, a documenL LhaL was dlsLrlbuLed aL Lhe Aprll 2009 C20 summlL ln London, albelL raLher quleLly. 29
1haL sLudy employed a verslon of Lhe convenLlonal sources and uses meLhods used by Lhe auLhor for global caplLal fllghL esLlmaLes publlshed ln hls 1996 sLudy of bankers and money launderers. 1o LhaL baslc meLhod, Lhe auLhor added Lhe noLlon of esLlmaLlng Lhe neL value of offshore wealLh accumulaLed by developlng counLrles, flrsL ouLllned aL a 1!n / AA8A Conference ln Lssex ln Lhe uk, ln 2006-2008.
1he sLudy esLlmaLed LhaL durlng Lhe perlod 1970 Lo 2007 aL leasL $130- $200 bllllon of unrecorded prlvaLe caplLal fllghL flowed ouL of Lhe developlng world each year. uslng conservaLlve assumpLlons abouL relnvesLmenL raLes and lnvesLmenL ylelds, Lhese flows suggesLed LhaL accumulaLed offshore wealLh sLock owned by developlng counLry resldenLs was worLh aL leasL $6.2 Lrllllon by 2007. 1hls lncluded only wealLh from developlng counLrles, whlch ls aL mosL 23 Lo 30 percenL of all offshore prlvaLe wealLh 30 .
1hls large flgure for developlng counLry wealLh alone lmplled LhaL $)$"( offshore prlvaLe wealLh was much hlgher Lhan oLher esLlmaLes, and lL lmplled LhaL developlng counLrles mlghL be loslng as much as $120-$160 bllllon per year ln losL Lax revenue on Lhe lnLeresL and oLher lncome generaLed by all Lhls unreporLed anonymous wealLh more Lhan Lhe enLlre global LoLal of forelgn ald from CLCu counLrles. MosL of Lhls unreporLed lncome was elLher reLalned abroad and relnvesLed or spenL on shopplng Lrlps ln arls, London or Mlaml.
lL was also lnLeresLlng Lo compare Lhls lnlLlal $6.2 Lrllllon esLlmaLe for developlng counLry fllghL wealLh wlLh Lhe $3.4 Lrllllon gross forelgn debL owed by all developlng (low- and mlddle-lncome) counLrles as of 2007. AfLer Laklng fllghL wealLh lnLo accounL,
"# $%& '()*+(,)-.(,/ 0"11234 $5* 6+-7* .8 988:5.+*4 ,) hLLp://www.Lax[usLlce.neL/cms/upload/pdf/rlce_of_Cffshore.pdf6 28 As dlscussed below, 1!n (2003) based lLs esLlmaLe on scallng up 8lS daLa on offshore deposlLs by nonbanks by an average raLlo of cash and bank deposlLs Lo all flnanclal asseLs LhaL was based on Merrlll Lynch/ CapCemlnls annual esLlmaLes of asseL allocaLlons for hlgh-neL worLh porLfollos. 29 1ax haven crackdown could dellver $120bn a year Lo flghL poverLy Cxfam, March 13, 2009, aL hLLp://www.oxfam.org.uk/appllcaLlons/blogs/pressofflce/?p=3912. 30 See Lhe reglonal breakdowns of Lhe offshore wealLh esLlmaLes by Lhe consulLlng flrms. 1helr esLlmaLes are conslsLenL wlLh Lhls markeL share for Lhe source counLrles ln our sample.
20 !ames S. Penry, 1!n 2012 Lherefore, |t became even c|earer that deve|op|ng countr|es as a who|e d|dnt rea||y face a debt prob|em, but a huge offshore tax evas|on prob|em.
Cn a neL of forelgn reserves debL basls, Lhe slLuaLlon was even more sLrlklng. AfLer allowlng for forelgn reserves owned by developlng counLrles, as of 2007, Lhe mosL recenL year for whlch Lhe daLa was Lhen avallable, low-and mlddle-lncome developlng counLrles as a group had +/,5# $413.1 bllllon of neL forelgn debL -- even before Laklng Lhls offshore fllghL wealLh lnLo accounL. 1he negaLlve number ls malnly because hlgher- growLh counLrles llke Chlna and lndla have managed Lo accumulaLe over $2.3 Lrllllon of reserves. Lven omlLLlng Chlna and lndla, by 2007 all oLher developlng counLrles owed [usL $979 bllllon of neL-of-reserves forelgn debL, compared Lo $3.3 Lrllllon of non- Chlnese offshore fllghL wealLh. 31
! 1ransfer M|spr|c|ng. AnoLher offshore acLlvlLy LhaL has recenLly been Lhe sub[ecL of lncreased esLlmaLlon efforLs by economlsLs ls corporaLe Lransfer mlsprlclng. 1hls work focuses on Lhe facL LhaL ln addlLlon Lo caplLal fllghL and offshore haven abuse by lndlvldual Laxpayers, Lhere has also been a Lrend for MnCs Lo slash Lhelr corporaLe Laxes by shlfLlng proflLs and royalLy paymenLs Lo low-Lax havens and losses and lnLeresL expense Lo hlgh-Lax [urlsdlcLlons. lor example, accordlng Lo Lhe l8S, reporL, from 1994 Lo 2004 uS companles more Lhan Lrlpled Lhelr forelgn proflLs, parklng nearly 60 percenL of lL ln Lax havens. 32
Some of Lhese Lransfer prlclng abuses have Lo do wlLh over-lnvolclng of goods lmporLs and under-lnvolclng goods exporLs, so as Lo mlnlmlze lncome ln hlgher-Lax counLrles and shlfL unreporLed proflLs abroad. CLher Lransfer prlclng abuses lnvolve Lhe (below- markeL) Lransfer of lnLellecLual properLy rlghLs (know-how, brand value, fllms, paLenLs, and sofLware) Lo low-Lax [urlsdlcLlons.
Accordlngly Clobal llnanclal lnLegrlLy (Cll) slnce 2006 have publlshed a serles of esLlmaLes of gross flnanclal ouLflows from developlng counLrles LhaL rely heavlly on esLlmaLes of goods Lransfer mlsprlclng. ln 2008 ChrlsLlan Ald esLlmaLed Lhe annual cosL of such corporaLe Lransfer mlsprlclng Lo developlng counLrles, ln Lerms of losL Lax revenues, aL $160 bllllon per year. 33 1hese losses are supposedly ln addlLlon Lo Lhe losL Lax revenue on unreporLed lndlvldual lncome, offshore lnLeresL and dlvldends due Lo unrecorded fllghL.
lor our purposes, Lhese esLlmaLes of Lransfer mlsprlclng *()B# do noL lead dlrecLly Lo esLlmaLes of Lhe slze and growLh of offshore prlvaLe wealLh #$)A?# )4 unreporLed lncome generaLed by Lhese sLocks. 1hey are sub[ecL Lo many Lechnlcal problems for example,
31 World 8ank (2009) daLa on gross exLernal debL and forelgn reserves by counLry, our analysls. 32 See uS l8S, 8emarks by Commlssloner uouglas Shulman before Lhe 21sL Annual Ceorge WashlngLon unlverslLy lnLernaLlonal 1ax Conference, uecember 8, 2008. 33 ChrlsLlan Ald (2008), ueaLh and 1axes: 1he 1rue 1oll of 1ax uodglng, aL hLLp:// www.chrlsLlanald.org.uk/caw08/8eporL_08_LxecuLlve20Summary.pdf
21 !ames S. Penry, 1!n 2012 Clls esLlmaLes perLaln only Lo flows, noL sLocks or lnvesLmenL earnlngs, Lhey leave ouL all negaLlve numbers from Lhe calculaLlons, and ln mosL cases Lhey only daLe back Lo Lhe year 2000. 34 Powever, Lhey have cerLalnly helped dramaLlze Lhls aspecL of Lhe offshore problem.
Summary: key L|m|tat|ons, Lx|st|ng Lst|mates
All Lold, Lhese early esLlmaLes of offshore acLlvlLy reflecL a growlng consensus LhaL offshore wealLh and Lhe lncome LhaL lL generaLes are slgnlflcanL and growlng, and LhaL Lhey are largely mlsslng from offlclal sLaLlsLlcs, convenLlonal esLlmaLes of lncome and wealLh lnequallLy, and mosL lmporLanL, Lhe global Lax base.
Powever, Lhe flurry of esLlmaLes produced Lo daLe have many llmlLaLlons.
Iuzzy Methods. ln many cases Lhe esLlmaLlon meLhodology has been ad hoc and someLlmes [udgmenL-based, maklng lL dlfflculL for lndependenL observers Lo repllcaLe and verlfy lL.
1echn|ca| Lrrors and Cm|ss|ons. ln some cases, whole caLegorles of offshore wealLh have been omlLLed for example, ln Lhe case of Lhe consulLlng sLudles noLed above, asseLs held ln LrusLs and foundaLlons for prlvaLe beneflclarles are omlLLed enLlrely. ln oLher cases, such as Clls Lransfer prlclng analysls, we have seen ouLrlghL lf frulLful Lechnlcal errors.
Under-Used Data. A varleLy of oLher poLenLlally relevanL, publlcally accesslble daLa Lypes have noL yeL been broughL Lo bear Lhe Lask of esLlmaLlng Lhe slze, growLh, and dlsLrlbuLlon of offshore acLlvlLy for example, cerLaln daLa on cross-border lnvesLmenLs ln 8lS reporLs, daLa on local holdlngs of currency, gold, and oLher llquld wealLh, Lhe composlLlon of prlvaLe banklng cllenL porLfollos, enabler acLlvlLles, sLafflng, and producLlvlLy by haven locaLlon, and prlvaLe banklng asseLs under managemenL, consldered below.
ropr|etary Mode||ng. arLly because of compeLlLlve behavlor among offlclal lnsLlLuLlons, academlcs and members of Lhe nCC communlLles, Lhere has been a Lendency Lo redo slmllar daLa analyses over and over agaln, wlLhouL Lhe klnd of open, collaboraLlve peer revlew LhaL mlghL be more conduclve Lo generaLlng cumulaLlve progress on models and meLhods. 1here has also been a Lendency Lo be somewhaL vague abouL key esLlmaLlon meLhods, daLa sources, and assumpLlons. llnally, Lhere has no doubL been a greaL deal of redundanL analysls of Lhe very same daLa seLs. 1hls conLrasLs sharply wlLh Lhe more open approach Lo models, esLlmaLes and daLa ln, say, envlronmenLal sclences.
34 See Lhe dlscusslon below LsLlmaLlon MeLhods.
22 !ames S. Penry, 1!n 2012
Cff|c|a| Secrecy. 1he key problem ls mlsslng daLa. Much of Lhe daLa requlred for Lhls klnd of analysls appears Lo be elLher already on hand or readlly avallable, should cenLral banks, Lreasurles, and mulLllaLeral lnsLlLuLlons llke Lhe 8lS and Lhe lMl declde Lo make lL avallable or aL leasL analyze lL lnLernally and aggregaLe lL a form LhaL ls lnLo accepLable for publlcaLlon. A key acLlon lLem ls for nCCs noL Lo menLlon developlng counLrles Lo press such lnsLlLuLlons for more Lransparency.
rocess Improvements Lst|mat|on kesearch
A key ob[ecLlve of Lhls pro[ecL, beyond generaLlng new esLlmaLes, ls Lo esLabllsh a more expllclL framework for maklng such esLlmaLes and a more open process for researchers Lo collaboraLe.
We have esLabllshed a web slLe on LsLlmaLlng Lhe rlce of Cffshore where all Lhe daLa seLs and models employed ln Lhls paper wlll be made publlcly avallable. 1hls wlll permlL researchers Lo develop varlaLlons of Lhelr own, avold redundancy, and even more lmporLanL, Lrade suggesLlons for lmprovemenLs, and share new daLa sources. 1he web page ls here. hLLp://www.Lax[usLlce.neL/cms/fronL_conLenL.php?ldcaL=148
We also plan Lo esLabllsh faclllLles for researchers ln Lhls area Lo collaboraLe.
23 !ames S. Penry, 1!n 2012 S. NLW LS1IMA1LS
Some background
endlng changes ln research collaboraLlon, we have Lrled Lo address all Lhe oLher key flaws ln exlsLlng esLlmaLes -- Lechnlcal errors and omlsslons, Lhe lack of expllclL meLhodologles, varlaLlons ln daLa sources, meLhods, and Llme perlods LhaL makes lL dlfflculL Lo compare alLernaLlve esLlmaLes, and excesslve rellance on a llmlLed range of daLa Lypes and esLlmaLlon meLhods.
Scope of the Lst|mates
AL Lhe ouLseL lL ls worLh resLaLlng our core focus, whlch ls on measurlng (),2D$'4+ 5,4'A)4&'& A4)##D@)4&'4 H4/!"$' */,",A/"( A"H/$"( *()B# ",& #$)A?# LhaL have conLrlbuLed slgnlflcanLly Lo Lhe eroslon of Lhe domesLlc Lax base, especlally ln developlng counLrles.
1here are several very dlfferenL sources for Lhese flows and sLocks LhaL ls one reason why l prefer Lo call Lhem unrecorded caplLal flows and sLocks raLher Lhan caplLal fllghL.
Cne key source ls underreporLed caplLal flows LhaL have been secreLed offshore and lnvesLed abroad beyond Lhe reach of domesLlc Lax auLhorlLles.
1hls broad deflnlLlon focuses on unrecorded caplLal flows wlLhouL pre-[udglng Lhe moLlves for lL. 33 Among Lhe posslble moLlves are (1) shorL-Lerm speculaLlon (hoL money), (2) longer-Lerm porLfollo dlverslflcaLlon, (3) asseL proLecLlon (lncludlng proLecLlon agalnsL pollLlcal rlsks and lllegallLy), and (4) more dublous moLlves, llke money launderlng, lncome Lax evaslon, round-Lrlpplng (Laklng money offshore, dresslng up ln secrecy sLrucLures Lhen preLendlng Lo be forelgn lnvesLors ln order Lo Lake advanLage of Lax breaks and exchange raLes only avallable Lo forelgners), back- Lo-back lendlng games, exporL subsldy fraud, avoldance of lmporL duLles, corrupLlon and more.
All Lhese moLlves have been aL work Lhrough Lhe perlod we are conslderlng, so Lhe besL explanaLlon ls all of Lhe above.
Powever, slnce neL ouLflows from developlng counLrles have conLlnued over susLalned perlods of Llme, and slnce llLLle offshore wealLh or Lhe earnlngs LhaL lL produces have been repaLrlaLed, Lhe mosL lmporLanL facLors drlvlng lL are noL Lhose LhaL drlve hoL money, buL long-Lerm de-caplLallzaLlon.
33 1hls ls ln conLrasL Lo oLher auLhors llke CuddlngLon (1986), who have llmlLed Lhelr aLLenLlon Lo flows LhaL are moLlvaLed by some speclflc facLor, llke shorL-Lerm speculaLlon. Slnce Lhere are so many oLher posslble moLlves for unrecorded caplLal, and slnce Lhe daLa needs so much work, lL seems more approprlaLe Lo be a llLLle more open-ended abouL Lhe measuremenL.
24 !ames S. Penry, 1!n 2012
lor Lhe lnLeresLed reader, Appendlx ll provldes my crlLlcal revlew of exlsLlng explanaLlons for caplLal fllghL ln Lhe economlcs llLeraLure. 36
As dlscussed below, our besL esLlmaLe ls LhaL leasL 23-30 percenL of Lhese funds, averaglng several hundred bllllon per year slnce Lhe1970s, have come from developlng counLrles.
Another key source ls under-Laxed corporaLe proflLs and royalLles LhaL have been parked offshore ln low-Lax havens by way of rlgged Lransfer prlclng schemes. Whlle esLlmaLes for Lhe value of such Lransfer prlclng abuses are more problemaLlc, Lhey are llkely Lo be slgnlflcanL. 37
A th|rd source ls a myrlad of llllclL acLlvlLles ln Lhe global underground economy corrupLlon, fraud, lnslder Lradlng, drug Lrafflcklng, blood dlamonds, and lnnumerable oLher for-proflL crlmes. Lven for source counLrles wlLh zero or very low lncome Laxes, llke 8ussla, Saudl Arabla, and mosL oLher Mlddle LasLern oll producers, havens provlde a convenlenL way Lo launder all Lhls llllclL looL. Whlle banks and oLher flnanclal lnLermedlarles are supposed Lo follow know your cusLomer rules LhaL prevenL Lhls klnd of chlcanery, ln pracLlce Lhe regulaLlons are full of loopholes -- raLher llke, approprlaLely enough, Swlss cheese. 38
Narrow scope, conservat|ve est|mates.
uesplLe lncludlng all Lhese sources, our focus so far ls sLlll much narrower Lhan Lhe full scope of Lhe prlce of offshore. 1here ls a long laundry llsL of economlc bads enabled by haven [urlsdlcLlons: noL only Lax evaslon buL also fraud, brlbery, lllegal gambllng, money launderlng, and Lrafflc ln conLraband: drugs, sweaLshops, human and sex Lrafflcklng, arms, Loxlc wasLe, confllcL dlamonds, endangered specles, booLlegged sofLwareLhe llsL ls vlrLually endless.
ln prlnclple, all Lhese bads deserve Lo be lncluded on Lhe soclal balance sheeL ln any overall assessmenL of Lhe offshore lndusLry. ln pracLlce, however, we wlll focus here on whaL we can geL a handlly on. And glven Lhe sheer scale of Lax evaslon faclllLaLed by offshore havens, however, lL ls clearly one of Lhe maln anchors for Lhe sysLem, whlch underplns all Lhese oLher dublous acLlvlLles.
36 See Appendlx ll: Lxplalnlng CaplLal lllghL. 37 See, for example, Lhe slgnlflcanL orlglnal research on corporaLe Lransfer prlclng abuses by ur. Slmon ak: hLLp://www.russla[ournal.com/node/18073, hLLp://porLal.acm.org/clLaLlon.cfm?ld=780978 , 38 Cf. Lhe LorLured hlsLory of Lhe quallfled lnLermedlary regulaLlons for ma[or banks ln Lhe uS: hLLp://www.hg.org/arLlcle.asp?ld=3934.
23 !ames S. Penry, 1!n 2012
We also omlL several lmporLanL Lypes of non-flnanclal wealLh LhaL are collecLlvely qulLe slzeable, as well as lmporLanL for Lax [usLlce and developmenL pollcy -- noLably human caplLal, neL clalms on real properLy (lncludlng land and naLural resources), and lnLanglbles llke clalms Lo paLenLs, Lrademarks, brands, Lechnlcal know-how, and oLher lnLellecLual properLy. ln all Lhese cases Lhe role of cross-border flows and offshore havens appears Lo have been lncreaslng: for example, recenL exposes concernlng redomlclllaLlons of lnLellecLual properLy Lo low-Lax [urlsdlcLlons by leadlng uS pharmaceuLlcal and compuLer companles, or conLlnulng concerns abouL 1hlrd World braln draln.
Do tax havens prov|de pub||c goods, as we|| as bads?
uefenders of Lhe sLaLus quo wlll argue LhaL Lhls paper only addresses Lhe prlce of offshore, noL Lhe value Lhey say lL provldes. 1hey would argue LhaL havens, for example, help people dodge noxlous governmenL rules and regulaLlons, provlde escape haLches for Lhe vlcLlms of oppresslve reglmes, help wealLhy ellLes evade conflscaLory LaxaLlon, dlverslfy Lhelr domesLlc porLfollos, en[oy Lhe frulLs of Lhelr hard labor undlsLurbed by Lhe lrrlLanLs of LaxaLlon and regulaLlon, and aL Lhe same Llme compelllng naLlon-sLaLes Lo engage ln Lax compeLlLlon, whlch (Lhey argue) force Lhem Lo become more efflclenL ln dellverlng governmenL servlces.
Such Lradeoffs are noL unknown. lor example, lL ls very hard Lo defend Lax enforcemenL aL any prlce when Lhe Lax collecLor ls Lhe 8urmese [unLa, Caddafls Llbya, or perhaps even Lhe ClLy of Chlcago.
Cur sense, however, ls LhaL mosL counLrles operaLe very, very far from Lhls hypoLheLlcal Lax compllance vs. freedom-and-prosperlLy margln, along whlch lncreased Lax compeLlLlon and reduced compllance auLomaLlcally leads Lo lncreased llberLy, enLrepreneurshlp, and growLh. We aL Lhe 1ax !usLlce neLwork feel we have, ln a varleLy of dlfferenL fora, effecLlvely demollshed preLLy much every one of Lhelr argumenLs for Lhese supposed beneflLs 39 .
1he argumenLs of Lhe defenders of Lax havens are especlally problemaLlc once we conslder Lhe facLs LhaL Lhe proceeds of non-compllance Lend Lo flow noL Lo Lhe besL and brlghLesL, buL Lo Lhe mosL unsavory, LhaL non-compllance ls conLaglous, so aggresslve non-compllance by ellLes promoLes non-compllance by everyone else excepL Lhe poor, who end up fooLlng Lhe blll, LhaL mosL of Lhe proceeds of caplLal fllghL and Lax evaslon are never repaLrlaLed Lo source counLrles buL slL ldle ln relaLlvely-low-yleld offshore lnvesLmenLs, and LhaL when Lhe publlc secLor has been sLarved for caplLal (perhaps
39 lor lnsLance, on Lhe common clalm LhaL Lax haven secrecy can help clLlzens escape un[usL despoLs and Lhelr llke, see our arLlcle 1he non-perlls of lnformaLlon exchange, !uly 2009. hLLp://Lax[usLlce.blogspoL.ch/2009/07/non-perlls-of-lnformaLlon-exchange.hLml
26 !ames S. Penry, 1!n 2012 havlng had Lo rely on hlgh-cosL loans or lnflaLlonary flnance raLher Lhan Lax revenue), Lhe raLe of reLurn on publlc lnvesLmenLs ls ofLen hlgher Lhan on prlvaLe lnvesLmenL. 1ax, by produclng beLLer roads and educaLed populaLlons and so on, can crowd ln prlvaLe lnvesLmenL, raLher Lhan crowd lL ouL as many people belleve.
Lst|mat|on Methods and Some kesu|ts
As noLed, Lo overcome Lhe llmlLaLlons of prevlous esLlmaLes, our sLraLegy has been Lo LrlangulaLe on esLlmaLes from several dlfferenL angles. As menLloned, our four key models are as follows: A sLandard verslon of Lhe sources-and-uses model for counLry-by-counLry unrecorded caplLal flows, An accumulaLed offshore wealLh model, An offshore lnvesLor porLfollo model, based on cross-border asseLs daLa, and ulrecL esLlmaLes of offshore asseLs under managemenL for Lhe worlds Lop 30 global prlvaLe banks.
1he paper supplemenLs Lhese wlLh oLher evldence, lncludlng
1. uaLa on so-called Lransfer mlsprlclng, 2. uaLa on Lhe cross-border demand for llquld maLLress money llke reserve currency and gold, parL of whlch may move Lhrough offshore markeLs 3. A revlew of markeL research by leadlng consulLlng flrms on Lhe slze of Lhe offshore prlvaLe banklng markeL
AnoLher reference polnL ls CredlL Sulsses global wealLh esLlmaLe for mld-year 2011, whlch ls probably Lhe mosL comprehenslve and mosL recenL esLlmaLe of global wealLh. lL puLs LoLal global wealLh aL $231 Lrllllon, lncludlng flnanclal asseLs and non-flnanclal asseLs (prlnclpally houslng and land) aL markeL value 40 . CredlL Sulsse does noL offer a flgure for offshore holdlngs buL Lhe raLlo of Lhls $231 Lrllllon flgure Lo 1!ns $21-32 Lrllllon flgure headllned above ls roughly 1:10, supporLlng our vlew LhaL our new esLlmaLes are reasonable and conservaLlve.
1hls secLlon of Lhe paper wlll revlew Lhe esLlmaLlon meLhods ln some deLall. 1he followlng secLlon on lllghL aLLerns summarlzes Lhe resulLs of Lhe caplLal flows and wealLh models.
40 1he CredlL Sulsse esLlmaLe uses Pousehold 8alance SheeL daLa for selecLed counLrles, comblned wlLh household lncome and expendlLure daLa, and supplemenLed by an esLlmaLlon of wealLh holdlng paLLerns wlLhln naLlons, lL also Look lnformaLlon gleaned from varlous rlch llsLs Lo flll lnformaLlon gaps on wealLh holdlngs aL Lhe Lop of Lhe dlsLrlbuLlon. lrom 7()@"( :'"($. 8'H)4$1 IKLL, CredlL Sulsse 8esearch lnsLlLuLe: hLLps://lnfocus.credlL- sulsse.com/daLa/_producL_documenLs/_shop/323323/2011_global_wealLh_reporL.pdf
27 !ames S. Penry, 1!n 2012
1he methods |n deta||.
(1) Unrecorded Cap|ta| I|ows: sources-and-uses.
1hls paper employs laLesL daLa from Lhe World 8ank/lMl, Lhe un, cenLral banks and naLlonal accounLs Lo expllclLly model caplLal flows for each counLry, for 139 key caplLal source counLrles, mosLly low-mlddle lncome counLrles, for whlch such Llme serles daLa ls publlshed.
As of 2010, Lhese counLrles ln our sample accounLed for 83 percenL of Lhe worlds 6.89 bllllon populaLlon, 31 percenL of lLs $76.7 Lrllllon -ad[usLed gross naLlonal lncome, $4.1 Lrllllon of forelgn debL, and $6.8 Lrllllon of forelgn reserves, or 73 percenL of Lhe worlds LoLal.
Data Sources. 1he speclflc varlables lncluded ln Lhe counLry models of unrecorded caplLal flows changes ln forelgn reserves, Lhe currenL accounL balance, gross exLernal debL sLocks (of more Lhan 1 year duraLlon), neL forelgn lnvesLmenL (8C basls), porLfollo lnvesLmenL (neL of coverlng LransacLlons), nomlnal gross naLlonal lncome, prlce deflaLors are all annual daLa, lnlLlally ln uS dollars. (See Lhe speclflc counLry spreadsheeLs).
ln mosL cases Lhese daLa are from Lhe World 8anks World uevelopmenL lndlcaLors daLabase. lor a few counLrles, as noLed on Lhe spreadsheeLs, daLa for recenL years had Lo be derlved from oLher sources, lncludlng cenLral banks and Lreasurles. LsLlmaLes of Lhe currency composlLlon of forelgn debL, debL reschedullngs, and changes ln arrears are also generally from World 8ank/ lMl sources, alLhough ln a few cases (e.g., Slngapore) cenLral bank daLa was used Lo supplemenL Lhese esLlmaLes.
1|me er|ods. As shown ln each counLry spreadsheeL, ln mosL cases lndlvldual counLry daLa are analyzed for Lhe enLlre perlod 1970-2010 lncluslve, alLhough Lhe avallable Llme serles are shorLer for a subseL of counLrles, noLably Lhe lSu sLaLes, Chlna, and some sub- Saharan counLrles.
Country I|ows Mode| Deta||s. 1he baslc counLry flows model provldes esLlmaLes of nomlnal and real offshore caplLal flows for Lhe perlod from 1970 Lo 2010 lncluslve, correcLlng Lhem for facLors.
Several alLernaLlve measures of unrecorded caplLal flows fllghL are avallable ln Lhe academlc llLeraLure on Lhls sub[ecL. 1hese are volumlnous, compared wlLh Lhe number of emplrlcal sLudles. 41
41 lor an acerblc summary of Lhe dlfference beLween armchalr and lnvesLlgaLlve economlcs, see Appendlx lll, p10. lor a revlew of alLernaLlve measures of caplLal fllghL, see llorlan kaufmann, op. clL., fooLnoLe 2. MosL of Lhe alLernaLlve measures of caplLal fllghL appear Lo be hlghly correlaLed. lor example, ln Lhe case of Mexlco, Lhe slmple correlaLlon of lederal 8eserve's neL currency payouLs from Lhe San AnLonlo and Ll aso branches (ln real $1980) wlLh a sources and uses measure of Mexlco's fllghL
28 !ames S. Penry, 1!n 2012
Cur own preferred measure ls an ad[usLed verslon of Lhe so-called sources and uses meLhod. 42 8aslcally Lhls adds up a counLrys measured sources of forelgn caplLal forelgn loans, neL dlrecL lnvesLmenL, and neL porLfollo lnvesLmenLs and compares Lhem wlLh recorded uses, lncludlng flnanclng currenL accounL deflclLs and lncreaslng offlclal reserves. A slmple 8usslan example ls provlded ln Appendlx lll. 43
ln prlnclple, Lhe dlfference beLween Lhese recorded sources of forelgn caplLal and uses of forelgn caplLal may be aLLrlbuLed Lo unrecorded neL caplLal ouLflows. Cf course each and every lngredlenL ln Lhls yardsLlck ls sub[ecL Lo measuremenL error. Powever, over Llme and across dozens of counLrles, Lhe errors should more or less cancel ouL.
We have preferred Lo develop our own esLlmaLes of caplLal fllghL Lo have conslsLenL esLlmaLes and a sLandard perlod for comparlson .. Cn Lhe sources slde, our measure sLarLs wlLh World 8ank daLa for dollar value of gross exLernal debL sLocks by year, as well as neL forelgn dlrecL lnvesLmenL (on a balance-of-paymenLs basls) and neL forelgn porLfollo lnvesLmenL, excludlng coverlng LransacLlons by forelgn CenLral 8anks. 1he World 8anks annual debL sLock esLlmaLes already lnclude an esLlmaLe of forelgn porLfollo debL lnvesLmenLs and Lrade debL.
Lxcept|ona| I|nanc|ng. Cne necessary ad[usLmenL Lo Lhe World 8ank debL sLock numbers ls for excepLlonal flnanclng, an accounLlng enLry made when counLrles have Lrouble servlclng Lhelr debLs -- Lhe sum of neL lnLeresL raLe arrears, debL forglveness, and caplLallzed lnLeresL. As noLed by kaufmann (2004), Lhls enLry ls flcLlonal flnance, ln Lhe sense LhaL lL amounLs Lo an arblLrary book-keeplng enLry, noL acLual cash flow. 1he World 8anks debL serles lncludes excepLlonal flnanclng ln lLs debL serles, buL noL ln reserves, whlle Lhe lMl lncludes lL ln lLs measure of reserves and relaLed lLems. lor Lroubled debLors llke 8razll, ArgenLlna, nlgerla, and 8ussla, Lhe numbers are large enough Lo lead Lo nonsenslcal resulLs lf excepLlonal flnanclng ls lncluded. We deducLed lL from Lhe World 8ank debL sLock serles.
Debt I|ows Vs. Changes |n Debt Stocks. 1he World 8ank also provldes a serles of debL flows by year for some counLrles, for 1989 on, buL Lhere are large unexplalned dlscrepancles ln Lhls daLa, compared wlLh flrsL-dlfferenclng our ad[usLed debL sLock numbers, and lL glves lmplauslble resulLs ln Lhe case of several counLrles, so we declded Lo rely on Lhe ad[usLed debL sLock numbers.
from l970 Lo l987 was .69. lor changes ln u.S. bank llablllLles reporLed as owned by non-bank Mexlcans, Lhe correlaLlon was .67. lor an errors and omlsslons measure of fllghL -- Lhe sum of shorL-Lerm prlvaLe caplLal ouLflows plus errors and omlsslons from Lhe balance of paymenLs accounL -- Lhe correlaLlon was .67. SLaLlsLlcal regresslon LhaL conLrols for oLher varlables ylelds even sLronger resulLs. 42 1he key assumpLlons for Lhls caplLal ouLflow model are summarlzed ln Appendlx lll, pp10, 11 43 See Appendlx lll, p17
29 !ames S. Penry, 1!n 2012 Lxchange kate Ad[ustments. WhaL ls forelgn debL, anyway? 1he World 8ank deflnes lL wlLh respecL Lo Lhe resldency of Lhe borrower, noL Lhe currency ln whlch lL ls denomlnaLed, so lL ls lmporLanL Lo keep an eye on currency varlaLlons ln dlsLlngulshlng real from nomlnal caplLal flows.
Some researchers on caplLal flows (8oyce and ndlkumana op.clL.) adopLed Lhe pracLlce of ad[usL Lhe World 8anks long-Lerm debL sLock numbers for changes ln exchange raLes, ln cases where Lhere ls a subsLanLlal amounL of non-dollar-denomlnaLed long-Lerm debL. lf. say, Lhe dollar appreclaLes relaLlve Lo Lhe yen, Lhe acLual dollar value of yen- denomlnaLed long-Lerm debL and Lhe flows of new prlnclpal assoclaLed wlLh lL would lncrease, whlle Lhe 8anks dollar-denomlnaLed serles would be unchanged.
l have grudglngly followed Lhls pracLlce here Lhough lL ls laborlous, glven Lhe number of counLrles and currencles lnvolved, and Lhough lL Lurns ouL noL Lo make all LhaL much dlfference Lo Lhe esLlmaLes. lndeed, Lhls resulL ls conslsLenL wlLh Lhe evldence from 8oyce and LpsLeln (op. clL.) own sLudy, whlch shows LhaL Lhe effecLs of exchange raLe ad[usLmenL Lo debL on caplLal flow esLlmaLes are small for mosL counLrles, malnly because movemenLs ln mulLlple currencles ofLen cancel each oLher ouL, especlally over long perlods of Llme. ln addlLlon, mosL large debLors now hedge agalnsL currency flucLuaLlons. neverLheless, for Lhe sake of academlc purlLy, weve done Lhe LheoreLlcally correcL Lhlng.
Debt keschedu||ngs and Changes |n Arrears. uependlng on Lhe counLry, Lhese can be very lmporLanL, alLhough mosL esLlmaLes Lo daLe have largely lgnored Lhem. lrom a balance of paymenLs accounLlng sLandpolnL, for example, decllnlng arrears are llke lncreased reserves or a rlslng currenL accounL deflclL a use of forelgn funds. We have lncorporaLed Lhem ln esLlmaLes for all 139 counLrles.
(2) 1he Accumu|ated Cffshore Wea|th Mode|.
8ulldlng on Lhe counLry models developed Lo esLlmaLe caplLal flows, our second key model uses a slmple framework Lo esLlmaLe how much Lhese accumulaLed flows mlghL be worLh over Llme.
1hls addresses Lhe obvlous problem wlLh flow-based esLlmaLes Lhey donL help us accounL for Lhe rlse of Lhe global offshore lndusLry and lLs asseLs under managemenL over Llme. nor do Lhey acknowledge Lhe reallLy LhaL Lhanks Lo Lhe prlvaLe banklng lndusLrys susLalned lobbylng -- offshore lnvesLors Lyplcally are permlLLed Lo en[oy Lhe Lax-free perks of non-domlcllaLlon or non-resldency ln counLrles llke Lhe uS and Lhe uk, aL leasL wlLh respecL Lo lnLeresL on debL and bank deposlLs.
1he base case verslon of Lhls model assumes LhaL a slgnlflcanL porLlon 30 Lo 73 percenL, on average -- of Lhese Lax-free earnlngs are noL repaLrlaLed Lo source counLrles,
30 !ames S. Penry, 1!n 2012 buL are relnvesLed abroad ln relaLlvely safe, low-yleldlng lnvesLmenLs, denomlnaLed ln LradlLlonal reserve currencles llke uS dollars. 1hls ls conslsLenL wlLh Lhe facL LhaL mosL haven lnvesLmenLs are made for longer-Lerm moLlves llke asseL proLecLlon, money launderlng, and dlverslflcaLlon, noL [usL shorL-Lerm speculaLlon, wlLh low Lurnover and hlgh relnvesLmenL raLes.
We furLher assume LhaL 100 percenL of Lhe lnlLlal ouLflows are neL of source-counLry lncome Laxes oLherwlse Lhey presumably would noL be unrecorded. SubsequenL offshore earnlngs are noL repaLrlaLed and are noL sub[ecL elLher Lo domesLlc Lax or forelgn Laxes.
8eflecLlng Lhelr appeLlLe for secure low-rlsk lnvesLmenLs, we assume offshore cllenLs earn a modesL 6-monLh Cu raLe on Lhelr accumulaLed forelgn caplLal, whlch Lhelr prlvaLe bankers have already grossed down Lo reflecL Lhe cosLs of offshore managemenL.
1hls baslcally assumes, conservaLlvely, LhaL over Lhe long haul, a large share of fllghL flows has been lnvesLed ln relaLlvely-secure porLfollos of uS and Luro-denomlnaLed asseLs, malnly bank Cus. unLll Lhe laLe 2000s, Lhe convenLlonal wlsdom among fllghL caplLallsLs was, WhaL could be safer Lhan Loo-blg-Lo-fall uS, Swlss and uk banks? Cf course Lhere was also lnvesLlng ln ark Avenue, Mayfalr, Ceneva, and SouLh 8each real esLaLe, hlgh-flylng hedge funds, lnLerneL sLarLups, anama ocean fronL, fllm producLlons, rap muslclans and drug deals. 8uL our assumpLlon permlLs us Lo esLabllsh a conservaLlve basellne for porLfollo reLurns. As one ClLlbanker ln Mexlco ClLy sald, 1he money my cllenLs puL offshore ls for safe-keeplng...When Lhey wanL 200 percenL reLurns, Lhey keep Lhe money here. 44
Cbvlously Lhere ls much more work Lo wlLh exacL modellng for speclflc groups of lnvesLors, Llme perlods, and raLes of reLurn. lor example, appeLlLes for rlsk and llquldlLy may vary conslderably by reglon, whlle Lhe low lf sLable -- nomlnal raLes of reLurns on offshore lnvesLmenLs assumed here may have been wlldly hlgh or low ln dlfferenL perlods.
Cn Lhe oLher hand, based on our lnLervlews wlLh prlvaLe bankers and oLher offshore lndusLry experLs, for Lhe medlan offshore prlvaLe banklng cusLomer, Lhese assumpLlons are noL a bad sLarL. 8ased on our lnLervlews wlLh prlvaLe bankers and oLher offshore lndusLry waLchers Lo daLe, Lhls klnd of ldeal Swlss cusLomer ls much less lnLeresLed ln maxlmlzlng shorL-Lerm reLurns Lhan ln securlng an offshore nesL egg Lyplcally he or she ls ofLen Laklng qulLe enough rlsk, Lhank you, back home.
8uL of course Lhese assumpLlons are easlly modlfled, assumlng we have evldenLlary reasons Lo do so.
44 1he ClLlbanker quoLe ls from my lnLervlew wlLh Mexlcan ClLlbanker, !uly 1988.
31 !ames S. Penry, 1!n 2012 Ch|nas Story. 1here ls one counLer-example, or aL leasL a poLenLlal quallflcaLlon Lo Lhls offshore wealLh accumulaLlon model, however. 1hls ls Lhe case of Chlna. As noLed below ln our revlew of Lhe resulLs, Chlna, lncludlng Lhe malnland and Lhe Pong kong SA8, have recenLly been aL Lhe Lop of Lhe llsL ln Lerms of boLh unrecorded caplLal ouLflows and esLlmaLed fllghL wealLh. As of 2010, for example, Chlna had accumulaLed real ouLflows (ln $2000) of $743 bllllon, Lhe hlghesL ln LasL Asla, whlle Lhe Pong kong SA8 recorded ouLflows of $123.9 bllllon. lf we slmply comblne Lhese flows and apply Lhe wealLh accumulaLlon model Lo Lhem, Lhe resulL ls LhaL Chlnese lnvesLors supposedly now accounL for nearly $1.2 Lrllllon of offshore prlvaLe wealLh an lmplauslble 13 percenL of Lhe global LoLal. 43
1he real problem here ls LhaL, [usL as ln Lhe case of Chlnas spurlous Lrade mlslnvolclng noLed above, much of lLs unrecorded caplLal LransacLlons wlLh Pong kong are probably double-counLed. A slgnlflcanL amounL of Chlnas apparenL caplLal fllghL ls acLually [usL round-Lrlpplng by way of lnLermedlary companles based ln Pong kong and a few oLher key havens, noLably Lhe 8rlLlsh vlrgln lslands.)
1hls sLlll counLs as parL of Lhe offshore lndusLry, buL for purposes of our wealLh model, lL begs Lhe quesLlon of how large a facLor such round-Lrlpplng ls noL only for Chlna, buL for oLher source counLrles, especlally for Lhose on our Lop 20 offshore wealLh llsL.
Powever, wed argue LhaL Chlnas slLuaLlon wlLh Pong kong may be unlque. Whlle, for example, key LaLln Amerlcan counLrles llke venezuela and Mexlco have long en[oyed close Lles wlLh Lhelr closesL haven, Lhe uS, and we are aware of some round-Lrlpplng by lnvesLors ln 8razll, we also know for a facL LhaL wealLhy lnvesLors from Lhese source counLrles accounL for a slgnlflcanL share of uS bank deposlLs owned by non-resldenLs.
llnally, we can use our esLlmaLes of Pong kongs unrecorded caplLal Lo place an upper bound on Lhe maxlmum amounL of round-Lrlpplng wlLh respecL Lo Lhe malnland aL mosL, lL ls abouL 17 percenL of malnland Chlnas gross unrecorded caplLal ouLflows. 1hls permlLs us Lo ad[usL Chlnas accumulaLed wealLh esLlmaLes accordlngly. lL also provldes a reasonable maxlmum esLlmaLe for Lhe llkely share of round-Lrlpplng by oLher source counLrles.
(3) Ana|ys|s of r|vate 8ank|ng Assets
As weve seen, Lhe accumulaLed offshore wealLh model opened Lhe door Lo many oLher posslblllLles, lncludlng an analysls of relnvesLmenL earnlngs on offshore wealLh.
lor decades, llrsL World prlvaLe bankers employed by Lhe Lop 30 or so lnsLlLuLlons have orchesLraLed Lhe sysLemaLlc eroslon of lncome and wealLh Lax bases ln hlgh- and low- lncome counLrles allke. 1hey have asslduously recrulLed Lhe worlds wealLhlesL people as
43 See Appendlx lll, p47
32 !ames S. Penry, 1!n 2012 Lhelr cllenLs, lncludlng Lens of Lhousands of developlng counLrles. 1hey have served as senlor plloLs ln CaplLal lllghL Alr, helplng Lhese cllenLs move a slgnlflcanL share more Lhan half, ln Lhe case of LaLln Amerlca and some Aslan counLrles of Lhelr llquld caplLal Lo offshore accounLs under Lhe cover of shell companles and LrusLs, beyond Lhe reach of domesLlc Lax auLhorlLles. 1hey have enabled cllenLs Lo move lL, hlde lL, lnvesL lL, manage lL, spend lL, and make use of lL remoLely on Lhe fly.
All Lold, Lhe sophlsLlcaLed Lax ln[usLlce neLwork LhaL Lhese lnsLlLuLlons have consLrucLed now employs fewer Lhan a mllllon people all over Lhe planeL. 46 8uL Lhls ls an lnfluenLlal mllllon Lhey are Lhe sysLems archlLecLs, operaLors, and managers of Lhe p|utonom|c system LhaL ClLlgroup analysLs descrlbed qulLe ln an lnfamous, now suppressed, 2003 memo. 47
Whlle Lhere are now over 300 prlvaLe banks, hedge funds, law flrms, accounLlng flrms, and lnsurance companles LhaL speclallze ln offshore, Lhe lndusLry ls acLually very concenLraLed. MosL of lLs employees work dlrecLly or lndlrecLly for Lhe worlds Lop 30 prlvaLe banks, especlally Lhe Lop 21 LhaL now each have prlvaLe cross-border asseLs under managemenL of aL leasL $100 bllllon each. 48
ln shorL, Lhls comparaLlve handful of ma[or prlvaLe banklng lnsLlLuLlons now accounLs for 62 Lo 74 percenL of all offshore prlvaLe wealLh. Many readers wlll recognlze Lhe names of Lhe domlnanL players, as Lhey have done for decades: u8S, CredlL Sulsse, ClLlgroup/SS8/Morgan SLanley, ueuLsche 8ank, 8ankAmerlca/Merrlll Lynch, !MorganChase, 8n arlbas, PS8C, lcLeL & Cle, Coldman Sachs, A8n Amro, 8arclays, CredlL Agrlcole, !ullus 8aer, SocleLe Ceneral, and Lombard Cdler.
1o address Lhls baslc facL abouL Lhe offshore markeL, we have underLaken a sysLemaLlc analysls of cross-border prlvaLe banklng asseLs under managemenL aL Lhe Lop 30 lnLernaLlonal prlvaLe banks for Lhe perlod 2003-2010. Cur daLa sources lnclude company annual reporLs and 10ks, lnvesLmenL analysLs, lnLervlews wlLh prlvaLe banklng lndusLry experLs, lndusLry waLchers llke WealLh 8rleflng news and Money Launderlng AlerL, and a survey of recenL markeL research sLudles for Lhe prlvaLe banklng lndusLry.
46 SwlLzerland probably has Lhe largesL number of dlrecL employees ln prlvaLe banklng abouL 200,000, accordlng Lo Lhe Swlss 8ankers AssoclaLlon. 1he global lndusLry esLlmaLe ls based on our deLalled analysls of lndlvldual havens and prlvaLe banklng lnsLlLuLlons. 1he lndusLry also creaLes lndlrecL demand for oLher buslness servlces, lncludlng law flrms, accounLlng flrms, offlce servlces, and Lravel. We have allowed for Lhls ln our esLlmaLe. 47 See www.scrlbd.com/.../ClLlgroup-CcL-16-2003-luLonomy-8eporL-arL-1. 48 As of !uly 2008, before Lhe crash, Luromoney esLlmaLed LoLal asseLs under managemenL for Lhe global prlvaLe banklng lndusLry as a whole aL $11.8 Lrllllon. See: hLLp://www.euromoney.com/ArLlcle/2093990/rlvaLe-banklng-MeLhodology.hLml. 8y SepLember 2009 Lhls flgure had fallen Lo $11.1 Lrllllon. Cur sample of Lhe Lop 30 banks accounLed for abouL $8 Lrllllon of Lhls. As noLed ln Lhe LexL, ln addlLlon Lo AuMs, of course, Lhere are also asseLs under cusLody, deposlLs, and cllenL brokerage asseLs under admlnlsLraLlon, whlch are on Lhe order of 30 Lo 80 percenL greaLer Lhan AuMs. So Lhese flgures are all conslsLenL wlLh our esLlmaLed $20 Lrllllon for 2009.
33 !ames S. Penry, 1!n 2012
1he resulLs are readlly summarlzed.
I|rst, as of uecember 2010, by our esLlmaLes, Lhe worlds Lop 30 global prlvaLe banks alone had $12.06 Lrllllon of prlvaLe cross-border flnanclal wealLh under managemenL. 1hls compared wlLh 2003, when our esLlmaLe ls LhaL Lhe Lop 30 managed $3.4 Lrllllon an average annual growLh raLe for Lhe lndusLry of nearly 16 percenL, desplLe Lhe world economys ups and downs.
nor are Lhese all Lhe cllenL asseLs LhaL Lhese lnsLlLuLlons handle. 1here are also bank deposlLs, whlch are usually lncluded under managed asseLs, as well as asseLs under cusLody and admlnlsLraLlon, lncludlng brokerage asseLs. uependlng on Lhe year, Lhese addlLlonal asseLs Lyplcally add aL leasL 23 percenL Lo Lhe LoLal. Allowlng for Lhls, as well as for underreporLlng and oLher daLa problems, Lhese flgures are conslsLenL wlLh our overall $21 Lrllllon Lo $32 esLlmaLe for global offshore flnanclal asseLs as of 2010.
Second, Lhe Lop Len banks ln Lhls group are remarkably sLable Lhe Swlss banks u8S and CredlL Sulsse occupled Lhe lead poslLlons ln boLh years, and 7 of Lhe Lop 10 remalned ln Lhe Lop Ller. noLable new addlLlons Lo Lhe leaders were 8arclays, Lhe ulLra-prlvaLe Ceneva bank lcLeL, and 8ankAmerlca, by way of lLs Merrlll acqulslLlon, noLable decllners were ClLlgroup and A8n Amro.
1h|rd, Lhe Lop Len banks grew even fasLer Lhan Lhe lndusLry as a whole, an AAC8 of over 20 percenL per year durlng Lhls perlod, sharply lncreaslng Lhelr share of Lhe groups asseLs under managemenL from 42 percenL ln 2003 Lo more Lhan 31 percenL ln 2010.
1he lrony here ls LhaL every one of Lhese leadlng global banks, excepL lcLeL, were deemed Loo blg Lo fall by Lhelr governmenLs ln 2008-2010, and collecLlvely recelved hundreds of bllllons ln Laxpayer-flnanced caplLal ln[ecLlons, sLandby credlLs, loan guaranLees, Loxlc asseL guaranLees, low-cosL loans, and Lhe uS 1reasurys lebruary 2009 swap deal wlLh SwlLzerland. 49 1hey beneflLLed greaLly from Lhe $80 bllllon AlC ballouL and Lhe vlrLually-zero real lnLeresL raLe envlronmenL esLabllshed by Lhe worlds cenLral banks. WlLhouL Lhese Loo blg Lo fall governmenL subsldles, several would have dlsappeared. 30
49 lrom AugusL 2008 Lhrough AugusL 2009, Lhe governmenL ald recelved by Lhe Lop 21 prlvaLe banklng lnsLlLuLlons as a group LoLaled aL leasL $393 bllllon of sLandby credlLs, $939 bllllon of Loxlc asseL guaranLees, and $392 bllllon of governmenL caplLal ln[ecLlons. CovernmenL ald Lo banks, Lurope and Lhe uS, uS lederal 8eserve, uS 1reasury, PM 1reasury, and oLher cenLral bank offlclal sources, our analysls.
30 lL ls lnLeresLlng Lo noLe LhaL Lhe precursors of ClLlgroup, !MorganChase, 8ankAmerlca, and Coldman Sachs were all deeply lnvolved ln lendlng Lo flnance sLock markeL speculaLlon ln 1928-29. ClLlbanks chalrman aL Lhe Llme, Sunshlne Charlle MlLchell, was descrlbed by a uS SenaLor as more Lhan any 30 men, responslble for Lhe sLock markeL crash.
34 !ames S. Penry, 1!n 2012 uld Lhe 1reasury ueparLmenLs around Lhe world noL undersLand LhaL Lhese very same banks are leadlng Lhe world ln enabllng Lax dodglng lndeed, Lo some exLenL, preclsely because offshore lnvesLors know LhaL Lhey have been under-wrlLLen by Lhelr 1reasury ueparLmenLs?
(4) 1he Cffshore Investor ortfo||o Mode|.
Cur lasL model ls anoLher relaLlvely slmple, daLa-focused model of offshore lnvesLor porLfollo behavlor. lL ls based on a comblnaLlon of 8lS daLa on cross-border deposlLs and oLher asseL holdlngs by non-bank lnvesLors, an analysls of porLfollo mlx assumpLlons made by wealLh lndusLry analysLs, and lnLervlews wlLh acLual prlvaLe banks.
All Lhls ylelds a range of porLfollo scale-up facLors LhaL can be comblned wlLh reporLed 8lS asseLs daLa Lo yleld anoLher esLlmaLe for Lhe slze of offshore wealLh.
koots of the mode|: 1INs Iru|tfu| Lrrors. Cur model ls rooLed ln 1!ns orlglnal 2003 esLlmaLe for offshore prlvaLe wealLh of $11.3 Lrllllon, whlch lncluded $2 Lrllllon of non- resldenLlal offshore real esLaLe and $9.3 Lrllllon of offshore flnanclal wealLh. unLll now, Lhls ls sLlll noL only Lhe largesL esLlmaLe Lo daLe, buL also Lhe only one LhaL was made wlLh even a rudlmenLary porLfollo model.
1!n based lLs rough esLlmaLe on Lhe slmple procedure of scallng up 8lS daLa on offshore deposlLs by nonbanks by an average raLlo of cash and bank deposlLs Lo all flnanclal asseLs LhaL was based on ML/CapCemlnls annual esLlmaLes of asseL allocaLlons for hlgh- neL worLh porLfollos.
unforLunaLely, 1!n made several slmple errors. 1hese conLrlbuLed, lf anyLhlng, Lo a subsLanLlal 5,&'4'#$/+"$/), of Lhe slze of offshore. Powever, Lhe baslc meLhodology can be reflned and exLended Lo dellver yeL anoLher LrlangulaLlon.
1!ns 2003 esLlmaLe conLalned several errors, all of whlch blased lL downwards. llrsL, lL relled on Lhe wrong llne lLem for offshore deposlLs, uslng a LoLal of uS$ 2.7 Lrllllon. 31
1he correcL comparable flgure for offshore deposlL llablllLles by nonbanks for !une 2003 was $4.68 Lrllllon. 8y !une 2007 Lhls had grown Lo $7.43 Lrllllon, dropplng back Lo $7.01 Lrllllon as of !une 2010, under Lhe lmpacL of Lhe economlc crlsls. 32
Second, Lo esLlmaLe LoLal offshore flnanclal asseLs, 1!n used a 3.3 llquldlLy raLlo Lo scale up Lhls cross-border deposlLs flgure Lo LoLal flnanclal asseLs. 1hls was based on a 2004 sLudy of global flnanclal asseL demand by Mcklnsey & Co., whlch found LhaL LhaL
31 See 1!n (2003), op. clL., 1he rlce of Cffshore, p. 1, whlch lncorrecLly use $2.7 Lrllllon for Lhe value of offshore deposlLs ln !une 2004. 1he acLual flgure for offshore deposlLs by nonbanks was $4.03 Lrllllon. (See 8lS CuarLerly Survey, uecember 2004, 1able 1.) 32 1hls 8lS daLa on cross-border deposlL llablllLles ls llmlLed Lo reporLs obLalned from [usL 30 offshore flnanclal cenLers, and may Lherefore be regarded as conservaLlve.
33 !ames S. Penry, 1!n 2012 Lhe raLlo of cash (bank deposlLs) Lo LoLal flnanclal neL worLh ln hlgh-neL worLh lnvesLor porLfollos had averaged 3.3 Lo 3.83 over Lhe precedlng 4 years.
1!n Lherefore assumed a 3.3 raLlo Lo scale up lLs deposlLs esLlmaLe Lo flnanclal neL worLh. 1hls was desplLe Lhe facL LhaL lL was hard Lo reconclle Mcklnseys esLlmaLes for deposlLs wlLh Lhose of 8lS, slnce Lhe Mcklnsey numbers lnapproprlaLely lncluded a huge amounL of lnLer-bank deposlLs, whose raLlos Lo flnanclal asseLs had noLhlng Lo do wlLh Lhe behavlor of nonbanks.
ln facL, ML/ CapCemlnl (ML/CC) has esLlmaLed acLual llquldlLy raLlos for hlgh-neL worLh lndlvlduals wlLh more Lhan $1 mllllon ln neL flnanclal asseLs dlrecLly for !une 1998, uecember 1998, !une 2002, and each successlve year Lhrough 2010. 1he medlan value for 1998-2010 was 4.6, and for 2004-2003 lL Lopped 4.9. (See ML/CC, World WealLh 8eporLs, op. clL.) Cnly ln years of decllnlng sLock markeLs llke 2002 and 2008-9 dld Lhe llquldlLy raLlo dlp below 4. lor purposes of our esLlmaLes here, 2010 was baslcally a year when all Lhe key sLock markeLs ln Lhe uS, Asla, and Lurope were recoverlng nlcely, slgnallng (premaLurely) a recovery.
1here +/2.$ @' a case LhaL ML/CCs esLlmaLes of Lhese average llquldlLy raLlos for PnWls should be Laken as an 5HH'4 @)5,& for offshore flnanclal lnvesLmenLs, on Lhe argumenL LhaL offshore funds are securlLy blankeLs for more rlsk-averse lnvesLors. Weve doubled-checked Lhls by examlnlng Lhe slze of prlvaLe banklng cllenLs lnvesLed asseLs under managemenL relaLlve Lo Lhelr cross-border deposlLs for several prlvaLe banks LhaL are leaders ln Lhe offshore markeL, and publlsh Lhls daLa, lncludlng u8S and CredlL Sulsse. lor Lhese lnsLlLuLlons, llquldlLy raLlos equaled or exceeded Lhe same- year llquldlLy raLlos lmplled by ML/CCs average PnWl porLfollo allocaLlons.
ln our esLlmaLes, we have LreaLed Lhe ML/CC-based llquldlLy raLlos as upper bounds, and have seL a conservaLlve lower bound aL 3.0 raLlo, much lower Lhan Lhe 3.3 orlglnally used by 1!n. 1hls 3.0 raLlo ls below Lhe ML/CC porLfollo allocaLlon esLlmaLes for all buL one year ln Lhe 1998-2010 perlod 2002, when lL equaled 3.0. lL Lherefore esLabllshes a super-conservaLlve floor under our non-banks deposlLs mulLlpller.
Cf less lnLeresL Lo our focus here -- whlch ls on flnanclal wealLh -- whlle 1!n also allowed LhaL non-resldenLlal real esLaLe holdlngs mlghL be anoLher lmporLanL componenL of offshore wealLh, Lhelr $2 Lrllllon esLlmaLe for 2003 seems low relaLlve Lo Lhe correcLed porLfollo slze.
llnally, 1!n also lgnored Lhe role of so-called alLernaLlve lnvesLmenLs and collecLlbles ln PnWl porLfollos, lncludlng hedge funds, arL, and prlvaLe equlLy. Whlle we lack speclflc lnformaLlon on Lhe share of )**#.)4' porLfollos H'4 #' accounLed for by such llllquld lnvesLmenLs, for real esLaLe and non-collecLlble alLernaLlves, Lhe respecLlve medlan porLfollo shares reporLed by ML/CC for Lhe perlod 1998-2008 were 17 percenL and 10 percenL respecLlvely, wlLh ranges of 13-24 and 7-20.
36 !ames S. Penry, 1!n 2012
ln sum, whlle furLher research on offshore PnWl porLfollo allocaLlons ls warranLed, Lhe baslc 1!n meLhodology ls worLh developlng. lor purposes of Lhls paper we have checked Lhe range of llquldlLy raLlos used above wlLh prlvaLe bankers and lndusLry sources, and Lhey have conflrmed LhaL Lhey are plauslble.
kev|sed ortfo||o Mode| Lst|mates. As for 1!ns orlglnal esLlmaLes, when Lhe errors are correcLed, Lhe mosL llkely quanLlLy of prlvaLe offshore flnanclal asseLs ln !une 2004 was noL $9.3 Lrllllon, buL $12.1 Lo $20 Lrllllon, dependlng on wheLher we use Lhe very conservaLlve 3.0 llquldlLy mulLlpller or someLhlng more reallsLlc.
Slnce Lhen, offshore deposlLs by nonbanks nearly doubled from 2004 Lo uecember 2007, when Lhe global economy Look a Lumble. uurlng LhaL perlod, offshore flnanclal asseLs may have grown Lo be worLh as much as $22 Lo 33 Lrllllon. Slnce Lhen Lhe model lndlcaLes LhaL Lhey have slumped sllghLly Lo Lhe $21 Lrllllon Lo $32 Lrllllon range, wlLh a plauslble mldpolnL of abouL $26 Lrllllon. 8uL Lhls sLlll represenLs enough growLh slnce 2004 Lo be conslsLenL wlLh Lhe growLh ln global prlvaLe banklng AuMs noLed above. 33
Assumlng a developlng counLry wealLh share of 23 Lo 30 percenL, Lhls range ls also conslsLenL wlLh Lhe resulLs of our accumulaLed wealLh model.
1he kev|sed G|oba| D|str|but|on of Wea|th
lL may be helpful Lo place Lhese esLlmaLes ln Lhe conLexL of Lhe overall dlsLrlbuLlon of global flnanclal wealLh. lL Lurns ouL LhaL Lhls dlsLrlbuLlon ls lncredlbly concenLraLed. 8y our esLlmaLes, aL leasL a Lhlrd of all prlvaLe flnanclal wealLh, and nearly half of all offshore wealLh, ls now owned by worlds rlchesL 91,000 people [usL 0.001 of Lhe worlds populaLlon.
1he nexL 31 percenL of all wealLh ls owned by Lhe nexL 8.4 mllllon, anoLher Lrlvlal 0.14 of Lhe worlds populaLlon. 34 As noLed, a Lhlrd of Lhls has been accumulaLed from Lhe 139 source counLrles ln our focus sample. AlmosL all of lL has managed Lo avold all lncome and esLaLe Laxes, elLher by Lhe counLrles where lL has been lnvesLed and or where lL comes from.
33 1hese esLlmaLes for 2010 exclude oLher non-flnanclal very lmporLanL forms of offshore wealLh llke non-resldenLlal real esLaLe, alLernaLlve lnvesLmenLs, and collecLlbles. As noLed, for 2004, 1!n added anoLher $2 Lrllllon Lo lLs $9.3 Lrllllon flgure for flnanclal asseLs Lo accounL for real esLaLe, buL noLhlng Lo allow for offshore alLernaLlve lnvesLmenLs and collecLlbles. Accordlng Lo ML/CapCemlnl, non-resldenLlal real esLaLe and alLernaLlve lnvesLmenLs may add anoLher 23 Lo 38 Lo PnWls wealLh porLfollos, alLhough Lhe offshore porLfollo shares are lower.
34 See Appendlx lll, p102
37 !ames S. Penry, 1!n 2012 6. 1kADL MISINVCICING -- AN IMCk1AN1 ASIDL
Cn Lhe uses slde, as noLed, we have employed Lhe World 8anks serles for counLry currenL accounL deflclLs and changes ln reserves. Some auLhors, llke Cll, LpsLeln (2003), and 8oyce and ndlkumana have also Lrled Lo ad[usL Lhe observed currenL accounL deflclL for alleged Lrade mlslnvolclng of forelgn Lrade.
l have experlmenLed exLenslvely wlLh such ad[usLmenLs. l slmply donL flnd Lhe rough rules of Lhumb LhaL are used Lo make Lhese ad[usLmenLs rellable. 33
lf one ls a greaL bellever ln Lhe lmporLance of aggregaLe as opposed Lo case-speclflc goods Lrade mlsprlclng, leavlng ouL Lhls ad[usLmenL generally Lends Lo make our unrecorded caplLal flow esLlmaLes even more conservaLlve.
33 LpsLeln (2003) ls hereLofore Lhe mosL comprehenslve caplLal fllghL esLlmaLes Lo daLe, buL hls esLlmaLes are sub[ecL Lo many ob[ecLlons. llrsL, LpsLeln (2003) omlLs key counLrles LhaL are leadlng source of fllghL caplLal noLably, 8ussla, Mexlco, venezuela, and ArgenLlna. Llke Cll, he also employs Lhe so-called 8oyce-ndlkumana ad[usLmenL for mls-lnvolclng, whlch l flnd very problemaLlc. lor several counLrles, llke Chlna and Malaysla, Lhls ad[usLmenL alone accounLed for several hundred bllllon ln addlLlonal lmpuLed fllghL. 1he ad[usLmenL, whlch ls based on Lhe lMls ulrecLlon of 1rade sLaLlsLlcs, assumes a consLanL normal 10 percenL clf/fob raLlo for all developlng counLry Lrade wlLh Lhe llrsL World. lL Lhen Lakes all Lhe llrsL World exporLs Lo a speclflc developlng counLry LhaL are reporLed ln Lhe lMls ulrecLor of 1rade SLaLlsLlcs, and scales Lhem up by Lhls raLlo, compares Lhe resulLlng ad[usLed exporLs flgure wlLh Lhe reporLed llrsL World lmporLs by a speclflc counLry, and aLLrlbuLes any resulLlng gap Lo lmporL over- lnvolclng. lL Lhen exLrapolaLes from Lhls llrsL World-developlng counLry gap Lo Lhe counLrys enLlre Lrade balance, on Lhe assumpLlon LhaL llrsL World Lrade daLa are accuraLe.
My lnspecLlon of Lhe uns CommodlLy 1rade SLaLlsLlcs daLa base (hLLp://unsLaLs.un.org/unsd/comLrade) whlch has deLalled lnformaLlon on acLual fob/clf raLlos by counLry and commodlLy, reveals LhaL Lhe 1.1 raLlo assumpLlon ls ofLen wlldly lnaccuraLe, and LhaL lL flucLuaLes over Llme, even for a glven commodlLy famlly. 1here may also be Llmlng and quallLy dlfferences LhaL affecL Lhe daLa for example, exporLs Lend Lo be recorded earller Lhan lmporLs, and wlLh less accuracy, because of Lhe lncenLlves Lo collecL lmporL duLles and enforce quoLas. 1hls blases our esLlmaLes of lmporL over-lnvolclng upwards. ln Lhe case of Chlna, as dlscussed ln Lhe LexL, Lhere are also serlous problems lnvolved ln separaLlng ouL Lrade wlLh Pong kong, a key LransshlpmenL polnL. See also lrank 8. CunLer (2003), CaplLal lllghL from Chlna 1984-2001, !une 2003, Lelgh unlverslLy. Much more research aL Lhe lndlvldual counLry level and commodlLy level ls needed Lo resolve Lhls lssue, so for Lhe momenL, we belleve LhaL Lhe safer course ls Lo omlL Lhe ad[usLmenL enLlrely.
LpsLeln (2003)s esLlmaLes for Chlna also appear Lo be vasLly oversLaLed, even wlLhouL Lhe mls-lnvolclng ad[usLmenL, because of Lhe way Pong kong ls LreaLed ln Chlnas balance of paymenLs sLaLlsLlcs. LpsLeln (2003) also aLLempL Lo correcL for Lhe facL LhaL some counLrles debLs are parLly denomlnaLed ln non- dollar currencles.
llnally, for mosL counLrles, LpsLelns esLlmaLes sLarL ln 1980 and leave off ln 2001, whlle ours exLend from 1970 Lo 2010.
38 !ames S. Penry, 1!n 2012 1hls ls as good a polnL as any Lo make a more general polnL abouL Lhe esLlmaLlon of Lransfer mlsprlclng, slnce lL has recenLly recelved more aLLenLlon Lhan ever ln Lhe offshore debaLe.
ln Lhe course of preparlng Lhls paper and anoLher presenLaLlon for 1!ns recenL Pelslnkl Lransfer prlclng conference, 36 l have closely examlned several recenL esLlmaLes of alleged Lransfer mlsprlclng ln goods -- by Cll ln parLlcular.
1hls analysls leads me Lo conclude LhaL we A",,)$ rely on Lhe sLandard approaches Lo esLlmaLlng Lransfer mlsprlclng abuse ln order Lo correcL errors ln Lhe currenL accounL, much less ln order Lo esLlmaLe aggregate flows of unrecorded caplLal flows due Lo Lransfer mlsprlclng. 1here are slmply Loo many problems wlLh Lhe daLa.
ln parLlcular, we have Clls well known clalms LhaL aL leasL $1 Lrllllon per year ls belng losL '"A. -'"4 by developlng counLrles because of goods mlsprlclng.
l flnd Lhls sensaLlonal clalm dublous aL besL. A close look aL Clls analysls of Lrade mlsprlclng reveals LhaL lL ls serlously flawed, for several reasons. 37
(1) Clls analysls leaves ouL all counLry cases where Lhelr measures of Lrade mlsprlclng are negaLlve desplLe Lhe facL LhaL such cases are qulLe common, especlally ln Afrlca, due Lo facLors llke unreporLed conLraband and Lhe parklng of proflLs ln offshore havens. 1hls facLor Lends Lo lnflaLe Lhe aggregaLe esLlmaLes enormously.
(2) More Lhan 70 percenL of Clls aggregaLe esLlmaLes for Lrade mlslnvolclng derlve from [usL Lwo counLrles Chlna and Mexlco. 8oLh of Lhese gaps also largely perLaln Lo Lhese counLrles Lrade wlLh Lhe uS. WhaLs golng on here?
8ecenL work on Lhese Lwo slzable lmbalances by professlonal Lrade experLs suggesLs LhaL Lhey were largely due Lo bureaucraLlc facLors llke Llmlng delays ln Lrade reporLs, exchange raLe varlaLlons, Lhe handllng of LransshlpmenLs Lhrough Pong kong, Chlnas key offshore Lrade lnLermedlary, and Lhe facL LhaL Mexlco does noL Lrack or ad[usL lLs daLa for re-exporLs. WlLhouL a more careful look aL speclflc bllaLeral Lrade relaLlonshlps aL Lhe level of Lradlng parLners and lndusLrles, we should LreaL Lhese esLlmaLes wlLh greaL cauLlon. 38
1he Need for Case Stud|es. Cn Lhe oLher hand, my recenL work on goods mlslnvolclng shows LhaL lL can be a serlous problem ln speclflc lndusLrles llke mlnlng, foresLry, Lobacco, beer, and agrlbuslness, -- especlally where lL provldes a condulL for caplLal ouLflows, parklng proflLs ln low-Lax havens, and Lax avoldance.
uslng prevlously-unanalyzed Lradlng parLner daLa from Lhe un ComLrade daLa base, we have done counLry-by-counLry Lrade mlsprlclng analyses for key source counLrles, lncludlng Chlna, Mexlco, lndla, and Zambla, a Llny rlch poor counLry ln sub-Saharan Afrlca LhaL heavlly on copper exporLs. 39
We have also looked closely aL Lransfer mlsprlclng by SwlLzerland, a key re-exporLlng haven ln Lhe global Lrade sysLem LhaL has hereLofore earned a nlce llvlng by parklng proflLs havens. 1hese careful case sLudles of parLlcular bllaLeral Lradlng relaLlonshlps acLually underscore Lhe unsavory role LhaL havens play ln our global Lradlng sysLem.
Cur analysls leads Lo several key concluslons:
(1) We slmply A",,)$ rely on Lransfer mlsprlclng comparlsons Lo correcL errors ln Lhe currenL accounL, for purposes of esLlmaLlng aggregate unrecorded caplLal flows. 1he garganLuan esLlmaLes of Lrade mlsprlclng LhaL we ofLen read abouL ln Lhe headllnes should be Laken wlLh a salL Lhumb and a shoL glass of Lequlla.
(2) Speclflc lndusLry case sLudles are Lhe way Lo go. Cur deLalled sLudles of mlsprlclng wlLh respecL Lo Mexlco, Zambla, lndla, and SwlLzerland show LhaL analyzlng bllaLeral Lrade closely underscores Lhe unsavory role LhaL havens play ln our global Lradlng sysLem. Clven Lhe serlous daLa problems for example, Lhe absence of up-Lo-daLe publlshed Cll-lC8 raLlos for speclflc Lradlng parLners and commodlLles -- Lhere ls no way rlghL now Lo esLlmaLe how large lL ls on a global scale, wlLhouL a greaL deal of furLher research.
lurLhermore, Lhe focus on goods Lransfer mlsprlclng obscures Lhe slgnlflcanL role played by corporaLe abuses of Lhe haven sysLem wlLh respecL Lo lnLanglbles llke brands, paLenLs, and knowhow, mlsprlclng of whlch has been growlng dramaLlcally. llnally, nolsy esLlmaLes of aggregaLe corporaLe goods mlsprlclng Lend Lo dlsLracL aLLenLlon noL only from Lhe klnd of haven abuse of mls-lnvolclng [usL noLed, buL also from Lhe serlous abuses by Lhe offshore lnvesLmenL and prlvaLe banklng lndusLrles LhaL are Lhe focus of Lhls paper.
2J H** E;;*(@-F '''4 B,+-.C: 75,+):
40 !ames S. Penry, 1!n 2012
7. IMLICA1ICNS
1hls reporL has many lmpllcaLlons for publlc pollcy and research. 1hls flnal secLlon summarlzes some of Lhe mosL lmporLanL ones.
I. M|ss|ng Wea|th, Understated Inequa||ty.
As noLed earller, Lhls reporL focuses our aLLenLlon on a huge black hole ln Lhe world economy LhaL has never before been measured prlvaLe offshore wealLh, and Lhe vasL amounLs of unLaxed lncome LhaL lL produces. Comlng aL a Llme when governmenLs around Lhe world are sLarved for resources, and we are more consclous Lhan ever of Lhe cosLs of lnequallLy, lL could noL be more Llmely.
uslng several lndependenL esLlmaLlon meLhods, and Lhe mosL comprehenslve daLa seL ever assembled, we have been able Lo LrlangulaLe on Lhe slze and growLh of Lhls black hole. Lven Lhough we Laken palns Lo err on Lhe conservaLlve slde, one key lmpllcaLlon ls LhaL Lhls hldden offshore secLor ls large enough Lo make a slgnlflcanL dlfference Lo mosL of our convenLlonal measures of lnequallLy.
1he reporL underscores Lhe facL LhaL all of our convenLlonal measures of lnequallLy sharply undersLaLe Lhe levels of lncome and wealLh lnequallLy aL boLh Lhe counLry and global level. (1he 1ax !usLlce neLwork reporL ;,'C5"(/$-M N)5 0),$ O,)B $.' F"(* )* ;$1 whlch accompanled Lhe 2012 release of Lhls one, explores Lhe lssue ln more depLh.) lor mosL counLrles, Lhe lnequallLy of flnanclal wealLh wlLhln counLrles ls noL only much greaLer Lhan we suspecLed aL any one polnL ln Llme, buL lL has been growlng much fasLer.
lndeed, slnce Lhe overwhelmlng share of unrecorded offshore prlvaLe asseLs LhaL we have ldenLlfled belong Lo a Llny ellLe, Lhe lmpacL on lnequallLy ls asLonlshlng We have esLlmaLed, for example, LhaL less Lhan 100,000 people, .001 of Lhe worlds populaLlon, now conLrol over 30 percenL of Lhe worlds flnanclal wealLh. 60
lurLhermore, slnce Lhe offshore lndusLry experlenced a Lake-off ln Lhe laLe 1960s, and has been growlng relaLlve Lo Lhe resL of Lhe world economy Lhrough 2010, lL appears LhaL ln-counLry wealLh lnequallLy may have lncreased rapldly durlng Lhls perlod.
60 Appendix III, p102
41 !ames S. Penry, 1!n 2012 Cf course global lnequallLy ls also Lhe resulL of dlsparlLles among "!'4"2' levels of wealLh among naLlons. 1he acceleraLlon of naLlonal growLh by leadlng developlng counLrles llke Chlna, lndla, and 8razll slnce Lhe laLe 1980s lmplles LhaL Lhe narrowlng of average dlfferenLlals ln naLlonal lncome and wealLh levels among counLrles may have offseL Lhese wlder ln-counLry lnequallLles Lo some exLenL.
Cn Lhe oLher hand, Lhe acceleraLed growLh raLes of Lhese forLunaLe few developlng counLrles has also wldened Lhe gap beLween Lhem and many oLher developlng counLrles.
lurLhermore, local, ln-counLry lnequallLy ls arguably by far Lhe mosL lmporLanL Lype of lnequallLy ln Lerms of soclal, pollLlcal, and economlc lmpacLs, as well as percepLlons of relaLlve well belng, and LhaL ls Lhe Lype LhaL has lncreased aL Lhe hands of Lhe global haven lndusLry.
llnally, from Lhe plraLe bankers markeL perspecLlve, whaL ls perhaps mosL lnLeresLlng abouL Lhe new landscape of global lnequallLy LhaL we have uncovered ls Lhe recenL emergence of a Lrue LransnaLlonal prlvaLe ellLe a relaLlvely Llny fracLlon of Lhe worlds populaLlon LhaL shares surprlslngly slmllar needs and lnLeresLs from Lhe sLandpolnL of flnanclal secrecy, banklng servlces, Laxes, and regulaLlon.
lncreaslngly, lndeed, Lhe lndlvldual members of Lhls prlvaLe ellLe may be assumlng many of Lhe same aLLrlbuLes as mulLlnaLlonal companles, even as MnCs have been becomlng more llke prlvaLe lndlvlduals, so far as pollLlcal rlghLs are concerned. 1hls means LhaL super-rlch lndlvlduals are lncreaslngly acLlng as clLlzens of mulLlple [urlsdlcLlons aL once, even Lhough Lhey may be resldenL nowhere for Lax purposes, LhaL Lhey are able Lo relocaLe qulckly across borders, and LhaL Lhey are able Lo acqulre represenLaLlon wlLhouL LaxaLlon, Lhe ablllLy Lo exerL local pollLlcal lnfluence ln mulLlple [urlsdlcLlons, lndependenL of whaLever Laxes Lhey pay ln any parLlcular [urlsdlcLlon.
lL also means LhaL as a group Lhls LransnaLlonal ellLe has, ln prlnclple, a sLrong vesLed lnLeresL ln pushlng for weaker lncome and wealLh LaxaLlon weaker governmenL regulaLlon, more open markeLs, and weaker resLrlcLlons on pollLlcal lnfluence and campalgn spendlng across borders wlLh a huge LransnaLlonal haven army of plraLe bankers, law flrms, accounLlng flrms, lobbylsLs, and 8 flrms ready Lo do Lhelr blddlng.
1hus Lhe )@P'A$/!' lncrease ln global wealLh lnequallLy aL Lhe lndlvldual level LhaL we have begun Lo measure here seLs Lhe sLage for asklng a wlde varleLy of quesLlons abouL Lhe resulLlng pollLlcal and soclal lmpacLs on Lhe LradlLlonal naLlon-sLaLe.
II. 1ax 8ase Lros|on. AnoLher key lmpllcaLlon of Lhls sLudy ls LhaL Lhe lmpacL on losL Lax revenue lmplled by our esLlmaLes may be huge -- large enough Lo make a slgnlflcanL dlfference Lo Lhe flnances of naLlons, especlally Lo developlng counLrles LhaL are now sLruggllng Lo replace losL ald dollars and pay for cllmaLe change.
42 !ames S. Penry, 1!n 2012
Assumlng, conservaLlvely, LhaL global offshore flnanclal wealLh of $21 Lrllllon earns a LoLal reLurn of [usL 3 percenL a year, and would have faced an average marglnal Lax raLe of 30 percenL ln Lhe home counLry, Lhls unrecorded wealLh mlghL have generaLed Lax revenues of $189 bllllon per year more Lhan Lwlce Lhe $86 bllllon LhaL CLCu counLrles as a whole are now spendlng on all overseas developmenL asslsLance.
Cf course calculaLlons are sub[ecL Lo all klnds of caveaLs Lhe mosL lmporLanL belng LhaL Lhey lmaglne a world ln whlch developlng and developed counLrles allke are noL only Lo locaLe Lhls offshore wealLh and Lax lL, buL are also able Lo agree on reasonable rules for dlvvylng up Lhe proceeds. ln pracLlce, of course, many key source counLrles donL even have domesLlc lncome Lax reglmes ln place, leL alone Lhe power Lo enforce such Laxes across borders.
8uL from Lhls angle, Lhls sLudy acLually conLalns some good news. llrsL, ln effecL, we have acLually [usL locaLed a huge plle aL leasL $21 Lrllllon -- of unLapped flnanclal wealLh LhaL mlghL now be called upon Lo conLrlbuLe Lo Lhe soluLlon of our mosL presslng global problems.
Second, we also now know LhaL a subsLanLlal fracLlon of Lhls wealLh ls belng managed by Lhe Lop 30 players ln Lhe global prlvaLe banklng lndusLry.
1ogeLher, Lhese flndlngs may provlde an opporLunlLy Lo Lhlnk creaLlvely abouL (1) how Lo prevenL Lhe abuses LhaL have lead Lo off-Lhe-books wealLh accumulaLlon ln Lhe fuLure for example, Lhrough auLomaLlc lnformaLlon exchange, counLry-by-counLry reporLlng, and beneflclal ownershlp reglsLraLlon, and (2) how besL Lo make use of Lhe huge sLock of accumulaLed, unLaxed wealLh LhaL ls already Lhere, as well as Lhe sLeady sLream of unLaxed earnlngs LhaL lL generaLes for example, by levylng a modesL CLCu- wlde wlLhholdlng Lax on anonymous asseLs under managemenL ln Lhe Lop 30 banks, wlLh Lhe proceeds devoLed Lo ald and cllmaLe change.
Agaln, such pollcy measures are especlally lmporLanL for Lhe ma[orlLy of developlng counLrles LhaL are noL ln a poslLlon Lo be able Lo Lax lncome or wealLh on Lhelr own. lndeed, once we Lake lnLo accounL Lhe growLh of hldden offshore asseLs and Lhe earnlngs Lhey produce lnLo accounL, many ersLwhlle debLor counLrles are ln facL revealed Lo be neL credlLors of Lhe wealLhy CLCu counLrles where much of Lhls prlvaLe flnanclal wealLh has been parked, off Lhe books.
lndeed, as noLed, Lhls reporL has shown LhaL developlng world as a whole has been a slgnlflcanL C8Lul1C8 of Lhe developed world for more Lhan a decade.
lor developlng counLrles, Lhen, Lhe Lrue so-called developmenL flnance problem ls preclsely LhaL all Lhls unrecorded wealLh ls now offshore, ln Lhe hands of prlvaLe bankers and Lhelr own rapaclous ellLes. 1haL means Lhelr debL problem has really become a
43 !ames S. Penry, 1!n 2012 tax [ust|ce problem one LhaL Lhe developed counLrles have a responslblllLy as well as Lhe capaclLy Lo help Lhem solve.
III. |rate 8ank|ng on the kampage. As dlscussed above, lL Lurns ouL LhaL Lhe secreLlve offshore secLor whlch essenLlally speclallzes ln Lax dodglng and Lhe launderlng of proceeds from a myrlad of oLher dublous acLlvlLles ls noL [usL an archlpelago of exoLlc unrelaLed havens, buL a very lucraLlve global lndusLry Lhe global plraLe banklng lndusLry. 1hls lndusLry has baslcally been deslgned and operaLed for decades, noL by shady no-name banks locaLed ln lsland paradlses, buL by Lhe B)4(&# ("42'#$ H4/!"$' @",?#, as well as leadlng law flrms and accounLlng flrms. All of Lhese lnsLlLuLlons are based, noL ln lsland paradlses, buL ln ma[or llrsL World caplLals llke new ?ork, London, Ceneva, lrankfurL, and Slngapore.
As weve seen, Lhls reporL suggesLs LhaL Lhe worlds largesL banks have, lf anyLhlng, been expandlng Lhelr haven-relaLed plraLe banklng operaLlons slgnlflcanLly. even whlle offlclal lnsLlLuLlons llke Lhe C20, Lhe 8lS, Lhe lMl, and Lhe World 8ank have baslcally Lurned a bllnd eye Lo lL.
lndeed, Lhe reporL remlnds us LhaL LhaL many of Lhe very same banks LhaL were mosL deeply lnvolved ln parklng Lrllllons offshore have recenLly recelved huge publlc ballouLs.
1hus, of Lhe Lop 10 players ln global prlvaLe banklng- a|| ten recelved subsLanLlal ln[ecLlons of governmenL loans and caplLal durlng Lhe 2008-2012 perlod. ln effecL, ordlnary Laxpayers have been subsldlzlng Lhe worlds largesL banks Lo keep Lhem afloaL, even as Lhey help Lhelr wealLhlesL cllenLs slash Laxes.
Many of Lhese markeL leaders ln global plraLe banklng Lhe pracLlce of hldlng and managlng offshore asseLs for Lhe worlds ellLe have also been ldenLlfled laLely as Lhe markeL leaders ln many oLher forms of dublous acLlvlLy, from Lhe lrresponslble morLgage lendlng and hlgh-rlsk securlLlzaLlon LhaL produced Lhe 2008 flnanclal crlsls, Lo Lhe very laLesL ouLrageous scandals lnvolvlng Llbor raLe rlgglng and money launderlng for Lhe Mexlcan carLel.
All Lhls begs Lhe quesLlon of why lL ls LhaL flnanclal regulaLors conLlnue all Lhls dublous acLlvlLy Lo be perpeLraLed by Lhls very same comparaLlve handful of glanL lnsLlLuLlons Lhe llvlng embodlmenL of corporaLe serlal offenders, produclng a seemlngly endless sLream of flnanclal chlcanery on a global scale, ln counLry afLer counLry afLer counLry. ls Lhere now such a Lhlng as 1oo 8lg 1o 8e PonesL?
44 !ames S. Penry, 1!n 2012 IV. Stemm|ng the 1|de. 1hls leads naLurally Lo Lhe nexL lmpllcaLlon -- naLlon-sLaLes need Lo work LogeLher Lo Lake sLeps now Lo conLrol over all Lhls ouL-of-conLrol global flnanclal polluLlon.
ln a sense, Lhls presenLs exacLly Lhe same pollLlcal quesLlon as Lhe Lax [usLlce and global ellLe problems LhaL were noLed earller. 8uL Lhe resulLlng flnanclal mlsbehavlor ls no longer [usL a Lax [usLlce lssue lL ls a fundamenLal problem of corporaLe malpracLlce LhaL goes Lo Lhe very hearL of Lhe global markeL economy, and Lhe key players ln Lhe offshore lndusLry are aL Lhe hearL of lL.
Pow should global bank regulaLors, as well as Lax auLhorlLles, work LogeLher Lo clean up all Lhls chlcanery? 1he flndlngs of Lhls 1!n reporL calls lnLo quesLlon Lhe clalms made by C20 leaders way back ln Aprll 2009, ln Lhe lmmedlaLe afLermaLh of Lhe flnanclal crlsls, when Lhey boldly declared LhaL Lhe era of bank secrecy ls over.
1hls reporLs underscores Lhe facL LhaL Lhe era of bank secrecy and Lhe mlsbehavlor lL proLecLs ls far from over LhaL Lhere ls an urgenL need for pollcy makers Lo Lake fundamenLal sLeps Lo sLem Lhe growLh of Lhe global haven lndusLry.
As noLed, among Lhe key pollcy measures LhaL 1!n supporLs are auLomaLlc lnformaLlon exchange among Lax auLhorlLles, counLy-by-counLry corporaLe reporLlng, and Lhe deploymenL of publlc reglsLrles for beneflclal ownershlp of companles, LrusLs, and foundaLlons.
Clven Lhe lead role played by leadlng banks, law flrms, and accounLlng flrms ln enabllng all Lhls dublous acLlvlLy, global auLhorlLles musL slmply adopL much sLlffer sancLlons for Lhe repeaL offenders ln Lhls lndusLry. Lven large scale flnes have noL been effecLlve deLerrenLs we need Lo adopL much sLronger sancLlons for Lhe lnsLlLuLlons LhaL engage ln plraLe banklng mlsbehavlor and Lhe managers LhaL run Lhem.
8eyond LhaL, Lhls reporL also suggesLs LhaL Lhe llne beLween offshore and onshore Lax dodglng has recenLly been blurred by Lhe rlse of llrsL World secrecy [urlsdlcLlons llke uelaware, nevada, and Slngapore, ln addlLlon Lo LradlLlonal blackllsL nomlnees llke SwlLzerland, MaurlLlus, LlechLensLeln, and Lhe 8ahamas. ln order Lo curLall Lhe klnd of global bads deLalled ln Lhls reporL,, lL wlll be vlLal Lo curb Lhls onshorlng of offshore secrecy.
V. Mu|t||atera| kegu|ators -- M|ss|ng kesearch, Data, and W||| ower. As noLed, desplLe Lhls well-documenLed exploslon ln Lhe slze and growLh of offshore prlvaLe wealLh, lL ls slmply a scandal LhaL offlclal lnsLlLuLlons llke Lhe 8ank for lnLernaLlonal SeLLlemenLs, Lhe lMl, Lhe World 8ank, Lhe CLCu, and Lhe C20, as well as leadlng cenLral
43 !ames S. Penry, 1!n 2012 banks, have devoLed so llLLle research Lo Lhls flnanclal black hole ln Lhe global economy.
ln prlnclple, lnsLlLuLlons llke Lhe World 8ank, Lhe lMl, Lhe uS lederal 8eserve, Lhe 8ank of Lngland, and Lhe 8ank for lnLernaLlonal SeLLlemenLs noL only have ample analyLlc resources, lncludlng scores of economlsLs. 1hey also have much of Lhe daLa needed Lo esLlmaLe Lhls secLor more carefully.
lor reasons of Lhelr own, however, Lhey have LoleraLed Lhe growLh of Lhe offshore secLor far Loo long -- lL has been lefL up Lo nCCS llke 1!n Lo supporL Lhe klnd of deLalled, palnsLaklng facLual analysls LhaL underlles Lhls reporL
lL ls Llme for Lhese lnsLlLuLlons Lo llve up Lo Lhelr promlses, and work wlLh organlzaLlons llke 1!n on a research and pollcy agenda LhaL flnally glves Lhls offshore secLor Lhe aLLenLlon lL deserves.
Colng forward, a key nexL sLep wlll be slmply be Lo demand more Lransparency from global publlc lnsLlLuLlons on Lhls sub[ecL. 1hey need Lo release more of Lhe daLa Lhey already have on lLs slze and growLh, and Lhey need Lo devoLe more serlous resources Lo lLs sLudy.
VI. Summary. lL ls common for researchers Lo conclude each and oLher every sLudy wlLh a call for more sLudles!
lndeed, we have done LhaL here, parLly [usL because we are exclLed abouL Lhe progress reporLed here, and parLly [usL because we have uncovered so many more lnLeresLlng puzzles.
Powever, by calllng for more careful analysls and sLudy, we donL wlsh Lo dlsLracL aLLenLlon from Lhe many lmporLanL flndlngs LhaL, we belleve, are already secure enough Lo [usLlfy acLlng upon Lhem.
lor example, we already know LhaL Lhe black hole represenLed by offshore flnanclal wealLh ls much larger Lhan anyone has prevlously deLermlned.
We already know LhaL lL has grown large enough Lo have a powerful lmpacL on lnequallLy, Lhe dlsLrlbuLlon of Lhe Lax burden, publlc flnances, and pollLlcal lnfluence across Lhe globe.
We already know LhaL Lhls secLor has been deslgned, operaLed, and pollLlcally defended by an lnfluenLlal, well-organlzed global plraLe banklng lndusLry.
46 !ames S. Penry, 1!n 2012 We already know LhaL Lhls lndusLry now operaLes, for all lnLenLs and purposes, off Lhe books, and beyond Lhe effecLlve reach of Lodays publlc regulaLors and Lax auLhorlLles.
ln shorL, Lhls huge, secreLlve offshore lndusLry has Lruly become Lhe dark slde of globallzaLlon.
We all share a collecLlve responslblllLy now Lo redouble our efforLs Lo geL lL under conLrol.