Implications of Inequality Aversion for International Climate Policy

Transcription

1 Discussion Per o Imictions of Ineuity Aversion for Interntion Cimte Poicy Crsten Vot nd Bodo Sturm

2 Discussion Per o Imictions of Ineuity Aversion for Interntion Cimte Poicy Crsten Vot nd Bodo Sturm Downod tis ZEW Discussion Per from our ft server: tt://ft.ew.de/ub/ew-docs/d/d050.df Die Discussion Pers dienen einer möicst scneen Verbreitun von neueren Forscunsrbeiten des ZEW. Die Beiträe ieen in einier Verntwortun der Autoren und steen nict notwendierweise die Meinun des ZEW dr. Discussion Pers re intended to mke resuts of ZEW reserc romty vibe to oter economists in order to encoure discussion nd suestions for revisions. Te utors re soey resonsibe for te contents wic do not necessriy reresent te oinion of te ZEW.

3 on-tecnic Summry Any soution to te robem of vountry cooertion in interntion cimte oicy does not ony ve to overcome te notorious free rider incentives, but so s to comy wit te notions of euity nd ustice. Ineuity version, i.e. te disutiity of vin ier or ower yoff tn oters, is one imortnt concet in tis re of oter-rerdin references. In tis er, we nyse te imictions of ineuity version for interntion cimte oicy. For tis urose, we extend te Fer nd Scmidt (999) mode of ineuity version to te context of fu eteroeneity, i.e. yers in our ubic ood mode re owed to differ wit resect to teir initi endowments, teir derees of ineuity version nd to teir mrin benefits nd costs of contributions to te ubic ood. Hence, te mode deveoed in tis er ctures essenti sects of re word interntion cimte oicy. Tere, eteroeneous countries fce te oortunity to contribute to te roduction of te ob ubic ood cimte rotection. Te urose of tis er is two-fod: Firsty, we derive sime nytic conditions tt ow us to identify te circumstnces under wic erfecty eteroeneous countries re wiin to contribute to te ubic ood cimte rotection. Secondy, we ceck weter tese conditions od for imortnt neotitin rties in te re word, e.. Cin, te EU, Russi nd te US. Te min teoretic resut is tt country s benefit s to exceed some ic vue s necessry condition for contributin to te ob ubic ood cimte rotection, wereby tis ic vue deends on country s deree of ineuity version, its osition in te yoff rnkin, its costs of contributin nd oter countries benefits. It is excty tis condition tt is used for te emiric test. We emoy dt from ordus RICE, reionised imct ssessment mode, to estimte emiric benefits nd costs of cimte rotection for different reions. Te dt inut for te deree of ineuity version is tken from te exeriment economics iterture. We ten sow tt for mor countries invoved in interntion cimte oicy suc s Cin, te US nd Russi, te necessry condition for contributin is vioted t est for te time orion unti Tus, tou ineuity version teoreticy ennces te rosects for vountry cooertion in rovidin ubic oods, it is dvisbe not to overestimte its otenti for overcomin cooertion robems in re word cimte oicy.

4 Ds Wictiste in Küre Eine Lösun des Kooertionsrobems in der oben Kimoitik erfordert neben einer deuticen Senkun der Treibussemissionen eine von den Beteiiten s fir emfundene Aufteiun der mit der Emissionsreduktion verbundenen Lsten. Ein wicties Konet ur Abbidun von Firnessräferenen ist Uneiceitsversion, d.. der utenverust, der ddurc entstet, dss mn eine öere oder niedriere Ausun s ndere Akteure t. In diesem Pier untersucen wir die Auswirkunen von Uneiceitsversion uf die interntione Kimoitik. Hieru erweitern wir ds Uneiceitsversion-Mode von Fer und Scmidt (999) uf den F voständier Heteroenität. Akteure in dem untersucten Öffentices-Gut-Sie untersceiden sic insictic irer Anfnsusstttun, dem Ausmß n Uneiceitsversion und den uten und Kosten des Beitrs um öffenticen Gut. Dmit bidet ds untersucte Mode e Asekte der reen Kimoitik b. Auc ier ben eteroene Akteure (Länder) die Möickeit, Beiträe ur Ersteun eines öffenticen Guts (Kimscut) u eisten. Im Pier wird unäcst teoretisc eeit, wece kritisce Bedinun erfüt sein muss, dmit eteroene Akteure bereit sind, um öffenticen Gut Kimscut beiutren. Im Anscuss wird überrüft, ob die ereeitete kritisce Bedinun für wictie Akteure in der reen Kimoitik,.B. Cin, EU, Russnd und USA, erfüt ist. As kritisce Bedinun knn ereeitet werden, dss der uten us dem Beitr um Kimscut für ein Lnd einen kritiscen Wert übersteien muss. Dieser kritisce Wert ist bäni von der Uneiceitsversion des Lndes, seiner Position in der Rnordnun der Pro-Kof- Einkommen, seinen Kosten des Beitrs und dem uten er nderen Länder. Diese kritisce Bedinun wird dnn einem emiriscen Test unteroen. Hieru nuten wir Dten des reionisierten Kim-Ökonomie-Modes RICE von Wiim ordus. RICE iefert reionisierte uten und Kosten für untersciedice Kimscutsenrien. Der Dteninut für Uneiceitsversion stmmt us der exerimenteen Wirtscftsforscun. Wir können s entres Resutt eien, dss für wictie Verndunsrteien in der interntionen Kimoitik wie Cin, USA und Russnd die kritisce Bedinun für den Beitr um öffenticen Gut Kimscut nict erfüt ist umindest bis Ds bedeutet: Obwo Uneiceitsversion teoretisc die Kooertionscncen bei der Bereitsteun öffenticer Güter eröt, ist in der reen Kimoitik nict dmit u recnen, dss sic die Aussicten uf Kooertion durc die Berücksictiun von Uneiceitsversion deutic verbessern. 2

5 Imictions of Ineuity Aversion for Interntion Cimte Poicy Crsten Vot nd Bodo Sturm b,c Dertment of Business Administrtion, Bocum University of Aied Sciences b Dertment of Business Administrtion, Leii University of Aied Sciences c Centre for Euroen Economic Reserc (ZEW), Mnneim E-mi: crsten.vot@s-bocum.de, bodo.sturm@wiwi.twk-eii.de Abstrct: In tis er, we extend te Fer nd Scmidt mode of ineuity version to sitution were te yers differ wit resect to teir benefits nd costs from contributions to non-iner ubic ood. A necessry condition for contributin to te ubic ood is tt te yers benefit exceeds some ic vue. Usin dt from te imct ssessment mode RICE nd estimtes for ineuity version from te exeriment iterture, we sow tt tis condition fis to od for mor countries invoved in interntion cimte oicy. JEL cssifiction: C72, D63, H4, 54 Keywords: Cimte oicy, ubic ood me, ineuity version, vountry cooertion Acknowedements: Te utors woud ike to tnk Andres Lne for efu comments. 3

6 Introduction Any soution to te robem of vountry cooertion in interntion cimte oicy does not ony ve to overcome te notorious free rider incentives, but so s to comy wit notions of euity nd ustice (e.. Posner nd Weisbc 200). Sever imortnt rties invoved in te cimte tks, ike te G77/Cin, strive for n euitbe burden srin. Witin te EU, notions of euity so yed roe wen te intern burden srin for EU members ws fixed fter te dotion of te over EU reduction tret of 8% under te Kyoto Protoco. Witin te cdemic sere, studies ike Lne et. (2007) rovide emiric suort for te ide tt euity is n imortnt issue in te neotition ren of cimte oicy. Furtermore, neotitors my be truy motivted by some notion of euity: Dnnenber et. (200) find emiric suort for te yotesis tt cimte neotitors sow retivey i derees of ineuity version s rominent exme of euity references. Moreover, brod evidence from exeriment economics suests tt sinificnt frction of subects is to some extent uided by euity concerns suc s ineuity version. Subects reect i mounts of money in te utimtum me (Güt et. 982, Oosterbeek et. 2003), nd tey mke ositive contributions in te dicttor me (Knemn et. 986, Ene 200). Fiscbcer nd Gäcter (200) rovide stron emiric evidence for te exntory ower of ineuity version. Teir study demonstrtes tt te tyic observtion of decinin contributions in reeted ubic ood mes cn esiy be exined since yers in knowede of teir co-yers disyed ineuity version in revious rounds, on wic future beiefs cn be bsed. Ineuity version s been introduced s n nytic concet in economics by Fer nd Scmidt (999, erefter ced F&S). In teir iy infuenti er, tey ssume tt ctors derive disutiity from uneu outcomes. Tereby, yers my sow n version to bot disdvnteous nd dvnteous ineuity. Te interestin imiction of tis ide is tt it rovides n esy nd eent exntion of wy yers soud vountriy cooerte in soci diemm. Te reson is uite sime: If yer is sufficienty verse to dvnteous ineuity, e wi simy bstin from exoitin te free rider oortunity. Te cooertion-enncin effect of tis tye of references s redy been sown for iner ubic ood me by F&S temseves. Lne nd Vot (2003) sow tt tis effect However, tese observtions my be mbiuous. As Lne et. (200) oint out, te use of euity eri in cimte neotitions my be motivted by ure economic sef-interest. I.e. te reference for some euity erion my to re extent be exined wit considertions of minimisin te costs of imementin trety obitions. In tis er, we tke te ide of ineuity version seriousy, i.e. we do not ssume tt it cn be erfecty exined by urey strteic considertions tt infuence te brinin outcome. 4

7 so ers in more comicted two-ste -yer coition me. Teir nysis is, owever, restricted to symmetric countries. Lne (2006) oints out tt more severe intriccies in sustinin cooertive outcomes re te conseuence if te yers re owed to be eteroeneous. In tis er, we extend te F&S mode of ineuity version to te context of fu eteroeneity, i.e. yers in our non-iner ubic ood mode re owed to differ wit resect to teir initi endowments, teir derees of ineuity version nd tis is new to teir mrin benefits nd costs of contributions to te ubic ood. Hence, te mode deveoed in tis er ctures essenti sects of re word interntion cimte oicy. Tere, countries fce te oortunity to contribute to te roduction of te ob ubic ood cimte rotection. However, countries in te re word re obviousy iy diverse wit resect to teir economic wet s we s teir benefits nd costs of cimte rotection. Tus, secific deree of ineuity version my ve different effects on te countries incentives to contribute. Te urose of tis er is two-fod: Firsty, we derive sime nytic conditions tt ow us to identify te circumstnces under wic erfecty eteroeneous countries re wiin to contribute to te ubic ood cimte rotection. Secondy, we ceck weter tese conditions od for imortnt neotitin rties in te re word, e.. te US, te EU, Russi nd Cin. We re rimriy concerned wit te time orion unti 2055 since tis time eriod domintes te current cimte neotitions. Te min teoretic resut is tt country s benefit s to exceed some ic vue s necessry condition for contributin to te ob ubic ood cimte rotection, wereby tis ic vue deends on country s deree of ineuity version, its osition in te yoff rnkin, its costs of contributin nd oter countries benefits. It is excty tis condition wic is used for te emiric test. We emoy dt from ordus RICE mode, reionised imct ssessment mode, to estimte emiric benefits nd costs of cimte rotection for different reions. Te dt inut for te deree of ineuity version is tken from te exeriment economics iterture. We ten sow tt for mor countries invoved in interntion cimte oicy, te necessry condition for contributin is vioted. Tus, tou ineuity version teoreticy ennces te rosects for vountry cooertion in rovidin ubic oods drsticy, it is dvisbe not to overestimte its otenti for overcomin cooertion robems in re word cimte oicy. 5

8 Tis er is structured s foows. In Section 2, we rovide te non-iner ubic ood mode wit eteroeneous yers nd derive te necessry condition for contributin to te ubic ood. In Section 3, we investite te rosects for vountry cooertion in cimte oicy usin te derived teoretic frmework nd emiricy estimted vues for ineuity version, nd benefits nd costs from cimte oicy. In Section 4, we discuss te resuts nd drw concusion. 2 Te mode In tis er, we de wit non-iner ubic ood mode in order to investite te rosects for n interntion cimte reement. In tis mode, ctors re endowed wit some initi oction wic cn be ket for rivte consumtion or be devoted to te roduction of some ubic ood. Let tere be i =,..., yers nd consider n rbitrry yer. Ec yer is endowed wit some mount of rivte ood. Let = x denote te mount of rivte ood tt yer devotes to te roduction of te ubic ood y. Hence, x is te mount of te initi endowment tt yer kees for rivte consumtion. Te ubic ood y is roduced ccordin to some non-iner roduction function y = were = = Te yers yoffs consist of te mounts of te rivte nd te ubic ood tey consume, i.e. ( x, y) = x y π. Te rmeter denotes te mrin oortunity costs for te investment in te ubic ood, i.e. te oss of rivte consumtion if yer decides to devote. one more unit of endowment to te ubic ood y. Te rmeter mesures te roductivity of yer s individu contribution nd te term ' is s benefit from investin one more unit of endowment to te ubic ood. Wit resect to rmeters nd te roduction function, we ssume nd ' > > ', were ' mesures te mrin soci benefit from n investment to te ubic ood by yer. Te yoff for yer ten resuts s ( ) = π (). ote tt, if decides to invest one unit of endowment to te ubic ood, e cretes ositive externity: A oter yers receive n ddition yoff ccordin to teir mrin benefit. Pyers wo ct economicy rtion nd sefis wi not tke tese externities into 6

9 ccount. Since te mrin soci benefit of contributin ' exceeds te mrin costs of costs of contributin, wic exceed te rivte mrin benefit of contributin ', s te mount of te ubic ood rovided coectivey in s euiibrium ( = 0 ) fs eff sort of te efficient eve ( = ). However, te resut sketced out bove ony ods for stndrd references. F&S introduce te ide tt ctors my be verse to ineuity. 2 Tey formise te ide of ineuity version by introducin te foowin utiity for yer, iven te yoffs for oter yers i: U π (2). (, π i ) = π mx { πi π, 0} mx { π πi 0}, i i For, > 0, (2) imies tt yer derives disutiity from ineuity. Te second term reresents disutiity from disdvnteous ineuity (in cse of π > π ), wie te tird term i refects disutiity from dvnteous ineuity (in cse of π > π ). Wit resect to i rmeter, F&S ssume in ddition <. 3 ow et us y tis utiity function to te non-iner ubic ood me. In tis er, we ow for fu eteroeneity of yers, i.e. te yers differ wit resect to teir derees of ineuity version, tey my fce different initi octions nd, most imortnty, te yers re crcterised by vryin benefits nd costs of contributions to te ubic ood. Let us consider n rbitrry yer witin te yoff order enerted by te vector of initi endowments nd te vector of contributions by yers enery iven by i =,...,. Pyer s utiity is U, (3). : π > π : π > π ( π π, π ) = π ( π π ) ( π π ) 2 Boton nd Ockenfes (2000) so roose n infuenti mode of ineuity version. Our coice of F&S mode is simy motivted by its reter mtemtic simicity. 3 Te ustifiction for tis restriction is uite usibe. For simifiction, ssume two-yer cse nd tt one yer s ier yoff tn te oter. Aowin = woud men tt te first yer is rered to trow one euro wy in order to reduce is dvnte. > imies tt e is even wiin to trow wy more tn one euro. 7

10 8 Tereby, denotes yers wo rnk bove yer in te yoff ierrcy nd denotes tose yers wo rnk beow. Hence, > π π π π : is te tot disdvnteous ineuity weited by nd normised by, wie > π π π π : is te tot dvnteous ineuity weited by nd normised by. By insertin yoffs ccordin to () nd (3), we cn derive yer s utiity s = U,, = (4). We re interested in te condition under wic yer decides to contribute vountriy to te roduction of te ubic ood. To mke contributin wortwie, it must not y off to devite from ny iven ositive contribution eve 0 >. In rticur, it must not y off to contribute ess tn. Tus, et us consider n rbitrry devition beow. Te utiity in te cse were yer cooses to devite is iven s = U,, (5). We use te foowin roximtion for : = =. Te utiity from devitin resuts s foows:

11 9 = U,, (6). ow, we ve to comre (4) nd (6). Cery, contributin ys off if > > U U 0,,,, (7). (7) cn be soved for sever vribes. If we tret,, nd s iven rmeters, it mit be interestin to sove te term for. Tis ows us to investite ow te osition witin te yoff ierrcy ffects te decision to contribute or not. Reconisin tt, we et 0 if, ) ( > < (8) nd 0 if, ) ( < > (8 ). (8) tes us tt yer wi ony coose to contribute if is sufficienty ow, i.e. yer rnks sufficienty i in te yoff ierrcy. However, second cse my so occur: If te mrin benefit exceeds te mrin costs, ten yer wi contribute ony if e rnks sufficienty ow. Surrisiny, tis cse is emiricy of rticur reevnce wen we consider time orion in cimte oicy unti 200. We wi come bck to tis oint ter. For our uroses, it is interestin to ve coser ook t te RHS of (8) nd (8 ). Consider

12 te first cse, i.e. > 0. We re ookin for ower bound of. Cery, for (8) to od, te RHS s to be ositive. Tus, > 0 ' > ( )( ) ' b (9). (9) is necessry condition for contributin. It sttes tt yer wi coose to contribute to te ubic ood if is mrin benefit from contributin, te ic mrin benefit ', is i enou, i.e. exceeds b. (9) is we suited to esiy study ow cnes in rmeter vues ffect te decision to contribute: An increse in mkes (9) more demndin ( b > 0 ), wie wit incresin (9) becomes ess restrictive. Te st conecture cn esiy be confirmed by ookin t te corresondin rti derivtive: b = ' [( )( )] ( ) [( )( )] 2 '. ote tt > ' imies <. Hence, b < 0. Conseuenty, if yers re iy verse to dvnteous ineuity, teir incentives to contribute wi increse. On te oter nd, stroner disdvnteous ineuity mkes vountry cooertion in te roduction of ubic oods more difficut. Moreover, (9) is so suited to discuss te roe of te osition witin te yoff ierrcy. Obviousy, te more yers rnk bove (or beow) yer in te yoff ierrcy, te ier (or ower) te vue nd te more (or ess) demndin is (9). Te intuition beind te comrtive sttics is uite esy: Devitin downwrd from some ositive contribution eve eds to ier bsoute yoff for te devitin yer. So ceteris ribus, te disdvnteous ineuity is reduced. Te ier, te stroner tis effect is. On te oter nd, devitin cretes ddition dvnteous ineuity. Tus, ier vues of mke it ess wortwie to devite. ow consider te second cse, i.e. < 0. Ten, for condition (8 ) to od, te RHS must not exceed, te corresondin uer bound of 0. Hence,

13 < ' > ( )( ) ' b 2 (0). Obviousy, te comrtive sttics run in te sme direction s for (9). 3 Prosects for vountry cooertion in cimte oicy In tis section, we use our mode deveoed in Section 2 to investite te rosects for vountry cooertion in interntion cimte oicy. It is we known tt references ike te ones roosed by F&S ve te otenti to rety fciitte cooertion in te roduction of ubic oods. 4 For exme, Lne nd Vot (2003) sow tt even te rnd coition of countries of te word cn be stbiised s s euiibrium in two-ste coition formtion me, if yers re sufficienty verse to dvnteous ineuity. Yet te uestion of weter ineuity version rey es to ese cooertion in rctice s to be nswered emiricy. In tis section, we use conditions (9) nd (0) in order to investite weter re word countries ve incentives to contribute to te ob ubic ood cimte rotection. Remember condition (9). It sttes tt iven te vues for te version rmeters nd nd iven country s osition witin te yoff ierrcy (i.e. iven te vues of nd ) te country s mrin benefit, of (9), ', must exceed te ic vue on te RHS b, in order to contribute. Hence, in order to evute (9) nd (0) emiricy, we need vues for nd s we s for ' nd '. 3. Estimtes for ineuity version Tere re ony few studies so fr wic try to estimte emiric vues for nd. F&S temseves use dt from Rot (995) nd oters nd derive te distribution for nd ccordin to te observed beviour of roosers nd resonders in te utimtum me. Prmeter cn be derived directy from te resonder s ccetnce beviour. Te distribution over is sown in Tbe. If rooser does not know te rmeter of te 4 Te use of euity references in te context of interntion neotitions t overnment eve cn be ustified by sime oitic economy resonin: Governments or teir deetes in interntion cimte neotitions my be forced to tke euity considertions into ccount if teir ivot voter t ome sows corresondin concerns.

14 resonder but beieves tt te robbiity distribution over is te one iven in Tbe, ten te resonder s otim offer cn be derived s function of is ineuity rmeter. Given te ctu offers in utimtum mes, te distribution of cn be comuted s deicted in Tbe. 5 Tbe : Distribution of Ineuity Prmeters s Assumed by F&S distribution of s distribution of s 0 < % < % 0.5 < 30 % < % < 4 30 % % 4 < 0 % [ 0.5] = [ 0.5] = otes: Te medin vues, [0.5] nd [0.5], re derived by iner roximtion ccordin to te intervs F&S roose ( ). In teir Tbe III, F&S ssume for bot rmeters few oints in te density wit mss. Tis eds to = nd = See Section 3.4 for sensitivity nysis. We use te distribution sown in Tbe in order to comute te medin by iner roximtion of te emiric cumuted density function. Tis eds to nd [ 0.5] = [ 0.5] = We roceed wit tese vues in our nysis nd ssume tt ec reion s te sme deree of ineuity version. 6 A sensitivity nysis of our resuts wit resect to te cosen vues for nd is iven in Section Estimtes for benefits nd costs Our estimtes for benefits nd costs of cimte rotection re bsed on te imct ssessment mode RICE ( reion interted mode of cimte nd te economy ) deveoed by Wiim ordus (ordus 200). RICE views cimte cne in te frmework of te economic rowt teory. In stndrd neocssic otim rowt mode, te society invests in cit oods, tereby reducin te current consumtion so s to increse consumtion in te future. RICE modifies te rowt mode to incude cimte investments. Te cit stock of te convention mode is extended to incude investments in te environment (ntur cit). 5 F&S sow tt te distribution in Tbe is consistent wit te bevior in brod rne of exeriments. 6 Tis ssumtion seems to be wrrnted since in Dnnenber et. (200) te estimted vues for F&S rmeters of neotitors in interntion cimte oicy do not differ cross different reions of te word. 2

15 Emissions reductions in te extended mode re noous to cit investments in te minstrem mode. Tt is, concentrtions of reenouse ses (GHGs) re modeed s netive ntur cit nd emissions reductions s owerin te untity of tt netive cit. Emissions reductions ower te current consumtion, but, by reventin economicy rmfu cimte cne, increse consumtion ossibiities in te future. Te mode divides te word into 2 reions. Some re re countries, suc s te United Sttes or Cin; oters re re muti-country reions, suc s te Euroen Union or Ltin Americ. Ec reion is ssumed to ve we-defined set of references, reresented by soci wefre function, nd to otimise its consumtion, GHG oicies nd investment over time. Te soci wefre function is incresin in te er cit consumtion of ec enertion, wit diminisin mrin utiity of consumtion. Te imortnce of enertion s er cit consumtion deends on its retive sie. Te retive imortnce of different enertions is mesured usin ure rte of time reference, nd te curvture of te utiity function is iven by te esticity of te mrin utiity of consumtion. Tese rmeters re cibrted to ensure tt te re interest rte in te mode is cose to te vere re interest rte nd te vere re return on cit in re-word mrkets. Te mode contins bot trdition economic sector, ike tose found in mny economic modes, nd eoysic retionsis desined for cimte-cne modein. Bsed on RICE, sever scenrios on future cimte oicy cn be investited. For our uroses, te scenrios Bseine nd Otim re reevnt. In Bseine, no ob cimte cne oicies re doted, i.e. comete inction on interntion cimte oicies is ssumed. In Otim, cimte cne oicies mximise economic wefre wit fu rticition by ntions strtin in 200 nd witout cimtic constrints. Tus, in tis cse te most efficient cimte-cne oicies re ssumed. Tereby, efficiency invoves bncin of te btement costs nd te benefits of reduced cimte dmes. Tere seems to be some inconsistency between RICE nd our ubic ood mode. Wie RICE ssumes overnments to mximise wefre functions bsed on neocssic stndrd references, ctors in our mode re rty motivted by ineuity version. ote, owever, tt we ony use RICE to obtin emiric estimtes for te benefits nd costs. By considerin te Otim versus te Bseine scenrio, we ssume rter otimistic cse of cimte oicy were te difference between ntion benefits nd costs is re. Any more mbitious mitition oicy, e.. mitition oicy imin t te 2-deree tret, woud necessriy ed to ess fvourbe benefit-cost rtio (ordus 2008). Tus, by considerin 3

16 te Otim scenrio from RICE, we ive our ic conditions (9) nd (0), resectivey, rter ood cnce of bein fufied. Tbe 2 sows benefits nd btement costs for te 2 reions in RICE s differences between te Otim nd Bseine scenrios for two different time orions, one unti 2055 nd te oter unti 200. Te benefit vues for reion corresond to ' in our mode, since we consider discrete coice between ony two scenrios. Anoousy, btement cost vues for reion corresond to. For exme, iven time orion unti 2055, te US s 90. bn$ 2005 of ier benefits nd 48.9 bn$ 2005 of ier btement costs in Otim comred to Bseine. Tis eds to benefit-cost rtio of 0.6. Since wit oner time orion, benefits from btement (i.e. reduced dmes of cimte cne) wi increse, te corresondin benefit-cost rtio for time orion unti 200 is.06. For te EU, mitition benefits re ier (0.4 bn$ 2005 ) nd te btement costs re ower (72.4 bn$ 2005 ), edin to more fvourbe benefit-cost rtio of.40 for te time orion unti For te time orion unti 2055, ony few reions wi in net benefit in fu cooertion: te EU, Indi, Afric, nd Ltin Americ. Tese re te reions tt wi be sinificnty it by n unmitited cimte cne in te medium term. On te oter nd, mor emittin countries suc s te US, Russi nd Cin ve benefit-cost rtio tt is beow one. Given te stndrd references, tese reions wi ose in fu cooertion, since teir benefits from mitition re smer tn teir btement costs. In oter words, te fundment incentives to rticite in n interntion cimte reement re not iven. Tbe 2: Benefits nd Abtement Costs in RICE bn $ 2005 Reion CHI US RUS JP EU ID LtA AFR MidE EurA OHI Oter Benefits trou Abtement costs trou Benefit-cost rtio otes: Benefits nd costs re differences between te scenrios Otim nd Bseine in RICE (ordus 200 nd RICE mode runs vibe t tt://nordus.econ.ye.edu/ricemodes.tm). Te RICE mode contins 2 reions: Cin (CHI), United Sttes (US), Russi (RUS), Jn (JP), Euroen Union (EU), Indi (ID), Ltin Americ (LtA), Afric (AFR), Midde Est (MidE), Eursi (EurA), Oter i income countries (OHI), Oter. Benefits nd costs re resent vues usin te consumtion weited interntion re interest rte in RICE. 4

17 3.3 Evution Te min uestion of te er is, weter ineuity version is be to imrove te incentives to cooerte in interntion cimte oicy. We cn nswer tis uestion by usin our estimtes from Sections 3. nd 3.2 in order to evute weter (9) nd (0) od emiricy. ote tt te yoff ierrcy witin te ubic ood me is determined endoenousy. Tis is due to te fct tt it deends on two determinnts, te initi endowment on te one nd nd te mount contributed to te ubic ood on te oter nd. Tus, we cnnot rue out tt te ex-nte osition of country (determined soey by te iven initi endowments) my cne due to te contribution decision. But tis imies tt tere is ue number of ossibe yoff ierrcies tt woud rinciy ve to be considered. We bbrevite te nysis by tkin ony te best cse scenrio into ccount, i.e. te constetion of rmeters tt mkes te RHS of (9) or (0) cieve teir minimum vues. Obviousy, tis is te cse wen te country under considertion rnks iest in te yoff ierrcy (i.e. = 0). trou ' Tbe 3: Critic Vues nd Benefits CHI US RUS JP EU ID b ' b ' b ' b ' b ' LtA AFR MidE EurA OHI Oter trou ' b ' b ' b ' b ' b ' otes: Benefits, ', nd ic vues, b, ccordin to (9) or (0) under te best cse scenrio ( = 0). Used vues for F&S rmeters re nd (F&S 999). [ 0.5] = As n exme, te ic vue for Cin in 2055, CHI b = CHI ( US ' RUS '... Oter ' ) ( 0.288) = [ 0.5] = b CHI, is (ccordin to (9)) comuted s foows: 0.288( ) = b b Tbe 3 sows te ic vues nd te mrin benefits for te time orions unti 2055 nd 200 usin te medin vues from F&S distribution of ineuity version rmeters. Consider first te time orion unti As we cn cery see, te emiricy estimted mrin benefits do not exceed te ic vues for imortnt countries in cimte neotitions. For Cin, te US nd Russi, te ic condition (9) for contributin to te 5

18 ubic ood cimte rotection is vioted. Remember tt tis resut s been obtined under best cse ssumtions, i.e. te country under considertion rnks on to of te ob yoff ierrcy. Hence, te difference between ' nd b woud be even bier for countries ike Cin or Russi wen more reistic ositions in te yoff order re considered. On te oter nd, benefits of reions suc s te EU, Afric nd Jn re i enou to offer incentives for cimte rotection. Tis resut my rty exin wy te EU s been rticury ened in cimte oicy nd used for interntion reements so ctivey in te st. Comrin te vues witout ineuity version (Tbe 2) nd wit ineuity version (Tbe 3), we find tt for Jn, te Mid Est, EurA, OHI nd Oter, ineuity version s ositive imct on te incentives to contribute to te ob ubic ood cimte rotection. For oter countries, te initi decision wit stndrd references remins te sme if we tke ineuity version into ccount. A brief ook t te dt for 200 reves tt te rosects for cooertion in cimte rotection re muc more fvourbe in tis cse. Tere is ony one country, Cin, wic s no incentives to contribute. In ener, for countries or reions, te necessry conditions for contributin re muc ess demndin. ote, owever, tt conditions (9) nd (0) re ony necessry conditions for contributin to cimte rotection. In order to investite weter countries rey woud be wiin to enter into n interntion reement on cimte mitition, more soisticted nysis woud be necessry, e.. witin coition formtion frmework. Te ubic ood mode is we suited to demonstrte tt rerdin te time orion unti 2055, fundment incentives for cimte rotection re very imited. But it so eves oen number of imortnt uestions wic certiny ve n imct on country s decision to contribute or not, e.. te country s sie or te comosition of cimte coition nd burden srin witin suc coition. 3.4 Sensitivity nysis We use te vibe estimtes for F&S rmeters in order to rove te robustness of our resuts. Firsty, F&S temseves resent men vues for te ssumed distribution in Tbe ( = nd = ). Secondy, Bnco et. (20) use modified utimtum nd dicttor mes to obtin vues for ineuity version rmeters. Tey resent mximumikeiood estimtions of te rmeters ( ˆ = 0. 9 nd ˆ = ). However, s Bnco et. 6

19 use student subects ( = 6), te uestion of ow fr it is rorite to y tese vues to te context of interntion cimte oicy is we wort skin. Tirdy, Dnnenber t. (200) try to fiure out emiric vues for nd by skin rticints ( = 55) of te interntion cimte tks. For tis urose, tey conducted n onine exeriment wit individus wo d been invoved in interntion cimte oicy. Te exeriment consisted of two sime non-strteic mes resembin te utimtum nd dicttor mes, suited to mesure individu ineuity version s ctured in te mode by F&S. 7 Tey find tt rticints, on vere, sow considerbe version to dvnteous ineuity, wie te version to disdvnteous ineuity is moderte. Te obtined men vues re = nd = Wie te medin for te weit of te disdvnteous ineuity version strony devites from te men ( 0 ), te medin for te weit of te dvnteous [ 0.5] = ineuity version is rter simir to te men ( 0. 53). Remrkby, te study cnnot [ 0.5] = confirm sinificnt differences wit resect to eoric vriety, i.e. neotitors from reions of te word sow rter simir derees of ineuity version. Moreover, Dnnenber et. (200) sked teir rticints to imine tt decisions in teir exeriment d to be mde by members of deetion of teir country on Conference of te Prties (COP) or in meetin of te Subsidiry Bodies. By nd re, resondents exected teir country deetes to ct in more sefis wy tn temseves. Tis mens tt te vues of nd obtined in tis study cn be rerded s uer bounds. If we use te men vues from Tbe (F&S 999), no cnes wit resect to te incentives to contribute cn be observed (Scenrio S, Tbe 4 in te endix). If we dot te ineuity version rmeters obtined in Bnco et. (20), most of our resuts from Section 3.3 remin robust. As Tbe 4 (Scenrio S2) sows, bsed on best cse nysis we sti cn concude tt Cin nd te US ve no incentives to contribute to cimte rotection. Tbe 4 (Scenrios S3 nd S4) so sows te resuts if we dot version rmeter vues from Dnnenber et. (200). We sti cn sfey concude tt Cin s definitey no incentive to rticite in interntion cimte reements. For oter countries, it is not ossibe to drw sfe concusions bsed on best cse scenrio. 7 In te modified utimtum me used by Bnco et. (20), ec subect rects to n ctu roos of is or er co-yer, i.e. tere is distinct eement of strteic interction in tis me. In te modified utimtum me used by Dnnenber et. (200), owever, tere is no direct interction between bot subects nd no room for strteic considertions. As te individu ineuity version is estimted witout strteic interction between yers, it is not surrisin tt te medin vue for is very ow nd te corresondin men vue is strony infuenced by some outiers on te rit ti of te distribution. Tus, one mit be sketic weter te vues of Dnnenber et. (200) re correct mesure of te version to disdvnteous ineuity. 7

20 ote, owever, tt Dnnenber et. (200) obtined n excetiony i vue for. Remember te notion of tis rmeter: If, e.., = 0. 5 in two-yer cse, ten yer woud be indifferent between keein one dor of income for imsef or ivin te dor wy. We cn rdy imine tt > 0. 5 crcterises te beviour of countries ened in cimte oicy. Pers te study suffers from sef-seection bis, nd strony euityoriented neotitors re overreresented in teir sme. Moreover, te obtined vue for is excetiony ow, nd, rticury, te study finds >. Te buk of emiric evidence, owever, suests tt eoe suffer more from disdvnteous tn from dvnteous ineuity (Loewenstein et. 989). 4 Concusion In tis er, we eborted on te ide tt te beviour of countries in interntion environment oicy my be motivted by considertions of ineuity version. We extended te mode of ineuity version introduced by F&S to context were yers my differ wit resect to teir benefits nd costs from contributin to ob non-iner ubic ood cimte rotection. Hence, our mode ctures essenti re word dimensions of eteroeneity in te fied of interntion cimte oicy: Countries obviousy differ wit resect to teir economic wet nd to teir exected dmes from cimte cne nd teir btement costs. We sow tt country cn ony be exected to contribute to te ob ubic ood cimte rotection if its benefit is i enou. We evute tis condition emiricy by tkin estimtes for benefits nd costs from te imct ssessment mode RICE nd ineuity version rmeters from te exeriment economics iterture. We sow tt emiricy, te ic condition for contributin fis to od for mor yers in interntion cimte oicy excet te EU nd Jn. Our resuts re uite robust wit resect to vritions of te deree of ineuity version. Even wit extreme vues, t est Cin, te word s biest CO 2 emitter, never contributes. Hence, tou ineuity version is teoreticy ein wy to sove te cooertion robem in soci diemms, we soud be crefu not to overestimte its otenti for sovin re word neotition robems in cimte oicy. Finy, we woud ike to discuss briefy wic oicy essons mit be erned from tis er. From te viewoint of interntion cimte oicy, te uestion of ow vountry cooertion in cimte rotection cn be cieved sti remins. Our mode suests tt country s benefit ys cruci roe comred to oter rmeters, rticury te 8

21 derees of ineuity version. Hence, n ccurte mesurement of country s benefits nd costs from cimte oicy is rereuisite for te country to mke rtion decision. Hence, it is iy imortnt tt overnments re we-informed bout te exected dmes fcin teir country. Te views on tis toic re eviy infuenced by te roress in cimte-reted reserc. A better understndin of te cimte system, its cnes nd its imct on societies s we s dttion otions my ter te views on dmes. For exme, if new insits from science suest tt te robbiity for ctstroic cimtic events is incresin, tis my cne te benefit-cost rtio nd, ence, so infuence country s contribution decision in te ubic ood me. On te oter nd, if e.. te btement costs decrese due to tecnic roress, tis my so cne country s decision to contribute or not to cimte rotection. Hence, it mit be wortwie to tink bout rorite reserc oicies to romote tecnic roress wic woud mke emissions reductions ceer. Tere is tird ess obvious determinnt of te benefits. Wen ccutin future cimte dmes, economists tyicy discount future vues. Tkin into ccount tt te dmes of n unmitited cimte cne re sinificnt in te remote future, te coice of te discount fctor cruciy ffects te benefit: Te ier te discount fctor, te ower te exected dmes nd, tus, te ower te corresondin benefits from cimte oicy. Atou economists widey ree wit ordus (2008) view tt discount fctors re rorite wic re consistent wit te returns on re word cit mrkets, te coice of te discount rte is not urey scientific robem but invoves normtive decision. Tis my ive some eewy for te benefit-cost rtio due to etic resonin. 8 Furtermore, tere re stron ruments for coosin discount rte ner te very owest exected rte of return in te on run (Weitmn 999). Future reserc woud ve to imrove our knowede bout te determinnts of te benefit-cost rtio of cimte oicy nd te corresondin effects on te contributions to te ob ubic ood cimte rotection. 8 It woud be n esy tsk to ccute discount rtes necessry for mkin contributin wortwie in our mode. 9

22 5 References Brrett, S. (2005), Environment nd Sttecrft. Te Strtey of Environment Trety-Mkin, Oxford. Bnco, M., D. Enemnn, nd H.-T. ormnn (20), A witin Subect Anysis of Oter-Rerdin Preferences, Gmes nd Economic Bevior 72, Boton, G.E. nd A. Ockenfes (2000), ERC A Teory of Euity, Recirocity nd Cometition, Americn Economic Review 90, Dnnenber, A., B. Sturm nd C. Vot (200), Do Euity Preferences Mtter for Cimte eotitors? An Exeriment Investition, Environment nd Resource Economics 47, Ene, G. (200), Dicttor Gmes: A Met Study, Discussion Per, Prerints of te Mx Pnck Institute for Reserc on Coective Goods 200/07, Bonn. Fer, E. nd K.M. Scmidt (999), A Teory of Firness, Cometition, nd Cooertion, urtery Journ of Economics 4, Fiscbcer, U. nd S. Gäcter (200), Soci Preferences, Beiefs nd te Dynmics of Free Ridin in Pubic Good Exeriments, Americn Economic Review 00, Güt, W., R. Scmittberer nd B. Scwre (982), An Exeriment Anysis of Utimtum Brinin, Journ of Economic Bevior nd Ornition 3, Knemn, D., J.L. Knetsc nd R.H. Ter (986), Firness nd te Assumtions of Economics, Journ of Business 59, Lne, A., A. Lösce, C. Vot nd A. Zieer (200), On te sef-interested use of euity in interntion cimte neotitions, Euroen Economic Review 54, Lne, A., Vot, C. nd A. Zieer (2007), On te imortnce of euity in interntion cimte oicy: n emiric nysis, Enery Economics 29, Lne, A. (2006), Te imct of euity-references on te stbiity of eteroeneous interntion reements, Environment nd Resource Economics 34, Lne, A. nd C. Vot (2003), Cooertion in Interntion Environment eotitions due to Preference for Euity, Journ of Pubic Economics 87, Loewenstein, G.F., L. Tomson nd M. H. Bermn (989), Soci Utiity nd Decision Mkin in Intererson Contexts, Journ of Personity nd Soci Psycooy 57, ordus, W.D. (2008), A uestion of Bnce: Economic Modein of Gob Wrmin, Ye. ordus, W.D. (200), Economic sects of ob wrmin in ost-coenen environment, Proceedins of te tion Acdemy of Sciences 07, Oosterbeek, H., R. Soof nd G. vn de Kuien (2004), Cutur differences in utimtum me exeriments: Evidence from met-nysis, Exeriment Economics 7, Posner, E.A. nd D. Weisbc (200), Cimte Cne Justice, Princeton. Rot, A.E. (995), Brinin Exeriments, in: Ke, J.H. nd A.E. Rot (Eds.), Hndbook of Exeriment Economics, Princeton. Weitmn, M.L. (998), Wy te Fr-Distnt Future Soud Be Discounted t Its Lowest Possibe Rte, Journ of Environment Economics nd Mnement 36,

23 Aendix trou ' Tbe 4: Critic Vues nd Benefits Sensitivity Anysis,S b,s 2 b,s3 b,s 4 b ',S b,s 2 b,s3 b,s 4 b CHI US RUS JP EU ID LtA AFR MidE EurA OHI Oter otes: Benefits, ', nd ic vues, b, ccordin to (9) or (0) under te best cse scenrio ( = 0). Scenrio S: Vues used for F&S rmeters re = nd = (F&S 999). S2: Vues used for F&S rmeters re ˆ = 0. 9 nd ˆ = (Bnco et. fortcomin). S3: Vues used for F&S rmeters re 0 nd (Dnnenber et. 200). S4: Vues used for F&S [ 0.5] = [ 0.5] = rmeters re = nd = (Dnnenber et. 200). 2