Ecological Economics

Transcription

1 Ecologcal Economcs 69 (2010) Contents lsts avalable at ScenceDrect Ecologcal Economcs journal homepage: Analyss Moral concerns on tradable polluton permts n nternatonal envronmental agreements Johan Eyckmans a,b, Snorre Kverndokk c, a Hogeschool-Unverstet Brussel HUB, Stormstraat 2, B-1000 Brussels, Belgum b Katholeke Unverstet Leuven, Centrum voor Economsche Studën Naamsestraat 69, 3000 Leuven, Belgum c Ragnar Frsch Centre for Economc Research, Gaustadalléen 21, 0349 Oslo, Norway artcle nfo abstract Artcle hstory: Receved 13 August 2009 Receved n revsed form 16 Aprl 2010 Accepted 23 Aprl 2010 Avalable onlne 22 May 2010 JEL classfcaton: D63 Q54 Keywords: Tradable emsson permts Internatonal envronmental agreements Non-cooperatve game theory Moral motvaton Identty We nvestgate how moral concerns about permt tradng affect an endogenous polluton permt tradng equlbrum, where governments choose non-cooperatvely the amount of permts they allocate to domestc ndustres. Poltcans may feel reluctant to allow permt tradng and/or may prefer that abatement s undertaken domestcally because of moral concerns. Ths wll have an effect on the ntal permt allocatons, and, therefore, on global emssons. The mpact on global emssons depends on the precse formulaton of the moral concerns, but under reasonable assumptons, we show that global emssons may ncrease. Thus, dong what s perceved as good does not always yeld the desred outcome. However, ths can be offset by restrctons on permt tradng when governments have moral concerns about ths trade Elsever B.V. All rghts reserved. 1. Introducton Economsts typcally prefer trade n polluton permts to conventonal, non market-based envronmental polcy nstruments, as both n theory (Montgomery, 1972) and n practce (Schmalensee et al., 1998), market-based polcy nstruments have been shown to foster cost-effectveness. However, many non-economsts such as envronmental organsatons and poltcal partes have objected aganst permt trade. Some consder t a way of avodng one's oblgatons, to pay others to clean up, or to pay ndulgence, see Goodn (1994), oras Hansen (2009) put t: [ ] t actually perpetuates the polluton t s supposed to elmnate. 1 Several exstng multnatonal tradable permt schemes contan restrctons on permt tradng. These may have been ntroduced as a consequence of sgnatores beng reluctant to allow full tradng. 2 In the Kyoto protocol on greenhouse gas emssons for nstance, trade n polluton permts s allowed, but only as a supplement to natonal Correspondng author. Tel.: ; fax: E-mal addresses: Johan.Eyckmans@HUBrussel.be (J. Eyckmans), Snorre.Kverndokk@frsch.uo.no (S. Kverndokk). 1 Carsten Helm ponted out to us that Luftverschmutzungsrechte tradable ar polluton permts was on place 3 n the 2004 votng for the Unwort des Jahres (uglest word of the year) n Germany, see 2 Competng explanatons based on market power arguments have also been put forward, see Ellerman and Wng (2000). mtgaton. 3 Also n the European Emsson Tradng Scheme (ETS), access to buyng emsson reductons n thrd party countres (JI Jont Implementaton for economes n transton and CDM Clean Development Mechansm for developng countres) s lmted. 4 Thus, the cost-effectve volume of trade may not be wthn reach. In ths paper we wll study the mplcatons of moral consderatons about permt tradng n an nternatonal cap-and-trade market. These moral consderatons are assumed to play a role only n the governments' decson makng process; not n the proft maxmzaton objectve of the permt tradng frms. In a frst stage of the model, governments are assumed to negotate on natonal emsson targets at an nternatonal forum. The emsson celngs are allocated freely to the natonal ndustres, whch can trade the permts n an nternatonal and compettve polluton permt market n the second stage of 3 Artcle 6.1 of the orgnal Kyoto Protocol text states The acquston of emsson reducton unts shall be supplemental to domestc actons for the purposes of meetng commtments under Artcle 3. However, later meetngs of the Conference of the Partes (CoP) have not been able to fnd a consensus on a more precse or quanttatve meanng of ths supplementarty requrement. 4 Under Phase II of the ETS ( ), some EU member states have lmted access to CDM credts for the nstallatons on ther terrtory. For Phase III ( ), a strcter lmtaton s n place requrng that no more than 50% of the total EU reducton effort over the perod can be covered by credts generated by project based mechansms n thrd countres. More detals can be found at europa.eu/rapd/pressreleasesacton.do?reference=memo/08/ /$ see front matter 2010 Elsever B.V. All rghts reserved. do: /j.ecolecon

2 J. Eyckmans, S. Kverndokk / Ecologcal Economcs 69 (2010) the model. The set-up s smlar to the one of Helm (2003) and mples that the overall emssons, and hence the total number of polluton permts, s determned endogenously and not gven exogenously as s the case n most of the exstng lterature on tradable permt markets. Our model s partcularly suted to descrbe the ongong negotaton process to forge a follow up agreement to the Kyoto Protocol and the Copenhagen Accord. Most lkely, a new agreement wll combne natonal emssons targets wth provsons for emssons tradng. Moral consderatons are captured n our model by means of dentty effects (Akerlof and Kranton, 2000, 2005; Brekke, et al., 2003). Governments are assumed to care about the dentty of ther naton as ths reflects the vews of ther voters. There may be dfferent reasons for moral concerns as dscussed below, and we focus on two ways to specfy dentty effects that we thnk capture many arguments that have been rased. Frst, we assume that a naton's dentty s negatvely affected by ts net permt trade volume. Ths reflects the dea that some do not lke tradng polluton permts smply because they beleve one should not trade an essentally non-desrable tem lke lcences to pollute. But, at the same tme, we assume that countres are wllng to trade off the better dentty aganst benefts from exchangng permts. In our second formulaton of moral consderatons, we assume countres' denttes depend on ther actual abatement effort at home. Thus, they beneft n dentty terms from abatng at home nstead of buyng emsson reductons elsewhere. Ths s specfed as the dentty dependng negatvely on the gap between countres' actual and deal emsson levels, the latter reflectng what they thnk they deally should do. Ths formulaton captures concern for the state of the global envronment and one's own responsblty for t. We study how moral consderatons may affect global emssons n an endogenous permt tradng equlbrum, and we fnd that the effect on global emssons depends on how these concerns are specfed. Wth dstaste for permt tradng, moral consderatons wll most lkely lead to hgher, nstead of lower, global emssons n an nternatonal clmate agreement wth permt tradng. The reason s that on average, permt mportng countres have a stronger ncentve to overallocate ther domestc ndustres to reduce permt purchases compared to the ncentve of permt exportng countres to underallocate thers. However, f the concern of countres s based on a preference for domestc abatement and global responsblty, global emssons wll go down. Ths concluson s based on an equal concern n all countres, whch wll probably not be the case. If only a few countres share ths concern, the benefcal envronmental effect may be rather small. Both these results show that good ntentons do not always yeld the desred results. Whle restrctons on permt tradng as ntroduced n the Kyoto Protocol and the ETS, may ncrease the costs of reachng an emssons target, we show that such restrctons may actually reduce global emssons when governments are morally concerned about tradng permts. The reason s that the ncentve of permt mporters to overallocate domestc ndustres to lmt permt purchases has been reduced. Thus, f the restrctons result from reluctance to trade, they may actually be good for the envronment. The paper s organsed n the followng way. We frst dscuss possble reasons behnd the reluctance to trade polluton permts. In Secton 3, we present the model, whle Secton 4 derves conclusons on how moral concerns about emsson permt tradng change governments' behavour n the permt tradng market and, therefore, global emssons. Secton 5 studes the mpacts on global emssons from restrctons on buyng permts, and the fnal secton concludes. 2. Ethcal reasonng, norms and dentty Markets for polluton permts have been recognsed by several authors as a case n whch there may exst some reluctance or even repugnance aganst transactons, see, e.g., Goodn (1994), Bénabou and Trole (2007) and Roth (2007). But why may there be reluctance aganst trade n polluton permts? We can dentfy several possble reasons for ths, some of whch are based on arguments rooted n a more procedural vew of justce, and others n a more consequentalst ethcs framework. Based on a procedural justce approach, one can argue that ndustralsed countres have created the global warmng problem, and that t s ther duty to reduce the consequences of t, even f ths does not mnmse overall costs of takng acton. Ths can be used as an argument aganst developng countres sellng permts to ndustralsed countres because the permt trade would not lead to abatement n the countres responsble for the problem. Another argument s based on unfar background condtons (see Kverndokk, 1995, and Eyckmans and Schokkaert, 2004). Even f two partes agree to trade permts, the trade may not be justfed on ethcal grounds. A voluntary agreement between two partes s not necessary far f s entered nto condtons that are not far (Pogge, 1989). Background justce s not preserved when some partcpant's basc rghts, opportuntes or economc postons are grossly nferor. 5 Under the Kyoto Protocol, for nstance, some may argue that ths s the case for some CDM contracts, as ths s a trade between poor and rch countres. 6 Another lne of arguments hnges on consequentalst ethcs. Standard economc analyss s bascally about consequences and f the consequences of a partcular polcy are postve (.e., ncreased welfare), economsts wll recommend t. Ths s the case wth emsson permts. The basc argument n favour of permt trade s that t s cost-effectve (Montgomery, 1972). Partes nvolved n permt trade would get lower abatement costs than f they had to mtgate the emssons wthn ther geographcal boundares. Thus, cost-savngs wll be welfare mprovng, everythng else equal. One possble explanaton of the resstance to permt trade s, therefore, that welfare mprovements followng from permt trade have not been communcated well enough,.e., ths pont of vew s based on lack of nformaton. However, explanatons based on allegedly negatve sde effects of a permt market, may also be plausble. Buyng CDM quotas,.e., greenhouse gasses polluton permts n countres that dd not subscrbe to bndng emsson lmts n the Kyoto Protocol, may have adverse effects based on lack of an emsson baselne, lack of ncentves to undertake emssons reductons by the developng countres, transacton costs and carbon leakages. As a result, CDM projects may not fully offset emssons, see Rosendahl and Strand (2009). Related to ths s cheatng or non-complance n the permt market that also has negatve envronmental mpacts, see van Egteren and Weber (1996). Further, hot ar, meanng that some countres receve an ntal emsson quota allocaton that exceeds ther actual emssons, has also been mentoned as a reason to avod emsson tradng as tradng hot ar wll not reduce emssons. Some papers argue that f the permt allocaton s set n a noncooperatve equlbrum, permt tradng may actually lead to hgher emssons, see Helm (2003) and Holsmark and Sommervoll (2008) Ths can partcularly be the case for sellers of permts that may allocate more permts than n the non-cooperatve case wthout permt trade. One example s hot ar as mentoned above, when several countres got allowances hgher than ther busness-as-usual emssons. The dea s that f the permt prce s not very senstve to ncreases n permts and f the margnal damage of the country s relatvely low, the benefts from overallocaton s hgher than the costs. 5 Some examples can be kdney trade or an agreement between a prosttute and her/hs customer. 6 Some argue that t s not far that the developed countres take all the low hangng fruts and the developng countres are left wth the more expensve mtgaton optons n a possble future agreement. An economc treatment of ths long hangng frut argument can be found n Naran and van 't Veld (2008).

3 1816 J. Eyckmans, S. Kverndokk / Ecologcal Economcs 69 (2010) Countres may also be opposed to permt trade because of nequalty averson. If for nstance they are concerned about ncome nequalty and f nequalty ncreases n a trade regme, they may prefer not to trade. Further, permt trade may reduce ncentves to behave green. AsHansen (2009) ponts out, ndvdual actons to reduce the carbon footprnt wll not have any mpacts, as all you do s to free up emsson permts for someone else as the total amount of emssons s gven the socety. Thus, there may be some crowdng out of moral motvaton. Abatng at home nstead of buyng emsson permts may also be perceved as a better polcy based on consequences. Arguments that have been rased n ths debate are the postve spllover effects of technology development by natonal abatement as well as the ancllary benefts (reducton n local emssons, traffc accdents, congeston etc.) of abatng at home. 7 It s further argued that unlateral abatement may lead to smlar behavour by other countres, t may affect postvely the negotaton clmate n the nternatonal polcy arena, and t may reduce the conflct of nterest wthn a country as t actually shows the true costs of abatement, a cost that economc agents have an ncentve to exaggerate; see Hoel (1991) and Golombek and Hoel (2004). The dscusson above provdes arguments, based on both procedural farness and consequentalst ethcs, aganst permt tradng. We summarze ths dscusson by reducng t to two basc statements that we wll use to formalze moral concerns: 1) People mght dslke permt tradng, and 2) People mght prefer to do abatement at home. The frst statement s weaker than the second as reluctance to trade permts does not necessary mean that countres care about the envronment. However, reasons to avod tradng are often based on a preference for envronmental values, and n the second clam, the major motvaton s to save the envronment, ndependent of nternatonal agreements and quota trade. We do not clam that these statements are true, nor that there are good ethcal arguments aganst permt tradng. Instead, we take an agnostc stand by takng these arguments as gven as we thnk they descrbe some of the reluctance observed n the poltcal debate on permt tradng. Also note that n the followng, we assume that governments perform the moral reasonng on behalf of ther voters, so that moral concerns by the people of a naton are reflected n government's polces. If a substantal share of voters has moral concerns about permt tradng, t may be reflected n the government's decson, for nstance because n democratc regmes poltcans care about ther re-electon chances. The statements above may consttute a norm aganst tradng polluton permts. A norm s defned as a standard of rght or wrong, and norms are closely related to the preservaton of dentty. Thus, by modellng a norm aganst permt trade, we buld on the theory of dentty and moral motvaton (Akerlof and Kranton, 2000, 2005; Brekke, et al., 2003). In ths way we can model both the reluctance to trade wth polluton permts, and gven that the agent does not follow the norm, the wsh to reduce the trade even f t s economcally proftable. As the reasons for reluctance to trade may be based on both procedural farness as well as consequences, we do not try to endogense the norm n the model, meanng that we do not determne why a socety chooses a certan norm aganst polluton permt tradng. 3. The endogenous permt market allocaton model Our framework s based on Helm (2003) who presents a model of nternatonal emssons tradng n whch countres choose the amount 7 Related to ths s the envronmental justce argument, that mnorty groups suffer from permt tradng because they lve close to pollutng facltes and wll, therefore, not beneft from the ancllary benefts of abatng at home (see, e.g., Kverndokk and Rose, 2008, for a survey). of permts they allocate to ther domestc ndustry n an endogenous and non-cooperatve way. 8 We expand Helm's model by ntroducng moral concerns (dentty consderatons). Countres are assumed to be reluctant to trade emssons permts and/or may dslke the fact that they emt more than what they thnk they deally should do. The basc buldng blocks of the model are as follows. There are n countres or governments engaged n negotatng a future nternatonal envronmental agreement ncludng provsons for polluton permt tradng. The governments represent the vew of ther voters who are, to some extent, reluctant to trade permts. However, they are also wllng to trade f the benefts from trade are large enough. As n Helm (2003), we make the assumpton that the permt tradng regme s establshed only by unanmous approval of all countres. 9 In the frst stage, governments choose non-cooperatvely ther ntal emssons allocaton. Ths set-up resembles closely the realty of nternatonal clmate negotatons, n partcular the process leadng to the 1997 Kyoto Protocol, the 2009 Copenhagen Accord, and presumably also future nternatonal clmate polcy negotaton rounds. Another example s the ETS where EU member states had to draft Natonal Allocaton Plans (NAPs) suggestng a permt allocaton for all nstallaton covered by the ETS drectve on ther terrtory for Phase 1 ( ) and Phase 2 ( ). In contrast to most of the lterature on natonal emsson permt schemes, the overall number of permts cannot be assumed to be exogenous n an nternatonal context. In the second stage of the game, ndvdual frms trade emssons n a compettve permt market. We assume that they obey the emsson celng and that they maxmze ther prvate proft only. Thus, frms do not have moral concerns about tradng. 10 Ths game s solved by backward nducton,.e., we start be solvng the second stage Stage 2: Frms tradng emssons In every country N={1,2,,n} there s a large number of dentcal frms that maxmzes profts due to emssons, π, takng as gven the emssons tradng scheme: max e π ðe ; p; ω Þ = B ðe Þ + p½ω e Š ð1þ Let ω 0 be the ntal allocaton of permts to the representatve frm n country and assume that permts are allocated for free. For smplcty, we set the emssons of one country equal to the emssons of ts representatve frm, e [0,e ], where e s busness as usual (BAU) emssons,.e., emssons n absence of an nternatonal treaty and any moral consderatons. Ths s the maxmum emssons that a country wll emt. 11 Emssons can be traded at a gven market prce p. Frms choose emssons n order to maxmze profts,.e., the sum of benefts and net permt trade revenues. The benefts of emssons, B, can be nterpreted as a producton functon. Producton requres nput of carbon emssons, and s assumed ncreasng and strctly concave n emssons: B e 0 and B ee b0. In the Appendx, the formal propertes of 8 For other applcatons of the model, see Holsmark and Sommervoll (2008) and Carbone et al. (2009). 9 Allowng for endogenous coalton formaton would substantally complcate the analyss, as the countres are not symmetrc n our analyss. Symmetrc or homogeneous countres are a common assumpton n studes of coalton formaton, see, e.g., Barrett (2005), but would be napproprate n our case snce the dfference between countres s essental n the analyss of polluton permt tradng. 10 Ths s consstent wth Sebert (1992, p. 130) and Rauscher (2006) who argue that a frm spendng resources on socal actvtes not rewarded by the market wll not reman compettve and wll be drven out of the market. However, f frms have market power, non-proft motves can survve. In contrast to ths, Brekke and Nyborg (2008) argue that corporate socal responsblty can lead to recrutment of hghly motvated workers and, therefore, be proftable. 11 Note that wth moral concerns on abatement at home, emssons wll be less than e even n the absence of a clmate treaty, see Secton 4.2.

4 J. Eyckmans, S. Kverndokk / Ecologcal Economcs 69 (2010) the beneft functon are derved from the standard propertes of an emsson abatement cost functon. From the frst-order condton for proft maxmzaton, we can derve the demand for emssons: B e ð Þ = p e = E ðpþ = B e 1 ðpþ e Proft maxmzng behavour by frms leads to cost-effectveness; margnal benefts of emssons are equalsed across frms. Thus, ths condton s vald even f countres have moral arguments aganst permt tradng as such consderatons are not taken nto account by the compettve frms. Comparatve statcs of ths expressons shows that emssons are decreasng n the permt prce: B ee de = dp e p = Ep = 1 B ee b 0 Usng frst-order condton (2), we can defne a net supply of permts functon that s ncreasng n the prce of permts: S ðpþ = ω E ðpþ wth S p = Sp = E p N 0 ð4þ A permt market equlbrum defnes a prce level such that total net supply of permts s nonnegatve: p o such that S j p o h = ω j E j p o 0 ð5þ Ths market equlbrum condton mplctly defnes a prce functon mappng a vector of emsson allocatons ω nto the market clearng prce level: p o =ρ(ω). We assume that the margnal beneft functons are such that for every vector of emsson allocatons, there exsts a unque equlbrum permt prce. 12 The permt prce functon can be shown to be decreasng n the ntal allocatons of permts to a country. ρ ω ðωþ = ρ = dp = 1 ω dω b 0 Hence, n accordance wth standard economc ntuton, hgher allocatons of permts lead to a decrease n the equlbrum permt prce Stage 1: Governments choosng ntal permt allocatons Gven the smoothly workng permt market n stage two, governments negotate n stage one on the ntal allocaton of permts, and we assume that they choose the number of permts ω as to maxmze the followng natonal welfare functon: W e ; ω ; ω = B e ð Þ + p½ω e ð2þ ð3þ ð6þ Š D ω + ω + I e ; e S ; ω ð7þ The functon D denotes polluton damages ncurred n country. These damages are convex n global emssons, defned by the total amount of permts dstrbuted (country 's permts are ω, whle all other countres' emsson permts are denoted by ω ). Thus, the envronmental problem s caused by a unformly mxng pollutant as n the case of global warmng. We assume that country maxmzes ts welfare, takng as gven the permt allocatons by all other countres 12 If more permts would be allocated than the net demand for emssons, we assume that the equlbrum prce s zero: ω j N E j (0) p o =0. (ω =ω ). Hence, we are lookng for a Nash equlbrum n permt allocatons among natonal governments. Our set-up s smlar to the model ntroduced by Helm (2003). However, n addton to the approach by Helm, we assume that countres have moral concerns about permt tradng defned as an dentty, I, whch adds postvely to ther welfare functon. Identty s usually defned as a person's self mage as an ndvdual or as a part of a group (Akerlof and Kranton, 2005). Identty has been recognsed as mportant for ndvdual behavour n felds as socal psychology and socology, but has only recently been adopted n formal economc models. In ths paper we defne a country's moral concerns n the same way as the dentty of an ndvdual as governments perform the moral reasonng on behalf of ther voters. The model also bears some resemblance to the lterature on voluntary provson of publc goods, see Bergstrom et al. (1986), and n partcular to the strand of lterature on the so-called warm glow of gvng, see Andreon (1990). In ths lterature, economc agents care about the overall level of the publc good and feel good about contrbutng a postve amount to ts provson. One can nterpret ths warm glow effect as a postve dentty effect (see Brekke et al., 2003). However, n our model, prvate contrbutons, n the sense of low allocaton of polluton permts to one's domestc ndustry, may lead to a cold feelng nstead of a warm glow as ths ncreases the ncentves for frms to enter the permt tradng market and buy large amounts of permts. Based on the statements made n Secton 2, the moral concern (dentty) of a country s a functon of ts actual emssons, e, ts permt allowances ω as well as ts deal emssons, e S. The latter s defned as the emsson level that the country would lke to am for based on ethcal reasonng. Ths deal s consdered exogenous. We return to ths n Secton 4. Usng the notaton ntroduced before, we can wrte the welfare functon of country, takng nto account the compettve permt tradng n stage two, as: W ðω ; ω Þ = B ðe ðρðωþþþ + ρðωþ½ω E ðρðωþþš D ðω + ω Þ + I E ðρðωþþ; e S ð8þ ; ω Each country wants to set ts ntal permt allocaton, ω N0, n order to maxmze ts welfare defned by Eq. (8). Let ΔI be the change n dentty for a change n ω. The frst-order condton for an nteror soluton s, therefore: 13 B e E p ρω + ρ ω ½ω e Š + ρ 1 E p ρω ω D + ΔI =0 ð9þ Usng the frst-order condton B e =ρ of compettve permt tradng among frms n stage 2 (Eq. (2)), the followng condton should be satsfed for all countres N: ρ ω ½ω e Š + ρ D ω + ΔI =0 ð10þ The frst term on the left hand sde (LHS) s the effect of addtonal permt allocatons on the emsson tradng revenue through the effect on the permt prce. A more generous permt allocaton s benefcal for permt mporters (the market prce goes down), whle a more restrcted permt allocaton s benefcal for permt exporters (the market prce goes up). We label ths effect the strategc permt tradng effect. 14 The second term s the drect prce effect of a more generous permt allocaton. Every addtonal permt s worth the prevalng market prce ρ. The thrd term stands for the addtonal polluton damage effect 13 All varables and functons are evaluated n the Nash equlbrum of permt allocatons,.e., the permt allocatons of other countres are taken as exogenous. 14 Note that ths effect cancels out n the aggregate and wll not have any effect on global emssons, see Helm (2003), p. 2741, or Eq. (14) below.

5 1818 J. Eyckmans, S. Kverndokk / Ecologcal Economcs 69 (2010) caused by a more generous permt allocaton. More permts lead, ceters parbus, to hgher global emssons and hence hgher polluton damages. Ths effect s therefore negatve. Fnally, the last term captures the change n the moral concerns of extra permts. Ths effect can be postve or negatve dependng on the precse specfcaton of the dentty functon. Rearrangng Eq. (10) and usng agan the frms' proft maxmzng frst-order condton B e =ρ, we obtan: B e = D ω ρ ω ½ω e Š ΔI ð11þ Hence, every country chooses an ntal permt allocaton such that ts margnal beneft from the last ton of emssons equals ndvdual margnal damages, corrected for a strategc permt trade effect and an dentty effect. 4. Dfferent formulatons of natons' moral concerns In ths secton we present the man results for dfferent formulatons of countres' dentty concerns. We start n Secton 4.1 wth the dentty functon based on reluctance to trade permts. Both the symmetrc (Sectons and 4.1.3) and asymmetrc formulaton (Secton 4.1.2) are explored. Secton 4.2 focuses on dentty consderatons based on preference for abatement at home. Fnally, Secton 4.3 combnes both formulatons of the dentty functon Identty as reluctance to trade Symmetrc reluctance to trade So far we have not specfed the dentty functon. Let us consder the two statements on moral concerns from Secton 2: Countres mght dslke permt tradng, and countres mght prefer to do all the abatement at home. However, for ease of exposton, we study the two statements separately before combnng them. We frst focus on the statement that countres dslke permt tradng. Assume a symmetrc formulaton of reluctance to trade,.e., countres dslke both sellng and buyng permts: I ðe ; ω Þ = F δ½ω E ðρðωþþš 2 f ω E ðρðωþþ 0 otherwse ð12þ Involvement n permt tradng represents a cost, both for buyers and sellers, due to the fact that one does not act n accordance wth one's moral convcton. Ths loss n dentty conssts of a fxed cost ndependent of the amount of permts traded, and a varable cost. The fxed cost, F N0, s the loss of gong from one regme to another, here represented by gong from a non-trade regme to a tradng regme. However, the volume of trade also matters. If a country decdes to trade, t feels less comfortable the hgher the volume of permt tradng s when δn0. 15 An example can be the supplementary condton n the Kyoto Protocol as well as poltcal dscussons n Norway on settng a lmt on how much one can reduce abatement abroad. Fnally, note that ths dentty functon has a maxmum at zero wthout emssons tradng. The specfcaton of the moral concerns, both the fxed and the varable term represent costs of tradng. Ths shares smlartes wth a permt trade system wth transacton costs as n Stavns (1995). Wth transacton costs, the volume of trade s lower and welfare s lower compared to a system wthout such costs. Also the ntal allocaton of permts may affect the outcome of tradng. However, moral consderatons affect permt tradng dfferently than transacton costs n several respects. Frst, the fxed dentty term affects the 15 In the case where δ=0,.e., there s an dentty cost of not followng the norm, whch s ndependent on the volume of trade as long as the volume s postve, we wll actually get the same frst -order condtons as when dentty does not matter,.e., ΔI =0. decson whether the country wants to take part n the permt tradng market, and second, the endogenous part of the dentty functon affects the allocaton of ntal allowances as countres do not want the allocatons to be very dfferent from actual emssons. However, for a gven level of aggregated allowances (global emssons target), the outcome of tradng s not affected by the ntal allowance allocaton. In our model, frms face no transacton costs n trade so they trade cost-effectvely,.e., margnal abatement costs among sources are equal. Ths s not the case n models wth transacton costs as these costs are usually modelled as a functon of the volume of trade. Usng ths explct dentty functon, we can derve the followng lemma: Lemma 1. If countres are reluctant to trade permts, and f dentty s symmetrc, then: f country s a permt seller (ω Ne ) t follows that B e ND e ; f country s a permt buyer (ω be ) t follows that B e bd e. Ths means that sellers underallocate and buyers overallocate emssons compared to a stuaton wthout permt tradng and wthout moral concerns. Proof. Note frst that f there were no permt tradng and no moral concerns, maxmzng welfare n Eq. (7) would mply B e =D e. Now ntroducng permt trade and moral concerns, the change n dentty from a margnal ncrease n ω s gven by: ΔI = 2δ[1 E p ρ ω ][ω e ]. Ths change s postve for permt buyers and negatve for permt sellers for the symmetrc dentty functon because 0 E p ρω = E p = 1. It follows from Eq. (11) that: B e De = ρ ω ½ω e Š +2δ 1 E p ρω ½ ω e Š = ½ω e Š ρ ω +2δ 1 E p ρω fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} þ and therefore B e ed e f ω ee. The lemma says that permt sellers allocate permts so that the margnal beneft from emssons s hgher than the margnal damage, whle t s the other way around for permt buyers. The ntuton s as follows. Net permt sellng countres tend to underallocate ther domestc frms, yeldng margnal beneft from emssons n excess of margnal costs, as ths makes permts scarce and drves up the equlbrum market prce. In addton, the underallocaton has postve dentty effects as the volume of trade goes down and the gap between permt allocaton and actual emssons shrnks. On the other hand, net permt buyng countres tend to overallocate ther domestc frms because ths makes permts more abundant and lowers the market prce. Further, the same dentty mechansm as descrbed for sellers s also vald for buyers; overallocatng permts has postve dentty effects as the volume of trade goes down. Summarzng, the moral concerns, f they only stem from reluctance to trade, renforce the strategc trade ncentves for both sellers and buyers of permts. Ths proves to be a useful result for the remander of the paper Asymmetrc reluctance to trade How does the ntroducton of moral concerns affect the global amount of permts ssued nto the market? We start wth a smple case based on the fact that the dscusson on the acceptablty of permt tradng s manly a topc n countres that are potental permt buyers. Ths means that the dentty functon s asymmetrc, where countres only suffer an dentty loss f they buy permts: I ðe ; ω ; βþ = F δ½ω E ðρðωþþš 2 f ω be ðρðωþþ 0 otherwse ð13þ

6 J. Eyckmans, S. Kverndokk / Ecologcal Economcs 69 (2010) Proposton 1. If only buyers are reluctant to trade (asymmetrc dentty functon), then every ndvdual country wll emt more and global emsson wll be hgher than n the endogenous permt allocaton equlbrum wthout moral concerns. Hence, h nρ = D e j 2δρ ω ω j e j ð14þ Proof. The frst-order condtons for governments ssung permts are dfferent for permt mporters (ω be ) and exporters (ω e ), see Eqs. (11) and (13): ( B e D e + ρ ω ½ω e Š 2δ 1 E p ρω ½ ω e Š =0 f ω be B e D e + ρ ω ½ω e Š =0 f ω e Summng over both types of countres and usng the market clearng condton from Eq. (5), t follows that: n oh B e j D e j 2δ mn 0; ω j e j 1 ρω =0 Let ω N be global emssons and top scrpts I and O refers to the equlbrum wth moral concerns and wthout such concerns respectvely. Then assumng, n contrast to the clam n the proposton, that ω I N bω o N, and usng the convexty of the damage functons, t follows that: D e j (ω I N )b D e j (ω o N ). Summng over the approprate frst-order condtons for both equlbra, see Eq. (10), t can be shown that (recall that 0 [1 ρ ω ] 1): ρ I ρ o b 2δ n oh n mn 0; ω j e j 1 ρ ω b0 Hence, the equlbrum permt prce would be lower wth asymmetrc moral concerns than wthout. Gven that the equlbrum prce functon s decreasng n the global permt allocaton, we get ω I N Nω o N, whch contradcts the ntal assumpton. Therefore, ω I N ω o N, and moral concerns wll lead to hgher global emssons than wthout such concerns. As we have shown that the equlbrum permt prce wll be lower wth asymmetrc moral concerns than wthout, every country's representatve frm wll emt more: B e (e I )=ρ I bρ o =B e (e o ) e I o Ne due to concavty of the beneft functons. The proposton s ntutvely clear. From Lemma 1 we know that permt buyers have an ncentve to overallocate ther domestc ndustres because of 1) strategc trade consderatons (drvng down the equlbrum permt prce), and 2) moral concerns (overallocatng domestc frms reduces the amount of permts that has to be mported). Snce only buyers' moral concerns are taken nto account n the asymmetrc dentty functon, global emssons 16 n the nternatonal Nash equlbrum wll be hgher than n the scenaro wthout such consderatons Symmetrc reluctance to trade revsted We now turn back to the more complcated case where both permt mporters and exporters dslke permt tradng,.e., the symmetrc dentty functon. In order to fnd the global effect, we summarse all countres frst-order condtons based on Eq. (10) and ΔI = 2δ[1 E p ρ ω ][ω e ], to obtan: n ρ ω ρ ω h h h ω j e j + ρ D e j 2δ ω j e j h ω j e j + nρ nh ω j e j +2δρ ω nρ D e j +2δρ ω D e j 2δ o =0 h ω j e j =0 1 ρω h ω j e j o =0 16 We want to remnd the reader that global emssons refer to actual emsson. At the global level, the net trades of permts cancel out n a permt market equlbrum, see Eq. (5). As the followng result shows, the outcome depends on the balance of power between permt exporters and mporters. 17 Proposton 2. If countres are reluctant to trade permts, dentty s symmetrc, and [ω j e j ] N(b)0, global emssons wll be lower (hgher) and every country wll emt less (more) than wthout moral concerns. Proof. Assume the clam s false,.e., [ω j e j ] N0 and ω I N Nω o N. Becauseofconvextyofthedamagefunctonsffollowsthat: D e j (ω I N )N D e j (ω o N ). Usng Eq. (14) and the approprate condton when there s no dentty functon (.e., the second term of the rght hand sde of Eq. (14) s equal to zero), t s easly shown that: h nρ I +2δρ ω ω j e j N nρ o ρ I ρ o N 2δρω n h ω j e j N 0 Hence, the equlbrum permt prce wth moral concerns would be hgher than the prce wthout such concerns. Gven that the equlbrum prce s decreasng n the global permt allocaton, ω I N bω o N, whch contradcts the ntal assumpton. Fnally, as we have just shown that f [ω j e j ] N0, global emssons wll be lower and, hence, the equlbrum prce of permts wll be hgher n case of moral concerns and, therefore, every country's representatve frm wll emt less: B e (e I )=ρ I Nρ o =B e (e o ) e I be o due to concavty of the beneft functons. Hence, f [ω j e j ] N0, reluctance to trade leads to lower overall emsson allocatons, and hence lower emssons, than n the absence of moral consderatons. But how should we nterpret the condton [ω j e j ] N0? It can be nterpreted as a weghted average of all permt trades, where permts exporters have [ω j e j ] b0 and mporters[ω j e j ] N0, and where the weghts,, are gven by the nverse of the slope of the margnal beneft of emssons functon (recall that =1/B ee j b0). Thus, n order for [ω j e j ] to be postve, permt exporters should, on average, have smaller absolute values of E p than permt mporters. Note that hgh absolute values of B ee (.e., steep margnal emsson abatement cost functons) mply small absolute values of E p. Therefore, the term s postve f permt sellers are predomnantly countres wth steep margnal abatement cost functons (see the Appendx). Ths s not very lkely n the Kyoto permt market. Most emprcal models predct the contrary,.e., that low abatement cost countres (.e., countres wth flat margnal beneft functons, B e ) wll export carbon emssons permts, see Böhrnger (2002), Den Elzen and de Moor (2002) or Eyckmans and Hagem (2008). Therefore, t s more lkely that moral concerns would lead to a hgher number of permts ssued. Ths means that the soluton wth moral concerns s lkely to result n hgher global emssons than n the absence of those consderatons. The result s agan due to the overallocaton of permts n permt mportng countres that follows from Lemma Note that n the followng propostons we do not do comparsons to a frst-best socal optmum as n general the frst-best soluton wth and wthout moral concerns would dffer. One excepton s the frst-best soluton n Secton 4.2 where the two solutons concde.

7 1820 J. Eyckmans, S. Kverndokk / Ecologcal Economcs 69 (2010) Identty as a preference for abatement at home Reluctance to trade s one aspect of a country's moral concerns, but the country could also have preferences for dong the abatement at home. To model ths, we assume that dentty depends on the relatonshp between actual emssons and the morally deal emssons, e S,.e., the amount of emssons the country thnks t deally should am for. Ths can be specfed n the followng way: h I e ; e S = γ e e S 2 ð15þ As the countres are concerned about the global envronment, t s reasonable to assume that the deal requres substantal abatement. One way to specfy ths s to follow Brekke et al. (2003) and assume that countres share an ethcal vew that global socal welfare should be maxmzed. Thus, deal emssons are found by maxmzng a utltaran global welfare functon where everybody follows the same general rule, namely to emt the amount that maxmzes global welfare: maxð e1 ;e 2 ; ;e n Þ B e B j h 2 e j D j e k k N γ e j e S j Ths gves rse to the followng frst-order condtons: e h 2γ e e S = D e j e k k N N We see that the global welfare functon s maxmzed for e S =e.thus, n ths case the dentty terms dsappear, and the frst-order condtons equal the well known Samuelson rule (after Samuelson, 1954) that defnes the frst-best allocaton of emssons wthout moral concerns: B e e = D e j e N k k N ð16þ Havng defned the deal reference level of emssons, we can now wrte the dentty functon n the followng way, where e s consdered exogenous: I e ; e = γ E ðρðωþþ e 2 ð17þ Usng ths functon n the maxmsaton problem defned by Eq. (7), we can show that every country wll emt more than the deal emssons level. Lemma 2. Even f countres care about ther deal emssons, every ndvdual country wll emt more than ts deal: e e and, therefore, total amount of permts allocated wll exceed the socally optmal level: ω N e N. Proof. Assume, on the contrary, that N:e be. From the strct concavty of the emssons beneft functon, t follows that B e (e )NB e (e ). Usng Eqs. (2) and (17) mples: ρω ð N Þ = B e ðe Þ N B e D e j e = At the same tme, we can derve: D e ðω N Þ +2γ e e p E ρω = B e ðe Þ N B e ð ω N Þ D e j e N e D ð ω N D e j e N Þ D e j ω N N e N = ρ e N ωn be N e N e = D e j e N N 2γ e e p E ρω 0 Ths contradcts the prevously establshed nequalty. Thus, for N:e e. Gven that every country emts more than the deal level, the total emssons n the endogenous permt allocaton equlbrum wll exceed the frst-best level: ω N Ne N. As every country always emts more than ts deal, we can easly sgn the dervatve of the dentty functon wth respect to ω : ΔI = 2γρ ω e e p E b0 ð18þ Based on ths, we fnd that when all countres care about ther deal effort, the endogenous permt allocaton equlbrum shfts towards the Pareto effcent frst-best allocaton of emssons, and they wll emt less than wthout moral concerns. Proposton 3. If countres care about ther deal effort, every ndvdual country wll emt less wthout moral concerns and global emssons wll be lower. Proof. Assume, n contrast, that ω I N Nω o N. Usng convexty of the damage functons, t follows that: D e j (ω I N )N D e j (ω o N ). As before, usng the approprate frst-order condtons for both equlbra, we fnd: ρ I ρ o N 2γρω n h e j e j N 0 Hence, the equlbrum permt prce would be hgher wth moral concerns than wthout. As the equlbrum prce s decreasng n the global permt allocaton, we fnd that ω I N bω o N whch contradcts the ntal assumpton. Therefore, we get ω I N ω o N. Also, as the equlbrum prce of permts wll be hgher wth moral concerns, every country's representatve frm wll emt less: B e (e I )= ρ I Nρ o =B e (e o ) e I be o due to concavty of the beneft functons. Ths result s n contrast to the prevous concluson when countres cared about ther level of emssons tradng and moral concerns could lead to hgher global emssons. The reason s that, when countres care about ther deal level of emssons, they feel a warm glow when dong more than what a strctly prvate cost beneft analyss would prescrbe. Ths shows that t matters what people's concerns are about. If the man concern s dstaste aganst tradng emsson permts, the envronment may be harmed, whle f the desre s to reduce emssons at home, the envronment wll beneft. Note, that n ths latter concluson we have assumed that all countres share the same moral standpont. Ths wll not necessary be the case. If only a few countres have ths vew, the addtonal abatement would share smlartes to unlateral actons taken by one sngle country, and t s known from the lterature that unlateral actons from one country can actually ncrease global emssons (Hoel, 1991) Combnng both dentty effects One may argue that people's moral concerns follow both from dstaste aganst tradng permts as well as the desre to reduce emssons at home. Thus, t seems approprate to combne the dentty functons from the prevous sectons. Taken together, the full dentty functon can be wrtten as: 18 8 I e ; e ; ω < F δ½ω E ðρðωþþš 2 γ½e ðρðωþþ e Š2 f ω E ðρðωþþ = : γ½e ðρðωþþ e Š2 otherwse ð19þ 18 Here we chose the symmetrc verson of the reluctance to trade dentty functon.

8 J. Eyckmans, S. Kverndokk / Ecologcal Economcs 69 (2010) The new dentty functon descrbes an nternal conflct; we can have e =e, but stll ω e, thus dong what s best accordng to one moral deal may not fulfll the other. We can now characterze the full mpacts of the moral concerns. Proposton 4. If countres are reluctant to trade permts and f they care about abatement at home, global emssons wll be lower f ether: [ω j e j ] N0, or the concerns about abatement at home are strong enough to compensate the reluctance to trade. Proof. The proof s trval by combnng Proposton 2 and Proposton 3. At frst sght, one mght thnk t s obvous that moral concerns would lead to lower global emssons. However, Proposton 4 shows that ths depends crucally on both the form of the dentty functon and the balance of power between permt mporters and exporters. As we have argued above, n the case of global clmate change, [ω j e j ] s lkely to be negatve. We can, therefore, conclude that the overall emsson level resultng from the nternatonal clmate negotatons outcome wll be lower f there s a relatvely strong concern about actual emsson levels n all countres. If ths concern s weak, or f few countres share ths concern, and n addton there s a strong dstaste aganst permt tradng, emssons may actually be hgher than wthout any concerns about permt tradng. Thus, amng for the good does not always brng the desred outcomes. 5. Restrctons on permt tradng As mentoned n the ntroducton, restrctons of permt tradng as found n the Kyoto Protocol and n the ETS, may have been ntroduced due to reluctance of tradng permts. From economc theory we know that restrctons on permt tradng are lkely to ncrease the costs of a treaty for a gven number of total allowances; whle there s no mpact on the envronment the, cost-effectve allocaton of emsson reducton may not be reached. Thus, restrctons are consdered to be bad from an economc pont of vew. However, ths result follows from standard economc assumptons where the permt allocaton s assumed exogenous to each country, and where countres are assumed to have no moral concerns about permt tradng. If these assumptons do not hold, wll the standard result stll preval? Overallocaton of permts by buyers was the reason for hgher emssons under moral concerns about permt tradng. Thus, we focus on the case wth symmetrc dentty functon and where buyers are restrcted from buyng permts, 19.e., ðω e Þ aðe ω Þ ð20þ Ths means that the amount of permt a country can buy should be less or equal to a certan share, a, of the necessary abatement. Assume now that ths restrcton s bndng for some countres,.e., these countres would lke to buy a larger share. In ths case we fnd: Proposton 5. If countres are reluctant to trade permts, dentty s symmetrc, and the restrcton on buyng permts s bndng for at least one country, global emssons wll be lower than n the case wthout such restrctons. 19 Note that ths restrcton s always satsfed for sellers as long as there s no hot ar. Proof. The proof follows the same lne as the proof of Lemma 1. As the frst-order condton from Eq. (11) gves B e D e = ½ω e Š ρ ω +2δ 1 E p ρω ; fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} þ we see that an effectve restrcton on buyng permts makes the term on the rght sde less negatve. Thus, the margnal beneft wll be closer to the margnal damage and there wll be less of an over allocaton. As there s a restrcton on trade, the permt mportng countres are restrcted and cannot overallocate emssons to the same amount as n the stuaton wthout the restrcton. Ths has a benefcal effect on the envronment. Thus, when allowances are set n a noncooperatve settng, and when countres have moral concerns on permt tradng, restrctons on tradng have a postve effect on the envronment,.e. global emssons are lower. Note that ths concluson also holds for the asymmetrc dentty functon. 20 Fnally, n the case wth moral preferences for abatng at home, a restrcton on permt tradng wll not affect the dentty functon. Thus, the effect on total emssons wll be the same as wthout moral concerns Dscusson and conclusons In ths paper we have analysed how moral concerns about permt tradng affect an endogenous polluton permt tradng equlbrum, n whch governments choose non-cooperatvely ther natonal permt allocatons, and therefore also the overall global envronmental objectve. Wthout moral concerns, such an equlbrum typcally leads to a cost-effectve allocaton of emsson reducton efforts (because of the permt market), but the global envronmental objectve falls short of the deal (.e., the global welfare maxmzng) envronmental ambton level, see Helm (2003). Ths modellng setup resembles closely the realty of nternatonal clmate polcy negotatons n the aftermath of the Copenhagen Accord. However, there may be dfferent reasons why people, and therefore also ther governments, have moral concerns about permt tradng. For nstance, countres may be reluctant to trade permts because they thnk t s a way to escape ther moral responsblty, or because of the assumed negatve consequences the trade may have n developng countres. Hence, both consequentalst and procedural ethcs arguments are used to justfy calls for lmts on the access to flexble mechansms lke CDM n the framework of the Kyoto Protocol or the European Emsson Tradng Scheme. Gven an nternatonally negotated permt tradng system, we fnd that moral concerns may ncrease or reduce global emssons dependng on the precse formulaton of these concerns. Ths s 20 Wth standard economc assumptons and no moral reasonng, we know that restrctons ncrease the costs of a treaty, and the ncentves to jon the treaty for ndvdual countres wll be lower. As a result, the treaty may fal or t may consst of fewer countres. The consequence may be hgher global emssons. In our settng we fnd that whle total emssons go down wth restrctons, welfare wll also go down, otherwse the restrctons would not have been bndng. Thus, also n ths case the costs of jonng the treaty wll be hgher, whch may gve a negatve ncentve to jon. However, t s not obvous that ths has negatve mpacts on total emssons n ths model, see footnote 21 below. 21 Restrctons on permt tradng may have an effect on emssons when allowances are set n a non-cooperatve equlbrum. As mentoned n Secton 2, Helm (2003) and Holsmark and Sommervoll (2008) found that emssons may be hgher n a permt trade system compared to busness-as-usual when allowances are set non-cooperatvely as n our model. Thus, restrctons on permt tradng may also lead to lower total emssons even wthout moral concerns. Ths s, however, outsde the scope of ths paper.