Energy Policy 32 (2004) 1109–1130

State responsibility and compensation for climate change damages—alegal and economic assessment

Richard S.J. Tola,b,c,*, Roda Verheyend

aCentre for Marine and Climate Research, Hamburg University, Troplowitzstrasse 7, 22529 Hamburg, Germanyb Institute for Environmental Studies, Vrije Universiteit, Amsterdam, The Netherlands

cCenter for Integrated Study of the Human Dimensions of Global Change, Carnegie Mellon University, Pittsburgh, PA, USAdResearch Unit Environmental Law (FORUM), Faculty of Law, Hamburg University, Hamburg, Germany

Abstract

Customary international law has that countries may do each other no harm. A country violates this rule if an activity under its

control does damage to another country, and if this is done on purpose or due to carelessness. Impacts of climate change fall under

this rule, which is reinforced by many declarations and treaties, including the UNFCCC. Compensation for the harm done depends

on many parameters, such as emission scenarios, climate change, climate change impacts and its accounting. The compensation paid

by the OECD may run up to 4% of its GDP, far exceeding the costs of climate change to the OECD directly. However, the most

crucial issues are, first, from when countries can be held responsible and, second, which emissions are acceptable and which careless.

This may even be interpreted such that the countries of the OECD are entitled to compensation, rather than be obliged to pay. State

responsibility could substantially change international climate policy.

r 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Climate change; Climate change impacts; State responsibility; Compensation

1. Introduction

Climate change has claimed its first victims asTuvalu’s entire population prepares for emigration.With rising sea levels, their homes and infrastructurebecome uninhabitable and unusable. As Leo Falcam,former President of the Federated States of Micronesiaput it, for those low lying islands in the Pacific andelsewhere ‘‘climate change is nothing less than a form ofslow death’’.1 This already led the Tuvalu PrimeMinister to declare that his country is seeking to suethe USA and/or Australia for damages.2 And indeed,Tuvalu’s inundation is only the first sign. Regardless ofefforts undertaken to reduce anthropogenic climatechange, mankind is committed to change. And while

the impacts resulting from climate change are stillrelatively poorly understood, they could have severeimplications for peoples and economies.The question of who will pay the damage and

compensate the refugees of tomorrow for the loss oftheir homelands, damage to their health and property aswell as potential casualties has been posed many times—mainly in a moral context. This paper provides thequestion with an international legal framework andpresents some estimates of its economic implications.Section 2 sets out what international law has on offer

to tackle the issue of damage due to impacts of climatechange. We argue that, despite remaining gaps and legalas well as factual problems, there will be a generalobligation of industrialised nations under internationallaw to compensate developing nations for damageresulting from anthropogenic climate change. Theanalysis is based on specific treaties, which may not beuniversally ratified, as well as on customary interna-tional law, which is equally applicable to all countries.We deal only with public international law, defined as

the law between nation states, as opposed to privateinternational law, which tackles issues of law between

ARTICLE IN PRESS

*Corresponding author. Centre for Marine and Climate Research,

Hamburg University, Troplowitzstrasse 7, 22529 Hamburg, Germany.

Tel.: +49-40-42838-7007; fax: +49-40-42838-7009.

E-mail address: tol@dkrz.de (R.S.J. Tol).1See http://starbulletin.com/2001/08/19/editorial/special.html2World Environment News, 30 August 2002 ‘‘US faces legal battles

as climate bogeyman’’, and ‘‘Tuvalu seeks help in US global warming

lawsuit’’, available at /http://www.planetark.orgS.

0301-4215/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved.

doi:10.1016/S0301-4215(03)00075-2

persons of different jurisdictions, and domestic law,such as tort law in distinct jurisdictions. We doacknowledge that domestic and private internationallaw might form a basis for compensation for climatechange damage, but a discussion of this is beyond thecurrent paper. For the same reason, we also disregardany potential private liability by companies or indivi-duals under either national or international law forfuture climate change damage inflicted on States orother natural persons or legal entities. This papertherefore does not prejudge the existence in law of suchprivate liability, nor do we take any position regardingthe possibility of claims seeking damages3 or injunctionsin any national jurisdiction. Instead, we argue thatStates are responsible as they (fail to) regulate emissions.Moreover, as we do not seek to analyse whether a

particular state could sue another in a court ofinternational law (such as the International Court ofJustice or the International Tribunal of the Law of theSea), we do not engage in the (valid) discussion aboutenforcement of public international law here (but seeTomuschat, 1999; Sands, 1995). Whether or not such aclaim could be successful depends purely on the specificcase and the (potential) state parties to it. It suffices tosay that the chances of a climate change related caseappearing before an international tribunal are notnegligible per se. Besides, States frequently recogniseresponsibility without being sued in court.Section 3 provides an analysis of these impacts in

economic terms. Sections 4 and 5 then sets out whatgeneral state responsibility for climate change damagewould imply in economic terms for OECD countries.Note that we disregard emissions from land use changesand potential carbon sink enhancing measures, becauseof measurement problems. We do not seek to argue thatany specific state is responsible for climate changedamage, nor do we try to define the level of responsi-bility. Instead, we present a number of alternatives,which span the range of not implausible interpretations.There are a large number of papers on responsibility

for climate change in the literature (see Banuri et al.,1996; Pinguelli-Rosa and Munasinghe, 2001; Toth,1999; Toth et al., 2001 for overviews), but most focuson moral responsibility, none are so specific on the basisfor liability in international law, and only few try toquantify potential compensations. Kemfert and Tol(2002) and Tol (2002c) look at the effect of the polluterpays principle on international climate policy, andPanayotou et al. (2002) use historical responsibility toassign emission reduction obligations. None of thesepapers has a sensitivity analysis around scenarios,

responsibilities, or compensations. Legal literature inturn has mostly avoided looking at state responsibilityfor climate change damage, exceptions being Cameronand Zaelke (1990) and Verheyen (2003), while someauthors have tried to tackle the similar legal problem oftransboundary damage from air pollution (Okowa,2000) as well as generally, the problem of internationalresponsibility for environmental harm (Francioni,1991).

2. State responsibility and environmental damage under

international law

One of the basic rules of international law is thatStates shall not inflict damage on or violate the rights ofother States. In environmental law, this is captured inthe so-called ‘‘no harm rule’’ which in turn has itsfoundations in the principle of good neighbourlinessbetween States formally equal under international law.4

Principle 2 of the 1992 Rio Declaration, which echoesPrinciple 21 of the 1972 Stockholm Declaration,reiterates this rule of customary international law,outlawing transboundary environmental injury: ‘‘States

have, the sovereign right to exploit their own resourcespursuant to their own environmental and developmentalpolicies, and the responsibility to ensure that activities

within their jurisdiction or control do not cause damage to

the environment of other States or of areas beyond the

limits of national jurisdiction.’’ (emphasis added).Although the exact boundaries and elements of thisrule are heavily discussed in international law, the basicrule exists, as emphasised by the ICJ in its 1996Advisory Opinion on the Legality of the Threat orUse of Nuclear Weapons5 and, for example, in the 3rdRestatement of US Foreign Relations Law.6 The no-harm rule also forms the basis of international environ-mental law, such as, inter alia, the 1992 UN FrameworkConvention on Climate Change (FCCC) and its 1997Kyoto Protocol.7 This rule also contains an obligationto minimise risk, i.e. to prevent harm when it isforeseeable. This is of particular importance in the

ARTICLE IN PRESS

3The term ‘‘damages’’ is used to connote the remedy sought, i.e. an

applicant would ask for damages to compensate for any damage

sustained. ‘‘Damage’’ is therefore used interchangeably with ‘‘harm’’,

‘‘injury’’ or negative ‘‘impact’’.

4This principle is accepted in international law since the famous

Trail Smelter Arbitration of 1941, Reports of International Arbitral

Awards (RIAA) III, p. 1905 and (1939) 33 American Journal of

International Law (AJIL) p. 132; (1941) 35 AJIL p. 684. See for further

analysis Perrez (1996).5 ICJ Reports (1996) at 241, para 29.6American Law Institute, Third Restatement (1987) St Paul,

Minnesota, Vol. II at 103, also available at/http://www.westlaw.orgS,section 601: State obligations with respect to environment of other

states and the common environment.7As of 18th December 2002, 187 States and the EC are Parties to the

UNFCCC, which entered into force in 1994. The Kyoto Protocol is

not yet in force, but as of 3rd January 2003, 102 States had ratified it,

representing 43.9% of the 55% CO2-emissions shares needed to put the

Protocol into force.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301110

context of adaptation to climate change, and thequestion of who has to bear the costs of such measures.Moreover, in international law, states are responsible

for violations of public international law and are obligedto compensate the indirectly or directly affected statesfor the damage caused.8 This rule forms the basis of thelaw of state responsibility, a body of law which hasrecently been codified by the International Law Com-mission (ILC, ‘‘Draft Articles on State Responsibilityfor Internationally Wrongful Acts’’9), a UN bodyentrusted with promoting the codification and develop-ment of international law.10 While the rules developedby the ILC do not automatically represent internationallaw but have to be accepted (e.g., ratified) by States, theycan serve as a useful tool to examine the conditions andconsequences of state responsibility for climate changedamage.The following section sets out the basic steps for

establishing state responsibility (II.2), which will thenprovide the basis for identifying and testing theapplicable international law (II.3). While, as notedabove, this paper does not attempt to discuss all legalproblems associated with a potential claim for compen-sation for climate change damage, some other issues areflagged in Section 2.4.

2.1. General steps for establishing state responsibility

A claim for damages under international must involvethe following steps: (i) Identifying the damaging activityattributable to a state, (ii) establishing a causal linkbetween the activity and the damage, (iii) determiningeither a violation of international law or a violation of aduty of care (due diligence), which is (iv) owed to thedamaged state. Step (v) in a court of law would be toquantify the damage caused and relate those back to theactivity. While this may be doable, we focus on the firstfour items, as they are sufficient to establish generalstate responsibility.

2.1.1. Damaging activity and attribution

To hold states responsible for climate change damage,it is either necessary to identify a legally relevantbehaviour by a state or to attribute the actions ofprivate persons to the state. In terms of state behaviour,(i) allowing emissions of greenhouse gases per se orduring a certain time, and (ii) not having put in place theregulatory means to arrest emissions over and above acertain threshold are both clearly legally relevant stateactions or omissions. Moreover, a breach of aninternational treaty such as the FCCC would beattributable to a state regardless of its reasons.However, it could be argued that millions of private

persons are responsible for greenhouse gas emissionsand not the state. The wording of Principle 2 RioDeclaration (see page 3) seems to be clear on this issue.It generally obliges states to ensure that no damage isdone from their territory to other states, and does notdifferentiate between state and private conduct. Or, asexpressed already in 1941 by the arbitrator in the Trail

Smelter case: ‘‘A State owes at all times the duty toprotect other states against injurious acts by individualsfrom within its jurisdiction’’.11 But many have alsoargued that states cannot assume ‘‘full accountabilityfor the actions of their citizens who, in the exercise oftheir human rights, are not subject to governmentalcontrol’’ (Tomuschat, 1999, p. 274).The customary law rules are somewhat unclear on this

issue. The ILC Draft Articles suggest that behaviour ofprivate persons, for example private industry and energyutilities emitting greenhouse gases, cannot generally beregarded as conduct of a State (Articles 4 ff). But themajority of emitting activities are subject to a licensingor permit procedure, be it in the energy or transportsector. However, Article 8 of the Draft Articles suggestthat, as soon as an activity is permitted or licensed by astate (under the control of ...), the resulting behaviour isattributable to the state because states must exercise duediligence in control of private persons is an acknowl-edged principle.12 This has been argued in particularwith regard to ultra-hazardous activities (e.g., Handl,1980; cf. Le Cahier, 1977). To argue that greenhouse gasemissions are not attributable to states would result inmajor inconsistencies. For example, it is undisputed thatemissions from state-owned electricity plants or otherindustrial plant are attributable to the state. In someparts of the world, power plants are now fullyprivatised, in others the main CO2 emitting sector isstill under state direct control. There is no reason whyinternational law would support exoneration of a state

ARTICLE IN PRESS

8See the landmark case Chorz !ow Factory, Permanent Court of

International Law (PCIL) Reports Ser. A Nr.7 (1927), p. 30 and for an

in-depth analysis of the current state of the law of State Responsibility

Tomuschat (1999).9The ILCs Draft Articles on State Responsibility are the first

comprehensive codification of the law on state responsibility. See

Chapter IV of the ‘‘Report of the International Law Commission, 53rd

session’’, General Assembly, Official Records, 56th session, Suppl. No.

10, UN Doc. A/56/10. Available on /http://www.un.org/law/ilcS. Seefor a commentary Crawford et al. (2001).10Article 13.1 of the UN Charta provides: ‘‘The General Assembly

shall initiate studies and make recommendations for the purpose of: a.

...encouraging the progressive development of international law and its

codification’’. The text of the Statute of the International Law

Commission can be obtained from /http://www.un.org/law/ilc/texts/statufra.htmS.

11See note 4, RIAA III, p. 1963.12See the findings in the Youmans Claim, IV RIAA, 110, the Janes

Claim, IV RIAA, 82, and the Massey Claim, IV RIAA, 155. These

cases were primarily concerned with the inadequate punishment of

individuals having committed crimes on foreign persons. See also

Brownlie (1983), p. 162.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1111

simply because of privatisation of the polluting activity.But more importantly, the discussion becomes academicif one accepts that monitoring and regulation of privateperson’s conduct is still a prime function of states, afunction states can fail to fulfil with the appropriatecare. This must have also been the underlying argumentfor Canada to assume responsibility in the Trail Smelter

case, where damage was caused by fumes from the‘‘Consolidated Mining and Smelting Company ofCanada, Limited’’, i.e. a private company.Thus, as a result, at least the failure to stop, reduce or

regulate emitting activities with due care can triggerstate responsibility.

2.1.2. Causation

It is useful to distinguish between general causationand specific causation, a distinction that is found inmost domestic tort laws and to some extent alsoapplicable in international law. The first type refers toa causal link between an activity and the generaloutcome. In our case, this concerns the general proofthat anthropogenic greenhouse gas emissions change theradiative forcing in the atmosphere, which results inglobal warming, which then leads to impacts onecosystems such as air temperature rise, sea level rise,shift of climatic zones, etc. This causation chain will befurther discussed below.Specific causation requires the proof that a specific

activity has caused a specific damage in order to put a‘‘figure to a claim’’ and to link this to a particular actor.This link would be needed to issue specific damageawards. As mentioned before, we cannot deal with thisissue here but will discuss general responsibility only.The following scientific facts can be taken as given

and will satisfy the requirements for general causation.First of all, there is almost universal internationalscientific consensus that anthropogenic emissions ofgreenhouse gases cause and have caused changes in theradiative forcing balance in the atmosphere, whichcauses climate change. In other words, already observedchanges are not just due to natural climate variability(Santer et al., 1996; Mitchell et al., 2001). Secondly,there is a high likelihood that global climate change willlead to impacts on ecosystems and human life. In fact,the IPCC has already found that recent regional changesin temperature have had discernible impacts on manyphysical and biological systems (Gitay et al., 2001;Smith et al., 2001) and that the observed warming in thepast 50 years has contributed significantly to global sealevel rise (Church et al., 2001). Thirdly, some damagewill occur, regardless of reduction efforts undertaken bythe international community in the framework of theFCCC or its Kyoto Protocol.Moreover, there is relative certainty about the extent

to which states as entities have contributed to emissionsof greenhouse gases and there are scenarios attempting

to reflect possible contributions for the future. Forexample, regarding historic emissions of carbon dioxidefrom fossil fuel combustion (1900–1990), Europe hascontributed 28%, the USA 30%, Japan 4% and theformer Soviet Union 14%, while Africa is responsiblefor only 3%, and South and Central America for only4%.13 While it is impossible to attribute specificemissions of a specific country to specific impacts (orharm), there is a causal link between each ton ofgreenhouse gas emitted and the change in radiativeforcing. Thus, even though the shares of contributionsdiffer and only lead to the resulting changes inaccumulation, they are equally causal in a legal sense.14

Climate science has also started to estimate damagefrom the impacts of climate change, thus providing somefigures on an aggregate level that can be linked to thelevel of general causation. In this framework, we presentimpact estimates in Section 3. One result of theseassessments is that damage in developing countries isprojected to be significant, while impacts in OECDmight overall be positive. On the basis of these aggregatefigures it would be possible to assign legal responsibilityto countries on the basis of their contributions. Whilethe figures in Sections 3–5 relate to the OECD as awhole, it is also possible to differentiate betweencountries for the assigning of responsibilities for futureharm.

2.1.3. Wrongdoing or due diligence

Once the damaging activity and causation has beenestablished, there is usually a requirement in tort orliability law to show some kind of wrongdoing orviolation of due diligence.15 With regard to environ-mental damage in international law, there are two basicviews: (i) a state has to violate a duty of care or a rule ofinternational law to incur responsibility or (ii) wheresignificant environmental injury is concerned, the causallink between injury and activity attributable to the stateis enough to trigger state responsibility and compensa-tion duties.View (i) is reflected in Article 1 of the ILC Draft

Articles on State Responsibility. It simply states that‘‘every internationally wrongful act’’ entails responsi-bility, while wrongful act is defined as a conduct thatconstitutes a breach of an international obligation

ARTICLE IN PRESS

13WRI, Contributions to Global Warming, July 2001. See for the

most recent figures FCCC/SB/2002/INF.2, National Communications

by Parties included in Annex I to the Convention, Report on national

greenhouse gas inventory data from Annex I Parties for 1990 to 2000,

available on /http://www.unfccc.intS.14This has been argued for acid rain by Ott and Paschke (1997).

National tort law has developed different types of and preconditions

for ‘causation’ that are applicable to international law only by analogy.15Negligence is a behaviour which disregards the due diligence in a

given case, i.e. deficient conduct. Due diligence must be defined

individually in response to the pertinent situation or conduct.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301112

(Article 2). When exactly an obligation is breacheddepends on the nature and character of the pertinentobligation, but for many international lawyers, theremust always be a due diligence test. That means thatstate responsibility can only occur if the respective statehas not acted with the appropriate care. In turn, thestandard of ‘‘appropriate care’’ must be determinedaccording to the actual circumstances and obligation inquestion.This is the traditional view of state responsibility; it

always includes some kind of fault on behalf of the state.However, there has been much discussion about director strict state responsibility, i.e. whether states should beliable for any damage caused by certain activities undertheir control regardless of negligence or fault (view (ii),this is sometimes also called ‘state liability’)16 (see interalia, Horbach, 1996; Arsanjani and Reisman, 1998;Lefeber, 1996; Okowa, 2000). This discussion wassupported by the endeavour of the ILC to develop ruleson harm caused by certain lawful but hazardousactivities. This project is entitled ‘‘international liability’’but has been reduced to suggest general rules on theprevention of environmental damage.17

There have also been attempts to apply the polluterpays principle, originally an economic allocation con-cept, between states. As a result, responsibility forenvironmental damage would be generated by the merecausation of damage. Interestingly, during the negotia-tions of the FCCC, it was suggested that the principlecould serve as an appropriate legal framework toaddress issues of liability and compensation.18 But sofar, the legal content of the polluter pays principleremains unclear and it is doubtful whether it can beregarded as customary international law rule applicablebetween states (Sands, 1995).Applied to the topic of this paper, the divergence

between the views is quite significant. Either, it sufficesto show that a state has contributed to causing globalwarming and thus to its (adverse) impacts. This wouldmean that one could take into account all historical

contributions, from all sources for calculating compen-sation—but would then all states be responsible, even ifthey have contributed to greenhouse gas emissions onlyto a negligible extent? Or, to trigger internationalresponsibility, a state must have disregarded duediligence, i.e. the polluting activity or allowing theactivity to take place must be shown to be negligent insome way. The challenge would then be to determineexactly what the standard of care or the threshold ofnegligence is in any given case.This paper does not solve the question of whether

damage from climate change might trigger direct stateresponsibility under international law. Rather, wepresent results for alternative legal findings, rangingfrom strict to fault-based responsibility.

2.1.4. Obligation towards a state

Finally, regarding the fourth element of the stateresponsibility test (obligation owed to a specific state), itis necessary to keep in mind that originally, publicinternational law was mostly concerned with delimitat-ing rights and duties of states, for example by definingstate boundaries and establishing the law of treatment ofaliens. Today, international law moves more and moretowards a law of cooperation. Many multilateral treatiesestablish obligations that do not correspond with pre-existing rights of specific states but that are much ratherowed to the international community as a whole.International environmental law, which in many in-stances protects global commons such as the oceansand the atmosphere, is the best example for such ‘‘erga

omnes’’ obligations (an obligation which is owedto a multitude of states and can thus be invoked bythese jointly or individually). We return to this issue inSection 2.3.

2.2. Different kinds of claims—costs for adaptation vs.

residual damage

Having shown that it is, in principle, possible to applyinternational law of state responsibility to climatechange damage, there is a need to identify the potentialtypes of damage or costs.To start with, an analysis of state responsibility

should distinguish between climate change mitigation,adaptation to climate change and residual climate changedamage. Mitigation is the prevention of anthropogenicclimate change at the source by either reducing green-house gas emissions or enhancing sink capacities(terrestrial or other). Mitigation is the focus of theFCCC and the Kyoto Protocol. In the framework of alegal analysis of responsibility for climate changedamage, mitigation can be viewed as ‘indirect’ damageprevention, concerned with actually preventing a risk ofdamage from anthropogenic climate change. As such,the costs or obligations to mitigate climate change do

ARTICLE IN PRESS

16The reason for the call for direct responsibility is the realisation

certain (lawful) activities entail a great risk for the environment (such

as transports of nuclear materials and waste, oil transport by ship, etc.)

and can cause damage to property and life in other states. Many

hazardous activities are today subject to specific treaty regimes, which

often impose strict liability on the private operators rather than the

respective state. Some also establish direct state responsibility for

potential damages, such as in the case of space objects. In many

regimes where the private operator is primarily liable for harm caused,

states also accept some form of residual responsibility, which implicitly

shows the acceptance of the concept of direct responsibility.17Draft Articles on the ‘‘Prevention of Transboundary Harm from

Hazardous Activities’’ were adopted in 2001, Report of the ILC, 53rd

session, note 9 above, p. 370 ff. See for a recent analysis Boyle (2000).18See Report of the INC, 1st session 4–14 February 1991, UN Doc.

A/AC.237/6, 6 f. and UN Doc. A/AC.237/Misc.1/Add.3 at 24,

submission by Vanatu on behalf of AOSIS.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1113

not fall within the scope of this paper. However,mitigation obligations can form the basis for stateresponsibility claims (see Section 2.3).Adaptation on the other hand can be legally defined

as direct damage prevention, as it can reduce residualdamage and thus reduce the risk of such damage.Residual damage occur when adaptation measures arenot possible or are not carried out due to economic ortechnical constraints. For legal purposes, the obligationto directly prevent damage corresponds with theobligation to compensate for any damage done.19 Sincehumankind is committed to some climate changeregardless of mitigation efforts, both are possibleresponses to a situation in which the rights or interestsof states or/and individuals are affected because ofongoing and past activities (in this case: emissions ofgreenhouse gases). Therefore, if general legal responsi-bility for climate change damage is established, suchobligation also covers adaptation measures (and costs)as direct damage prevention measures.Note that we do not engage in a discussion about

the legal definition of environmental damage andrecoverable costs. The impacts and costs listed inSection 3 and compensation calculated in Section 5 doinclude some damage (e.g., ecological damage)that might not recoverable in a legal sense. However,since the estimates used are highly uncertainanyway and differentiation of cost categorieswould be beyond this paper, we operate on theassumption that most of the damage calculated wouldbe recoverable.

2.3. Testing the applicable law

In general, international obligations can be found ineither treaty law or customary international law.Customary International Law is developed by statepractice and opinio juris, i.e. the perception of states thata certain behaviour actually reflects a rule of interna-tional law. The already mentioned no-harm principlebelongs to this category. Treaty law is the main sourceof law in international environmental law, containingmuch more defined rules as well as differentiatedobligations regarding implementation control and en-forcement—elements that are largely lacking for rules ofcustomary law. The FCCC and Kyoto Protocol, as wellas other treaties such as the 1982 UN Convention on theLaw of the Sea (UNCLOS) are relevant sources for thispaper. We will not attempt to look at any other treaties,which does not, however, preclude their potentialapplicability.

2.3.1. The climate treaties

We scrutinise the FCCC and the Kyoto Protocol bothin terms of existing direct provisions regarding climatechange damage and adaptation, and in terms of whetherthey contain obligations that would give rise to a claimfor reparation under the traditional law of stateresponsibility.

2.3.1.1. Explicit provisions regarding state responsibili-

ty. The FCCC and Kyoto Protocol provide only apartial answer to the issue of responsibility for damage.The Convention does not address the issue directly.Rather, during the negotiation process of the FCCCindustrialised nations emphasised that they would notaccept any treaty provisions hinting at state responsi-bility (Bodansky, 1993; Ott, 1996; Sands, 1992; Ver-heyen, 1997). This caused several States, upon signatureof the FCCC and the Kyoto Protocol, to make thefollowing declaration, which refers to State responsi-bility: ‘‘(y) signature of the Convention shall in no wayconstitute a renunciation of any rights under interna-tional law concerning state responsibility for the adverseeffects of climate change (y)’’.20 This reservation hadbeen proposed by the Alliance of Small Island States forinclusion in the Convention itself during the negotia-tions21 but was not included in the final document. TheKyoto Protocol contains no pertinent provisionseither.22

2.3.1.2. Explicit provisions on adaptation—direct da-

mage prevention. Both FCCC and Kyoto Protocolcontain some unique provisions relating to adaptationand funding for adaptation. Firstly, Article 4.1 (b)FCCC obliges all Parties to ‘‘formulate and implementnational or regional programmes containing measuresto mitigate climate changey and measures to facilitateadequate adaptation to climate change’’. Thus, adapta-tion is not a voluntary undertaking but a substantiveobligation on all Parties with a view to reducing futureclimate change damage. However, there is uncertainty asto what constitute ‘‘adequate’’ adaptation measures andwhen and exactly how the obligation must be met.Secondly, by ratifying the Convention, OECD coun-

tries (Annex II countries) have accepted a generalobligation to assist developing countries in meeting the

ARTICLE IN PRESS

19 In national legal systems, damage prevention is a duty upon the

person responsible for future damage, be it either as an active

obligation or only in the form of a duty to bear the costs for damage

prevention incurred by the victim or third persons.

20Declarations made by the Governments of Nauru. Tuvalu, Fiji,

and Papua New Guinea, etc. See status of ratification of the FCCC

and the Kyoto Protocol, available on /http://www.unfccc.intS.21See submission of Vanatu on behalf of the Alliance of Small Island

States (AOSIS), Elements for a Framework Convention on climate

change, in: Set of informal papers provided by delegations, related to

the preparation of a framework convention on climate change, UN

Doc. A/AC.237/Misc.1/Add.3, at 22.22The same holds true for the Intergovernmental Panel on Climate

Change (IPCC). Some OECD countries consistently opposed refer-

ences to responsibility in the IPCC reports, and removed any such

references from their summaries for policy makers.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301114

costs of adaptation under certain circumstances (e.g.,Articles 4.3, 4.4, 4.8, 4.9 and 11 FCCC). According toArticle 4.3 Annex II countries ‘‘shall provide new andadditional resources to meet the agreed full incrementalcosts of implementing measures y’’, which covers‘‘preparing for the adaptation to the impacts of climatechange’’ (Article 4.1 (e) FCCC) as well as the dutiesunder Article 4.1(b) mentioned above. Article 4.4 statesthat Annex II Parties ‘‘shall also assist developingcounty Parties that are particularly vulnerable to theadverse effects of climate change in meeting costs ofadaptation to those adverse effects’’.23 Thus, while theobligations are neither capped nor time restricted, thetreaty does not oblige Annex II countries to bear the fullcosts of adaptation in all developing countries.Thirdly, the Kyoto Protocol establishes a special

adaptation fund that will receive revenues from theoperation of the Clean Development Mechanism (CDM‘share of proceeds’) and from voluntary commitmentsby Parties. New funds for (inter alia) adaptationactivities were established at the 7th Conference of theParties to the FCCC (COP7)24 (see Verheyen, 2002).Summing up, the wording of the treaty only foresees

partial funding of adaptation measures by Annex IIcountries. The general legal framework of responsibilityas well as burden sharing issues are not addressed.Funding pledges made are not directly connected to anyconcrete assessment of the actual aggregate adaptationneeds of developing countries. Even though the fundingprovisions of the FCCC are mandatory, thus far,funding is made available on a political basis withoutattaching it to legal responsibilities, even though the‘‘polluter pays principle’’ was suggested by countries asthe basis for determining respective countries’ sharesduring the COP6bis negotiations.25

Thus, the FCCC and Kyoto Protocol do not resolveissues of state responsibility for adaptation and residualdamage. We now proceed to look at specific law rulesand treaty provisions which could serve as the basis for

showing that a state has ‘‘done wrong’’ or actednegligent, an important element of a state responsibilityclaim.

2.3.1.3. Breach of obligations leading to state responsi-

bility. The FCCC formulates the objective of preventing‘‘dangerous anthropogenic interference with the climatesystem’’ (Article 2) but does not contain a rule thatprohibits greenhouse gas emissions per se. The muchdiscussed Article 4.2 FCCC which contains the ‘‘aim’’ ofreturning Annex I countries’ emissions of greenhousegases to 1990 levels by the year 2000 is not anenforceable duty under international law due to itsambiguous wording, and might in particular not attachany responsibility to historic emissions before the year1992 (adoption of the FCCC). Thus, even though manyAnnex I countries have not met the year 2000-target,there are no direct legal consequences attached. If,however, a country would continue to increase itsemissions continually since its ratification of the FCCC,this could amount to a breach of treaty.26

The Kyoto Protocol is not (yet) in force, but does setlegally enforceable targets for countries (although theaccounting of emissions is still ambiguous). Once it is inforce, a country that does meet its target at the end ofthe 1st commitment period (2012) has breached aninternational law obligation. However, primarily, thespecial compliance and enforcement mechanism of theProtocol would be triggered and might even preclude theapplication of general law on state responsibility,depending on the will of the parties to the Protocol.As envisaged in the accords of COP7,27 a state that doesnot meet its targets will be subject to, inter alia, a‘‘penalty’’ rate for excess emissions, i.e. will have toreduce 1.3 times as much in the next commitment periodto make good for the non-compliance. The agreementdoes not, however tackle the issue of damage and doesnot formally preclude damage claims to be based on theinfringement of the treaty obligation.The Kyoto obligations are erga omnes obligations, i.e.

obligations that can be invoked by one State on behalfof all (see Section 2.1) and it is conceivable that evennon-Parties could challenge non-compliant Kyoto Par-ties. There would be no need to show negligentbehaviour of the respective state, rather the breach ofobligation would in itself constitute the required‘‘wrongdoing’’ to trigger the right to reparation, i.e.compensation for damage in as much as they areattributable to the State exceeding its Kyoto target.However, such a claim could only encompass the excess

ARTICLE IN PRESS

23Para. 19 of the preamble to the Convention lists as particularly

vulnerable: low lying or small island countries, arid and semi-arid areas

or areas liable to floods, drought and desertification, and developing

countries with fragile mountainous ecosystems. This is not an

authoritative definition as it is only contained in the preamble.24See FCCC/CP/2001/13/Add. 1 and Add. 4, available at /http://

www.unfccc.intS Already at COP6bis had some Parties pledged

substantial increases in climate change funding for developing nations,

including for adaptation purposes (US$410 million per year by 2005,

to be revised in 2008 were committed by the EU, Canada, New

Zealand, Switzerland and Iceland, see FCCC/CP/2001/Add.13/Add.1

p. 43).25The President’s paper ‘‘New Proposals by the President of COP6’’

of 9th April 2001 paper in preparation for COP6bis contains the

phrase that ‘‘contribution targets should be based on Annex I Parties’

relative share of CO2 emissions in 1990.’’(p. 4) This clearly is a kind of

polluter pays approach, even though it only relates to CO2 and

adaptation funding unrelated to the actual needs for damage

prevention.

26All OECD nations have ratified the FCCC, and indeed most

nations have. Most OECD nations have ratified or will likely ratify

(e.g., Russia) the Kyoto Protocol, with the exception of the USA and

Australia who are politically unwilling to follow suit.27See FCCC/CP/2001/13/Add.3, available at /http://www.unfcc-

c.intS.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1115

emissions over and above the target as agreed by therespective country. We illustrate this scenario andresulting compensation claims taking the average 5%reduction by 2012 compared to 1990 emissions as ourguidance (see Sections 4 and 5).

2.3.2. The UN convention on the law of the sea

(UNCLOS)

The IPCC predicts a 9–88 cm sea level rise during the21st century as well as regionally varying increases ofocean temperatures. Sea level rise, coral bleaching andfish migration due to water temperature changes areonly some examples of potentially disastrous effects ofglobal warming on oceans, coastal and island states.UNCLOS both regulates rights and duties of states withregard to the specific jurisdictional regimes of maritimezones and the protection of the maritime environment.Article 194.2 UNCLOS implicitly prohibits unlimited

emissions of greenhouse gases by obliging States to ‘‘(..)ensure that activities under their jurisdiction and controlare so conducted as to not cause damage by pollution toother States and their environment (y)’’. Furthermore,Article 235 UNCLOS stresses that state responsibility istriggered if States do not fulfil their environmentalduties under UNCLOS: ‘‘States are responsible for thefulfillment of their international obligations concerningthe protection and preservation of the marine environ-ment. They shall be liable in accordance with interna-tional law.’’We first turn to the disappearance of territory and

impact on maritime zones under UNCLOS due to sealevel rise and secondly discuss how the impacts ofclimate change on the maritime environment itself arecovered by UNCLOS.28

2.3.2.1. Maritime zones. UNCLOS defines differentmaritime zones, including the territorial waters (up toB22 km off the coastline or ‘baseline’), the exclusiveeconomic zone, EEZ (B370 km) and the continentalshelf area. In the EEZ, the coastal state has sovereignrights for the purposes of exploring and exploiting,conserving and managing the natural resources. Theserights are invaluable for many coastal states. Forexample revenues from fishery concessions and permitsfor the exploration of minerals represent a substantiveelement in many state budgets. Beyond these maritimeareas defined by UNCLOS, on the high seas all stateshave equal rights (freedom of the high seas).29

The maritime zones are generally determined bydrawing a line around land territories (the baseline)according to a special methodology, and it is likely thatsea level rise would affect the positioning and/or scale ofState’s maritime zones (Soons, 1990; Freestone, 1991). Iffor example an (inhabited or inhabitable) island30

disappears, it can no longer serve for baseline determi-nation and the potential loss of maritime zones is quitesubstantial in some instances. And if a country/islandstate disappears altogether, it will lose its maritime zonesand accordingly all sovereign rights while maintaininglegal personality.31 The alteration of a coastline due tosea level rise will otherwise lead to a shift of the EEZwhile most likely leaving unaffected the continental shelfzone. The overall size of a country’s EEZ would remainthe same, but it would shift landwards. Since many fishstocks are dependent on the topography of the seabedrather than the distance to the coast, this could lead tofish stocks becoming high seas stocks that formerly werelocated in or straddled the EEZ of a given country. Sincecoastal states would lose their (restricted) sovereignrights over these stocks, such shift of EEZ could also bedefined as a damage.If such damage would occur due to anthropogenic

climate change, Article 194.2 would apply as a generalprohibition against such damage ‘‘by pollution’’ to otherstates. We will examine conditions for invoking thisprovision more closely in the next section.

2.3.2.2. Maritime environment. Another effect of cli-mate change on coastal states and fishing rights are thepredicted temperature changes in the ocean. Alreadyminor changes can cause the migration of stocks andthus affect coastal states’ fishing economies severely.Rising water temperatures are also identified as causefor coral bleaching which has severe impacts on coastalecosystems, the development of fish stocks and tourismin island and coastal states. Both the migration of fishstocks to areas outside a country’s EEZ and coralbleaching and other detrimental impacts on the marineenvironment could qualify as transboundary environ-mental damage and give rise to compensation claims.

ARTICLE IN PRESS

28 It should be noted that the Convention is not universally

applicable as for example the US has not ratified it. However, many

of its provisions qualify as customary international law (Brownlie,

1996).29Treaties such as the UN Straddling Stock Agreement of 1996 and

regional fisheries agreements seek to restrict unlimited fishing rights

also in the high seas to prevent overfishing and unsustainable

management of stocks.

30Article 121 UNCLOS provides that rocks ‘‘which cannot sustain

human habitation or economic life’’ have no EEZ or continental shelf.31This is an issue of uncertainty in international law. While the 1933

Montevideo Convention on Rights and Duties of States (165 LNTS

19) lists ‘‘a defined territory’’ as necessary element to qualify as a

person of international law, there is also the view that a state which

‘‘disappears’’ must still qualify as a state under international law so as

to protect its citizens. They could then be regarded as special (sui

generis) international persons. The general interest of a state in its

survival was recognised in the ICJ Advisory Opinion on Nuclear

Weapons, ICJ Rep. 1996, 263 para. 96; see also: Declaration of

President Bedjaoui, ICJ Rep. 1996, 273 para. 22. UNCLOS does not

protect state territory per se, however, this does not preclude claims

under customary law.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301116

According to Article 193 UNLCOS States have the‘‘obligation to protect and preserve the maritimeenvironment’’. This is supplemented by Article 207 ff.UNCLOS which contain more detailed duties regardingall sources of pollution to prevent harm and to protectthe marine environment, for example: ‘‘States shalladopt law and regulations to prevent, reduce andcontrol pollution of the marine environment fromland-based sources y. taking into account internation-ally agreed rules, standards and recommended practicesand procedures’’ (Article 207.1).First, rising water temperatures would have to qualify

as maritime pollution, as greenhouse gases obviously donot directly affect the marine environment, other thanincreasing the amount of CO2 available for uptake.UNLCOS defines pollution broadly. While this could besubject to a debate, in our view, indirect pollutingactivities such as emitting greenhouse gases are covered,since this activity, over time ‘‘results or is likely to resultin such deleterious effects as harm to the living resourcesand marine life, hazards to human health, hindrance tomarine activities, including fishing and other legitimateuse of the sea, (y)’’ (Article 1.1). This is supported bythe fact that States explicitly addressed pollution ‘‘fromor through the atmosphere’’ (Article 212) and that,during the negotiations, states were aware of thepotential threat of climate change to marine life.32

Despite the wording of Article 193 and 194, theseobligations in respect of the maritime environment donot represent an absolute prohibition to pollute. Rather,they represent due diligence obligations with the goal tominimise rather than eliminate pollution.33 Thus, tocomply with UNCLOS obligations, a state must actwith appropriate care.34 There are different methods todetermine this crucial standard of care. Firstly, it isnecessary to decide what the point of reference is. Thefact that climate change is caused by many polluters andover time and its impacts only manifest themselves muchlater makes this determination more difficult than if wewere dealing with, for example, a specific oil spill.Negligence could be linked to a specific international

standard. For example, could Article 4.2 FCCC serve asan indirect point of reference to determine negligenceeven though it is not phrased as a direct obligation? In

that case, all countries that have not met the year-2000target would be liable for damage under UNCLOS fromthe year 1994 onward depending on the time ofratification by the state. Another standard could beone of best available techniques, i.e. has a state taken allappropriate measures to prevent further emissions ofgreenhouse gases by employing best energy efficiencyand energy conservation as well as carbon-free technol-ogies. But to date, there exists no such universallyagreed technical standard and an international tribunalwould have to venture to define what is ‘‘appropriate’’itself.Foreseeability is also a frequently used concept: a

state acts negligent if it could have or has foreseenpotential damage. However, it is unclear whether theknowledge of climate change that could result fromgreenhouse gas emitting activities fulfils this criterion orwhether a state must have foreseen (i) the general or (ii)the precise nature of the damage that such changes cancause. The ICJ in the Corfu Channel case35 for exampleasked whether the Albanian authorities had known ofmines in the channel, which would later damage Britishships. It then found that Albania had acted negligentbecause it failed to warn the British despite havingknowledge of the mines. (This judgement was deliveredon the basis of circumstantial evidence, which showsthat positive proof of knowledge of a state of damage isprobably not necessary.) Spinning this further, Albaniawould in fact have been obliged to remove the mines,which represented such a risk from its territorial waters.Thus another question arises, which is linked to the issueof causation (see Section 2.1): is it a prerequisite that astate could (have) prevent(ed) the damage from occur-ring?In the framework of climate change damage, almost

every state36 has the opportunity to take measures toprevent damage or to minimise the risk of such damage.But due to the cumulative effect of the greenhouse gasemissions it would in fact be difficult to prove that oneor several states would have actually prevented a certaininjury. Therefore, in our view, in instances of severalpolluters, the question must be whether States were andare able to take action to significantly reduce emissionsof greenhouse gases, which would significantly reducetheir contribution to future climate change damage.Otherwise, there would, per definition, not be any legalresponsibility in instances where several actors cause aninjury. Such a defence is not accepted in national legal

ARTICLE IN PRESS

32See UN Secretary General ‘‘Law of the Sea—Protection and

preservation of the marine environment’’, 18 September 1989, UN Doc

A/44/461, para 114.33General opinion, see UN Secretary General ‘‘Law of the Sea—

Protection and preservation of the marine environment’’, 18 Septem-

ber 1989, UN Doc A/44/461, para 30.34Some have argued that UNCLOS contains direct or strict state

responsibility in Article 139 which is concerned with sea bed activities.

However, a state can exonerate itself if it has taken all ‘‘necessary and

appropriate measures to ensure effective compliance’’(Article 139.2

2nd sentence UNCLOS). This article is not therefore an example for

direct state responsibility, but sets rather a special due diligence

threshold.

351949 ICJ Reports p.4, Corfu Cannel (UK v Albania) at 22. Also

on the ICJ web pages at http://www.icj-cij.org/icjwww/idecisions/

isummaries/Iccsummary490409.htm (merits and assessment of com-

pensation).36Note that some states like many small islands are net-carbon sinks

as they emit almost no greenhouse gas and have a rich supply of

terrestrial biosphere taking up carbon.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1117

systems and could be a general legal principle applicablealso in international law.Another methodology to determine negligence seems

to look at the risk involved. If the risk is obvious, statesmust prove that they have taken all necessary measuresto prevent the risk from materialising into actualdamage. The quantity or severity of risk thus sets thethreshold for the due care duty. This is closely linkedwith the question of what exactly states would have tobe able to foresee. At what point in time could or shoulda Government have known that its conduct, for examplethe enforcement of fossil-fuel based energy plans, thelicensing of oil exploration or the authorisation of fossilfuel based power plants will cause damage to otherstates in the future? And what are the requirements interms of scientific certainty—is it enough to establishfault in this context that states knew of the risks or is itonly now, that scientific certainty has grown to virtualconsensus that states are responsible for failing to reactto these findings?The answer to these questions has bearings on when

responsibility starts. For example, as early as 1827 thefirst scientific studies (by Jean Baptiste Fourier)appeared which showed the relationship between con-centrations of greenhouse gases, in particular CO2, andthe radiative forcing of the atmosphere. Arrhenius(1896) argued that increasing CO2 concentrations wouldlead to warming, calculated the economic impacts ofthis, and even warned governments that regulation waswarranted. In the 1950s, monitoring data from theAntarctica and Hawaii allowed for in-depth research inatmospheric greenhouse gas concentrations and the firststudies on possible implications (climate change im-pacts) appeared. Global warming was a topic ofdiscussion at the first UN Conference on HumanDevelopment in Stockholm 1972, and the first worldclimate conference took place in 1979. Governments andpolicy makers were publicly alerted to the climatechange problem by the Villach Conference in 1985 andin 1990 the IPCC presented its first assessment report,outlining potential climate change scenarios and im-pacts; estimates of economic impacts of changes in theclimatic system were no longer ignorable. Finally, in1992 an overwhelming majority of the members of theinternational community adopted the FCCC, whichexplicitly recognises adverse impacts of climate change.Again, these issues would require much deeper

analysis. But we think that at the latest in 1992 withthe adoption of the FCCC states knew that theirbehaviour (or omission to regulate their economies)would contribute to future damage. Thus, if foresee-ability is accepted as measure for due diligence, stateswould be responsible for their respective emissions as ofeither the beginning of the 19th century, as of the 1950sor as of 1992. These timeframes are reflected in themodel runs in Section 3.

It should be noted that any claims under UNCLOSwould only cover damage due to maritime pollution orchanges in maritime zones, fishing rights, etc. Recover-able damage under UNCLOS would thus not encom-pass agricultural or health damage, but wouldencompass all coastal territory and adaptation/protec-tion costs.

2.3.3. The no-harm rule

The obligation to prevent injury to another state andto minimise the risk thereof has been mentioned aboveand is clearly a norm of customary international law.According to the majority view amongst legal scholars,it would be necessary to prove a breach of a due careduty, or negligence. We have already discussed theproblems associated with this in the section above. Itshould be noted, however, that, without discussingdamage quantification and coverage in detail, all kindsof climate change damage and direct prevention costscould come under the umbrella of the no-harm rule. Wereiterate that, in view of the figures and calculation thatwe want to present here, we do not attempt to associatea specific damage in a specific country with a pollutingactivity.

2.4. Other issues

Some final issues should be noted. Firstly, especially ifstate responsibility is established on the grounds offoreseeability, not only OECD countries might be liablefor future damage. While it seems that (on an aggregatelevel) negative impacts will only occur in developingcountries, it is theoretically possible that OECDcountries will try to recover damage costs from adeveloping country for its contribution to climatechange. Again, this could be subject to much legaldiscussion. In defence, developing countries could arguethat (i) their historic contributions are negligibleand (ii) the principle of common but differentiatedresponsibility, as enshrined in Article 3.1 FCCCrepresents a modification of the formal equality ofStates and would prevent them from being subject toany compensation duty. Yet, from the point ofview of a small island state or other least developedcountries it could well be argued that large emitters suchas India, Brazil and China cannot be freed of stateresponsibility for injuries abroad per se. We thereforealso show results in which developing countries areliable for damage.Furthermore, regarding burden sharing between

developed countries we point to Article 39 of the ILCDraft Articles on State Responsibility, which stipulatesthat in determining reparation, account shall be taken ofthe contribution of the state to the injury. This couldresult in a burden sharing according to emission shares.Another possibility for a court could be to hold one

ARTICLE IN PRESSR.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301118

state responsible for the whole injury to enable theclaimant state or states to recover its or their damagesfully.Moreover, even though we have not discussed various

problems associated with actually bringing a claim forcompensation before an international tribunal or court,two remarks are appropriate. One, under internationallaw, every injury capable of being measured by factualand objective standards is compensable, includingpersonal injury. More problematic is damage to thenatural environment. However, in many cases, theeconomic value of natural systems will equally beassessable. Two, it has been said that States wishing tomake a claim (at a minimum before the ICJ) would haveto bring claims against all possible defendants, i.e. allOECD countries together since those countries would bejointly liable for the damage caused (joint and severalliability). However, in the 1992 Nauru case37 theInternational Court of Justice allowed Nauru to bringa claim against Australia despite the fact that Naurucould have challenged two other nations, New Zealandand the UK. The court argued that, as long as theinterests of a state not party to a dispute before the courtwill not be affected by the judgement, for example byfinding illegal behaviour of this third party as aprerequisite for the claim in question, the plaintiff isnot required to challenge every possible defendant.38 Inthe case of greenhouse gas emissions contributing toclimate change and thus causing damage to States andindividuals of those states, each State’s independentbehaviour can be judged, even though in practice, thesum of greenhouse gas emissions leading to increasedconcentrations of greenhouse gases in the atmosphereare causing the climate system to change. This case lawhas now also been incorporated in Article 47 of the ILCDraft Articles on State Responsibility.

2.5. Preliminary conclusion

This section has argued that general state responsi-bility for climate change damage can be established. Thegeneral causal link between emissions of greenhousegases and climate change impacts is closed to a sufficientextent to make a legal case. Therefore, a claim is muchless difficult to envisage than only 10 years ago.However, further research into issues such as specificcausation and foreseeability as a legal standard for duediligence is needed. Moreover, the analysis has nottouched upon the problems of jurisdiction of interna-

tional courts and tribunals, and enforcement of awardsand judgements.

3. The impacts of climate change

There are various reasons why one may worry aboutclimate change. Some people argue it is a problembecause it could cause unacceptable hardship forparticularly vulnerable populations (e.g., those livingon small island states). Others are concerned about thepotential threat to certain unique and valuable systems(such as coral reefs). Still others worry that climatechange will increase the probability of large-scaleclimate instabilities (e.g., a shutdown of the GulfStream), and will have costly impacts on economiesthrough floods and storms. A fourth group wondersabout the total (or aggregate) impacts of climate changeon well-being and development (cf. Smith et al., 2001).A key challenge when assessing the impacts of climate

change is synthesis, i.e., the need to reduce the complexpattern of local and individual impacts to a moretractable set of indicators. The challenge is to identifyindicators that can summarise and make comparable theimpacts in different regions, sectors or systems in ameaningful way. If the aim is to integrate the impacts ofclimate change with standard national accounts, or tocompare the benefits of avoided climate change with thecosts of emission reduction, or to pay compensation, it isnecessary to express the costs climate change in the samemetric, that is money (for estimates, see Ayres andWalter, 1991; Cline, 1992; Downing et al., 1996a, b;Hohmeyer and Gaertner, 1992; Fankhauser, 1995;Nordhaus, 1991, 1994; Mendelsohn and Neumann,1999; Titus, 1992; Tol, 1995).Money is a particularly well-suited metric to measure

market impacts, that is impacts that are linked tomarket transactions and directly affect GDP. The costsof sea level rise can be expressed as the capital cost ofprotection plus the economic value of land andstructures at loss or at risk; and agricultural impactcan be expressed as costs or benefits to producers andconsumers. Using a monetary metric to express non-market impacts, such as effects on ecosystems or humanhealth, is more difficult. There is a broad and establishedliterature on valuation theory and its application,including studies (mostly in a non-climate changecontext) on the monetary value of lower mortality risk,ecosystems, quality of life, etc. (e.g., Freeman, 1993).But economic valuation can be controversial, andrequires sophisticated analysis that is still mostly lackingin a climate change context (e.g., Pearce et al., 1996).Global warming damage is often estimated for the

impacts of a doubling of the concentration of green-house gases in the atmosphere on the present economy.Most of the research on the assessment of these impacts

ARTICLE IN PRESS

37Certain Phosphate Lands in Nauru (Nauru v Australia), 1992 ICJ

Reports p. 240.38 In the Nicaragua case the ICJ expressed as a general rule that the

ICJ would ‘‘decline y to exercise the jurisdiction conferred upon it

where the legal interest of a State not party to the proceedings would

not only be affected by a decision, but would form the very subject-

matter of the decision’’ (1954 ICJ Reports, p. 32).

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1119

is carried out in and for the United States, but a handfulof studies have estimated the total impact of climatechange (disaggregated across sectors) in differentregions of the world. Table 1 shows aggregate,monetised impact estimates for a doubling of atmo-spheric carbon dioxide on the current economy andpopulation from three recent studies and summarises the‘first generation’ of studies (Pearce et al., 1996). Thenumerical results remain speculative, but they canprovide insights on signs, orders of magnitude, andpatterns of vulnerability. Results are difficult tocompare because different studies assume differentclimate scenarios, make different assumptions aboutadaptation, use different regional disaggregation andinclude different impacts. The Nordhaus and Boyer(1999) estimates, for example, are more negative thanothers, partly because they factor in the possibility ofcatastrophic impact. The Mendelsohn et al. (2000) andTol (2002a, b) estimates, on the other hand, are driven

by optimistic assumptions about adaptive capacity andbaseline development trends, which results in mostlybeneficial impacts.Standard deviations are rarely reported, but likely

amount to several times the ‘best guess’. They are largerfor developing countries, where results are generallyderived through extrapolation rather than direct estima-tion. This is illustrated by the standard deviationsestimated by Tol (2002a, b), reproduced in Table 1.These estimates probably underestimate the true un-certainty, because they exclude omitted impacts andsevere climate change scenarios.Overall, the current generation of aggregate estimates

may understate the true cost of climate change becausethey tend to ignore extreme weather events; under-estimate the compounding effect of multiple stresses;and ignore the costs of transition and learning.However, studies may also have overlooked positiveimpacts of climate change and not adequately accounted

ARTICLE IN PRESS

Table 1

Estimates of the regional impacts of climate changea

First generation Mendelsohn et al. Nordhaus/Boyer Tolb

2.5�C 1.5�C 2.5�C 2.5�C 1.0�C

North America �1.5 3.4(1.2)

USA �1.0 to �1.5 0.3 �0.5OECD Europe �1.3 3.7 (2.2)

EU �1.4 �2.8OECD Pacific �1.4 to �2.8 1.0 (1.1)

Japan �0.1 �0.5Eastern Europe and fUSSR 0.3 2.0 (3.8)

Eastern Europe �0.7fUSSR �0.7Russia 11.1 0.7

Middle East �4.1 �2.0c 1.1 (2.2)

Latin America �4.3 �0.1 (0.6)Brazil �1.4South and Southeast Asia �8.6 �1.7 (1.1)India �2.0 �4.9China �4.7 to �5.2 1.8 �0.2 2.1 (5.0)d

Africa �8.7 �3.9 �4.1 (2.2)DCs 0.12 0.03

LDCs 0.05 �0.17World

Output weighede �1.5 to �2.0 0.1 �1.5 2.3 (1.0)

Population weighedf �1.9At world average pricesg �2.7 (0.8)Equity weighedh 0.2 (1.3)

Source: Pearce et al. (1996); Mendelsohn et al. (2000); Nordhaus and Boyer (2000); Tol (2002a, b).aFigures are expressed as impacts on a society with today’s economic structure, population, laws, etc. Mendelsohn et al.’s estimates denote impact

on a future economy. Estimates are expressed as per cent of Gross Domestic Product. Positive numbers denote benefits, negative numbers denote

costs.bFigures in brackets denote standard deviations. They denote a lower bound to the real uncertainty.cHigh-income OPEC.dChina, Laos, North Korea, Vietnam.eRegional monetary impact estimates are aggregated to world impacts without weighing.fRegional monetary impact estimates are aggregated to world impacts using weights that reflect differences in population sizes.gRegional impacts are evaluated at world average values and then aggregated, without weighing, to world impacts.hRegional monetary impact estimates are aggregated to world impacts using ‘‘equity weights’’ that equal the ratio of global average per capita

income to region average per capita income.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301120

for how development could reduce impacts of climatechange. Our current understanding of (future) adaptivecapacity, particularly in developing countries, is toolimited, and the inclusion of adaptation in currentstudies too varied to allow a firm conclusion about thedirection of the estimation bias.Market-impacts are low, and may be positive in some

countries and sectors—at least in developed regions.This is largely due to adaptation. Efficient adaptationreduces the net costs of climate change because the costof such measures is lower than the concomitantreduction in impacts. However, impact uncertaintyand lack of capacity may make efficient and error-freeadaptation difficult. Even so, market impacts could besignificant in some conditions, such as a rapid increasein extreme events, which might lead to large losses and/or costly over-adaptation (if random fluctuations aremistaken for a trend) (see Downing et al., 1998).Developing countries are more vulnerable to climate

change than developed countries because their econo-mies rely more heavily on climate-sensitive activities (inparticular agriculture), and many already operate closeto environmental and climatic tolerance levels (e.g., withrespect to coastal and water resources). Developingcountries are poorly prepared to deal with the climatevariability and natural hazards they already face today(World Bank, 2000). If current development trendscontinue, few of them will have the financial, technical,and institutional capacity and knowledge base to dealwith the additional stress of climate change.Estimates of global impact are sensitive to the way

figures are aggregated. Because the most severe impactsare expected in developing countries, the more weight isassigned to developing countries, the more severe areaggregate impacts. Using a simple adding of impacts,some studies estimate small net positive impacts at a fewdegrees of warming, while others estimate small netnegative impacts.Net aggregate benefits do not preclude the possibility

of a majority of people being negatively affected, andsome population groups severely so. This is due to thefact that developed economies, many of which couldhave positive impacts, contribute the majority of globalproduction but account for a smaller fraction of worldpopulation. However, there are no studies so far thathave consistently estimated the total number of peoplethat could be negatively affected by climate change.The value judgments underlying regional aggregation

are discussed and made explicit in Azar (1999), Azar andSterner (1996) and Fankhauser et al. (1997, 1998). Weunderline the importance of aggregation by using fouralternative ways of computing world total impacts fromregional impact estimates in Table 1.The impact estimates of Table 1 are very uncertain,

and the studies upon which they are based suffer frommany shortcomings. We list the most important. A

major difficulty in impact assessment is our stillincomplete understanding of climate change itself, inparticular the regional details of climate change (Mahl-man, 1997). Impacts are local, and impacts are related toweather variability and extremes. Current climatechange scenarios and current climate change impactstudies use crude spatial and temporal resolutions, toocrude to capture a number of essential details thatdetermine the impacts.Knowledge gaps continue at the level of impact

analysis. Despite a growing number of country-levelcase studies (e.g., US Country Studies Program, 1999),our knowledge of local impacts is still too uneven andincomplete for a careful, detailed comparison acrossregions. Furthermore, differences in assumptions oftenmake it difficult to compare case studies acrosscountries. Only a few studies try to provide a coherentglobal picture, based on a uniform set of assumptions.The basis of many such global impact assessments tendto be case studies with a more limited scope, oftenundertaken in the United States, which are thenextrapolated to other regions. Such extrapolation isdifficult and will be successful only if regional circum-stances are carefully taken into account, includingdifferences in geography, level of development, valuesystems and adaptive capacity. Not all analyses areequally careful in undertaking this task. While ourunderstanding of the vulnerability of developed coun-tries is improving—at least with respect to marketimpacts—good information about developing countriesremains scarce.Non-market damage, indirect effects (e.g., the

effect of changed agricultural output on the foodprocessing industry), horizontal interlinkages (e.g., theinterplay between water supply and agriculture;or how the loss of ecosystem functions will affectGDP), and the socio-political implications of change arealso still poorly understood. Uncertainty, transienteffects (the impact of a changing rather than a changedand static climate), and the influence of change inclimate variability are other factors deserving moreattention.Another key problem is adaptation. Adaptation will

entail complex behavioural, technological and institu-tional adjustments at all levels of society, and not allpopulation groups will be equally adept at adapting.Adaptation is treated differently in different studies, butall approaches either underestimate or overestimate itseffectiveness and costs. Impact studies are largelyconfined to autonomous adaptation, that is, adaptationsthat occur without explicit policy intervention from thegovernment. But in many cases governments too willembark on adaptation policies to avoid certain impactsof climate change, and may start those policies wellbefore critical climatic change occurs—for example, bylinking climate change adaptation to other development

ARTICLE IN PRESSR.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1121

and global change actions, such as on drought anddesertification or biodiversity.The analysis is further complicated by the strong link

between adaptation and other socio-economic trends.The world will substantially change in the future, andthis will affect vulnerability to climate change. Forexample, a successful effort to roll back malaria couldreduce a climate change induced spread of malaria risks.A less successful effort could introduce antibiotic-resistant parasites or pesticide-resistant mosquitoes,increasing vulnerability to climate change. Thegrowing pressure on natural resources from unsustain-able economic development is likely to exacerbate theimpacts of climate change. However, if this pressureleads to improved management (e.g., water markets),vulnerability might decrease. Even without explicitadaptation, impact assessments therefore varydepending on the ‘type’ of socio-economic developmentexpected in the future. The sensitivity of estimatesto such baseline trends can in some cases be strongenough to reverse the sign, i.e., a potentially negativeimpact can become positive under a suitable develop-ment path or vice versa (Mendelsohn and Neumann,1999).Global warming impact under the 2�CO2 scenario is

hypothetical impact on the present society. The realimpact of global warming will occur in the future, andsociety will have changed in terms of populationnumbers, economic size and structure, technology andin terms of socio-cultural and political factors. Thesechanges may affect the vulnerability of society to globalwarming. If we are interested in the absolute size ofimpact (for example per ton of CO2 emissions) it isnecessary to establish the future size of the populationand stock at risk. The timeframe of global warmingimpact is too long, however, to predict such future

developments with any measure of precision. Hence,scenarios are used that describe possible futures but donot claim to describe the most likely future.One of the main challenges of impact assessments is to

move from this static analysis to a dynamic representa-tion of impacts as a function of shifting climatecharacteristics, adaptation measures and exogenoustrends like economic and population growth. Ourunderstanding of the time path aggregate impacts willfollow under different warming and developmentscenarios, is still extremely limited. Among the fewexplicitly dynamic analyses are Sohngen and Mendel-sohn (1999), Tol and Dowlatabadi (2001) and Yohe et al.(1996). These studies are highly speculative, as theunderlying models only provide a very rough reflectionof real-world complexities. Fig. 1 shows examples fromthree studies. While some analysts still work withrelatively smooth impact functions (e.g., Nordhausand Boyer, 2000), there is growing recognition (e.g.,Tol, 1996, 2002a, b; Mendelsohn and Schlesinger, 1999)that the climate impact dynamics—the conjunction ofclimate change, societal change, impact, and adapta-tion—is certainly not linear, and might be quitecomplex.Impacts in different sectors may unfold along

fundamentally different paths. Coastal impacts, forexample, are expected to grow continuously over time,more or less in proportion to the rise in sea level. Theprospects for agriculture, in contrast, are more diverse.While some models predict aggregate damage alreadyfor moderate warming, many studies suggest that undersome (but not all) scenarios the impact curve might behump-shaped, with short-term (aggregate) benefitsunder modest climate change turning into losses undermore substantial change (e.g., Mendelsohn and Schle-singer, 1999).

ARTICLE IN PRESS

Fig. 1. The impact of climate change as a function of the global mean temperature, according to Mendelsohn et al. (1996), Nordhaus and Boyer

(2000), and Tol (2002a, b). Mendelsohn et al. aggregate impacts across different regions weighted by regional output. Nordhaus and Boyer aggregate

either weighted by regional output or weighted by regional population. Tol aggregates either by regional output or by equity, that is, by the ratio of

world per capita income to regional per capita income.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301122

The hypothetical compensation paid in the calcula-tions below is based on the output-weighted estimates ofTol (2002a, b), as calculated with the FUND 2.0 model(Tol, 1997, 1999a-e, 2001). The scenario used is IS92a(Leggett et al., 1992). Because of the uncertainties, wealso present results for seven alternative impact esti-mates. First, we use the output-weighted impactestimates of Nordhaus and Yang (1996). The other fivealternatives are based on Tol (2002a, b). We include theIS92d and IS92f scenarios as alternatives to IS92a. Wealso look at equity-weighted impact estimates (Fan-khauser et al., 1997, 1998), which reflect utility ratherthan income losses. We look at market impacts only,excluding the hard-to-measure non-market impacts. Welook at negative impacts only, excluding impacts onthose sectors that are positively affected by climatechange. Finally, we consider the case that willingness toaccept compensation (WTAC) rather than willingness topay (WTP) is the appropriate way of valuing non-market impacts, assuming that WTAC is 4 times higherthan WTP (Pearce and Turner, 1990). Fig. 2 displays theeight alternative impact estimates. We consider totalimpacts in less developed countries, as measured inpercentage of GDP.The base case estimate of LDC climate change impact

has slightly positive impacts of climate change up to theyear 2050, primarily due to CO2 fertilisation ofagriculture. After 2050, impacts become negative, risingup to some 1% of GDP. Impact estimates are hardlysensitive to the scenario used. Aggregate market impactsare positive throughout the period. The negative impactsof climate change run up to some 8% of GDP. The

equity-weighted and WTAC impact estimate emphasisethe strongly non-linear nature of the non-marketimpacts, leading to rather speculative total impactestimates. Nordhaus’ impact estimates are considerablymore pessimistic than those of Tol (2002a, b).

4. Responsibilities

We use cumulative emissions as our measure ofresponsibility for impacts of climate change. Emissionsare the established metric in international climate policy.Although harder to measure than atmospheric concen-trations, emissions are more easily attributable tocountries. We ignore greenhouse gases other that carbondioxide, and also exclude CO2 from land use change.Carbon dioxide emissions from industry, householdsand transport are easily and verifiably measure, andconstitute about 75% of the problem.As our base case, we consider only emissions after

2000, the year in which governments can be expected toknow about climate change and have had the time tostart emission abatement programmes. The first OECDcommitments to reduce emissions (part of the UNFCCC) were for the year 2000. We do this for the IS92ascenario as well as for IS92d and IS92f. As analternative, we only consider excessive emissions, thatis, those emissions above a reasonable emission reduc-tion policy. We assume that a 5% emission reductionpolicy (from 1990 levels) is reasonable; 5% is approxi-mately what was agreed in the Kyoto Protocol. As asecond alternative, we look at cumulative emissions

ARTICLE IN PRESS

Fig. 2. Eight alternative estimates of the impact in less developed countries as a percentage of their GDP. The Nordhaus estimate is due to Nordhaus

and Yang (1996), the other estimates are due to Tol (2002a, b). The Equity estimate weights the regional impact estimates by the ratio of world

average income and regional average income, the other estimates add up dollars without weighting. The IS92f and IS92d estimates are based on the

scenarios of the same name, the other estimates are based on IS92a. The Market estimates excludes non-market impacts, the WTAC estimates

multiplies non-market impacts by 4, the other estimates include non-market impacts without weighting. The Only Negatives estimate excludes

impacts on those sectors that, on average, benefit from climate change, the other estimates aggregate positive and negative impacts. The Output

estimate is the base case.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1123

since 1800, the starting point of emissions data. In allcases, cumulative emissions in the OECD are comparedto world cumulative emissions since 1800. Fig. 3 displaysthe resulting ‘‘responsibility’’ of the OECD. Forcomparison, Fig. 4 displays the ‘‘responsibilities’’ ofnon-OECD countries, looking at cumulative emissionssince 2000.If cumulative emissions are measured since 2000, the

OECD is responsible for 20–25% of climate changefrom 2025 onwards. These numbers hold for IS92a aswell as for IS92d and IS92f. If only emissions in excessof the ‘‘Kyoto’’ emission reduction policy are counted,the OECD’s contribution to climate change slopes up toabout 15% in 2200. If cumulative emissions since 1800are counted, the OECD starts with a contribution of

about 65% falling to 25% by 2200. Since 2000, the shareof non-OECD emissions increases, in particular fromAsia. China and South and Southeast Asia account formore than half of cumulative emissions by 2200 in theIS92a scenario.Section 3 argues that there are alternative dates to our

choice of 1800 and 2000, particularly 1827, 1896, 1950,1972, 1985, 1990 and 1992. This would lead to assignedresponsibilities that are somewhere in between theextremes presented in Fig. 3. We omit those intermedi-ate cases for ease of exposition.An alternative to emissions in excess of ‘‘Kyoto’’

would be emissions in excess of some per capitaemissions allotment, again starting in the year 2000(for example). The result would a graph very similar to

ARTICLE IN PRESS

Fig. 3. The share of cumulative OECD emissions of industrial carbon dioxide in cumulative world emissions. Cumulative OECD emissions are

measured since 2000 for IS92a (the base case), IS92d and IS92f. Cumulative OECD emissions are also measured since 1800, and in deviation from a

reasonable emission reduction policy (relative to Kyoto).

Fig. 4. The share of cumulative non-OECD emissions of industrial carbon dioxide in cumulative world emissions. Cumulative emissions are

measured since 2000.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301124

the ‘‘Since 2000’’ graph in Fig. 3, albeit lower. Thereason is that the ‘‘Since 2000’’ scenario, counting allemissions, in fact uses a zero allotment; any positiveallotment would shift the graph downwards.

5. Compensations

In Section 3, we present alternative estimates of theimpact of climate change on developing countries. InSection 4, we present alternative estimates of thecontribution of the OECD to global warming. Thecompensation, hypothetically paid by the OECD todeveloping countries,39 follows from multiplying theestimated impact with the estimated contribution ofcumulative OECD emissions to cumulative worldemissions.Fig. 5 displays the results for one view on responsi-

bility (cumulative emissions since 2000) and six alter-native impact estimates. Compensations vary betweenzero—if only market impacts are compensated (asmarket impacts are positive)—and 4% of the GDP ofthe OECD—if only negative impacts count—towardsthe end of the 22nd century. Using the output-weightedestimates of Tol, compensation is small in the 21stcentury but amounts to 0.5% GDP in the 22nd century.For equity-weighted damage, compensation is morethan twice as high in the later years, while alsosubstantial compensation is paid in the earlier years.Compensation is also paid in the earlier year if WTACrather than WTP is used as the basis for non-market

valuation. In the later years, compensations increase toabout 2% of GDP.Fig. 6 displays compensations according to one

impact estimate (output-weighted Tol) and three alter-native views on responsibility. If OECD responsibility ismeasured as cumulative OECD emissions since 2000relative to cumulative world emissions since 1800,compensation amounts to an annual 0.4% of GDP inthe long run. This number goes up to 0.5% of GDP ifemissions since 1800 matter, and down to 0.3% ifemissions in excess of ‘‘Kyoto’’ matter.Fig. 7 displays the results for three alternative

development scenarios. Note that the scenarios differboth in the their impact estimates (output-weighted Tol)and in their responsibilities (cumulative emissions since2000). Compensation is about a 0.4% of GDP per yearin the long term for the IS92a scenario. This numberincreases to 0.5% for IS92f, and falls to 0.3% for IS92d.Fig. 8 shows the compensation split out over the

OECD, using the output-weighted Tol estimates, IS92a,and cumulative emissions since 2000. The USA andCanada (OECD-America) would pay up to 0.6% oftheir GDP, Western Europe (OECD-Europe) up to0.5% of GDP, and Japan, Australia and New Zealand(OECD-Pacific) up to 0.2% of GDP.Fig. 9 compares the compensation paid by the OECD

to the climate change impacts on the OECD. The paidcompensation is relatively small, so that we should notexpect a great difference on climate change policy (butsee Kemfert and Tol, 2002; Panayotou et al., 2002; Tol,2002c). Fig. 10 compares the compensation paid todeveloping countries to their climate change impacts. Asfollows from Fig. 3, only a minor part of the damage iscompensated. Figs. 9 and 10 also include the case inwhich currently developing countries are held liable forclimate change damage. As developing countries are

ARTICLE IN PRESS

Fig. 5. The compensation paid by the OECD to less developed countries as a percentage of the GDP of the OECD. The scenario is IS92a, the

compensation is paid on the basis of cumulative emissions since 2000. Six alternative impact estimates are used, as in Fig. 2.

39 In reality, compensation would be paid by one country to another,

or even at a lower level of aggregation (e.g., company to individual).

The model’s regional resolution does not allow an analysis of that.

Besides, an ex ante analysis in that detail makes little sense.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1125

responsible for a large share of future emissions, OECDcountries could claim a lot of compensation, particularlyin the 22nd century. In the particular scenario shown,the received compensation exceeds the paid compensa-tion from 2160 onwards. This result is easily overturnedif responsibility would extend further back than the year2000, or if an allowance would be made for a certainamount of per capita emissions. Note that the calcula-tions do not take into account emission shares relativeto population size but only look at industrial emissionsof CO2 per se. Note also that per capita emissions intoday’s developing countries are likely to be far belowthose of today’s industrialised countries even in 2050and beyond, which might indeed have an impact on alegal analysis.

6. Discussion and conclusion

International law provides a basis for responsibilityfor climate change impacts, while also many gapsremain. This basis lies in widely accepted customarylaw, such as the no harm rule. This is reinforced bydeclarations and precedents, but also by internationalagreements, some of which are almost universallyratified, such as the United Nations FCCC.The crucial issue is whether countries are responsible

for all greenhouse gas emissions and the resultingdamage, or only for emissions in excess of a reasonableemission abatement target. In the latter case, the OECDcarries only limited responsibility for climate change, asin the past emissions were unregulated and in the future

ARTICLE IN PRESS

Fig. 6. The compensation paid by the OECD to less developed countries as a percentage of the GDP of the OECD. The scenario is IS92a, the

damage is the output-weighted impact estimates of Tol. Compensation is paid on the basis of three alternative views on responsibility, as in Fig. 3.

Fig. 7. The compensation paid by the OECD to less developed countries as a percentage of the GDP of the OECD. The damage is the output-

weighted impact estimates of Tol, the compensation is paid on the basis of cumulative emissions since 2000. Three alternative development scenarios

are used, as in Figs. 2 and 3.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301126

non-OECD emissions are much larger. If, however,countries are responsible for all emissions, then itmatters when one starts counting. If one counts allemissions from the time governments could have knownabout climate change, OECD responsibility is large. Ifone starts counting at the time climate change wasofficially recognised as a policy problem, OECDresponsibility is much smaller.These questions have a much greater influence on

OECD responsibility than do the uncertainties aboutfuture emissions. Moreover, irrespective of compensa-tion duties, states might still have a duty to reduce therisk of climate change damage by undertaking andsupporting adaptation measures. The (independent)

costs of such measures were not calculated in ourscenarios, but could be substantial.Developing countries may be held responsible for

their future emissions, which cannot be excused by alack of knowledge about the consequences or a lack oftechnological alternatives. Scenarios under which theresponsibilities of developing countries exceed those ofdeveloped countries, and the net compensation flows arefrom South to North are not inconceivable.Assuming that the OECD will be held accountable

and will compensate developing countries for damageincurred, the question is for what and how much. Thisquestion cannot be answered with any accuracy, asestimates of the impacts of climate change are very

ARTICLE IN PRESS

Fig. 8. The compensation paid by the various OECD regions to less developed countries as a percentage of the GDP of the OECD. The damage is

the output-weighted impact estimates of Tol, the compensation is paid on the basis of cumulative emissions since 2000, and the scenario is IS92a.

Fig. 9. The compensation paid by the OECD to developing countries, the compensation received by the OECD, and the impact of climate change on

the OECD, all as a percentage of the GDP of the OECD. The damage is the output-weighted impact estimates of Tol, the compensation is paid on

the basis of cumulative emissions since 2000, and the scenario is IS92a.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1127

uncertain, as are future law suits. Important issues arethe magnitude of climate change impacts; whether non-market impacts on health and ecosystems are alsocompensated; whether compensation is paid for adapta-tion costs and residual damage only, or whethercompensation also includes reparation for imposeddamage; and whether positive impacts offset negativeones or not. These issues make billions of dollars ofdifference in the actual compensation paid.The compensation by the OECD to developing

countries would imply a transfer of up to 0.25% ofGDP in the short run and up to 4% of GDP in the longrun. This compares to 0.25% currently paid in devel-opment aid, the 0.70% UN goal for official developmentaid and 0.75% percent foreign direct investment by theOECD in developing countries. Potential compensationpaid could also be much higher than the climate changeimpacts on the OECD.This means that, should state responsibility for

climate change impacts be taken seriously, it wouldsubstantially affect not only the climate change billfooted by the OECD, but also north–south investmentand trade patterns. The wider implications of thisrequire considerable further analysis. On the policylevel, states might wish to start developing globalsolutions to tackle the issue of future climate changedamage (see Whitmore, 2000; AOSIS, 1991) as legally,ignoring the problem will not exonerate polluters in thefuture.

Acknowledgements

G. Wiser (Washington, DC), M. Buck (Hamburg),Prof. H. Rittstieg (Hamburg) and an anonymous refereeprovided valuable comments on earlier versions of this

paper. The US National Science Foundation throughthe Center for Integrated Study of the Human Dimen-sions of Global Change (SBR-9521914) the MichaelOtto Foundation, and the German Scholarship Foun-dation (Studienstiftung des deutschen Volkes) providedwelcome financial support. All errors and opinionsare ours.

References

AOSIS, 1991. Proposal for an ‘‘Insurance Mechanism’’, submitted by

Vanatu on behalf of the Alliance of Small Island States

(AOSIS), contained in UN Doc. A/AC.237/WG.II/CRP.8,

reprinted in Report of the 4th INC session, UN Doc. A/AC.237/

15, 126 ff.

Arrhenius, S., 1896. On the influence of carbonic acid in the air upon

the temperature on the ground. The London, Edinburgh, and

Dublin Philosophical Magazine and Journal of Science [fifth series]

41 (251), 237–276.

Arsanjani, Reisman, 1998. The Quest for an international liability

regime for the protection of the global commons. In: Wellens, K.C.

(Ed.), International Law: Theory and Practice; Essays in honour of

Eric Suy. Nijhoff Publishers, The Hague, p. 469.

Ayres, R.U., Walter, J., 1991. The greenhouse effect: damages, costs

and abatement. Environmental and Resource Economics 1,

237–270.

Azar, C., 1999. Weight factors in cost-benefit analysis of climate

change. Environmental and Resource Economics 13, 249–268.

Azar, C., Sterner, T., 1996. Discounting and distributional considera-

tions in the context of global warming. Ecological Economics 19,

169–184.

Banuri, T., Maeler, K.-G., Grubb, M.J., Jacobsen, H.K., Yamin, F.,

1996. Equity and social considerations. In: Bruce, J.P.,

Lee, H., Haites, E.F. (Eds.), Climate Change 1995: Economic

and Social Dimensions, Working Group III of the Intergovern-

mental Panel on Climate Change. Cambridge University Press,

Cambridge.

Bodansky, D., 1993. The UN framework convention on

climate change: a commentary. Yale Journal of International

Law 18, 451.

ARTICLE IN PRESS

Fig. 10. The compensation paid by the OECD to the developing countries, the compensation received by the developing countries, and the impact of

climate change on the developing countries, all as a percentage of the GDP of the developing countries. The damage is the output-weighted impact

estimates of Tol, the compensation is paid on the basis of cumulative emissions since 2000, and the scenario is IS92a.

R.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301128

Boyle, A., 2000. Codification of International Environmental Law and

the International Law Commission: injurious consequences revis-

ited. In: Boyle, Freestone (Eds.), International Law and Sustain-

able Development—Past Achievements and Future Challenges,

OUP, Oxford, p. 61.

Brownlie, I., 1983. System of the Law of the Nations, State

Responsibility, Part I. Clarendon Press, Oxford.

Brownlie, I., 1996. Principles of Public International Law. Clarendon

Press, Oxford, 1998.

Cameron, J., Zaelke, D., 1990. Global warming and climate change—

an overview of the international legal process. American University

Journal of International Law & Policy 5, 249.

Church, J.A., Gregory, J.M., Huybrechts, P., Kuhn, M., Lambeck, K.,

Nhuan, M.T., Qin, D., Woodworth, P.L., 2001. Changes in sea

level. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van

der Linden, P.J., Dai, X., Maskell, K., Johnson, C.A. (Eds.),

Climate Change 2001: The Scientific Basis, Report of Working

Group I of the Intergovernmental Panel on Climate Change.

Cambridge University Press, Cambridge.

Cline, W.R., 1992. The Economics of Global Warming. Institute for

International Economics, Washington, DC.

Crawford, J., Peel, J., Olleson, S., 2001. The ILC articles on state

responsibility for internationally wrongful act: completion

of the second reading. European Journal of International Law

13, 963.

Downing, T.E., Eyre, N., Greener, R., Blackwell, D., 1996a. Full Fuel

Cycle Study: Evaluation of the Global Warming Externality for

Fossil Fuel Cycles With and Without CO2 Abatament and for Two

Reference Scenarios. Environmental Change Unit, University of

Oxford, Oxford, pp. 1–72.

Downing, T.E., Eyre, N., Greener, R., Blackwell, D., 1996b. Projected

Costs of Climate Change for Two Reference Scenarios and Fossil

Fuel Cycles. Environmental Change Unit, Oxford.

Downing, T.E., Olsthoorn, A.A., Tol, R.S.J. (Eds.), 1998. Climate,

Change and Risk, 1st Edition. Routledge, London.

Fankhauser, S., 1995. Valuing Climate Change—The Economics of

the Greenhouse, 1st Editions. EarthScan, London.

Fankhauser, S., Tol, R.S.J., Pearce, D.W., 1997. The aggregation of

climate change damages: a welfare theoretic approach. Environ-

mental and Resource Economics 10, 249–266.

Fankhauser, S., Tol, R.S.J., Pearce, D.W., 1998. Extensions and

alternatives to climate change impact valuation: on the critique of

IPCC working group III’S impact estimates. Environment and

Development Economics 3, 59–81.

Francioni, F. (Ed.), 1991. International Responsibility for Environ-

mental Harm. Graham & Trotman, London.

Freeman III, A.M., 1993. The Measurement of Environmental and

Resource Values. Resources for the Future, Washington, DC.

Freestone, D., 1991. International law and sea level rise. In: Churchill,

Freestone (Eds.), International Law and Global Climate Change,

Kluwer Law International, London, p. 109.

Gitay, H., Brown, S., Easterling, W., Jallow, B., Antle, J., Apps, M.,

Beamish, R., Chapin, T., Cramer, W., Frangi, J., Laine, J., Lin, E.,

Magnuson, J., Noble, I., Price, J., Prowse, T., Root, T., Schulze,

E., Sirotenko, O., Sohngen, B., Soussana, J., 2001. Ecosystems

and their goods and services. In: McCarthy, J.J., Canziani, O.F.,

Leary, N.A., Dokken, D.J., White, K.S. (Eds.), Climate Change

2001: Impacts, Adapation and Vulnerability, Report of Working

Group II of the Intergovernmental Panel on Climate Change.

Cambridge University Press, Cambridge.

Handl, G., 1980. State liability for accidental transnational environ-

mental damage by private persons. American Journal of Interna-

tional Law 74, 525.

Hohmeyer, O., Gaertner, M., 1992. The Costs of Climate Change—A

Rough Estimate of Orders of Magnitude. Karlsruhe: Fraunhofer-

Institut fur Systemtechnik und Innovationsforschung.

Horbach, N., 1996. Liability versus responsibility under international

law—defending strict state responsibility, Ph.D. Leiden.

Kemfert, C., Tol, R.S.J., 2002. Equity, international trade and climate

policy. International Environmental Agreements 2, 23–48.

Le Cahier, 1977, Le probl"eme de la responsibiltit!e pour risqu!e en droit

international. International Relations in a Changing World 16,

411–432.

Lefeber, R., 1996. Transboundary Environmental Interference and the

Origin of State Liability. Kluwer Law International, The Hague.

Leggett, J., Pepper, W.J., Swart, R.J., 1992. Emissions scenarios for

the IPCC: an update. In: Houghton, J.T., Callander, B.A., Varney,

S.K. (Eds.), Climate Change 1992—The Supplementary Report to

the IPCC Scientific Assessment, 1st Edition. Cambridge University

Press, Cambridge, pp. 71–95.

Mahlman, J.D., 1997. Uncertainties in projections of human-caused

climate warming. Science 278, 1416–1417.

Mendelsohn, R.O., Neumann, J.E. (Eds.), 1999. The Impact of

Climate Change on the United States Economy. Cambridge

University Press, Cambridge.

Mendelsohn, R.O., Schlesinger, M.E., 1999. Climate-response func-

tions. Ambio 28 (4), 362–366.

Mendelsohn, R., Morrison, W., Schlesinger, M.E., Andronova, N.G.,

2000. Country-specific market impacts of climate change. Climatic

Change 45, 553–569.

Mitchell, J.F.B., Karoly, D.J., Hegerl, G.C., Zwiers, F.W., Allen,

M.R., Marengo, J., 2001. Detection of climate change and

attribution of causes. In: Houghton, J.T., Ding, Y., Griggs, D.J.,

Noguer, M., van der Linden, P.J., Dai, X., Maskell, K., Johnson,

C.A. (Eds.), Climate Change 2001: The Scientific Basis, Report of

Working Group I of the Intergovernmental Panel on Climate

Change. Cambridge University Press, Cambridge.

Nordhaus, W.D., 1991. To slow or not to slow: the economics of the

greenhouse effect. Economic Journal 101, 920–937.

Nordhaus, W.D., 1994. Managing the Global Commons: The

Economics of Climate Change. The MIT Press, Cambridge.

Nordhaus, W.D., Boyer, J.G., 1999. Warming the World: Economic

Models of Global Warming. Anonymous.

Nordhaus, W.D., Boyer, J.G., 2000. Warming the World: Economic

Models of Global Warming. MIT Press, Cambridge, US.

Nordhaus, W.D., Yang, Z., 1996. RICE: a regional dynamic general

equilibrium model of optimal climate-change policy. American

Economic Review 86 (4), 741–765.

Okowa, P., 2000. State Responsibility for Transboundary Air

Pollution in International Law. Oxford University Press, Oxford.

Ott, H.E., 1996. V .olkerrechtliche Aspekte der Klimarahmenkonven-

tion. In: Brauch, Klimapolitik, Springer, Berlin, 61.

Ott, C., Paschke, M., 1997. Ausgleichsw .urdige Summations—und

Distanzsch.aden am Beispiel der neuartigen Waldsch.aden, Um-

weltbundesamt, Text 89/97.

Panayotou, T., Sachs, J.D., Zwane, A.P., 2002. Compensation for

meaningfull participation in climate change control: a modest

proposal and empirical analysis. Journal of Environmental

Economics and Management 43, 437–454.

Pearce, D.W., Cline, W.R., Achanta, A.N., Fankhauser, S., Pachauri,

R.K., Tol, R.S.J., Vellinga, P., 1996. The social costs of climate

change: greenhouse damage and the benefits of control. In: Bruce,

J.P., Lee, H., Haites, E.F. (Eds.), Climate Change 1995: Economic

and Social Dimensions—Contribution of Working Group III to

the Second Assessment Report of the Intergovernmental Panel

on Climate Change. Cambridge University Press, Cambridge,

pp. 179–224.

Pearce, D.W., Turner, R.K., 1990. Economics of Natural Resource

and the Environment. Harvester Wheat Sheaf, New York.

Perrez, F.X., 1996. The relationship between ‘‘permanent sovereignty’’

and the obligation not to cause transboundary environmental

damage. Environmental Law 26, 1187.

ARTICLE IN PRESSR.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–1130 1129

Pinguelli-Rosa, L., Munasinghe, M. (Eds.), 2001. Ethics, Equity and

International Negotiations on Climate Change. Edward Elgar,

Cheltenham.

Sands, P., 1992. The United Nations framework convention on climate

change. Review of European and International Environmental

Law 1, 270.

Sands, P., 1995. Principles of International Environmental Law.

Manchester University Press, Manchester.

Santer, B.D., Wigley, T.M.L., Barnett, T.P., Anyambe, E., 1996.

Detection of climate change and attribution of causes. In:

Houghton, J.T., Meira Filho, L.G., Callander, B.A., Harris, N.,

Kattenberg, A., Maskell, K. (Eds.), Climate Change 1995: The

Science of Climate Change. Cambridge University Press, Cam-

bridge, pp. 407–433.

Smith, J.B., Schellnhuber, H.-J., Mirza, M.Q., Lin, E., Fankhauser, S.,

Leemans, R., Ogallo, L., Richels, R.G., Safriel, U., Tol, R.S.J.,

Weyant, J.P., Yohe, G.W., 2001. ‘Synthesis’. In: McCarthy, J.J.,

Canziani, O.F., Leary, N.A., Dokken, D.J., White, K.S. (Eds.),

Climate Change 2001: Impacts, Adapation and Vulnerability,

Report of Working Group II of the Intergovernmental Panel on

Climate Change. Cambridge University Press, Cambridge.

Sohngen, B.L., Mendelsohn, R.O., 1999. The timber market impacts of

climate change on the US timber market. In: Mendelsohn, R.O.,

Neumann, J.E. (Eds.), The Impact of Climate Change on the US

Economy. Cambridge University Press, Cambridge, pp. 94–132.

Soons, A., 1990. The effect of a rising sea level on maritime limits and

boundaries, 1990. Netherlands International Law Review 37, 207.

Titus, J.G., 1992. The costs of climate change to the United States. In:

Majumdar, S.K., Kalkstein, L.S., Yarnal, B.M., Miller, E.W.,

Rosenfeld, L.M. (Eds.), Global Climate Change: Implications,

Challenges and Mitigation Measures. Pennsylvania Academy of

Science, Easton, pp. 384–409.

Tol, R.S.J., 1995. The damage costs of climate change toward more

comprehensive calculations. Environmental and Resource Eco-

nomics 5, 353–374.

Tol, R.S.J., 1996. The damage costs of climate change towards a

dynamic representation. Ecological Economics 19, 67–90.

Tol, R.S.J., 1997. On the optimal control of carbon dioxide emissions:

an application of FUND. Environmental Modeling and Assessment

2, 151–163.

Tol, R.S.J., 1999a. Safe policies in an uncertain climate: an application

of FUND. Global Environmental Change 9, 221–232.

Tol, R.S.J., 1999b. Spatial and temporal efficiency in climate change:

applications of fund. Environmental and Resource Economics 14

(1), 33–49.

Tol, R.S.J., 1999c. The marginal costs of greenhouse gas emissions.

Energy Journal 20 (1), 61–81.

Tol, R.S.J., 1999d. Time discounting and optimal control of climate

change—an application of FUND. Climatic Change 41 (3–4),

351–362.

Tol, R.S.J., 1999e. Kyoto, efficiency, and cost-effectiveness: applica-

tions of FUND. Energy Journal Special Issue on the Costs of the

Kyoto Protocol: A Multi-Model Evaluation, 130–156.

Tol, R.S.J., 2001. Equitable cost-benefit analysis of climate change.

Ecological Economics 36 (1), 71–85.

Tol, R.S.J., 2002a. New estimates of the damage costs of climate

change, Part I: benchmark estimates. Environmental and Resource

Economics 21 (1), 47–73.

Tol, R.S.J., 2002b. New estimates of the damage costs of climate

change, Part II: dynamic estimates. Environmental and Resource

Economics 21 (1), 135–160.

Tol, R.S.J., 2002c. Welfare specification and optimal control of climate

change: an application of FUND. Energy Economics 24 (4), 367–

376.

Tol, R.S.J., Dowlatabadi, H., 2001. Vector-borne diseases, climate

change, and economic growth. Integrated Assessment 2, 173–181.

Tomuschat, C., 1999. International law: ensuring the survival of

mankind on the eve of a new century. Recueil des Cours 281,

1–269.

Toth, F.L. (Ed.), 1999. Fair Weather? Equity Concerns in Climate

Change. EarthScan, London.

Toth, F.L., Mwandosya, M., Carraro, C., Christensen, J.M.,

Edmonds, J.A., Flannery, W.B., Gay-Garcia, C., Lee, H., Meyer-

Abich, K.M., Nikitina, E., Rahman, A., Richels, R.G., Ye, R.,

Villavicencio, A., Wake, Y., Weyant, J.P., 2001. Decision-making

frameworks. In: Metz, B., Davidson, O., Swart, R.J., Pan, J. (Eds.),

Climate Change 2001: Mitigation, Report of Working Group III of

the Intergovernmental Panel on Climate Change. Cambridge

University Press, Cambridge.

US Country Studies Program, 1999. Climate Change: Mitigation,

Vulnerability, and Adaptation in Developing Countries. US

Country Studies Program, Washington, DC.

Verheyen, R., 1997. Der Beitrag des V .olkerrechts zum Klimaschutz,

In: Koch/Caspar, Klimaschutz im Recht, Nomos, Baden–Baden.

Verheyen, R., 2002. Adaptation to the impacts of anthropogenic

climate change—the international legal framework. Review of

European Community and International Environmental Law 11

(2), 15–28.

Verheyen, R., 2003. The legal framework of adaptation and adaptive

capacity. In: Huq, Klein, Smith (Eds.), Climate Change, Adaptive

Capacity and Development. Imperial College Press, London, 2003,

in press.

Whitmore, A., 2000. Compulsory environmental liability insurance as

a means of dealing with climate change risks. Energy Policy 28,

739–741.

World Bank, 2000. Attacking Poverty. World Development Report

2000/2001. Washington, DC.

Yohe, G.W., Neumann, J.E., Marshall, P., Ameden, H., 1996. The

economics costs of sea level rise on US coastal properties. Climatic

Change 32, 341–387.

ARTICLE IN PRESSR.S.J. Tol, R. Verheyen / Energy Policy 32 (2004) 1109–11301130