Elements of Change 1995: Session II Summary

This report reflects the deliberations of an expert group that considered ways to improve the long term effectiveness of the global warming regime. Issues covered range from the transfer of financial resources to the management of non-compliance, negotiation of additional commitments, the participation of developing countries, and ways to improve systems that provide scientific advice. It develops several dozen specific recommendations for policy and research.

First, the existing state of knowledge is systematically incomplete in the area of non-market impacts, impacts of all types in developing countries, and potential catastrophes. Policymakers should be aware of these limitations, which probably understate the overall potential consequences of global warming. Further, in many countries political attention to global warming is driven by fear of potential catastrophes and surprises lurking in the climate system. Useful information is needed on the real probabilities and possible consequences of major potential catastrophes. We suggest a method and six case studies that can provide a first crucial analysis.

Second, the international system for negotiating and implementing commitments should be seen as an ongoing process of bargaining. If that process is to be well-informed, and if agreements reached are to be connected to what is actually implemented, the process must be served by mechanisms for exchanging and reviewing information about implementation and handling problems of non -compliance. Most multilateral environmental agreements do not benefit from such mechanisms and are thus less effective than they could be. In the spirit of improving the effectiveness of the Climate Convention, it is essential that policymakers give continued attention to building the system of national "communications" and regular reviews that now exist in nascent form in the Convention.

Third, participation of developing countries is essential to the long term effectiveness of the regime if global emissions are controlled stringently. But the interests of developing countries are different from those of the developed nations that have put the global warming issue on the agenda. Principals are identified to guide efforts to include developing countries in the regime. Because the policy priorities are different, although the developing countries are more vulnerable to climate change the developed world will still need to increase climate-related resource transfers by several orders of magnitude if they want to control global emissions stringently over the long term. Further, it is likely that there will be a need for a practical system that makes a finer distinction among developing countries as efforts are made to tar get control measures and distribute the currently small resources to the countries where they are most needed and effective.

Fourth, there is a massive increase under way in the participation and influence of non-state actors in making and implementing climate -related policies. Devices are explored to expand non-state influence further, but there are many problems that are already evident in the democratic societies that have opened themselves to non-state influence in the making of domestic environmental policy. Researchers need to conduct more comparative case studies of different styles of participation and influence in order to develop some governing principles. In the interim, policy makers should experiment with some different systems that allow more extensive participation and influence by non-state actors. Some of the problems can be managed - for example, the fear that NGOs are not accountable because they don't have responsibilities could be addressed by requiring NGOs to compile reports and submit them in parallel with the government reports, perhaps when the next round of reports under the Climate Convention are due in 1997.

Fifth, virtually every aspect of climate science is marked by uncertainty and the need for scientists to make subjective assessments of probability and consequences. Improvement in the communication of such assessments is badly needed - without it, policymakers have little feel for the range of possible outcomes and how they compare with other risks that society spends scarce resources to manage.

The existing state of knowledge is systematically incomplete in the area of non-market impacts, impacts of all types in developing countries, and potential catastrophes. Policymakers should be aware of these limitations, which probably understate the overall potential consequences of global warming.

In addition to these five major issues, the AGCI group proposes some detailed changes to policy and research on the basis of its evaluation of the long term needs of the climate regime and analysis of the factors that have made other areas of international cooperation effective. These include, for example, actions needed to identify technologies for verification of future climate commitments. And they include ways to improve existing systems for transferring resources and strategies for keeping the geo-engineering option viable in case it is needed in a crisis.

Two dozen experts met over two weeks in August 1995 at the Aspen Global Change Institute (AGCI). They considered ways to improve the effectiveness of the international agreements and institutions designed to manage the causes and consequences of global warming. The group included scholars and practitioners from industrialized and developing countries and trained in international law, political science, physics and economics. Following the long term, cumulative nature of the global warming issue, the group took a long term perspective on the climate regime. They sought to identify what can be done now with policy and research to manage climate change effectively over the next several decades.

This essay reflects the eight major themes and conclusions from the AGCI discussions:

Industrialized countries are relatively invulnerable to market impacts from global warming, yet they are the countries that (currently) have the highest emissions and are being urged to take the lead by implementing perhaps costly measures to control emissions.

The essence of the global warming problem is that emitters of greenhouse gases are forcing the complex natural climate system. The global, regional and local consequences of this forcing are unpredictable. Research can and has made a major contribution to under standing some of the possible and likely outcomes.

Policy on global warming is fundamentally driven by concern about the impacts of changing climates; thus the focus here is on the results of impacts research, leaving aside the costs and benefits of mitigating emissions of greenhouse gases that lead to global warming, which have been reviewed extensively by many others. The impacts fall into three categories: gradual and roughly predictable impacts on marketable goods and services, gradually and roughly predictable impacts on non-market goods and services, and potential catastrophes. The categories reflect how interdisciplinary research is done to measure the impacts. Figure i.1 shows these three categories and assesses the state of knowledge in each for "developed" and "developing" countries.

Two observations are striking. First, current research is focused almost exclusively on the impacts on marketable goods and services. The results of such research suggest that relatively little money will or should be spent to slow global warming. Industrialized countries are relatively invulnerable to market impacts from global warming, yet they are the countries that (currently) have the highest emissions and are being urged to take the lead by implementing perhaps costly measures to control emissions. In contrast, the impacts on developing countries are probably much higher, but those countries have other more pressing demands than to spend scarce resources on abating emissions of greenhouse gases. In other words, the countries being urged to control their emissions don't have much rational incentive to do so. The countries most at risk have higher priorities than global warming.

(Of course, there is an important ethical and legal argument against developed countries transmitting harm to developing countries without some form of compensation or effort to limit those harms. That issue is considered again in a later section on compensation and insurance schemes. Here, the focus is on the distribution of costs and benefits.)

The second observation is that there are sharp differences between what is known about climate impacts, what information is needed, and the forces driving policy. Quantified and monetized results of impacts research are systematically refined in the areas where the impacts of global warming are likely to be lowest: the gradual market impacts on the industrialized countries (cell 1 of figure i.1). This reflects the fact that doing research in all the other cells is extremely difficult. It is also the product of a research monoculture that is equipped only to answer that narrow set of market-oriented questions.

Some answers are needed in the other cells. The impacts on developing countries (cells 4 -6 of figure i.1 ) are crucial. Because resources are scarce in these countries, the decision to spend money on slowing and adapting to global warming requires much better information on the potential consequences. Presumably, the increasing awareness of the intrinsic value of nature will give growing weight to the non-market consequences of global warming in all countries (cells 2 and 5 of figure i.1 ). Yet there is little systematic knowledge about these potential consequences, just some illustrative examples such as the loss of biological diversity in marginal areas. Finally, the politics of global warming have and will be driven by the fear that a nasty surprise is lurking in the climate system (cells 3 and 6 of figure i.1 ). But researchers lack systematic knowledge about possible catastrophes and their consequences. At least one simple inventory now exists (see Elements of Change 1994, Part 2, Anticipating Global Change Surprises , Aspen Global Change Institute), but research has not attached probabilities and consequences to the possible events.

Later in this report these failings are considered in more depth, especially issues related to developing countries and potential catastrophes.

The countries being urged to control their emissions don't have much rational incentive to do so. The countries most at risk have higher priorities than global warming.

The observations here lead to several simple conclusions. For policy, it is essential that decisionmakers understand that current knowledge about potential impacts is improving but remains remarkably narrow. There are systematic differences in the accuracy and completeness of research results; they probably understate the consequences of global warming by emphasizing the market impacts of global warming in developed countries, which is likely to be small. Nobody knows how the understatement compares with what is known, in part because it is difficult to compare non-market and potential catastrophic consequences with gradual changes in marketable goods and services.

For research, the areas where knowledge is weakest are those where the need for information is greatest: in making policy choices in developing countries, in assessing whether the non-market consequences of global warming provide a substantially more compelling case to slow global warming, and in assessing the consequences of possible catastrophes.

For policy and research, it is essential that current efforts to integrate expert assessment continue and that much greater attention be given to the need to communicate judgments about potential impacts to the policy process, especially in the areas where current assessments are weakest but needed most. The missing areas of research all occur in areas where research is extremely difficult to do it is likely that such research will be in complete for the foreseeable future, and thus the policy community will long be in the position of needing information that is incomplete and thus surrounded by large margins of error. Scientists will be asked to make incomplete and subjective assessments because at least some information is better than none when making policy. How to gather and convey such assessments in a useful and accurate manner is discussed in the final section of this report.

Scientists will be asked to make incomplete and subjective assessments because at least some information is better than none when making policy.

Catastrophe is an effective motivator. It is also a legitimate concern since possible catastrophes merit serious policy attention even when their probability is low and uncertain. Societies often spend a premium on avoiding catastrophic losses such as war and nuclear meltdowns. This issue is focused upon here because the other major gaps in impacts assessment identified in the last section are known, and some efforts are under way to fill them. In contrast, the research community has not gone far to identify and study empirically potential climate catastrophes. Yet in some countries, fear of catastrophe in the climate system may be the single most important factor driving the politics of global warming. That may be especially true in industrialized countries which are mostly not particularly vulnerable to gradual impacts. Fear of catastrophe may prove a strong motivator for these countries to implement costly measures to slow global warming. Over the long term it is essential to know whether these fears are warranted and how they compare with other policy priorities.

The way forward on catastrophe research is to conduct some detailed case studies on major potential catastrophes using a common method. Research in this field will always be unsatisfying because the list of completed case studies will always be shorter than all the conceivable catastrophic outcomes. But some studies will give a sense of the issues at stake and create a research community that is informed to make the subjective assessments about catastrophic risks that are badly needed. An initial set of cases and a research method are proposed below. Each of the cases is paired with at least one historical analogue. None of the analogues is perfect, but all help illustrate the hypothetical case with real stakeholders, outcomes and responses.

Six cases are identified that span the range of major climatic catastrophes now discussed in policy and research circles. Two concern catastrophes in the classic sense - they are "tipping points" where a complex system rapidly shifts from one mode of operation to another:

Four cases reflect extreme versions of the gradual and roughly predictable impacts of climate change:

The way forward on catastrophe research is to conduct some detailed case studies on major potential catastrophes using a common method.

It is not argued that these scenarios of potential catastrophe are accurate or complete, only that they are the major catastrophes currently under discussion. They are the symbols being used to motivate more action on global warming; it would be good for the research community to provide more useful information on their likelihood and consequences.

First, each case should consist of a prescribed physical scenario. The main problem with current research on climatic surprises and catastrophes is that the social and institutional causes and responses are not adequately explored; yet they are the most important factors in most (but not all) cases. A prescribed physical scenario will help increase the focus on the non-physical factors. The scenario should reflect the range of possible outcomes and also underscore the role for natural scientific research to improve its predictive capacity.

Second, as noted, it is crucial that cases have empirical focus. The careful use of historical analogues can help sensitize the researcher to the range of stakeholders, causes and responses.

Third, the physical scenario should include rapid and slow modes, and research into social and institutional responses should investigate both. The rapid mode of change helps to explore how societies might react to catastrophes when they are on the doorstep. The slow mode helps illustrate how catastrophe might be anticipated. In practice, any catastrophe might catch some people by surprise while others anticipate and prepare for the outcomes. The distribution of costs from catastrophe might reflect which populations were able to anticipate and adapt.

Fourth, the entire exercise must be conducted in the spirit of imagining the world 50 to 75 years in the future when these possible catastrophes might have full effect. It is inappropriate to imagine these changes imposed on the current vision of the world or to assess all catastrophic outcomes as "caused" by global warming. Rather, the full story is the interaction of warming and other pressures with social change. Weather prediction will likely improve in the future. In turn, the improved capacity to predict the weather could make it easier to anticipate and adapt to some extreme events such as storms. Growing populations will settle new lands, thus fragmenting habitats even more than today. Climate change could exacerbate the effects of fragmentation, but affluence and growth would be the principal destroyers of natural habitats, not climate change.

The identified needs apply mostly to the research community. The immediate need is to conduct some case studies according to a common method, perhaps under a single integrated research project to allow comparisons across cases. The case research will demand interdisciplinary teams, great care to integrate the relevant but disparate existing research and historical studies, and extreme care to forecast the future societies that will respond to the possible catastrophes. After conducting this research, attention must be paid to communicating the results in a way that conveys what is at stake as well as the (wide) range of possible errors and outcomes (considered in the final section of this report). Implementing this style of research and communication will be extremely difficult, but without it the research community will lack answers to critical policy questions.

A global solution to climate warming requires international flows of resources to help poorer countries control emissions.

International regimes perform many functions. They help coordinate otherwise independent national actions. They establish systems for exchanging and assessing information. They offer fora for continued negotiations. These tasks are crucial; all are being performed in the climate regime, and this group is optimistic they will be performed well.

Here, the focus is on four other functions: (a) managing the international flow of relevant financial resources; (b) technological planning; (c) possible insurance schemes; and (d) managing geo-engineering. Each of these conceivably has a role in the long term climate regime but has received less attention than the other functions of international regimes. They are discussed in order, from the most immediately relevant and needed to the most distant.

Virtually every analyst agrees that a global solution to climate warming requires inter national flows of resources to help poorer countries control emissions. In general, explicit flows of resources for environment and development are very small today. The Marshall Plan transferred money and other benefits measured in a few percent of U. S. gross domestic product to war-torn Europe. Today, explicit flows for environmental purposes, as a percentage of GDP, are several orders of magnitude smaller. Total Official Development Assistance (ODA) for all purposes is only (0.4%) of the world product. Donors today are unwilling to allocate substantial sums for international transfer through formal multilateral institutions.

This context is not ideal for considering potentially very large resource flows that might be needed to slow global warming, for example to pay the incremental costs that developing countries incur while implementing projects. If global warming proves to be a concern that merits stringent action over the next few decades those flows might ultimately be measured in hundreds of billions of dollars annually - a large sum but only a percent of world product. The regime needed to transfer those sums is radically different from what we have in place today. The group acknowledged this major shift that might be needed but did not consider its consequences further. It did explore what, in the interim decades, could be done to improve existing systems that transfer much smaller but still valuable sums.

The experience with financial mechanisms to date offers four lessons, some of which contravene conventional wisdom.

First,it is essential to couple donor and recipient interests. The large supply of new and old resources programmed for "environmental" purposes reflects mainly the greening interests of the donor community. In practice, many projects funded under "environmental" programs are camouflaged with environmental components by recipients to make them more attractive to donors. Formally that camouflage is discouraged, but in practice it is tolerated. Within limits, it should be encouraged because it leads to a larger supply of potential projects that meet donor and recipient interests.

Second, the international community discourages aid that is tied to particular conditions such as the use of certain technologies or national firms, but in practice the ties are numerous. There is some danger that formal prohibitions are not reducing ties but rather making them less transparent and perhaps exacerbating the problem. In principle, some ties might actually be welcome. They improve the fit with donor interests and thus increase the flow of resources available for environment and development purposes. Some conditions must be met for tied aid to be effective, including that ties don't undermine multilateral lending and coordinating mechanisms and that ties don't lead to inappropriate technologies. These conditions are well established (if not always followed), but the point here is simply that tied aid might actually be (cautiously) welcomed.

Third, often many different bilateral and multilateral lending activities programs operate in a given issue area. Competition can be healthy, but some coordination is probably also essential. Mechanisms such as the Project Planning Committee (PPC) for aid programs in Central and Eastern Europe offer useful lessons for improving coordination. The PPC encourages co-financing, dialogue and faster project cycles. Areas where the PPC has not been successful are also illustrative of the needed policy changes: it does not involve recipients, and it is unclear if the coordinating mechanism also improves the setting of priorities or supply of new resources. Nonetheless, this and many other examples underscores the role of multilateral coordinating mechanisms.

Fourth, a large number of other lessons are being learned in the systems for international resource transfers. Governance is expanding to include recipients and non-state actors both in setting policy and implementation. And there is increasing attention to "capacity building;" increasingly, the concept of "capacity" is shifting from one focused on government institutions to include non-state actors and institutions. All these lessons underscore that the international system for resource transfers is dynamic and, in general, appears to be learning how to become more effective.

The issue of making financial transfers more effective is enormous. Some of the more de tailed observations from experience are discussed in the individual reports. (See Gutner, Gyawali, Najam, Ponce-Nava). Among the many finance-related topics that are not discussed here is the large number of market-oriented systems, such as joint implementation and tradeable permits, that might substantially increase the flows of resources related to global warming.

The international system for resource transfers is dynamic and, in general, appears to be learning how to become more effective.

Technology touches every aspect of the global warming issue. Yet few studies approach the technological dimension of climate change in a systematic way. Most of the international attention to technology has concerned "technology transfer," yet donors rightly note that nearly all technologies are "transferred" through normal market channels. The long history of attempts at overt technological transfer repeatedly underscore that successful transfers envision technologies not just as hardware but also as a complex of institutional capacity and training. Technical change could be a large part of the solution for reducing emissions of greenhouse gases, but government-driven efforts to plan technology policies are mixed in their efficacy. Normal market-driven innovation is poorly understood. It is unclear which incentives best spur the innovation and diffusion of appropriate new technologies. All of these issues are at the center of current economic and technological research related to climate change.

Given the magnitude and uncertainty of the technological dimension, it is easy to despair that no single suite of technology policies seems right to implement. In the short term, at minimum, one sensible approach is to continue what nearly all industrial countries are already doing: implement so-called "no regrets" policies that help reduce market failures and reduce greenhouse gas emissions at little or no cost. Often these consist of diffusing new, efficient technologies into use. The "no regrets" options are well-known. But the success stories need more attention. Demand-side management programs such as EPA's "Green Lights" appear clearly cost-effective yet are under budgetary threat because their success is not well known.

In addition, it would be useful to identify existing technology that is applicable to monitoring and verifying compliance with possible climate commitments and for research. These include technologies developed for other purposes, such as military intelligence. Much of this is being done already, but special attention is needed to explore the functions of monitoring and verification so that a foundation of experience can be built in time for its use in the coming decades if substantially more stringent climate commitments are agreed to. As the stringency of international commitments increases, so do the incentives to cheat and the need for verification.

The resource transfer and technological dimensions of the global warming problem are well known. Next we turn to two issues insurance and geo-engineering that could affect the long term climate regime but are less well known.

Insurance is used in many settings as part of a package of policies and incentives for managing risk. Risks can be avoided; losses can by limited and controlled; but some risk remains. Insurance is purchased in an effort to transfer some of the remaining risk to insurance providers. Insurance premiums help sensitize risk bearers to the costs and consequences of their behavior.

In domestic and international settings the law has tended to focus on incentives to avoid risks, such as through injunctions and liability. The Framework Convention on Climate Change and the general debate on climate change have mostly adopted the same philosophy of risk avoidance. But a package of avoiding some and retaining other risks might be a more integrated, efficient and fair way to deal with climate change. A climate insurance scheme would be a central part of a system to sensitize risk bearers; premiums would highlight activities vulnerable to climate change and attempt to limit those while the climate regime also attempted to limit the sources of risks: greenhouse gas emissions.

At minimum, nations should implement so-called "no regrets" policies that help reduce market failures and reduce greenhouse gas emissions at little or no cost.

Appealing in principle, insurance schemes currently face a large number of practical problems. These include the fact that any initiative must be implemented in the con text of large existing (mostly private) insurance markets. Further, one does not insure against "global warming" but rather against particular impacts. Those impacts are not known with precision, and it is difficult to identify what fraction of a particular hazard is caused by the normal climate system and what part is the result of anthropogenic global warming. Separating the two will require research and, ultimately, impartial expert judgment. Nonetheless, separating the sources of risks is done in other contexts and not inconceivable in the case of some global warming impacts.

An interesting possibility is to couple the anthropogenic portion of the risk to a system of international compensation such as a premium subsidy. Compensating some risks of climate change could be a lot cheaper than avoiding risks entirely through abatement of greenhouse gases, and compensation would also be more fair than the current practice of offering no compensation for international environmental hazards. Politically, compensation might be impossible to implement, but coupling it with an insurance scheme might make it more attractive.

Clearly an international insurance scheme is not ready for practical application in the near future, but the contribution to efficient management of the climate problem is potentially enormous. One conclusion for the research community would be to elaborate and test some possible schemes. In the interim, one conclusion for policy is to make existing insurance markets more efficient and rational so that risk bearers in marginal areas ( e. g., coastal zones) are appropriately more sensitive to possible hazards. Indeed, climate politics might be driven by premiums set in normal insurance markets by an industry wary of hazards such as storms and flooding. This is not a full solution to the climate problem, but it may help improve the connection between global warming policy and some hazards.

Finally, many observers note that geo-engineering has a potential role in the suite of policy responses to global warming, especially if a climate catastrophe occurs or is imminent. Geo-engineering options such as putting mirrors in space, sulfate in the stratosphere, or iron in the Antarctic Ocean have many problematic features. Some merely postpone climate change and demand perpetual intervention in the climate system. None of these is clearly devoid of potential harm. Thinking about such options may breed complacency. But in a crisis, geo-engineering might be needed.

Based on earlier treaties, international guidelines and customary international law, there are some loose norms that may govern the appropriate use of geo-engineering. These include the duty to prevent transboundary harm, the duty to undertake assessments prior to actions that might cause environmental harm, and the duty to consult with potentially affected states.

The Framework Convention on Climate Change says nothing directly about the use of geo-engineering. Presumably the Subsidiary Body for Scientific and Technological Advice (SBSTA) and the Conference of the Parties would handle such issues. Other relevant international agreements such as those on Outer Space or Antarctica which are relevant to the space mirrors and iron-enrichment solutions, respectively also say little.

Given that the existing suite of international norms is only loosely relevant, would it be wise to consider developing additional norms specifically dedicated to geo-engineering now, before the geo-engineering options might really be needed? The sense of the AGCI group is that the answer is "no." Unless the climate system faces a crisis, it is unlikely that geo-engineering options could be seriously evaluated in terms of their real need and function; presented with the choice only in the abstract, most countries would appropriately seek to ban geo-engineering. A ban might prove irrelevant in the future if geo-engineering were really needed, but in the interim it could dissuade research specially targeted at geo-engineering options. That, in turn, would leave the world less prepared.

The range of functions to be performed by the international climate regime is enormous. No effort is made here to be comprehensive. Rather, several conclusions that emerged from the discussions are highlighted.

Regarding financial transfers, many lessons are being learned about how to improve the existing system. Some of these are unexpected: under some conditions, tied aid and "camouflaged" projects might actually be welcome because they improve the fit between donor and recipient interests and in crease the international flow of resources targeted to the environment.

Nonetheless, the overall amount of money now transferred through official aid programs is very small compared with the amounts that might be required if there is stringent global action to manage climate change. An important observation for policy is that all the adjustments to improve the operation of bilateral and multilateral funding mechanisms discussed here are marginal changes to the existing system. There is no reason to believe that if large-scale transfers are needed that it would be wise to scale up the existing system.

One conclusion for research is that some exploratory studies might be done on the possible design of a large-scale resource transfer system, how it could fit with (or replace) the existing systems, and the relationship to private markets through which most international resources flow. Historical analogues include the Marshall Plan. Such research contrasts with most work on financial mechanisms which uses the existing system as a starting point.

Regarding technology, our main conclusion concerns policy: successful programs, such as demand-side management that promotes energy efficient technologies, need more publicity. At present there are few clear ideas of what technological responses are best. Some clear and well-documented success stories can help show the way over the short term. The longer term needs include the identification of technologies for monitoring and verification purposes and to build a verification system that can be used if substantially more stringent climate commitments are adopted.

Regarding insurance schemes, there is a strong need for the research community to develop and test some workable ideas.

Regarding geo-engineering, we have already indicated that no policy response is appropriate now it is better to wait for a clearing indication of crisis. However, research should continue. Although many participants in the climate debate fear that geo-engineering options will be treated as a panacea, it makes sense to sponsor research into the options so that they may be available if needed. Fortuitously, because geo-engineering is intervention in the climate system, most of the needed research is exactly that also needed to under stand the climate system generally.

Regarding geo-engineering, no policy response is appropriate now it is better to wait for a clearing indication of crisis. However, research should continue.

It is helpful to examine other cases of successful international cooperation to draw lessons for building an effective climate regime. Several studies are underway that do exactly that. Here we look only at the two cases acid rain in Europe and depletion of stratospheric ozone that are cited most often as "successful" and draw some lessons for climate.

The regime to control acid rain in Europe formally dates to the 1979 Convention on Long Range Transboundary Air Pollution. Five protocols have elaborated the terms of the Convention: a protocol to fund a system for gathering and exchanging data and model results (1984); two protocols to limit emissions of sulfur dioxide (1985 and 1994); a protocol on emission of oxides of nitrogen (1987); and a protocol to limit emissions of volatile organic compounds and other precursors of ground-level ozone (1991). Although cooperation is formally codified in these agreements, collaborative monitoring programs date much earlier a network of scientific research and monitoring activity was in place at least since the early 1970s, waiting for a political window of opportunity.

As in most regimes, the "success" of the European acid rain agreements is a function of luck, narrow interests, and ingenuity. Here the focus is on the element of ingenuity because it informs most directly what types of policies are feasible and effective. (The full story is told in the summary of Marc Levy's presentation.) Ingenuity's contribution to the acid rain regime's "success" can be traced primarily to two elements.

First, the international legal agreements to control emissions were a framework within which national efforts to control emissions were coordinated. The framework also sponsored and legitimized scientific research, modeling and assessment activities; it helped build shared knowledge and concern about the acid rain problem.

Second, the working groups on impacts and the concept of critical environmental loads introduced in the late 1980s helped focus attention on the potential harms of acid rain. Research on the impacts on forests was not conclusive; in contrast, impacts research on lichen was better able to demonstrate harm. Even though the economic consequences of declining lichen might have been small, the demonstrable link gave human populations in affected areas a tangible reason to press for political action to control acid rain. The regime was effective in changing interests because it was effective in connecting scientific information on impacts to the political process.

Once the regime was in place, it became the focal point for a political process driven by the fear of damages and the desire to be a responsible neighbor. The regime's research and assessment activities were able to demonstrate that pollution crossed borders and caused harm. Faced with that information, nations felt an obligation and domestic pressure to control their emissions.

The acid rain regime also faces many unsolved problems. It has no operational system for reviewing implementation and dealing with problems of non-compliance. It has made little use of financial transfers despite strong evidence that transfers to poorer countries would improve implementation and overall effectiveness in a cost-effective manner. Generally, the regime has not given central attention to comparing costs and benefits - its control actions have been economically sub-optimal.

The Montreal Protocol (1987) on Substances that Deplete the Ozone Layer is the legal agreement at the core of the most extensively studied international environmental regime. The regime is well-studied probably because it is widely seen as a success. However, there are many dangers of using the ozone regime as a model for international cooperation in other areas. The rapid phase-out of CFCs can be explained by the distribution of costs and benefits of the phase-out. The firms that stood to benefit from the creation of markets for CFC substitutes were the same firms that produced the original CFCs. Those who paid the costs of the CFC phase-out were CFC users and consumers - a diffuse and poorly organized group. The phase-out was rapid because this special set of industrial interests happened to overlap with interests in protecting the environment.

A more common situation (including global warming) is that those who pay the costs of abatement are well known, concentrated and well-organized (or become so), and generally opposed to stringent controls. The beneficiaries of environmental protection are diffuse and often not well organized. (However, non-governmental organizations acting on behalf of the environment are an increasingly well-organized representative of environmental interests.) This situation leads to an imbalance of political pressure and systematic failure to establish and implement environmental policies that are collectively beneficial. One strategy is to ensure that diffuse interests are well-organized and represented in the political process. (That and many related issues are considered in a later section on transparency and governance.)

Another strategy is to make more systematic use of compensation to pay the extra costs of parties that would not otherwise implement the needed changes in behavior. In contrast with extortion, where a party purposefully does something that it wouldn't do otherwise in order to extract a payment, compensation is an under-utilized tool for reaching more ambitious collective agreements. Compensation may be essential in order to build viable coalitions that favor certain environmental policies where opposition is well organized. It is most useful in environmental problems such as global warming where policymakers can predict that narrow interests will be opposed and well organized.

The ozone regime provides an example of multilateral compensation: a Fund pays the incremental cost of implementing projects to comply with the Montreal Protocol. Countries otherwise reluctant to join the Protocol have now done so because the Fund pays these costs. The Fund also illustrates some of the difficulties of international and multilateral funding: contributions to the Fund have fallen behind schedule, and the agreement that established the Fund is so delicate that few Parties want to raise these problems formally. The Fund is based on expert assessment of country programs and individual projects; such assessment help avoid inflated claims and extortion.

In addition to the Fund, some other forms of ingenuity have also contributed to the over all effectiveness of the ozone regime. Notably, multiple economic, scientific and technological assessment panels are connected to the regime, provide periodic assessments, and are responsive to particular questions that arise during the negotiations. Those panels help connect up-to-date scientific information to the actual operation of the regime - they provide legitimate answers to crucial questions such as which uses of CFCs are "essential" and thus eligible for exemptions under the Protocol. Further, the regime is highly flexible and thus able to respond to new information - the control measures agreed upon in 1987 have undergone a major strengthening in 1990 and again in 1992. Unlike the acid rain regime, the Protocol also includes a dedicated fund for financial transfers which has improved implementation of the control measures in developing countries. Finally, the Protocol has a system for handling problems of non-compliance - since 1990 that system has significantly improved implementation of the Protocol's requirement to report data on production and consumption of ozone depleting substances. That system is just now starting the more difficult task of improving implementation of the control measures - the early signs are positive.

There are many dangers of using the ozone regime as a model for international cooperation in other areas. The rapid phase-out of CFCs can be explained largely by the distribution of costs and benefits of the phase-out.

Only some aspects of only two other cases are discussed here. The range of relevant lessons from "success" stories and from failures is enormous.

For policy, these cases both illustrate the importance of connecting science to the regulatory regime. The acid rain case illustrates how connecting information on impacts helps to sensitize people and nations to the issues at stake and makes abstract notions of environmental damage more concrete. The ozone case illustrates how responsive assessment panels helped to connect detailed technical and economic information to the detailed operation of the regulatory regime. In both cases, the independence of working groups and panels gave their information credibility; the formal connection to the regulatory regime gave it legitimacy. This group believes that the SBSTA of the Climate Convention is headed in this direction; however, there are dangers that the climate regime will be dominated by political representatives rather than more independent experts, which could reduce the credibility of its products.

Some body will be needed to perform the responsive assessments that proved very useful in the ozone regime. The panels being established under SBSTA will and should per form this function. They should be designed with the effective ozone panel system in mind. A critical issue that must be resolved early is the relationship to the IPCC. If IPCC works on the same issues that are on SBSTA's agenda, there is danger both of duplication and that participants in the climate debate will look past SBSTA to the more independent (and perhaps more credible) IPCC for answers.

Policy will also benefit from more attention to compensation schemes. Unlike the phase -out of CFCs, global warming and most other environmental issues are marked by well-organized opposition to regulation. Compensation strategies can convince opposing forces and lead to better package deals while also avoiding extortion.

For research, there remains much to be learned by comparative studies of other international environmental regimes. Much of that work is already underway. The two cases discussed here suggest at least one gap that researchers might fill. It is easy to over-state the importance of the lucky overlap of industrial and environmental interests. But the same situation did not necessarily exist for the phase-out of halons, which was completed on an even faster timetable than for CFCs. What were the industrial interests at stake? Did the halon phase-out proceed rapidly for the same reasons (which seems unlikely), were halons caught in the coattails of the CFC phase-out, and/or did ingenuity in the design of the regime and strong domestic pressure lead to the rapid phaseout? Answers to these questions and comparisons with the voluminous existing literature on the CFC phase-out will help identify the different roles of industrial interests, domestic pressure and regime design in the overall operation of the ozone regime. That case study has not been undertaken.

Unlike the phase-out of CFCs, global warming and most other environmental issues are marked by well-organized opposition to regulation. Compensation strategies can convince opposing forces and lead to better package deals while also avoiding extortion.

The following pages review what policies and measures are being implemented by different countries and then focus on four interlocking aspects of future commitments and their effective implementation: a) the content of future climate commitments; b) translation of international commitments into domestic action; c) systems for managing compliance; and d) the relationship between climate commitments and other international commitments, primarily trade law.

The main substantive commitment of the Convention is a loose pledge that all industrialized countries return their emissions of greenhouse gases to 1990 levels by the year 2000.

The story of which countries are implementing what policies is told at substantial length in individual national communications submitted to the Climate Convention as well as a synthesis report and independent evaluations by NGOs and an especially useful report by OECD. We briefly discuss one country, the United States, to give a sense of the ways that it is trying to meet its climate commitments and then broadly review implementation in other countries.

The main substantive commitment of the Convention is a loose pledge that all industrialized countries (OECD members minus Mexico) return their emissions of greenhouse gases to 1990 levels by the year 2000. Although the commitment is modest, the United States is finding it difficult to create a "climate policy" in the current political con text which is wary of government intervention. Within this constraint, many sensible actions are being implemented:

Individually, none of these measures makes a substantial dent in U. S. emissions. It seems likely that the U. S. will fail to meet the Convention's loose target. However, it is difficult to predict whether a particular target will be met at a distant date, or to plan definitively for that goal. Emissions are intrinsically tied to economic growth and industrial structure, so the future path of emissions is not predictable. Nor is it possible to predict exactly what policies will be implemented with what efficacy.

The developed countries, which are already implementing many "no regrets" options, will probably agree to even deeper cuts in emissions while low-cost options remain plentiful in other countries. Joint Implementation allows both the high- and low-cost countries to gain an advantage.

All other developed countries are also implementing at least some policies; a few ( e. g., Denmark, Norway and Sweden) have implemented significant carbon taxes; most are relying primarily on no regrets options such as demand-side energy management. The overall level of abatement is uneven. Most share the U. S. experience that it is difficult to meet specific quantified emission targets. The only countries that will clearly meet the Convention's target are those that will over-comply.

There is no clear way forward for the negotiation of future commitments, a process that was set in motion by the Conference of the Parties in April 1995 and is slated to have a protocol of the Convention ready for signature perhaps in 1997. It is virtually impossible to calculate what types of commitments are optimal because the costs and benefits of mitigating global warming are so uncertain. Thus as with the acid rain regime and so many other areas of international cooperation (e. g. , the GATT), the style will and must be one of collective learning and consensus-building. What happens in the next few years on detailed policy measures matters less than whether the foundation is in place to allow learning and periodic review and adjustment of international commitments and domestic actions over time. In that spirit, the following principles should guide the next few formative years:

The debate over adopting new quantified targets and timetables for controlling greenhouse gas emissions is not the most important activity. Indeed, it may be especially important not to lock in a target and timetable system because that system for expressing commitments is incompatible with the low ability of countries to predict their future emissions and needed policies exactly;

The system of national communications and international review is essential be cause it offers a foundation for determining whether countries are on track in the implementation of long term policies and measures needed to reduce emissions of greenhouse gases from the economy. The style of reviewing policies rather than focusing on end-of-pipe emissions is more consistent with the types of actions that governments will actually implement to control emissions.

The JI pilot phase should be run to build some shared experience with and trust of tradeable offset schemes. If climate commitments are made substantially more stringent, then a viable and operational JI scheme will be enormously valuable. JI is most helpful in reducing the cost of controlling greenhouse emissions when the overall emissions caps are tight and countries that are obligated to make the deepest cuts have already exhausted their "no regrets" options. That is exactly where the climate regime is now headed: the developed countries, which are already implementing many "no regrets" options, will probably agree to even deeper cuts in emissions while low-cost options remain plentiful in other countries. JI allows both the high- and low-cost countries to gain an advantage. But the crucial first step is to build trust in this concept which is currently shrouded in suspicion and to demonstrate in a pilot phase that it works.

The AGCI group feels that the climate regime is broadly headed in the right direction. How ever, there remains a central danger: because governments are responsive to domestic pressure and some pressure groups have pushed hard for more targets and timetables, it will be easy for governments to slip into another round of target-making. In 1991, while the Convention was being negotiated, nearly every OECD country unilaterally announced a domestic target to limit greenhouse gas emissions; that round of target-setting was in tended to induce the Convention to adopt a quantified target (it didn't) and show that these countries were taking to lead to slow global warming. Yet in the intervening years it has become clear that there is actually little relationship between the declared targets and the policies these countries actually implemented. Over the next two years while a 1997 protocol is being negotiated it seems likely that countries will engage in another round of unilateral target-setting in an effort to influence the content of the Protocol. That would not be intrinsically bad, provided that it does not detract from the main mission of building a long-term system for learning and adjustment rather than chasing short-term political advantage through good-looking targets.

The process by which international commitments are translated into domestic action varies across countries. Some of the major issues are discussed elsewhere. (See Raustiala) The following observations relate to how the international system of setting and adjusting climate commitments should be viewed.

First, nearly all of the actions required to control greenhouse gas emissions are at the domestic level where legislation, programs and incentives are created and implemented. Effective international commitments must be negotiated on the basis of whether they can be given domestic effect. Otherwise a gap will develop between international law and domestic action; countries will be reluctant to ratify binding international agreements that they can't implement; international law will become empty of influence. Consequently, building a system for gathering and reviewing information about domestic implementation is crucial; without a shared base of information there is no foundation for setting international commitments that can be effective domestically. As international commitments become more stringent, the process of domestication will become even more difficult yet crucial.

Second, the state-centric international system requires that states give their consent for international commitments to have effect. Although it is complicated and cumbersome, the process of domestication is the source of legitimacy for international law. It is now popular to discuss ways to avoid domestication, especially the formal process of ratification - systems of adjustable annexes, for example, offer a quick fix to the often long delays in ratification. It seems appropriate to use these devices where detailed or technical matters are at stake and ratification may not be essential, but we caution that for larger issues such quick fixes are counter productive.

The process of domestication might benefit from more transparency which will assist in comparisons of national efforts and compliance. Making domestication part of a system of periodic reviews of policies and measures will promote more complete learning about the types of commitments that governments can actually undertake.

The existence of some system for handling compliance problems is essential to an effective international regime that keeps international commitments connected to domestic action. Parties that don't implement their commitments are scrutinized and cajoled by the operation of compliance reviews into doing more. Systems for managing compliance are part of a style of effective international governance that sets, adjusts and re views commitments on a regular basis.

International environmental agreements have generally not given much attention to the management of compliance. Experience elsewhere shows how the system can operate. The OECD conducts regular policy reviews of its members, including reviews of environmental policies. The GATT Trade Policy Review mechanism operates in the same spirit of reviewing overall trade performance and highlighting specific areas for improvement. The International Labour Organisation (ILO) has a long standing system of supervision that reviews national implementation of ILO commitments and has been effective in identifying cases of inadequate implementation and in nearly all instances convincing countries to bring domestic actions into conformity with their international commitments.

Two environmental agreements have some significant form of compliance review. The Convention on International Trade in Endangered Species (CITES) handles about 100 cases per year of suspected non-compliance. In each, the Convention's secretariat communicates with the affected party and then forwards a report to the Conference of the Par ties for action.

The Montreal Protocol has had a non-compliance procedure and Implementation Committee in operation since 1990; it is active on most issues of compliance and has convinced many parties to meet their international obligations. In all these systems the style of managing non-compliance is less to confront an accused and more to discuss and manage problems of compliance as they arise. Tough actions such as expulsion from a treaty are rare; the goal is to develop plans and time tables for bringing problem states back into compliance and to preserve the integrity of international commitments. Some observers lament that these systems are not tough enforcers of the law, but in practice tough systems are rarely used in international law. Dispute resolution systems, for example, are omnipresent in international environmental treaties but have never been used.

The climate regime is developing a compliance review system in the same spirit as discussed here. The process for elaborating that system will be slow because Parties are unlikely to agree to compliance review without knowing more about their commitments - willingness to adopt tough commitments may be inversely related to the perceived effectiveness of compliance review. However this issue evolves, for now it seems on the right track. Two functions are underscored as essential foundations for compliance review:

Climate change is primarily caused by burning fossil fuels for energy, which is (presently) an integral part of modern economies. Not surprisingly, climate policies overlap with many other aspects of national and international policy. Here we focus on one: the international trading system.

Trade agreements are a crucial part of the international economic system. Given their important role, there are many potential linkages. Rewards of membership or other privileges in trade agreements could be linked to membership in the climate regime. Trade sanctions could be threatened to enforce compliance with climate commitments (although that is rarely done in other environmental agreements). Commitments in trade and climate agreements can be coupled to avoid competitiveness problems from implementing climate commitments while trading partners do not.

There are also many potentially beneficial common interests between free trade and climate protection. Reduction of subsidies on dirty fossil fuels (especially coal) help level the terms of trade while also favoring cleaner and more economic energy sources. Harmonized international standards for technologies or agricultural practices can create uniform reductions in greenhouse gas emissions while also leveling technical barriers to trade. Trade liberalization promotes growth and the efficient use of resources, perhaps also lowering the intensity of greenhouse emissions per unit of economic output.

Unfortunately, in practice few constructive links actually exist. Conflict between environmental and trade regimes has been the norm and is likely. The "trade and environment" debate has had a chilling effect on the willingness of environmental treaties to use trade measures and links. Concern about the trade consequences of using trade measures in environmental agreements has led to a conflict between regimes. In general, the trade community has prevailed.

In managing non-compliance, the goal is to develop plans and timetables for bringing problem states back into compliance and to preserve the integrity of international commitments.

The issues of setting future commitments, domestication of international commitments, managing compliance and building links with other international agreements have been considered together because they are inseparable. They must be viewed as a system for establishing, implementing and adjusting international commitments and domestic policy. The fundamental goal of the system is primarily to change domestic policies and actions; doing so in a cost effective manner requires that the international system be able to gather, review and assess information about domestic policies and measures in parallel with negotiating and adjusting international commitments.

The most important conclusion for policy is that in the absence of a clear way forward on climate commitments it is important to view the climate regime as a focal point for learning, negotiation and coordination. Building this spirit over the long term depends crucially on a good foundation of information about what countries are implementing. A future system for managing non-compliance also requires such a foundation. Thus the central tasks for the next few years must be to ensure that the system of national communications remains useful and informative and that a system for regular reviews of those communications develops properly. Other tasks such as negotiating new commitments can also be pursued, but it is important that they not interfere with the foundation. Nearly every other international environmental agreement has failed to develop active and useful systems for reviewing and exchanging information.

In addition, both the policy and research communities are aware of the potential linkages between trade and environmental measures. So far, efforts to forge the links have been incomplete. If those communities were to elaborate some criteria to define when trade linkages are useful for environmental purposes without impeding the trade regime. That could help identify some common ground.

Finally, the policy and research communities should elaborate some criteria to define what types of international commitments demand what styles of domestication. At one extreme, international agreements are not well served if any change even to a technical annex requires time-consuming national ratification. At the other, fast track procedures that allow major substantive changes to international commitments without domestic consent are likely to be illegitimate. (They are also likely to be ineffective because consent through ratification often includes the creation of legislation that gives domestic effect to international commitments.) But what mix of commitments is likely to be most effective and legitimate in the middle ground? The research community can contribute to this task by conducting case studies and comparisons - there are many historical examples of different types of international commitments and styles of domestication.

At one extreme, international agreements are not well served if any change even to a technical annex requires time-consuming national ratification. At the other, fast track procedures that allow major substantive changes to international commitments without domestic consent are likely to be illegitimate.

The most important long-term issue in climate management is the involvement of developing countries. Fifty years in the future the largest share of emissions will come from these countries. The poorest nations suffer the greatest vulnerability to changing climate. Aware of the hazards of discussing hundreds of different national circumstances as a single group of "developing countries," some basic issues and trends are identified below.

Efforts to involve developing countries in the climate regime must begin with acute awareness of the context in which they would participate. There is a rapid increase in concern about environmental issues, reflected in the creation and rising status of environment ministries. Non-governmental organizations have also arisen to address these issues. But the phenomenon is recent in comparison with the developed countries. Many environmental issues, especially global ones such as greenhouse warming, have been put on the agenda by outside pressure. Local and immediate issues such as air pollution and access to safe water have a much stronger domestic constituency.

Most important is that the policy priorities appropriately reflect mostly local needs and interests. Local pollution and resources (e. g. , water) dominate the environmental agenda. Climate change may resonate strongly with some of these priorities such as controlling burning of dirty fossil fuels or managing water resources but in general the climate issue is mostly disconnected. In many developing countries, the productivity of safe water and agricultural resources is al ready under stress and thus especially vulnerable to additional stresses caused by climate change. There is at least some need to start planning for possible impacts of climate change, and especially there is a need for the developed countries that are causing global climate change to assist in that prudent planning.

The prospects for decision making are relatively good. Science is generally held in high esteem, and scientists are increasingly engaged in making policy. As in the developed countries, setting coherent priorities for managing environmental risks remains a problem.

Seven principles should guide and condition efforts to improve participation of these countries in the climate regime over the next several decades:

In developing countries, global warming has been put on the agenda by outside pressure. Local and immediate issues such as air pollution and access to safe water have a much stronger domestic constituency.

For policy, a long term goal is to improve substantive participation of developing countries in the climate regime. That does not preclude much swifter and more stringent action by developed countries. But over the long term an integrated approach to climate change increasingly can't ignore the contributions and vulnerability of the developing world. The seven principles identified above should be kept in mind as policymakers adapt the international regime and national policy to include developing country interests.

It is striking that every analysis underscores that "developing countries" are not a homogeneous group. Yet international environmental treaties have not made much progress in differentiating countries beyond "developed" and "developing," often in the form of long lists of countries in each category. (Now there is a third group of "economies in transition" that view themselves as part of the "developed" category but deserving special treatment.) In the context of managing global warming, it could be useful to develop some additional categories within "developing countries" to distinguish countries that should be on a faster track to implementing some climate policies in comparison with other developing countries. Currently the concept of "middle income" countries is used, but in practice it is unclear which countries fit into that category. It might be productive to extend the type of system used in Article 5 of the Montreal Protocol: to define countries according to whether they are above or below a per capita emissions threshold. Unlike the Protocol, where there is a clear dividing line and relatively few countries on the borderline, a more continuous measure might be useful than a single threshold.

In addition to increasing the involvement of developing countries, the system of international governance increasingly faces the need to involve a wider range of non-state actors from all countries. At the national level, many failures of governance can be traced to the failure of government to respond to its citizens. The consequences are misplaced priori ties and resources and unfair distributions of costs and benefits in the society. Failure to involve constituents often leads to policies that are ineffective - broad participation gives policy legitimacy, especially where the consent of constituents is crucial to effective implementation. Democratic states have embraced these notions of expanding participation.

The same arguments may apply at the international level. International governance could benefit from broader participation and influence in at least three areas:

The rapid increase in participation by non-governmental organizations (NGOs) in all three of these areas reflects these ideas at work. Some of these issues are explored be low. First, ways to improve participation and influence are considered. Later some of the many unanswered questions and problems of opening the policy process are reviewed.

In practice, participation depends on access to the process of making and implementing policy as well as information. There are several ways to increase both:

Participation does not automatically lead to influence. However, at least three devices might increase the influence of non-state actors:

In practice, the system of international environmental governance is only at the earliest stages of addressing issues of participation. Participation of non-state actors has grown dramatically, but influence remains generally low and variable.

Among the many problems and challenges that remain, some are: First, open participation easily leads to gridlock and inefficiency. Indeed, exclusion of certain groups is often "effective" in achieving particular policy out comes. Systems with no formal capacity to exclude typically do so through informal de vices such as back room deals. If open participation leads to gridlock then the use of informal devices may become more frequent; on balance that could undermine the goal of good governance that wider participation has sought to achieve. Second, open systems are vulnerable to regulatory capture. Third, there are dangers of allowing groups to influence outcomes when they don't have responsibilities as well. States enjoy the privilege of control over the international system because they also have the responsibility of implementing what they agree to and representing their citizens. Fourth, allowing NGOs or other representatives to influence international policy immediately raises the question of who they represent. Are some groups more "representative" or democratic than others? Should they be granted greater influence?

The extent to which wider participation leads to more or less environmental protection, and at what cost, depends on the particular case.

The goal here is to make the system of governance more responsive to a broad range of interests, not just those interests expressed by state representatives. If states were perfectly representative then wider non-state participation and influence would be superfluous, but states are not.

Many opportunities for wider participation and influence, but these attractive principles need practical expression. The obstacles are numerous. There are no clear design principles and no consensus that a more open system of governance would be more efficient, more effective, or even more just. More experimental programs and comparative re search are needed.

Regarding policy, international regimes might experiment with different criteria and devices for overcoming the most serious problems of wider non-state participation and influence. For example, lack of transparency about NGO activities, constituencies and responsibilities might be offset by requirements that NGOs submit reports. Whether that system could work in practice is worth an experimental effort, perhaps on a voluntary basis in con junction with the next round of reports due from the governments of developed countries in 1997.

Regarding research, the major need now is to conduct some case studies of governance at the domestic and international level and to ask at least two questions in each study:

The goal of these case studies would be to gain a sense of the criteria that affect participation and influence. For example, how do different rules of access lead to different out comes? Case studies conducted with a common method and research questions would allow comparisons that are crucial to building more general conclusions. Ultimately the goal may be a set of constitutional principles to govern non-state participation; some might be drawn from theory, but at present the weak link is systematic information about what works in practice.

The goal is to make the system of governance more responsive to a broad range of interests, not just those interests expressed by state representatives.

Expert information is essential to the long term efficacy of the climate regime. In particular, three functions must be performed:

The Climate Convention establishes subsidiary bodies to perform each function. What should be their relationship to the IPCC? For the first function it is possible to have only a thin formal relationship. In the past the IPCC's assessments have been influential through many diffuse channels. Its credible assessments have been the centerpiece of debates within and among governments, firms, and NGOs. IPCC's reviews of the natural science issues have been consistently good; its reviews of impacts research are improving; the social science assessments remain problematic. This range in quality reflects the fields of research themselves and not a systematic failure of IPCC. The IPCC process has focused attention on the nature of the global warming problem, possible consequences and policies. This can continue with periodic assessments that have little or no formal relationship to the Convention.

The other two functions demand a closer formal relationship to the Convention. Ad equate developing country participation will require such a relationship to avoid a proliferation of uncoordinated meetings. A formal link will also help keep work timely and relevant. Such a link brings the danger of institutional monoculture and rigidity, but on balance it is essential for sustaining the legitimacy of expert assessment (which so far has been high). The European acid rain regime began with monitoring and research programs initially separate from the Convention but brought within the Convention within a few years. The link helped make the science legitimate and influential. It also helped to make data and models more transparent, allowing complementary independent efforts to flourish (e. g., the RAINS model at International Institute for Applied Systems Analysis (IIASA), which played a crucial role in analyzing the critical loads concept). The expert panels in the ozone regime gave useful and timely advice because they were formally connected to the regulatory system in the Montreal Protocol. Independent expert roles for panel members sustained the credibility of the panels despite the close connection to the political elements of the Protocol.

There is a need for more high forcing scenarios. The model benchmark is 2xCO₂, but virtually every emission scenario projects actual forcing to be much higher.

For policy, a close relationship to the Convention can work well, especially for question -specific advice and technical support. That relationship will be most effective if:

In addition, the research community can make several changes that would yield research and assessments more useful to policy:

The research community must do a better job of gathering, evaluating and communicating the subjective assessments of experts to policymakers.

Finally, and this is one of the most important conclusions relevant for policy and research, the research community must do a better job of gathering, evaluating and communicating subjective assessments. Most of the major impacts of global warming are poorly known, but the policy process needs at least some information on likely outcomes as well as the range of possibilities. The IPCC has made some efforts along these lines through careful wording and assigning "stars" to statements according to confidence levels.

But much more can be done to convey the full range of subjective information that scientists are developing. Notably, more information is needed on the tails of the distribution of potential impacts, especially catastrophic impacts. Those low probability but high consequence events are crucial, but few policymakers have a good feel for what is at stake (good or bad) and the probabilities. Nor do they know how these risks compare with other hazards on the policy agenda. Some graphical representation of these issues would complement the current technique that relies primarily on negotiated text in IPCC reports.

It is probably too late in the current (1995) IPCC assessment to develop and implement such a system. But it would be constructive for some independent group to convene some lead authors from the IPCC process and develop a methodology for conducting and presenting subjective assessments. That group could apply its method to the major conclusions of their IPCC chapters. The potential benefits of developing an effective system for communicating information to policymakers are enormous. Notably, it could improve on the current system where advocacy science communicates through the media and offers alternative views, but non-experts are left only with the sense of controversy but little idea about the range of opinions and likely outcomes. The method is perhaps most useful to the natural sciences, but social sciences would also benefit from the rigor of having to assess the mode and range of their research conclusions.