Efficiency issues and the Montreal protocol on CFCs

� е �' � � � � � �� �: �� � �у � � ��уΡ �, g о ® l�'J � €� � � • � � $ � ,.. ��� � ле � s� я' $ . (D � � �� � � � �� � � о �� � � � � � � � � � `� �О � о �� � � � � � .� � � � � � � � � о � �Q й � � � х � � � � � �+ � �� � � � � � �' � � е L/� о `'' � � �_ � � .� с � � � G � �. . L - This paper is written by Peter Bohm, Professor of Economics, University of Stockholm. Most of his work on the paper was conducted while he was a Visiting Research Fellow at the World Bank in the spring of 1990. The paper reports on findings that are particularly significant in the context of the overall work program of the Environ.mental Policy and Research Division. The author is indebted to A.P. AlLdbusan, Scott Barrett, Lorenzo Brown, Herman Daly, John Dixon, Salah El Serafy, Gunnar Eskeland, Robert Hahn, Carl Gustaf LundLn, Mohan MunasLnghe, Kenneth PiddLngton, Stephen Seidel, Tom Waltz and Jeremy Warford for helpful comments on an earlier draft. Departmental Working Papers are not formal publications of the World Bank. They present preliminary and unpolished results of country analysis or research that are circulated to encourage discussion and comment; citation and the use of such a paper should take account of its provisional character. The ftudLngs, interpretations, and conclusions expressed in this paper are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organization, or to members of its Board of Executive Directors or the countries they represent. Because of the informality and to present the results of research with the least possible delay,the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. . LL - astract The Montreal Protocol, designed to regulate CFC emissions, is particularly interesting now as a checklist or blueprint for future treaties on global pollution other than CFCs. This paper takes a look at the properties of the Protocol from the point of view of global efficiency. Although it cannot be proven, of course, that the Protocol deviates from the feasible "best" observing all political constraints as well as all real or perceived transactions costs, it can be shown that the Protocol has a couple of inefficiency properties that could have been avoided through package design: The Protocol is inefficient even if signed by all nations, e.g., because allocation of pollution among signatories is inefficient, especially through the way of compensating developing countries for their adjustments in CFC use. Also, the control variables of the Protocol are imperfectly linked with the relevant target variables, CFC emissions. The Protocol implicitly punishes signatories who like to speed up the process of reducing CFC use. Moreover, the existence of nonsignatories creates loopholes in the Protocol. The paper also comments on a couple of attempts to model more rigorously some efficiency related issues relevant to the Montreal Protocol. * iii * EFFICIENCY ASPECTS OF IMPERFECT TREATIES ON GLOBAL PUBLIC BADS: LESSONS FROM THE MONTREAL PROTOCOL Table of Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 1. The Perfect Treaty . . . . . . . . . . . . . . . . . . . . . . . . 4 2. A Simple Model . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3. The Montreal Protocol . . . . . . . . . . . . . . . . . . . . . . . 6 4. The Montreal Protocol as an Optimun Treaty for a World of Identical Countries . . . . . ............ 7 5. Would the Montreal Protocol Be Efficient if Signed by All? . . . . . . . . . . . . . . . . . . . . . . . 9 6. Would Signatories Respond to Incentives to Cheat on the Protocol? . . . . . . . . . . . . . . . . . . . . . . . . 12 7. The Role of the Existence of Non-Signatories for Signatory "Cheating' Behavior . . . . . . . . . . . . . . . 13 8. Intersignatory Trade Discouraging Unilateral Reduction of CFC Use . . . . . . . . . . . . . . . . . . . . . . 15 9. Does the Montreal Protocol Preclude Reaching a Future Optimal-Feasible CFC Treaty? . . . . . . . . . . . . . 16 10.Does the Montreal Protocol Provide Efficient Compensation to Devaloping Countries? . . . . . . . . . . . . . 18 ll.Summary and Concluding Remarks . . . . . . . . . . . . . . . . . . 23 Appendix 1. The Identical-Country Model . . . . . . . . . . . . . . . 25 Appendix 2. Constant Marginal Abatement Benefits . . . . . . . . . . 27 Appendix 3. An Equilibrium Number of Signatories Below Full Cooperation . . . . . . . . . . . . . . . . . . . . . . . 28 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 1 Introduction In principle, global public bads (GPBs), such as depletion of the ozone layer or global climate change, affect everybody and can be influenced by anybody on the globe. Hence, a constitution, aimed at an efficient level of GPB reduction, would need to involve everybody in order both to heed their valuations and to control their actions. However, even if we had such a constitution or "world government" and designed it according to some definition of efficiency, it could still fail due to deviations from a fully cooperative behavior on all sides. There are two kinds of uncooperative behavior: the individual (or the national government) may cheat or he may openly break the rules of the constitution. In the absence of a "world government", we can have a treaty among national governments. As one extreme, the implications of a treaty could match that of a perfect world constitution. As another, the treaty could run into three kinds of failure in terms of uncooperative behavior: In addition to individual signatories (1) cheating or (2) openly breaking the rules, we have that parties can (3) choose not to be signatories. The third item may imply that cheating options grow with the possibility of signatories forming coali- tions with non-signatories. Furthermore, it may complicate the process for setting rules and, as a consequence, increase the cheating options. Since a treaty among national governments is the best institution available for controlling GPBs, it would be interesting to know whether there are treaty options with systematic differences in their ability to lead the world economy to a state of efficiency. Relevant information could be acquired from an analysis of existing treaties related to GPBs. More specifically, checking the efficiency problems of existing treaties would be important for two reasons. It would allow an evaluation of the treaty, revealing if possible how well it performs in its role of controlling the GPB; such an evaluation may contribute to a discussion of ways in which the treaty may be improved. In addition, investigating the properties of an existing treaty could contribute to formalizing a checklist that would be useful in drafting similar treaties for other GPBs. In this paper, we investigate from the economist's perspective -- essentially disregarding "political constraints" as well as various enforcement aspects -- whether there are any incentives to uncooperative or inefficient behavior to be found in the 1987 "Montreal Protocol on Substances that Deplete the Ozone Layer" (MP). This treaty was drafted in an attempt to reduce the releases of CFCs (chlorofluorocarbons or 'freons') and halons into the atmosphere, hence reducing the risks of depleting the ozone layer. Initially, when CFC technology seemed hard to replace in a number of important applications, especially in foam blowing and as a solvent or refrigerant, there existed a strong reason for analyzing how effective and efficient this treaty version would be in providing incentives for reducing global CFC use. The costs for CFC substitution have now gone down so much (see e g Markandya, 1990) that there may no longer be any strong incentives for uncooperative or 2 distorting behavior here.' Therefore, and as long as new applications of CFCs do not arise, for which costs of substitution would be significant, the primary reason for analyzing the MP and its efficiency aspects, would be to improve our understanding of the problems facing the design of a future treaty combatting other GPBs. Specifically, it could provide an input to a treaty on global climate change, an issue which in many ways is similar to the ozone depletion problem but likely to entail compliance costs of a much higher order of magnitude. The design of a treaty like the MP can be seen as the result of a negotiation game to which game theory could be employed. That is, all clauses of the treaty could be regarded as being parts of some equilibrium outcome of negotiations among a group of players with given objectives. This kind of analysis can be found in the literature (see Barrett, 1989; and Hoel, 1990) and we shall also make reference to parts of that literature in what follows. But as just indicated, the topic of this paper in simply one which tries to identify "counterproductive" incentives and other causes of inefficiency in the MP, regardless of why this treaty was given the design it now actually has. Had there been only one conceivable hypothesis for the actual treaty design, this choice of topic could be questioned, since it would not lead us to observe that some imperfections' actually were the result of intentional compromises during the negotiations. It seems pretty safe to argue, however, that there are far more than one realistic and sufficiently well specified version of a negotiation game leading up to the MP. One special reason for this uncertainty is the fact that external conditions -- in panticular, the perception of the damage to the ozone layer and of the size of the substitution costs -- changed during the negotiation period to an extent that also the relevant game model may have changed over the negotiation period as a whole. Moreover, at least some parties seem to have felt that time for action was running short and that being able to use as much CFCs as possible in one's own country, given the global CFC use, was no longer a top priority. In addition, the role of "domestic political considerations" -- emphasized by Hahn-Richards (1989) -- in shaping the positions of the negotiators is hard to incorporate in a negotiation model. Hence, the treaty as a whole may not be entirely the result of negotiations for which a well-specified game- theoretical framework could be profitably used. The main conclusions of the paper will probably sound familiar to those well acquainted with efficiency aspects of national environmental policy. The conclusions may be summarized by stating that, although the MP, at the time it was drafted and all things considered, may have been the "best" feasible treaty for controlling CFC emissions, there seems to be some incentives missing, to which the Protocol could have been adjusted to the (potential) benefit of all parties. More specifically, the main points are as follows: (a) The loss of CFC control due to a large number of countries not signing the MP is probably small 1 This does not rule out of course that a change in the global political climate could give rise to uncooperative behavior among countries which are now signatories to the MP, and that such a behavior could spread. 3 compared to the inefficiency of CFC control allowed by the Protocol itself. The main reasons are that most non-signatories are developing countries (DCs) which the MP anyway allows a substantial increase in CFC use in the foreseeable future, and that non-CFC- technologies developed in the signatory industrialized countries (ICs) to replace CFC-based technologies are likely to become highly competitive in the long run. (b) The MP does not take into account the differentes in net costs of CFC-use reductions among signatories; especially when ICs and DCs are compared, these differences are likely to be significant. (c) The Protocol tries to control CFC use, not CEC emissions; some emissions occur with certainty at the time when the CFCs are used, while others occur with uncertainty some 20 years or so later. Control of crude estimates of emissions, based on the frg= of CFC use (thus avoiding prohibitively high costs of administration), would not only be more relevant, but also help to control emissions from CFCs which remain in products and which, sconer or later, may prove to be cost-efficient to have safely destroyed. (d) The MP allows trade in CFC-based products among signatories in a way that discourages them from doing better than required by the MP. (e) Although CFC emissions from countries which are not signatories to the MP are relatively small, the existence of non-signatories creates a possibility for signatories to counteract the spirit of the MP by importing certain CFC-based products from the non- signatories. Sections 1 and 2 present the general benchmark-case of a "perfect" or efficient treaty controlling a global public bad and compare this case with that of having no treaty at all. Section 3 summarizes the main properties of the MP. Section 4 reviews previous work that attempts to explain why all potential signatories have not signed the MP.2 The q-Aestion of whether the treaty would have been efficient had all countries actually been signatories is addressed in Section 5. How the existence of non-signatories can distort signatory behavior is discussed in Sections 6 and 7, while Section 8 shows how unconstrained trade among signatories can stop CFC use from being reduced below the level required by the Protocol. Section 9 addresses one aspect of the question of whether the MP may actually prevent a better global treaty from being reached. The MP solution to providing compensation to DC 2 We disregard here that the MP is in a dynamic process of being redrafted and -- especially for this reason -- of having new countries joining the treaty. 4 signatories is analyzed in Section 10. (Table 1 gives an overview of the main efficiency issues to be discussed in the paper.) Conclusions are presented in Section 11. Table 1. EFFICIENCY ISSUES OF PARTIAL COOPERATION (TREATY NOT SIGNED BY ALL): The case of the Montreal protocol: Section 1. Inefficient even if all had signed? 5 Specifically: 2. Is compensation (redistribution to DCs) inefficiently designed? 10 3. Are signatories deterred from doing more than required? 8 4. Nonsignatories make loopholes effective? 6,7 5. Does a Protocol not signed by all now make future global CFC treaties less effective? 9 1. The Perfect Treaty A vreaty which is signed by all nations and which is "efficient" and "fair" will here be called a perfect treaty. For a treaty to promote a globally efficient reduction of a global public bad, it would need to observe abatement costs as well as benefits for all potential signatories in the world. An efficient treaty may then be defined as one which maximizes a weighted sum of glbal benefits net of global costs, using e.g. some definition of exchange rates as international weights and assuming the benefits of future generations to be incorporated in the benefits as estimated by the present national governments. Alternatively, if it is not possible to even approximate the benefits of reducing the global public bad, an efficient treaty would be one that minimizes the global abatement costs for any given reduction. Efficiency objectives such as the two now indicated do not necessarily mean that all countries should take abatement actions. Abatement costs in a given country may be so high that the globally optimal abatement policy would require no actions to be taken there. Another case is that national abatement costs as part of a globally efficient abatement package do not need to be paid for by the country where the abatement occurs. Strictly speaking, the efficient distribution of abatement activity and the distribution of the financing of this activity are two separate matters. Thus, a country that has low-cost options for abatement may certainly have to use them as part of the efficient abatement activity of a perfect treaty but need not -- as a consequence of the chosen definition of fairness or optimal international income distribution -- have to pay for (all of) the abatement costs. Conversely, countries that do not have enough low-cost abatement options to meet the abatement requirement for which they have a financial responsibility in a perfect treaty may still have to bear the costs of abatement and, presumably, would only welcome being able to utilize less costly abatement options in another country. 5 Although it may be required that a perfect treaty takes account of equity or fairness issues such as those now indicated, we shall reserve the term "perfect" for global tr3aties that are efficient, thus assuming them to be independent of the distribution of the financial burden among countries. 2. A Simle -odel Contributions to a global public bad are, under all circumstances, likely to differ among countries. For the specific cases of greenhouse gases and substances that deplete the ozone layer, the relevant country characteris- tics seem to be primarily population size and stage of development (or GNP per capita). Taking these characteristics into account when modelling optimal or equilibrium behavior in the global economy places a heavy burden on making such a model transparent and tractable. A model of the global economy that may provide a simple starting-point for analyzing perfect treaties as well as imperfect ones is that where all countries are assumed to be identical. Here, we shall make use of a slightly revised version of a model of this kind presented in Barrett (1989, 1990). Following Barrett, we formulate the global efficiency problem as one where total abatement benefits minus total abatement costs are maximized for all countries. (For details see Appendix 1.) Marginal abatement costs, MCL, increase with the level of abatement activity as shown in Fig. 1. Furthermore, Barrett assumes that marginal abatement benefits, when all countries increase their.abatement level concurrently, fall as indicated by MB, for country i, and by MB.., when country i acts alone. The optimal national abatement level at full international cooperation (q,) is that where marginal costs equal aggregate marginal benefits (which in this model of identical countries equal the individual country's marginal benefits of the coordinated abatement activity of all countries), MB . Optimal individual behavior in the case of no cooperation (qa) is where marginal costs equal marginal benefits of the country acting alone, MB.L As shown in the Appendix and as illustrated in the Figure for two different marginal cost curves, the effect of cooperation on abatement activity, q. - q, as well as the gains from cooperation, shown by the shaded areas in the Figure, entuallv decrease the lower the marginal costs, given the marginal benefits associated with abatement. Fig. 1 M. 6 The Montreal Protocol (UNEP 1987) is so far the only treaty that exclusively deals with a global public bad. Let us now see to what extent this treaty approaches the optimum using this simple model. 3. The Montreal Protocol The aim of the MP is to control the emissions of CFCs and halons, hereafter CFCs for short. The treaty deals with the CFC asistasA, not with the products made from these substances. In essence, the treaty states that a. CFC consumption (i.e., production plus imports minus exports of bulk chemical3) by a signatory should not exceed certain levels after certain dates; b. ditto for CFC production; c. a developing country is entitled to delay its compliance by ten years and to reach a consumption level of 0.3 kg per capita during the first ten years4 to "meet its basic domestic needs"; d. parties may transfer CFC roduc.tion reduction requirements among them5; e. imports of controlled substances from non-signatories are banned, and products containing CFCs should be listed and no later han January 1993 be banned from being imported from non- signatories; and f. no developing country may export controlled substances to a non- signatory (beginning January 1993). There is a number of additional rules in the Protocol, most of them rather vague. jose rules are likely to have been amended at the time when this is being read. However, the original rules indicated above are the ores which ou discussion will focus on. Referring to the model presented above, we may note that, in the Protocol, countries are distinguished as being developing (DCs) and in- dustrialized (ICs). Hence, countries are not treated as being identical. Moreover, countries are distinguished also in that some may choose to be signatories, others not. The fact that not all countries have signed (and ratified) the MP is of course a major reason why this treaty deviates from a perfect one. When the MP was executed in January 1989, almost all ICs, among them the most important user and producer countries, and a minor part of the S After January 1, 1993 exports to non-signatories are no longer to be subtracted. This, however, means that up to four years after the protocol took effect, January 1989 being the earliest possible date, signatories are free to export any amount to non-signatories. 4 W*rl average CFC consumption at the time the treaty was negotiated was some 0.25 kg per capita. 5 The limits to these transfers that originally existed were removed in 1990. 7 DCs had signed the MP, covering more than 80 percent of global CFC use. Barrett (1989) uses his model to explain this relative success of the MP. Before addressing the questions (a) whether or not the MP would have been a perfect treaty had all countries signed it and (b) what the existence of non- signatories means for the performance of the MP signatories, let us analyze the validity 9f this explanation. 4. The Montreal Protocol as an Optimum TreatX for a World of Identical Countries Barrett shows how it may be in the interest of a stable subset of the identical countries to cooperate, subject to the non-cooperative behavior of the nonsignatories. He points out that even if net benefits from (limited) cooperation are small, many countries are likely to cooperate and sign the treaty. In the case of the MP, part of the reason for this behavior could be that recent evaluations have indicated that marginal abatement costs are small. The fact that so many countries have signed the MP may well be explained by Barrett's argument that the non-cooperative abatement effort is not too far from the globally optimal level. In Figure 1, this is illustrated by case 2, where q. and q. are quite close (assuming the number of signatories to approach the maximum). The relevance of the model used hire may however be questioned. One prerequisite for the very small effect of cooperation shown in this model is that marginal abatement benefits are significantly reduced by increased abate- ment and eventually reach zero (see Fig. 1). However, marginal benefits are unlikely to approach zero even at a 100 percent abatement level, i.e., with no new CFC emissions. This is because there are (a) significant amounts of earlier emissions that will continue to deplete the ozone layer for a long period to come and (b) other substances affecting the ozone layer that will continuously be released since they are outside the control of the (original) MP or any similar treaty. If the marginal benefits of reduced CFC emissions are nowhere near zero, the result of a reduction of marginal abatement costs may well be that the gains from cooperation will increase. In Appendix 2, this is shown to be consistently so for marginal benefits being constant over the relevant range of abatement activity. The case of an equilibrium number of actual signatories being smaller than the number of potential signatories can arise in the revised model as well. This is illustrated in Appendix 3. However, hardly any model, in which countries are assumed to be identical, is likely to be suitable for explaining the actual number of MP signatories. While simple models assuming countries to be identical could give instructive results for specific issues, there are other issues for which actual differences among countries now and in the past play a decisive role. More specifically, observing various differences among countries would not add much insight into the two extreme cases of full cooperation and no cooperation; in characterizing these two cases and the possible existence of stable equilibria for cases of partial cooperation, models with identical countries will do (see Appendices 2 and 3). By contrast, the impact of (large) differences among countries is likely to be significant when analyzing specific outcomes between these two extremes, i.e., 8 for imperfect treaties such as the MP. At least four reasons may be highlighted: First, although the opportunity costs of using CFCs are small for most uses, relative opportunity costs may still differ a great deal among uses.' Thus, differences with respect to -volume and sectoral distribution of CFC uses may givc rise to substantial differences in abatement costs among countries. In some cases, efficient abatement of CFC use can cause considerable social or political costs in terms of labor layoffs -- e.g., when a depressed region is affected and no suitable replacement for CFC-based production can be found. Hence, real costs -- esRecially relative to the size of the economy -- may vary considerably among countries. Second, countries may differ as well, and even more, in their benefit valuations of changes in the ozone layer. In particular, the health effects of such changes differ depending on the pigment of the skin and average longevity in the country. (Such differences may be particularly significant between industrialized and developing countries.) In addition, certain parts of the world, e.g., countries close to Antarctica, seem to be more at risk of ozone depletion than others. Governments may also differ in their evaluation of the probabilities of health effects or effects on productivity in agriculture. Hence, differences in benefit estimates may be substantial. Third, an aspect related to the preceding one is that benefits from gn reductions of emissions may be significant for some (heavy-user) countries, primarily the U.S., and completely negligible for others. In the early phases of negotiations concerning the ozone depletion problem, the U.S. accounted for more than 50 percent of global CFC emissions. Differences in this respect are likely to influence the outcome of the negotiations and the willingness to sign a treaty like the MP. Finally, there are likely to be relevant differences among countries in areas that are entirely outside the simple model and relate to things other than the costs and benefits discussed so far. For historical or political reasons, some countries may choose to act as followers of actions taken by certain other countries and these countries only. As a case in point, DCs have been reluctant to follow the reductions proposed by the ICs since the CFCs which have now accumulated in the stratosphere are the results of actions almost exclusively taken by the ICs for a number of years. Therefore, an invitation to introduce a bound to CFC use in all countries may be unacceptable to those which have used almost no CFCs in the past. For the reasons now mentionad, the willingness to cooperate may vary a great deal among countries. Modelling negotiation behavior concerning global 6 At the time of the original drafting of the MP, appropriate examples would be CFC for aerosol use, where opportunity costs were around zero (with the exception for certain special uses), and automobile air conditioning, for which CFCs were clearly superior. Ia here that some signatories already had abolished the first type of CFC use by 1986, the base year for CFC reductions according to the Protocol, while others had not. 9 pollution and pollution abatement when countries differ in relative abatement costs, benefit valuations and even in their past use of CFC is obviously not an easy task. In addition, the relevance of any such attempt would be difficult to evaluate. Hence, instead of trying to model the behavior toward ozone layer protection and remain uncertain about the relevance of such a model for determining the efficiency of the actual design of the MP, we now turn to investigating the efficiency properties of the MP by identifying what incentives to non-cooperative or globally inefficient behavior the MP actually contains. Does the MP make signatories willing to keep uncontrolled CFC-based imports from non-signatories at a minimum? Does it make signatories willing to use as little as possible of their CFC quotas or does it stop them from unilaterally increasing their abatement activity? And, more generally, does it make signatories attain the resulting reduction in global CFC emissions in the most efficient way? A treaty which does all these things would obviously be a better one than one that does not. 5. Would The _P Be Efficient If Sined By All? There can be widely diverging views on what are the national, and hence, the global, benefits associated with global CFC reduction. A number of governments may argue that the MP is imperfect in the sense that its cutbacks of CFC production and consumption are based on an underestimate of these benefits. Conversely, there are likely to be governments that do not consider the benefits large enough to outweigh any noticeable sacrifice. These governments would of course find it efficient to abstain from signing the MP. Still, the question can be raised: Cold the MP have been a perfect treaty, if it had been signed by all nations?7 At least three reasons can be given why the answer is in the negative: (1) Noting the clause which allows countries to trade reduc- tions of CFC production among one another (see point d above), we can see that the MP allows a given (reduced) level of CFC output for the world as a whole and with all nations as signatories to be reached at minimum cost. However, the MP does not allow a similar arrange- ment for trade in quotas of CFC consMption.8 As is perhaps the most common lesson taught by environmental economics, regulating the size of pollution from different sources without heeding the abatement costs involved, will (except by pure chance) lead to a situation where marginal abatement costs differ among polluters. Thus, it follows that there is no guarantee 7 Here, we disregard the fact that the MP does not cover all substances, the emissions of which contribute to the depletion of the ozone layer, and that the use of substances not covered may increase as a result of certain CFC uses now being controlled. 8 Trade in CFC consumption quotas is allowed among member countries of the European Community (EC), the reason being that the EC in this respect is regarded as one party to the MP. 10 that global abatement activity would be efficient even if all nations had loined and abided by the Protocol. In fact, as we shall return to in Section 10, the MP is supporting CFC use in DCs where abatement costs are relatively small, hence making it unlikely that efficiency will be attained, regardless of the number of signatories. (2) The MP does not provide incentives for all kinds of CFC release control. CFC releases can be controlled in four ways: by limiting the use of CFCs in the production of new products, by recovering and reusing CFCs from products under repair or servicing, by recovering and reusing CFCs in products when scrapped, and by safe destruction of used CFCs. The major part of CFC releases occurs during manufacturing or when repairing/servicing CFC-based products, and this may continue to be the case for some time. Such releases are controlled by the MP through the its limits on CFC use. Recovery for reuse of CFCs is encouraged by the mere existence of binding constraints imposed on CFC use; this creates -- if not a formal price in the form of a charge or a permit price -- a shadow price on CFCs as long as the signatory has not chosen to use only such regulatory instruments (design standards) which leave no room for incentive effects on CFC reuse. Recovery for destruction would be encouraged in an efficient manner if countries chose to subject CFC destruction to a subsidy equal to the CFC charge (or shadow price). The composition of the four sources to CFC releases would be efficient with such a pricing policy. 'When charges or permits are levied on the use of new CFCs, this policy can be seen to amount to a deposit- refund system with a deposit (charge/permit price) on new CFCs and a refund on all CFCs recovered for reuse (automatically via the absence of a charge/permit price for muse) and destruction (the subsidy).' Destruction of CFCs in scrapped cooling equipment and closed-cell foams does not seem to have been an attractive proposition so far, although pilot plant testing of various destruction methods have met with some success (UNEP, 1989). This may change over time when CFC use is further reduced, increasing the marginal costs of CFC use reductions, hence increasing the optimal subsidy rate for destroyed CFCs. More important, perhaps, is that BW- and profitable methods for safe destruc'ion (e.g., of halon products, or CFCs in cooling units in areas with a high density of such equipment) are not likely to come forth, rapidly or at all, without a subsidy on this kind of CFC release control. The MP, however, provides no incentive for signatories, who use but do not produce CFCs, to submit destruction technologies for used CFCs for approval and to use such technologies to attain an efficient level of CFC 9 Cf Bohm, 1981 and 1988. 11 release control.10 By contrast, deducting CFCs destroyed - using technolog- ies approved by the signatories -- is accepted for determining net CFC production (see MP, Article 1.5). If signatories were given the right to deduct from their consumption of new CFCs those CFCs which had been safely recovered from old equipment by methods and monitoring as accepted by the MP signatories -- thus defining a value of net CFC consumption close to the real target variable, CFC emissions -- this could have led to a significant change in the composition of those emissions in two ways: First, incentives would then have been provided for CFCs used in products made now. and those remaining in the products until scrapped, to be recovered for safe destruction; this could also influence the design of such products, facilitating future CFC recovery and destruction. Second, the significant amount of CFCs in products existing at the time the MP took effect would also have been subject to an incentive for safe CFC destruction when scrapped.n In sum, CFC emissions from old sources are only partly covered by the MP, leading to an inefficient composition of measures to reduce CFC emissions in the long run. Providing incentives for efficient reductions of CFC emissions from old sources would allow a higher use of new CFCs without exceeding the limits of total CFC emissions implied by the MP or allow total CFC emissions to be reduced more than has been possible so far. (3) For a treaty to be perfect it must take account of the fact that different activities involving controlled substances may have different potentials for polluting the environment. The levels of production and consumption allowed for in the MP are,calculated based on the varying Ozone Depletion Potentials (ODPs) of the different controlled substances. However, the MP does not observe that the dates at which CFCs are released into the atmosphere differ among CFC uses. In some cases, the CFCs are released at the time they are used as inputs in commodity production. In other cases (e.g., in foams and as refrigerants), a substantial' amount of CFCs remains enclosed in the products for many years, say, up to 20 years. Obviously, the environmental costs of releases today differ from those 20 years later. Aside from ordinary time preferences, one reason may be that we may become more efficient in recovering CFCs from existing products 20 years from now; another is that over time efficient ways may be found to 10 This can be contrasted with the incentive effect that the MP has had on R&D of non-CFC technology. This effect, which has been in force at least since the time the MP was drafted, has most likely contributed a great deal to the reduction in costs of such technology. 11 We may add here that, when the MP is amended so that CFCs are to be phased out at some date, CFC releases from old equipment would continue after that date at least as long as there are no incentives for reducing these releases through reuse or destruction. 12 reduce the detrimental effects of increased UV radiation. This could very well warrant significant differences in value among CFC uses, even larger than some of the ODPs which are observed in the Protocol. Since the MP refers to CFC production dates and use dates, we have that the Protocol is neutral in dealing with the expected release dAtes for CFCs used for manufacturing and consequently, it does not discourage CFC uses that involve early CFC releases. In addition, as discussed in Section 8, the MP does in fact discourage signatories from Xeducing CFC use below the level required in the Protocol. Thus, we may conclude, that even if the Protocol had been signed by all nations, it would not operate in a way that could be called efficient. We now turn to the question whether or not the existence of non-sig- natories increases the malfunctions of the Protocol, over and above those now observed. More specifically, does the Protocol in the presence of non- signatories make signatories behave less favorably than if all nations had signed? 6. Would Sinatories Respond to Incentives to Cheat on the Protocol? Signatory CFC related behavior is of course likely to differ a great deal from that of non-signatories. In a formal sense, a non-signatory country can do whatever its government wants to do, although in actual practice its behavior may be influenced by considerations to keep the option open to join later or to avoid running the risk of being economically hurt by "retaliatory" trade policies by signatory countries. Thus, non-signatory behavior may to some extent be -- from an ozone-protection point of view -- favorably influenced by signatory behavior. Signatory behavior, as we shall see in the subsequent section, can be unfavorably influenced by the existence of non- signatories, given the loopholes that the MP has created for signatories to "cheat" on the Protocol. Except for those countries that keep their CFC use strictly bel the accepted maximum level (likely to be true at least for some DCs), signatories are subjected to free-rider incentives to cheat. Why would countries, choosing to join the MP, be willing to respond to such incentives? At least three reasons can be given: (1) A signatory who would willingly abide by a stringent MP, only if all countries had cooperated and signed, may no longer be willing to do so when this proviso is not fulfilled and esoecially when countries in a similar position (say, cner NICs) have not signed the Protocol. (2) Signatory behavior may differ according to the reason why the country has chosen to be a signatory. In Barrett (op cit) and Hoel (1990), a country is assumed to join the treaty if abatement costs fall short of estimated environmental benefits. But in addition, a reason for joining may be that countries simply want to 13 be part of a cooperative movement in order to benefit from its side-effects (e.g., avoiding risks of losing partners for other forms of international cooperation). The side-effects of joining the MP could conceivably be greater than the net abatement costs, i e the excess of abatement costs over environmental benefits. (3) Another case is that where a country has chosen to join the treaty, but all along has been endorsing its own version of an optimal treaty -- one which allows an abatement level well below that of the MP. Such a signatory may be tempted to cheat, owing to pressing domestic problems and if the ways to cheat are subtle and palatable. We shall not deal here with cheating in the form of false reporting of CFC use rather focus on the "loopholes" for circumventing the spirit of the MP. This can happen through CFC-related trade between non-signatories and signatories which the MP has failed to consider. 7. The Role of the Existence of Non-Signatories for Signatory "Cheating" Behavior? Being able to trade with non-signatories opens up some possibilities for signatories to circumvent the spirit of the MP. To illustrate, let us take a version of case 3 just mentioned. This is the case of a country that has accepted having its CFC-based production activity constrained by the MP and therefore has joined the treaty, but finds the MP requirements "too demanding". Say, the government of this country would have liked the MP to be less stringent towards countries of the type this signatory belongs to, e.g., the DCs, and thus would like to do less of a reduction of net CFC use if it somehow could "get away with it". According to the MP, a signatory can, at least for the time being, import CFC-based products from non-signatories. Here, we focus on such products which are produced with, but do not contain, any CFCs.12 The MP states (Article 4, point 4) that signatories shall, no later than 1993, determine "the feasibility of banning or restricting the imports" of such products from non-signatories. However, the Protocol does not state when actions would be taken, should banning or restricting imports be found feasible. Thus, for a significant period of time, the MP allows signatories to meet the MP requirements by reducing CFC use in this type of production and instead import the products from a non-signatory.1 12 The monitoring problem is much smaller and the period for unregulated imports is much shorter for products containing CFCs (see MP, article 4, point 3). Therefore, we abstract from this case here. 13 Production activities possibly relevant for this kind of trade are cleaning or sterilisation of electronic products and the manufacturing of certain foam products. 14 Over a longer period of time, this trade activity could also encourage producers to leave the signatory home country and establish production in a non-signatory country, then from there export the products back to the signatory home country or other signatories. The kind of trade activities now mentioned can take place without the government of the signatory country knowing it or doing anything about it. Or, the government could be actively supporting it. One reason for doing so may be that influential domestic producers and/or employment in the country are dependent on the supply of the CFC-based products and that the government is hard pressed to accommodate such interests. Could this loophole have been eliminated? There is no doubt that policy actions of the type referred to in the MP, Article 4.4, are feasible, although all parties may not agree to implement them now. One example is to have potentially CFC-based products be treated as actually CFC-based products as long as signatory importers cannot present proof of the contrary as obtained from their trading partners. (Given that the MP bases compliance checks in general on self-reporting, enforcement techniques and enforcement costs would not be a major problem here.) Such a rule should probably have been introduced in the "package" of the original MP, since later on it may not be feasible to get sufficient agreement on a separate point like this one. (This, of course, does not necessarily mean that the original number of signatories would remain the same). Another way, similar to that now discussed, to circumvent the MP is for signatories to replace CFC use in the manufacturing of certain products -- which are to be exported to non-signatories -- by having the products, or relevant parts in them, manufactured in the importing countries.14 Also this loophole could have been reduced by trade restrictions, e.g., banning signatories from exporting products designed for CFC use to nonsignatories. 14 OTA (1988) gives an interesting example, where air-conditioned cars exported to other countries could have the air-conditioners charged in the importing country. Another example is when ships from signatory countries have servicing operations, involving CFC (and halon) use, moved to harbors in non- signatory countries. 15 It should be noted that this problem need not vanish completely, if all countries were signatories. With signatories having excess CFC use capacity -- especially DCs since they were given extra capacity by the MP -- it would not. Excess CFC-use capacity would provide an incentive for a country to start CFC- based production for exports. (In fact, this is still a possibility even after the explicit proposal by some signatories to let DCs expand their CFC use also for exports was rejected.) Although the MP formulation, saying that DCs may delay compliance by ten years "to meet its basic domestic needs" (Article 5, para. 1), now has been clarified to mean "not for exports", this can hardly guarantee that additional CFC use is not directly or indirectly linked to increased exports. -- What has been said now also forms a basis for arguing against providing compensation in the form of extra room for future use of CFCs to countries that 15 To sum up the main argument here, we have seen that the existence of non-signatories presents some trade opportunities involving CFC use that signatory countries may want to take advantage of, and by doing so obstruct the spirit of the MP. We will now show that the opportunities for trade between signatories allowed by the MP also present obstacles for an efficient interrational abatement policy. 8. Intersignatory Trade Discouraying Unilateral Reduction of CFC Use An efficient treaty (disregarding enforcement costs) would see to it that signatories are not prohibited from undertaking CFC abatement, the full costs of which they are willing to bear. As a case in point, Sweden has decided to do better than the MP requires it to do by speeding up its reduction of CFC use and to phase out CFCs by 1995; to this end, it is instituting bans and similar regulatory instruments to stop Swedish firms from using CFC-based technology. Firms using CFCs, e.g., as a blowing agent for making foams, are required to shift to non-CFC-based techniques, of which only more expensive or less attractive substitutes are available. Since the CFC- based products now c uld be imported from other signatory countries, albeit at a higher price than what previously charged by the Swedish producers, but still below the competitive price of the new substitutes, firms as well as environmental policy administrators complained. Not allowing a signatory in this situation to control imports from other signatories thus threatens to eliminate the new substitute products that the signatory is trying to encourage in the spirit of the MP. Not only does this outcome deter others from following the example set for substituting certain CFC-based products, which must be the main purpose of the measure taken by the signatory, it may also discourage signatories from taking similar steps for other CFC-based products. It should be noted here that this is a problem which obviously would not disappear even if all countries were signatories. Thus, a treaty being complete in this sense would not have helped. But could the MP have been designed so as to avoid the counteracting effect on initiatives for promoting the replacement of CFC-&ased technology by substitutes? If it could, the treaty design would not have stopped countries from exceeding the ambitions of the MP and possibly from playing a lead role in increasing the speed by which CFC use is cut back, assuming that the aggregate reductions advocated by the MP fall short of a globally efficient level. Allowing the use of an import ban would promote efficiency in this case and remain in line with the GATT rules which recognize the rights of countries to use trade barriers to protect its environment as long as this does not favor similar goods produced domestically (cf. OECD, 1989). The MP could therefore have explicitly stated the rights of signatories to introduce trade barriers to protect their attempts to reduce CFC, as long as such trade restrictions are compatible with existing treaties. In addition, given the "vague and cursory wording" (OECD, 1989) of the relevant GATT Article XX, the interpretation of this rule may so far have not used much CFCs (see Section 10). 16 have turned out more favorable for the objective of the MP, had the Protocol provided this backing of trade control. 9. Does the MP Preclude Reaching a Future Optimal-Feasible CFC Treaty? A potential problem with the MP is that the early commitments of the signatories may have put them in a worse position when bargaining with the non-signatories to achieve a truly global CFC treaty. As shown by Hoel (1990) for a two-country model, where the "unselfish" country 1 has taken unilateral actions to reduce its emissions, it may no longer be possible to reach, in a second step, the cooperative solution with the "selfish" country 2 that would have been attainable in the absence of the unilateral action and that would have offered a lower level of total emissions. As a simple illustration of this outcome, let group S represent the group of early MP signatory countries (essentially ICs) and groug NS represent the non-signatory countries (mainly DCs).16 Let us say that, in 1986, before the MP was drafted, group S accounted for 90 emission units and group NS for 10 units. Furthermore, let us assume that the predicted levels for 1995, are 50 units for each of the two groups, whereas the emission levels in the absence of the MP would have been 100 units and 50 units, respectively (see Table 2). Now, assume that, for year 2000 and on, group NS can offer at most a 50 percent cut in its emissions (i.e., from 50 units to 25) in exchange for a reduction by group S from 50 to 5 units (i.e., by 90 percent from the 1995 level). If no unilateral action had been taken, i.e., in the absence of the 1987 MP, group NS would have accepted to undertake, we assume, an 80 percent reduction (from 50 to 10 units) in exchange for a 95 percent reduction by group S (from 100 units to 5). Thus, with the 1987 MP, the eventual global emission level would be 30 units in contrast to only 15 units for the case without the MP. Table 2: EMission levels (in "units") for two groups of countries, the original MP signatories (S) and the non-signatories (NS). for two cases. (I) with and (II) without the 1987 MP Case I II Group S NS S NS 1986 90 10 90 10 Predicted for 1995 50 50 100 50 Global treaty for 2000 and on 5 25 5 10 Global emission level 30 15 16 Hoel acknowledges that his analysis also covers the case where a "country" is 'a group of countries which have coordinated their policies". 17 While there is a risk that the outcome of the MP is negative as illustrated here, the relevant question is what role, if any, should be given to this risk when drafting treaty proposals like the MP. For example, should the requirement of a minimum number of signatories for the MP to take effect have been increased, say, to cover all "potentially important" countries? Hoel's conclusion is that a country should give "careful consideration" to this risk before it "departs from following its own self interest". Putting this advice into practice is nonetheless hampered by the presence of a number of other possible implications of "unilateral" actions, most of which are cited by Hoel but, for obvious reasons, left out from his bargaining model: (a) The action would set an example, influencing the behavior of the rest of the world in the direction of increasing its willingness to take similar actions. This may be particularly relevant for our application of the Hoel model to the MP case, given the significance of the "founding fathers" of the MP in terms of CFC production and consumption. (Behavior is, of course, likely to be influenced also by considerations working in the opposite direction, implying that if "country 1" reduces its emissions there is less need for *country 2" to do so, assuming marginal benefits of reduced emissions are decreasing; this effect, however, is already incorporated in Hoel's model.) (b) The fact that a country or a group of countries documents a willingness to make sacrifices is likely to improve the climate for international negotiations and the chances for reaching a cooperative solution later on. This would be particularly important where all those countries which have been significantly contributing to the ozone depletion problem over the years are signatories to the MP and those which have not are non-signatories (see also Bohm, 1982). (c) Taking action makes it clear to others that the actor estimates the costs of reductions in CFC use to be in some sense surmountable. Furthermore, the practical implications in this respect would become more conspicuous as time passes. (Points (a) to (c) are observed in Hoel, 1990). (d) "Country 2", or the non-signatories of the MP, may eventually become aware of the retaliatory power built into the Protocol (concerning CFC-related trade and certain forms of technology transfer to non-signatories, see MP, article 4) and other unfavorable trade measures that signatory countries may take. (e) Most important perhaps is that, since CFC emissions have accumulative environmental effects, lasting far into the 18 (e) Most important perhaps is that, since CFC emissions have accumulative environmental effects, lasting far into the future, an early start in reducing CFC emissions is valuable in itself. Even if there are assurances of reaching a treaty better than the MP, the value of this time gain from unilateral action, may still outweigh its drawbacks. The aspects now mentioned are not only of potential importance in the case of the MP, but rather likely to have been among the driving forces behind the drafting of the MP. Even more relevant for our discussion is that these aspects must be considered when evaluating the efficiency of the Protocol. In that perspective, it is hardly seems justified to criticize the MP for making an eventual, truly global, treaty less efficient. 10. Does the MP Provide Efficient Compensation to DCs? Forming a treaty to control CFC emissions raises, as mentioned earlier, a distribution problem regarding the size of the sacrifices to impose upon the different signatory countries. The MP deals with this problem in three ways: (1) through the choice of basic control measures, in particular the pace at which CFC use and production are to be reduced, implying a specific distribution of sacrifices among the countries to which these basic measures apply (MP, Article 2); (2) by singling out a set of countries, viz. DCs as defined by the Protocol, for a less demanding set of the control measures, essentially a ten-year delay of compliance, during which period annual CFC use may rise up to 0.3 kg per capita (average DC use being at most some 0.05 kg per capita in the mid-80s, see Markandya, 1990); (3) by calling upon the Parties to the Protocol to facilitate bilaterally or multilaterally the provision of aid to signatory DCs for a transfer to technologies alternative to those based on CFCs (MP, Article 5). Here, we shall discuss the form of compensation mentioned under point (2), given the choice of basic measures under point (1) and given the group of countries to which the less demanding set of rules should apply. But first, it should be noted that the extent of the concessions mentioned under point (2) is quite significant in terms of the additional CFC use permitted. Thus, even if all DCs were to sign the MP, their CFC use could increase substantially such that global CFC use by 1998 could be much larger than that of 1986 -- the year on which the MP bases its reductions of CFC use for ICs 19 approximately that of the whole world for 1986. Second, it should be noted that the reductions for DC signatories after 1998 are to be based on their average annual CFC use for the 1995-97 period. In this way, the MP provides incentives to DC signatories, who like to keep their CFC use options on a high level after 1998, to exceed their otherwise optimal CFC use level for 1995-97, wherever this level is not effectively bound by the 0.3 kg per capita limit. An alternative to the MP would have been to let post-1998 CFC use be kgsd on (but possibly being many times larger than) some historical CFC use figure, say, the CFC use level in 1986 (as is the case for IC signatories). However, in the discussion below, we assume for simplicity that, without the special treatment described by point (2), DC signatories would have had to follow the same rules as those for IC signatories. With the MP giving DCs a compensation in the form of the special treatment stated in point (2) above, the question to ask is whether this is in fact an efficient form of compensation. As an alternative, we shall consider offering signatory DCs a money transfer equal to the estimated value of the delayed compliance to the Protocol. (We abstract here from effects on CFC- related trade with non-signatories.) Assume, to begin with, that an accurate prediction could be made of the CFC use in individual DCs as well as of the cost of replacing CFC by non-CFC technology.1a Given that the objective is to improve the welfare of the signatory DCs from what they would have attained if they were not given the "favor" of delayed compliance, we can note that a money transfer equal to the cost difference would represent an gM=compensation. The reason is that the recipients of the transfer are likely to prefer using part of the transfer for other things than the end products (primarily refrigeration) for which the CFCs would have been used. Given that this overcompensation amount is easier to compute than the exact compensation and that this amount is likely to be the only feasible version of a money transfer of the type now suggested, the question is, would the (signatory) ICs prefer to pay that amount in exchange for an increase in their CFC use equal to the reduced CFC use of the signatory DCs under this scenario? 19 The answer is yes, if the IC willingness to pay for the extra amount of CFC use exceeds the DC's extra costs for giving up that amount of CFC use (i.e., if IC costs of giving tp that amount of CFC use is higher than that of the DCs). The efficiency problem now presented is illustrated in Fig. 2 a-b. Here, we have the total global signatory CFC use, say for the period 1989 to 1998, shown by q*2 in the absence of the MP (Fig. 2a) and by q when the MP is in force (Fig. 2b). In the latter case, CFC use for the ICs are limited to qlc*, while assuming that the room for DC use of CFCs provided by the Protocol 1o For a recent elaborate calculation of this kind, covering the period up to 2008, see Markandya, 1990. 1 We assume that the set of DC signatories will be at least the same as that of the present MP after a change in the form of compensation. 20 is still large enough for them to use CFCs up to the point, q l * qz where the marginal benefits of such use is zero. From this it follows that there is at least a volume q*" - qzc* that should be transferred from the DCs to the ICs to obtain an efficient distribution of the given global CFC use volume, qe. At least part of this efficiency gain would have been achieved by identifying the uniform CFC use reduction path of the alternative version of the MP that would have approached Lhe efficient CFC use distribution, qop and q - qopt. Hence, the change of compensation system considered here would be called for. Fig. 2A NoMP MvTIC... lVlWrPDC -+ qNMP IqC Marginal I willingness to *I prCFC-use 4m I IO I II IA Fig. 2b. MP at work IB ~4 q q lc %~4 21 It should be noted that in addition to the reason already pointed out, there is at least one other that contributes to making marginal use (net) benefits relatively much larger for ICs than for DCs over the relevant CFC use range. Before the implementation of the MP, ICs already had in place a production capacity adjusted to CFC technology, part of which would become unused as a consequence of joining the MP. DCs, in contrast, will have to increase their capacity of production using CFCs to obtain their optimal CFC use level under the MP. On the other hand, costs of shifting to the new non- CFC technology would probably not differ very much between ICs and DCs, at least not to the extent that would outweigh the short-term capital cost difference now discussed. Returning to one of the crucial assumptions made earlier, it is of course not possible to estimate the exact CFC use in DCs for long periods like 10-15 years or more. It would be even more difficult to estimate cost differences for CFC and non-CFC technologies over such long periods (see, however, Markandya, op. cit.). Thus, had the parties that drafted the MP in the mid-1980s tried to do this, the result would have been much less accurate than now assumed. However, this uncertainty also affects the MP as it came to be designed in the sense that the implications of Article 5 (see points (2) and (3), p. 18) for DC CFC use, and hence for global CFC use, could not be known. Since the uncertain outcome of the MP in this respect proved to be acceptable to the Parties, the uncertain adequacy of a money transfer mecharism of the type discussed here might also have been accepted. This is especially so, since the alternative arrangement would have the important effect of making the treaty effect on global CFC use from signatories known with certainty (disregarding the problem of inaccurate international statistics in this field, see Office of Technology Assessment, 1988). The compensation approach discussed here involves estimating the costs saved for ICs being able to raise its CFC use to an extent equal to the extra DC CFC use arising from point (2). These cost savings should have been relatively easy to estimate with reasonable precision, granted that the proposed MP was already available and was confronted with the alternative discussed here. Given then an approximate estimate of the willingness to pay from individual ICs as their minimum contributions (aside from those envisaged under point (3)) and given the advantage of knowing the "exact" CFC use to be expected from DC signatories, amending the MP in the way discussed here can be expected to have been politically acceptable. Two further aspects need to be observed. One is that the relevance of the estimates of what individual ICs should be ready to provide as a minimua to an international fund for transfers to DC signatories, would be jeopardised if the ICs withdraw part or all of that money from ordinary aid to DCs. While this is possible, similar incentives for withdrawal of aid could be expected to arise from the relatively tougher standards that the present MP imposed on its IC signatories and the preferential treatment accorded to DCs. The other is a question of how the fund should be allocated among the DCs. This distribution problem, however, will also be encountered under the present version of the MP. As noted in point (3) above, the MP parties are to 22 decide how much aid should be provided to signatory DCs in order to assist them in shifting to non-CFC technologies. Thus, while the distribution problem may be less significant and arise at a later stage, it is an issue that also the MP must address. The difficulties now suggested would be significantly reduced, if not eliminated, by letting the CFC use quotas allowed by the MP be tradeable. The DC quotas should be based on some estimate of their expected CFC use under the present version of the MP and not on the maximum of 0.3 kg per capita. To illustrate, say that there only two countries, IC and DC, who the MP allows qlc* and qe - qlc, respectively. IC may now buy q*P - q1C. from DC at a total price, at least equal to the minimum compensation required by DC. In the real world, with a large number of traders/signatories (and, if necessary, some protection against imperfect competition), a market price equal to p in Fig. 2b is established, determining the actual compensation to DCs as equal to area B in the Figure. (Area A is the gain of the ICs.) Here, we accomplish both the certainty of the volume of total CFC use for signatories and an efficient allocation of this volume. Two other possible implications of this approach should be mentioned. First, it is possible that DCs which have remained non-signatories may have been attracted by this version of the treaty, if they perceive that it offers them overcomensati (as defined) in convertible-currency. Second, for signatories who under these new circumstances are willing to make the same sacrifices as those they accepted when they joined the MP, further reductions of total CFC use (below qe) could be accomplished. To sum up, we have shown that ICs, as far as economic principles go, would be willing to pay an amount sufficient to compensate DCs for lowering their CFC use from the level permitted by the MP; this is in addition to the aid to facilitate the replacement of CFC-based technologies required by the MP. As a "by-product", better information would be available concerning CFC emissions from signatory DCs. Thus, from the point of view of efficiency, it is hardly possible to find the MP solution to the distribution problem to be efficient. 11. Summary and Concluding Remarks While the Montreal Protocol is admired by many (including the present author) as a remarkable achievement of international negotiation for controlling global pollution, it has come under some criticism. Thus, Barrett (1989) argued that the MP was made possible mainly because its effects on the environment and the sacrifices required to implement its rules were so small. Hoel (1990) has developed a model showing the risks that unilateral actions by a country or a group of countries, which may be interpreted as the MP signatories, can worsen the prospects for a true global treaty. The relevance of these two results has 23 been questioned here, however. The first result is questioned because the benefits are unli#ely to have been represented in an appropriate fashion in Barrett's model and also because the uniform-country model does not permit the proper evaluation of the HP, considering that there are major differences among countries as far as CFC use is concerned. The relevance of the second result is questioned because the negative effects implied by Hoel's model are likely to be outweighed by the positive effects of a "unilateral" action that * can come about through the HP. Here, the efficiency of the MP has been questioned on a number of * points, most of them summarized in Table 3. In particular, we have seen that the MP discourages signatories from reducing their CFC use below the maximum level permitted by the MP and from reducing their = CFC use by bringing about efficient methods for destruction of CFCs in scrapped products. Table 3. SummaX. of Main Conclusions The Montreal Protocol: 1. Inefficient even if all had signed? Yes, allocation of CFC use among (IC) signatories is inefficient (5.1) and - control variables (use and production) are not firmly linked to target variables (emissions) (5.2 - 3) 2. Is compensation (redistribution to DCs) inefficiently designed? Yes, future CFC use quotas are based on data the DC signatory can control and * DC compensation could have been made larger and/or IC sacrifices smaller by allowing quota transfers or quota trade 3. Are signatories deterred from doing more than required? Yes, signatory trade counteracting unilateral attempts to further reduce CFC use is not allowed to be regulated 4. Non-signatories make loopholes effective? Yes, non-signatories can increase the output of products, made with but not containing CFCs, for trade with signatories, hence allowing signatories to use their CFC quotas for other products 5. Does a Protocol not signed by all now make future global CFC treaties less effective? No, hardly to an extent that would make the Montreal. Protocol not vorthwbile We have seen that there are both short*run and long-run incentives to avoid doing better and even to do worse (thU trade) than that Intended by the MP. In the short run, signatories are likely to refrain from being more ambitious than strictly. required by the HP due to competition from CFC-based production in other signatory countries; moreover, they are induced to buy CFC*based products from non-signatories and from signatory DCs with CFC use 24 quotas in excess of their own requirements. In addition, incentives to develop methods, and invest in plants, for CFC destruction have been kept at a suboptimal level, while at the sams time incentives have been present for CFC- based firms in signatory ICs to move to non-signatories and signatory DCs. Moreover, we have pointed out that by basing permitted DC use of CFCs after 1998 on the level of use during a future period (1995-97), instead of a historical period, the P could induce DC signatories to. increase their CFC Use. The preferential treatment given to DCs which is allowed for in the MP in tackling the distribution problem, has been found to be inefficient. It is inefficient because the resulting global CFC emissions could have been obtained at smaller costs to the ICs or with larger compensations to the DCs or both, by transferring CFC emissions from DCs to ICs so that the marginal willingness to pay for CFC use would tend to be equalized between the two groups. Alternatively, for the same level of national sacrifices that the MP imposes on the Parties, a lower global CFC emission volume could have been attained. Compensating the DCs by money transfers equal to the value of CFCs for redistribution to the ICs would also permit that the maximum future CFC use by signatories be known with some certainty; this is not possible under the concessions currently used in the MP. This last observation suggests that the 1P does not attempt to ainismie the costs of global CFC use reductions in a way that would have been formally possible, e.g., with a system of globally tradeable CFC use permits. There may have been strong political reasons for not doing so, but it is far from clear that the Implicit costs of these reasons have been fully taken into account. It may. be argued, of course, that since the MP is still in the process of being redrafted, obvious Inefficiencies can be corrected later. However, it must be .borne in mind that, given the initial design of the treaty package, some piecomeaadjustments for Improving the package may no longer be feasible. FiLing the starting point for future negotiations by choosing a particular imperfect treaty design creates a set of vested interests that may preclude such adjustments from being feasible. Therefore, it remains Important to have the efficiency properties of alternative initial treaty designs carefully investigated before one is selected. 25 APPENDIX 1 The Identical-Country Model Barrett (1989) formulates the global efficiency problem as one where total abatement benefits minus total abatement costs, Max N (BL(Q) - CL (qj)) (1) are maximized for the given number of countries, N. Here, benefits of the in- dividual country, B,, are a function of total abatement Q - Nqi , whereas abatement costs, C., are a function of the country's abatement level, q,. Using for simplicity quadratic cost and benefit functions, Bi (Q) b ( aQ - Q2/2N) and C,(qi) = cq.2/2 (- cQ2/2N2 ), respectively, we maximize (1) with respect to Q for (qj - Q/N), producing the familiar first-order condition of public goods (here abatement of global pollution) saying that the marginal global benefits should equal marginal national abatement costs: R MBi - Nb (a - Q/N) - cQ/N - MCj (2) Thus, under full cooperation (co), each country should be required to set its level of abatement at qO - Q/N - a/(1 + c/bN). This blueprint for a perfect treaty can be contrasted to the treaty-free Cournot equilibrium, where every country maximizes its abatement level given that of all other countries. Max (qi) BL (qL + Rjnot.s qj) - CL (qL) z3 which yields the noncooperative abatement level, qn, where each country's marginal benefits of its own abatement activity equal its marginal abatement costs, i.e. MBj - b(a - qL) - Cqi - MC. That is, qu - a / (1 + c/b), which is clearly less than q. for N > 1. The difference in abatement level between the perfect-treaty and no-treaty cases is 4 - q, - aN (1/(N + c/b) - l/(N + Nc/b)). We may note right away that this difference is small for c/b close to zero or to infinity, i e for relatively insignificant marginal abatement costs or relatively insignificant abatement benefits.20 In Fig. Al, where MB., are the marginal benefits of unilateral abatement adjustments (i.e. when all qj are given, equal to qa) and where MB,, are the marginal benefits when all countries adjust under full cooperation (i.e. q, - qj), we show q. and q. for three levels of c, given b (see cases (1), (2) and (3)). (Each noncooperative equilibrium is shown by the intersection of MBi and MCL; the cooperative equilibrium is shown by the intersection of MB.1. and MCI.) Here, we can see that country net benefits from cooperation are small when c is small or large relative to b. (Compare the shaded areas with the striped one, illustrating the case of middle range c/b values, given a and N). 20 Note that changes in c and b imply gproortional changes in marginal costs and benefits. � . � л .� ' � >:.�r :� .: ti:,�ti.:>q,л5{... ,,.:,�4ч: �� � '� л а � 8 � :� �� �у � � � �» � N : � а �. � � . р �. .. э � в. _а .� V � � � �ti}i V � �/� iJ 1 .д о .. .� 27 APPENDIX 2 Constant Marginal Abatement BenefIts Here, we treat total abatement benefits for the individual country as equal to bQ. Hence, in a world of identical countries, global benefits equal NbQ. With costs as in Appendix 1, we maximize N ( B (Q) - C (q)) - NbQ - Q c/2N with respect to Q to get the optimal abatement activity under full cooperation: qa. - Nble For the case of no cooperation, we maximize 2 j, b(Q-i + qj) - qj cl with respect to qL1 with optimal abatement activity without cooperation given by %, -b/c. From this we have that the abatement effect of co2geration, q.. - q. - b (N - 1) / c, grows consistently larger the smaller the marginal costs or the cost parameter c. (See Fig. A2 for N - 3 and c(2) > c(l)). This is contrary to the case with decreasing marginal benefits and perhaps intuitively more appealing. (We can easily see here that no additional information of relevance for the issues now discussed would be obtained by assuming that the countries were different, say, with N - 3 and b, > b2 > b3 and C1 < C2 < CO Fig. A2 Fig. A3 MO)SON Me.^ me me 3balMS in -It-A .3b 2b 1 2) IV b' b q lp q qft qM q. qw 28 APPENDIX 3 An Equilibrium Number of Signatories Below Full Cooperation A simple illustration of the case where we have an equilibrium number of signatories No, where 1 < Ns < N, is shown here. Assume N - 3. Benefits (B) for each country is as in Appendix 2 B - bQ Marginal costs are a*(q) + c*(q)q - max (a, + cl., a2 + c2q; a, > a2, c1 < C2), see Fig A3 for an example. (1) Full cooperation implies, as before, marginal global benefits equal to marginal costs and hence q%o - (3b - a*(q.))/c*(qe,) (2) No cooperation is, as before, given by marginal costs equal to marginal benefits of the country moving alone, i.e. qn - (b - a*(q))/c*(q,) (3) If two countries departing from qi(l) - qa, i - 1,2, would start to cooerate with no. 3 not cooperating, thus q3 remaining at q., q1 would shift to qi(2) - (2b - a*(qi(2)))/c*(qi(2)) See Fig A3 for an illustration. For qL(2) as in this illustration, this shift is worthwhile for nos. 1 and 2 (and even better for no. 3, the non-cooperating country). (4) If no, 3 wants to join the two others, it would first have to take the step up to the abatement level nos. 1 and 2 have agreed on, qL(2). This implies a net loss (I) for no. 3, of course, since the new level deviates from the optimum level, q; see the shaded area (I) in the Figure. As the next step all three move to qe., the optimum abatement level when all countries cooperate, which implies a gain (II) for each country; see the shaded area (II) in the Figure. Hence, it pays country no. 3 to join if the net effect of the two steps is positive. If the kink in the marginal cost curve is outside the q. - q. interval the net effect is zero. If it is inside this interval, it follows from simple geometry that the loss in step I exceeds the gain in step II. The Figure illustrates such a case. We can use this model to show how to analyze the issue, when it pays a country to defect from a treaty. Assume to begin with full cooperation among the three countries. Defection (a move from 9. to qn) pays, say, for country no. 3 (a) if the remaining signatories adjust to their now optimal abatement level (q(2)), exactly for the reasons stated under point (4) above, (b) even more, if the remaining signatories would stay put at their now suboptimal abatement level q..* 29 However, defection does not pay if cooperation breaks down altogether as a result of the defection (a variation of the prisoners' dilema); each country now loses an amount corresponding to areas III and IV in the Figure. The model can be used also to show how it could pay signatories to bribe a non-signatory party to loin the treaty or a particular signatory not to defect. Each signatory, nos. 1 and 2, gains areas II + III from no. 3 joining (or not defecting in case (a)) which is far more than needed for compensating no. 3. The results presented in this Appendix hold even for the case of more countries than three and even if there are non-signatories aside from the countries under discussion. 30 Barrett, Scott, On the Nature and Significance of International Environmen tal.Agreements, Working Paper, London Business School, 1989 " The Problem of Global Environmental Pollution, Oxford Rev. Econ. Poliy, Vol. 6, No. 1, 1990 Bohm, Peter, CFC Emissions Control in an International Perspective, in Th. Economics of Managing Chlorofluorocarbons (Ed. by J Cumberland et Al) RFF, Washington, 1982 Deosit-Refund Systems - Theory and Anolications to Environmental. Conservation and Consumer Policy, Johns Hopkins Univ. Press, 1981_ * Economic Instruments for Reducing CFC Emissions, Nordic Council of Ministers, Copenhagen 1988 Grubb, Michael, The Greenhouse Effect: Negotiating Tarrets, Royal Institute of International Affairs, London 1989 Hahn, R.W.and A.M. McGartland, The Political Economy of Instrument Choice: An Examination of the U S Role in Implementing the Montreal Protocol, Northwestern University Law Review. Vol 83, Spring 1989 Hahn, R.W. and K.R.Richards, The Internationalization of Environ mental Regulation, Harvard International Law Journal, Vol. 30, no.2, Spring 1989 Hoel, Michael, Global Environmental Problems: The Effects of Unilateral Actions Taken by One Country, Memorandum from Department of Economics, University of Oslo, 1989 * Efficient International Agreements for Reducing Emissions of C02, Department of Economics, University of Oslo, 1990 Markandya, A., The Costs to Developina Countries of Enterin_ the Montreal Protocol, Report to UNEP, January 1990 Morrisette, P.M., The Evolution of Policy Responses to Stratospheric Ozone Depletion, Nat. Resources J., Vol 29, Summer 1989 UNEP, Montreal Protocol on Substances that Deplete the Ozone Laver, Final Act, 1987 Renort of the Technology Review Panel, 1989 31 , Transfer of Technology and the Financing og Global En vironmental Problems: The Role of Users' Fees, A Note by M. Tolba, January 1990 U S Congress, An Analysis of the Montreal Protocol on Substances that Deplete the Ozone Layer, Staff Paper, Office of Technology Assessment, Washington, D.C., February 1, 1988