Proceedings of the UNESCO - University of Tsukuba International Seminar on Traditional Technology for Environmental Conservation and Sustainable Development in the Asian-Pacific Region, held in Tsukuba Science City, Japan, 11-14 December, 1995.
Editors: Kozo Ishizuka, D. Sc. , Shigeru Hisajima, D. Sc. , Darryl R.J. Macer, Ph.D.
Scientific opinion has converged strongly around the view that greenhouse-gas emissions caused by human activity have raised global temperatures over the past two centuries, and will cause accelerated warming in future unless emissions are abated and atmospheric concentrations of the various GHGs stabilised. The central hypotheses from world climate models are that business-as-usual emissions of carbon dioxide, methane, and other greenhouse gases will raise mean global temperature by between 1 and 4 degrees Celsius by 2100, leading to a sea-level rise of between 25 and 90 cm, with the range of these estimates reflecting various assumptions specified in the IPCC's 1992 set of scenarios for the world economy (Intergovernmental Panel on Climate Change, Working Group II Second Assessment Report: Summary for Policymakers , September 1995, p.SPM1.). Large-scale scientific experiments to test these hypotheses are being designed and put in place, and results from those experiments should become available during the first two decades of next century. Comprehensive surveys of the present state of climate science are available in the recently approved Second Assessment Report from Working Group I of the Intergovernmental Panel on Climate Change (Forthcoming 1996 from Cambridge University Press).
Meanwhile, policy discussion has also progressed steadily under the aegis of the IPCC, the United Nations, and the OECD. Two central issues in the debate have been
(a) What criteria should be used by the world community in choosing the scale and timing of policy response to climate change?
(b) Which policy instruments should best be used to achieve global abatement of greenhouse gas emissions?
Debate on these issues remains heated, but there are growing signs of consensus on some basic themes:
(a) So-called "economic instruments" - carbon taxes and tradeable permits - are preferable to command-and-control techniques on a global scale, both because of the efficiency with which economic instruments can reveal and promote the least-cost abatement options, and because of the severe constitutional and political objections in the way of establishing a world authority with the required powers to enforce a command-and-control system.
(b) While a worldwide carbon tax is the economic instrument usually assumed for the purpose of economic modelling of emission abatement, the practical advantages of tradeable permits have come increasingly to the fore. The UNCTAD secretariat in 1992 came out in favour of tradeable permits as the preferred instrument, and Working Group III of the Intergovernmental Panel on Climate Change has recently moved towards the same view (Fisher et al 1995). Especially if international policy continues to be developed in an incremental fashion, the likely outcome is an international system which requires countries to acquire and hold some sort of tradeable emission quotas or permits to match their contribution to atmospheric concentrations of GHGs.
(c) In a sense, the first step towards evolution of a world tradeable-permit system was so-called "Joint Implementation", under which developed countries promoted investment by their own national firms in abatement projects in the "South". Usually this has been done by offering companies in OECD countries exemptions from domestic abatement obligations or taxes, in return for those firms carrying out recognised abatement projects in developing countries. Suspicions of rent-seeking and capture by large multinational firms have led to a bad press for JI, but for all its shortcomings this approach of seeking to create a world market for GHG abatement points the way forward to a less discriminatory and more comprehensive tradeable permit or tradeable quota system. The present paper therefore looks "beyond Joint Implementation".
Assume that the international community adopts a tradeable permits regime of the sort outlined by Bertram (1992), Barrett (1992), Larsen and Shah (1994), Stavins (1995), and by a number of papers presented at the 1994 IPCC Workshop on Policy Instruments and their Impacts at Tsukuba University, Japan, January 17-20 1994. An annual global emissions budget is broken down into quotas each of which is specified as a percentage of the total, and these quotas are then allocated across the world community by some agreed rule. I assume that the chosen rule is equal per capita entitlements based on population in the year 2000 (cf Bertram 1992), but other rules could equally apply (see below).
To motivate the discussion which follows, it is assumed that the allocation of tradeable permits is such that OECD countries (the "North") receive fewer permits than they require to meet their obligations, while non-OECD copuntries (the "South") receive more permits than required. A number of studies have estimated the distributional consequences of allocating equal emission entitlements per capita across the world community (Bertram 1992a, 1992b, Barrett 1992, Grubb 1989, Larsen and Shah 1994; Kosobud et al 1995).
The paper utilises recently-published 1991 country-by-country emission data together with the Penn World Tables estimates of PPP GDP per capita and population about 1990 to present estimates of the order of magnitude of emission trading under a per-capita permit allocation rule, and uses a simple model to consider the incentives which a permit-trading regime presents in relation to technological progress in the North and South, under various time-paths for the global emission budget.
Figure 1 shows for fossil fuel use and cement manufacture the familiar relationship between GDP and carbon emissions and the much looser relationship between GDP and the carbon intensity of GDP. The first panel emphasizes that economic growth is the driving force behind rising emissions, and hence rising atmospheric concentrations, of CO2 in the long run. The second panel emphasizes that technology varies extremely widely across countries, so that a country which combines rapid economic growth with rapid technological progress may be able to avoid raising its emission profile. Figure 2 gives the corresponding picture for all main GHGs (CO2, methane and CFCs).
Recent IPCC Working Group II work on technology has concluded that already-known technologies would enable the world economy to grow on a business-as-usual path through to 2100 while bringing total GHG emissions down from the present 6 gigatonnes of carbon per year to 4 GtC by 2050 and 2 GtC by 2100 (IPCC forthcoming). The feasibility of abatement is therefore not an issue; the cost of abatement, however - and hence its economics - is fundamental.
A major survey of studies on abatement cost has recently been completed for the IPCC (Hourcade, Halsnaes et al, forthcoming). The general conclusions are between 10% and 30% of the world's current energy consumption can be costlessly eliminated; this part of world CO2 emissions therefore reflects pure market failure.
the costs of stabilizing the OECD's emissions of CO2 over the next half century at 1990 levels are probably of the order of a 1-2% sacrifice of GDP, and a number of analysts using CGE models have extrapolated this to the world economy (see, e.g., Dean and Hoeller 1992).
Figure 2: Emissions of Three Main Greenhouse Gases, in CO2 Warming Equivalents
The issue is shown schematically in Figure 3 which contains two alternative Average Abatement Cost curves. One of these, CT, corresponds to the "top-down" perspective on technological change; the other, CB, corresponds to the "bottom-up" perspective. The top-down view is that of economic modellers extrapolating from past behaviour with a conservative perspective on the extent to which, and the rate at which, new technologies will in fact be adopted by the world's firms. The bottom-up perspective is that of visionary engineers who focus on what could be achieved if the world community put a concentrated effort into forcing down its emissions of GHGs. (For discussion of the two schools see Hourcade, Richels et al, forthcoming).
In a theoretically perfect world economy, the true abatement cost curve would be revealed by the social experiment of imposing a uniform carbon tax and then raising it steadily, observing for each level of tax the corresponding amount of emission abatement. Top-down modellers conduct the same exercise of "scanning" their model economy with a carbon tax, but on models which embody estimated actual elasticities of substitution which are assumed to remain unaltered across a wide range of tax rates and abatement effort. The relative unresponsiveness of the model structure to huge changes in the incentives which economic agents face is the key shortcoming of top-down work to date, and the common practice of inserting judgementally some "backstop" technology threshold at which the model economy's resistance to technical change suddenly breaks (cf Manne and Richels 1993) provides only an unsatisfactory ad hoc fix.
Bottom-up work, on the other hand, is apt to slip into technological utopianism, ignoring the real-world economics of innovation and substitution. Integrated models which combine the two represent the current frontier of modelling work in this area.
(Incidentally, the convention of naming the two schools of thought in this way seems to me quite wrong. The "top-down" school rests its case on the argument that, left to themselves, individual economic agents will abate reluctantly, so that strong incentives such as a world carbon tax of several hundred dollars per tonne will be required to achieve abatement targets if the instruments used rely on decentralised behaviour, which is what I would regard as bottom-up adjustment. The "bottom-up" school relies on the imposition of technological progress by a strong central planner with an engineer's knowledge of the technical possibilities, which I would regard as a top-down approach to abatement.)
What is missing from most modelling work to date is scenario work on dynamic technological progress and adoption of innovations. The history of the world energy economy is dominated by technical change on a very large scale as wood fuels were replaced by coal, then coal by oil. Currently the early stages of a new wave of energy technologies are apparent; and once the cumulative tendencies of such technological "waves" take over, the transition away from fossil fuels towards renewable energy systems - which harvest the earth's current incoming energy from the sun by means of organic processes (biomass growing and burning) and physical processes (wind, tidal, direct solar) - may occur quite rapidly. The main unknown is the future of nuclear energy - a technology which seekes to transfer the source of solar energy from the sun to the earth. The race between "renewables" technologies which locate the required nuclear reactions on the sun, and "nuclear" technologies which relocate the reactions to the earth, remains undecided. There do, however, seem to be grounds for believing that in the long run the sun may have a comparative advantage in nuclear energy production relative to the earth, and that the optimal strategy is to rely on planetary imports.
In terms of the costs and benefits of various global responses to greenhouse warming, there are three relevant dimensions of technological progress towards sustainable energy: the speed, cost, and location of innovations. To some extent all of these are unknown until after the event, but it is possible to analyse some tendencies. In the next section of this paper I consider the way in which choice of a particular economic instrument (tradeable emission poermits) might affect the innovation incentives faced by agents in different parts of the world.
Figure 3: Average Abatement Cost: Two Scenarios
the initial allocation of scarce permits is of fundamental importance in determining the effect of permit trading on the world distribution of income and wealth (Barrett 1992, Bertram 1992);
the time-path followed by the permit price depends heavily on the nature of the property right created and the constraints imposed (Kosobud et al 1995)
transaction costs are the main potential source of distortions in a permit-trading regimen (Sandler and Sargent 1995, Stavins 1995).
The discussion which follows concentrates on the first two of these. The tradeable-permit instrument works by imposing a global quantity constraint on some valued activity such as emissions, and then allowing the market mechanism to reveal the shadow price of the constraint at each point in time, as all countries respond to the emission constraint by abating their emissions to the point where marginal abatement costs are equalised across all countries. In Figure 4, the length of the horizontal axis between ON and OS represents the global emission budget and two marginal-abatement-cost schedules are drawn, for the "North" and "South" respectively. The North's MAC is drawn with respect to origin ON and shows how lower volumes of emissions from the North are associated with higher marginal costs of abatement. "Business as usual" (zero abatement) would locate the North at point N with emissions ONN. Similarly, in the absence of abatement the South's emissions are OSS. Total business-as-usual emissions (ONN + OSS) violate the global emission limit. Economic instruments (a carbon tax or a tradeable permit system) work by pushing both parties back up their MAC curves until the global contraint is met at least cost; this brings them to point E in Figure 4, with a shadow price on emissions of P* which is made explicit to individual economic agents by means of a corresponding carbon tax equaL to P*, or permit price of P*.
Figure 4: Optimal Abatement to Meet a Global Emissions Limit
Because the world economy starts out with very unequally distributed purchasing power, it is important to ensure that the initial allocation of permits does not lead to abatement behaviour being dictated by financial rather than technical constraints, and also to organise the allocation of permits to secure agreement from all countries to participate. In practice this means that permits would be allocated disproportionately to poorer countries, so that the onus to buy-in emission rights lies with rich countries which possess the means to pay. There is general agreement in the literature that simply "grandfathering" emission permits on the basis of current emissions, or on the basis of present GDP, could not provide the basis of a workable international agreement, because it would impose heavy costs on non-OECD countries, while enabling OECD countries to capture rents from the shortage of atmospheric carbon storage space for which the OECD countries themselves have been responsible through high past and present emissions (Hourcade, Halsnaes et al forthcoming section 9.1.5.1.4; Edmonds et al unpublished; Bertram 1992; Barrett 1992; Larsen and Shah 1994).
As Barrett (1992 p.86) has observed,
Since accession to a treaty is voluntary, and since the rich countries have a particular interest in having the poor participate in a treaty, the rich have an incentive to offer the poor countries a treaty proposal that makes them better off.
Two alternative permit allocation principles have emerged which could be the basis for agreement. The first, based on a simple and universally intelligible ethical principle, is the per-capita approach, which allocates permits at the outset in proportion to the population of each country (Bertram 1992). The principle here is that the global atmosphere is a common, in which all individuals of the world community have equal stakes. In issuing property rights to this common, there is a presumption in favour of giving an equal share to each individual.
The second option is to allocate emission permits to the South on the basis of their business-as-usual emissions, with the North receiving the residual. This is an application of the Pareto principle: the South is no worse off with the agreement than without it, while the North has the option of undertaking its own abatement programmes to meet the constraint, or buying-in permits from the South (thus creating financial transfers from North to South). (See Edmonds 1993, Larsen and Shah 1994).
The per-capita principle is based on the ethical view that all individuals have equal rights in the global commons. The no-regrets-for-the-South (NRFTS) principle is purely pragmatic (game-theoretic) in its quest for agreement. The usual argument against per capita allocation has been that the financial transfers from North to South would be greater than the North's willingness to pay, because the South would be "rewarded" for foregone per capita emissions rather than for foregone total emissions, which means much greater financial transfers for any given level of global abatement.
Figure 5 shows the two cases. With the global emission budget fixed as the length of the horizontal axis, the per capita rule would allocate the North ONB of permits, while the South would receive OSB. The North would then abate to point E and buy-in BA* of permits from the South, in the process paying the South the sum P*.BA*.
Figure 5: Two Possible Permit Allocation Rules
If we assume that the South has no willingness to pay for global abatement, so that the NRFTS rule is the minimum requirement for agreement to be reached, then the quest for economic efficiency would point to this regime. A per-capita rule would apply if equity considerations were also important. The abatement incentives to the two parties are theoretically the same under either rule, but the per-capita rule obviously increases the total volume of rents which the South secures from its low-emissions status, and might thereby accelerate technical progress in the South somewhat. The per capita rule has the operational advantage that the objective criterion used (population) can be readily and unambiguiously measured, and is (to a first approximation) invariant to abatement effort, whereas actual emissions are hard to measure with precision (Problems are well-known. Are landuse changes to be included? Is methane included? Are emissions measured on a net or a gross basis? What conversion factors are used to bring GHGs to a common unit of measurement?), and become an endogenously-moving target once abatement gets underway.
Discussion of permit allocation to date has mainly taken a static cross-sectional view of the story. Bringing in dynamic elements makes the task of instrument design potentially easier as well as more sophisticated. Two dynamic processes which stand out are the optimal time-path for global abatement, and the rate of technical progress.
The abatement time path sets the size of the global budget for each period (that is, the length in each period of the horizontal axis in Figure 5). Over time, the aim of emission abatement is to stabilise not the rate of annual emissions, but the concentration of GHGs in the global atmostphere. As several authors have pointed out, this means that abatement effort can be distributed over time in such a way as to minimise the long-run cost. Kosobud et al (1994) have shown that a time path with initially low abatement effort but a rising shadow price over time leading to progressively-increased abatement, may give a superior dynamic path compared to a "big bang" focus on early abatement at high cost. Numerous commentators have recognised that abatement costs will fall as technology progresses, and this may be one reason for deferring abatement effort.
A possible permit allocation that would start the process running while limiting the early abatement cost would be one which combined a per-capita permit allocation rule with a global budget set at such a level that the South's allocation exactly matches its actual emissions. This would give incentives for abatement and trading with a relatively low shadow price. The budget could then be reduced over time on an agreed path, and the shadow price would rise or fall depending on the rate at which the tightening budget constraint was met by technological progress rather than by adoption of the abatement possibilities built into the MAC curves drawn in Figure 5.
Figure 6: Technical Progress in the South
It can be seen that the downward shift of the South's MAC curve has had three effects which are relevant to gains and losses from technical progress for a given global emissions budget:
abatement cost in the South has fallen, which frees up resources for other uses
the volume of permits sold by the South to the North has increased
the world price of permits has fallen
If the North's MAC is sufficiently steep that its elasticity is less than unity, then the combined effect of the second two will be zero or negative for the South; the fall in revenue from permit sales offsets or outweighs the direct gains to the South from its lower abatement costs. The South is then transferring technology rents to the North. If, on the other hand, the North's MAC curve is flat over the relevant range, then technical progress in the South both achieves direct resource-cost savings and brings in net revenue.
An implication of this analysis is that the South gains most from technical changes at the lower end of its MAC curve. If we think of the South as a monopolist seller of permits, its optimal strategy would be to maximise the revenue from permit sales by aiming for the most profitable point on MACN (the North's demand curve for permits) while minimising the size of the abatement cost triangle A*ES. This implies an incentive for the south to focus its R&D effort on the adoption of known low-cost abatement technologies at the lower end of its MAC curve, but to avoid radical leftward shifts of the whole MAC curve by the sort of fundamental research and innovation involved in a global move to sustainable energy. The North, in contrast, as the buyer of permits, has a generalised incentive to shift its MAC curve left, while at the same time trying to steepen it at the tail. A global tradeable-permit regime, in other words, may generate incentives which reinforce the existing concentration of fundamental research and technological progress in the North, with the South in the more passive role of adopter of the resulting innovations.
This situation may be acceptable, to the extent that the net flow of technology rents runs from North to South (that is, as the South shifts its MAC down its total revenue from permit sales rise faster than its payments of royalties on the Northern technologies being adopted). However, to the extent that countries in the South wish to promote their own endogenous technical progress, the effect of a tradeable-permit regime on rent distribution could prove a significant problem.
The size of North-South transfers, however, will depend not just on the trend of technology but also on the trend of the global budget, and here it is possible to see a possible feedback mechanism to ensure that the South does not suffer from rapid technical progress. Tightening the global budget by the amount of extra abatement achieved by the South - that is, by the horizontal distance EJ in Figure 6 - while maintaining a per-capita allocation rule, would ensure rising permits revenues and falling resource costs for the South. A rule of this sort may have some potential as the "carrot" to hasten technical progress in the South and hence minimise the long-run cost of global abatement.
The budget is then tightened and the path of the world price of permits (P* in Figure 4) mapped for various technology-change scenarios. Results of the modelling exercise are not available at the time of preparing this version of the paper, but will be presented at the seminar.
=0.25, and for the rest of the world
=0.06. These figures have no very strong credibility at this stage of research (fortunately, the model results are not particularly sensitive over the range of values shown), but they do provide us with the ability to generate benchmark estimates from the model. With these assumed values for
and using the actual 1990 data for 125 countries' population, real GDP. and fossil fuel CO2 emissions, we conduct experiments by imposing various global quota limits on CO2 emissions and observing the outcomes under a permit trading system, with various amounts of technical progress in North and South (modelled simply by dropping "free" abatement into the individual countries' models and thus freeing-up part of their permit allocation to be sold to other countries). For simplicity, we assume that permits are allocated on a per-capita basis.
First we ignore technical progress and simply tighten the global quota progressively, forcing all countries to cut back their emissions in line with their Marginal Abatement Cost Curves. Table 2 shows the resulting pattern of response for 10% and 20% abatement, and Figure 7 shows the Marginal Abatement Cost curves for North and South, and the permit allocation, for the 20% abatement case.
Then taking the 20% abatement case as a benchmark, we introduce technical progress in the South to see who gains and who loses. The answer is that the permit trading regime as specified here causes the South to lose all the benefits from its own technical progress, but to gain from technical progress in the North. The model results are in Table 3.
These results are still only preliminary findings from a model in its early stages of development. However, they point clearly to an issue which has not yet figured in the international policy debate on economic instruments. This is that if an international agreement is struck on a basis which is favourable to the South in terms of permit allocation (using a rule such as the per-capita rule modelled here), then this permit allocation sets a benchmark compared to which the South could lose from technical progress wherever it occurs, because of the threat which technical progress poses to the value of allocated permits.
On the face of it, this seems to provide an additional argument for the NRFTS allocation rule against the per capita rule: starting from an allocation which reflects the status quo, all parties should gain more from the reduced abatement costs that technical progress brings than they lose from the falling value of their permits.
The issue is one which needs further exploration with this and other models.
Figure 7: Marginal Abatement Cost Curves from the Model