You are on page 1of 30

Climate change impacts, energy, and development

Professor Michael Grubb1


Paper to World Bank

Annual Bank Conference on Development Economics


Tokyo, 30 May 2006

ABSTRACT Climate change both through its direct impacts, and the implications of measures to tackle the problem through reducing greenhouse gas emissions is likely to be a major influence upon global socio-economic development during the 21st Century. This paper addresses three dimensions of the challenge: the risks associated with changing climatic patterns, particularly in relation to key vulnerabilities and scope for adaptation, and economic approaches to evaluating the scale of those risks; the relationship between emissions and development, including potential opportunities between climate change policies and development at the project and sector levels innovation and macroeconomic dimensions of emissions mitigation in the national and global context, including the role of economic instruments The broad conclusion is that climate changes holds opportunities as well as threats to development, and that the balance between them will to a large degree be a function of how public policy responds the most dangerous and damaging approach being to ignore the problem and hope that it goes away.

Chief Economist, The Carbon Trust, 3 Clements Inn, London WC2A 2AZ michael.grubb@carbontrust.co.uk; Senior Research Associate, Faculty of Economics at Cambridge University; Visiting Professor, Centre for Environmental Policy at Imperial College, London.

Introduction The general acceptance that climate is a real and pressing problem is beginning to move the issue from one of scientific debate and observation towards questions about the impact of climate change on economic development, and the implications of measures to tackle it. This paper briefly summarises the scientific evidence and nature of the problem, and moves on to discuss the implications and relationship to economic and development policy. The paper comes at an interesting time in relation to international assessment processes. The Fourth Assessment report of the Intergovernmental Panel on Climate Change is in its final analytic stages, with drafts either already circulating or soon to be issued for global governmental and expert review. In addition, the Stern review of the Economics of Climate Change, commissioned by the UK Treasury as an assessment of the global economic issues, is due to report this autumn. Thus a much richer base of material will shortly become publicly available. Obviously, this paper reflects my view of the debate. The paper is divided into three main parts: Part I summarises the scientific evidence, projections, and discusses key points around evaluation of impacts particularly in relation to developing countries; Part II discusses the relationship between emissions and economic development, presenting four facts and four opportunities around the CO2-development relationship Part III analyses the macroeconomics of mitigation policy, including the role of innovation and economic instruments

This paper as submitted for the ABCDE conference contains only Parts I and II. Part III will be included for the final publication.

PART 1: SCIENCE AND THE NATURE OF THE CHALLENGE2 Is Climate Change Real?
Emissions of various gases from industrial and other human activities are changing our atmosphere. Climate change encapsulates the wide variety of accompanying impacts on temperature, weather patterns and other natural systems. Despite decades of research, important things remain uncertain, but much is also now established beyond reasonable doubt. The fundamental science of climate change. The fundamentals of climate change have long been well understood because they involve the same basic physics that keeps the earth habitable. Heat-trapping greenhouse gases in the atmosphere (of which the two most important are water vapour and carbon dioxide, CO2) let through short-wave radiation from the sun but absorb the long-wave heat radiation coming back from the earths surface and re-radiate it. These gases act like a blanket and keep the surface and the lower atmosphere about 33 deg. C warmer than it would be without them. The Earths greenhouse blanket is a good balance between the extremes of our neighbours: Mars, exposed without any greenhouse gases, is a frozen wasteland; whilst Venus remains trapped in a dense blanket of hot CO2. Primarily through the burning of fossil fuels and deforestation, humans have been increasing the concentration of CO2 and other greenhouse gases in the atmosphere since the industrial revolution began, thickening the greenhouse blanket. The world has been warming. Surface warming in recent decades is established beyond doubt. So too is cooling of the stratosphere (the layer above the main blanket), as would be expected from greenhouse warming that traps more heat near the surface. Direct temperature records back to the middle of the last century are considered to be reliable enough to establish that recent temperatures are warmer than any since direct measurements began. Since the 1980s, due in part to the clean-up of other industrial pollutants (some of which had masked underlying warming), the underlying, long-term greenhouse warming has emerged more clearly - all of the 10 warmest years have occurred since 1990, including each year since 1995. Better accounting for these and other factors can now generate a good fit between the observed temperature trend and the results of computer simulations that incorporate these multiple factors (Figure 1).

The first half of this section draws heavily upon a report prepared by the author for the Carbon Trust, published as Grubb (2004). The source data draws heavily upon the Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC, 2001). The IPCC is currently engaged in its Fourth Assessment, which will update such analyses.

Figure 1. Global average temperatures since 1860.

Source: IPCC, Third Assessment Report. The charts compare observed temperature trends (red) with model predictions without (left hand side) and with (right hand side) greenhouse gas emissions Although debate continues about the exact temperatures during mediaeval times, a wide variety of proxy indicators (such as tree rings, coral layering, glacier records, etc) give a high confidence that the warming observed is unprecedented; indeed it appears that global average temperatures have varied by less than a degree C for thousands of years, and probably during the entire post Ice-Age period during which human civilisation has developed, so that recent years are probably the warmest seen for more than 100,000 years. Scientists have been unable to identify natural factors that could explain either the degree or the pattern of the surface warming and stratosphere cooling observed over recent decades. Understanding is still incomplete, but the fundamentals are clear and supported by a long list of other accumulating impacts. Other observed indicators and impacts of our changing climate The list of observed changes other than temperature and sea-level is growing rapidly. These include the thawing of permafrost, later freezing and earlier break-up of ice on rivers and lakes, lengthening of mid to high-latitude growing seasons, poleward and altitudinal shifts of plant and animal ranges, declines of some plant and animal populations, and earlier flowering of trees, emergence of insects, and egg-laying in birds.3 Perhaps the most clear, prominent and consistent indicator of warming is the retreat of mountain glaciers (e.g. Chart 2) which has been a worldwide phenomenon. Impacts on ice are also clear around the poles. The Arctic ice cap is shrinking and the Larsen Ice Shelf
Cited from IPCC Third Assessment, Climate Change 2001, Report on Impacts, Adaptation and Vulnerability (Policymakers Summary).
3

around the Antarctic peninsula has undergone unprecedented calving. Another widelyobserved impact is the bleaching of coral reefs caused at least in part by rising sea-surface temperatures. Changes in extreme weather events potentially have considerable impact on humans, but since by definition they occur infrequently, trends are hard to prove. Warming increases evaporation and precipitation, and both aggregate rainfall and occurrences of heavy precipitation events in northern mid-latitudes (e.g. Europe and the US) the principal cause of flooding has increased in recent decades. In tropical regions, the potential for more intense hurricanes and typhoons increases in a warmer world, but the data are sufficiently sparse and complex that the observational trend remains in dispute. (a) c. 1900

(b) Recent

Figure 2. Alpine Glacier: comparison of present to 1900 Photo Source: Munich Society for Environmental Research

The impact on some other extremes is better established. Many areas have seen fewer long cold spells and more long hot spells, in ways that are consistent with the predictions of climate models. But unlike the general trends of temperature, ice and sea level, it may always be questionable to attribute any one particular weather event to climate change,

because all weather events have multiple causes. So the question was X due to climate change? cannot be answered simply whether X was record temperatures, exceptional storms, floods or droughts. Nevertheless science may increasingly be able to estimate how much have past emissions increased the risk of such events? and the chances, at least of extreme high temperatures, and in some areas droughts and flood events, are rising.4 Insurance data (Figure 3) show a dramatic rise in the economic costs due to extreme weather events, though a major part of this is probably due to changes in demographics, property valuation and insurance practices.

Figure 3. Global costs of extreme weather events (inflation-adjusted) since 1950.

Note. The economic losses from catastrophic weather events have risen globally 10-fold since the 1950s,
after accounting for inflation. Part of the trend is linked to growing wealth and population, which increases economic vulnerability to extreme events, and part is linked to regional climatic factors (eg. changes in precipitation and flooding). Source: IPCC Third Assessment, Climate Change 2001: Synthesis report (Figure 2-7)

The distinction between climate and weather is itself a bit like that between sea-level and waves. Sea-level sets average conditions which vary locally according to tides and coastline, but even understanding all these does not mean one can easily pick out trends from individual waves, or predict them in detail. But the complexities and uncertainties around climate change should not obscure the basic facts. The fundamental mechanics of climate change are well understood; the world is warming; and much of the warming is due to human emissions of greenhouse gases. The next section explains why climate changes seem set to accelerate in the future, and the varied impacts this may bring.

The IPCC Third Assessment (IPCC, 2001) detailed observations, trends and projections, and in particular the Working Group II assessed observed and projected impacts. The Fourth Assessment will present considerably enhanced data on impacts and projections including at the regional level.

Projected impacts of climate change over the next few decades


Introduction and overview Some of the big, persistent trends indicated for example in glaciers, which embody a lot of inertia also due to past warming, can already be projected with confidence. The Snows of Kilimanjaro, for example, already much shrunk, are expected to disappear entirely within the next few decades it is already too late to avert this (Alverson et al, 2001). Glaciers and sea ice will continue to shrink, and there may be no Arctic sea ice in summer by the end of this century. The Antarctic ice sheet, being in a much colder climate, is less likely to lose mass, notwithstanding some shrinking ice shelves around it. Existing zones of preferred vegetation and associated crops will migrate towards the poles, requiring farming practices and ecosystems to adapt. However, many species and ecosystems have limited scope to move, because of a wide variety of barriers. The most comprehensive study to date estimates that about a quarter of the worlds known animals and plants more than a million species will eventually die out because of the warming projected to take place in the next fifty years.(Thomas et al. 2004) In addition to the broad physical and biological trends of warming and glacier retreat, sealevel rise, and the migration and loss of species and ecosystems, other predicted impacts of climate change are many and varied, and as research continues and experience begins to accumulate the list grows longer. Scientists rate the following other changes to be very likely (with more than 90% confidence):5 Higher maximum temperatures, with more hot days and heat waves over nearly all land areas. This would increase heat-related deaths, as well as heat related stresses on crops, livestock, etc; Higher minimum temperatures, fewer cold days, frost days and cold waves over nearly all land areas. This would reduce cold-related deaths and crop and livestock-related stresses associated with frost and other cold conditions. The balance between this and the first set of effects obviously depends on the starting conditions, but also on the rate and degree of change. Tentative estimates predict net agricultural gains for the US and Europe for equilibrium global changes up to 2.5 deg.C (this does not include transitional effects), the balance becomes negative for greater changes; More intense precipitation events, resulting in increased floods, landslide, avalanche, and mudslide damage, with increased soil erosion and increased flood runoff.

The following changes are rated as likely (with confidence greater than two-thirds):

This list is drawn from the IPCC 3rd Assessment, Report on Impacts, Adaptation, and Vulnerability, Table SPM-1.

Summer drying over most mid-latitude continental interiors and associated risk of drought; More intense tropical cyclones (in terms of both wind and rainfall); Intensified droughts and floods associated with the Pacific El Nio events, in many different regions; and More variable Asian summer monsoon, obviously of particular relevance to the half of the worlds population that live in China, India and surrounding countries.

Human impacts: an overview of the debate There are several broad approaches to thinking about the potential implications of such impacts for human economies and societies. Research during the 1990s tended to divide between scientific emphases on physical impacts, and economic studies some of which were far more optimistic. The economic debate was largely stimulated by Nordhaus (1991), who argued that quantifiable impacts of a warmer climate would be modest and justified only very modest action to mitigate emissions, with a contrary view by Cline (1992) who adopted broadly comparable methods but found quite different results depending upon parametisation. Mendelsohn and others (1994) developed more detailed analyses of US agricultural impacts, which based on comparative static analysis concluded that moderate levels of climate change could boost US agricultural output. Over the 1990s he and colleagues extended the work to other sectors and other countries. These studies are considered further below, but one feature to which they draw attention is a prediction that impacts will be highly diverse, with the brunt of damages falling upon low latitude developing countries, whilst high latitude countries may gain and, according to his studies, mid-latitude countries may be little affected in aggregate over the Century. These optimistic analyses were based primarily on projections of aggregate, average warming patterns, and effective adaptation to these, and have come under extensive criticism for these reasons. Certainly, any evaluation of human implications needs to start from a more comprehensive understanding of the likely nature of impacts than displayed in these early economic evaluations. Moreover, human impacts depend on specific changes in regions and localities. Localised changes are likely to be both more varied, and harder to predict, than the global averages. All projections are thus still quite speculative. Nevertheless, two regional examples help to illustrate possible consequences. Summer drying and heatwaves in and around the Mediterranean could further stress water supplies in some regions that are already politically sensitive and heavily dependent upon irrigation for agriculture. The suffering could also drive expanded migration into northern Europe that which might itself come under growing pressure from increased floods and heatwaves. On the Indian subcontinent, Bangladesh and north-east India could face a number of diverse pressures: rising seas and storms inundating the Ganges delta region; a more variable monsoon undermining the agricultural foundations that feed a quarter of a billion people; and changing patterns of river flow as climate change impacts the Himalayan glaciers that feed the rivers, with corresponding international tensions across already volatile borders.

These are just tentative examples; the possible human consequences of climate change are only just beginning to be seriously considered. A particularly complex consideration is that whilst most scientific studies have focused upon the possible impacts of a warmer world, most human impacts may flow from the nature of a warming world, in which change often hard to predict at the local level may be the most difficult characteristic for societies to handle. Farming practices, water industries, and innumerable other social and infrastructural systems designed for the last Centurys climate will not necessarily adapt easily to the accelerating change now in prospect, particularly as some of the underlying natural systems are also pressured by global economic and population growth. Such considerations inform the risk assessment-led approach to considering impacts. One form of this is illustrated in Figure 4, in which the impacts of projected climate changes have been summarised in terms of five risk categories. This suggests that even at the most optimistic end of projections, some unique and threatened ecosystems will disappear and some regions will be exposed to adverse impacts. In the mid range, many unique systems may be at risk and the impact of extreme events would rise, with the developing countries hurt the most although impact on the aggregate global economy could still be modest. Change towards the upper end pose significant risks to all and the risk of planetary-scale abrupt disruptions becomes significant. To date, this debate on impacts between economists quantifying specific, potentially measurable and monetiseable impacts, and scientists focused on risk indices and scenarios, has been largely a dialogue of the deaf. The next section sets out more formally a structure for thinking about these different dimensions.
Figure 4 Five risk indicators associated with projected global temperatures changes

The chart shows how the projected range of temperature changes for this century (left hand panel) would affect the risks posed in terms of five generic reasons for concerns (right hand panel). In the columns, white indicates neutral or small impacts, yellow indicates negative impacts for some systems or low risk, and red means negative impacts or risks that are more widespread and/or greater. The five columns concern (i) risks to unique and threatened ecosystems, (ii) the risks from extreme climate events; (iii) the risks posed to specific regions, with only the most vulnerable being affected in the white/yellow zone but most in the red zone; (iv) the aggregated impact on the global economy, and (v) the risk from large-scale climatic discontinuities (such as collapse of ocean circulation patterns). The assessment took account only of the magnitude, not the rate, of change.

Source: IPCC, IPCC Third Assessment, Report on Impacts, Adaptation, and Vulnerability, Figure SPM-2.

Economic evaluation of climate change impacts


How costly may climate change really be? This is a natural question for economists in particular to ask, but an extraordinarily difficult one to answer. Continuing scientific uncertainties about the detailed nature, timing and severity of natural impacts are multiplied by many layers of uncertainty about how society will cope with growing impacts and how to quantify these. The impacts literature is dominated by natural scientists. But it is natural for economists to ask, for any given emissions pathway, how costly may climate impacts be taking account of human capacity to adapt and hence how much might the reduced impacts in lower emission pathways compare against the presumed costs of lowering emissions? Rather than answering this question quantitatively, this section attempts to introduce a comprehensive structure into the debate on impacts quantification, and to draw some basic conclusions. Attempts by economists to quantify impacts in monetary terms have tended to concentrate on a few measurable dimensions, using one of two approaches: either model simulations, or comparative-static (cross-sectional) studies that attempt to compare indices such as land value and other indicators as a function of temperature. Some of the most widely-cited studies have been those of Mendelsohn (1994), who referred to the latter as the Ricardian method. These methods are applied to specific sectors such as timber, energy, water supply etc. Mendelsohn and Williams (2004) present a recent collection of such studies. The essential foundation of such studies is that the explicitly climate-vulnerable sectors of the economy account for a relatively limited share of GDP, and in particular Ricardian approaches suggest that there are optimum temperatures for most of the sectors, which lie somewhat above the average temperatures typical in mid-latitude regions. This drives the principal findings that climate damages are found to be modest in mid-latitude regions, averse in low latitude regions, but with net gains in high latitude regions. Since mid-latitude countries which dominate world GDP, the net impact of climate change is modest across the Century. The Mendelsohn and Williams (2004) results suggest that globally aggregated damages and benefits are comparable for the next several decades; damages start to dominate

10

after about 2050 and get worse thereafter, but in an aggregated present value analysis these have little impact due to discounting. Table 1, drawn from a major review study on the "social cost of carbon" (Downing et al., 2005) helps to set such studies in perspective in terms of a 3x3 risk matrix covering the dimensions of uncertainty in both climate change impacts (rows) and the scope of evaluation (columns). As the authors note, over 95% of the studies that seek to put a monetized value on climate impacts have focused mostly on only two out of the nine elements of the matrix, namely the market and non-market costs associated with smooth projected change. This is true for example of most of the studies cited above, whilst Nordhaus estimates try to quantify the most measurable impacts, and then assumes by extrapolation that other impacts are correspondingly small. Even more limited are the comparative static (Ricardian) analyses which compare the costs of two assumed climates, since these neglect transitional costs of shifting systems from one climate to another (which is itself changing). However since the bulk of existing work lies in this domain, I start by discussing the metrics of evaluation within the boundaries of existing comparative-static and projection-based quantification studies. I then try to lay out considerations around the missing elements, respectively: the dynamic and socially contingent issues, focused upon the actual capacity of societies to prepare for, and to tackle, climate-related changes; and how as a result societal constraints might affect the welfare consequences of impacts (or limit adaptation) in ways beyond the evaluation of direct market and non-market measures currently employed; the risks that may arise from regionally variable (non-trend) changes within the broad envelope of projected climate variation including extreme events (bounded risks), and larger scale system surprises

The essential conclusion of these sections is that there is no a priori ground for believing these elements to be insignificant compared to those that economists have sought to quantify.
Uncertainty in valuation

A. Market
Uncertainty in Climate Change 1. Projection 2. Bounded risks 3. System change and surprise

B. Non-market

C. Socially contingent
Plausible effects have been posed but not adequately valued nor included in the marginal SCC No credible studies

Over 95% of the studies are in this category; with a bias toward market costs. Some models have explicit scenarios but most are tied to benchmark 2xCO2 scenarios and do not cover local changes in weather. A few exploratory studies, but not sufficient to provide robust estimates of the marginal SCC

Table 1. Locating the 'social cost of carbon' literature in a risk assessment framework
Source: Downing et al. (2005).

11

a) Issues in aggregating quantified impacts: discounting and contingent methodologies. Because of the long timescales of climate change, discounting over time is a critical determinant of the present cost from impact assessments. The discounting literature is enormous, and has yielded consensus that market-based discount rates are not appropriate for evaluation of very long-term issues like climate change. Indeed it is now general practice to use discount rates for public policy evaluation that are well below market interest rates, particularly for longer-term endeavours; uncertainty around future economic growth rates lowers applicable rates further (eg. Weitzman 1998). Indeed, the literature increasingly questions the form as well as the number used for discounting: whilst the argument of Heal (1998) for a logarithmic form has not been generally accepted, Groom et al (2003) do conclude that the classical single exponential form is not tenable, and the UK government itself has adopted a rate that declines over time (UK Treasury Green Book (2004)). All these revisions tend to amplify the present value of climate change impacts, which predominate in the longer term. This in itself has important implications, as sketched in the Downing et al report, establishing that the long-term cumulative impacts of climate change cannot be wholly "discounted away" in evaluation of climate damages. I return to some implications at the end of this section. In the pursuit of treatments that are ethically consistent across space as well as time, similar scrutiny needs to be applied to the evaluation of transboundary impacts. Contingent valuation methodologies based upon 'willingness to pay' lead for example to valuation of mortality based upon national Value Of Statistical Life (VOSL), which is heavily constrained by national income and can differ by a factor of well over ten between countries (IPCC, 1996) a fact which has already led to political dispute due to the apparent unequal valuation of life contingent upon the wealth of the country impacted. In aggregate there is a huge 'NorthSouth' asymmetry between the principal emitters and biggest potential victims: it is the rich countries whose mitigation expenditures would be most influenced according to the estimated global damages, not the poor. Hence the case for using national VOSL (and other willingness-to-pay based measures) is unclear. A logical link can only be maintained by appeal to the argument that abatement expenditure in rich countries would displace foreign assistance for adaptation or other aid (an indirect opportunity cost argument). But there is no evidence that mitigation expenditure does or would displace foreign aid. Moreover foreign adaptation assistance (partly due to institutional constraints and the dynamic uncertainties documented below) is likely to be an imperfect substitute for reduced climate variability. Equity weightings introduce a multiplier for VOSL or other willingness-to-pay based impact measures in poorer countries to increase their weighting in global economic aggregation indices (Groom et al, 2003), and thus attempt to correct for the apparent inequities arising from such approaches. However the basis and derivation for such weights is unclear. In my view this reveals some genuinely complex ethical issues underpinning global aggregation of damages that have yet to be resolved. Simply taking a purely egalitarian approach (for example, assuming a constant VOSL across humanity at the level of rich countries) does

12

vastly amplify quantified estimates of climate change impacts, but is similarly riddled with inconsistencies. Economics is thus struggling to provide an objective approach to quantifying global impacts, let alone one that is accepted as such across the most relevant parties. In the absence of this, the only ethically defensible approach to developing responses has to include negotiations based upon fair representation of both victims and emitters. b) Dynamic and socially contingent impacts6 Unlike issues of discounting and even international comparison, the literature exploring the socially continent aspect of impacts is extremely limited. Concerning this, three particular dimensions need consideration: (i) Transitional impacts. Spatial and time aggregation over the long run may mask the bulk of social costs, which are far more likely to be those associated with transitions and extremes: adaptation to a changed climate, predicted ex-ante, may be very different from adaptation to a changing climate, with attendant changes in the distribution and scale of extremes. Both theory and recent experiences (such as Asia and New Orleans) suggest that what matters is the joint effect of climate impacts and constraints upon preparation, reconstruction and adaptation capabilities. Consequently the scale of losses may be sensitive to the pre-existing conditions of the economy on which climate change impacts may fall. Hourcade (2005) argues that impacts may be aggravated by constraints on: (i) reconstruction capabilities; (ii) failure of costsharing mechanisms including insurance and international assistance; (iii) local obstacles including rigid agricultural practices; (iv) knock-on economic impacts arising from depreciation of this share of capital stocks (through real estate and property ownership); and (v) ecological constraints. Drawing in part on wider development literature on the economics of natural disasters (Benson and Clay, 2004), Hallegate and Hourcade (2005) present a model in which poor societies are unable to recover from one extreme climate event before the next disaster strikes, leaving such countries trapped a cycle of under-development. Finally, mechanisms for adaptation, compensation and cost-sharing are inevitably weaker at international levels. This may increase the probability of adverse effects propagating across regions (including through migration), blurring any distinction between winners and losers. (ii) Uncertainty. The most obvious conclusion is that adaptive capacity needs greatly to be strengthened. This is unquestionably true but incomplete, not least because of the uncertain nature of impacts (particularly extremes) combined with the demonstrated incapacity of societies to prepare adequately on the basis of risk warnings (such as with the Asian Tsunami, and Hurricane Katrina). The main impact of climate change may arise from the interplay between climate uncertainty and the constraints and sources of inertia in social and economic systems. The dilemma is neatly illustrated by the juxtaposition of two papers in Climate Policy (Olsen (2006) and Butt et al. (2006)): one an agricultural modelling study of the
This section draws heavily on a Powerpoint presentation by Dr Jean-Charles Hourcade, posted on the World Bank website.
6

13

capacity of optimal adaptation to yield net benefits in Mali, the other a political economy study of the fact that a decade of international assistance efforts have made little headway in influencing practical policy in Uganda. Three big factors constrain the capacity for preparatory adaptation: (i) uncertainty in regional climate predictions, probably an order of magnitude greater than that in global average predictions; (ii) the masking effect of natural climate variability, which means that climate change signals may be undetected, ignored or misinterpreted; and (iii) the capitalintensive nature and inertia of adaptation strategies (eg. dams, building norms, etc). In combination these factors combined create a significant risk of maladaptation. The first lesson from comparing optimal control models is that costs and responses can be very different between perfect foresight and decision-making under high uncertainty. (iii) Aggregation and the value of climate stability. Estimates of GDP losses associated with climate change in existing literature are thus questionable to the extent that they neglect issues of transition, uncertainty, and human capacity. When extrapolated to global estimates they also face questions about the ethical basis of aggregation indicated above, and moreover implicitly assume perfect compensation between winners and losers, which human institutions are unlikely to deliver. c) Beyond smooth change: risks, surprises and the very long term Particularly in the longer term, these difficulties are amplified by the remaining elements in the risk matrix, namely larger-scale risks and surprises in the climatic system, particularly when combined with inertia. The 'Burning Embers' diagram of the IPCC WGII Third Assessment report (reproduced in Figure 4 above), attempted to depict how different levels of atmospheric change may affect risk levels across five impact metrics (ecosystem impacts, risks from extreme events, impacts focused upon vulnerable regions, globally significant impacts, and risks of surprises).

14

Figure 5: Potentially sensitive switch point areas in which local effects might trigger larger-scale changes The chart shows regions in which specific local phenomena may result in points of sensitivity for larger-scale and possibly rapid changes in regional or global climatic conditions. Source: J. Schellnhuber and H. Held, adapted from How Fragile is the Earth System?, in J.Briden, and T.Downing, T.(Eds.), Managing the Earth: the Eleventh Linacre Lectures, Univ. Press., Oxford (2002).

Scientists studying the interaction between different components of the climate system, and related natural systems, express concern about various possible instabilities. The north Atlantic ocean circulation is the best known, but is by no means the only example. Some studies question the stability of monsoon patterns particularly on the Indian subcontinent. The UK Hadley Centre projects that climate changes over Amazonia will lead to loss of the rainforest during this century. Other very long term possibilities include the melting or collapse of the Greenland and West Antarctic Ice sheets. Figure 5 illustrates some of these potential points of instability. The scale of threats posed by structural disruption; for example, to Indian monsoon or African rainfall patterns are extremely hard to evaluate, but clearly should not be ignored in any quantification that claims to be reasonably comprehensive. There are also feedbacks which concern scientists. Drying of the Amazonian rainforest system would feed more carbon back into the atmosphere. Thawing permafrost in the far north is likely to release pent-up methane (another and potent greenhouse gas). Far larger amounts of methane are currently locked on the sea bed and could ultimately be released, though only over very much longer time periods (Centuries or Millennia, if and as warming penetrates to ocean floors). There are inherent uncertainties about such systems; the dynamics that keep them stable, and their limits, are not well understood. When it comes to such big questions about complex systems, uncertainty is endemic. But especially given the inertia in all these systems including the inertia in economic systems discussed further in Part II - by the time limits are fully understood they may already be unavoidable. Several of the examples noted above systemic changes in monsoon patterns; desertification of the Amazon; and perhaps collapse of salinity shifts in the Thermohaline, may only be clearly identifiable through observational 15

data (on rainfall change, soil moisture / forest cover, and saline water movement); but by the time changes can be observed in the data with sufficient statistical certainty to understand and project much further, the inertia in all the systems concerns may well mean that the transition involved can no longer be avoided - potentially with dramatic consequences. Long term trends and risks This is because one fundamental characteristic of the climate problem is the inertia involved. Atmospheric greenhouse gas concentrations will not stabilise until global greenhouse gas emissions are reduced to a small fraction of todays levels, which few expect before the end of the century. Even after the atmosphere stabilises, other effects will continue to accumulate. Global temperatures will continue to rise for decades as the oceans slowly adjust to the higher heat input. Sea levels will rise due to both thermal expansion and ice melt effects which will cumulate over hundreds to thousands of years respectively (see Figure 6): over centuries, sea levels would rise many metres if and as the Greenland and/or west Antarctic ice sheets disintegrate. Although these seem far away, there are economic reasons discussed further in Part II why choices over the next few decades will affect emissions and concentrations for decades beyond that, and thus do much to determine the degree of commitment to a range of temperature, sea-level and other kinds of risks and instabilities noted below.
Figure 6 Accumulating impacts of climate change over the long term

The chart shows how key aspects of climate change will continue to accumulate long after global emissions are reduced even to low levels. Temperatures would continue to rise slowly for a few centuries as the oceans continued to warm; but sea level rise would continue for hundreds to thousands of years, due to the continuing impact on ice sheets in addition to thermal expansion of the oceans. The impact on species might also continue for centuries as the effects on ecosystem viability play out (not shown).

Source: IPCC Third Assessment Synthesis Report, SPM-5.

16

Conclusions Against this background, it is not hard to understand the conclusion of Downing et al. (2005) that the social cost of carbon the present value of damages associated with a tonne of carbon emissions - is characterised by huge uncertainties. They suggest that values could span a range potentially from 1 to 1000/tC, though also argue that the very low values in this range are unlikely. Hourcade (2005) offers a complementary perspective that would have similar practical implications. He argues that considerations such as those set out above could be used to argue that from an economic standpoint, climate stability is a component in utility functions that should be explicitly represented; and that given loss aversion (one of the most stable findings in behavioural economics) there is an intrinsic value to avoiding an unstable climate. Moreover, although it is poorer societies that may suffer most from an unstable climate, the decreasing marginal utility of income means that high-income populations / generations should be more willing to spend resources on protecting the climate (in addition to the fact that they have been and remain the major emitters of greenhouse gases). Climate stability is thus a superior good and this may influence some of the policy insights, including those relating to the cost-effective distribution of mitigation investments. To conclude, these various factors suggest a strong evidence base that:
1. Climate change will result in significant social costs, adaptation is very important but cannot be expected to negate such costs. All the discussion above underlines why trying to cost the

impacts of climate change is a daunting task. Early efforts to do so were criticised on the grounds that they assumed little or no adaptation a dumb farmer. Since a substantial degree of climate change is already unavoidable, there is no question that far greater efforts are needed to help societies adapt to its likely impacts, and that this has the potential to lower the cost of such impacts. But assuming that adaptation can radically reduce the costs of impacts is itself questionable, not least because of the fundamental nature of the uncertainties at the micro-level, where adaptation is actually relevant but where the uncertainties are greatest: in economic terms, it is not certain that replacing assumptions of dumb farmers (no adaptation) by assumptions about clairvoyant farmers (perfect adaptation) is necessarily more realistic. 2. Adaptation is only likely to contain adverse impacts if combined with serious moves towards slowing atmospheric change and ultimately stabilising concentrations. The risks
associated with uncertainties and irreversibilities are considerable and also constrain the ability of adaptive measure to avoid adverse impacts: a stable climate has characteristics of an intrinsic good. Moreover the whole issue is set in the context that climate change is not a discrete

phenomenon with an identifiable end-point to which we need to adapt; to the contrary, the projected growth of global emissions means simply that it will be an ongoing and accelerating process of continual climatic change, without any identifiable prospect of stability, and growing risk of planetary-scale disruption.
Very few studies have looked explicitly at the relationship between emission pathways and impacts, 7 but the two facts above make it plain that reducing global emissions ultimately to low levels - has to
ONeill and Oppenheimer (2004) offer one of the few studies of this nature. They examine the implications of three types of pathways: rapid change that follows reference until at least 2030 before departing and
7

17

be part of a balanced response. Hence with this background, I turn to consider emission prospects and the scope for emissions mitigation.

stabilising in 2100; slow change pathways that depart from 2005 and remain on a lower trajectory that finally reaches and stabilises at the target level in 2200; and overshoot pathways that exceed the target level by 100ppm in 2100 and then decline to the target in 2100. For a range of different ultimate outcomes (target concentrations 500, 600 and 700ppm) they find that the range of temperature outcomes in 2100 across the three types [of pathway] is about 0.5-1.2 deg.C, as large as or larger than the difference in long-term outcomes for different stabilisation levels. Moreover, the slow change pathways lead to median rates of temperature change that decline over time from an initial rate of 0.16deg.C/decade (for their central climate sensitivity), contrasting with peak rates of around 0.2deg.C/decade for the rapid change 500ppm case, and approaching 0.3deg.C/decade for the rapid change (and overshoot) scenarios with higher stabilisation levels. The authors conclude that the faster rates of change would both be harder to adapt to, and also would increase the risk of abrupt climate changes such as impacts on the Thermohaline circulation and polar ice melting.

18

PART II. CLIMATE, EMISSIONS AND DEVELOPMENT Despite the emerging efforts to tackle the problem, global CO2 emissions are widely projected to grow. If industrialised countries fail to limit their emissions and energyintensive, fossil fuel-driven energy systems remain a foundation of economic growth, it is hard to see rapid emissions growth in the rest of world being much curtailed as other countries aspire to the same levels of economic development. Yet, the link between wealth and emissions is far less strong than generally supposed. In this section I first present and probe four facts around the relationship between global economic and emissions growth, and then outline four opportunities that arise in the context of considering lower emitting development paths.
A. Four Facts

1. Large disparities in present emissions combined with population and economic growth create huge potential for global emissions growth if countries pursue existing models of development. Figure 7 plots per-capita emissions against population in different countries and regions (so that annual emissions in each are represented by the area of the blocks). Per capita emissions in the industrialised countries are typically as much as ten times the average in the more populous developing countries, particularly Africa and the Indian subcontinent. The potential for global emissions growth is thus huge, even if and as leading countries start to embark upon more serious action:

19

Figure 7. CO2 emissions per capita and population by region in 2000 Note. The chart shows the global distribution of CO2 emissions in terms of three major indices: emissions per capita (height of each block); population (width of each block); and total emissions (product of population and emissions per capita = area of block). Per capita emissions in the Industrialised countries are currently as much as ten times the average in developing countries, particularly Africa and the Indian subcontinent.

Source: Grubb (2004)

Population. Recent debates have tended to lower populations projections for this Century,
due to sharply declining birth rates, but most still involve projections that global population (the horizontal axis in Figure 7) will expand by around 50%.8 Economic growth and per-capita emissions. Recent debates about CO2-GDP relationships and projections focused on metrics of measurement,9 and about expectations of economic convergence vs a continued bimodal distribution of world per-capita income levels (eg. Jones 1997; Quah 1993, 1996; Barro and Sal-i-Martin, 1997; Riahi 2005), do not change the big picture. Almost all scenarios involve considerable economic growth in developing countries that, in the absence of counteracting policies, would tend to take their per-capita emission levels closer to those of the present industrialized world.

Mainly as a result of these two forces, the vast majority of non-intervention scenarios in the peer-reviewed literature (as reviewed for the IPCC Fourth Assessment) result in global CO2 emissions typically close to doubling around mid-Century, and reaching between two and four times current levels by 2100 taking the world far beyond the doubled CO2 concentrations scenarios that were the traditional focus of climate change modeling.

Out of 115 population scenarios collated recently by IIASA for the IPCC (Lutz et al, 2001; UN 2005; World Bank 2005; US Census Bureau 2005), the majority project population in the second half of the Century to be moderately stable at around 9-10 billion people, and only one scenario involves global population declining below 5 billion by the end of the Century. 9 An extensive debate has focused upon use of Purchasing Power Parity (PPP) vs Market Exchange Rates (MER)) and assumptions about economic convergence. Whilst the choice of GDP denominator would not be expected to have a first order impact on projections of a physical quantity such as emissions (Holtsmark and Alfsen, 2004a and b), it may have second-order impacts due to structural effects. Nordhaus (2005) concludes that the jury is still out and recommends a hybrid treatment using PPP base-year calibration with MER growth rates; Dixon (2005) presents evidence that PPP treatments could lower emission projections due to differential structural effects and associated sectoral emission intensities and elasticities.

20

9 Emissions Per Capita (Metric Tonnes) 8 7 6 5 4 3 2 1 0 0 10000 20000 30000 40000 50000 60000 Income Per Capita (1995 US$) EIT OECD OPEC NICs Other

Figure 8. CO2 emissions per capita for different economic categories as a function of income

Source: Grubb, Butler and Feldman (2006)

2. Beyond the stage of basic industrialization, there are large differences in per-capita emissions and huge variability in the CO2-GDP relationship. Figure 8 plots per-capita emissions against per capita GDP. At present no country with income beyond c.US$10,000 per capita emits less than about 1.5tC/cap. This reflects the emissions inherent in building basic industrial and urban infrastructures a fact which if extrapolated forward, and comparing against Figure 7, implies considerable growth in developing country emissions on almost any scenario for the world economy. Nevertheless there are big variations across the richer countries. Per-capita CO2 emissions in the new world developed economies, of 5-6tC/cap, tend to be around twice the levels typical in old world economies (Figure 7). Looked at more closely (Figure 8), the differences are even more extensive. This diversity is a modest source of hope, even based on current patterns; that itself implies a large degree of freedom over long run emissions even in the absence of radical technological breakthroughs or major lifestyle changes.10 A world in which most countries by the end of the Century emit 1.5-2.5 tC/cap clearly has far lower climate risks than one in which they emit at up to 3 times those levels. Yet Fig.8 illustrates that the difference between these is not primarily to do with wealth. It is to do largely with technology and infrastructure choices around the development,
10

Eg. The econometric evidence is mixed. If cross-country data show the predicted relationship (aleit with controversities, country-level analysis show relatively weak relationship between levels of GDP and emissionsn .. econometric anlayiss does not support an optimistic interpretation of the hypothesis that the problem will take care of itself with economic growth, bt the pessimistif interpretation, that gorwht and CO2 emissions would be irrevocably related, is not supported by the data either. Case studies confirm that there are major degrees of flexibility. (F. Lecocq, RFF/DEFRA workshop on the Economics of Climate Change: Understanding Transatlantic Differences, Washington, March 2-3 2006).

21

scale and efficiency of buildings, industrial and transport systems, and the supply systems (particularly electricity) with these are supplied. To these we now turn. 3. Emissions arise from a wide diversity of activities that embody huge inertia but many of which offer a correspondingly wide array of higher or lower emitting technology options It is widely recognized that the climate problem overall requires us to tackle a number of different gases and sources in addition to fossil fuels; in addition to the energy sector, greenhouse gases also emanate from agriculture, land use and direct industrial process emissions. For some developing countries, non-energy sources (particularly deforestation and other land use activities) dominate, and the desirability of addressing these for a multitude of reasons is widely recognized. But the relative role of energy-related emissions tends to grow with development. Yet even fossil-fuel related emissions result from several different systems each of which involve fundamentally different processes. Specifically, they are driven by energy demand in three main components (buildings, industry, and transport), supplied increasingly through three main systems (electricity, refined fuels, plus direct fuel delivery ( Figure 9)). Figure 9 Main components of global energy system and CO2 emissions

End Users
Buildings, Appliances & other (36%)

Supply Systems

Resources
Coal (37%) Gas (21%) Biomass Solar & geothermal

Direct Fuels & Heat (24%)

Industry
(Manufacturing and Construction)

Electricity System (36%)

Nuclear Hydro Wind, wave & tidal Solar PV

(35%)

Transport (29%)

Refined Fuels System (40%) Flow of economic value

Petroleum (42%) Biofuels

Notes: The data show the % of global energy-related CO2 emissions associated with the different parts of the energy system (including emissions embodied in fuels and electricity). Some small flows that comprise under 1% of global energy flows (eg. electricity and natural gas contributions to transport) are not shown. Note that patterns vary between regions (eg. industry is lower and transport higher in developed economies), and the sectors are growing at different rates (over past 30 years, energy demand for buildings : industry : transport has grown at

22

2.6%:1.7%:2.5% annual average. Non-electric energy industries (emissions from refineries, gas etc) cited as 7% of total, are allocated here in ratio 4:1:2 to transport : industry : buildings & other. Refined fuels taken as petroleum less input to elec; direct fuels and heat is the residual. Source data: Resources CO2 from EIA (2002); supply systems and end-use data from IEA (2002

The options for different technologies and systems are in most cases extensive. Buildings differ radically in the efficiency with which they consume energy. Urban planners are regularly faced with choices between road and rail investments. Electrical power is generated from a wide array of technologies, from the highest to the lower carbon emitters. A recent contribution (Pacala and Socolow, 2004) sought to frame the debate in terms of technology wedges that could each deliver savings of 1GtC/yr by mid century, and listed fifteen possible such wedges. Even in terms of energy resources, most options are not seriously limited in total; nor are low-carbon options including renewable energy sources. Although constraints limit what is feasible, the estimated global potential for tidal, wave and hydro are comparable to the scale of global electricity consumption, whilst most estimates of practicable wind and solar resources are substantially greater still (Figure 10 summarises various estimates). As with natural gas, key issues for delivery include the systems, and the fact that (with the minor exceptions of direct solar heating and lighting, and geothermal heating) all but one biomass - produce primary electricity. Rather, the constraints concern the economics of matching sources and systems to demands. It is often said that countries will not leave their domestic energy resources (such as coal) in the ground have coal, so will use it. There is no fundamental reason why the same logic should not apply to the renewable energy resources which sweep most countries: the reality is that the options developed are a matter of cost, technological capacity, and political choices.
Figure 10 Global renewable energy potential estimated by various studies compared to current global energy and electricity demand.

23

Source: Neuhoff (2005). See source for references and explanation


Electricity and Thermal Renewable Sources Fuel Renewable Sources Global Energy Consumption
450

Electricity Only Renewable Sources


125000

100000 Electricity G eneration (TWh/yr)


Fuel Use (Other) Fuel Use (Transport) Fuel Use (Buildings) Fuel Use (Industrial) Current Electricity Output Losses in Electricity Generation

360

50000

180

25000

90

-25000

-90

-50000 Wind Tidal Wave Hydro Geothermal Solar Biomass Fuel and Electricity Use

-180

Bonn TBP (2004) Greenpeace (1993) IEA (2002) x

WEA (2000) Grubb & Meyer (1993) WEC (1994) y

RIGES (1993) WBGU (2004) IPCC (1996)

Shell (1996) Fischer & Schrattenholzer (2001) Hall & Rosillo-Calle (1998)

4. The biggest determinants of future emissions will be the combination of the patterns set by industrialized countries and the capacity of developing countries to leapfrog towards higher-income but lower-emitting patterns of development I outline specific issues surrounding the economics of mitigation below. Here I note wider issues around the relationship of emissions to development. Development economics has increasingly emphasized the scope of development choices and their dependence upon institutional capacity in developing countries (Meier, 2001). The same is likely to be true with respect to emissions, and since one impact of weak institutions is that economies operate further from the efficiency frontier it cannot a priori be concluded that stronger institutions and resulting higher economic growth will result in higher emissions. Higher dependence upon fossil fuels is not intrinsically good for development and carries numerous attendant problems ranging from other environmental impacts to exposure to international fossil fuel price variability. Institutional capacity to accelerate efficiency improvements and foster lower fossil fuel paths could put countries on pathways that are both lower emitting and better for development the ideal being capacity to leapfrog the inefficiencies displayed in traditional models of development (the IPCC Fourth Assessment, WG3 report, contains a chapter (12) covering these issues in depth). The concept of developing countries leapfrogging to more advanced conditions is not new, but it tends to have been confined to academic discussion far from the realities of the ongoing struggle for economic development. It is a central conclusion of this paper that leapfrogging can neither be confined to the margins of debate any longer, nor does it represent one simplistic view of what could in theory be done. It is rather a necessity that represents a set of specific opportunities, to which I now turn.

Thermal Equivalent (E J/yr)

75000

270

24

B. Four opportunities 1. Opportunities for enhancing energy and economic efficiency There is a long-standing literature on the apparently favourable economics of improving energy efficiency (eg. IPCC Third Assessment, WGIII). Global studies date back to Goldemberg (1988). Even in developed countries that had made large strides during the 1980s and 1990s, considerable cost-effective potential remains.11 Numerous World Bank studies have highlighted the potential in developing countries tends to be even greater [WDR/which are best refs?]. Many factors explain the wastage the literature on barriers to energy efficiency is similarly enormous. One specific factor is the continuing degree of energy sector subsidies, which are generally recognized to be macroeconomically detrimental.12 Reforming subsidies, or introducing stronger regulatory measures around energy efficiency, is not easy. There are strong obstacles of political economy. In such conditions it is not uncommon that additional issues can offer leverage to achieve reforms that would anyway be desirable. It is perfectly possible that climate change could help to play such a role blaming the medicine on the need to tackle emissions make be one factor in making it easier to swallow. 2. Co-benefits Whilst stronger measures around energy efficiency and removal of fossil fuel subsidies may simply improve the internal efficiency of the energy sector, there are wider potential cobenefits, in both forestry and energy / transport sectors. Energy is a source of multiple emissions, for example sulphur dioxide. Higher energy consumption also means greater exposure to the impacts of price volatility in international fuel markets. The list is surprisingly extensive and studies indicate that such co-benefits could themselves justify a significant degree of measures that also reduce CO2 emissions (covered in depth in IPCC Fourth Assessment, Chapters 11 and 12). 3. Leapfrogging in infrastructure In the context of developing countries, the most important single consideration may be harnessing the pace of construction of new capital stock. Most of the sectors indicated in Figure 9 are also characterised by huge inertia. Industrial equipment that consumes, generates or processes energy has a lifetime typically measured in decades. The buildings that consume energy, the road and rail systems that determine transport demands, and even the pipeline and port infrastructures required for direct fuel delivery, can set infrastructure patterns for a Century or more. We have learned a great deal since rich countries started locking themselves in to higher emitting patterns of infrastructure. The wasteful nature of the UK building stock remains one of greatest headaches for the governments energy policy. North Americas/ exceptional energy intensity, and resulting dependence on oil imports, is to an important degree driven by choices in the transport sector made in the first half of the 20th Century.
11

eg. the UK improved its national energy productivity (ratio of GDP to energy consumption) by over 20% during the 1990s, but the governments Performance Intelligence Unit still estimated the potential net value of additional energy savings to the UK economy in 2000 at over 2bn/yr). 12 WDR refs? Cite data or is this all too well known anyway now?

25

Inefficient industrial equipment installed during those decades is frequently still operating, with continual cycles of refurbishment that can rarely bring performance up to the standard of new plant (Alic et al, 2003). Leapfrogging in infrastructure, by trying to make choices at the leading edge for the long term, is thus a huge opportunity in the course of development.

4. Leapfrogging in technology Finally, in some of the major developing countries there is the possibility of moving to the frontier of technological developments, in both domestic investment and capturing a growing share of the global market for energy efficient and low CO2 technologies. Indias embrace of IT is a classic example, but in the energy sector, Brazils dominance of biofuel technology now reaping large rewards - is perhaps more relevant. Technological development based on the large investment needs in key areas is a real opportunity, with solar PV technology perhaps being the biggest prize of all, because of the almost unlimited resource in which developing countries excel. Time is not on our side. In energy use and supply, emission patterns will be set by how the world chooses to invest tens of US$trillions capital over the next few decades, investments that will have irreversible impacts throughout the century. The uncertainties around global emissions growth trends, and the extent to which it depends upon choices made about the deployment of capital, are underlined by the IEA's World Energy Outlook (IEA, 2004), which estimates that about US$16tr will be invested in energy supplies up to 2030 (about US$10tr of this in the power sector), divided roughly equally between industrialised and developing countries. In their reference scenario most of the generation investments are in carbon-intensive stock; their alternative scenario involves more rapid growth in less carbon intensive investments, and although this is more expensive per unit, the scenario actually requires less capital investment overall because of the increased efficiency of end use (even when the end-use investments are included). The choice of path out to 2030 will have profound implications for the structure of capital stock, and its carbon intensity, well into the second half of this Century and even beyond.13 None of this that should deflect attention from the need for industrialized countries to get their emissions on a declining course. Indeed, as emphasized by a leading Chinese researcher (Zhou, 2005), it will be much harder for developing countries to achieve progress unless the worlds industrial powerhouses simultaneously develop the requisite lower-carbon technologies, businesses, capacities and institutional models. But a debate which ignores the crucial importance of emissions growth in developing countries is simply not a mature debate. And, nor would ignoring the opportunities be in the interests of developing countries themselves, for the reasons set out here, over and above their exposure to climate change.

Ulph (1997) noted that 'information acquisition and irreversibility could make a significant difference to policy advice, but models of irreversibility effects (Pindyck (2000), Kolstad (1996a and 1996b)) appear to have treated carbon and non-carbon intensive investments asymmetrically, assuming only the latter to be irreversible.13 In practice both embody considerable inertia: every major investment is a decision with irreversible consequences. The dominant net irreversibility is then carbon in the atmosphere and associated damages. Uncertainty about impacts (relative to best estimate) consequently increase the attractiveness of low carbon paths, to a degree that depends on the potential damages, risks and degrees of irreversibility (see also section 3).

13

26

The brake on embracing such opportunities appears to be partly political (the position in the global negotiations that developing countries have no responsibility to act), partly institutional (the sheer difficulty of thinking long-term in the crush of development pressures), and partly motivated by fears of economic consequences. In the final part of this paper, I turn more formally to the policy and macroeconomic issues.

Part III: Policy and macroeconomic dimensions Family illness has prevented completion of the third part of this paper prior to the ABCDE conference; it will be completed for the final published proceedings. It will draw i.a. upon studies of the global economics of atmospheric stabilisation published in Vurren, Weyant and de la Chesnaye (2006), Edenhofer et al, eds (2006), and studies of the economics of emissions trading and the future of the European Trading System (Grubb and Neuhoff, eds, 2006)

Annex: References
Alic, J.A., D.C. Mowery and E.S. Rubin, (2003): U.S. Technology and Innovation Policies- Lessons for Climate Change. Prepared for Pew Center on Global Climate Change. Alverson, K., R. Bradley, K. Briffa, J. Cole, M. Hughes, I. Larocque, T. Pedersen, L. Thompson and S. Tudhope (2001) Letter: A Global Paleoclimate Observing System. Science 293(5527): P.47 - 48. Baron, R. and ECON-Energy, (1997): Economic/fiscal instruments: competitiveness issues related to carbon/energy taxation. OECD/IEA. Barro, R.J., and X. Sala-i-Martin, (1997): Technological diffusion, convergence, and growth. Journal of Economic Growth, 2(1), pp. 1-26.

Benson, C., Clay, E., (2004).Understanding the economic and financial impact of natural disasters. The International Bank for Reconstruction and Development. The World Bank, Washington D.C.
Butt, T. A., B. A. McCarl and A. O. Kergna (2006) Policies for reducing agricultural sector vulnerability to climate change in Mali. Climate Policy 5(6).

Carbon Trust (2005), The UK Climate Change Programme: potential evolution for business and public sector, www.thecarbontrust.co.uk/carbontrust/about/publications/CTC518_CCPR2.pdf
Carbon Trust (2004), The European Emissions Trading System: implications for industrial competitiveness, [online] http://www.thecarbontrust.co.uk/carbontrust/about/publications

Carraro, C. and D. Siniscalco, (1998). International Environmental Agreements: Incentives and Political Economy, European Economic Review, 42: 561-572.
Carraro, C., (1999). The Structure of International Agreements on Climate Change, in C. Carraro, ed., International Environmental Agreements on Climate Change, Kluwer Academic Pub.: Dordrecht. Carraro, C. and B. Buchner,(2005). Economic and Environmental Effectiveness of a Technologybased Climate Protocol, Climate Policy, 4(3).

27

Carraro, C. and B. Buchner, (2005).Regional and Subglobal Climate Blocs. A Game-Theoretic Perspective on Bottom-up Climate Regimes, FEEM Nota di Lavoro 21.05 and CEPR DP.5034. Forthcoming in C. Carraro and C. Egenhofer, eds., Bottom-up Approaches towards a Global Climate Agreement, E. Elgar. Dadi (2005), We need a development model of low emissions, presentation to Exeter conference on avoiding dangerous climate change, http://www.stabilisation2005.com/day3/Zhous_Dadi.pdf Delay T. and M.Grubb, (2005). Solutions to climate change that also make money, Financial Times, 5 Oct 2005. Dixon, P. and Maureen T. Rimmer (2005), Analysing convergence with a multi-country computable general equilibrium model: PPP versus MER, Energy and Environment, 16/6. [online] <http://www.usitc.gov/ind_econ_ana/research_ana/seminars/Convergence_Dixon_Rimmer.pdf> Downing, T.E., Anthoff, D., Butterfield, B., Ceronsky, M., Grubb, M., Guo, J., Hepburn, C., Hope, C., Hunt, A., Li, A., Markandya, A., Moss, S., Nyong, A., Tol, R.S.J., and Watkiss, P. 2005. Scoping uncertainty in the social cost of carbon. London: Defra (http://www.defra.gov.uk/environment/climatechange/carbon-cost/sei-scc/index.htm). Edenhofer O., C. Carraro, J.Kohler and M. Grubb (eds), Endogenous technological change and the ecnomics of atmospheric stabilisation, Energy Journal Special Issue 2006. Edenhofer et al (2006), Induced Technological Change: Exploring its Implications for the Economics of Atmospheric Stabilization Synthesis Report from the Innovation Modelling Comparison Project, Energy Journal. Ellis, J. and M. Bosi, (2005) Exploring Options for Sectoral Crediting Mechanisms, OECD/IEA Publication. Grubb M. (2004), The Climate Change Challenge, Carbon Trust, London, 2004. http://www.carbontrust.co.uk/Publications/publicationdetail.htm?productid=CTC502 Grubb, M., (2005), Technology Innovation and Climate Change Policy: an overview of issues and options, Keio Economic Studies, 41/2:103-132. Grubb, M., (2004) Kyoto and the Future of International Climate Change Responses: From Here to Where?, International Review for Environmental Strategies, 5/1:15-38. Grubb, M. 2001, "Who's afraid of atmospheric stabilisation? Making the link between energy resources and climate change", Energy Policy, 29:837-845. Grubb M. and K.Neuhoff (eds), Allocation and competitiveness in the European Emissions Trading System, Climate Policy, Vol.6:1 Special Issue, 2006. Grubb, M. & Ulph, D. (2002), "Energy, the Environment and Innovation", Oxford Review of Economic Policy, 18(1):92-106. Groom, B., C. Hepburn, P. Koundouri and D. Pearce (2003) Valuing the Future. World Economics 4. Heal, G. M. (1998): Valuing the future: economic theory and sustainability, Columbia University Press. Hallegatte, S. and J.-C. Hourcade, (2005), Why economic growth matters in assessing climate change impacts: illustration on extreme events, submitted to Ecological Economics. Holtsmark, B.J., and K.H. Alfsen, (2004a): PPP correction of the IPCC emission scenarios does it matter? Climatic Change, 68(1-2), pp. 11-19.

28

Holtsmark, B.J., and K.H. Alfsen, (2004b): The Use of PPP or MER in the construction of emission scenarios is more than a question of 'metrics'. Climate Policy, 4(2), pp. 205-216. Hourcade, J.C., R. Crassous and O. Sassi (2006). Endogenous Structural Change and Climate Targets: Modelling experiments with Imaclim-R, Energy Journal Special Issue, forthcoming. International Energy Agency, (2004): World Energy Outlook 2004. www.worldenergyoutlook.org. Jaffe, A. B., R. G. Newell and R. N. Stavins (2003): Technological change and the environment. pp. 461-516 in K.-G. Mler and J. R. Vincent (eds), Handbook of Environmental Economics, Elsevier Science B.V. Jones, C., (1997), Convergence revisited. Journal of Economic Growth, 2(2), pp. 131-153. Kolstad, C. D., (1996b): Fundamental irreversibilities in stock externalities. Journal of Public Economics 60(2): 221-233. Kolstad, C.D. (1996a): Fundamental irreversibilities in stock externalities, Journal of Public Economics 60(2): 221-233. Lutz, W., and W. Sanderson, (2001): The end of world population growth. Nature 412(6846), pp. 543-545. Meier, G.M., (2001): The old generation of development economists and the new. In: Frontiers of development economics. The Future in perspective. Oxford University Press, New York Mendelsohn R., W.Morrison, M.Schlesinger and N.Adronova, Country-specific market impacts from Climate Change, Climatic Change, 45 (2000), 553-569 Mendelsohn, R., W. D. Nordhaus and D. Shaw (1994) The Impact of Global Warming on Agriculture: A Ricardian Analysis. American Economic Review 84: P.753-71.

Mendelsohn, R. and L. Williams (2004) Comparing Forecasts of the Global Impacts of Climate Change. Mitigation and Adaptation Strategies for Global Change 9(4).
Nordhaus W.D. (1991): To slow or not to slow: the economics of the greenhouse effect. Economic Journal, 101(407) 920-937. Nordhaus, W.D., (2005): Notes on the use of purchasing power parity or actual market prices in global modeling systems. IPCC Seminar on Emission Scenarios. IPCC, Washington DC. Olsen (2006), National ownership in the implementation of global climate policy in Uganda, Climate Policy, Vol.5:6 Pindyck Robert S. (2000), Irreversibilities and the timing of environmental policy, Resource and Energy Economics (2000), Vol.22:233-259 Quah, D., 1993: Galton's fallacy and tests of the convergence hypothesis. Scandinavian Journal of Economics, 95(4), pp. 427-443. Quah, D.T., 1996: Convergence empirics across economies with (some) capital mobility. Journal of Economic Growth, 1(1), pp. 95-124. Riahi, K., 2005: Present and expected research activities (EMF-21) and gaps in knowledge in stabilization/emissions scenarios. IPCC Expert Meeting on Emissions Scenarios. Intergovernmental Panel on Climate Change, Washington, D.C.

29

Schellnhuber J. and H. Held, adapted from How Fragile is the Earth System?, in J.Briden, and T.Downing, T.(Eds.), Managing the Earth: the Eleventh Linacre Lectures, Univ. Press., Oxford (2002). Shrestha, R. M., (2004): Technological Implications of the Clean Development Mechanism for the Power Sector in Three Asian Countries. International Review for Environmental Strategies 5(1): 273288. Ulph, A. and D. Ulph, (1997): Global Warming, Irreversibility and Learning. The Economic Journal 107(442): 636-650. UK Treasury, (2004, The Green Book: appraisal and evaluation in central government, http://greenbook.treasury.gov.uk/ United Nations Department of Economic and Social Affairs Population Division, (2005): World population prospects : the 2004 revision. CD-Rom edition - extended dataset. United States Bureau of the Census (USBC), 2005: International data base - data updated. Data updated 4-26-2005. Vurren v., J.Weyant, F.de la Chesnaye (2006), Multigas scenarios to stabilise radiative forcing, Energy Economics, 28 : 102-120. World Bank, (2005): World development indicators - CD-ROM. Weitzman, M. L. (1998) Why the Far-Distant Future Should Be Discounted at Its Lowest Possible Rate. Journal of Environmental Economics and Management 36(3): P.201-208. Zhang, Z. and A. Baranzini, (2004): What do we know about carbon taxes? An inquiry into their impacts on competitiveness and distribution of income. Energy Policy 32 (2004): 507-518.

Zhou, Dadi (2005), We need a development model of low emissions, presentation to Exeter conference on avoiding dangerous climate change, http://www.stabilisation2005.com/day3/Zhou_Dadi.pdf

30