IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group II: Impacts, Adaptation and Vulnerability

18.4.1 Trade-offs and synergies in global-scale analysis

Analysts working on global-scale climate analyses remain apart in their formulation of the inter-relationships between adaptation and mitigation. Some consider them as substitutes and seek the optimal policy mix, while others emphasise the diversity of impacts (with little scope for adaptation in some sectors) and the asymmetry of social actors who need to mitigate versus those who need to adapt (Tol, 2005a). Yet others maintain that adaptation is the only available option for reducing climate-change impacts in the short to medium term, while the long term has a mix of adaptation and mitigation (Goklany, 2007). Note that these positions are not contradictory; they just emphasise different aspects of the same problem.

Cost-benefit analyses (CBAs) are phrased as the trade-off between mitigation costs, on the one hand, and adaptation costs and residual damages on the other. As a recent example, Nordhaus (2001) estimates the economic impact of the Kyoto-Bonn Accord with the RICE-2001 model. Without the participation of the USA, the resulting emissions path remains below the efficient reduction policy (which equates estimated marginal costs and benefits of emissions reductions) whereas the original Kyoto Protocol implied abatement that is more stringent than would be suggested by this CBA. Note that RICE-2001, like all models, has assumptions, simplifications and abstractions that affect the results. Nonetheless, this is a common finding in the cost-benefit literature, driven primarily by relatively low estimates of the marginal damage costs (Tol, 2005b). Cost-benefit models are recognised by many as sources of guidance on the magnitude and rate of optimal climate policy (for a wide range of definitions of what is ‘optimal’ see Azar, 1998; Brown, 1998; Tol, 2001, 2002; Chapter 2), while others criticise them for ignoring the sectoral (economic and social), spatial and temporal distances between those who need to mitigate versus those who need to adapt to climate change. CBA requires conversion of many different damages to a common metric through monetisation, for example, by polling people’s values of different benefits, and the use of discount rates, which is controversial over long time-scales like those of climate change but common practice for other issues. Discounting implies that long-time-scale Earth-system transitions, such as melting of ice sheets, slowdown of the thermohaline circulation or the release of methane, have small weight in a CBA and therefore tend to attach little weight to adaptation costs (see also Chapter 17).

CBA is a special form of multi-criteria analysis. In both cases, policies are judged on multiple criteria, but in CBA all are monetised, while multi-criteria analyses use a range of mathematical methods to make trade-offs explicit and resolve them. Multi-criteria analysis has relatively few applications to climate policy (e.g., Bell et al., 2003; Borges and Villavicencio, 2004), although it is more common for adaptation (e.g., the National Adaptation Programmes of Action).

The Tolerable Windows Approach (TWA) adopts a different approach to integrating mitigation and impact/adaptation concerns and deals with adaptation indirectly in the applications. The ICLIPS (Integrated assessment of CLImate Protection Strategies) model identifies fields of long-term greenhouse-gas emissions paths that prevent rates and magnitudes of climate change leading to regional or sectoral impacts without imposing excessive mitigation costs on societies, either of which stakeholders might consider unacceptable or intolerable. This ‘relaxed’ cost-benefit framework can be used to explore trade-offs between climate change or impact constraints, on the one hand, and mitigation cost limits in terms of the existence and size of long-term emissions fields, on the other hand. For any given impact constraint, increasing the acceptable consumption loss due to emissions-abatement expenditures increases the emissions field and allows higher near-term emissions but involves higher mitigation rates and costs in later decades. Conversely, for any given mitigation cost limit, increasing the tolerated level of climate impact also enlarges the emissions field and allows higher near-term emissions (Toth et al., 2002, 2003a, b). This formulation allows the exploration of side-payments for enhancing adaptation in order to tolerate impacts from larger climate change. The TWA is helpful in exploring the feasibility and implications of crucial social decisions (acceptable impacts and mitigation costs) but, unlike CBA, it does not propose an optimal policy.

Cost-effectiveness analyses (CEAs) depict a rather remote relationship between adaptation and mitigation. They implicitly assume that some sort of a global climate change target can be agreed upon that would keep all climate-change impacts at the level that can be managed via adaptation or taken as ‘acceptable losses’. Or, cost-effectiveness analyses consider a range of hypothetical targets, but remain silent on the appropriateness of these targets. Global CEAs have proliferated since the publication of the TAR (e.g., Edmonds et al., 2004). In addition to exploring least-cost strategies to stabilise CO2 concentrations, CEAs are applied to analysing the stabilisation of radiative forcing (e.g., Van Vuuren et al., 2006) and global mean temperature (Richels et al., 2004). While most analyses are deterministic in the sense that they implicitly assume that we know the true state of the world, there is also a body of literature that models the ‘act, then learn, then act again’ nature of the decision problem, but primarily for mitigation decisions. See the WGIII AR4 Chapter 3 for details (Fisher et al., 2007).

The competition of adaptation measures, mitigation measures and non-climate policies for a finite budget has not been studied in much detail. Schelling (1995) questions whether the money that developed countries’ governments plan to spend on greenhouse-gas emissions reduction, ostensibly to the benefit of the children and grandchildren of the people in developing countries, cannot be spent to greater benefit. As a partial answer to that question, Tol (2005c) concluded that development aid is a better mechanism to reduce climate-change impacts on infectious disease (e.g., malaria, the best-studied health impact) than is emissions abatement. This analysis implies that the concern about increases in these infectious diseases is not a valid argument for greenhouse-gas emissions reduction (there are of course other arguments for abatement). The same study also shows that this result does not carry over to other impacts. More broadly, Goklany (2003, 2005) shows that the contribution of climate change to hunger, malaria, coastal flooding and water stress (as measured by the population at risk for these hazards) is usually small compared with the contribution of non-climate-change-related factors. He argues that, through the 2080s at least, efforts to reduce vulnerability would be far more cost-effective in reducing these problems than would any mitigation scheme. Other studies estimate the change in vulnerability to climate change due to emissions abatement; for instance, a shift to wind and water power or biofuels would reduce carbon dioxide emissions, but increase exposure to the weather and climate (e.g., Dang et al., 2003).

Some studies estimate the change in greenhouse-gas emissions due to adaptation to the impacts of climate change (Berrittella et al., 2006, for tourism; Bosello et al., 2006, for health). They find that emissions increase in some places and some sectors (making mitigation harder), and decrease elsewhere (making mitigation easier). The disaggregated effects are small compared with the projected growth in emissions, while the net effect is negligible. Similarly, Fankhauser and Tol (2005) show that the impact of climate change on the growth of the economy and greenhouse-gas emissions is small compared with the economy as a whole and because economic adjustment processes would dampen the impact. Note that they only include those climate-change impacts that affect economic performance; they do not use monetisation techniques. Fisher et al. (2006) reach a similar conclusion for population projections, because the net increase in mortality is small. As there are so few studies, focusing on a few sectors only, these conclusions are preliminary.

Although some industries (e.g., wind farm and solar panel manufacturing) may benefit, emissions reduction is likely to slow economic growth, but this effect is probably small if smart abatement policies are used (Weyant, 2004; Barker et al., 2007; Fisher et al., 2007). However, small economic losses in the member states of the Organisation for Economic Co-operation and Development (OECD) may be amplified in poor exporters of primary products (i.e., many African countries). Tol and Dowlatabadi (2001) use this mechanism to demonstrate an interesting trade-off between adaptation and mitigation. Taking malaria as a climate-related disease, they observe that countries with an average annual income per capita of US$3,000 or more do not report significant deaths from malaria and that all world regions surpass this threshold by 2085 in most IPCC IS92 scenarios (IPCC, 1992). Progressively more ambitious emissions reductions in OECD countries gradually decrease the cumulative malaria mortality if one considers only the impact side; that is, the biophysical effects of climate-change mitigation on malaria prevalence. However, if the economic effects of mitigation efforts (i.e., the slower rate of economic growth) are also taken into account, then, according to the FUND model, the malaria-mortality improvements due to slower global warming will be gradually eliminated and eventually surpassed by the losses due to the reduced rate of income growth, unless health care expenditures are decoupled from economic growth. Note that FUND has somewhat high costs of emissions reduction (see the SAR), and also assumes a large impact of slowed growth in the OECD on the rest of the world. Barker et al. (2002), Weyant (2004), Edenhofer et al. (2006), Köhler et al. (2006) and Van Vuuren et al. (2006) show that there is a wide range of estimates of mitigation impacts on economic growth, but these studies did not explore the link between mitigation and vulnerability. In fact, the impact of mitigation on adaptive capacity has not been studied with any other model. More generally, the capacity to adapt to climate change is related to development status, although the two are not the same (Yohe and Tol, 2002; Tompkins and Adger, 2005). The earlier studies used ‘adaptive capacity’ and ‘development’ in a generic and broad sense. Tol and Yohe (2006) use more specific indicators of adaptive capacity and development without changing the general conclusion. Emissions reduction policies that hamper development would increase vulnerability and could increase impacts (Tol and Yohe, 2006). Based on this contingency, Goklany (2000b) argues that aggressive mitigation would fall foul of the precautionary principle.

The literature assessed in this sub-section indicates that initial studies tended to focus on the relationship between mitigation and damages avoided, but our knowledge of this subject is still limited and more research needs to be undertaken. More recently, the literature has begun to focus on the relationship between adaptation and damages avoided. Ultimately, better knowledge about the interaction between adaptation and mitigation actions in terms of damages avoided would be useful. However, such research is at a very rudimentary stage. Moreover, large-scale modelling of adaptation-mitigation feedbacks is needed but still lacking. A necessary first step will be improved modelling of feedbacks from impacts, which is currently immature in most long-term global integrated assessment modelling. Adaptation modelling can follow with modelling structures that permit the reallocation of production factors and budgets in response to the changing climate. The adaptation responses therefore redefine the circumstances for mitigation. However, current impact modelling capability is undeveloped and modelling of adaptation responses to climate-change impacts has only just begun. In the above assessment we do not distinguish adaptation by actors (e.g., individuals, government departments) as the conclusions generally hold for all types of adaptation.