2Decision SupportMethods for ClimateChange Adaptation

Cost–EffectivenessAnalysis

Summary of Methods and Case Study Examples from the MEDIATION Project

• There is increasing interest in the appraisal ofoptions, as adaptation moves from theory topractice. In response, a number of existingand new decision support tools are beingconsidered, including methods that addressuncertainty.

• The FP7 MEDIATION project has undertaken adetailed review of these tools, and has testedthem in a series of case studies. It hasassessed their applicability for adaptation andanalysed how they consider uncertainty. Thefindings have been used to provideinformation and guidance for the MEDIATIONAdaptation Platform and are summarised in aset of policy briefing notes.

• One of the tools that has been widely appliedto climate change mitigation, and is alsobeing considered for adaptation, is Cost-Effectiveness Analysis (CEA).

• CEA can be used to compare and rank therelative attractiveness of different options, andto identify the least cost combination ofoptions to achieve pre-defined targets usingcost curves.

• CEA has been widely used in climate changemitigation. However, the MEDITION reviewhighlights the application to adaptation ismuch more challenging. This is becauseadaptation is a response to many local,regional or national level impacts, rather thanto a single global burden, i.e. there is no singlecommon metric. The application of CEA toadaptation is therefore much moredemanding, in terms of analysis detail andresources.

• A key issue for CEA is the choice of cost-effectiveness metrics. The MEDIATION reviewhas identified a set of potential metrics bysector, however, it is stressed it can bedifficult to pick these, especially whenconsidering complex or cross-sectoral risks.

• Further, most applications of CEA do notconsider uncertainty, working with single costcurves. However, the use of central estimatesfor future climate change is notrecommended. For adaptation, uncertainty

has the potential to alter the ranking ofoptions and the overall cost curve, and thusneeds consideration.

• The review has considered the strengths andweakness of the approach for adaptation.

• The key strength of CEA is that it avoidsvaluation of economic benefits, and can thusbe used where valuation is difficult orcontentious (e.g. ecosystems). The approachis also relatively easy to apply, and the resultsare concise and easy to understand.

• The potential weaknesses relate to the use ofa single common metric and the considerationof uncertainty, both critical issues foradaptation. Further, CEA tends to focus ontechnical options, and often omits capacitybuilding and non-technical options, while thelinear sequencing adopted contradicts theadaptation focus on portfolios of options andinter-linkages.

• Previous applications of CEA to adaptationhave been reviewed, and a case study ispresented for the biodiversity in Finland.

• The review and case studies provide usefulinformation on the types of adaptationproblem types where CEA might beappropriate, as well as data needs, resourcerequirements and good practice lessons.

• The review identifies CEA is most useful fornear-term assessment, particularly foridentifying low and no regret options, in areaswhere monetary valuation is difficult. It is mostapplicable where there is a clear headlineindicator and where climate uncertainty is low.

• A number of good practice lessons arehighlighted. The most important of these areto ensure that adaptation CEA does not focusonly on technical options, and that itconsiders uncertainty through multiple costcurve analysis. Furthermore, the need toconsider all attributes of options ishighlighted. Finally, it is considered goodpractice to undertake CEA within an iterativeplan, to capture enabling steps, portfolios andinter-linkages, rather than using the outputsas a simple technical prioritisation.

Key Messages

IntroductionThere is increasing policy interest in the appraisalof options, as adaptation moves from theory topractice. At the same time, it is recognised thatthe appraisal of climate change adaptationinvolves a number of major challenges,particularly the consideration of uncertainty. Inresponse, a number of existing and new decisionsupport tools are being considered for adaptation.

The European Commission FP7 fundedMEDIATION project (Methodology for EffectiveDecision-making on Impacts and AdaptaTION) islooking at adaptation decision support tools, inline with its objectives to advance the analysis ofimpacts, vulnerability and adaptation, and topromote knowledge sharing through aMEDIATION Adaptation Platform (http://www.mediation-project.eu/platform/). To complementthe information on the Platform, a series of PolicyBriefing Notes have been produced on DecisionSupport Methods for Climate Change Adaptation.

An overview of all the decision support toolsreviewed is provided in Policy Briefing Note 1:Method Overview, which summarises eachmethod, discusses the potential relevance foradaptation and provides guidance on theirpotential applicability. The methods consideredinclude existing appraisal tools (cost-benefitanalysis, cost-effectiveness analysis and multi-criteria analysis), as well as techniques that morefully address uncertainty (real options analysis,robust decision making, portfolio analysis anditerative risk (adaptive) management). It alsoincludes complementary tools that can assist inadaptation assessment, including analyticalhierarchic processes, social network analysisand adaptation turning points. Additional

information on each method is presented in aseparate Policy Briefing Notes (2 – 10).

This Policy Brief (Note 2) provides a summary ofcost-effectiveness analysis. It provides a briefsynthesis of the approach, its strengths andweaknesses, the relevance for adaptation, how itconsiders uncertainty, and presents case studyexamples. It is stressed that this note onlyprovides an overview: more detailed information isavailable in MEDIATION deliverables, and sourcesand links on the MEDIATION Adaptation Platform.

Description of the MethodCost-Effectiveness Analysis (CEA) is a widelyused decision support tool. It comparesalternative options for achieving similar outputs(or objectives). In this regard it is a relativemeasure, providing comparative informationbetween choices. It has been widely used inenvironmental policy analysis, because it avoidsmonetary valuation of benefits, and insteadquantifies benefits in physical terms.

At the technical or project level, CEA can beused to compare and rank alternative options. Itdoes this by assessing options in terms of thecost per unit of benefit delivered, e.g. cost pertonne of pollution abated. This identifies thoseoptions that deliver highest benefit for lowestcost (i.e. the most cost-effective). As well asranking different options, such an analysis canbe used for benchmarking, see box.

At the project, policy or programme level, wherecombinations of options are needed, CEA can beused to assess the most cost-effective order ofoptions, and so identify the least-cost path forachieving pre-defined policy targets. This is

Cost-Effectiveness Analysis

1

Box 1. Benchmarking using CEA

Cost-effectiveness analysis can be used to benchmark options, i.e. by setting thresholds. Thisapproach is often used in considering new treatments in health service provision (e.g. in the UK),where the clinic effectiveness of new interventions are compared against a cost-effectivenessthreshold, measured as the cost (£) per Quality Adjusted Life Year (QALY). New treatments or drugsare considered cost-effective if they are lower than £20 000 to £30 000 per QALY (NICE, 2010). Suchan approach is needed because publicly funded healthcare systems cannot pay for every newmedical treatment that becomes available. As there are limited resources, choices have to be made,cost-effectiveness analysis helps provide the largest benefits with the available resources.

undertaken through the use of marginalabatement cost (MAC) curves. The approach canalso identify the largest benefits possible with theavailable resources, and can even be used tohelp set targets, by selecting the point wherecost-effectiveness falls significantly (i.e. wherethere are disproportionately high costs for lowbenefits).

Cost-effectiveness analysis has been widelyused in European and Member State policyappraisal, and was previously the main approachused for air quality policy (Watkiss et al, 2007).Air quality concentration or deposition targetlevels were set on a scientific basis, and thecosts of alternative ways of achieving thesetargets (or progressing towards them) wereassessed using CEA.

It has also been used in risk-based floodprotection assessment, particularly for coastalzones (e.g. RIVM, 2004) assessing the cost-effectiveness of achieving flood protectiontargets (defined as a level of acceptable risk,such as protection against a 1 in 10000 yearreturn period).

Most recently, cost-effectiveness analysis hasbecome the main appraisal technique used for

climate change mitigation, as it allows thecomparison and ranking of greenhouse gas(GHG) abatement options, using the cost-effectiveness metric of cost per tonne of GHGabated (€/tCO2). There has also beenwidespread use of marginal abatement costcurves for mitigation. These show the relativeranking of options in order of cost-effectiveness,and can be used for identifying the least costway of achieving emission reductions includingcross-economy targets (e.g. CCC, 2008).

Because of the widespread use of CEA forclimate change mitigation, many commentatorshave also highlighted its potential use for climatechange adaptation. However, the application ofCEA to adaptation involves a number of majordifferences, see Box 2.

Cost-effectiveness analysis involves a series ofcommon methodological steps.

• Establish the effectiveness criteria.

• Collate a list of options.

• Collect cost data for each option – noting thisinvolves the full costs over the lifetime of theoption, including capital and operating costs –and thus requires all values to be expressedon a common economic basis (in equivalent

DecisionSupport 2

2

Box 2. Comparing Mitigation and Adaptation

Cost-effectiveness analysis has become the primary appraisal method for mitigation. However,adaptation is very different to mitigation, for the following reasons:

Mitigation involves a single, common metric for benefits, i.e. tonnes of GHG emissions. This metricrelates to a global burden, so a reduction in a tonne of GHG emissions is treated the same,irrespective of the technology, sector or location. This means a tonne of CO2 abated from roadtransport in an urban area has the same benefit (and unit of effectiveness) as a tonne abated fromthe electricity sector in a rural location. This allows equivalent cross-economy analysis of optionsusing €/tCO2.

In contrast, adaptation is a response to a local, regional or national level impact, rather than to aglobal burden, and involves different types of risks in different sectors. There is therefore nocommon single metric which allows cross-sectoral cost-effectiveness analysis. Furthermore, thereare often many risks even within a single sector, which can make even sectoral CEA studieschallenging: for example, adaptation to sea level rise (SLR) can involve protecting people from floodrisk, reducing coastal erosion, conserving coastal ecosystems, etc. all of which involve differentmetrics. Finally, the analysis of impacts in adaptation (rather than burdens in mitigation) means thattechnology, location and time period are important: for adaptation, it makes a different how, whereand when cost-effectiveness is assessed.

3

terms using discount rates and either anequivalent annualised cost or a total presentvalue).

• Assess the potential benefits (effectiveness) ofeach option. In many but not all, these areexpressed as an annual benefit, relative to abaseline or reference case.

• Combine these to estimate the cost-effectiveness, by dividing the lifetime cost bythe lifetime benefit (or annualised costs byannualised benefit).

At this point, all the options can be expressed inequivalent terms, as a cost per unit ofeffectiveness. This allows the ranking ofmeasures, identifying the most cost-effectiveness options, i.e. those that deliver highbenefits for low costs.

This information can then be used as an input toform a marginal abatement cost curve. Costcurves have been used for many decades inpolicy (and mitigation) analysis. In graphicalterms, they are often presented as cumulativebar charts.

In simple terms, a cost curve presents all optionsin order of unit cost-effectiveness analysis,starting with the most cost-effective. At the sametime, it also assesses the total cumulativeeffectiveness of each option, as it is added.When considered together, this allows theestimation of the least-cost way to achieve aplan, programme or policy target /objective.

An example is included below, showing a typicalmitigation cost curve. Each bar represents aspecific option. The options are arranged inorder of cost-effectiveness (left to right), asmeasured on the vertical axis by the cost per unitof abatement – €/CO2. The width of each barindicates the total abatement potential of eachoption (i.e. the total effectiveness, in tonnes ofCO2) – noting this could be for a local plan or anational level analysis. Wider bars show optionsthat can achieve larger total benefits, i.e. whichreduce more emissions. In the example, themarginal abatement costs of some cost-effectiveoptions are negative, showing these achievebenefits at negative cost, so called no regretoptions (e.g. energy efficiency).

To undertake a policy CEA, a target level of totaleffectiveness is first set and the cost curve isgenerated. As the graph presents options inorder of cost-effectiveness, it estimates thecumulative least-cost pathway to achieve thetarget, because it implements those options thathave high benefits for low cost first. By contrast,if the least cost-effective options wereimplemented first (those on the right of thefigure), it would cost far more to achieve thesame target level.

The combination of options needed to achieve thetarget can thus be read off the graph. A similarapproach can be used to derive the total costs ofdifferent levels of ambition. In practice, CEA ismore involved, and further checks are needed toensure that options can be implemented together,and to consider other criteria.

The Application to AdaptationCost-effectiveness is already used in manysectors that are relevant to adaptation, such ashealth and flooding, and therefore has potentialfor appraising options to address future climatechange. The MEDIATION project has reviewedthe application of CEA to adaptation, includingexisting case studies in the academic and greyliterature.

The first issue that the review has identified is thechoice of cost-effectiveness metrics foradaptation, and the related sector policyobjectives. This recognises there are a widerange of potential risks, across and betweensectors that could be considered. As part of thereview, MEDIATION has identified possible bysector, presented in Table 1.

Cost-Effectiveness Analysis

Emissions potential (cumulative tCO2)

Example cost curve.

Abatem

entcost(€/tCO

2)

DecisionSupport 2

4

Table 1. Possible cost effectiveness metrics / objectives for adaptation.

Sources used in compiling the table: Nicholls et al (2006); Boyd et al, 2006: Rosenzweig and Tubiello (2007): Watkiss, Huntand Markandya (2009); UNFCCC (2009).

Sector Possible Metric

Health Cost per DALY, cost per fatality or cost per lifeyear saved (impact metrics).

Health thresholds (maximum occupationaltemperatures, comfort levels)

Sea level rise Cost per reduction in land area at risk ornumber of people at risk (exposure metric) ore xpected annual damages (economic metric)

Cost per ha. For the measure relative to valueof land protected per ha (impact metric).

Pre-defined acceptable risks of flooding asobjective / threshold level for adaptation.

Floods As above.

Agriculture Impact based metrics include cost per unit ofcrop yield, production or land value, butdepends on risk (e.g. could be reduction inwater stress).

Possible headline indicator is cost per changein value added as a result of adaptationmeasures.

Water Impact metrics for water availabilityresources (household) and cost per M3 of water

provided.

Possible thresholds in terms of environmentalquality (Directives) or acceptable flows.Possible thresholds for risk of supplydisruption.

Ecosystems Critical targets (sustainable levels) andand standards (overall objective).Biodiversity

Possible cost per unit of ecosystem services.

Business Possible headline indicator is cost per& industry change in value added as a result of

adaptation measures. Could also includeacceptable risk levels for infrastructure orservice supply.

Extreme Possible metric in terms of cost per level ofEvents risk reduction, or pre-defined acceptableincluding to levels of risks as objectiveinfrastructure

Issues

Different cost per life year used acrossEurope.

Consistency issues with other sectorswhere health a part of wider risks (e.g.floods, transport)

Land area and ha only covers a sub-setof SLR impacts. Issue of non-marketvalues, loss of biodiversity andecosystem services.

Very different levels of acceptable riskand protection across Member States

As above.

Issue of capturing wider environmentaland multi-functionality of agriculture.

Highly aggregated and only one elementof potential impacts.

Issues with wider attributes of waterincluding quality (environmental).

Issue of multi-functionality and multipleusers and sectors (agriculture, industry,etc.)

Issue if standards are available (andcomplex and contentious to set).

Broad nature of sector and potential risk.

Very different levels of acceptable riskand protection across Member StatesVariability in risk acceptability acrossdifferent extremes, and for differentinfrastructure.

This does highlight a particular issue for theapplication of CEA to adaptation, namely that itcan be often be difficult to identify a singlecommon metric for analysis, because there aremany types of risks across and even betweensectors. In the case of sea level rise for example,using a headline metric of the number of peopleat risk, or an objective of acceptable levels ofrisk, will omit consideration of coastal erosionand coastal ecosystems. This means such a CEAwill not consider all relevant costs and benefitsfor coastal adaptation and may not identify themost holistic option. For this reason, CEA is lesssuitable for complex or cross-sectoral risks.

The second key area identified in the applicationof CEA to adaptation is the consideration ofuncertainty, one of the key areas of investigationin the MEDIATION review.

Most previous CEA applications, e.g. in areassuch as environmental policy and mitigation,ignore uncertainty, presenting single cost curves.Early applications of CEA to adaptation havealso followed this approach, largely presentingindividual cost curves, or at best, a few costcurves (each representing a central estimate fora different emission scenario, or a central andhigh scenario). As highlighted in Policy Note 1,the use of central estimates for future climatechange assessments can often providemisleading results for adaptation.

Indeed, the range of climate model outputs for agiven scenario, whether from the degree oftemperature change, or for precipitationprojections where even the sign of the change isuncertain, will alter the cost-effectiveness ofoptions, their relative CEA ranking, and their totaleffectiveness and the cost curve. It is possible toaddress this by sampling across multiplescenarios/model outputs, or using stochasticapproaches, but this has resource implications.

Strengths and WeaknessesA key part of the MEDIATION project has been toidentify the strengths and weaknesses ofdifferent approaches.

The key strength of CEA is that it avoidsvaluation of economic benefits, enhancingapplicability where valuation is difficult orcontentious (e.g. ecosystems). The approach isalso relatively simple to apply, and thecommunication of results is concise and easy tounderstand – helped by the widespread use ofCEA in mitigation.

The potential weaknesses relate to the need tochoose a single common cost-effectivenessmetric and the consideration of uncertainty (seeprevious section). These are both critical issuesfor adaptation. It is also highlighted that theanalysis of adaptation benefits (effectiveness) is

Cost-Effectiveness Analysis

5

Key strengths

Benefits expressed in physical terms, thereforedoes not require monetary valuation of benefits.Increases applicability to non-market sectors.

Relatively simple approach to apply and provideseasily understandable ranking and outputs thateasy to understand.

Frequently used for mitigation, and thus approachwidely recognised and has resonance with policymakers.

Use of cost curves can assess different policytargets and how to achieve these at least cost,look at how to achieve greatest benefits foravailable resources, or look at the costimplications of progressively more ambitiouspolicies.

Potential weaknesses

Optimises to a single metric, which can bedifficult to pick. Less applicable for cross-sectoralor complex risks.

The focus on a single metric omits importantrisks, and does not capture all costs and benefits(attributes) for option appraisal.

Tends to work best with technical options, andcan therefore omit or give lower priority tocapacity building and soft (non-technical)measures. Sequential nature of cost curvesignores portfolios of options and inter-linkages.

Does not lend itself to the consideration ofuncertainty and adaptive management, tending towork with central tendency.

location and technology specific, often requiringanalysis of (local) impacts, which change foreach baseline and for each scenario considered.These impacts – and the resulting benefits ofoptions – also vary over time, and thus multiplecost curves are needed to address different timeperiods. All of this means that the application ofCEA to adaptation has much higher resourceneeds for adaptation than for mitigation.

Finally, CEA tends to focus on technical options,because these can be easily assessed in termsof costs and benefits (effectiveness). However,adaptation is now seen as a process as well asan outcome, and capacity building and non-technical (soft) options are considered animportant and early priority. Such non-technicaloptions do not lend themselves easily to thequantitative analysis in CEA, thus they tend to be

given lower priorities (or omitted). This issue iscompounded by the strict linear sequencingadopted in cost-effectiveness analysis, whereoptions are considered as discrete optionsimplemented in turn: this contradicts theemphasis in the adaptation literature forportfolios of options and the need to explicitlyconsider inter-linkages.

A summary of some of the key strengths andweakness of the approach is presented below.

Case StudiesThe MEDIATION study has applied CEA to anumber of adaptation case studies, as well asreviewing existing literature examples. A numberof these case studies are summarised in the boxbelow.

DecisionSupport 2

6

Box 3. Case Studies

Boyd et al (2006) undertook a detailed application of cost-effectiveness analysis for the South-Eastof England, looking at the impact of climate change (including potential scenarios of reducedprecipitation) and socio-economic growth (increased demand) on water resource zones and thepotential adaptation response to address household water deficits. The study undertook detailedbasin modelling for the water catchment (Wade et al, 2006) and assessed baseline 30-year averagehousehold water deficits in three future time periods (2011-2040, 2041-2070 and 2071-2100) for fourseparate climate-socioeconomic scenarios.

The cost of addressing the projected water deficits was analysed through a cost-effectivenessanalysis, looking at a range of options for managing public water supply (including options thatreduced demand and options that increased supply). Detailed cost-yield curves (cost-effectivenesscurves) were produced to estimate how to eliminate the household water deficit at minimum cost,providing cost curves for each scenario, for each of the three future time periods in an inter-linkedanalysis. This addresses many of the issues raised above, by working with multiple projections andmultiple time periods. An example of one of the cost curves is presented below.

Source: Boyd et al (2006)

Cost-Effectiveness Analysis

7

Case Study 2: Cost-effectiveness analysis for biodiversity in Finland (MEDIATION)

Semi-natural grasslands and wooded pastures are among the most species-rich habitats in Finlandand include numerous red-listed species. However, intensification of agriculture and abandonment oftraditional management practices have reduced the area of valuable grasslands and their biodiversity.Climate change is projected to cause additional challenges, as species may need to shift their rangesto follow the changes in climatically suitable areas. However, the success of species moving to newareas depends on their dispersal ability and the availability of suitable habitats.

The Mediation case study investigated thee impacts of climate change on grassland biodiversitythrough a case study on grassland butterflies, which are a key indicator species group as well ashaving a high amenity value. The study compared the sufficiency of habitat in the areas projected tobe climatically suitable in the future or where ecological corridors might be constructed to aiddispersal, and then considered alternative adaptation options to enhance grassland biodiversity inFinland under a changing climate.

The analysis first looked at future bio-climatic envelopes to explore which areas would besimultaneously suitable for various species, considering a range of climate outputs and a series ofdifferent modelling methods to explore their uncertainty. The model predictions show large variationsin the suitable areas for many species, (see figure below), as a consequence of different modellingmethods and climate scenarios being used. However, a common finding is that the current extent ofgrassland habitats in Finland is far lower than the target level estimate required to sustainablysupport current populations, as well as to secure species dispersal.

Figure 1. Projected suitability of future climate for the Parnassiusmnemosyne butterfly averaged over the time slice of 2051-2080. Thered colour indicates the most suitable areas.

The projections are based on bioclimatic envelope models developedwith three different modelling methods (GAM, GLM, GBM) and fivedifferent climate scenarios.

The study then considered alternative conservation measures (adaptation options) which couldmaintain the biodiversity of Finnish semi-natural grasslands under a changing climate. Three majoradaptation options were considered: agri-environmental scheme (AES) measures (which are alreadyused in Finland), construction of ecological corridors, and species translocation. The results showthat management of traditional biotopes by cattle grazing is the most efficient measure for butterflies,but when costs are taken into account buffer zones appear to be the most cost-effective AESmeasure.

Cost-effectiveness of Agri-Environmental Scheme Measures.

AES Type Total PVC Effectiveness (Y) CE measure (PVC/Y)Environmental Fallow 710 0,7 1014Buffer Zone 944 1,16 858Traditional biotope 2520 2,29 1096

The analysis was complemented with a farmer survey to understand farmer attitudes to biodiversityconservation and how different factors affect farmers’ willingness to implement different AESmeasures, thereby providing useful complementary information on the attractiveness of managementoptions and the possible barriers to implementation.

Source: Tainio et al. (2013).

Discussion and ApplicabilityThe review and case studies provide a number ofpractical lessons on the application of cost-effectiveness analysis to adaptation. Theyprovide useful information on the types ofadaptation problem types where CEA might beappropriate, as well as data needs, resourcerequirements and good practice.

CEA is considered most useful for near-termadaptation assessment, particularly for identifyinglow and no regret options. The approach can beapplied to both market and non-market sectors,but it particularly relevant for areas that aredifficult to value in monetary terms, e.g.biodiversity, health. The use for long-termassessment is considered most appropriate whenused as part of an iterative adaptive managementanalysis, rather than as a tool on its own.

It is most applicable (and relevant) where there isa clear headline indicator and a dominant impact– and less applicable for cross sectoral andcomplex risks, because it works with a singlemetric. It is thus more applicable when there isalready agreement on sectoral objectives andeffectiveness criteria. It is more appropriatewhere climate uncertainty is low, and good dataexists for cost/benefit components.

The key data inputs vary on the use of the tool, i.e.whether a cost-effectiveness ranking (cost per unitbenefit) or a full policy analysis. An initial ranking ofoptions can often work with generic data onburdens to identify promising options. However, afull policy analysis usually requires some form ofscenario-based impact outputs, to assess uniteffectiveness accurately, and total effectivenessagainst a baseline. Full cost data is needed (capitaland operating costs, expressed in equivalenteconomic terms) as well as data on uniteffectiveness. For policy applications, additionalinformation is needed in the form of baseline risksand the total potential for each option.

In considering the application of CEA toadaptation, a number of good practice lessonsare highlighted:

• A good starting point for an adaptation CEA isto consider the cost-effectiveness of optionsto current climate variability, and then to

assess cost-effectiveness in a number ofdefined future periods.

• The application of CEA to adaptation willideally be context and location specific. It isimportant to identify appropriate sector andrisk specific metrics, and stakeholderconsultation can help this step.

• The application of CEA to adaptation shouldconsider non-technical options and capacitybuilding as well as technical options, notingthese are more difficult to quantify.

• The application of CEA to adaptation needs toconsider uncertainty. This should involve asampling (multiple cost curves) across a rangeof socio-economic and climate modelprojections (even if low/high ranges). The use ofsingle central estimates and single cost curvesshould be avoided. To capture the issues oftiming and dependencies, analysis is likely torequire a minimum of two future time periods.

• The CEA baseline should take account ofcurrent conditions and existing and plannedpolicy. Future baseline projections shouldconsider socio-economic as well as climatechange, and ideally autonomous adaptationand existing/planned adaptation measures.

• The analysis of inter-dependencies betweenoptions is important, i.e. how one optionmight affect another. It is also preferable toundertake CEA within an iterative plan, tocapture enabling steps and portfolios ofoptions, rather than using outputs as a simpletechnical prioritisation.

• Due to the focus on a single metric, there is aneed to assess wider attributes of options, i.e.their wider environmental, social andeconomic characteristics, as well practicality,acceptability, etc. These should alter theranking of options.

• Recent applications of CEA have tried to applythe cost curve concept to adaptation using fullcost-benefit analysis. The MEDIATION reviewdoes not recommend such an approach. Moredetails are provided in the box.

Finally, due to the widespread application of cost-effectiveness analysis in the mitigation domain,some more advanced lessons have beenidentified. These are summarised in the box.

DecisionSupport 2

8

Cost-Effectiveness Analysis

9

Box 3. More Advanced Lessons from Mitigation Cost-Effectiveness Analysis

A number of lessons of relevance for adaptation have emerged from the widespread use of CEA inmitigation.

• CEA tends to work with technical costs, omitting important policy and/or transaction costs, whichneed to be factored in when moving to policy implementation. For this reason, they underestimatethe costs of options (and overestimate the relative cost-effectiveness). These policy costs shouldbe factored into analysis.

• Cost curves can be divided into expert-based and model-derived curves. Expert-based curvesassess the cost and reduction potential of each single abatement measure, while model-derivedcurves are based on a range of partial- or general-equilibrium models. For adaptation, most initialassessments are likely to be expert based, but there may be potential for modelling in some futureareas.

• Cost-effectiveness usually optimises to one attribute, but in practice, policy options need toconsider many elements, whether expressed in monetary or non-monetary terms. There havebeen some applications of CEA which seek to build in ancillary effects, either through the use ofcost-effectiveness adjustments or through multi-optimisation analysis. These involve a stepchange in complexity and resources, but do provide much more robust results.

• A key area of discussion has centred on discount rates, and whether to use a social or privatesector discount rate. Recent examples have undertaken sensitivity analysis with both to examinewhether this alters the ranking and overall costs of compliance.

• Baseline assumptions, including the technology and reference costs (e.g. future energy prices),have a significant impact on the cost-effectiveness analysis. Such socio-economic drivers areknown to be as important as climate drivers for adaptation, and need to be taken into account inanalysis.

• Most MACC assessments have limited feedback between sectors or even time periods.Furthermore, they are defined with respect to a certain year. These issues are more important foradaptation.

• There has also been a debate around learning curves and innovation, which are important indetermining the balance of current versus future options. This is something which requiresconsideration in the adaptation domain, albeit in more complex assessments.

Building Adaptation Cost Curves Using Economic Valuation

• A number of recent assessments of adaptation have taken the marginal abatement cost curveconcept used for mitigation, but used monetary values to define effectiveness. In essence, thisjust undertakes cost-benefit analysis, but presents results so that they look like a mitigation costcurve.

• The MEDIATION project has reviewed this approach and does not recommend it for adaptation.

• This is because the approach tries to force adaptation to fit a decision framework targeted formitigation. It does not solve any of the issues raised above on cost-effectiveness analysis, i.e. ittreats adaptation as a simple linear process, focused on technical options, and most importantly,it has little consideration of uncertainty.

• Furthermore, it introduces a new problem with respect to the challenge in estimating monetaryvalues for many sectors of interest to adaptation (health, ecosystem services), as well as capacitybuilding and non-technical options, which are a priority for early adaptation.

ReferencesBoyd, R Wade, S. and Walton, H (2006) in Metroeconomica et al (2006). Climate Change Impactsand Adaptation: Cross-Regional Research Programme Project E. Final report to Defra, London, UK.

CCC (2008). Building a low-carbon economy – the UK’s contribution to tackling climate change.Committee on Climate Change December 2008. Published by the Stationary Office (TSO), London,UK.

ECA (2009). Shaping Climate-resilient Development a framework for decision-making. A report of theeconomics of climate Adaptation working group. Economics of Climate Adaptation.

Nicholls et al (2006). Metrics For Assessing The Economic Benefits Of Climate Change Policies: SeaLevel Rise. ENV/EPOC/GSP(2006)3/FINAL. OECD, 2006.

NICE (2010). National Institute for Public Health and Clinical Excellence (NICE),http://www.nice.org.uk/newsroom/features/measuringeffectivenessandcosteffectivenesstheqaly.jsp

RIVM (2004). Dutch dikes, and risk hikes. A thematic policy evaluation of risks of flooding in theNetherlands Extended summary Netherlands Environmental Assessment Agency National Institutefor Public Health and the Environment (RIVM). 2004.

Cynthia Rosenzweig and Francesco Tubiello (2007). Metrics for Assessing the Economic Benefits ofClimate Change Policies in Agriculture. ENV/EPOC/GSP(2006)12. OECD, 2007.

Tainio, R.K. Heikkinen, J. Heliölä, Leikola, N., S. Lötjönen, O. Mashkina, T. R. Carter, S. Bharwani, A.Hunt, R. Taylor and P. Watkiss (2013). Conservation of grassland butterflies in Finland under achanging climate. Mediation Case Study. Submitted to the Mediation Special Edition.

UNFCCC (2009). Potential costs and benefits of adaptation options: A review of existing literature.UNFCCC Technical Paper. F CDCeCce/mTPb/e2r0 20090/29. Available at:http://unfccc.int/resource/docs/2009/tp/02.pdf Markandya, A. and Watkiss, P. (2009).

Paul Watkiss, Alistair Hunt and Anil Markandya (2009). Common Metrics for Adaptation. Report forthe UK Defra.

Watkiss, P., Holland, M., Hurley, F., Hunt, A., and Pye, S. (2007). Assessing the Costs and Benefits ofthe European Air Pollution Policy (CAFE): Results and Lessons from Experience. p175-199. In.Economic Appraisal of Environmental Legislation. Scottish Environmental Protection Agency (SEPA).Jean Le Roux, Tshering Sherpa and Evan Williams (Eds). Published by the Stationary Office (TSO),London, UK.

Further Reading and Reference SourcesMEDIATION Policy Briefing Note 1: Method Overview.

DecisionSupport 2

10

These notes provide a brief introduction to traditionaland novel approaches to assess options to adapt toclimate change, covering their strengths andweaknesses, conditions for applicability, case studyexamples and suggestions for further reading. Thefull set can be found at http://mediation-project.eu/.

The research leading to these results has receivedfunding from the European Union SeventhFramework Programme (FP7/2007- 2013) under grantagreement n° 244012.

To find out more about the MEDIATION project,please visit:http://mediation-project.eu/

For further information on the MEDIATION project:contact Rob Swart atrob.swart@wur.nl

For further information on economic decision supporttools, contact Paul Watkisspaul_watkiss@btinternet.com

The views expressed in this publication are the soleresponsibility of the author(s) and do not necessarilyreflect the views of the European Commission. TheEuropean Community is not liable for any use madeof this information.

Citation: Watkiss, P. and Hunt, A. (2012). Cost-effectiveness analysis:: Decision Support Methodsfor Adaptation, MEDIATION Project, Briefing Note 2.Funded by the EC’s 7FWP

Copyright: MEDIATION, 2013

First published May 2013

The MEDIATION project is co-ordinated byAlterra, Stichting DLO and involves 11 teams fromacross Europe

Cover Photo:Green Roof,I-Stock. ©iStockphoto.com

Further information