55470 2 forthcoming titles in this series Beyond the sum of its Parts -- Blending financial instruments to support low-carbon Development monitoring climate Finance and ODA D E V E L O P M E N T A N D C L I M A T E C H A N G E Economic Evaluation of Climate Change Adaptation Projects Approaches for the Agricultural Sector and Beyond ii © 2010 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org/climatechange E-mail: feedback@worldbank.org All rights reserved. This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The find- ings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. RIGHTS AND PERMISSIONS The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400; fax 978-750-4470; Internet: www.copyright.com. Images by Shutterstock Images LLC and The World Bank Photo Library iii taBle of contents acKnoWleDgments vii acronyms vii executive summary ix 1 introduction -- scope and concepts underlying this Paper 1 2 challenges in evaluating adaptation initiatives in agriculture 3 2.1 Climate change and adaptation in agriculture 3 2.1.1 adaptation and no-regret investment 4 2.1.2 classification of adaptation 6 2.1.3 adapting to changes in climate variability and to medium-long term climate change 7 2.2 Assessing the Costs and Benefits of Adaptation 7 2.2.1 Dealing with uncertainty in the economic analysis of adaptation 8 2.2.2 Deciding between investing now or later 9 3 approaches and methodologies for evaluating adaptation 11 3.1 Assessing the impacts of climate change on agricultural projects 11 3.1.1 agronomic or crop models 12 3.1.2 ricardian or hedonic method 15 3.1.3 Probabilistic methods for impact assessment of extreme events 20 3.2 Evaluating costs and benefits of planned adaptations 22 iv 3.2.1 methods for assessing economic benefits and costs 22 3.2.2 a non-economic assessment approach -- multi-criteria decision analysis 30 3.2.3 Dealing with uncertainty 31 4 conclusions -- some Basic steps for Project-level economic evaluation of adaptation 39 references 45 Boxes 1 What is meant by "adaptation deficit" and "maladaptation"? 5 2 implications of climate change on food security in Bangladesh 13 3 climate change impacts in drought and flood affected areas in india 14 4 impacts of climate change in the agriculture sector in morocco 16 5 an example of new generation ricardian models 17 6 adaptation for irrigated agriculture in china 19 7 uncertainty and probability functions 21 8 estimating the rate of adoption of agricultural innovations 23 9 Deriving soft and hard adaptation costs for irrigation 24 10 eliciting adaptation cost information from local communities and institutions 25 11 calculating additional adaptation costs for irrigation modernization in china 27 12 multi-criteria priority setting for adaptation decisions in latin america 32 13 application of real option analysis to an irrigation project in mexico 35 14 rDm for adaptation decisions in the water sector 38 15 toward a more straightforward application of ricardian and crop models to project-level impact assessment 40 v figures 1 stylized schematic illustration of a soil-water-crop module (based on sWaP-Wofost models) 12 2 effect of climate change on an intensity-probability function 20 3 application of real option to an irrigation project in mexico 36 taBles 1 summary of adaptation categories by type 6 2 Disaggregated unit costs 24 3 summary of methodologies 41 vi vii acKnoWleDgments We would like to thank the members of the climate change team of the environment Department -- especially Kseniya lvovsky and ian noble -- who have helped in many ways to facilitate the preparation of this paper. Peer reviewers John nash (World Bank), Pradeep Kurukulasuriya (unDP) and urvashi narain (World Bank) have provided valuable inputs throughout the preparation of this report. in addition, the following people have made helpful suggestions and supplied additional information: Jock anderson (World Bank), Jan Bojo (World Bank), francesco Bosello (fondazione eni enrico mattei, venice), mark cackler (World Bank), raffaello cervigni (World Bank), David corderi (World Bank), svetlana edmeades (World Bank), timothy essam (World Bank), Willem Janssen (World Bank), craig m. meisner (World Bank), emanuele massetti (catholic university of milan and fondazione eni enrico mattei, milan), nicholas Perrin (World Bank), apurva sanghi (World Bank), Pasquale scandizzo (university of tor vergata, rome), Paul siegel (World Bank), William sutton (World Bank), Johannes Woelcke (World Bank) and Winston yu (World Bank). the findings, interpretations and conclusions are the authors' own and should not be attributed to the World Bank, its executive Board of Directors, or any of its member countries, or to those who provided comments on this paper. acronyms gef global environment facility ghg greenhouse gas mcDa multi-criteria decision analysis nPv net present value rDm robust decision making viii ix EXECUTIVE SUMMARY This paper identifies key challenges and solutions potential future climate change trends must be for carrying out project-level economic analysis of taken into account when development outcomes adaptation to climate change, both stand-alone depend on how the climate will change over the and integrated into broader development projects. next few decades. For example, the design of a Very few projects addressing adaptation thus far new irrigation system calls for consideration of have been subject to in-depth and rigorous the expected water availability during the lifetime economic analysis for a variety of reasons, includ- of the project; and water availability will be influ- ing a lack of guidance on how to deal with enced by, e.g., melting of glaciers threatening to assessments of the impacts of climate change, as compromise water availability in entire water- well as with estimating costs and benefits of sheds in the Andean region. Adaptation needs to adaptation under uncertainty. Our focus is on the deal with the medium- to long- term changes in agricultural sector, where the impacts of climate overall climatic conditions, as well as changes in change have the potential to disrupt the liveli- the variability of climate conditions. hoods of rural populations in many regions and where adaptation must be given urgent consider- The main challenges faced in carrying out proj- ation. Nevertheless, some of the approaches ect-level economic evaluations are briefly discussed are suitable to projects in other sectors discussed below: as well. 1. Many, if not most, of the needed investments Over the next few decades, climate change for adaptation, especially in the agricultural impacts on agriculture are likely to be felt due to sector, will also bring benefits irrespective of greater climate variability, and increased how much the climate changes. First, adap- frequency and intensity of extreme events, as well tation investments could increase resilience to as from changes in average climatic conditions. current climate variability, while preparing Individuals, communities and institutions often for a future increase in variability due to cli- make strengthening shorter-term responses to mate change. Moreover, many responses will current climate variability a priority. Nevertheless, provide benefits beyond managing climate x risks (e.g., improving water-use efficiency in 4. Decision makers have a choice about when to areas that are already water-scarce due to invest, as well as how much and in what non-climatic pressures, such as increased forms. Where investment has high co-bene- water demand from different sectors). On the fits in reducing a current adaptation deficit, other side of the spectrum, some responses the argument for more rapid investment is largely provide benefits only in the context of strengthened. More generally, however, climate change risks, such as infrastructure deciding how much to adapt now versus projects (e.g., dams and dikes) that proac- waiting to do more after gaining additional tively respond to projected changes in factors information on the impacts of climate change such as runoff and sea level rise. The latter and the options for ameliorating those must explicitly factor the uncertainty of cli- impacts is not an easy decision given the mate change, as well as the costs and benefits uncertainties discussed above. of adaptation, into the evaluation. 5. The choice of discount rate for evaluating 2. Development projects focus on public invest- future benefits and costs is often controver- ments in adaptation. Planned adaptation-- sial in many other contexts, as well as in involving action by a local, regional and/or adaptation. Debates exist on the proper rates national government to provide needed pub- of return for evaluating projects given uncer- lic goods and incentives to the private sector tainties, distortions from taxation and incor- to fit the new conditions--is therefore the rect market prices, and incomplete or poorly focus of this paper. Nevertheless, autono- functioning capital markets. A more particu- mous adaptation -- involving actions by lar concern in evaluating adaptation invest- farmers, communities and others in response ments with long time horizons (e.g., 50­100 to the threats of climate change perceived by years) is how to value the long-term benefits. them, based on a set of available technology One common but ad hoc approach is sensi- and management options--must be taken tivity analysis using a lower discount rate to into account in defining the "baseline" or see how sensitive the project evaluation "without-project" scenario. Moreover, in might be to benefits accruing only in the project evaluation, it is important to consider more distant future. Other approaches that how planned adaptation may influence the try to assess the relative benefit of a project private sector's capacity to undertake autono- in reducing long-term uncertainty for an mous adaptation. affected population should be considered, even if the valuation of such benefits can be 3. Evaluating the economic benefits of hard undertaken only heuristically. investments is relatively straightforward (although, in practice, it is not trivial) 6. For a stand-alone adaptation project, both because a direct relationship can be con- benefits and costs can be assessed relative to structed between inputs provided by the a no-project alternative. For a project with physical investment (i.e., water supply from a adaptation components undertaken within a dam) and production output. Soft adaptation, broader set of activities, the comparison on the other hand, is more complicated would be made relative to a business-as-usual because the benefits, to a great extent, must project without adaptation components. In be inferred from resulting changes in private either case, but especially in the latter case, sector behaviors and prices. there is an inherent subjectivity and need for xi expert judgment in defining the hypothetical econometrically estimated using cross-sectional alternative as a basis for comparison. data. An important characteristic of this method- ology is that the findings on longer-term climate The problem of economically evaluating adapta- change impacts are net of whatever autonomous tion to climate change at the project level can be adaptation responses to climate change individual disaggregated into two distinct subproblems, farmers are able to make over the long term. namely: Both approaches have specific strengths and weaknesses that need to be carefully considered a. Evaluating the potential impacts that climate when choosing which method to use in project change could have on agricultural productiv- evaluation. ity in the project area, assuming only autono- mous adaptation. The literature and practice in the disaster risk b. Evaluating costs and benefits of possible reduction field suggest another method for esti- planned adaptations, including the implica- mating expected economic losses due to climate tions of uncertainty with respect to the change, as well as economic benefits of adapta- choice of specific adaptation options. tion measures. This method was developed for application to natural disasters and, hence, is These assessment stages are common to the eval- immediately applicable to impacts of climatic uation of adaptation in any sector. The specific extremes (i.e., floods), although it may be possible approaches and methodologies used to deal with to adapt the approach to evaluate other impacts each subproblem, on the other hand, can be of climate change. different, depending on the sector and the specific project's characteristics. Possible method- The challenges in evaluating costs and benefits of ologies for addressing each subproblem are briefly hard and soft adaptation investments are similar summarized below. to challenges in evaluating such investments in other types of development projects. For example, For the evaluation of climate change impacts on the approaches used in the past for estimating agriculture, two approaches in particular -- the ex-ante the economic benefits of agricultural agronomic (or crop) models and the Ricardian innovations can be applied to some soft adapta- (or hedonic) models -- have become the most tions. As for adaptation costs, different methods widely used in applications to country studies and can be applied. One approach consists of piggy- projects dealing with climate change impacts and backing the costs of adaptation measures from an adaptation in agriculture. Agronomic models are in-depth analysis of documentation of past proj- biophysical representations of crop production ects that financed the same types of interven- simulating the relevant soil-plant-atmospheric tions, which would be needed for adaptation components that determine plant growth and purposes (i.e., irrigation, agricultural extension, yield. They can be used to assess the impacts of flood protection, etc.). Another possible approach climate change on agricultural productivity, as is based on the solicitation of information directly well as to investigate the potential effects of from the local communities that are vulnerable to different adaptation options. The Ricardian climatic risks and that take adaptation-relevant method is based on the idea that the long-term decisions. productivity of land is reflected in the land's asset value. The impacts of different influences on land In the case of no-regret adaptation investments value, including climatic differences, are and broader development projects that fully xii integrate adaptation into their design, isolating evaluate alternatives across a range of different the costs and benefits of the adaptation compo- and potentially incommensurate criteria. This is nent might not be feasible, as such decisions are especially true in the context of agriculture and also simultaneously conditioned by a whole range climate change where an adaptation project can of other factors. While it might be possible in help reduce the negative effects of climate change principle to consider a hypothetical alternative on a number of social and environmental, as well project designed with less adaptation integrated as economic, indicators. There also may be many into it, such an effort would have little meaning instances, as already noted, when information on and it will be more valuable to compare alterna- the monetary value of potential benefits or their tive project designs per se. For stand-alone adap- likelihood of being realized is scarce and signifi- tation projects or projects with a distinct cant amounts of informed judgment must be adaptation component included, additionality of substituted. In such cases, multi-criteria decision costs and benefits of adaptation may be useful to aiding approaches can be useful. estimate in some cases. In particular, this can be important when there are alternative projects or Economic evaluation with uncertainty usually component designs with different benefits and takes the form of considering certain scenarios costs that can then be compared. One can also judged to have various degrees of likelihood. attempt to indirectly identify the costs of an More sophisticated extensions of this approach adaptation activity linked to an existing develop- postulate more explicit probability distributions ment project through a "gap analysis" to pin down for key factors. For some adaptation initiatives, which additional investments are needed in order especially when a main focus of concern is with to increase its resilience to climate change by a the impacts from climatic extremes, it may be certain degree. possible to economically evaluate how the project reduces the risks and expected monetary losses The presence of co-benefits in adaptation projects associated with an uncertain adverse agricultural is particularly important in the economic evalua- impact. tion if they otherwise would not be reflected in the project appraisal. This would typically be the Another possible approach is "real option analy- case if the co-benefits have the nature of public sis," which reflects the state of the art in goods. For example, where investment in economic evaluation under uncertainty but, thus improved water management for adaptation in far, remains difficult to apply in concrete cases. agriculture also conveys benefits for other catego- Real option analysis is based on the idea that ries of users (e.g., municipalities), estimates of some real investment projects can be evaluated as these benefits can be included and strengthen the a set of compound options. For example, a water overall case for the project. These co-benefits can, management project may help a community at least in some cases, be quantified and would preserve the option of remaining in place rather increase the overall economic attractiveness of the than migrating if future climate change makes adaptation investments. local livelihoods infeasible. Evaluating a project through this approach can be considered a new Alternatives to economic approaches for project form of risk analysis, where risk is identified both evaluation exist, which may allow bypassing some positively, as the contingent wealth of opportuni- of the specific challenges of an economic evalua- ties created by the project, and as a cost, in terms tion. Often, decision makers need or want to of contingent liabilities the project may generate. xiii Finally, robust decision making (RDM) can Although time, budget and data limitations provide an alternative quantitative decision constitute obvious constraints in using the meth- analytic method that avoids subjective probability ods discussed, a good reason for investing in more assessments and scenario predictions. RDM in-depth economic evaluation of adaptation is creates hundreds or thousands of plausible that it can be very useful to inform project design futures, in the judgment of the analyst, that are (i.e., to select the crops most suitable to the local then used to systematically evaluate the perfor- climate conditions, or to design project compo- mance of alternative actions. This approach facili- nents that are likely to maximize benefits for tates identifying the set of conditions under local communities according to their own judg- which any particular alternative adaptation ment). Moreover, despite the complexity of these performs well or poorly, according to various approaches, options exist for employing simplified evaluation criteria based on the decision maker's versions of some methodologies for project-level judgment. The decision maker can identify analysis. A series of steps for carrying out the "robust" alternatives that, compared to other economic analysis of an adaptation project, as alternatives, perform reasonably well across a wide well as a summary table of the methods discussed, range of plausible futures. can be useful tools for project teams. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 1 1. InTRodUCTIon SCopE And ConCEpTS UndERlYIng ThIS pApER The economics of adaptation has become a hot broader development projects. While our focus is topic over the past few years, since the adverse on the agricultural sector, we also highlight some impacts of climate change are raising important general approaches that are suitable to projects in concerns about the future livelihoods of many other sectors as well. We concentrate on assess- people around the world. In the very near term, ing adaptation at the level of specific projects, as vulnerable communities will need to accelerate opposed to sector-level or economy-wide assess- adaptation in order to mitigate the additional ments of adaptation potential encountered in the burdens of climate change. This is especially research and policy literatures (IPCC 2007 important in the context of agriculture, given the provides a comprehensive review of the climate critical role of that sector in the livelihoods of change impacts and adaptation literature, includ- populations throughout the developing world. ing for agriculture). At the same time, investments in adaptation For our purposes, adaptation projects are activities compete with other development priorities. undertaken to ameliorate anticipated or actual Economic evaluation of adaptation options can losses in output and/or increases in cost of agri- provide decision makers with important informa- cultural production as a consequence of climate tion for evaluating alternative uses of scarce change. Our particular emphasis here is on antic- resources, as well as on when and how to make ipatory adaptation, though the same basic adaptation investments. Unfortunately, very few concepts can also be applied to coping measures adaptation projects or project components thus taken after adverse impacts are realized. The far have been subject to in-depth and rigorous climate change drivers of the adverse impacts on economic analysis that would contribute to output or cost include both changes in longer- weighing these trade-offs. term conditions (average temperature, rainfall) and increased variability of climatic conditions. This paper identifies key challenges and solutions This scope does not include investments to raise for carrying out economic analyses of adaptation productivity under existing climatic conditions or projects and adaptation components within to increase resilience to existing climatic 2 D e v e l o P m e n t a n D c l i m at e c h a n g e variability, though in practice many of the poten- economic analyses, which also include "satellite tial activities will be the same (see also Box 1). 1 assessments" of other indicators. We focus on economic analysis as a means for We are concerned here with adaptation initia- assessing the benefits and costs of investments in tives whose outcomes have the attributes of adaptation, as distinct from financial analysis of "public goods" in varying degrees. These can "additionality" in adaptation costs vis a vis "busi- flow from investment in physical infrastructure ness as usual," e.g., for accessing dedicated adap- and natural capital ("hard" adaptation efforts, tation financing sources. Still, we offer basic such as irrigation and land terracing), as well as suggestions on how to approach additionality of in human capital ("soft" adaptation, including adaptation costs and benefits. developing knowledge and skills and institu- tional strengthening for responding to a chang- The emphasis on assessing benefits and costs in ing climate). Each type of investment presents project evaluation may invoke a perception of a different challenges in assessing potential narrowly focused economic analysis of aggregated impacts and valuing benefits. Social and knowl- net economic benefits over time. In principle, edge investments generate benefits through the however, the ideas we are addressing can be way they change the actions of individuals applied more broadly (see also Heltberg and throughout the sector. Thus, the value of such others 2009). Thinking of adaptation benefits in investments must be inferred by attempting to the context of reduced vulnerability, benefits can project and evaluate the economic gains from be enumerated in several ways--reduced food these behavior changes. The benefits of invest- insecurity, greater capacity to maintain diversified ment in physical infrastructure flow more assets, less stress on social relationships, reduced directly from its use; one of the key challenges, dread--not all of which reduce so readily into in this case, is evaluating long-term benefits monetary equivalents. Benefits can also be from infrastructure investments.2 assessed in terms of mitigating adverse distribu- tional impacts of climate change. That said, we A number of environmental, technical and imagine that the most immediate application of economic uncertainties, which need to be the ideas discussed would be in more traditional factored into economic analysis of adaptation activities, loom over these considerations. While this paper does not provide detailed descriptions or guidance on specific techniques for addressing 1 terminology in the literature on adaptation (and related literature such as disaster risk management) is not well these uncertainties, it presents and discusses standardized, which can be a source of confusion. possible approaches for addressing them and heltberg and others (2009) construct a "risk-vulnerabili- ty" chain for social risk management generally and provides references for obtaining more detailed show how it applies to climate change adaptation. in information. their framework, risk is the chance of loss (which can be measured using various metrics) for households or other social units stemming from an external force like climate change. exposure to risk depends on the size and distribution of assets, the mix of strategies and 2 heltberg and others (2009) emphasize the importance activities for livelihoods, and external shaping influenc- of a broad asset-based approach to adaptation that es (government policies, cultural influences). expected more systematically formalizes the ideas presented losses, after taking into account ex-ante and ex-post here. they argue that household well-being depends risk management strategies, depend on risk, exposure, on both the assets available to the household, broadly and the nature and effects of risk management strate- defined, and the livelihood strategies that reflect use of gies taken. in this context, climate change adaptation is these assets. assets in turn can be broken down into a risk management strategy. as noted in section 2, standard measures of: physically accumulated wealth; adaptation can be further divided into autonomous knowledge and human capital; natural assets, including activities undertaken by households and other social ongoing benefits derived from being in a particular units, and planned activities undertaken at a more col- location; and those related to social and political institu- lective level by governments. tions. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 3 2. ChAllEngES In EVAlUATIng AdApTATIon InITIATIVES In AgRICUlTURE 2.1 climate change dry sub-humid regions in the developing world, where high rainfall variability and recurrent anD aDaPtation in droughts and floods regularly disrupt food agriculture production, and where poverty is pervasive. Only a few regions, including northern China, Eastern Since the challenges of climate change for agri- Europe, northern North America, and the culture have already been extensively documented, Southern Cone of South America, might benefit we provide only a quick summary here (see from a poleward shift in agriculture under a Padgham 2009 for more details). During the last limited degree of future warming. Other areas several decades, we have seen higher average may benefit, at least for a time, from the carbon temperatures across the globe, an increased occur- fertilization3 effect, which could compensate rence of heavy rainfall events and floods, and negative impacts on yields due to temperature longer and more intense droughts in many increases and changes in rainfall (see Cline 2007). regions of the world. These occurrences have often led to reduced crop yield levels and disrup- The risks that climate change poses for agricul- tions in agricultural production, especially in the ture are both direct and indirect. Potential direct most vulnerable and least prepared countries. impacts include the effects of temperature rise and changes in precipitation frequency and inten- Over the next few decades, climate change sity on crop growth. Temperature rise alone is impacts on agriculture are likely to increase due to greater climate variability, and increased frequency and intensity of extreme events, not 3 carbon fertilization is defined as an increase in plant only from changes in average climatic conditions. growth attributable to a higher-than-normal carbon dioxide concentration in the environment. the benefits In the longer term, these systemic climatic for agricultural productivity from carbon fertilization are difficult to gauge because they depend on many vari- changes are likely to reshape the geography of ables (i.e., crop type, latitude, soil conditions and man- agricultural land worldwide. The most vulnerable agement practices, etc.). as a consequence, impact estimates accounting for this effect are lower than agricultural systems occur in arid, semi-arid and those that do not account for it, but are affected by high uncertainty. 4 D e v e l o P m e n t a n D c l i m at e c h a n g e likely to result in reduced food production within broader strategies that benefit livelihoods and the next couple of decades in areas already facing mitigate other risks. First, adaptation investments food insecurity, especially in low-latitude regions. could increase resilience to current climate vari- Temperature rise, combined with changes in ability, while also preparing for a future increase timing, magnitude and distribution of precipita- in variability due to climate change. This possi- tion, is likely to increase moisture and heat stress bility reflects the presence of an "adaptation defi- on crops and livestock, with the subtropical cit" that diminishes the efficiency of the regions being among those most impacted. agriculture even in the context of current climate Potential indirect impacts include: heightened conditions (see Box 1). Second, many responses risks of soil erosion, runoff and landslides; will have benefits beyond managing climate risks decreased river flows in the dry season caused by (e.g., improving water-use efficiency in areas that reduced glacier runoff; and increased crop losses are already water-scarce due to non-climatic pres- from insects, diseases and weeds. These impacts sures, such as increased water demand from are likely to be very acute without any adaptation different sectors). In both cases, these adaptations (the so called "dumb farmer" syndrome). In real- are referred to as "no-regret" investments. ity, some degree of autonomous adaptation (see Examples of no-regret adaptation responses in 2.1.2) will occur, especially where adaptation agriculture include (Padgham 2009): capacity is higher, which will reduce productivity losses. Still, the residual damage from climate · Improving access to new crop varieties and change, net of autonomous adaptation, may be other production factors, which can help substantial in a number of areas, especially those farmers improve overall production and bet- with the poorest populations (World Bank ter manage risks from droughts and floods. 2008a). · Enhancing resilience of the resource base to extreme climate events through conservation agricultural practices that protect soils 2 .1 .1 Ad A ptAtion And against runoff and erosion, promote biodiver- no- regret investment sity and conserve water. Adaptation in agriculture entails sustaining rural · Modernizing irrigation systems, which can development in the context of risks from a increase water-use efficiency, bring greater changing climate.4 However, many, if not most, flexibility to water delivery for agriculture, of the needed investments and other activities and help farmers diversify to better manage will also bring benefits, irrespective of how much climate risks. the climate changes, for one of the following · Improving coordination around the contain- reasons. In other words, actions identified as good ment and management of invasive alien spe- risk management strategies for adaptation to cies, which is needed for managing both climate change also can be valuable parts of current risks from invasive species and for building the capacity to cope with an expected increase in this risk with climate 4 from a narrow economic perspective, this may not be change. true in some areas, especially marginal areas. When investments to sustain livelihoods in marginal areas are · Creating opportunities for rural livelihood not economically justifiable, one may argue that aban- donment of rural marginal areas and migration is a bet- diversification, which can lead to increased ter adaptation strategy. But in this case, other issues economic security and less reliance on cli- (i.e., overpopulation in urban areas leading to public health problems and/or social unrest) may arise. mate-sensitive agricultural activities. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 5 Box 1 What is meant By "aDaPtation Deficit" anD "malaDaP- tation"? "adaptation deficit" refers to circumstances in which even under existing climatic conditions, the agricul- ture sector is less productive, less efficient and less resilient to unanticipated shocks than it could be. adaptation deficits have arisen, for example, where: agricultural development has been neglected for a number of years (i.e., in drylands and other marginal areas that have not benefitted from investments and subsidies, generally targeting high potential areas); lack of access to markets (including due to pro- tectionist policies in other countries) limits economic returns to increased crop diversity; and lack of access to knowledge or credit constrains the use of more efficient practices and resilient crops. in the presence of an adaptation deficit, policies and investments that improve efficiency and resilience today will also contribute toward making agriculture more adaptable to future climate change. in this respect, they are "no-regrets" measures for both current and future agricultural activity. for this reason, it is becoming more common to refer to the problem as a "development deficit" rather than just as an "adap- tation deficit." "maladaptation" refers to interventions that, in addressing specific development objectives, end up being counterproductive with respect to adapting to climate change or supporting the adaptive capacity of local communities. an example is the presence of wasteful water subsidies that damage the environment (e.g., by reducing environmental flows) and create incentives for cultivation of water-intensive crops, irrespective of water-use efficiency considerations. a more subtle case of maladaptation exists in proj- ects that aim to implement some type of planned adaptation, but may end up lowering local adaptive capacity and/or creating disincentives to autonomous adaptation. an example is an agricultural project that supports monoculture of a high-value crop, with the objectives of maximizing the irrigation system efficiency, water productivity and yields ("more crop per drop"), and, ultimately, of boosting income gen- eration. although such a project might be designed taking into account the effects of climate change on the local climate and hydrological conditions, in the absence of insurance against yield losses, it would lower the adaptive capacity of farmers by making their income generation base more volatile. in the case of a bad harvest, farmers' income would be greatly affected, i.e., the ultimate impact of the project would be one of increased vulnerability to climate risks. Source: authors. On the other side of the spectrum, there may be important (see 3.2). Of course, timing matters: responses whose benefits stem mainly from adaptation measures can be separated according addressing climate change risks, such as infra- to a time dimension, e.g., with reference to short- structure projects (e.g., dams, dikes) designed term, medium-term and long-term temporal specifically to proactively respond to projected horizons (Kurukulasuriya and Rosenthal 2003). changes in factors such as runoff and sea level rise Moreover, the decision to act now or later is an as a consequence of climate change. The latter important aspect of project evaluation, particu- investments could have "higher regret," meaning larly for higher-regret investments (see 2.2.3). that explicit consideration of the uncertainty of climate change in the evaluation is even more 6 D e v e l o P m e n t a n D c l i m at e c h a n g e While adaptation strategies, policies and activities Planned or public sector adaptation involves action can take place as stand-alone measures, they may by a local, regional and/or national government to be more effective when integrated into broader provide needed public goods and incentives to efforts designed to improve the livelihoods of the private sector to fit the new conditions. For communities dependent on agriculture (e.g., example, if climate change is expected to affect expansion of extension-type services, introduction water availability (i.e., runoff ) and demand, water of more cost-effective cultivation methods). harvesting infrastructure can be built and/or However, a risk that must be addressed is that water can be reallocated among users. Referring projects pursuing broader development objec- again to Table 1, the first intervention (water tives could be counterproductive with respect to harvesting infrastructure) is an example of a adaptation to climate change, including support- "hard" adaptation investment, while the second ing the adaptive capacity of local communities. (water reallocation) is an example of a "soft" adap- Box 1 further discusses this risk. tation investment via modified institutions and incentives. Soft adaptation actions alter the circumstances in which private sector decisions 2 . 1. 2 Cl AssifiCAtion of are made (in particular, autonomous adaptation A dAp tAti on decisions) and their value must be assessed in that light (Agrawala and Fankhauser 2008). Table 1 below provides a summary of the types of adaptation activity relevant to our purposes. Other examples of planned adaptation (taken Autonomous or private adaptation involves actions from Rosenzweig and Tubiello, 2007) include: by farmers, communities and others in response to the threats of climate change perceived by · modernization or development of new irriga- them, based on a set of available technology and tion infrastructure management options. Autonomous adaptation is · transport and storage infrastructure implemented by individuals only when consid- ered cost effective by those implementing it, i.e., · land-use arrangements and property rights when adaptation is in their self-interest · economic incentives for sustainable land uses (Mendelsohn 2006). Potential examples include selecting different technologies, changing crops, · water pricing inputs and management practices suited to the · watershed management institutions new environment, shifting crop calendars, and · training for the private and public sector/ changing irrigation schedules. capacity building taBle 1 summary of aDaPtation categories By tyPe Adaptation classification Examples sectoral change crops, crop calendars, irrigation schedules autonomous Private sector economy-wide market adjustments in crop prices reflect new production levels hard "climate proof" infrastructure, including irrigation systems and rural roads Planned Public sector seasonal climate forecasts, capacity building, research and extension on drought soft resistant crops, local institutions, economic incentives for efficient water use Source: compiled from material in agrawala and fankhauser 2008. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 7 · economic incentives for efficient water-use be taken into account when development technologies outcomes depend on how the climate will change in the next few decades. For example, the design · agricultural research on drought-resistant of a new irrigation system warrants consideration crops of the expected water availability during the proj- · financial services (microcredit, insurance) ect's lifetime, which is generally 20-30 years. All but the first two are other examples of soft Some longer-term climate-change related risks adaptation. Although each of these activities already seem very likely in the next few decades, could also be part of "business-as-usual" agricul- with high expected impacts, (e.g., melting of tural development initiatives, the common glaciers threatening to compromise water avail- denominator for our purposes is that they repre- ability in entire watersheds in the Andean sent responses to anticipated changes in climate region). In these cases, adaptation needs to deal (including increased variability). with medium-to-long term changes in overall climatic conditions, as well as changes in Development projects focus on public invest- variability. ments in adaptation. Therefore, planned adapta- tion is the focus of this paper. Nevertheless, autonomous adaptation must be taken into account when defining the "baseline" or "without- 2.2 assessing the project" scenario. Moreover, in project evaluation, costs anD Benefits of it is important to consider how planned adapta- tion may influence the private sector's capacity to aDaPtation undertake autonomous adaptation. Project economic analysis calls for defining the "baseline" or "without-project" scenario. For a stand-alone adaptation project, both benefits and 2 . 1. 3 AdApting to ChAnges in costs can be assessed relative to a no-project Cl im Ate vA riA bility And to alternative. For a project with adaptation compo- me di um- l ong term Cl imAte nents undertaken within a broader set of activi- ChA nge ties, the comparison would be made relative to a business-as-usual project without adaptation Partly as a consequence of uncertainty over future components. In either case, but especially in the climate change impacts (see 2.2.1), individuals, latter, there is an inherent subjectivity and need communities and institutions often put a priority for expert judgment in defining the hypothetical on strengthening shorter-term responses to alternative as a basis for comparison. Indeed, current climate variability (Callaway 2004). Given unless specifically called for to isolate and value the impacts of current climate variability on adaptation components (see 3.2.1), it may be development outcomes and projections of more useful simply to value alternative project increasing variability and extremes in the coming designs, including different adaptation compo- decades, many developing countries are likely to nents, without differentiating between adaptation aim first at making communities and natural and broader objectives. systems more resilient to both current and future climate variability (including, for example, Another important aspect in economic analysis is increased frequency of extreme events). the consideration of "co-benefits." The economic Nevertheless, future climate change trends must 8 D e v e l o P m e n t a n D c l i m at e c h a n g e assessment of any agricultural development proj- Climate variability and change, and responses to ect can and should consider adaptation co-bene- them, add other dimensions of uncertainty to fits of investments that help facilitate project evaluation, even over a medium-length autonomous adaptation or increase adaptive time horizon (20-30 years). Specific sources of capacity as a by-product. One example is an agri- uncertainty include the following: cultural project which aims to increase agricul- tural productivity through improved water · Uncertainty over the underlying physical or efficiency in an area that is already water-scarce.5 ecological processes. Longer-term climate On the other hand, when carrying out an change impacts remain uncertain, particu- economic assessment of any stand-alone adapta- larly for use in most project-level planning tion project, it is always important to consider and management decisions, for several rea- co-benefits, in addition to the specific benefits sons. First, future greenhouse gas (GHG) associated with climate change adaptation (see emissions are unknown, as they critically 3.2.1). depend on global economic growth and miti- gation efforts. In addition, the relationships between GHG concentrations, temperatures 2 . 2. 1 deAl ing with unCertAin t y (regional or global), and climate patterns are i n t he eConomiC AnAlysis of complex and uncertain (Pindyck 2007). AdA ptAti on Different global-scale models assuming the same emission scenarios often disagree about The estimation of costs, benefits and effectiveness scale and sometimes even about the direction of any investment project generally raises a of climate change impacts, particularly at the number of methodological issues. Even without regional and subregional levels. Projections considering climate variability and change, for are provided via a range of estimates, fre- example, the economic analysis of an agricultural quently with limited information about con- project will depend on assumptions made on fidence intervals. So, even if we could future crop, input and energy prices, development determine GHG concentrations in the next of export markets, and patterns of rural-urban 20-50 years, estimating expected impacts on migration. By the same token, for many invest- precipitation, biodiversity, agricultural yields, ments--particularly those involving environmen- etc. would be challenging. Furthermore, tal or social capital--uncertainty exists regarding information remains sparse regarding how the economic value of the non-market benefits. climate changes and socioeconomic changes might interact, even though individual and institutional responses are critical determi- nants of climate change damages. · Uncertainty over the damages avoided or 5 actions to increase a society's fundamental or "raw" mitigated through adaptation. Additional adaptive capacity (sen 1999)--for example, invest- ments in nutrition, education and health services--may uncertainty arises from the relative lack of also, in principle, be included within the purview of cli- experience in evaluating the benefits of adap- mate change adaptation because they contribute to making communities less vulnerable to climate risks. tation measures. We have already alluded to for example, education allows new generations to two components of this challenge. One is the engage in income-generating opportunities other than agriculture. this type of investment could be consid- challenge of tracing through the impacts of ered an extreme case of "no-regret" adaptation. however, our focus here is on adaptation measures interventions, particularly those related to more directly related to resilience to climate variability soft investments in knowledge and and change. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 9 institutions whose benefits are realized by a As noted, the controversy is sharpened in cases of range of changes in private behavior (see long-lived investment projects. 3.2.1). The other is the continuing challenge of how to evaluate physical or ecological When to invest also depends on the time profile impacts in monetary terms. This may be rel- of benefits. Soft adaptation projects may yield the atively manageable in examining the value of greater share of their benefits over a relatively changes in tangible resource availability, such short term (a few years). Investments in local as water. It is more difficult when ecosystem infrastructure that have a somewhat longer changes (i.e., land degradation) might affect economic life (e.g., 10-30 years) may also deliver agricultural productivity in several ways that the greatest benefits in the near term. Where remain poorly understood. In these cases, a such investments have high co-benefits in reduc- non-economic evaluation approach might be ing a current adaptation deficit, the argument for recommended (see 3.2.2). more rapid investment is further strengthened. Nevertheless, the real challenge for the eco- Deciding how much to adapt now, versus waiting nomic evaluation of adaptation goes beyond to do more in the future, also depends on difficult the lack of climate change data at the "square to evaluate tradeoffs related to uncertainty. In centimeter level" or uncertainty surrounding particular, waiting can deliver a benefit from which climate change scenario is likely. It has gaining additional information on the impacts of more to do with the absence of a systematic climate change and the options for ameliorating approach to explicitly make informed deci- those impacts. However, the magnitude of this sions under uncertainty (see 3.2.3 for a dis- benefit is uncertain and needs to be weighed cussion of possible solutions). against the cost of delaying adaptation. For exam- ple, in circumstances where the impacts of 2 .2 .2 deC iding between climate change or increased climate variability i nve sti ng now or l Ater pose serious threats to the livelihoods of whole communities, an adaptation measure imple- Decision makers have choices about when to mented now might give the affected population invest as well as how much and in what form. the possibility of remaining in place versus the When making a decision, a key issue regarding need to relocate when climate change hits hard in the timing of adaptation interventions is the eval- the future. On the other hand, large commit- uation of benefits and costs over time. Standard ments of fixed capital to adaptation-oriented economic net present value (NPV) analysis infrastructure investments may foreclose options discounts future costs and benefits to a common to pursue more gradual or different types of base year using a specified rate of discount. adaptation in the future (see Fankhauser 2006 for Numerous debates exist with respect to the more discussion of these issues). How one might choice of this discount rate in project assessment. try to gauge the value of such options is one of Conceptually, one seeks a discount rate that our topics in the next section (see 3.2.2), where reflects the social opportunity cost of capital we also address the related issue of long-term (Bosello and others 2007), but in practice there is discounting under uncertainty (see 3.2.1). much controversy over what that rate should be. 10 D e v e l o P m e n t a n D c l i m at e c h a n g e e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 11 3. AppRoAChES And METhodologIES foR EVAlUATIng AdApTATIon The problem of economically evaluating adapta- current assessment methods more suited to appli- tion to climate change at the project level can be cation at the project assessment level. disaggregated into two distinct subproblems, namely: 1. Evaluating the potential impacts that climate 3.1 assessing the change could have on agricultural productiv- ity in the project area, assuming either no imPacts of climate adaptation at all or only autonomous change on agricul- adaptation. tural ProJects 2. Evaluating costs and benefits of possible In the last few years, the research community has planned adaptations, including the implica- developed a few alternative methodologies (either tions of uncertainty with respect to the new approaches or adjustments of existing ones) choice of specific adaptation options. that can be used to carry out an economic analy- These assessment stages are common to the eval- sis of climate change impacts on agriculture. Two uation of adaptation in any sector. The specific approaches in particular -- one from the agro- approaches and methodologies that can be used nomic field and one from the economic field -- to deal with each subproblem, on the other hand, have become the most widely used in applications can be different depending on the sector and the to country studies and projects dealing with specific project's characteristics. In the remainder climate change impacts and adaptation in agri- of this chapter, we will describe some possible culture. These are the agronomic (or crop) models methodologies for addressing each subproblem and the Ricardian (or hedonic) models.6 A third from the perspective of an agricultural project, and illustrate their application by referring to specific project assessments. We will also under- 6 Pradeep and mendelsohn (2008) further divide the line the need for more applied research to make approaches into four categories: agronomic, panel data, agroeconomic and ricardian. 12 D e v e l o P m e n t a n D c l i m at e c h a n g e approach, developed in the engineering field for Agronomic models assess vulnerability to climate the estimation of disaster risk and based on prob- change, in terms of expected yield losses, of local ability functions, may be promising for applica- or regional agricultural production systems. tion to extreme events. We briefly describe each Seasonal dynamics and inter-annual variability of these in the following sub-sections, providing can be accounted for by some models. Some some examples in text boxes. recent applications aim to model the impacts of flood extremes (see Box 2), as well as long-term crop production under conditions of increased 3 . 1. 1 AgronomiC or Crop climate variability (i.e., more frequent dry spells mode ls or more intense rainfall). A summary table describing the main characteristics of some These models are biophysical representations of commonly used agronomic models in climate crop production simulating the relevant soil- change applications is provided in Padgham plant-atmospheric components that determine (2009). plant growth and yield (see Figure 1). They can be used to assess the impacts of climate change Agro-economic models include an economic on agricultural productivity, as well as to investi- gate the potential effects of different adaptation module and can be used to assess the economic options. Examples are planting and harvesting impact of climate change on agriculture, and methods, fertilization, irrigation, change of crops reduced economic losses for farmers from imple- and cropping mix, and timing and/or amount of mentation of particular adaptation practices. irrigation. Crop models can be part of more Costs of autonomous adaptation that fall on indi- complex "integrated models," where different vidual farmers can be accounted for (i.e., cost of components (i.e., climate, water balance, crop fertilizers, energy costs for irrigation, etc.), while production and economic modules) interact with costs of planned adaptation (i.e., the investment each other. cost of a water reservoir for irrigation serving a figure 1. stylizeD schematic illustration of a soil-Water- croP moDule (BaseD on sWaP-Wofost moDels) Irrigation Precipitation Transportation Evaporation Surface runoff Unsaturated zone Transport - water - heat - solutes Saturated Drainage/ zone subsurface infiltration Properties - water retention - hydraulic conductivity Deep groundwater Source: nkomo and gomez 2006. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 13 Box 2 imPlications of climate change on fooD security in BanglaDesh this World Bank study in Bangladesh has the objective of assessing future food security issues associ- ated with climate change at the country level (primarily focused on 2030 and 2050 time frames), taking into account both changes in mean climate variables and climate extremes. given the comprehensive purpose of the study, many different models have been applied, whose results have been integrated to provide a comprehensive picture of the degree to which climate change is likely to pose a risk to food security in the coming decades. among these, a crop model (Dssat) is being used to derive estimates of crop production throughout the country. the model is being calibrated with realistic local-level information on soils, crop manage- ment practices, weather data, cultivars used, planting schedules, etc. in addition, an analysis of histori- cal climate risks is being undertaken to examine empirical relationships to crop production. upstream water demand changes as a result of climate change, as well as flood damage yield functions, have been factored into the crop models. more specifically, basin and national-level hydrologic models (namely, the Dhi mike 11 model and an in- house ganges-Brahmaputra-meghna Basin regional model) are being employed to produce information on future characteristics of floods in the country. these hydrologic models are being calibrated to global circulation models parameterized to track 20th century historical scenarios. (these large-scale comput- er simulation models are designed to reproduce key features of the very complex processes making up the global climate system.) the Dssat crop model has built in flood damage yield functions that utilize output flood characteristics from the hydrologic models. Source: World Bank, 2009a. vast area) cannot be included in such farm-level climate change on future daily weather, and then assessments. on agricultural production. At this point, yield outputs can become inputs to economic models An important advantage of these models is their that calculate the economic value of production flexibility, particularly due to the possibility of or farmers' income (see Box 3 on an application in adding or removing specific modules and repre- India). senting local conditions in some detail.7 This allows for tailoring to specific local conditions. By far the most important issue related to the use For example, they can easily be linked to global of these models for project-level assessment is the or downscaled circulation models, and outputs fact that they are calibrated using historical rela- from these models can be used as inputs to a tionships between independent variables (i.e., soil "weather module" to simulate the effects of profile, climate data, management practices) and production outputs. However, these relationships are likely to overstate the longer-term potential 7 for example, the Dssat model is comprised of the future impacts of climate change, since they do following modules: land, management, soil, Weather, not adequately allow for autonomous adaptation soil-Plant-atmosphere and Plant growth modules. 14 D e v e l o P m e n t a n D c l i m at e c h a n g e Box 3 climate change imPacts in Drought anD flooD affecteD areas in inDia this World Bank study in india aims to enhance the understanding of climate and climate-related issues in the indian agricultural sector, focusing on areas particularly vulnerable to droughts and floods. the integrated modeling system (ims), developed for the purposes of this study, consists of three subcom- ponents--a regional climate model (hadrm3), a hydrological model (sWat) and an agro-meteorological simulation model (ePic)--and their functional links. these subcomponents are, in turn, linked to an eco- nomic model. Climate data: precipitation, temperature, solar radiation, Crop yields Climate humidity, etc. module: HadRM3+ local weather generator Agro- Crop mix and meteorlogical Farm economic financial model: EPIC model impacts Hydrological model: SWAT Surface water Farm data management techniques in particular: · the starting point for the ims is the generation of regional climate data based on iPcc emissions scenarios (iPcc 2009); climate projections have a spatial resolution of 50 km x 50 km and are gen- erated for 2070 to 2100. · a stochastic weather generator projects these climate impacts to the local level. · the resulting climate data is then used in the hydrological model, or sWat, to generate surface water data, required as inputs to run the agro-meteorological model. · the agro-meteorological model ePic integrates water and climate data into an agricultural output estimation framework. · finally, a custom-built farm-level economic model interacts with ePic to assess the financial impacts of climate change on farmers and to determine effective adaptation strategies. the basic assumption is that farmers respond to the actual weather by adopting management techniques that maximize their payoffs (for instance, in dry years it may be necessary to irrigate some crops more intensively and reduce water allocations for other crops. if this occurs, it will also be necessary to adjust fertilization rates). the ePic module predicts yields under different management regimes, while the corresponding economic module computes the associated payoffs. Source: World Bank 2008b. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 15 by the affected farmers whose activities are being the annual net revenues from production of the modeled. Moreover, the "without-project" cultivated crops or livestock. damages are overestimated to the extent that they cannot incorporate the effects of future techno- The impacts of different influences on land value, logical change. By the same token, crop models including climatic differences, are econometrically can overstate the positive impact of a planned estimated using cross-sectional data (i.e., data on adaptation initiative by not considering how agricultural land at different locations at a given autonomous adaptation already partly offsets the time). The effect of various other influences, such adverse climatic impacts. How serious this bias is as socioeconomic conditions, soil and geographic will depend on available opportunities to those characteristics, can be controlled to provide esti- covered in the analysis for autonomous adapta- mates of the effect of climate variables on land tion. A similar problem arises in trying to gauge values. After estimating how climate conditions the contribution of soft adaptation efforts. These (i.e., changes in temperature or precipitation) efforts are, in fact, designed to change the param- affect land values, it is possible to use climate eters related to farm-level inputs and outputs. For scenarios to infer the impact of climate change example, training on more effective fertilizer use on the value of farmland and, hence, on its will increase the yield per application of fertilizer. productivity. Operationally, data requirements (i.e., soil profile Ricardian approaches have been used to provide data, weather data, local management informa- analyses of the longer-term economic vulnerabil- tion, etc.) can be demanding for these models, ity of agriculture to climate change in: especially for project-level applications. If data availability is a constraint, an option is to apply · large countries -- India and Brazil (Sanghi less data-intensive agro-meteorology techniques, and Mendelsohn 2008), China (Wang and where the impact on yields is based only on others 2007) and the United States changes in crop evapotranspiration (see Box 4 for (Schlenker and others 2006); an application in Morocco). In terms of time and resources, the costs to benchmark and run a · small and medium countries -- Cameroon model may be considerable. (Molua and Lambi 2007) and Egypt (Eid and others 2007); · small islands -- Sri Lanka (Kurukulasuirya 3 . 1. 2 ri CArdiAn or hedoniC and Ajwad 2006); method · continents as a whole -- Africa (Kurukulasuriya and others 2006) and Latin The Ricardian method was pioneered by America (Seo and Mendelsohn 2008a). Mendelsohn and others (1994) to estimate the longer-term effects of differences in climatic These approaches have also been applied to esti- conditions on agricultural land values, and is mate impacts on the livestock sector (Seo and based on the idea that long-term land productiv- Mendelsohn 2008b). ity is reflected in the land's asset value. Given that the farmland is being used in the best possible An important strength of this methodology is way, and given environmental conditions, factor that the findings on longer-term climate change prices and other constraints, observed market rent impacts are net of whatever autonomous adapta- on the land (or farmland value) will be equal to tion responses to climate change individual 16 D e v e l o P m e n t a n D c l i m at e c h a n g e Box 4 imPacts of climate change in the agriculture sec- tor in morocco this World Bank study in morocco purports to evaluate climate change impacts and identify adaptation options by means of a combined climatological, agronomical, hydrologic and economic approach. the structure of the study includes: the construction of scenarios of mean daily temperature and precipitation based on climate change projections (derived from a statistical Downscaling model); a yields impact module based on agro-meteorology techniques (unlike crop models, these techniques estimate the impact on yields based only on changes in crop evapotranspiration, without considering soil characteris- tics); a hydrological study to evaluate climate change impacts on surface and ground water; and an eco- nomic general equilibrium model disaggregated by agro-climatic conditions, access to irrigation and farm type diversity. selected outputs of the study include: · yield impacts of climate change on rainfed and irrigated major crops at 9 regular time intervals for 4 climate scenarios. these scenarios are generated by combining results from two different global circulation models, and two projections of future economic activity and emissions from the iPcc scenarios (iPcc 2009). · the additional demand of water to offset the yield effects of climate change. · availability of water for irrigation from dams, taking into account the growth in demand for municipal and industrial use, the reduced inflows that will result from increased temperature and reduced rain- fall, and the rules that govern allocation of water across different uses (estimates disaggregated by major dam). · change in groundwater recharge due to change in rainfall and temperature, and increase in costs of groundwater extraction due to a combination of reduced aquifer recharge and over-extraction (estimates disaggregated by major aquifers). · adjustment of the economy as a whole (and of the agricultural sector in particular) to changing cli- matic conditions. for example, changes in agricultural value added, employment and agricultural trade will be estimated under different scenarios (incorporating assumptions regarding demograph- ics, labor force, savings and investment behavior, and productivity). · evolution of both the "median" equilibrium solution for the variables of key policy interests and of extremes (e.g., with probability of 5% or 10%) in order to inform policy on low-risk, high-impact events. · optimal mix of adaptation responses based on net marginal benefits. Source: World Bank 2009b. farmers are able to make over the longer term. In Earlier applications of Ricardian methods tended applying this approach, it is assumed that over to produce models that were, to some extent, the longer term, a new climate regime will induce "black boxes" with respect to the identification of geographic redistribution of agricultural activity actual adaptations by farmers. This also poses an and other behavioral changes that are reflected in obvious constraint on the analysis of specific how farmers have already adapted to different planned adaptation measures. However, the most climate conditions in diverse geographical areas. recent models are increasingly able to provide e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 17 Box 5 an examPle of neW generation ricarDian moDels Kurukulasuriya and mendelsohn (2008) examine the impact of climate change on primary crops grown in africa. they propose an innovative approach that aims to bridge the gap between agro-economic and traditional ricardian models, and label it a "structural ricardian" model. a simple model of the farm is developed where a farmer first chooses a desired crop or crop combina- tion and then earns a conditional income based on the crop chosen. By modeling crop choice across dif- ferent climates and measuring the role that climate plays in these choices, this approach reveals one of the explicit adaptations that farmers make, thus overcoming a primary limitation of the traditional ricardian models. the resulting model can be used to predict the effect of climate change scenarios on expected net revenue, both with or without changing crops. the model is estimated using a sample of over 5000 farmers across 11 countries in africa, with the analysis concluding that farmers shift the crops they plant to match the climate they face. according to the authors, by accounting for crop switching, the damages from climate change are not overestimated and the benefits not underestimated. Source: Kurukulasuriya and mendelsohn 2008. information for answering these kinds of policy- necessary, for example, to somehow convert the relevant questions. In particular, the latest genera- increased availability of water to an equivalent tion of "structural Ricardian" models change in the climate. However, this would be (Kurukulasuriya and Mendelsohn 2008, and Seo more subjective and prone to error. More complex and Mendelsohn 2008a and 2008b) can model issues arise in evaluating soft adaptation invest- crop/livestock, irrigation and farm-type choices ments. For example, if a measure of know-how using a multinomial probability setting, and are was included in the estimate, the application more capable of distinguishing among different would be reasonably straightforward. If not, then agro-ecological zones (see Box 5). it would be difficult to separate the influences of the capacity-building investment from the The Ricardian method has not traditionally been unmeasured autonomous adaptations. applied for assessing planned adaptation projects, but it could be, in principle. Consider a hard Therefore, in some cases, the Ricardian approach investment like water storage or irrigation. To the can be applied to assess the "with-project" extent that similar kinds of infrastructure invest- scenario. The counterfactual "without-project" ments were included as explanatory variables in scenario would call for a different approach, such the equation for the land value, it is possible to as the application of a crop model. Although a look at how an increased availability of infra- quantitative comparison of the two models is not structure services combined with projections of a feasible, a qualitative comparison can indicate changing climate would affect land values, and roughly the value of adaptation measures in thereby deduce a value for the benefit of the response to climate change, if the Ricardian investment. If infrastructure were not included as model is correctly specified in order to reflect, to an explanatory variable, then it would be the extent possible, the adaptation measures 18 D e v e l o P m e n t a n D c l i m at e c h a n g e promoted by the project. Crop models, in particu- change and the impacts of specific adaptation lar, can provide a baseline case corresponding to measures. In principle, the effects of changes in an assumption of current farming practices and climate variability and frequency/severity of environmental conditions except for the changes extreme events could also be assessed with a in temperature and precipitation specified in the Ricardian approach. This is possible as long as crop simulation models. The Ricardian model, on differences in these climatic attributes are the other hand, "allows" farmers to move from included in the cross-sectional data and their one set of crops and/or technology to another as effects can then be reflected in land values inde- the climate changes, providing an estimate of the pendent from long-term climatic differences. benefits derived from adaptation (although However, in practice, this may prove to be a diffi- limited to the observed ones in the particular cult task. region). An example of the joint application of a crop and a Ricardian model for project evaluation An important caveat of the Ricardian analysis is is provided in Box 6. that variations in the amount of water available at the farm level are not considered; the approach Operationally, the Ricardian approach relies on implicitly assumes that sufficient water is avail- data collected through surveys among farmers, able at a particular location to accommodate the where questions about farm types, crop cultiva- specific adaptation measures undertaken by farm- tion and other activities during a farming period ers. Crop models are better equipped to deal with are asked. These surveys need to be carried out in water availability issues if linked to an upstream districts and villages chosen to get a wide repre- hydrological module that calculates water avail- sentation of farms across climate conditions in able for plant growth. Another issue with the area of interest. Moreover, local data on Ricardian models is that the carbon fertilization climate, soil and hydrology are needed, and, above effect cannot be addressed, while the crop model all, reliable independent measures of land values analysis can account for it. are required. As a consequence, costs of obtaining the data for carrying out this analysis may be Institutional and technical constraints to autono- high, especially for a smaller versus national-scale mous adaptation also may be difficult to measure project. For smaller geographical levels (i.e., proj- in a number of cases. These constraints could be ect level), this approach can be successfully rooted in local culture and habits, or the lack of applied if a national-scale study is already avail- know-how in some regions. In a Ricardian analy- able (see Box 6). sis, this is not a problem if these factors stay constant over time. If they change, however, those For a full list of strengths and weaknesses of the omitted influences could lead to biased estimates Ricardian approach, see Kurukulasuriya (2006). of the potential effectiveness of autonomous For our purposes, a few advantages and disadvan- adaptation. Similarly, future technology advances tages of this method with respect to the crop are not factored in. If these changes imply less model approach are discussed here. climate change sensitivity, then omitting them leads to an overestimation of impacts from An advantage of the Ricardian approach over climate change despite the incorporation of crop models for assessing climate change impacts autonomous adaptation. Finally, by their very is that the economic impacts can be modeled nature, Ricardian models do not provide insight even in the absence of a full understanding and regarding how autonomous adaptation practices modeling of the biophysical impacts of climate would be phased in over time. Crop model e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 19 Box 6 aDaPtation for irrigateD agriculture in china the ongoing "mainstreaming climate change adaptation in irrigated agriculture" project in china (World Bank 2008b) used a two-model methodology to assess the benefits of the proposed adaptation mea- sures in the project area (3h Basin). in particular: · a set of biophysical process-based crop models was used to simulate the impact of expected cli- mate change on yields of major crops in pilot locations (the "without-project" or without adaptation case). · a farm-level, statistically-based economic (ricardian) model estimated the impact of farmer adapta- tion to climate change on farm income (the "with-project" or with adaptation case). the ricardian climate estimates were taken from a recent statistical analysis of agricultural sensitivity in china (Wang J., mendelsohn r., Dinar a., huang J., rozelle s., and zhang l. 2007). climate change scenarios (assessed in both the crop and the ricardian models) included a combination of: (a) increased temperature of 0, 2 and 5 degrees; and (b) increased and decreased precipitation of 0 per- cent, +15 and -15 percent, and +30 and -30 percent mm/year. changes in variations by season were not simulated. to estimate the economic impact of adaptation at the local county level, the present climate for each county was inserted into the ricardian function in order to calculate the present net revenue per hectare. such net revenue captures a crop mix as well as technologies and management practices that farmers already undertake in response to various perceived signals, including responses to present changes in climate. next, climate change scenarios were incorporated into the ricardian models to calculate the change in net revenue resulting from change in climate for each country. the comparison of results from the two models provided a "qualitative" justification for the project. indeed, the comparison of the biophysical crop simulation models showed that reductions in maize would occur as a result of climate change without adaptation ("without-project" case), while the ricardian model indicated that, with adaptation, net farm income can increase ("with-project" case). these results assumed that the farmers have the knowledge, guidance and support needed to imple- ment adaptation. given that the project aims to provide farmers with the necessary knowledge and favorable conditions to adapt, these results helped to justify the investment. in addition, the implicit mod- els' assumption that water is non-limiting, compared to the actual situation where water supplies are very limited in the 3h Basin, provided a justification for the project's strong focus on improving "real water saving" as a means for climate adaptation. Source: World Bank 2008c. analysis, which relies on field data and expert An obvious final constraint that applies to both judgment on current and future farming practices, Ricardian and crop models is that farmers' behav- can in principle better control for these factors, iors are influenced by variables that are generally but assumptions over techniques and technologies not accounted for in the models, such as policies available to farmers in the present and future and subsidies. For example, in northern Mexico, need to be made explicit and justified. wheat and corn are the most commonly grown 20 D e v e l o P m e n t a n D c l i m at e c h a n g e crops, not because of the region's climatic charac- probability that the event has equal or greater teristics, but because the government subsidizes intensity than the corresponding point on the irrigation and pays a kind of support price for x-axis. Thus, as the potential intensity of an event these products. In principle, these variables might increases along the x-axis, the probability that an be accounted for, but this has not been done in actual event will have impacts exceeding that level the Ricardian studies published so far. declines (ultimately falling toward zero as the target intensity grows without limit). With addi- tional probabilistic information about the 3 .1 .3 prob A bilisti C methods frequencies of occurrence of the climatic events for i mpACt Assessment of of each degree of intensity over time, and about ex treme events the associated economic losses, it is possible to put together an estimate of the expected cost of The literature on and the practice in the field of the occurrence of extreme events (and its vari- disaster risk reduction suggest another method ance). For high-intensity, low-frequency, events, for the estimation of expected economic losses estimating probability distributions is difficult, due to climate change, as well as of economic particularly since few historical observations are benefits of adaptation measures (see 3.2.3). These available. methods were developed for application to natu- ral disasters and, hence, are most immediately For many extreme events, climate change will applicable to impacts of climatic extremes (i.e., have the effect of translating the curve towards floods), although it may be possible to adapt the the right side (due to increasing frequency and/or approach to evaluate other impacts of climate increasing magnitude of natural hazards). This change. has the effect of raising the expected impact. However, the reliance of this method on proba- An exceedance curve showing the relationship bility functions makes its use in climate change between intensity and probability of a certain applications challenging, particularly because event (i.e., flood) is at the core of this technique, estimating probability functions of extreme which allows for the probabilistic estimation of climate events proves very difficult under climate monetary losses due to natural disasters. In Figure change (see Box 7). 2, each point on the y-axis indicates the figure 2 effect of climate change on an intensity-ProBaBility function Climate Change Frequency / Probability Intensity e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 21 Box 7 uncertainty anD ProBaBility functions the typical view of uncertainty assumes that the distribution of possible outcomes takes the shape of a bell curve (called normal or gaussian distribution), with equal probability that the actual outcome could be either smaller or greater than the predicted average (gulledge 2008). a situation with no climate change is represented in figure (a), where the left figure represents the normal distribution of the cli- mate variable, while the right figure represents variability around the mean. observed trends of climate variability and projections of climate change on different climate-related variables suggest that the shape of the initial probability distribution may change in the near or distant future. in particular: Box 7 Uncertainty and probability functions · in most locations, the mean values of the main climatic variables will (a) change, i.e., the probability distribution will shift towards the left or the right side (the left side in figure (b), left col- umn), and climate will vary around a (b) different mean (a lower mean in Impulsive change of central tendency figure (b), right column). this is what we refer to when we say, for example, that, according to a specific climate model, by 2050 the average precipita- (c) tion is likely to decrease by 10% in a Increasing variability certain location. unfortunately, differ- ent models may generate slightly or significantly different projections (d) Change in central tendency depending on the area. as a conse- and increasing variability quence, in many cases, it is not possi- ble to estimate with sufficient confidence which direction and by how much the distribution will shift. · future probabilities of above- or Source: scandizzo and notaro 2008. below-average values of many climate variables will be higher than those suggested by the initial bell curve. in other words, either the right or the left tail (or both) of the probability distribution of climate variables will become "fatter" (figure (c), left column), i.e., the climate variable will more often be farther away from the mean (right col- umn). this is what we refer to when we say that, as a result of climate change, climate extremes will become more frequent. Depending on the climate variable, a fatter right or left tail of the distri- bution means higher frequency of climate extremes (i.e., if the variable is "mean annual precipita- tion," a fatter left tail can be interpreted as increasing droughts, while a fatter right tail means increased probability of storms/floods). once again, in many cases, it cannot be estimated with rea- sonable approximation by how much the tail will get fatter. · the two effects described above can play together (figure (d)). Source: authors and W. yu (personal communication). 22 D e v e l O P M e N T a N D C l i M aT e C H a N G e Although some methods exist to help glean investments are similar to challenges faced in information about what future probabilities may evaluating such investments in other types of look like (i.e., trend analysis, multimodel ensem- development projects. We have mentioned bles used as probabilistic climate change fore- previously that evaluating the economic casts8 or more sophisticated spectral approaches), benefits of hard investments is relatively possibly the best approach for impact assessment straightforward (although in practice it is not in project level analysis is a Monte Carlo-type trivial) because a direct relationship can be simulation embedded in a "weather generator" constructed between inputs provided by the linked to climate projection scenarios. A recent physical investment (i.e., water supply from a application in a study in India (World Bank dam) and production output. Soft adaptation, 2008b) has proven that this method is promising on the other hand, is more complicated because and reasonably easy to apply. the benefits, to a great extent, must be inferred from resulting changes in private sector behaviors and prices. Assumptions based on experience and informed judgment must be made about how specific interventions ­ e.g., 3.2 evaluaTiNG COSTS agricultural innovations, training programs or aND beNeFiTS OF policy reforms ­ could alter farmers' decision PlaNNeD aDaPTaTiONS making, outputs and economic returns (see Box 8 for methods for ex-ante evaluation of We assume that readers of this paper are familiar agricultural innovations). with the general application of cost-benefit anal- ysis to project appraisal. In this section, we focus With respect to costs, estimates can be made of on: options for assessing benefits and costs of the direct costs of undertaking both hard and adaptation measures in agriculture within a cost- soft interventions. Once again, costs of hard benefit economic framework, including issues interventions are easier to compute (i.e., store- related to the discount rate; non-economic proj- houses for food stocking or irrigation systems), ect evaluation; and approaches for dealing with while estimating costs of autonomous adapta- uncertainty. tion, as well as of planned, soft adaptations, is more challenging. In climate change impact studies in the agricultural sector, "albeit adapta- 3 .2 .1 methods for Assessi n g tion processes either autonomous or planned e Conomi C benefits And Cos ts are considered among the main drivers of final The challenges in evaluating unit costs and climate change impacts on agriculture, they are benefits of hard and soft adaptation mainly examined under the potential benefits rather than under the costs side" (Bosello and others 2007). The underlying rationale is that 8 With this method, probabilities for different events related to climate change are inferred by counting the costs of adaptation measures that can be imple- 8 number ofmethod, in which the event occursevents With this models probabilities for different (the meth- related to climate change are inferred by counting the mented autonomously by farmers (i.e., change odology is described in räisänen and Palmer 2001). number of models in which the event occursregarding although in the literature some doubts exist (the meth- in crops or calendar shifts), as well as of those odology is described in räisänen and Palmer 2001). the theoretical foundations of this method (i.e., model although in the literature some doubts exist regarding measures that consist in policy incentives, are runs may not be independent from each other), this the theoretical been appliedof this method (i.e., model technique has foundations in different studies, includ- insignificant. Unfortunately, this assumption ing the assessment of climate change impacts this runs may not be independent from each other),and technique has been applied in different studies, includ- is only partly correct because, especially in adaptation solutions in the uS Metropolitan east Coast ing the (Gornitz and rosenzweig C. 2007). region assessment of climate change impacts and developing countries, there can be significant adaptation solutions in the uS Metropolitan east Coast region (Gornitz and rosenzweig C. 2007). transition costs in changing agricultural e C O N O M i C e va l u aT i O N O F C l i M aT e C H a N G e a D a P TaT i O N P r O J e C T S 23 bOx 8 eSTiMaTiNG THe raTe OF aDOPTiON OF aGriCulTural iNNOvaTiONS economic benefits of agricultural innovations require ex-ante insight into the likely rate of adoption. in order to estimate ex-ante what this rate might be, different dimensions that influence it must be ana- lyzed. Generally speaking, adoption decisions depend on both sociological and economic factors. For example, rogers (1962) suggests five dimensions (relative advantage, compatibility, complexity, divisibil- ity and communicability), which determine the adoption rate. Relative advantage relates to the extent to which a new technique or product is preferred to the existing technology. Generally, the superiority of an innovation is measured by its profitability (crucially depen- dent on assumptions on output prices) or risk-reducing potential. Compatibility is the extent to which a new innovation is consistent with existing norms, values and prior experience of prospective adopters. also to be considered is the extent to which it is physically and managerially compatible with existing practices. Complexity is the extent to which new techniques and their consequences are easy or difficult to understand. in general, less complex ideas are more quickly and widely adopted. Divisibility is the extent to which an innovation can be used on a limited basis. The importance of divisibility stems from the potential risks involved in trying a new innovation. if trials can be done on a limited basis, earlier adopters, in particular, are able to limit their exposure to losses. Finally, communicability is the ease with which knowledge of an innovation can be passed along to potential users. This concept includes both the complexity of the innovation, as well as the rapidity and tangibility of benefits. Other important vari- ables that may influence the rate of adoption are the innovation's age, the initial investment required by the adoption decision and the riskiness of the undertaking (agriculture Canada 1984). Operationally, one possible procedure for coming up with adoption rates is described in lesser and oth- ers (1986). This procedure is based on questionnaires to potential adopters of a new technology and involves providing a sample of producers with facts about the effects of the product. respondents are then asked a series of specific questions about their own plans based on the provided information. Potential diffusion rates are projected based on responses to a question like, "Overall, on how many hectares in your field would you expect to utilize technology x?" Moreover, respondents are asked when they plan to adopt the new technology, i.e., by choosing the most likely time between 6 months and 10 years from the innovation's availability. This approach can be applied in circumstances where respon- dents have no problems comprehending factual information of a hypothetical nature and responding to it in a meaningful way. Source: authors. practices (including opportunity costs of time agriculture more climate-resilient (i.e., new seeds and/or travel costs for participating in training or or water-saving irrigation devices). Hence, it is other capacity building programs, irrespective of becoming increasingly important to get a grasp of who actually bears such costs). In addition, new adaptation costs to calculate the net benefits of technologies might be required to make adaptation. 24 D e v e l o P m e n t a n D c l i m at e c h a n g e Box 9 Deriving soft anD harD aDaPtation costs for irrigation to establish a range of global irrigation unit costs, the economics of adaptation to climate change (eacc) study team has undertaken a broad literature review of World Bank, food and agriculture organization (fao) and international Water management institute (iWmi) documents, project reports and meta-evaluations directly related to completed and ongoing irrigation projects. from these docu- ments, the eacc team extracted project component costs and benefits into a database, which currently covers 622 projects in 66 countries. the review process focused on extracting hard and soft irrigation costs. for each World Bank document reviewed, it was possible to disaggregate engineering (hard) and institutional (soft) costs, as well as identify the total number of hectares involved in the project. engineering costs cover all project expendi- tures directly related to the physical construction, rehabilitation or modernization of an irrigation system. for example, land leveling, ditch construction and irrigation piping are all covered under engineering costs. on the other hand, institutional costs include all "soft" components of a project, such as water-user groups, train- taBle 2 DisaggregateD unit ings, irrigation management staff training costs or farmer capacity building. unit costs Unit Costs have been calculated by dividing the Region Statistic Institutional Development Engineering investment costs by the total land area africa mean 7,761.00 2,792.78 affected by the project (see table 2). std. dev. 15,224.92 3,145.88 obs. 7 9 this information will then be used to esti- eaP mean 161.29 750.22 std. dev. 202.95 616.28 mate needed investments in irrigation to obs. 7 9 adapt to climate change. a methodology eca mean 83.20 883.43 is being developed to estimate irrigation std. dev. 55.80 1,148.39 obs. 5 7 needs following changed climatic condi- lac mean 2,991.00 2,125.82 tions worldwide by using the imPact std. dev. 4,871.19 1,579.02 model (international food Policy obs. 10 11 mena mean 619.71 2,663.33 research institute). the related invest- std. dev. 524.45 3,690.43 ment costs for hard and soft interventions obs. 7 9 will be derived by multiplying the area in s. asia mean 454.13 2,401.27 std. dev. 790.01 3,577.98 need for irrigation by the historical cost, obs. 8 11 differentiated by region. total mean 2,130.75 1,997.13 std. dev. 6,693.47 2,650.48 Source: essam 2009. obs. 44 56 For these purposes, different methods can be of interventions, which would be needed for applied. One approach consists of piggybacking adaptation purposes (i.e., irrigation, agricultural the costs of adaptation measures from an extension, flood protection, etc.). Unit costs can in-depth analysis of the documentation of past then be applied to additional investments needed projects that financed the same types to adapt to climate change, estimated by means e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 25 Box 10 eliciting aDaPtation cost information from local communities anD institutions the costing adaptation through local institutions (cali) project seeks to identify the perceived costs of adaptation options in rural areas from the perspectives of both rural households and the institutions through which the adaptation options are channeled. four types of data collection methods, including household questionnaires, focus group discussions, institutional stakeholder interviews and expert inter- views, are utilized. for households, the study assesses ranges of past costs for households to adapt their strategies to cli- mate-related hazards. for institutions, the amount of money and resources used in order to perform their tasks to assist households in adapting to particular hazards is assessed. these estimates serve as a basis for judging how much investments or aid would be needed from governments or donors to pro- mote particular adaptation interventions in rural areas. in both cases, the information collected through stakeholder interviews is cross checked with information from the focus group discussions and expert interviews. since for most respondents, assessing the actual costs related to adaptation options is difficult, partici- patory appraisal methods are applied. in particular, the respondents are asked to allocate a fictitious income over different adaptation options by, for example, asking them to divide a number of coins or stones over a number of cups. this method provides two insights. first, more units are allocated to the more expensive than to the cheaper options. thus, the number of tokens provides an estimate of the cost of the different adaptation options as perceived by rural households. the monetary value of a token can be determined by comparing actual prices of the adaptation options for which the prices are known with the number of tokens allocated to these options. in this way, to what extent perceptions differ from reality can be verified, offering a possible explanation on why particular options may or may not be adopted. this also allows for estimating the perceived monetary value of the options for which no mar- ket prices are known. second, it forces respondents to rank the options, showing which options are con- sidered more "valuable" than others, providing information on the perceived economic benefits of different options. on the basis of the data collected through the questionnaires and interviews, an econometric and statis- tical analysis is performed to identify the cost elements of the different adaptation strategies. the costing framework used for this purpose indicates which costs have to be made in order to implement the differ- ent options, according to the following typology: · household monetary costs; · household labor requirements; · household training requirements; · required help from the community; · required help from institutions, such as authorities or ngos; and · monetary needs of institutions, which are necessary to implement work. Source: agrawal, Kononen, and Perrin 2009. 26 D e v e l o P m e n t a n D c l i m at e c h a n g e of specialized models or expert judgment. For are also simultaneously conditioned by a whole example, a review of irrigation projects, under- range of other factors" (Agrawala and taken as part of a broad study on the Economics Fankhauser 2008). While it might be possible, in of Adaptation to Climate Change (World Bank, principle, to consider a hypothetical alternative forthcoming), has come up with ranges of invest- project designed with less adaptation integrated ments per hectare for this specific hard adapta- into it, such an effort would have little meaning tion measure (see Box 9). and it will be more valuable to compare alterna- tive project designs per se.9 A second possible approach is based on the solic- itation of information directly from local commu- For stand-alone adaptation projects or projects nities that are vulnerable to climatic risks and with a distinct adaptation component included, that take adaptation-relevant decisions. An inter- additionality of costs and benefits of adaptation esting methodology, based on participatory may be useful to estimate in some cases. This can appraisal methods, is presented in Box 10. be important, particularly when alternative proj- ect designs exist with different benefits and costs As previously noted, estimating costs and benefits that can then be compared. One also can attempt of adaptation may be complicated by: (i) chal- to indirectly identify the costs of an adaptation lenges in measuring additional costs and benefits activity linked to an existing development project of adaptation compared to development activities through a "gap analysis" to pin down which addi- without the element of adaptation, when the tional investments the adaptation project needs in project is not stand-alone; (ii) the difficulty in order to increase its resilience to climate change identifying the co-benefits of adaptation in stand- by a certain degree (Box 11). alone projects; (iii) debate over which discount rate to use, particularly when adaptation is expected to (ii) Evaluating co-benefits of adaptation have long-term effects; and (iv) the high level of A project designed for other purposes may also uncertainty in evaluating costs and benefits of deliver increased climate change resilience as a adaptation due to uncertain future climate change co-benefit, even without a specifically identified and related impacts. The first three issues are adaptation component. For example, improved discussed below, while the other major challenge water management may add to yields in the near related to uncertainty is discussed in Section 3.2. term and generate additional value in the longer term by reducing climate-related risks if climate (i) Evaluating additional costs and benefits of change is expected to decrease water supplies or adaptation make them more erratic. In addition, adaptation In the case of no-regret adaptation investments activities themselves can yield co-benefits. For and of broader development projects that fully example, improved agricultural land management integrate adaptation into their design, "adaptation actions are embedded within responses under- taken by private and public actors to a broader set 9 if, for particular reasons, additional adaptation costs of an integrated project must be evaluated and the project of social and environmental stimuli. For example, design does not make it possible to directly identify them, one can try to make an educated guess of the farming practices, land use planning and infra- percentage of project costs that can be allocated to structure design might all reflect some consider- adaptation. for example, the integrated national adaptation project in colombia (World Bank 2006a) ations of current and anticipated climate, but it calculated the additional costs of adaptation by com- may not be feasible to isolate the costs and bene- paring the total project costs with the costs of existing projects with similar purposes implemented in the same fits of the climate component, as such decisions areas, but without consideration of climate change. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 27 Box 11 calculating aDDitional aDaPtation costs for irrigation moDernization in china the development objective of the "mainstreaming climate change adaptation in irrigated agriculture" project is to enhance adaptation to climate change in agriculture and irrigation water management prac- tices through awareness raising, institutional and capacity strengthening, and demonstration activities in the huang-huai-hai river plain (3h Basin) in china. the project is linked to an ongoing irrigated agriculture intensification project (iail3), whose main components, including water saving in irrigation, drainage, and environmental protection in different agro-ecological zones, are important for adaptation to future climate change. however, iail3 components did not take climate change into account when designed, so that a number of vulnerabilities due to increasing climate variability and change must be addressed. the gef/sccf (special climate change fund) financed project has been designed as a gap-filling and programmatic operation that would focus on the 3h region with possible expansion to other regions as appropriate, based on initial experience with the project. the project design was based on a preliminary gap analysis of iail3 from a climate change adaptation perspective. this approach allowed for clearly defining the baseline and additionality of the gef project as follows. Baseline Scenario iail3 (the baseline project) finances sustainable development of modem irrigated agriculture in five proj- ect provinces in the 3h Basin, with the following development objectives: (a) increasing water and agri- cultural productivity in low and medium yield farm land areas; (b) raising farmers' income and strengthening their competitive capacity; and (c) demonstrating and promoting sustainable participatory rural water resources and agroecological environmental management in the 3h Basin. Summary of the IAIL3 "Adaptation Gap Analysis" a number of specific weaknesses in iail3 with regard to climate change have been identified, including the following: (a) Public awareness of issues relating to adaptation to climate change is very limited, as is the under- standing and capacity of staff, officials and decision makers regarding climate change and adapta- tion. (b) in the design of water saving works, the concept of collecting and storing natural precipitation was not integrated with irrigation and drainage works and there are few, if any, works and facilities in the field to collect and store rainfall runoff for more effective use of available rainfall. (c) some agricultural measures for climate change adaptation have not been fully considered, such as planting nitrogen fixation crops, adjusting sowing times of double cropped areas, staggered maturity of crops to reduce peak water demand and the more widespread development of agriculture facili- ties, such as greenhouses. (d) farmer and water-user associations are weak and have only limited ability to popularize new variet- ies, practices and technologies, which are better adapted to climate change. (continued) 28 D e v e l o P m e n t a n D c l i m at e c h a n g e (Box 11 continued) Alternative (GEF/SCCF Project Enhancement of IAIL3) the gef/sccf climate change adaptation project will increase iail3 sustainability and, more broadly, the resilience of chinese irrigated agriculture in the face of climate change, and support global environ- mental objectives. specifically, the project focuses on those iail3 activities that were identified as being at risk from climate change. the gef/sccf project will review and refine the original iail3 technical design for all adaptation-related activities and adjust the iail3 Project implementation Plan to respond to the effects of both short-term climate variability, as well as long-term climate change in each region. Financing arrangements Due to the gap-filling approach in designing this project, the totality of project costs, estimated at us$55.5 million, can be considered "additional adaptation costs". nevertheless, the total costs were fur- ther divided into two parts: (a) us$50.5 million, cofunded under the ongoing iail3 Project, to increase resilience of those activities that are potentially most affected by climate change; and (b) us$5 million, funded by gef/sccf, to support additional adaptation activities not directly linked to the baseline proj- ect, namely: · identificationandprioritizationofadaptationoptions; · demonstrationandimplementationofadaptationmeasures;and · mainstreamingadaptationintonationalprogramsandinstitutionalstrengthening. Source: World Bank 2008c. practices to prepare for climate change can also such investments may also convey "public" bene- lead to reduced erosion/siltation and carbon fits for other categories of users (e.g., municipali- sequestration.10 ties). Estimates of these co-benefits can be included and strengthen the overall case for the Co-benefits become particularly important in the project. economic evaluation if they otherwise would not be reflected in the project appraisal. This is typi- (iii) Choosing the discount rate for evaluating cally the case if the co-benefits have the nature of longer-lived adaptation benefits public goods. A private investment in improved As noted, the appropriate discount rate to apply water management for agriculture, for example, in benefit-cost analysis is one of the most often- can yield a stream of "private" benefits, including debated topics in economics, regardless of the the reduction of farmers' longer-term, climate- project's anticipated economic life. Concerns have related risks (adaptation co-benefit). Additionally, been raised about the evaluation of long-term net benefits, more specifically, in the literature on climate change economics. Two distinct argu- 10 some also may entail negative spillovers (e.g., ments have been advanced for using a "special" increased irrigation upstream may limit water availabili- approach to discounting in this case, one that ty downstream). e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 29 does not make long-term benefits of limited discount rate, with the size of the reduction consequence in the economic calculus. depending on the extent of uncertainty, among other factors.12 The first argument is related to intergenerational distributional equity: a high discount rate can These arguments at the level of the economy as a trivialize the potential value of a long-term whole have limited direct relevance to valuing climate adaptation investment for the well-being individual adaptation projects. However, two of future generations (for example, see Stern other observations broaden their potential rele- 2007). The counterargument is that applying vance. The conclusion above is not that uncer- such a special discounting procedure to certain tainty should lead to an adjustment in the classes of investments, but not others, distorts the discount rate; rather, it is that any broader risk- allocation of scarce resources among investments reduction benefits should be incorporated in with different future benefit streams, even to the assessing an adaptation project's value. Moreover, point of crowding out alternative investments in principle, these risk-reduction benefits could with large near-term benefits.11 arise in the near term as well as in the longer term. The other argument is connected to uncertainty over long-term future rates of return, which is Thus, we can first recommend that whatever addressed in Section 3.2.3. The argument is discount rate is used for other projects in a usually framed in terms of economy-wide particular country or region should also be the impacts of policies or investments on future default rate for adaptation projects, with excep- economic growth and well-being. Suppose that a tions made sparingly on a case-by-case basis. In long-term adaptation activity has a particular assessing the benefits of a project, however, atten- expected flow of net benefits over time, but the tion should be paid to the possibility that the actual flow could be above or below the expected project will help smooth out fluctuations in over- net benefit stream because of uncertainty about all well-being from climate change, as well as the magnitude of adverse impacts of climate provide more direct benefits. Such impacts are change. If climate change impacts are more severe likely to be difficult to quantify, but it is useful at than expected, economic growth will be more least to identify them heuristically. Realizing such adversely affected. However, this situation also benefits still requires that a project be "large occurs when the benefit stream from adaptation enough" relative to total output and economic is more likely to be particularly large. Thus, the well-being that its success can have more than a adaptation activity delivers both an expected trivial effect on these aggregate variables. stream of its own benefits, and a reduction in the long-term variability of total income and well- Arguments for using lower longer-term discount being. This second benefit is what is known in rates based on intergenerational equity require the finance literature as a risk-reduction that the project has significant value to the well- premium. In some cases, it can have an impact on being of future generations, and that few, if any, evaluation similar to using a lower effective alternative investments can accomplish this end. For most adaptation projects, this condition is unlikely to be met. Use of lower longer-term 11 one suggestion to avoid such distortion is to apply, to all long-term projects, an annual discount rate that declines over time (referred to as "hyperbolic" discount- ing). however, this approach is still being debated in 12 an illustrative example is provided in Pindyck (2007); economics literature. see also howarth (2003) and Weitzman (2004). 30 D e v e l o P m e n t a n D c l i m at e c h a n g e discount rates in this context should be limited to Similar to the economic analysis methods, those large-scale and long-impact projects that MCDA is a utility-based approach, where the meet the conditions sketched above. In principle, "best" alternative is the one that results in the concern for the well-being of future generations most preferred probability-weighted outcome are better addressed in the project assessment by (Kenney and Raiffa 1993; Morgan and Henrion considering the tradeoffs and ethical constraints 1990). MCDA has taken many forms, but each that individuals are prepared to address in the approach has the same general steps: present.13 A more conventional present-value calculation would show how sensitive the evalua- · Identify the broad objective of the decision tion might be to the valuation of long-term bene- maker and operationalize it through multiple fits; this information can help decision makers qualitative and quantitative criteria, which form judgments regarding the emphasis that need to be both comprehensive and measur- intergenerational concerns might be given in able (Kenney and Raiffa 1993). project selection. · Assign weights to decision criteria based on the decision maker's subjective preferences. This can be done in a participatory setting, 3 .2 .2 A non- eC onomiC by eliciting the subjective preferences of A ss essment ApproACh--multi- stakeholders and trying to reach a common Cri t eri A deCision AnAlysis set of weights among different stakeholders through a consensus-reaching process led by Often, decision makers need or want to evaluate a facilitator. alternatives across a range of different and poten- tially incommensurate criteria. This is especially · Identify the utility function as a function of true in the context of agriculture and climate decision criteria and their associated weights. change, where an adaptation project can help It is common to assume a linear and additive reduce the negative effects of climate change on a function, but other functional forms are pos- number of social, environmental and economic sible as well (Kenney and Raiffa 1993). indicators. There also may be many instances, as · Identify the alternatives to be considered. already noted, when information on the monetary · Identify likely states of the world in which value of potential benefits or their likelihood of the alternatives might play out and the likeli- being realized is scarce and significant amounts hood of those states. This can be based on of informed judgment must be substituted14. In empirics or expert judgments (e.g., the such cases, multi-criteria decision analysis Delphi method, see Dalkey 1969, and (MCDA) can be useful. Kenney and Raiffa 1993). · Estimate the payoffs of the alternatives for 13 for more detailed consideration of these and other each state. issues related to discounting, see Portney and Weyant (1999). · Choose the alternative with the preferred 14 for example, in gauging the impact of climate change outcome (i.e., maximum expected utility). on ecosystem services and the benefits of adaptation measures (i.e., to combat land degradation), one approach might be to conduct structured interviews This method helps decision makers structure with affected local citizens who collectively could pos- complicated problems and systematically evaluate sess a great deal of qualitative information on how prior changes in ecosystem conditions affected productivity. alternatives. Other advantages include the fact this may be more useful than seeking to directly gauge that the method is easy to administer and trans- an economic value of avoided ecosystem damages through survey-based methods. parent, and allows for active involvement by e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 31 diverse participants and for qualitative values. In this section, we examine three alternative Clearly, this method relies heavily on individual approaches to deal with uncertainty. First, we judgments and subjective probability assessments. return to the use of probabilistic methods, which In judging its potential applicability, one needs to can be used for addressing reduction in risks from consider how explicit decision makers are about extreme events. We then discuss "real option their objectives and values, and the consequences, analysis," which reflects the state of the art in if a lack of consensus exists among the various economic evaluation under uncertainty, but, thus stakeholders. Box 12 discusses the application of a far, remains difficult to apply in concrete cases. simplified form of MCDA for identifying and Finally, we take up "robust decision making," an evaluating adaptation options in three Latin approach based on more heuristic evaluation American countries. methods, but able to be applied even in situations of high uncertainty over future states of the world. Each of these approaches addresses some 3 . 2. 3 deAl ing with unCertAin t y of the abovementioned limitations, but none addresses all of them. In drawing our conclusions We have already noted how environmental, tech- in Section 4, we offer a few specific suggestions nical and economic uncertainties permeate the on how to address uncertainty in evaluating evaluation of climate change adaptation. adaptation when practically available methods Economic evaluation with uncertainty usually and data are limited. considers scenarios judged to have various degrees of likelihood. For example, "high impact" and (i) Cost-benefit analysis of risk reductions "low impact" scenarios implicitly are deemed less For some adaptation initiatives, it may suffice to likely than an "anticipated impact" scenario. More be able to economically evaluate how the project sophisticated extensions of this approach will reduces risks and expected monetary losses asso- postulate explicit probability distributions for key ciated with an uncertain adverse agricultural factors, construct an implied distribution of impact. This might be the case, for example, results (in terms of NPV), and examine the mean when the impacts from climatic extremes are a (or median) and variability of the net benefits. primary concern. As noted in Section 3.1.3, work in the field of disaster risk reduction suggests There are three drawbacks to these approaches some potential approaches along these lines when evaluating adaptation. First, they assume (Proventium Consortium 2008b). Just as one can knowledge of probabilities about which we may assess, with this method, the ways that climate in fact know fairly little. Second, they typically change might alter the probabilities and expected treat probabilities as given, when the purpose of consequences of impacts of varying size and some adaptation is to reduce risks (defined as the frequency, one can examine how these factors probability of occurrence of threatening events). might be reduced by different resilience-increas- Finally, they do not incorporate the possibility of ing interventions (e.g., stronger flood protection) decisions that would, as in real life, unfold over and compare that to the cost of the interventions. time as circumstances change and new knowledge is gained. In such conditions, there is, in fact, an This method is better suited to disaster-oriented economic value to being able to maintain a larger adaptations versus adaptation to less extreme set of options, over and above whatever expected climate change impacts. Aside from the problem NPV would be calculated in scenario-based of estimating probabilities already noted, one approaches. challenge with this approach is how to identify 32 D e v e l o P m e n t a n D c l i m at e c h a n g e Box 12 multi-criteria Priority setting for aDaPtation Decisions in latin america the focus of this study -- carried out in three latin american countries with very different agro-climatic and socioeconomic characteristics -- was on identifying relevant climate changes in selected agro-eco- systems in latin america, and formulating adaptive response options that can be used to develop local action plans, which will, in turn, support informed responses in the future. the study adopted a "bottom- up" approach, in which response options were identified and prioritized by local stakeholders. this approach was chosen because it maximizes the likelihood that the adaptation measures, which are ulti- mately chosen, will be realistic and feasible to those who are familiar with local circumstances and will make resource management decisions. in a series of three workshops and intervening work by local teams in each country, local stakeholders ­ farmers, farmer organization representatives, agronomists and technical experts, extensionists and other stakeholders ­ were closely involved in: · identifying current climate changes and their implications for local agricultural systems, rural liveli- hoods and local people. · identifying possible response options ­ technical, institutional and policy ­ to support local adapta- tion strategies to climate change. · Prioritizing these possible response options in the form of activities and initiatives that will form local action plans. a formal priority-setting methodology, very similar to a mcDa approach, was used to establish priorities among alternative adaptation options. the completion of the prioritization exercise involved three com- ponents: 1. identification and weighting of a number of criteria: workshop participants in each country were asked to allocate 100 points among eight impact criteria and another 100 points among six viability criteria. 2. elaboration of the characteristics of each response option: "profiles" of each of the response options identified by stakeholders were developed, including information on: (i) the underlying need for the response option; (ii) technical characteristics; and (iii) a rough indication of costs and bene- fits. 3. assigning values to each of the criteria as applied to each of the response options in order to gen- erate a final prioritized ranking of the options: participants were given a matrix and asked to assign a value from 1 to 10 based on the extent to which they believed each criterion was effectively addressed by each response option. the participants' ratings of each response option were then weighted by criteria weights previously elicited; the impact criteria were proportionately assigned 50% of the overall score, while the other 50% was assigned proportionately to the viability criteria. as a result, a ranking of adaptation options was obtained. Source: World Bank 2009c. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 33 potential autonomous adaptations and their options created and destroyed by the project. This impacts on the benefits of the planned adaptation approach is particularly useful for: initiatives. Influences of autonomous adaptation would have to be specifically incorporated into · Evaluating investments under dynamic the "risk-exceedance" curves discussed in 3.1.3. uncertainty, i.e., investments whose value is highly sensitive to uncertainty over the future (ii) Real option analysis state of the world, and where the degree of uncertainty may vary (generally decrease) The real option methodology is based on the idea with time. that some investment projects can be evaluated as a set of compound options. Just as a financial · Evaluating investment decisions that can be option is defined as the ability, but not the obli- phased, i.e., when decisions such as making a gation, to buy or sell an underlying security at a certain investment, scaling up the project, fixed price, a real option can be defined as the abandoning the project or switching to dif- ability to undertake a future economic action or ferent activities can be delayed to a future project. An adaptation measure may maintain time (Knudsen and Scandizzo, 2004), such existing options or even create additional ones. as when more information on climate change For example, a water management project may impacts and on consequences of different help a community to preserve the option of project alternatives becomes available. remaining in place rather than migrating if future · Taking into account different irreversibilities climate change makes local livelihoods infeasible. (in future world conditions, as in certain Even though it is not possible to know in types of climate change impacts, and in the advance the severity of future climate change or range of future choices determined by current the ability of the investment to forestall reloca- investment decisions) and future options that tion, the project provides a choice that otherwise are conditioned by present choices (Ambrosi, would not be available. 2004). As indicated above, a number of adaptation proj- Because an investment often commits scarce ects can be considered natural candidates for the resources in an irreversible way under uncertainty, application of this methodology, since: another option for the decision maker to consider is whether to undertake a project now or to wait. Once the project has been implemented and the 1. their economic value crucially depends on the investment cost has become a "sunk cost," the future state of the climate, which is unknown waiting option is eliminated. The economic value when the decision to undertake a project is of waiting can, in principle, be calculated and made; compared to the economic value of the project 2. irreversible impacts may materialize in the (net of the waiting option value). Thus, option future if no action is taken today to prevent value considerations can affect the timing, as well this from happening (i.e., desertification of as the nature and scale, of adaptation initiatives. agricultural land); Real option analysis makes it possible to extend 3. some investments, especially in infrastruc- the standard NPV methodology through ture, can be considered irreversible in the combining traditional cost-benefit analysis under sense of locking in capital for decades (i.e., a uncertainty with value estimates of the real water reservoir of a predetermined capacity); 34 D e v e l o P m e n t a n D c l i m at e c h a n g e 4. phased projects constitute a common and adoption, beyond the standard cash flows consid- reasonable choice, when it is recognized that ered by the NPV approach. The economic value learning during the initial phase of the proj- of the project is calculated by adding the value of ect will allow improving project design for the options created by the project (i.e., option to subsequent phases, which is often the case in adapt, to abandon, to scale up, etc.) and subtract- new development fields, such as adaptation ing the options destroyed by the project (i.e., to climate change; and option to wait or to make an alternative invest- ment) to the standard NPV (Knudsen and 5. adaptation investments are meant to create Scandizzo 2004). As a consequence, a positive future options for its beneficiaries (i.e., "extended NPV" result is an indication that the capacity building on drought resistant crops, project is economically viable even if the standard water harvesting techniques, use of seasonal NPV is negative (Box 13, Figure 3). climate forecasts, or alternative livelihoods to agriculture are examples of soft adaptations Investment evaluation according to real option that generate a sustained livelihood/climate analysis can be done according to two different resilience option for farmers in the future). technical procedures, namely the decision tree Evaluating a project through real option analysis analysis and the binomial analysis. The main can also be considered as a new form of risk anal- problem associated with the option evaluation ysis, where risk is identified both positively, as the methods is that they require estimating probabili- contingent wealth of opportunities created by the ties associated with future outcomes. When this project, and as a cost, in terms of the contingent cannot be done based on scientific grounds, such liabilities that the project may generate as in the case of climate change, subjective proba- (Scandizzo 2008). One example is in the form of bilities are solicited through stakeholder consulta- threats generated by future climatic conditions, tions, by identifying possible future scenarios and which are different from ex-ante projected rain- assessing the likelihood of each, according to the fall data on the project area. An important stakeholders' viewpoints. Given the crucial element of project evaluation through real option importance of probabilities in determining the analysis is thus the estimate of risk, which may be economic value of the project, a main pitfall of counteracted by project actions, whose costs are this approach is the reliability on subjective prob- also evaluated. This form of risk analysis there- abilities. On the other hand, the participatory fore takes the form of an examination of the process of scenario creation and evaluation facili- project structure, and a largely qualitative tates the identification of a flexible project design appraisal of assets and liabilities corresponding to that expands the capabilities to deal with climate a number of project alternatives. Such a process variability and change, especially for the poor leads to the identification of those alternatives, (Scandizzo 2008). which increases the likelihood of project success in the context of uncertainty. As a consequence, Because of the constructive nature of evaluation, evaluation becomes an interactive and construc- a full application of real option analysis requires tive task, particularly in the early phases of proj- engaging the methodology along the entire proj- ect preparation. ect cycle, from the very beginning of project conception and design. In general, in order to Hence, real option evaluation involves recogniz- identify and evaluate the options within a project, ing and identifying capabilities and opportunities the following three steps are necessary (Scandizzo created or destroyed as a consequence of project 2008): e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 35 Box 13 aPPlication of real oPtion analysis to an irriga- tion ProJect in mexico a case study on an adaptation project in the agricultural sector in mexico has been conducted to show the potential of applying real option analysis for the evaluation of adaptation to climate change. the development objective of the rio conchos Basin project is to improve the efficiency of irrigation water use, thus promoting no-regret adaptation to climate change in irrigated agriculture. one of the project's components is the modernization of the existing irrigation infrastructure. the irrigation modernization component alone would create economic benefits solely from water sav- ings, estimated in us$5.6 million. the net present value (nPv) of the project, also considering operating costs, would be about us$25 million against investment costs of about us$317 million. hence, this proj- ect component does not appear acceptable on traditional grounds, since the nPv is negative by us$292 million. however, because of uncertainty and the possible impact of climate change, however, the analysis is extended to evaluate the adaptation options opened by this basic project. the most important of these options consists of the fact that fine tuning irrigation systems gives the opportunity to profitably extend cultivation to higher-value crops, if warranted by sufficiently favorable circumstances and provided that further investment is undertaken in the form of plant protection by plastic coverage (plasticulture). this "growth option" would create three major benefits respectively from water and energy savings and from increasing revenues. a more important way in which the project can help adaptation to climatic changes for the farmers of the project area is by reducing the threat of the lowering of the water table and the increasing danger of saline intrusion. this is a catastrophic threat linked to increasing aridity that climate change would be likely to cause. even though irrecoverable damages to the water table would occur for natural causes, these contingent damages correspond to a "liability option," held by an impersonal agent (e.g., mother nature). its value can be estimated as that of an option to significantly reduce or even destroy farmers' income (the underlying asset). the underlying asset of such an option is the opportunity cost of the water saved with the implementation of the basic project plus the value of the adaptation option (because they would both be lost if the water table were contaminated). the value of the strike, on the other hand, is the threshold of water utilization (in terms of its economic value) at which it is reasonable to expect that water contamination would occur. figure 3 summarizes the results of the analysis, assuming different volatilities, i.e., degrees of uncertain- ty (x-axis). it shows that, despite a negative nPv for volatility < 50%, the project appears highly profit- able on economic grounds, by virtue of its potential effect on active adaptation (the growth option), and also of its possible effectiveness in removing a liability option, which is a consequence of climate change. (continued) 36 D e v e l o P m e n t a n D c l i m at e c h a n g e (Box 13 continued) figure 3 aPPlication of real oPtion to an irrigation ProJect in mexico Traditional NPV vs. Extended NPV 1,200 1,000 800 600 Million US $ 400 200 0 20% 30% 40% 50% 60% 70% 80% 90% 100% -200 -400 Volatility NPV NPVes Source: scandizzo and notaro (2008). 1. Identifying the options. This first step aims analysis, the project team and other consul- to determine the consequences of project tants who carry out project preparation implementation on the capabilities (i.e., studies. higher capacity to respond to climate change) · Assessing the opportunities and threats that and opportunities (i.e., new technologies climate change can generate, such as and/or possible favorable consequences of cli- increases in aridity, changes in the level and mate changes) of its stakeholders. This phase distribution of precipitation, water contami- generally requires stakeholder involvement nation, changes in temperature throughout (i.e., through focus groups) in defining sce- the year, increased likelihood of extreme narios and identifying capabilities and meteorological events. Probabilities need to options created by the project. be assigned to future possible outcomes. 2. Analyzing the options. This second phase · Applying the option algorithms to calculate consists of: the option values. This step requires choos- ing among two alternative option evaluation · Designing the different components of the methods, depending on the circumstances, project, taking into account the results of the and calculating the project value. first phase, and identifying benefits and costs for the project's duration. This step is identi- 3. Evaluating the opportunity of acting on or cal to a standard feasibility study based on exercising the option. This third phase is the NPV method, and requires close collabo- aimed at determining whether or not to ration of the team working on the real option "exercise the option". For example, if the e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 37 option under consideration is the option to RDM is an iterative evaluation process. Once an wait, the assessment will indicate whether it ensemble of scenarios is generated, each alterna- is economically more efficient to implement tive action can be systematically compared the project (or a specific activity) now, or to according to a range of criteria, as in MCDA. For wait for additional information about input example, adaptation efforts could be evaluated or output data. The evaluation process should according to anticipated effects on yields given a incorporate a sensitivity analysis to test the climate scenario and an assumption about the sensitivity of option values to the parameter productivity and cost of the intervention, the estimates. differential effects across different economic subgroups of farmers, and performance if climatic (iii) Robust decision making conditions turn out much worse than anticipated All methods discussed above rest on some form in the scenario under consideration. of an expected benefit or utility analysis with a quantitative characterization of the uncertainties. Of particular interest is the identification for any In the context of climate change, however, often given alternative action of the set of conditions uncertainties are so profound that there is little where it performs poorly according to the various information or consensus on what probability criteria, reflecting, again, the judgment of the distributions to consider for input variables, how decision maker. A "robust" alternative is one that, to rank alternatives and what scenarios to compared to other alternatives, performs reason- consider in the analysis. The term "deep uncer- ably well across a wide range of plausible futures. tainty" has been used by Lempert and others In other words, it is one whose payoff is insensi- (2003) to describe the situation where decision tive to poorly characterized uncertainties makers lack the knowledge or consensus about (Lempert and others 2006). the system model that relates alternative courses of action to outputs of interest, distribution prob- A strong advantage of RDM over other methods abilities on the inputs to the system model, or for dealing with uncertainty is that it provides a value functions that rank the desirability of the means to evaluate alternatives even when there is outcomes of interest. lack of knowledge or disagreement on prior prob- abilities and benefit estimates. RDM also allows Robust decision making (RDM) can provide an decision makers to make better informed alternative quantitative decision analytic method tradeoffs in deciding on the desirability of alter- that avoids subjective probability assessments and natives (Box 14). The key disadvantage of RDM scenario predictions (Lempert and others, 2003). at this stage is that it is still a research tool In RDM, uncertainties are not framed with requiring the use of complicated computer algo- prespecified probability distributions over input rithms and software, as well as depending on the parameters to the system model. Moreover, the ability to construct a large range of plausible future "states of the world" that are considered in future scenarios from whatever information is the analysis are not limited to few subjective available. Significant work will be needed to scenarios. Instead, RDM creates hundreds or adapt this approach for use in evaluating specific thousands of plausible futures, in the judgment of projects. the analyst, that are then used to systematically evaluate the performance of alternative actions (Bankes and others, 2001). 38 D e v e l o P m e n t a n D c l i m at e c h a n g e Box 14 rDm for aDaPtation Decisions in the Water sector the inland empire utilities agency (ieua) is a water-supply agency in a rapidly growing area of south- ern california. Because of rapid population growth in an already dry area, ieua confronts the prospect of major investments in acquiring additional water supplies and in replenishing groundwater through recycling. the prospects of climate change add to the complexity of the planning task. the effects of future climate change on precipitation and runoff in the ieua area are uncertain; it is necessary to weigh the possibility of wetter than historical conditions with more natural recharge of aquifers, and hotter, dryer conditions with more rapid evaporation and less recharge. an analysis of "robust" water management options for ieua in the face of climate change uncertainty found that because the cost of water shortage was high if the latter conditions occurred, it made sense for the agency to invest in more water conservation as well as in recycling as a kind of hedge, if decision makers perceived the chance of those conditions to exceed 25 percent. this extra investment was not warranted when one evaluated alternative plans by their expected present value of net benefits, without factoring in hedging value. the approach taken in the study was highly interactive, with decision makers working in tandem with analysts to ascertain policy-relevant scenarios to consider and the costs of short- age. Source: groves and others 2008. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 39 4. ConClUSIonS SoME BASIC STEpS foR pRojECT-lEVEl EConoMIC EVAlUATIon of AdApTATIon The main purpose of this paper was to discuss activities and appraisal may pay significant divi- methodological options to assess the economic dends in improved project design. Options exist soundness of adaptation investments at the for employing simplified versions of some meth- project level, with a focus on agriculture. The odologies, as discussed below. choice of specific approaches to evaluate a particular adaptation project will be dictated We propose below a series of steps for carrying by the characteristics of the project, specific out the economic analysis of an adaptation proj- questions of interest, existence and accessibility ect, and discuss the potential utilization of the of data, and skills of available experts. Table 3 methodological options presented in this paper provides a summary of the main characteristics of for project-level analysis. Many of the evaluation the proposed methodological options. steps listed here (namely steps 3, 5, 6 and, even- tually, 8), can greatly benefit from knowledge Although time, budget and data limitations exchange with stakeholders. constitute obvious constraints in the use of the methods discussed (especially those involving 1. Develop information on relevant climate more technically complex modeling, statistical risks for the project area and specify the tem- assessments and/or probability calculations), they poral horizon of the analysis, clarifying, in suggest concrete steps in assessment that, if particular, the extent to which the focus of applied early enough in project preparation, can the intervention is on dealing with increasing be useful to inform project design (e.g., to select climate variability and extremes, or longer- the most suitable crops to local climate condi- term change in climatic mean values. tions and management practices, or to design Guidance on climate risk assessment can be project components that are likely to maximize found in World Bank (2009d). benefits for local communities according to their own judgment). In cases where the stakes are 2. Assess several possible without-project sce- large in terms of project resilience, a higher narios by estimating the impacts of different budget than initially envisaged for preparatory climate variability/climate change projections 40 D e v e l o P m e n t a n D c l i m at e c h a n g e on agricultural productivity and other rele- without-project case. Depending on the type vant measures of output and benefit. The of adaptation envisaged under the project, Ricardian approach has generally been better both crop and Ricardian models can be con- suited to country-scale assessments, while sidered for estimating the effects of adapta- crop models may be more easily scaled to tion, but their different ways of accounting local-level analysis. Developments in the near for autonomous and planned adaptation future might both reduce the complexity of need to be factored into the choice of meth- methods and increase their applicability to odology. Although using the same approach project-level analysis (Box 15). applied to the without-project scenarios 3. Identify the types of adaptation projects (e.g., maintains consistency, it may also be possi- stand-alone or integrated within a broader ble to utilize different methods to evaluate development project), and possible adaptation the two cases (see Box 6), when the specific measures (soft/hard) that the project could circumstances suggest that this is a better support and implement. way of comparing with- and without-project scenarios (e.g., to better reflect impacts of 4. Estimate potential reductions in projected autonomous adaptation). productivity losses (i.e., adaptation benefits) in with-project situations, under the same 5. Quantitatively and qualitatively assess, as multiple scenarios used for examining the appropriate, any co-benefits and negative Box 15 toWarD a more straightforWarD aPPlication of ricarDian anD croP moDels to ProJect-level imPact assessment ricardian analysis can already be applied to project-level assessment, when a ricardian function, devel- oped within a country-level study, can be utilized for estimating local impacts (i.e., the function's parame- ters estimated for the country remain the same as in the original study, while the value of the climatic variables and other control variables is substituted by local data). additional research is needed to allow using ricardian functions developed for other countries with similar characteristics (a sort of "benefit transfer" approach), in the absence of a study for the same country. an even coarser approach is to sim- ply apply country-level impact estimates to the local level. such estimates are readily available (cline 2007), but their direct application is not recommended, as local topography and land use can greatly affect both how climate change will materialize and how these climatic changes will affect local produc- tivity. two different directions of future research in crop modeling might help develop more readily applicable tools for local level analysis. first, a simplification of the models themselves and the development of a more user-friendly interface could constitute welcome advancements toward a more widespread applica- tion of these tools at local level, despite the obvious trade-off with the precision of results. second, simi- larly to the ricardian methods, possibilities of benefit transfer approaches between different areas with similar local characteristics should be explored. Benefit transfer for both ricardian and crop model esti- mates would call for an easily accessible and user friendly database with enhanced features. Source: authors. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 41 spillovers that the project may bring about scenarios, based on estimated adaptation compared to a non-project situation. costs and benefits from previous steps. Stakeholder consultations may be particu- Different approaches may be advisable larly useful at this step. depending on the types of adaptation measures: 6. Consider opportunities that the project may create in the future (i.e., through knowledge · A probabilistic benefit-cost approach may be development or capacity building), as well useful if a primary focus is adaptation to any options that the implementation of the extreme events. project may destroy, and the effects that the · For some types of soft adaptation or in other project may have on autonomous adaptation cases where monetization of benefits is espe- and adaptive capacity. cially challenging, a multi-criteria approach 7. Attempt some economic estimation of may be useful, though its subjectivity needs future options maintained or lost. Future to be recognized and incorporated into the research may make the application of real appraisal process (e.g., through participation option theory easier; in the meantime, some of different evaluators). insights can be gained by examining several 9. "Stress-test" the project to identify particular scenarios with and without different choices investments and soft adaptation initiatives available in order to get a rough idea of whose benefits are particularly vulnerable to what, if any, options are more important to changes in conditions, and investigate poten- be maintained or created by the project. tial project modifications that can reduce 8. Assess how different alternative project vulnerability to climate and other future options perform under different climate shocks. taBle 3 summary of methoDologies Methodology Suitability of the methodology with respect to: Increased cli- Economic Evaluation of Evaluation mate variabili- evaluation at autonomous of soft ty/extremes the project and planned and hard and climate Modeling of Precision of level adaptation adaptation change uncertainty results Crop models High if an eco- Medium for both High for both High for both Medium can Medium: nomic module autonomous and hard adapta- climate change simulate the although capa- is integrated in planned adapta- tion (i.e., (i.e., changes effects of ble of generat- the model. tion: specific mea- increased in average tem- future climate ing very precise sures are decided water avail- perature and scenarios field-level yield by the analyst, with ability due to precipitation) through estimates, crop no reliance on new dam) and climate weather gen- models are empirical data. and soft variability/ erators affected by over- adaptation extremes (can (monte carlo­ estimation of cli- (i.e., the model the type simula- mate change effect of effects of tion). impacts and training can droughts and, either under- or be modeled to some extent, over-estimation to some floods). of the effects of extent by adaptation assuming a (depending on change from the analyst's suboptimal assumptions). to optimal management practices). (continued) 42 D e v e l o P m e n t a n D c l i m at e c h a n g e (Table 3 continued) Methodology Suitability of the methodology with respect to: Increased cli- Economic Evaluation of Evaluation mate variabili- evaluation at autonomous of soft ty/extremes the project and planned and hard and climate Modeling of Precision of level adaptation adaptation change uncertainty results Ricardian Medium: the High for autono- Medium to High for climate Low account- Precision of method method is mous adaptation high for both change/climate ed for only results depends more suited (i.e., crop switch- types, variability; low through differ- on the method for evaluation ing, change in irri- depending to medium for ent future cli- (i.e., structural at the region- gation practices); on what climate mate approaches al/country medium to high for explanatory extremes (i.e., scenarios. accounting for level, but planned adaptation variables are recent applica- agroecological application at (i.e., new irrigation in the model tions have a zones are more the project systems). (see previ- built-in flood precise) and on level is possi- ricardian models ous column). damage func- country charac- ble if a are well suited to tion). teristics (i.e., if a ricardian forecasting "with- sufficiently wide function has out-project" scenar- range of cli- been estimat- ios, accounting for mates already ed at the a comprehensive exist in the country level range of autono- country, the and local cli- mous adaptation; impacts of cli- matic vari- the capacity to mate change in ables can be assess the impacts a particular area substituted in of planned adapta- are more easily the equation. tion depends on estimable). what kinds of explanatory vari- generally pro- ables are captured vides lower in the model. impact esti- mates than crop models since autonomous adaptation is built in. MCDA Generally low High for both, but High for High for both. Medium, Precision in current more meaningful both. through prob- depends on how applications, for planned. ability weight- project perfor- but possibly ed scenarios. mance with high, if eco- respect to each nomic criteria decision criteri- (costs and on is estimated: benefits) are increasing preci- included sion from stake- among the holder-based, to decision crite- expert-based, to ria. model-based estimates. Probability- Low: this High for some High for Potentially high Currently low, Low due to based method is planned adaptation hard, low for for climate vari- but potentially imprecise proba- approach more suited measures (mainly soft adapta- ability and high, through bility distribu- for evaluation hard). tion. extremes. probability dis- tions (because of specific tributions of of scarcity of adaptation Low for soft adap- Low for climate climatic vari- empirical data measures, not tation. change. ables under on extremes) of a compre- future climate and uncertain- hensive proj- scenarios. ties of future ect. probabilities under climate change. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 43 (Table 3 continued) Methodology Suitability of the methodology with respect to: Increased cli- Economic Evaluation of Evaluation mate variabili- evaluation at autonomous of soft ty/extremes the project and planned and hard and climate Modeling of Precision of level adaptation adaptation change uncertainty results Real Option High, espe- High for planned High for hard High for both. Medium: dif- Low due to the Analysis cially if the adaptation, includ- interventions ferent future many assump- method is ing interventions with irrevers- states of the tions necessary properly that increase ible invest- world can be to calculate the applied early autonomous adap- ments. considered, extended net in project tive capacity. with related present value preparation. Potentially (subjective) (strike, volatility, high for soft probabilities. value of underly- interventions. ing asset, etc.) Robust High: can be Medium to low for Medium to High for both. High: repre- Medium: able to Decision scaled to proj- autonomous adap- high for hard sentation of highlight vulner- Making ects of varying tation: normally investments uncertainty is abilities of differ- sizes. very reduced-form and medium a critical ele- ent project plans models are used to for soft adap- ment of the but does not generate scenarios tation. approach. provide precise for assessment. measures of Medium for payoffs. planned adaptation Source: authors. 44 D e v e l o P m e n t a n D c l i m at e c h a n g e e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 45 REfEREnCES Agrawal, A., M. Kononen, and N. Perrin. 2009. "The Role Callaway, J.M. 2004. "The benefits and costs of adapting to of Local Institutions in Adaptation to Climate climate variability and climate change." In: The Change." Social Development Department Working Benefits of Climate Change Policies, Chapter 4, pp. 123- Paper 118 ( June). Washington, DC: The World Bank. 169. Paris: OECD. Available at: Cline, W.R. 2007. "Global Warming and Agriculture. http://siteresources.worldbank.org/ Impact Estimates by Country." Washington, DC: EXTSOCIALDEVELOPMENT/ Center for Global Development and Peter G. Peterson Resources/244362-1164107274725/sdp118.pdf. Institute for International Economics. Agrawala, S. and S. Fankhauser, eds. 2008. Economic Corderi, D. 2009. "Climate Change and Agricultural Aspects of Adaptation to Climate Change -- Costs, Impacts: Methodology for Measuring Adaptation Benefits and Policy Instruments. Paris: OECD. Options." Background paper for the report "The Economics of Adaptation to Climate Change." Agriculture Canada, Regional Development Branch. 1984. Washington, DC: The World Bank (in preparation). Technology Transfer in Agriculture: What It Is and How It Occurs. Ottawa: Agriculture Canada (August). Dalkey, N.C. 1969. The Delphi Method: An Experimental Study of Group Opinion, RM-5888-PR. Santa Monica, Ambrosi, P. 2004. Amplitude et Calendrier des Politiques de CA: RAND Corporation. Réduction des Émissions face aux Risques Climatiques-- Leçons des Modèles Intégrés [Magnitude and Timing of Eid, H. M., S.M. El-Marsafawy, and S.A. Ouda. 2007. Mitigation Policies: Lessons from Integrated "Assessing the Economic Impacts of Climate Change Assessment Models]. PhD dissertation, Ecole des on Agriculture in Egypt. A Ricardian Approach." Hautes Etudes en Sciences Sociales (EHESS), Paris. Policy Research Working Paper 4293. Washington, DC: The World Bank. Available at: http://intranet.world- Bankes, S.C., R.J. Lempert, and S.W. Popper. 2001. bank.org/servlet/main?pagePK=64161651&theSitePK "Computer-assisted reasoning." Computing in Science =84798&piPK=64161652&menuPK=64166238&entit & Engineering, Vol. 3, Issue 2, March/April 2001, pp. yID=000158349_20070731143402. 71 ­ 77. Essam, T. 2009. Untitled background document for the Bosello, F., O. Kuik, R. Tol, and P. Watkiss. 2007. "Costs of report "The Economics of Adaptation to Climate adaptation to climate change: A review of assessment Change": The World Bank (in preparation). studies with a focus on methodologies used." Final Report to the European Environmental Agency. 6th European Commission. 2008. "Policy Incentives for Specific Agreement No. 3602/B2005. EEA under the Climate Change Mitigation Agricultural Techniques Framework Contract No. EEA/AIR/04/004. (PICCMAT)." 6th Framework Programme of the 46 D e v e l o P m e n t a n D c l i m at e c h a n g e Research DG. Available at: http://www.climate- Knudsen, O.K., and P.L. Scandizzo. 2006. "The Artful changeintelligence.baastel.be/piccmat/index.php. Face of Uncertainty: How to Manage Opportunities and Threats." Washington, DC: The World Bank. Fankhauser, S. 2006. The Economics of Adaptation. Background Document for the Stern Review. http:// Kurukulasuriya, P., and S. Rosenthal. 2003. "Climate webarchive.nationalarchives.gov.uk/, and http://www. Change and Agriculture: A Review of Impacts and hm-treasury.gov.uk/media/5/B/stern_review_support- Adaptations. World Bank Climate Change Series." ing_technical_material_sam_fankhauser_231006.pdf. Washington, DC: World Bank Environment Department. Gornitz, V., K. Jacob, and C. Rosenzweig. 2007. Kurukulasuriya, P. 2006. Estimating Impacts of Climate "Vulnerability of the New York City metropolitan area Change on Agriculture. PhD thesis, Yale University. to coastal hazards, including sea-level rise: Inferences for urban coastal risk management and adaptation Kurukulasuriya, Pradeep, Robert Mendelsohn, Rashid policies." In Managing Coastal Vulnerability. L. Hassan, James K.A. Benhin, Temesgen Deressa, McFadden, R. Nicholls, and E. Penning-Rowsell, eds. Mbaye Diop, Helmy Mohamed Eid, K. Yerfi Fosu, Elsevier, pp. 139-156. Glwadys Gbetibouo, Suman Jain, Ali Mahamadou, Renneth Mano, Jane Kabubo-Mariara, Samia Groves, D.G., D. Knopman, R.J. Lempert, S.H. Berry, and El-Marsafawy, Ernest Molua, Samiha Ouda, Mathieu L. Wainfan. 2008. Presenting Uncertainty About Climate Ouedraogo, Isidor Séne, David Maddison, S. Niggol Change to Water-Resource Managers. Technical Report Seo, and Ariel Dinar. 2006. "Will African agriculture TR-505. Santa Monica, CA: RAND Corporation. survive climate change?" World Bank Economic Review, 2006, Vol. 20, Issue 3, pp. 367-388. Groves, D.G., and R.J. Lempert. 2007. "A new analytic method for finding policy-relevant scenarios." Global Kurukulasuriya, P., and M.I. Ajwad. 2007. "Application of Environmental Change, Vol. 17, No. 1, pp. 73-85. the Ricardian technique to estimate the impact of cli- mate change on smallholder farming in Sri Lanka," Gulledge, J. 2008. "Climate Change Risks in the Context Climatic Change, Vol. 81, No. 1, March, 2007. of Scientific Uncertainty." Global Politics of Energy, The Aspen Institute. Kurukulasuriya, P., and R. Mendelsohn. 2007. "Endogenous Irrigation: the Impact of Climate Heltberg, R., P. B. Siegel, and S. L. Jorgensen. 2009. Change on Farmers in Africa." Policy Research Working "Addressing human vulnerability to climate change: Paper 4278. Washington DC: The World Bank. towards a `no regrets' approach." Global Environmental Available at: http://intranet.worldbank.org/servlet/mai Change, Vol. 19, No. 1, pp. 89-99. n?pagePK=64161651&theSitePK=84798&piPK=6416 1652&menuPK=64165676&entityID=000016406_20 Howarth, R. B. 2003. "Discounting and uncertainty in cli- 070706160116. mate change policy analysis." Land Economics, Vol, 79, Kurukulasuriya, P., and R. Mendelsohn. 2008. "Crop No. 3, pp. 369-81. switching as a strategy for adapting to climate change," IPCC. 2007. Climate Change 2007: Impacts, Adaptation and African Journal of Agricultural and Resource Economics, Vulnerability. Contribution of Working Group II to the Vol. 2, No. 1 (March), Pretoria. Fourth Assessment Report of the Intergovernmental Panel Lempert, Robert J., Steven W. Popper, and Steve C. on Climate Change, M.L. Parry, O.F. Canziani, J.P. Bankes. 2003. Shaping the Next One Hundred Years: Palutikof, P.J. van der Linden and C.E. Hanson, Eds. New Methods for Quantitative, Long-Term Policy Cambridge, UK: Cambridge University Press. Analysis, MR-1626. Santa Monica, CA: RAND Corporation. IPCC. 2009. Carbon dioxide: projected emissions and con- centrations. http://www.ipcc-data.org/ddc_co2.html. Lempert, Robert J., David G. Groves, Steven W. Popper, and Steve C. Bankes. 2006. "A general, analytic meth- Keeney, R. L., and H. Raiffa. 1993. Decisions with Multiple od for generating robust strategies and narrative sce- Objectives. Cambridge, UK: Cambridge University narios." Management Science, Vol. 52, No. 4, April. Press. Lesser W., W. Magrath, and R. Kalter. 1986. "Projecting Knudsen, O.K., and P.L. Scandizzo. 2004. "Real Options adoption rates: application of an "ex-ante" procedure to and Project Evaluation: A Primer." Washington, DC: biotechnology products." North Central Journal of The World Bank. Agricultural Economics, Vol. 8, No. 2, July, pp. 159-174. e c o n o m i c e va l u at i o n o f c l i m at e c h a n g e a D a P tat i o n P r o J e c t s 47 Mendelsohn, R. (2006). "The role of markets and govern- Rosenzweig, C., and F.N. Tubiello. 2007. "Adaptation and ments in helping society adapt to a changing climate." mitigation strategies in agriculture: an analysis of Climatic Change, Vol. 78, No. 1, pp. 203-215. potential synergies." Mitigation and Adaptation Strategies for Global Change, Vol. 12, Issue 5, pp. 855- Mendelsohn, R., W. Nordhaus, and D. Shaw. 1994. "The 873. impact of global warming on agriculture: a Ricardian analysis." American Economic Review, Vol. 84, No. 4, Sanghi, A., and R. Mendelsohn. 2008. "The impacts of pp. 753­771. global warming on farmers in Brazil and India." Global Environmental Change, Vol. 18, Issue 4, pp. 655-665.. Molua E., and C.M. Lambi. 2007. "The Economic Impact of Climate Change on Agriculture in Cameroon." Scandizzo, P. L. 2008. "Adaptation Projects and Real Policy Research Working Paper 4364. Washington, Options." Background paper for the report Mitigating DC: The World Bank. Available at: http://intranet. and Dealing With Climate Change in Latin America and worldbank.org/servlet/main?pagePK=64161651&theS The Caribbean. Washington DC: The World Bank (in itePK=84798&piPK=64161652&menuPK=64165676 preparation). &entityID=000016406_20070706160116. Scandizzo, P. L., and C. Notaro. 2008. "Adapting to Morgan, M. Granger, and Max Henrion. 1990. Climate Change: A Case Study of Project Evaluation Uncertainty: A Guide to Dealing with Uncertainty in through Real Option Theory." Background paper for Quantitative Risk and Policy Analysis. Cambridge, UK: the report "Low Carbon, High Growth: Latin-American Cambridge University Press. Responses to Climate Change--An Overview," by Augusto de la Torre, Pablo Fajnzylber and John Nash Nkomo, J. C., and B. Gomez. 2006. "Estimating and com- (2009). Washington, DC: The World Bank. paring costs and benefits of adaptation projects: Case studies in South Africa and the Gambia." Washington, Schenkler, W., W.M. Hanemann, and A.C. Fisher. 2006. DC: START Secretariat. Available at: http://www. The impact of global warming on U.S. agriculture: an aiaccproject.org/Final%20Reports/Final%20Reports/ econometric analysis of optimal growing conditions. FinalRept_AIACC_AF47.pdf. Review of Economics and Statistics, Vol. 88, No. 1, February, pp. 113-125. Padgham, J. 2009. "Agricultural Development Under a Changing Climate: Opportunities and Challenges for Sen, A. 1999. Development as Freedom. New York: Random Adaptation." Environment and Agriculture and Rural House. Development Department Joint Discussion Paper. Seo, S. N., and R. Mendelsohn. 2008a. "An analysis of crop Washington, DC: The World Bank. choice: adapting to climate change in South American Pindyck, R.S. 2007. "Uncertainty in environmental eco- farms." Ecological Economics, Vol. 67, pp. 109-116. nomics." Review of Environmental Economics and Seo, N., and R. Mendelsohn. 2008b. "Measuring impacts Policy, 2007 Vol. 1, Issue 1, pp. 45-65. and adaptation to climate change: A structural Portney, P.R., and J.P. Weyant, eds. 1999. Discounting and Ricardian model of African livestock management." Intergenerational Equity. Washington, DC: Resources Agricultural Economics, Vol. 38, Issue 2, pp. 151-165. for the Future. Seo N., R. Mendelsohn, A. Dinar, P. Kurukulasuriya, and Proventium Consortium. 2008a. "Disaster Risk Reduction R. Hassan. 2008. "Long-Term Adaptation: Selecting and Cost-Benefit Analysis." Available at: http://www. Farm Types across Agro-Ecological Zones in Africa." proventionconsortium.org/?pageid=26. Policy Research Working Paper 4602. Washington DC: The World Bank. Available at: http://intranet.world- Proventium Consortium. 2008b. "From Risk to Resilience bank.org/servlet/main?pagePK=64161651&theSitePK --Benefits and Costs of Disaster Risk Reduction." =84798&piPK=64161652&menuPK=64166238&entit Available at: http://www.proventionconsortium.org/?p yID=000158349_20080417145933. ageid=37&publicationid=158#158. Seo, N., R. Mendelsohn, A. Dinar, and P. Kurukulasuriya. Räisänen, J., and T.N. Palmer. 2001. "A probability and 2009. "Adapting to climate change mosaically: an decision-model analysis of a multi-model ensemble of analysis of African livestock management by agro-eco- climate change simulations." Journal of Climate, Vol. logical zones." Berkeley Electronic Journal of Economic 14, Issue 15, pp. 3212-3226. Analysis and Policy. Berkeley, CA: The Berkeley Electronic Press. Rogers, E.M. 1962. Diffusion of Innovations. New York: The Free Press. 48 D e v e l o P m e n t a n D c l i m at e c h a n g e Stern, N. H. 2007. The Economics of Climate Change: The World Bank. 2008a. Development and Climate Change. A Stern Review. Cambridge, UK: Cambridge University Strategic Framework for the World Bank Group. Press. Washington, DC: The World Bank. UNFCCC. 2008. "Compendium on Methods and Tools to World Bank. 2008b. Climate Change Impacts in Drought Evaluate Impacts of Vulnerability and Adaptation to and Flood Affected Areas: Case Studies in India. South Climate Change." Available at: http://unfccc.int/adap- Asia Region. Washington, DC: The World Bank. tation/nairobi_work_programme/knowledge_resourc- es_and_publications/items/2674.php. World Bank. 2008c. "China Mainstreaming Climate Change Adaptation in Irrigated Agriculture Project." Wang, J., R. Mendelsohn, A. Dinar, J. Huang, S. Rozelle Project Appraisal Document. Annex 9: Economic and and L. Zhang. 2007. "Can China Continue Feeding Financial Analysis. Washington, DC: The World Itself? The Impact of Climate Change on Agriculture. Bank. Policy." Research Working Paper 4470. Washington DC: The World Bank. Available at: http://intranet.world- World Bank. 2009a. "Implications of Climate Change on bank.org/servlet/main?pagePK=64161651&theSitePK Food Security in Bangladesh." Sustainable =84798&piPK=64161652&menuPK=64166238&entit Development Department Agriculture and Rural yID=000158349_20080303090028. Development Unit, South Asia Region. (Forthcoming.) Washington, DC: The World Bank. Weitzman, Martin L. 2004. Why the far-distant future should be discounted at its lowest possible rate. Recent World Bank. 2009b. "Adaptation to Climate Change in the Developments in Environmental Economics, Vol. 1, Agriculture Sector in Morocco." Middle East and pp. 429-36. Elgar Reference Collection. International North Africa Region. (Forthcoming.) Washington, Library of Critical Writings in Economics, Vol. 177. DC: The World Bank. Cheltenham, UK, and Northampton, MA: Elgar. World Bank. 2009c. "Vulnerability to Climate Change in Weitzman, M. L. 2007. "The Role of Uncertainty in the Agricultural Systems in Latin America: Building Economics of Catastrophic Climate Change." AEI- Response Strategies." Latin America and Caribbean Brookings Joint Center for Regulatory Studies, Region. (Forthcoming.) Washington, DC: The World Working Paper 07-11, May. Bank. World Bank. 2006a. "Project Appraisal Document on a World Bank. 2009d. "Mainstreaming Climate Risk Proposed Grant From the Global Environment Management and Adaptation to Climate Change in Facility Trust Fund for the Benefit of the Republic of Development Projects." Quick Reference Guidance Notes Colombia for the Integrated National Adaptation with Special Focus on Agricultural, Natural Resource Project." Report No: 34058-Co. Annex 9: Economic Management and Rural Development Projects. and Financial Analysis. Washington, DC: The World (Forthcoming.) Washington, DC: The World Bank. Bank. World Bank. 2009. "The Economics of Adaptation to World Bank. 2006b. "Not If But When: Adapting to Climate Change." In preparation. Draft report avail- Natural Hazards in the Pacific Islands Region." Policy able at: http://beta.worldbank.org/climatechange/con- Note, East Asia and Pacific Region, Pacific Islands tent/ Country Management Unit. Washington, DC: The economics-adaptation-climate-change-study-homep- World Bank. age. THE WORLD BANK The World Bank Group 1818 H Street, NW Washington, D.C. 20433 USA Tel: 202-473-1000 Fax: 202-477-6391 Internet: www.worldbank.org/climatechange