Concepts & Trends

A framework for the economic evaluation and selection of sustainability indicators in agriculture

A framework for the economic evaluation and selection of sustainability indicators in agriculture
of 15
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Ecological Economics 33 (2000) 135–149 ANALYSIS A framework for the economic evaluation and selection of sustainability indicators in agriculture David J. Pannell *, Nicole A. Glenn Agricultural and Resource Economics , Uni  6 ersity of Western Australia , Nedlands 6907  , Australia Received 20 April 1998; received in revised form 9 September 1999; accepted 4 October 1999 Abstract In recent years, the concept of ‘sustainability indicators’ has become prominent in agricultural science. The idea isthat particular characteristics of resources are monitored and recorded, with the intention that this information servesas an aid for decision making by farmers and / or policy makers. A great many sustainability indicators have beenproposed by agricultural scientists. However, there is no guidance currently available as to which of the possiblesustainability indicators provide information of economic value. In this paper we present a conceptual framework forthe economic valuation and prioritisation of sustainability indicators. The framework is based on Bayesian decisiontheory, particularly its use to calculate the value of information under conditions of uncertainty. We present anillustrative numerical example. Based on this example and the theoretical framework, we identify a number of important insights about the practical use of sustainability indicators. © 2000 Elsevier Science B.V. All rightsreserved. Keywords :  Sustainability indicators; Economic valuation; Bayesian decision theory; Management / locate / ecolecon 1. Introduction The concept of sustainability apparently hasgreat appeal with regard to environmental andresource management, yet its applicability in prac-tical decision making is hampered by the ambigu-ity of its meaning, and the multiplicity of definitions that have been proposed (Pannell andSchilizzi, 1999). The idea of sustainability indica-tors seems to have grown out of a recognitionthat sustainability cannot be condensed to a singlesimple definition. Its multifaceted nature can bedealt with by monitoring a range of indicators of different types. Hence, in recent years terms suchas ‘sustainability indicators’ and ‘environmentalindicators’ have increasingly been used (e.g.Lefroy and Hobbs, 1992; Standing Committee onAgriculture and Resource Management, 1993;Mannipieri, 1994; Oades and Walters, 1994; * Corresponding author. Fax: + 61-8-98928496. E  - mail address : (D.J. Pannell)0921-8009 / 00 / $ - see front matter © 2000 Elsevier Science B.V. All rights reserved.PII: S0921-8009(99)00134-2  D . J  . Pannell  , N  . A . Glenn /  Ecological Economics 33 (2000) 135–149  136 Pankhurst et al., 1995; Kellogg, and Goss, 1997;Ministry for the Environment, 1997).The idea of promoting the monitoring of sus-tainability indicators has been embraced by manypeople concerned with promoting adoption of more sustainable land-management practices.Pannell and Schilizzi (1999) argued that sustain-ability indicators are a practical and reasonablevehicle for attempting to deal with the multi-faceted nature of the ambiguous term ‘sustainabil-ity’. A great many sustainability indicators havebeen proposed in the existing literature. Examplesrelevant to agriculture include:  Microbial biomass within the soil  Organic matter in soils  Protein levels of crops  Diversity of production  Earthworm density in soil  Pesticide usage  Soil pH  Effective crop root depth  Depth to groundwater tableMost of the proposed indicators are stronglytechnical in focus, with no close link to manage-ment decisions. It has been recognised that thetypes of indicators most useful for differentgroups of users (e.g. on-farm and off-farm) arelikely to differ, but the nature of the differenceshas not been explored. This appears to reflect alack of emphasis on actual decision making in theliterature so far.There have been attempts to persuade farmersto monitor and use sustainability indicators, but itis clear that the attempts have failed. 1 Given thelack of a management focus of most indicatorsproposed so far, this is not surprising. It may alsobe related to the ad hoc nature of the criteria forselection of indicators that have been proposed sofar. The criteria do not appear to have been basedon a consistent, underlying conceptual frame-work.In this paper we argue that the value of asustainability indicator springs from its potentialto improve decision making, and so it is bestthought of as a source of information. We presenta conceptual framework, based on decision theoryand standard information theory, to help guidethinking about the values of potential indicators.We present a simple numerical example fromAustralian agriculture to illustrate the issues andhighlight a number of insights arising from theexample and the framework itself. Finally, wemake initial, tentative suggestions about criteriathat are most likely to identify indicators thatwould be the most valuable to monitor. As aresult of our conceptual framework, the set of criteria we suggest is somewhat different to anyalready in the literature. The scope of our analysisis intended to be broad. Although illustrated withspecific examples, it should be apparent that theprinciples identified are relevant to any monitor-ing that is used to inform decisions involvingcontinuous or approximately continuous decisionvariables. We do not believe that it is limited inrelevance to developed countries or to primarilyeconomic decisions. 2. Conceptual framework for choice of indicators From the perspective of agricultural policy,there are two broad decisions to make: whichindicators to recommend and promote to farmers,and which indicators to collect to assist in policymaking. As recognised by some in the literature,these sets of indicators are likely to differ. Giventhe differences in decision problems faced by thesetwo sets of decision makers, we believe that thesets of indicators are likely to differ substantially,potentially with little or no overlap between them.In choosing indicators to recommend to farm-ers, it has to be recognised that whatever is rec-ommended to them, farmers will make their own,independent choices based solely on their ownperceptions about whether indicators are worthmonitoring. Even so-called ‘minimum’ indicatorsets published so far seem a long way from thisrecognition in that they make unrealistically largedemands on farmers’ time and energy. In later 1 For example, a survey of farmers in the upper Kent Rivercatchment revealed that very few farmers monitor theirgroundwater depths, despite dryland salinity being amongtheir main land-degradation problems (Kington and Pannell,1999).  D . J  . Pannell  , N  . A . Glenn /  Ecological Economics 33 (2000) 135–149  137 parts of this paper, we will be exploring thefactors that are most likely to promote a percep-tion that an indicator is worth monitoring. 2  . 1 . Costs and benefits of monitoring sustainability indicators The fundamental criterion for choosing to mon-itor an indicator is that the benefits from doing somust exceed the costs. A decision maker (eitherfarmer or policy maker) should choose to monitorthe set of indicators for which the total benefitsexceed the total costs by the greatest absoluteamount.Costs are relatively straightforward to concep-tualise. They include:  direct financial costs of equipment, materialsand analysis;  the opportunity cost of the time spent on themonitoring process (meaning the benefitswhich are foregone by virtue of one not beingable to spend that time in some other way).Benefits potentially include:  improved decision making, either by farmers orpolicy makers;  increased awareness or understanding amongfarmers of the potential importance of issuesfor which monitoring is being conducted.Even the benefit of ‘increased awareness orunderstanding’ is desirable primarily because itmay lead, in the longer term, to improved deci-sion making. It seems extremely unlikely thatsociety as a whole or individual farmers would bewilling to invest in a program of monitoringsustainability indicators purely for the sake of interest. Although they may well find the issues of interest from an intellectual point of view, theywould be interested primarily in averting damageto or degradation of natural resources; in otherwords, in improved decision making about themanagement of those resources.The benefits of improved decision making areindirect. They may spring, for example, fromgreater economic returns, from better achieve-ment of social welfare objectives (e.g. intergenera-tional equity) or from protection of ecologicalsystems which may have extrinsic value (e.g. byincreasing economic output) and / or intrinsic value(e.g. through intrinsic ‘rights to exist’). We willtend to focus on economic benefits in the remain-der of this paper. While this objective is impor-tant, we do not imply that other types of benefitsare unimportant or unattainable. Indeed, ourframework could, in principle, be applied in thepursuit of non-financial payoffs.Whatever the underlying benefits, the key pointis the indirectness of benefits arising from im-proved decision making. This is fundamentallydifferent to benefits arising, for example, from anew production technology. The benefits of moni-toring sustainability indicators arise solely fromchanging decisions about which of the existingproduction technologies should be used. This hasprofound implications, as we will see below.With this clarified, we can recognise that thequestion of the value of monitoring a sustainabil-ity indicator is, at heart, a question about thevalue of information. There is a well-developedliterature on the value of information that ap-pears so far to have been completely ignored indiscussions of the choice of sustainability indica-tors. This is unfortunate because we believe thatthis literature offers a number of important in-sights that greatly help to clarify thinking aboutthe issue. In Section 3, we outline a conceptualframework for valuing a sustainability indicatorbased on the standard literature on informationvalue. 3. A conceptual model for estimating the value of a sustainability indicator The benefits of monitoring a sustainability indi-cator are conceptually no different to the benefitsof other types of monitoring which are conductedroutinely by farmers and governments. For exam-ple, farmers monitor their yields, weed problems,market prices, their bank balance, their equity,and interest rates. Governments have long moni-tored yields, areas under cultivation, prices, farmincomes, debt levels and so on. It may be notedthat these are all variables with short-term, directeconomic implications. Many variables of thistype are often included in lists of ‘sustainabilityindicators’ which ‘should’ be monitored by farm-  D . J  . Pannell  , N  . A . Glenn /  Ecological Economics 33 (2000) 135–149  138 ers or government. We see no point in this. To theextent that they are perceived to be of value, theyare already widely monitored, and their inclusionin sustainability-indicator programs is most un-likely to change this. The point of a sustainability-indicator program is surely to focus on issues thatare much more of a long-term nature, on the basisthat this may have been neglected historically.That seems to us to be the only basis on which arenewed effort to promote monitoring can be justified.To a farmer, the gross benefit of monitoring asustainability indicator depends primarily on thescale of production to which it is relevant (e.g. thearea of land for which the information is useful)and the benefit per unit of production (e.g. thebenefit per hectare of relevant land). For a gov-ernment, there is an additional consideration inthe level of adoption that is achieved (e.g. thenumber of farmers who choose to monitor theindicator and the area over which they apply theresults). However, it has been found that adoptionis strongly affected by the economic benefits of the technology or practice (e.g. Lindner, 1987), soadoption and benefit per unit are closely related.Section 3.1 outlines a conceptual framework forvaluing benefit per unit. 3  . 1 . The 6 alue of information from asustainability indicator Throughout this paper, the term ‘value’ is usedin the sense of an economic or financial value. Toavoid confusion, we will refer to the ‘level’ of anindicator to denote its measured physicalmagnitude.Anderson et al. (1977) provided a description of the standard theoretical framework for calculatingthe value of information, focusing on its applica-tion in agriculture. The framework is based on‘Bayesian decision analysis’ or ‘Bayesian decisiontheory’. In this approach, the value of informa-tion arises from its capacity to reduce uncertaintyabout the state of the world, leading to decisionsthat have a higher expected payoff (using ‘ex-pected’ in the statistical sense of a weighted aver-age and ‘payoff’ to mean the contributiontowards any desired outcome, not necessarily Fig. 1. Use of knowledge prior to obtaining additional infor-mation. financial). Uncertainty is represented explicitly asprobabilities of different possible states of theworld being the true state. For example, a farmermay be uncertain about the current depth of arising saline, ground-water-table. Each of the pos-sible depths is assigned a probability, with theprobabilities of all possible depths summing to 1,thus representing a probability distribution. Therelevant probabilities are those estimated subjec-tively by the decision maker (the farmer or policymaker). Clearly, the probability distribution of possible states of the world is representative of thedecision maker’s perceptions, not necessarily of any objective reality. If the objective reality couldbe determined, it would be represented as a prob-ability of 1 for the single true state of the worldand 0 for all others. In practice, the objectivetruth is never known; we can only deal withdifferent degrees of uncertainty.Fig. 1 represents the first part of the process of estimating the value of a piece of information. Afarmer’s decision on applying lime to reduce soilacidity is used for illustration. Based on currentknowledge and perceptions, K  0 (‘prior’ to thecollection of more information) the farmer canidentify the best-bet strategy [ S  ( K  0 )]: the strategythat maximises the expected value of the payoff given current knowledge [ p ( S  ( K  0 ), K  0 )]. For exam-ple, it might be to apply a particular rate of limeevery five years. This strategy is called the ‘prioroptimal act’.  D . J  . Pannell  , N  . A . Glenn /  Ecological Economics 33 (2000) 135–149  139Fig. 2. Use of additional information to modify strategy andincrease expected payoff. tions) of the prior optimal act. In other words, thenew level of knowledge is used to evaluate theprior strategy ( S  0 ), even though the old knowl-edge was used to select that strategy. That is whythe second term in Eq. (1) is p ( S  ( K  0 ), K  1 ), ratherthan p ( S  ( K  0 ), K  0 ). If we did not do the evaluationthis way, the decision maker might appear betteror worse off as a result of the information, even if it did not change the optimal decision. This wouldclearly be nonsensical, since the objective truthabout the biophysical world is not altered bychanging a decision maker’s level of uncertaintyabout it.Now, GVOI  1 is the result for just one possibleobserved level of the sustainability indicator. Inreality, there are many possible levels, and beforemaking an observation a decision maker does notknow what the level is going to be. In order to puta likely value on the information before we ha 6 e it ,we have to rely on subjective estimates of theprobabilities that the indicator will take each of its possible levels. Suppose that there are n possi-ble levels of the indicator and P ( I  i  ) gives thesubjective probability of observing level I  i  . Thenthe expected gross value of collecting informationon I  (evaluated before actually collecting the in-formation) is: E  ( GVOI  ) = % i  = 1... n ( y ( S  i  , K  i  ) − y ( S  0 , K  i  )) · P ( I  i  )(2)This forward-looking perspective, incorporatinguncertainty about what the level of the indicatorwill be once it is observed, is the correct one touse for a decision maker attempting to decidewhether or not it is worth investing in monitoringa sustainability indicator. 4. Requirements to operationalise the model This section is an outline of the steps that arenecessary to calculate the value of information inpractice. It is intended to provide a more concreteunderstanding of the process so that its implica-tions can be readily understood.Now imagine that the farmer observes a sus-tainability indicator, such as the soil pH in thesurface layer. The indicator could conceivablytake any of a large number of levels. 2 Supposethat when the farmer observes it, the level is I  1 .Having made this observation, the farmer reviseshis or her perceptions of the problem to K  1 (Fig.2). The result is a new, ‘posterior, optimal act’[ S  ( K  1 )] and a new expected payoff based on thenew strategy and the new knowledge [ p ( S  ( K  1 ), K  1 )].How much better off is the farmer given that heor she has observed I  1 ? Call this the gross value of information for level I  1 ( GVOI  1 ). GVOI  1 = y ( S  ( K  1 ), K  1 ) − y ( S  ( K  0 ), K  1 ) (1)Note that in calculating the improvement in pay-off, it is necessary to allow for the fact that thenew information may have changed perceptionsabout the expected payoff for the prior optimalact. One makes this allowance by comparing (a)the expected payoff for the new best-bet strategywith (b) the revised payoff (based on new percep- 2 It could take any of an infinite number of levels if it is acontinuous variable. For convenience, the model is presentedhere for a discrete variable indicator.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks