what does the public want from agriculture and the countryside? a review of evidence and methods
TRANSCRIPT
Journal of Rural Studies 20 (2004) 211–225
ARTICLE IN PRESS
*Correspondin
E-mail addres1This debate
commissions and
Environment, Fo
UK Agriculture i
http://www.publi
fru/550/550.pdf
Reform: Respons
2000, http://ww
ses.pdf.
0743-0167/$ - see
doi:10.1016/j.jrur
What does the public want from agriculture and the countryside?A review of evidence and methods
Clare Hall, Alistair McVittie, Dominic Moran*
Scottish Agricultural College, Kings Buildings, West Mains Road, Edinburgh, EH9 3JG, UK
Abstract
Multifunctional agriculture attempts to establish a new balance between traditional commodity support and payment for
the production of non-market goods and services that are increasingly demanded by the public. Supplying non-market
goods presents particular problems for optimal policy design, not least the elicitation of consumer demand for those goods.
The resulting configuration of support policy has potentially enormous implications for rural areas and yet surprisingly little is
known about how the public would prefer public support to be allocated. This seems to have more to do with the political
expediency than true public preferences. We review the evidence of consumer demand for non-market goods and consider the
methodologies used for eliciting public preferences regarding the policy tradeoffs that are likely to characterise the agri-industry
reform debate.
r 2003 Elsevier Ltd. All rights reserved.
1. Introduction
Multifunctionality has been identified as the wayforward for European agriculture with emphasis on theproduction of appropriate market goods and non-market or public goods and services (see ScottishExecutive, 2002). This shift in emphasis has beenongoing for a number of years as OECD countriesface up to the mounting costs of farm support and thedistorting impacts that assistance has on global trade.In the UK, the reform agenda has also been influencedby a series of food scares that have increased publicawareness of agricultural practices and heightenedexpectations of what farmers should supply in returnfor continued public support. A public debate1 aboutthe future of farming has thrown up a muddledagenda of demands including food safety, quality and
g author.
s: [email protected] (D. Moran).
has been prompted by the findings of policy
consultations—see for example, House of Commons
od and Rural Affairs Committee, 2002, The Future of
n a Changing World, Ninth Report of Session 2001-02,
cations.parliament.uk/pa/cm200102/cmselect/cmenv-
and DEFRA, 2002, Common Agricultural Policy
es to Consultation on Mid-term Review of Agenda
w.defra.gov.uk/corporate/consult/capreform/respon-
front matter r 2003 Elsevier Ltd. All rights reserved.
stud.2003.08.004
environmental amenity. Delivering on public expecta-tions will be challenging, not least because many of thesedemands are contradictory. This raises the questionabout the role of public preferences in the reformprocess.
But it is the non-market nature of much of this‘‘new post-productivist agriculture’’ that poses thebiggest policy challenge for both farmers and govern-ment. While market signals can guide patterns ofcommodity production, the nature of public preferencesover the range of essentially non-market (publicgood) outputs is not always identified in markets.Public goods have specific consumption attributesof non-rivalness and non-excludability that can validategovernment involvement in their supply. Because ofthis, people do not habitually transact for many ofthe attributes that farmers are expected to considerwhen they configure their plans for farm management.Indeed, there is no clear template or set of indicatorsthat suggest how we might measure what these outputsare and how the public weights them. Deliveringon growing public demands raises two main questionsfor farmers and policy makers. First, what mightthe public actually want? Second, if farmers can deliveron public preferences for non-market goods, how canthey be compensated in order to produce the rightamount? These two questions lie at the heart of theincreasingly prominent value for money and public
ARTICLE IN PRESSC. Hall et al. / Journal of Rural Studies 20 (2004) 211–225212
good2 debate surrounding farming. By extension theyare central to the reform of existing support schemes.3
The question of efficient compensation has beenexamined in an extensive literature on the design ofspecific support schemes (Colman et al., 1991), andtheoretical papers examining government to farmercompensation transactions under conditions of informa-tion asymmetry (Ozanne et al., 2001). But these papersdo not consider the match between public demand andscheme prescriptions.
Clarifying the demand (or consumer) side of thisdebate is important for the development of instrumentsthat have clear objectives and that target any potentialpublic support to the sector (Randall, 2002; OECD,2002). This paper evaluates recent evidence of measuredpublic (consumer) demand for the environmental andsocial attributes of farming. We consider the nature ofany tradeoffs that the public recognises as revealedthrough surveys, and the robustness of the methods usedto clarify the extent of consumer preferences in thecurrent reform debate.
The structure of the remainder of the paper is asfollows. First, we consider the general question of whatdoes the public want regarding the provision of publicgoods in rural areas and agricultural support payments.We then review public opinion surveys relating to thesetwo points. Next we review economic valuation studiesthat have attempted to value the preferences relating tothe provision of environmental goods in both the UKand EU and the US. Finally, we evaluate the potentialuse of another environmental decision-making tool,multicriteria analysis (MCA), and suggest how we canuse the details from the public opinion polls in acombined methodological approach that will use eco-nomic valuation and MCA.
2. What does the public want?
Given the importance of the question, it is surprisingto note that the multifunctionality debate has rarelystopped to consider the balance of consumer preferencesfor market and non-market outputs. Much of what isknown or assumed about the demand for public goodsappears to be rooted to a policy reform scenario that hasalready been decided and which favours widely flung
2We note the meaning of public good here is different to the received
economic definition. Indeed recent policy discussions of the House of
Commons Committee for Environment, Food and Rural Affairs
appear to have become side-tracked on the meaning of public good.
See http://www.parliament-uk/pa/cm200202/cmselect/cmenvfru/550/
55015.htm#n274.3The Countryside Stewardship Scheme and Environmentally
Sensitive Areas (ESAs) will be replaced by a high-tier ‘‘narrow and
deep’’ scheme. If trials are successful, it will complement the entry-level
broad and shallow scheme, due to be rolled out in England in 2005.
systems of farming. We have little idea of what publicdemand (and by extension ‘‘optimal’’ public goodsupply) might look like in other policy reform scenarios.OECD (2002) notes the challenges in decouplingagricultural support and switching payment to non-market goods. But the discussion reflects a continuingpreoccupation with trade effects rather than confrontingthe fuzziness of public preferences or the methods usedto uncover them. Reflecting the domestic vested inter-ests, countries are adopting the reform agenda indifferent ways. Few seem to be explicit about the roleof public preferences in decision-making. Such prefer-ences may in fact highlight the stark policy tradeoffsthat some architects of multifunctionality seem reluctantto contemplate (Potter and Burney, 2002). But theoutcome of the redirection of aid may be as contro-versial as commodity support if it is unsubstantiated bypublic demand. Policy therefore requires more preciseanswers to a number of questions that will increase therelevance of new farm support instruments:
* What are the attributes that characterise the publicperception of agriculture?
* What are the specific conventional market and non-market (environmental attributes) that are mostrelevant to public attitudes towards farming?
* What is the balance or tradeoff between specificattributes?
* How detailed can surveys be (broad and shallow ornarrow and deep)?
* Over what duration should public preferences beconsidered to be stable?
* How much of each attribute does the public want?* Do people prefer some benefits sooner than others?* Do some features fit in some parts of the country and
not elsewhere?* What is the role of use versus non-use values (or the
balance of private and public goods)?* Should payments go to farmers as the only means of
providing public goods?
Partial answers to some of these questions can befound in some of the valuation methods that haveinformed value for money questions in the design ofagri-environmental policy to date. However, the appli-cation of these methods has been unsystematic and onecan question whether the combined information actuallycontributes to answering the broad question aboutpublic preferences. The general public, consisting ofmillions of individuals, cannot transact directly for theright supply of public goods. Instead, public choicetheory states that enfranchised voters will expresspreferences and sanction expenditure by choosingbetween policy plans on offer when voting in elections.The same theory states that this relationship will bedistorted by constituency interests that can capturefunding, thereby opening up a divergence between what
ARTICLE IN PRESSC. Hall et al. / Journal of Rural Studies 20 (2004) 211–225 213
the public wants and what suppliers provide. So,obtaining value for money in the provision of publicgoods is complex even when governments are respon-sible for supply. In the case of agriculture theseproblems are multiplied because public money has tobe channelled through thousands of suppliers whooperate as the custodians of rural land used foragriculture. In addition, there is a range of difficultiesassociated with monitoring inputs and outputs to andfrom agriculture that give rise to positive and negativeexternalities. Given these difficulties, it is easy to see whypayment schemes can become inflexible in their responseto changing public preferences. It is only whenprompted by major shocks such as the foot and mouthoutbreak that public pressure can be the catalyst to anyreview of priorities.
However, even in periods of heightened publicawareness, establishing consistent preferences is notstraightforward. The reality is that public preferencesfor complex goods are hard to identify. If they are wellformed in the first place then preferences for agricultureappear to fluctuate with news events. Some people havemore knowledge of the issues than others and anysurvey makes strong assumptions about how existingand new information is perceived and processed. Anyobjective survey exercise must seek to take this intoaccount. In the end, we have to make strong assump-tions about what is relevant and about what the publiccan process. Axioms and rules from economic andpsychological theory can also provide guidance but theydo not guarantee stable and consistent preferences.
Table 1 outlines the range of suggested outputs thatthe public might want from agriculture and the ruralsector in general. Several observations are pertinentwhen considering value for money in meeting thedemand for public goods. As recognised by someexisting agri-environmental schemes, some attributesare complementary, while others are mutually exclu-sive.4 As noted by Hellerstein et al. (2002), it is possibleto make a distinction between the demand for agricul-
tural public good supply as opposed to a supply thatderives from any rural area but not strictly farming. Inthe context of multifunctionality negotiations, thisdistinction is particularly sensitive. A further distinctionapparent in the table is between public goods andprivate goods. The supply of public goods is, in theory,the sole reason for government involvement to rectifymarket failure. Again, much reform discussion seems tomake the unstated assumption that vibrant traditionalagriculture is itself a valued public good.5 It is aswell not to lose sight of the fact that even withgovernment withdrawal from the sector there would
4Water quality and pastures for example.5 In the sense that consumers derive welfare from the passive use or
existence value of farming traditions.
remain some supply of public goods. But there is littledocumented evidence testing the weight of preferences infavour of potential private supply scenarios (of publicgoods).
We can only speculate about the relevance of theseissues to broad public opinion because no studies haveconfronted the future of farming in such stark terms.But they do prompt an examination of the central rolethat farming should have in providing the attributessuggested by public opinion. Given the range of sociallydesirable agricultural outputs, it is hardly surprising tonote that policy is somewhat muddled in terms of preciseobjectives for support (OECD, 2002).
3. Surveying public opinion
Some of the distinctions referred to above are implicitin several opinion polls that we review. The objective isto tease out any common themes that characterisepublic preferences for policy reform that tradeoffconventional economic support and newer environmen-tal objectives. Ultimately, we would like to see whetherthe existing evidence helps us to narrow down theinformation in Table 1.
We distinguish three sorts of survey exercises. First,we review a range of polls and surveys conducted byconservation organisations, government departmentsand the EU. These provide general information onpublic preferences. Second, we review more rigoroussurveys that have tried to quantify public preferencesthrough structured tradeoff methods using willingnessto pay (WTP) methods or alternative weighting andscoring. We are more interested in these surveys from amethodological point of view. That is, we want to assesshow appropriate these methods are for framing thebroad policy tradeoffs in question. Third, we allude tothe merits of deliberative survey methods as a compro-mise between polls and valuation methods. (A tablesummarising the approaches we consider is included inAppendix A.)
Table 2 summarises a number of surveys that havebeen conducted in the UK to test opinion about a rangeof agri-environmental and countryside issues. All ofthese surveys have been undertaken in a period ofheightened public awareness of rural issues, so we canexpect the public opinion to be particularly focussed.The rigour of some of these surveys is questionable.Only those conducted by the Royal Society for theProtection of Birds and Friends of the Earth frame thepolicy question in an objective and unambiguous way toelicit environmental and economic tradeoffs for spend-ing limited public funds. While neither survey can beconsidered robust in terms of deriving statisticallysignificant results, the information suggests that thepublic does value a role for farming in the production of
ARTICLE IN PRESS
Table 1
Amenities, services and products provided by rural areas
Aspect Amenity, service or product
Requires active local agriculture
Agrarian cultural heritage Maintenance of the ‘traditional’ agricultural character of the land
Continuation of farming as a way of life in the rural community
New agricultural economy Production of organics and other niches Controlled appellation
Local and regionally produced food
Farm shops
Farmers markets
Added value farm food (cheeses, etc.)
Farm produce
Traditional agricultural
economy
Food quality (taste and nutritional value)
Adequate supply of food
‘‘Cheap’’ food
Agricultural employment
Income from agricultural exports
Farm incomes
Agriculturally related employment
Environmental Farmland landscape
Farmland habitats
Biodiversity—species associated with agriculture
Rural leisure activities Walks in pastoral settings
Visiting local farms
Cultural Agricultural landscape
Farm-based educational activities
Local food
Does not require active local agriculture
Traditional (non-agricultural) industries
Rural character
Viable rural communities
A diversified rural economy
Social Local employment and economy
Vibrant communities
Tourism
Educational resources
Social cohesion
New (non-agricultural)
economy
Rural tourism
Recreation
Access
Recreational opportunities Fishing
Swimming
Scenic drives in the countryside
Birdwatching
Environmental Biodiversity
Watershed protection
Flood control
Landscape
Soil conservation
Water quality
Biodiversity
Habitats
C. Hall et al. / Journal of Rural Studies 20 (2004) 211–225214
non-market goods and that this role legitimises publicspending in supporting farming as a way of life.
This is further confirmed by questions in the Depart-ment of Environment, Food and Rural Affairs (DE-FRA) survey, which reveals that a high percentage ofrespondents agreed slightly or strongly with theproposition that subsidies should be given to farmersonly if they are pursuing environmental goals.
We have considered a range of other sources of surveyinformation, some of which overlaps with other relatedconcerns such as food safety. Our sources also help toillustrate the point that public opinion is both unstableand somewhat suggestible. Thus the prominence ofanimal welfare concerns in the DEFRA study isproblematic given the findings from the recent ScottishExecutive Environment and Rural Affairs Department
ARTICLE IN PRESS
Table 2
Survey results on public opinion regarding agri-environmental and countryside issues
Survey sponsor and respondents Key issues Key findings
Wildlife and Countryside Linka
1450 respondents through link websites, and
RSPB agricultural conference – July/Aug 2001
What role should the production of
‘conservation friendly’ food have in the British
countryside?
Respondents were largely in favour of:* conservation friendly food;* support for farmers being dependent on
minimum environmental standards;* paying farmers to carry out conservation
work.
Respondents wanted to see:* more organic farming;* protection of hedgerows and woodlands;* a return to traditional farming practices;* A thriving rural economy.
Of respondents:* 50% would buy food produced on
conservation friendly farms;* 75% would pay 10% more for
conservation friendly food.
Friends of the Earthb What should farmers receive public money for? Respondents believe that farmers should
receive public money to help them:
793 respondents through FoE website – Nov
2002
* protect the environment (36%);* produce food (6%);* both (47%);* neither (11%).
Royal Society for the Protection of Birdsc
997 respondents through face-to-face
interviews – Jan 2002
What aspects of the British countryside do
people value?
What should be the priorities for government
spending on the countryside?
Respondents value the countryside for:* attractive landscapes (71%)* places where wildlife live (70%);* places for recreation (63%);* as a source of food (33%).
Respondents believe that priorities for
government spending on the countryside
should be:* job creation (63%):* improving services (60%):* paying farmers to protect wildlife and
environment (46%):* paying farmers to produce food (32%).
Dorset Agenda 21 Forumd Among a wide range of sustainability issues
those relevant to this study were
Statements relating to food and agriculture
adopted by LA21 Strategy:
1000 respondents through Local Agenda 21
household survey and LA21 conference
attended by 100+ community and public
sector organisations (LA21 Strategy published
1999)
What should the future vision be for food and
agriculture in Dorset?
What are the key issues relating to food and
agriculture in Dorset that need to be addressed?
* need to support sustainable farm practices;* need to promote agri-environment
schemes;* need to increase the number of jobs in
agriculture through reduced intensification;* need to extend organic food production.
Food Standards Agencye
1003 respondents through telephone survey—
September 2001
What concerns do people have about food
production?
Respondents were concerned about:* the way food is produced nowadays (32%)
(very concerned);* how animals are treated and raised (23%);* the use of chemicals and preservatives in
food production (18%);* Genetically modified crops (11%);* mass production (should be free range)
(8%).
Sustain and UK Food groupf
Network of 14 non-governmental
organisations—July 2002
How should the Common Agricultural Policy be
reformed?
UKFG want to see:* existing CAP resources redirected towards
environmentally and socially beneficial
farming and sustainable rural
development;
C. Hall et al. / Journal of Rural Studies 20 (2004) 211–225 215
ARTICLE IN PRESS
Survey sponsor and respondents Key issues Key findings
* public funds used to support the delivery of
public goods that are not delivered
adequately by the market;* support for small-scale and family farmers.
Department for the Environment, Food and
Rural Affairsg
3736 respondents through at home interviews –
Jan–April 2001
Among a wide range of quality of life issues
those relevant to this study were:
How worried do people feel about certain
environmental issues?
What about the British countryside make it a
place where people want to spend time?
To what degree do people support various
environmental policy options?
Respondents stated that:* they want to spend time in the British
countryside because of the scenery (46%),
the plants and wildlife (36%) and the way
of life (9%);* they strongly or slightly support the policy
of only paying agricultural subsidies to
farmers if they protect the environment
(74%)* they strongly or slightly support the policy
of paying farmers to protect and regenerate
threatened landscapes and habitats (69%)* they strongly or slightly support the policy
of planting trees and hedgerows where
possible (92%).
Scottish Executiveh
4119 at home interviews – 2002
Among a wide range of questions about the
environment in Scotland, those relevant to this
study were
What activities do people think present the
greatest threats to wildlife and habitats?
Who do people think should play the most
important role in protecting wildlife and
habitats?
Respondents stated that:* current methods of farming present the
greatest threat to wildlife (22%);* farmers play an important role in
protecting wildlife and habitats (33%);* farmers should play an important role in
protecting wildlife and habitats (17%);* paying farmers and foresters is a very good
or good way in which wildlife and habitats
can be protected or improved (60%).
Country Landowners Associationi How important is farming? Respondents stated that they agree that:
1001 respondents through telephone survey –
Nov 2002
How should CAP money be spent?
Is it worth paying more for environmental
standards in food production?
* farming is an important part of our
national life and economy (92%);* farmers play an important role in providing
an attractive and well-managed
countryside (87%);* CAP money should be switched from
subsidising food production into schemes
that support the environment and rural
development (61%);* it is worth paying more for quality UK-
produced food with high welfare and
environmental standards (88%).
Eurobarometrej What are Europeans’ perceptions of the
following
Respondents believe that the CAP should:
16,041 respondents in 15 member states
through a survey—February–April 2002
The benefits of the CAP for consumers and
farmers?
The role of the CAP?
How the CAP fulfils its role?
The evolution of the CAP?
* ensure that the food you buy is safe to eat
(40%);* ensure that you have information about the
origin of food (25%);* promote the respect of the environment
(88%);* improve life in the countryside (77%);* encourage agricultural diversification
(73%);* favour methods of organic production
(72%).
Respondents believe that the CAP is currently
successful in* promoting respect of the environment
Table 2 (Continued)
C. Hall et al. / Journal of Rural Studies 20 (2004) 211–225216
ARTICLE IN PRESS
Survey sponsor and respondents Key issues Key findings
(41%);* favouring methods of organic production
(37%);* encouraging agricultural diversification
(34%);* improving life in the countryside (31%).
Respondents believe that* development of the CAP away from
subsidising agricultural products and
towards the protection and development of
the rural economy is a very or fairly good
thing (62%).
aWildlife and Countryside Link, 2001, Future of the Countryside Web Survey—Findings, Wildlife and Countryside Link, Londonbwww.friendsoftheearth.org.ukchttp://www.rspb.org.uk/caffairs/archive/582.htmdDorset Agenda 21 Forum, 1999, Dorset in the 21st Century: An Agenda for Action, Dorset Agenda 21 ForumeCOI Communications, 2001, Food Concerns Omnibus Survey, Prepared for Food Standards Agency by COI CommunicationsfSustain & The UK Food Group, 2002, The CAP Doesn’t Fit, Sustain & The UK Food GroupgDEFRA, 2002, Survey of Public Attitudes Towards the Environment and Quality of Life—2001, http://www.defra.gov.uk/news/2002/020726b.htmhScottish Executive, 2002, Public attitudes to the environment in Scotland—2002 Survey.iCountry Landowners Association, 2002, Image of Agriculture SurveyjThe European Opinion Research Group, 2002, Eurobarometre 57.0 Europeans and the Common Agricultural Policy 2001–2002, EORG, http://
europa.eu.int/comm/agriculture/survey/2002/rep en.pdf
Table 2 (Continued)
C. Hall et al. / Journal of Rural Studies 20 (2004) 211–225 217
funded focus groups on food,6 which suggested that theissue was rated much lower than food safety concerns.In the Eurobarometre survey much of the focus is onhaving the respondents validate the Common Agricul-tural Policy as an institutional approach to agriculture.While this is the most extensive survey of public opinion,the form of questioning does not invite clear preferencetradeoffs. All of these polls confirm the difficultiesinvolved in eliciting objective statements about what thebalance of agricultural outputs should be. Specifically,response interpretation depends on
* who is questioned and who is asking the questions—most surveys considered here were conducted by non-representative pressure groups with specific agendasthat are not always pro-farming
* the type of question asked and what is not mentioned* how questions are framed7
* prevailing background news stories* different levels of information provided as part of the
exercise* whether or not the respondents are allowed multiple
response options rather than being forced intospecific close ended (either/or) tradeoffs.
6Public perceptions of food and farming in Scotland.7There is an obvious difference between asking what the public
wants from the countryside and recording the spontaneous weight put
on farming or joint goods from farming, or a more specific question
asking what do you want agriculture to produce.
Overall, we have located very few alternative UK orEuropean polls or research projects that take anobjective standpoint in trying to elicit policy tradeoffs.This is hardly surprising since such surveys rely on alevel of disinterest in the issues to be conductedobjectively. Ultimately, the surveys in Table 2 shouldbe viewed with care. At the very least, the attitudes mayneed validation with reference to surveys that are morerealistic in the forms of economic tradeoffs that thepublic faces for the production of outputs. Suchevidence may be contained in a body of non-marketvaluation literature.
In the US, Hellerstein et al. (2002) consider the natureof public preferences in the context of changing ruralland use. While suggesting the need to increase the inputof public preferences in designing rural amenity (fromboth farmland and other non-agricultural state lands),Hellerstein et al. note the absence of information onwhat the public really want. Side stepping direct elicita-tion, the authors assume a second best manifestationof preferences as represented by the explicit objectivesset down in state legislation to protect farmland. Byimplication, public preferences are reflected in theprograms that get enacted through the purchase ofdevelopment rights, for example. The study thereforeconducts a detailed review of farmland conservationlegislation in a number of states and concludes that ‘‘thepublic’’ values rural amenity—farming as a way of life,open space and scenic beauty. The intensity of legisla-tion is also related unsurprisingly to the amount of
ARTICLE IN PRESSC. Hall et al. / Journal of Rural Studies 20 (2004) 211–225218
farmland in the state. That is, the frequency ofpurchased development rights is greatest in the mosthighly urbanised states. The ‘‘farming as a way of life’’finding is consistent with results from previous polls thatsuggest farming is a part of a cultural landscape. Thatthe demand for rural amenity is highest in more denselypopulated areas is intuitive and perhaps warrants moreattention in the design of agri-environmental schemes.
Other findings from this study are of interest. First, aspreviously noted, the authors conclude that manyenvironmental attributes need not be provided exclusivelyfrom farming. However, they find that most statelegislature ensures that farming remains an importantvehicle for delivering some of them. There is thussomething of an emotional attachment to an agriculturalbase, which we assume to be common whether we aredealing directly with public expression or—as in thiscase—with state legislation. Second, the design of theHellerstein study raises the question of how the appro-priate level of public demand can be practically consideredfor policy purposes. Finding out what the public wants bysurveying them can be insightful but impractical toconduct on a regular basis. On the other hand, usinglegislation as a proxy for public choice is too crude toreflect preferences adequately. A possible middle waymight therefore be found using more deliberative methodssuch as Citizens Juries and Consensus Conferences.
4. The valuation of public preferences for non-market
goods
4.1. UK studies
There has been an extensive body of applied workeliciting economic values for non-market goods andenvironmental services related to UK agriculture. Mostof this work has used economic techniques to exploreindividuals’ WTP in order to derive a mean value forenvironmental change for society in general. Thiseconomic valuation research typically has the aim ofinforming cost–benefit decisions for agri-environmentalspend by uncovering the demand side or economic valueassociated with a range of environmental features,landscapes and production practices. The importantdifference between this research and the qualitativesurveys in Table 2 is that monetary valuation is based onunderlying economic theory that provides axioms orrules that allow tests of consistency of respondent WTPstatements. The validity of these rules is subject toongoing debate (Sagoff, 1988). Suffice to say that wellexecuted surveys make respondents behave as if theywere constrained by their income and produce realmeasures of welfare change. They avoid the hypotheticalcost-unconstrained nature of opinion polls. Addition-ally, in theory these studies offer a better measure of
social welfare to guide policy decisions. In contrast, withopinion polls, what we are looking at is whetherresponse options are statistically significant and havecategories where more than 50% respond favourably orunfavourably. In economic jargon the difference isbetween the measurement of a Pareto improvement—where a policy change occurs and aggregate welfare isincreased to the extent that everyone (were loserscompensated by gainers) could still be better off—andthe median voter criterion for justifying policy change.The latter says nothing about aggregate welfare and adecision supported by over 50% of a population samplemay still result in the negative welfare of the minorityoutweighing the majority gain.
We are interested in the economic valuation studiesbecause those of a stated preference nature (i.e.contingent valuation methods (CVM) and choiceexperiments (CE)) are direct elicitation methods usingsurveys of the general public. Some of the methodolo-gical approaches and findings are instructive in terms oftheir potential to address policy tradeoffs. The standardprocedure for both CVM and CE is to develophypothetical scenarios in which specific features arevalued. The content of a scenario is defined after focusgroups and survey pre-tests. In the former, selectedparticipants discuss open-ended themes that iteratetowards more specific discussion of the issue of interestto the researcher. In the context of both CVM and CE,the researcher is interested in identifying and represent-ing the key attributes that describe the issue or changebeing proposed by a policy. The CVM and CE methodsrepackage the information in different ways. In bothcases, the design of the survey has to respect thecognitive limits of the respondents.
In reviewing the content of existing agri-environmen-tal studies however, it is as well to note that by the timethe study appears in published form, it is typicallyimpossible to discern the content of discussions thattook place in any design (for example, focus group)stage. Thus while respondents may be valuing specificitems, we have little idea about the strength of thesepreferences relative to attributes or scenarios that werenot deemed to be central to the specific study. A lack ofco-ordination in conducting studies means that eachstudy will be predicated on a different baseline story thatrespondents were invited to consider. Caution musttherefore be exercised in lumping studies together to finda central message in the WTP values that arise.
In the UK context, we note that there is large body ofagri-environment valuation studies (for a review seeStewart et al., 1997). Most studies are of the CVMvariety that invite respondents to express a WTP inresponse to a specific scenario and some policy on/offalternative. A number of CE-type studies have beenapplied to agri-environmental issues. These studies aremore explicit in the way change scenarios are broken
ARTICLE IN PRESSC. Hall et al. / Journal of Rural Studies 20 (2004) 211–225 219
down into two or three key attributes that can varyacross different choice packages. Some studies are moreambitious in the list of attributes that respondents areinvited to consider although none frame the issue interms of the policy choice between income support anddecoupled environmental payments.
As an example of the typical breadth and depth ofenvironmental attributes, Hanley et al. (1996) report onthe application of CE (a CVM exercise was undertakenconcurrently) to the Breadalbane and Machair Envir-onmentally Sensitive Areas (ESA). The BreadalbaneESA CE involved presenting respondents with choicesof attribute bundles concerning the protection ofwoodlands, archaeological sites, moorlands, wetlands/herb-rich grassland and dry stone dykes, each taking a‘‘policy on’’ or ‘‘policy off’’ level. For the MachairESA survey the attributes were the protection ofbirds, flora and archaeology. There was also a tax(price) attribute with eight levels, which allowed thecalculation of WTP for different attribute levels and forwhole policy options. Respondents were presented witheight choice sets of two options, the choice being eitheroption A, option B, neither or do not know. The use ofseveral choice sets with varying levels allows theresearcher to infer four pieces of information (Hanleyet al., 1998):
1. attributes with a significant influence on choice;2. the implied ranking of the attributes;3. the marginal WTP for a change in any significant
attribute; and4. the implied WTP for a programme that changes one
or more attributes simultaneously.
Table 3 shows the implied rankings and implicit prices(WTP for a one unit change in each attribute) for eachof the attributes used in the CE. The implicit priceestimates can be used individually or in combination toassess the benefits of different combinations of policyoutcomes. More sophisticated experimental designs cancontain a wider range of attribute levels to allow a fuller
Table 3
Implied rankings and implicit prices for environmental attributes for the Br
Attributes Breadalbane
Protection of Implicit price (d per household)a
Woodlands 82.86
Moorlands 37.14
Wetlands/herb-rich grassland 34.29
Dry stone dykes 18.57
Archaeology 10
Birds —
Flora —
aCalculated from estimated coefficients reported in Hanley et al. (1996).bAll estimated coefficients were significant at the 95% level with the exce
analysis of utility functions for different levels of eachattribute. A priori, we might expect the implicit pricesfor an attribute to increase and then decrease due todiminishing marginal utility. That is, the more of anattribute that is supplied, the less we value the last unit.
A similar study by Foster and Mourato (2000)considers a different range of attributes that mix publicand private good considerations. Their CE studyconcerns the impacts of pesticide use in wheat cultiva-tion. The study is of particular interest as it considersboth the effects of pesticides on biodiversity, in terms ofnumber of bird species in decline, as well as impacts onhuman health. Whereas many valuation studies involvea price attribute (to estimate WTP) presented in terms ofan increase in tax, this study uses the price of a loaf ofbread; a payment vehicle suggesting a private paymentobligation for the mixture of a private good (health) andthe public good (conservation of species).
The CE approach differs from CVM in that althoughthe statistical design process is identical, respondents areasked to rank a number of treatments rather thanchoosing one. In this case respondents were asked torank four alternatives, one of which was the currentproduction practice, as shown in Table 4. Foster andMourato (2000) used two different model specificationsto analyse their survey responses. The first was the mostpreferred alternative (MPA) model, which considersonly the attribute levels of the highest ranked alter-native. The second, known as the ranks data (RD)model, uses the ranking of all the alternatives thus fullyexploiting the information gained from the survey.These different specifications result in divergent WTPestimates, ranging from 0.7 to 1.2 pence per loaf toavoid a case of ill health and from 5.3 to 7.4 pence perloaf for one less bird species in decline for the MPA andRD specifications, respectively. These results relate tomodels aggregated over the entire sample. Table 5summarises WTP across different population segments.
These results are consistent with a priori expectations,particularly in that bird watchers and those respondents
eadalbane and Machair ESAs
Machair
Implied ranking Implicit price (d per household)b Implied ranking
1 — —
2 — —
3 — —
4 — —
5 13.42 1
— 6.05 2
— 3.68 3
ption of ‘‘flora’’.
ARTICLE IN PRESS
Table 4
Sample contingent ranking question from Foster and Mourato (2000)
Process Aa Process B Process C Process D
Price of bread (pence per loaf) 60 85 85 115
Health effects on general public (cases of ill health per year) 100 40 40 60
Biodiversity effects on farmland birds (no. of species in decline) 9 2 5 2
Ranking (example) 2 1 3 4
aCurrent technology for wheat production.
Table 5
WTP estimates across different population segments
Gender Concern for environment Birdwatcher
Male Female High Low Yes No
WTP for human health 0.63 0.75 0.87 0.17 0.83 0.66
WTP for biodiversity 5.45 5.06 6.56 1.96 8.03 4.44
C. Hall et al. / Journal of Rural Studies 20 (2004) 211–225220
expressing high environmental concerns have muchlarger WTP for biodiversity. It is important to identifysuch segments in valuation studies, (as they may beover- or under-represented in the survey sample), inorder to avoid bias in aggregated WTP estimates.
4.2. US studies
Table 6 summarises a number of relevant US studiesand is taken directly from the review of valuation studiesby Hellerstein et al. (2002). We note that the US has adifferent rural tradition to the UK and our review ismore interested in methodological findings. In thisregard WTP valuation studies are prevalent. But theuse of non-monetary pilot surveys and focus groups inthe design of valuation scenarios is instructive. As wefind with the European literature, the authors indicate amajor disconnection between the extent of this specificvaluation literature and the more general question ofwhat the public wants.
Of methodological interest in this list are the studiesby Kline and Wichelns (1994, 1996) and Duke et al.(2002). Kline and Wichelns (1994, 1996) use the outputof focus groups to develop lists of attributes thatcharacterise public preferences for farmland preserva-tion in Rhode Island. These are subsequently used in thedesign of a preference list that is presented as a surveyfor the general public to derive a mean rank (1–10 with10 being the most important). The rankings are thenused in a factor analysis that produces more specificattributes that are used as input attributes to CEs.
Duke et al. (2002) draws on this approach and theirstudy is a template for exploring public preferences. LikeKline and Wichelns they undertake a CE. The attributeset includes simple photographs and symbols to
characterise the three environmental attributes: growthcontrol (urban sprawl), transitional agriculture andopen space (roughly landscape amenity). These attri-butes also form the basis of a parallel MCA. Basically,these two methods provide alternative ways of present-ing attributes to the public to derive more robuststatistical insights into public preferences and tradeoffsbetween farming and environmental attributes.
The body of these studies appears to conclude infavour of preferences for a traditional cultural role forfarming even if the environmental attributes can bedelivered by other means. The control of urban sprawland the role of farmland in protecting water sources andquality are also highlighted.
5. Multicriteria analysis (MCA) and the analytical
hierarchy process (AHP)
The limits of monetary valuation methods have beenhighlighted in an extensive literature that suggestscognitive barriers arise after presenting respondentswith certain levels of information. Faced with difficultchoice exercises, respondents can take mental short cutsand base all choices on the price variable. MCA is analternative non-monetary preference elicitation methodthat is common in environmental decision-making andhas been extended to assess public preferences overmany different environmental attributes (see for exam-ple, Tzeng et al., 2002). The essential distinction betweenMCA and monetary valuation is that MCA is notcircumscribed by the strict utility theoretical designrequirements. For some, the consequent removal of amonetary choice attribute facilitates the choice exercise.
ARTICLE IN PRESS
Table 6
US valuation and preference survey studies for farmland
Authors Region Findings
Halstead (1984) Hampden County, MA There are strong preferences for protecting remnant farmlands that
increase with size of program, and seem to be positively influenced by
the proximity to farms
Furuseth (1987) Mecklenberg County, NC There is broad support for farmland protection; farmland heritage,
environmental reasons, and protection of future food supply were
important reasons
Variyam et al. (1990) National Support for a variety of agricultural programs suggests that
preservation of family farms is important, but respondent self-interest
also influences support for agricultural policies
Dillman and Bergstrom (1991) Greenville County, SC Positive, though small, benefits to protection of farmland, with the
benefits of such protection stated as being limited to changes in rural
amenities. The low values are attributed to the large amount of
agriculture in the study region
Kline and Wichelns (1994, 1996) Rhode Island, PA Environmental reasons are most important, followed by local food
concerns, preservation of rural communities, and slowing development
Bowker and Didychuk (1994) New Brunswick, Canada Willingness to pay for farmland protection is correlated with
membership in environmental organizations and ‘‘visiting the land’’
and is negatively correlated with distance to farmland
Ready et al. (1997) Kentucky Positive difference between survey-derived compensating variation
measures and house-price/wage-rate hedonic measures of the value of
protecting horse farms suggests that these farms have an existence
value
Rosenberger and Walsh (1997) Routt County, CO Protection of ranchland yields small overall per acre values. These
values may be substantially larger if preferences of summer visitors are
considered
McLeod et al. (1999) Sublette County, WY Residents prefer continued agriculture on some lands, and wildlife/
recreational uses on others, with development never a preference
Krieger (1999) Chicago The support for rural land protection (which includes farmland
protection) seems suburbs to be derived from quality of life concerns,
especially those related to sprawl reduction. Compared with other rural
land protection programs, the most important reasons stated for
supporting farm protection were protecting family farms and
maintaining food supplies
Boyle et al. (2001) Several states Focus groups suggest that the public favours protection of family
farms, protecting land with water on it, and favouring land with active
farming
Duke et al. (2002) Delaware Delawareans seem to be most concerned with keeping farming as a way
of life, having access to locally grown agricultural commodities,
protecting water quality, and preserving rural character
C. Hall et al. / Journal of Rural Studies 20 (2004) 211–225 221
The Department for Transport, Local Governmentand the Regions (DTLR) manual on MCA (DTLR,2001) distinguishes several MCA variants, including theanalytical hierarchy process (AHP) that has been usedby Duke et al. (2002) and Duke and Aull-Hyde (2002).The method uses a number of pairwise comparisonsbetween quantitative or qualitative criteria to assess therelative importance of each criterion. These can bearranged in a hierarchical manner known as a value treeto form sets of attributes and qualities (levels) withinthese attributes. The simplicity of the AHP approach isthat unlike CEs, the qualities (or levels) of differentattributes are not directly compared, thus removing theneed for complex survey designs and associated impactson sample size. Instead, respondents first make pairwisecomparisons of the qualities within each attribute beforecomparing each of the attributes. Cognitive burden is
also reduced as comparisons are between two qualitiesor attributes rather than a larger bundle of attributesand levels. As a consequence, respondents are less likelyto adopt mental short cuts by concentrating dispropor-tionately on one attribute. As noted above, in CEs thiscan result in a choice directed by the price variable.
The pairwise comparison is framed in the form of thequestion: how important is criterion A (say, farmincome) relative to criterion B (say, number of birdspecies on farm)? The responses to these questions aretypically coded along a nine-point scale as set out inTable 7.
So for example, from Table 7, if B is considered to bemore important than A, then the reciprocal of therelevant rating is assigned (i.e. 1/7 as opposed to 7 whichwould be assigned if A were strongly more importantthan B). As it is assumed that a respondent is consistent
ARTICLE IN PRESSC. Hall et al. / Journal of Rural Studies 20 (2004) 211–225222
in judgements about any one pair of criteria, this use ofthe reciprocal allows only 1=2nðn � 1Þ comparisons to bemade where there are n criteria. The ratings, and theirreciprocals, are then collected in a comparison matrix.So for three attributes this might look like
1 7 9
1=7 1 2
1=9 1=2 1
264
375:
This is then used to derive weights that are consistentwith the relativities between the attributes or qualitiescontained in the matrix. Although there is consistency inthe judgements made between any pair of criteria, this isnot guaranteed in judgements between pairs, so theestimated weights aim to provide the ‘‘best fit’’ of theobservations (DTLR, 2001). This can either be achievedusing complex matrix algebra or by calculating thegeometric mean of each row and normalising these by
Table 7
Typical AHP rating scheme
How important is A relative to B?
Rating Explanation of relative importance
1 Equally important
2 Between 1 and 3
3 Slightly more important
4 Between 3 and 5
5 Moderately more important
6 Between 5 and 7
7 Strongly more important
8 Between 7 and 9
9 Overwhelmingly more important
Table 8
AHP state-level results
Attribute Quality Quality w
Attribute
Agriculture 0.334
Providing locally grown food
Keeping farming as a way of life
Important industry
Environmental 0.270
Protecting water quality
Protecting wildlife habitat
Preserving natural places
Growth control 0.213
Slowing development
Preserving rural character
Open space 0.184
Preserving scenic quality
Breaks in the built environment
dividing by the sum of geometric means for each row.For the above matrix the weights would be
Geometric mean Weight
Criterion 1 ð1� 7� 9Þ1=3 ¼ 3:9791 0:7926;
Criterion 2 ð1=7� 1� 2Þ1=3 ¼ 0:6586 0:1312;
Criterion 3 ð1=9� 1=2� 1Þ1=3 ¼ 0:3816 0:0760;
Sum ¼ 5:0193 1:000:
Thus for policy purposes, the method provides anunambiguous weighting from a three-way comparison.
The majority of existing applications of AHP toenvironmental and natural resource management issueshave involved small samples of experts, resourcemanagers and stakeholders. In these cases, the aimwas to reach consensus on management decisions andpriorities in a manner similar to Delphi exercises, but ina way that also elicits the relative ‘‘utilities’’ of differentmanagement options. The fact that AHP is notstatistically designed in the same sense as a CE alsomeans that it lends itself to applications with smallnumber of participants.
As previously noted, Duke et al. (2002) and Duke andAull-Hyde (2002) report on a study concerning publicpreferences for land preservation in Delaware, USA.The study departs from previous applications in that alarge-scale sample of 129 residents of four counties wasobtained. Respondents were asked to make pairwisecomparisons between a number of attributes and alsothe qualities or levels within these attributes. These areillustrated in Table 8. In this case each respondentneeded only to make 14 pairwise comparisons. This ismuch less onerous than the equivalent choice set in aCE, which would require 32� 22=36 combination
eight Quality rank
weight Within attribute Overall AHP rank
1
0.416 0.141 1
0.422 0.140 2
0.163 0.053 10
2
0.483 0.132 3
0.316 0.084 7
0.202 0.054 9
3
0.462 0.099 6
0.538 0.114 4
4
0.605 0.111 5
0.395 0.073 8
ARTICLE IN PRESSC. Hall et al. / Journal of Rural Studies 20 (2004) 211–225 223
matrices (of attributes times qualities) to be presented torespondents.
Table 8 illustrates the results obtained by Duke et al.(2002) and Duke and Aull-Hyde (2002). The table givessome idea of the potential output from an AHP study.Separate sets of weights are calculated for pairwisecomparisons between the qualities and then the attri-butes. Overall weights are then calculated by multiplyingattribute and quality weights. This further allows theimplied ranking of attributes and qualities.
The implied ranking of attributes suggests that‘‘agricultural factors’’ are the most important, while‘‘providing locally grown food’’ and ‘‘keeping farmingas a way of life’’ are the two highest ranked qualities.Interestingly, agriculture as an ‘‘important industry’’ isranked lowest. This may indicate that industry concernsare not seen as overriding public concerns such as‘‘protecting water quality’’ (3), ‘‘preserving rural char-acter’’ (4) and ‘‘preserving scenic quality’’ (5). Thesepublic priorities also appear to be anthropocentric innature as water quality has a greater direct impact onpeople rather than ‘‘protecting wildlife habitat’’ (7) and‘‘preserving natural places’’ (9).
6. Proposed methodological approach for scoring public
preferences
It is our view that a combination of the AHPapproach and a CE offers scope for including an arrayof attributes and their constituent qualities. Bothmethods offer potential for mixing attribute levels (i.e.broad and shallow and narrow and deep). In the CEliterature, the Foster and Mourato (2000) study showsthis by mixing human health, the price of farm outputs(bread) and a farmland environmental attribute (birdspecies). Layton and Brown (2000) use a CE to evaluatepreferences for greenhouse gas emission scenarios andresultant forest loss in Colorado over time horizons upto 150 years into the future. Both studies concludethat the method provides internally consistent re-sponses. This conclusion gives us confidence that withappropriate communication devices8 CEs lend them-selves to a general policy experiment that proposesbroad environmental and economic categories forrespondents to consider. Thus it will be possible for usto test whether attributes such as farm incomes, foodprices and other environmental attributes can bebundled for consideration in the same way. Of coursethe price attribute is the key that allows us to infermarginal values (MV) of attributes (by extensionrevealing how MV diminishes with rising supply).
8That is, clear explanation and photographic information if
necessary.
An AHP application allows some scope in terms ofattribute ‘‘depth’’ but the pairwise methodology is notconstrained by the strict design criteria of a CE. Whatthis method loses (in terms of its economic rigour) itgains in terms of the ease of flexibility in the way we canrepresent attribute sets. CEs are often criticised for beingover-optimistic about the way tradeoffs are made. AnAHP study would provide insurance that we can derivea preference ordering over a final set of attributes. Bothmethods allow for tests of sensitivity to suggestionsabout the time preference for realisation of benefits andto investigate spatial (for example, regional) variation inthe strength of preferences for specific attributes.
7. Conclusion
This review has attempted to see whether and how thefundamental question of what the public might wantfrom agriculture has been adequately addressed in theexisting agriculture and environmental literature. Weconclude that we cannot derive meaningful quantitativeconclusions from the existing literature. Existing studiesaddress public preferences in a very partial way and nostatistically robust UK study has attempted to evaluatepublic preferences from scratch. Existing studies havetypically evaluated a feature or several features of someform of countryside designation. It would be invalid toinfer public preference from comparing the magnitudeof WTP values between studies of a partial nature.
We have nevertheless sought to identify someevidence of public preferences from polls and WTPstudies. Our aim has been to narrow down the widearray of attributes presented in Table 1. A range of pollshas been undertaken and some of the topics featuredoverlap. While these methods are not statisticallyrobust, they do suggest that the public see a definiterole for farming as an intrinsically valued provider ofrural environmental public goods. Additional themesthat reoccur are farmland as cultural heritage,biodiversity and environmental quality. Of the morestatistical studies, none addresses public preferences insuch a broad ranging fashion. However, we havetaken methodological guidance from a study conductedin the United States, which combines MCA and CEsin a novel way to address a similar question to ourown for the state of Delaware (Duke et al., 2002).
Of the two methods, MCA has the weaker theoreticalfoundation but offers flexibility in permitting a mix ofattributes at different levels. CEs on the other hand posita much tighter set of rules that define what we can offerpeople in our choice sets and how they should be tradingthem off if they are consistent. What we note about thisliterature is that on the whole the attribute sets havebeen quite narrow. Few studies have used choice sets ofcomparable dimension to the policy tradeoffs implicit in
ARTICLE IN PRESSC. Hall et al. / Journal of Rural Studies 20 (2004) 211–225224
agricultural reform. A notable exception is Layton andBrown (2000), who attempted to test preferences forclimate change mitigation. The dimensions of thisproblem are possibly even more extreme than the oneswe are tackling. The development of CEs has reached astage where they can be appropriately used to test thepublic policy implications of agri-environmental reform.
Appendix A
Approach C
Opinion polls and consumer surveys St
LeLi
NOwFowsuco
Proxy for public preferences St
Eg review of legislation(see for example, Hellerstein et al., 2002)
CLi
CIg
Deliberative methods St
Eg focus groups PaCitizen’s juries So
anInterviewsDelphi methods Li
Consensus conferences N
Monetary valuation St
Eg Contingent valuationChoice experiments
SucoasCfoCchvainLi
StGsuUM
Multi-criteria analysis St
Eg AnalyticalHierarchy
Ctiv
Acknowledgements
This review derives from the project ‘‘Beauty Beastand Biodiversity: What does the Public Want fromAgriculture?’’ The project is funded by the ScottishExecutive Environment and Rural Affairs Department.
omment (strengths and limitations)
rengths
ss challenging in terms of respondent effortmitations
ot always statistically robustften carried out by non-representative pressure groupsith specific agendasr policy purposes, unclear how outcomes or results link
ith any theory of public choice—i.e. may use result ofrvey where more than 50% say yes or no. This does notnsider overall social welfare
rengths
an be useful if primary data collection not feasiblemitations
rudenores other influences on regulatory decisions
rengths
rticipatorymewhat open-ended formats. Do not present questionsd statements that limit the scope of the studymitations
ot statistically robust
rengths
rvey methods based on recognised axioms and rules ofnsumer choice for the derivation monetary valuesigned by respondents to attributes under considerationV uses a general verbal (sometimes graphic scenario)llowed by a direct WTP question to the respondentEs use design sets of several key attributes that describe aange to be valued. One of the attributes is a priceriable. Repeat choices of favoured sets allow thevestigator to derive WTP indirectlymitations
rict design criteriarounded in economic theory (which not everyonebscribes to)sually requires focus groups to establish scenarioust recognise cognitive limits of respondents
rengths
ombination of qualitative (participatory) and quantita-e (can be statistically robust)
ARTICLE IN PRESS
Procedure Offers flexibility in permitting a mix of attributes atdifferent levelsAHP allows attribute depthThe pairwise methodology of AHP is not constrained bythe strict design criteria of choice modelling (includingmonetary values)Limitations
Time consumingHas weaker theoretical foundation (than CEs)Can be cognitively challenging for respondents
C. Hall et al. / Journal of Rural Studies 20 (2004) 211–225 225
References
Bowker, J.M., Didychuk, D.D., 1994. Estimation on non-market
benefits of agricultural land retention in eastern canada. Agricul-
tural and Resource Economics Review 23 (2), 218–225.
Boyle, K., Peterson, R., Aheran, M., Alberini, A., Bergstrom, J.,
Libby, L., Welsh, M., 2001. Improved information in support of a
national strategy of open land polices: summary of focus group
findings, Summer 2000. Staff paper—Maine Agricultural and
Forest Experiment Station, February 2001.
Colman, D., Crabtree, B., Froud, J., O’Carroll, L., 1991. Comparing
effectiveness of conservation instruments. Report of the Country-
side Commission, Department of Agricultural Economics, Uni-
versity of Manchester.
Dillman, B., Bergstrom, J., 1991. Measuring environmental amenity
benefits of agricultural land. In: Hanley, N. (Ed.), Farming the
Countryside: An Economic Analysis of External Costs and
Benefits. Redwood Press, UK (Chapter 14).
DTLR, 2001. Multicriteria analysis: a manual. http://www.dtlr.go-
v.uk/about/multicriteria/index.htm.
Duke, J.M., Aull-Hyde, R., 2002b. Identifying public preferences for
land preservation using the analytic hierarchy process. Ecological
Economics 42, 131–145.
Duke, J.M., Ilvento, T.W., Hyde, R.A., 2002a. Public support for land
preservation: measuring relative preferences in Delaware, FREC
research reports, FREC RR02-01. Department of Food and
Resource Economics, University of Delaware. http://www.udel.e-
du/FREC/PUBS/RR02-01.pdf.
Foster, V., Mourato, S., 2000. Valuing the multiple impacts of
pesticide use in the UK: a contingent ranking approach. Journal of
Agricultural Economics 51 (1), 1–21.
Furuseth, O., 1987. Public attitudes toward local farmland protection
programs. Growth and Change 18 (3), 49–61.
Halstead, J., 1984. Measuring the non-market demand value of
Massachusetts agricultural land: a case study. Journal of North-
eastern Agricultural Economics Council 13 (1), 12–19.
Hanley, N., Simpson, I., Parsisson, D., MacMillan, D., Bullock, C.,
Crabtree, B., 1996. Valuation of the conservation benefits of
environmentally sensitive areas. Draft final report to SOAEFD.
Environmental Economics Research Group, University of Stirling
and Macauly Land Use Research Institute.
Hanley, N., MacMillan, D., Wright, R.E., Bullock, C., Simpson, I.,
Parsisson, D., Crabtree, B., 1998a. Contingent valuation versus
choice experiments: estimating the benefits of environmentally
sensitive areas in Scotland. Journal of Agricultural Economics 49
(1), 1–15.
Hellerstein, D., Nickerson, C., Cooper, J., Feather, A.P., Gadsby, D.,
Mullarkey, D., Tegene, A., Barnard, C., 2002. Farmland protection
and the role of public preferences for rural amenities report 815.
USDA Economic Research Service, November, http://www.ers.us-
da.gov/publications/aer815/.
Kline, J., Wichelns, D., 1994. Using referendum data to characterize
public support for purchasing development rights to farmland.
Land Economics 70 (2), 223–233.
Kline, J., Wichelns, D., 1996. Public preferences regarding the goals of
farmland preservation programs. Land Economics 72 (4), 538–549.
Krieger, D., 1999. Saving open spaces: public support for farmland
protection, Working paper series wp99-1. Centre for agriculture in
the environment, April 1999.
Layton, D., Brown, G., 2000. Heterogeneous preferences regarding
global climate change. The Review of Economics and Statistics 82
(4), 616–624.
McLeod, D., Woirhaye, J., Menkhaus, D., 1999. Factors influencing
support for rural land use control: a case study. Agricultural and
Resource Economics Review 28 (1), 44–56.
OECD, 2002. Agricultural policies in OECD countries: a positive
reform agenda. Joint working part on agriculture and trade,
November.
Ozanne, A., Hogan, T., Colman, D., 2001. Modelling moral hazard in
agri-environmental policy: the case of standard fixed contract
payments. European Review of Agricultural Economics 28 (3),
:–347.
Potter, C., Burney, J., 2002. Agricultural multifunctionality in the
WTO—legitimate non-trade concern or disguised protectionism?
Journal of Rural Studies 18, 35–47.
Randall, A., 2002. Valuing the outputs of multifunctional agriculture.
European Review of Agricultural Economics 29 (3), 289–307.
Ready, R., Berger, M., Blomquist, G., 1997. Measuring amenity
benefits from farmland: hedonic pricing vs. contingent valuation.
Growth and Change 28 (4), 438–458.
Rosenberger, R., Walsh, R., 1997. Nonmarket value of western valley
ranchland using contingent valuation. Journal of Agricultural and
Resource Economics 22 (2), 296–309.
Sagoff, M., 1988. Some problems with environmental economics.
Environmental Ethics 10, 55–74.
Scottish Executive, 2002. Custodians of change: report of the
agriculture and environment working group, www.scotland.go-
v.uk/library5/agri/aewg-00.asp.
Stewart, L., Hanley, N., Simpson, I., 1997. Economic valuation of the
agri-environment schemes in the United Kingdom. Report to HM
Treasury and MAFF. EERG, University of Stirling, September
1997.
Tzeng, G.-H., Tsaur, S.-H., Laiw, Y.-D., Opricovic, S., 2002. Multi-
criteria analysis in Taipei: public preferences and improvement
strategies. Journal of Environmental Management 65, 109–120.
Variyam, J., Jordan, J., Epperson, J., 1990. Preferences of citizens for
agricultural policies: evidence from a national survey. American
Journal of Agricultural Economics 72 (2), 257–267.