Research and Development - CAS · John Poulsen, Boru Douthwaite, and Douglas White 41 Assessing...

International Conference on Impact of AgriculturalResearch and Development:

Why has Impact Assessment Research not MadeMore of a Difference?

Proceedings of a conference organized by the

Standing Panel on Impact Assessment (SPIA) of the Interim Science Council,Consultative Group on International Agricultural Research (CGIAR)

and

The Economics Program,The International Maize and Wheat Improvement Center (CIMMYT)

4-7 February 2002Tryp Corobici Hotel, San José, Costa Rica

David J. Watson, editor

ii

CIMMYT® (www.cimmyt.org) is an internationally funded, nonprofit, scientific research and training organization.Headquartered in Mexico, CIMMYT works with agricultural research institutions worldwide to improve theproductivity, profitability, and sustainability of maize and wheat systems for poor farmers in developing countries. Itis one of 16 food and environmental organizations known as the Future Harvest Centers. Located around the world,the Future Harvest Centers conduct research in partnership with farmers, scientists, and policymakers to helpalleviate poverty and increase food security while protecting natural resources. The centers are supported by theConsultative Group on International Agricultural Research (CGIAR) (www.cgiar.org), whose members include nearly60 countries, private foundations, and regional and international organizations. Financial support for CIMMYT’sresearch agenda also comes from many other sources, including foundations, development banks, and public andprivate agencies.

Future Harvest® builds awareness and support for food and environmental research for a world with less poverty, ahealthier human family, well-nourished children, and a better environment (www.futureharvest.org).

© International Maize and Wheat Improvement Center (CIMMYT) 2003. All rights reserved. The opinions expressedin this publication are the sole responsibility of the authors. The designations employed in the presentation ofmaterials in this publication do not imply the expression of any opinion whatsoever on the part of CIMMYT or itscontributory organizations concerning the legal status of any country, territory, city, or area, or of its authorities, orconcerning the delimitation of its frontiers or boundaries. CIMMYT encourages fair use of this material. Propercitation is requested.

Correct citation: Watson, D.J. (ed). 2003. International Conference on Impacts of Agricultural Research and Development:Why has Impact Assessment Research not Made More of a Difference? Proceedings of a conference organized by theStanding Panel on Impact Assessment (SPIA) of the Interim Science Council, Consultative Group on InternationalAgricultural Research (CGIAR), and the Economics Program, the International Maize and Wheat ImprovementCenter (CIMMYT), 4-7 February 2002, San José, Costa Rica. Mexico D.F.: CIMMYT.

Abstract: This document contains the proceedings of an international conference on impact assessment research heldin San José, Costa Rica, from 4-7 February 2002. Co-organized by the Economics Program of the International Maizeand Wheat Improvement Center (CIMMYT) and the Standing Panel on Impact Assessment (SPIA) of the ConsultativeGroup on International Agricultural Research (CGIAR), the conference was attended by approximately 150 impactassessment practitioners and users. Participants included representatives from CGIAR Centers, national agriculturalresearch organizations, public and private universities, multilateral lending organizations, development assistanceagencies, non-governmental organizations (NGOs), philanthropic foundations, private corporations, and the media.The goal of the conference was to examine why many impact assessment studies have had limited influence on theway that agricultural research is conducted and have failed to ensure adequate and sustainable financial support forresearch. Participants concluded that the success of future impact assessment studies could be improved by: (1) bettermatching impact assessment results to the needs of decision makers; (2) making impact assessments more credibleand more understandable, without losing rigor; and (3) improving methods for assessing a broader array of impacts,beyond traditional economic measures.

ISBN: 970-648-104-4.AGROVOC descriptors: Economic situation; Economic growth; Environmental factors; Policies; Poverty; Health;Human nutrition; Agricultural resources; Agricultural research; Agricultural development; Production; Internationalorganizations; Developing countries; Costa Rica.Additional key words: CGIAR; CIMMYT.AGRIS category codes: E14 Development Economics and Policies; A50 Agricultural Research.Dewey decimal classification: 338.91.

Printed in Mexico.

iii

Contentsvii Foreword

viii Abbreviations and acronymsix Executive summaryxi Conference papers published to date

xiii Acknowledgments1 Conference program7 Discussion session presentations7 Discussion session 1: Donor perspectives7 1. IFAD’s contribution

Rodney Cooke9 2. USAID’s contribution

Dana G. Dalrymple12 3. GTZ’s contribution

Stephan Krall13 Discussion session 2: Media perspectives13 1. Future Harvest: Science for Food, the Environment, and the World’s Poor

Barbara Alison Rose16 2. Forging an alliance to win, win, win all the way

G. Venkataramani18 3. Placing Stories in the Media

Gideon Litchfield21 Discussion session 3: What next for impact assessment?21 1. Where Do We Go From Here?

Hans Gregersen and Tim Kelley23 Abstracts of conference papers23 Milestones in impact assessment research in the CGIAR, 1970-1999

Prabhu L. Pingali24 The Green Revolution: An end-of-century perspective

Robert Evenson and Douglas Gollin25 Under-investment in agricultural R&D revisited

Johannes Roseboom26 Impact of agricultural research in Bangladesh: Productivity, economic returns,

and varietal replacement issuesJoseph G. Nagy and Md. Ferdous Alam

27 Herding cats: Is impact assessment the ultimate exercise in futility?Patti Kristjanson and Philip Thornton

28 Resolving conflicting evidence about the impact of the Green RevolutionPeter Hazell

29 Professional error, critical awareness and good scienceRobert Chambers

30 Why has the Green Revolution passed by sub-Saharan Africa?Simeon Ehui, Samuel Benin, John Pender, and Mark Rosegrant

31 Measuring the benefits of international agricultural economics researchDavid E. Schimmelpfennig and George W. Norton

32 Evaluating the impact of economic policy research: Concepts, practices and lessonsJim Ryan

34 Why impact assessment research has not made more of a difference:Lessons from eastern and southern Africa

P. Anandajayasekeram and M. Rukuni35 Estimating the benefits of plant breeding research: Methodological issues and practical challenges

Michael L. Morris and Paul W. Heisey37 The impact of agricultural maintenance research: The case of leaf rust resistance breeding in

CIMMYT-related spring bread wheatC.N. Marasas, M. Smale, and R.P. Singh

iv

38 Economic impact of post-harvest research in potato and sweetpotato in developing countriesKeith O. Fuglie and Thomas S. Walker

39 Impact assessment in natural resource management researchJohn Poulsen, Boru Douthwaite, and Douglas White

41 Assessing environmental impacts of productivity-enhancing crop research:Concepts, evidences, and challenges

Mywish K. Maredia, Prabhu Pingali, and Michael Nelson42 Measuring the impact of user participation in natural resource management research

Nancy Johnson, Nina Lilja and Jacqueline Ashby43 Golden Rice: What role could it play in alleviation of vitamin A deficiency?

David Dawe, Richard Robertson and Laurian Unnevehr44 The impact of iron bio-availability-enhanced diets on the health and productivity of school children:

Evidence from a mungbean feeding trial in Tamil Nadu, IndiaKatinka Weinberger

45 Agriculture and nutrition: Adversaries, bedfellows or allies?Lawrence Haddad

46 What GIS can (and can’t) bring to impact assessment: Novel data, analysis, and insightsGerald C. Nelson

47 GIS tools: They’re not just for experts anymoreD.P. Hodson and J.W. White

48 Lost in space: Enhancing the role of spatial analysis in strategic impact assessmentStanley Wood and Jordan Chamberlin

49 Assessing the impact of agricultural research on poverty with the sustainable livelihoods framework:Concepts and methods

Michelle Adato and Ruth Meinzen-Dick50 Impact of rice research on poverty alleviation: The Bangladesh case

Mahabub Hossain, David Lewis, Manik Lal Bose, Alamgir Chowdhury, Ruth Meinzen-Dick51 Impact of improved vegetable and fish pond technologies on poverty in Bangladesh

Kelly Hallman, David Lewis, Suraiya Begum, and Agnes Quisumbing52 The impact of improved maize germplasm on poverty alleviation:

The case of Tuxpeño-derived material in MexicoMauricio R. Bellon, Javier Becerril, Michelle Adato, and José Alfonso Aguirre Gómez

53 The economic costs and benefits of a participatory project to conserve maize landraces on farmsSmale, M, Bellon, M, Aguirre, A. Manuel, J. Mendoza, A.M. Solano, R. Martinez, and A. Ramirez

54 The distribution of benefits from public international germplasm banks:The case of beans in Latin America

Oswaldo Voysest, Nancy Johnson, and Douglas Pachico55 Endowing future harvests: The long-term costs of conserving genetic resources at the CGIAR centers

Bonwoo Koo, Philip G. Pardey, and Brian D. Wright56 The simple econometrics of impact assessment: Theory with an application to milk-market

participation in the Ethiopian HighlandsGarth Holloway and Simeon Ehui

57 Measurement and sources of efficiency in Argentina’s agricultural sciences research system:A stochastic frontier analysis

Daniel Lema, Susana Mirassou, and Martin Guppy58 Making an impact with impact assessment: The tradeoff analysis approach and lessons

from the tradeoffs project in Ecuador and PeruJohn Antle, Jetse Stoorvogel, Walter Bowen, Charles Crissman, and David Yanggen

59 Factors affecting the adoption of maize technologies in the hills of NepalJ. K. Ransom, K. Paudyal, and K. Adhikari

60 Variety characteristics, transactions costs and maize adoption in HondurasHernando Hintze, Mitch Renkow, and Gustavo Saín

61 Factors affecting the adoption of selected wheat production technologies by farmers in Njoro andRongai Divisions of Nakaru District, Kenya

Alice Ndiema, W. Ongo’ndo, A. A. Abdallahi, and M. G. Kinyua62 Expanding the use of impact assessment and other evaluation research evidence

Ronald Mackay and Douglas Horton

v

63 Why does impact assessment continue to neglect institutional sustainability?Daniel Ticehurst, Simon Henderson, and Alistair Sutherland

64 Demand-driven technological change and the traditional cereals in sub-Saharan Africa: The Malian caseJeff Vitale and John H. Sanders

65 An assessment of IPGRI’s impact on international policy-making: A case study of the internationalundertaking on plant genetic resources (1996-2001)

Raphaël Sauvé and Jamie Watts66 International R&D spillovers and productivity growth in the agricultural sector:

A panel co-integration approachLuciano Gutierrez

67 Impact of modern technology adoption on output growth and sustainability of major cerealproduction in Bangladesh

S.M. Fakhrul Islam68 Component analysis of total factor productivity in high potential rice-wheat systems in the Indian Punjab

Joginder Singh and M. Hossain69 Effects of innovative wheat breeding for marginal environments

Maximina A. Lantican, Prabhu L. Pingali, and Sanjaya Rajaram70 The impact of bean research in Honduras

David Mather, Richard Bernsten, Juan Carlos Rosas, Abelardo Viana, Danilo Escoto, and Julio Martinez71 A study of Philippine peanut farming communities: Impacts of new peanut CRSP technology and

influences on sustainabilityRobert L., Moxley, Aida Librero, and David Alston

72 Impact of public sector and private sector in R&D and technology generation: The case of maize in AsiaRoberta V. Gerpacio

73 Winding up the impact pathway: An approach to strengthen the impact orientation of nationalagricultural research

Andreas Springer-Heinze, Frank Hartwich, Simon Henderson, Douglas Horton, and Isaac Minde75 The importance of impact assessment studies for the Brazilian agricultural research system

Antonio Flavio Dias Avila and Geraldo da Silva e Souza77 Potato production and pesticide use in Ecuador: Linking impact assessment research and rural

development intervention for greater eco-system healthCharles Crissman, Donald Cole, Steve Sherwood, Patricio Espinosa and David Yanggen

78 The welfare effects of maize technologies in marginal and high-potential regions of KenyaDaniel D. Karanja and Mitch Renkow

79 Agricultural research and urban poverty: The case of ChinaShenggen Fan, Cheng Fang and Xiaobo Zhang

80 Impact of the adoption of modern varieties of rice on productivity gains and income distributionfor the irrigated and the rainfed ecosystem

Mahabub Hossain, Manik Lal Bose, Tran Thi Ut, A.G. Agarwal, Jawhar Thakur, and Esther B. Marciano81 Household resource endowments and the impacts of soil fertility management

Meredith J. Soule and Keith D. Shepherd82 Socio-economic, ecological, and policy impact assessment in the introduction of a transgenic staple crop

variety to the developing world: The Insect Resistant Maize for Africa Project, KenyaAdrian Ely, Hugo de Groote, Steven Mugo

83 A farm-level evaluation of the impact of IPM on pesticide use: A Comparative analysis of IPMand non-IPM trained farmers in Zimbabwe’s smallholder sector

S. Siziba and M. Mekuria84 Impact assessment of biological control in Africa—20 years experience of the

International Institute of Tropical AgriculturePeter Neuenschwander, Hugo de Groote, and Boru Douthwaite

85 A socioeconomic analysis of farmer field schools implemented by the national programin integrated pest management in Thailand

Suwanna Praneetvatakul and Hermann Waibel86 Network approach in soil management research: IWMI’s experience in southeast Asia

Amado R. Maglinao, Djoko Santoso, and Frits Penning

vi

87 The impact of Rockefeller funded forum for natural resources management program inEast and southern Africa

Bharati K Patel, Paul Woomer and Alexander R. Phiri88 Impact of the coordinated rice improvement program on movement of improved germplasm and

productivity gains across Indian statesAldas Janaiah, Mahabub Hossain, and E. Cabrera

89 Impact of regional agricultural research networks: The case of the regional maize program forCentral America and the Caribbean

Miguel I. Gómez and Prabhu L. Pingali90 Can impact analysis be used for research evaluation?

Javier M. Ekboir91 An evaluation approach for achieving and attributing impact for INRM and IPM

Boru Douthwaite, Thomas Kuby, Elske van de Fliert, and Steffen Schulz92 From measuring impact to learning institutional lessons: An innovation systems perspective on

improving the management of international agricultural researchAndrew Hall

93 Disciplines, institutions, and organizations: Impact assessments in contextRajeswari S. Raina

94 Tradeoff analysis as a tool for assessment of economic and environmental impacts of agricultural researchDavid Yanggen, John Antle, Jetse Stoorvogel, Walter Bowen, and Charles Crissman

95 Agricultural development and impacts on the environment: Experiences from IndiaD.D. Naik and Archana Godbole

96 Impact of salinity management research in Northwest IndiaK.K. Datta, Laxmi Tewari, and P.K. Joshi

97 Economic and institutional factors affecting the adoption of soil conservationtechnologies in Central America

Gustavo Saín and Monika Zurek98 Evaluating capacity development of the Plant Genetic Resources Center in Bunso, Ghana

Samuel Bennett-Lartey, Raymond Vodouhe, and Jamie Watts99 Evaluating capacity development in research and development organizations:

Emerging results of a global studyDouglas Horton and Nancy Alexaki

100 The costs of transforming public extension services towards participatory approaches:The case of Egypt

Gerd Fleischer, Hermann Waibel, and Gerd Walter-Echols

101 List of participants

vii

This document reports on the outcome of aninternational conference on impact assessment inagricultural research. The conference, which took placein Costa Rica in February of 2002, was co-organized bythe Standing Panel on Impact Assessment (SPIA),which is part of the interim Science Council of theConsultative Group on International AgriculturalResearch (CGIAR), in collaboration with the EconomicsProgram of the International Maize and WheatImprovement Center (CIMMYT).

In recent years, the CGIAR and other agriculturalresearch organizations have confronted two closelyrelated problems. First, concern is growing within thedonor community relating to the effectiveness ofexisting impact assessment research in guidinginternational agricultural research and technologytransfer. Second, donor support for agriculturalresearch is declining, despite the credible assessmentsshowing that investment in this area indeed has hadhigh returns. In response to these problems, SPIAconvened a workshop in Rome in May, 2000, on thetheme “The future of impact assessment in the CGIAR:Needs, constraints and options” (for a summaryproceedings from the Rome workshop, see TAC 2001).The objective of the workshop, which was attended byrepresentatives of all 16 CGIAR centers, was to improvethe efficiency and effectiveness of impact assessmentresearch within the CGIAR.

The idea for the 2002 Costa Rica conference emergedfrom the 2000 Rome workshop. In addition to scientistsfrom the CGIAR centers, the Rome workshop wasattended by key representatives from the donorcommunity, a number of whom suggested that theCGIAR needed to spend more time and effortaddressing the question, “Why has the great volume of

Foreword

credible, positive impact assessment research not mademore of a difference in terms of decision making—bothat the center level in terms of program reformulationand at the donor level in terms of bolstering support/resources for agricultural research?” A follow-upconference on that theme subsequently was proposedby the then head of the CIMMYT Economics Program,Prabhu Pingali.

The Costa Rica conference was attended by 145 persons,including a variety of experts in impact assessment andevaluation research, representatives from 15 CGIARcenters, many national research organizations, publicand private universities, multilateral lendingorganizations, development assistance agencies, NGOs,philanthropic foundations, private corporations, and themedia. The interactions and debates among theparticipants were productive and at times heated,reflecting the different perspectives, backgrounds, andexperiences of those in attendance.

Progress in designing and implementing impactassessment research, and in increasing the usefulness ofsuch research, requires close and sustained interactionbetween the professionals who do the research andthose who use the results as an input into decisionmaking. It is the belief of the co-organizers of the CostaRica conference that the conference succeeded infostering such interaction. However, what was achievedin Costa Rica was only a beginning. SPIA, CIMMYT andthe rest of the CGIAR centers believe that the CGIARsystem – its sponsors, management, scientific advisorsand clients – must continue the interaction and thedebate. In that regard, the organizers would welcomeany comments or suggestions from conference attendeesand those others who read these proceedings.

Hans Gregersen, Chair Michael MorrisSPIA, interim Science Council Director, CIMMYT Economics Program

viii

Abbreviations and Acronyms

ADB Asian Development BankASARECA Association for Strengthening Agricultural Research in Eastern and Central AfricaAVRDC Asian Vegetable Research and Development CenterBGMV Bean golden mosaic virusCGIAR Consultative Group for International Agricultural ResearchCIAT Centro Internacional de Agricultura TropicalCIFOR Center for International Forestry ResearchCIMMYT International Maize and Wheat Improvement CenterCIP Centro Internacional de la PapaCONICET National Research Council of Science and Technology of ArgentinaCRSP Collaborative Research Support ProjectDICTA Dirección de Ciencia y Tecnología Agropecuaria, HondurasEMBRAPA Empresa Brasileira de Pesquisa AgropecuariaGIS Geographic information systemGMO Genetically modified organismGRENEWECA Genetic Resources Network of West and Central AfricaGTZ Deutsche Gesellschaft Fur Technische ZusammenarbeitHYV High-yielding varietyIBSRAM International Board for Soil Research and ManagementICAR Indian Council of Agricultural ResearchICARDA International Center for Agricultural Research in the Dry AreasICLARM International Center for Living Aquatic Resources ManagementICRAF International Centre for Research in AgroforestryICRISAT International Crops Research Institute for the Semi-Arid TropicsIDRC International Development Research CentreIDS Institute of Development Studies, UKIFAD International Fund for Agricultural DevelopmentIFDC International Fertilizer Development Center, USAIFPRI International Food Policy Research InstituteIITA International Institute of Tropical AgricultureILRI International Livestock Research InstituteINRM Integrated Natural Resource ManagementINTA Instituto Nacional de Tecnología Agrícola (Argentina)IPGRI International Plant Genetic Resources InstituteIPM Integrated pest managementIRRI International Rice Research InstituteiSC interim Science Council of the CGIARISNAR International Service for National Agricultural ResearchIWMI International Water Management InstituteKARI Kenyan Agricultural Research InstituteMV Modern varietyNARS National agricultural research systemNCAP National Centre for Agricultural Economics & Policy Research, ICAR, IndiaOPV Open pollinated varietyPCARRD Philippine Council for Agriculture, Forestry and Natural Resources

Research and DevelopmentPGRC Plant Genetic Resources Center, GhanaPRM Programa Regional de Maíz para Centroamérica y el CaribePROFRIJOL The Central American Bean Network (supported by SDC)SDC Swiss Agency for Development and CooperationUSAID United States Agency for International DevelopmentUSDA/ARS United States Department of Agriculture/Agricultural Research ServiceWARDA West Africa Rice Development Association

ix

Executive Summary

The CGIAR is an informal association of 58 public andprivate sector members that supports 16 internationalagricultural research centers known as Future HarvestCenters. CIMMYT is a Future Harvest Center. TheFuture Harvest Centers share a mandate to increaseagricultural productivity in developing countries,alleviate poverty, and enhance the sustainability of thenatural resource base. Successful achievement of thismandate depends on the ability of the Centers toidentify appropriate research priorities, effectivelymanage ongoing research, account for resourcesinvested in research and development activities, andbuild and maintain public support for internationalagricultural research. Effective impact assessmentshould play a key role in achieving the above. Theconference reported herein was organized to enhancethe effectiveness of CGIAR impact assessment efforts.Specific objectives included:

1. Providing a forum for publicizing “best practices” inimpact assessment research and for disseminatingresults of recent impact assessment research.

2. Fostering dialogue between impact assessmentpractitioners, both within the CGIAR andthroughout the larger research and developmentcommunities.

3. Demonstrating to donors that Future HarvestCenters are committed to organizational learning.

The program included invited and contributed paperssessions, panel presentations, free discussion sessions,and poster sessions. Participants highlightedexperiences and case studies of impact measurement inagricultural productivity; equity, poverty, social health,and nutrition; the environment; and institutions andhuman capital. Participants also described novelapproaches for assessing hard-to-measure impacts insuch areas as training and capacity-building,institutional strengthening, networking, participatoryresearch, and policy research. Many presentations,focusing principally on non-conventional approaches,analyzed the factors constraining the effective use ofimpact assessments research. Critical insights weregained regarding the design, process and use of impactassessment studies. Discussion also took place betweenthose practitioners conducting impacts research and the

potential users of such information. This includedrepresentatives from the donor community, the media,national research systems, NGOs and those involved inresearch prioritization in the CGIAR. Insights weregained regarding how to improve the effectiveness ofimpact assessment results, particularly applying thoseresults to research priority setting. Extensive discussionstook place between impact assessment practitioners andrepresentatives of the media and donor communitiesregarding the most appropriate ways to build andmaintain public support for international agriculturalresearch.

Major ideas for improving the success of future impactassessment studies included the following points:(1) better matching impact assessment results to theneeds of decision makers; (2) making impactassessments more credible and more understandable,without losing rigor; and (3) improving methods forassessing a broader array of impacts, beyond traditionaleconomic measures. To achieve the above, participantssuggested diverse actions, among them: interactingmore frequently with the impact assessment users;communicating with decision makers about findingsand their possible implications; going beyond a cost-benefit framework to provide richer information on thefactors influencing impacts; effectively publicizing anddisseminating results; recognizing that attribution maynot always be important; not limiting studies to successstories; establishing credible counterfactuals fortransparency and plausibility purposes; fullyinstitutionalizing impact assessment; and formulating aset of principles and strategic guidelines for future, ex-post impact assessment in the CGIAR.

To sustain the momentum generated during theconference, SPIA and CIMMYT planned to pursue thefollowing activities. Work on several aspects is alreadyunder way:

Conference website. The conference website would bemaintained through the end of 2002, and would beupdated with additional abstracts and links to all papersreceived. The complete list of conference participantswill be posted, including names, institutionalaffiliations, and contact information.

x

Publications. The conference program committeeconcluded that a conventional proceedings volumecontaining all 75 papers presented at the conferencewould not attract a great deal of attention. Therefore,efforts have been made to publish several themedcollections of conference papers targeted at specificgroups of users. The program committee has reviewedall conference papers and assembled several setsgrouped around four themes: (1) assessing the impactsof agricultural research; (2) impacts of agriculturalresearch—theory and evidence; (3) learning by doing—innovative approaches to evaluating agriculturalresearch; (4) returns to investment in plant geneticresource conservation and crop improvement research.The themed collections have been submitted toprofessional journals for publication as special issues. Alist of those published to date follows this summary.

Impact assessment website. An impact assessmentwebsite is being created. The aim is to provide a user-friendly portal to an extensive collection of informationof potential interest to the impact assessmentcommunity worldwide. It will include a bibliographicdatabase of impact assessment literature; links to the fulltext of selected impact assessment documents; capsulesummaries containing non-technical information aboutimpact assessment concepts, methods and results thatcould be used for public awareness messages; asearchable database containing a range of researchimpacts indicators; information about how to accessother on-line databases containing information ofrelevance to impact assessment research; a photo galleryof downloadable images related to impact assessmentcase studies; and facilities for electronic conferencingand on-line discussions. Much initial work has beendone on several of the above components. Completionand full implementation of the website will depend onthe availability of additional resources.

xi

Quarterly Journal of International Agriculture, vol. 42 (2003), No. 2Assessing the Impact of Agricultural Research: Theory and Evidence

Assessing the impact of agriculture research: An introductionPrabhu Pingali and Michael Morris

Assessing the impact of agriculture research: An overviewMichael Morris, Prabhu Pingali, Hans Gregersen, and Tim Kelley

Assessing the impact of agricultural research on poverty and livelihoodsMichelle Adato and Ruth Meinzen-Dick

Enhancing the role of spatial analysis in strategic impact assessment: Improving data resolution for regional studiesStanley Wood and Jordan Chamberlin

The tradeoff approach: Lessons from Ecuador and PeruJohn Antle Bozeman, Jetse Stoorvogel, Water Bowen, Charles Crissman, and David Yanggen

Measuring the benefits of international agricultural economics researchDavid E. Schimmelpfennig and George W. Norton

Special issue of Agricultural Economics, forthcoming in 2003Returns to investment in plant genetic resource conservation and crop improvement research

Part 1. Concepts and methodsEstimating the benefits of plant breeding research: Methodological issues and practical challenges

Michael Morris and Paul HeiseyThe economic impact of productivity maintenance research: Breeding for leaf rust resistance in modern wheat

Carissa Marasas, Melinda Smale, and Ravi SinghPart 2. Economics of genetic resource conservationThe economic costs and benefits of a participatory project to conserve maize landraces on farms in Oaxaca, Mexico

Melinda Smale, Mauricio Bellon, Javier Aguirre, Jorge Mendoza, Ana Maria Solano, Rafael Martinez, and AlejandroRamirez

The distribution of benefits from public international germplasm banks: The case of beans in Latin America NancyJohnson, Doug Pachico, and Osvaldo Voysest

Saving seeds: The cost of conserving genetic resources at the CGIAR centresBonwoo Koo, Philip Pardey, and Brian Wright

Part 3. Factors affecting the adoption of modern varietiesAdoption of improved maize varieties in the hills of Nepal

Joel Ransom, Kamal Paudyal, and Krishna AdhikariVariety characteristics and maize adoption in Honduras

Hernando Hintze, Mitch Renkow, and Gustavo SaínThe roles of public sector vs. private sector in R&D and technology generation: the case of maize in Asia

Roberta GerpacioPart 4. Distributional impacts of plant breeding researchWelfare effects of maize technologies in marginal and high potential regions of Kenya

Daniel Karanja, Mitch Renkow, and Eric CrawfordThe economic impact of bean disease resistance research in Honduras

David Mather, Richard Bernsten, Juan Carlos Rosas, Aberlardo Viana, Danilo Escoto, and Julio MartinezIs research on marginal lands catching up? The case of unfavourable wheat growing environments

Maximina Lantican, Prabhu Pingali, and Sanjaya Rajaram

Conference Papers Published to Date

xii

Special issue of Agricultural Systems, forthcoming in 2003Learning for the Future: Innovative Approaches to Evaluating Agricultural Research

Foreword Peter MatlonLearning for the future: Using evaluation to enhance organizational learning and improve performance

Douglas Horton and Ronald MackayExpanding the use of impact assessment and evaluation in agricultural research and development

Ronald Mackay and Douglas HortonWhy impact analysis should not be used for research evaluation and what the alternatives are

Javier EkboirDisciplines, institutions and organizations: Impact assessments in context

Rajeswari S. RainaFrom measuring impact to learning institutional lessons: an innovation systems perspective on improving the

management of international agricultural researchA.J. Hall, V. Rasheed Sulaiman, N.G. Clark, and B. Yoganand

Impact pathway evaluation: an approach for achieving and attributing impact in complex systemsBoru Douthwaite, Thomas Kuby, Elske van de Fliert, and Steffen Schulz

Using impact pathway analysis to strengthen the impact orientation of agricultural researchAndreas Springer-Heinze, Frank Hartwich, J. Simon Henderson, Douglas Horton and Isaac Minde

Measuring the impact of user participation in agricultural and natural resource management researchNancy Johnson, Nina Lilja, and Jacqueline A. Ashby

An analysis of IPGRI’s influence on the International Treaty on Plant Genetic Resources for Food and AgricultureRaphaël Sauvé and Jamie Watts

EIARD views on impact assessment and evaluation in agricultural research for developmentEuropean Initiative For Agricultural Research For Development (EIARD)

xiii

Funding for the conference was provided by the following sponsors:

• International Maize and Wheat Improvement Center (CIMMYT)• Interim Science Council (iSC) of the CGIAR• Norwegian Ministry of Foreign Affairs• The Rockefeller Foundation• International Fund for Agricultural Development (IFAD)• Deutsche Gesellschaft Fur Technische Zusammenarbeit (GTZ)• Japanese Ministry of Foreign Affairs• Cooperacion Cientifica International (Spain)

We also gratefully acknowledge the editorial assistance of CIMMYT science writer, Mike Listman, and the layout andcover contributed by CIMMYT designer, Antonio Luna Ávila.

Acknowledgments

1

Conference Program

Plenary Paper Session 1Impact Assessment in the CGIAR: The Road TraveledChairperson: Edgardo Moscardi (IADB)

Pl 1.1 Milestones in CGIAR impact assessment researchPrabhu Pingali (CIMMYT)

Pl 1.2 Assessing the impacts of international crop geneticimprovement research: Some lessons learnedRobert Evenson (Yale University) and Douglas Gollin(Williams College)

Contributed Papers Session 1Measuring Research Impacts: Quantitative MethodsChairperson: Richard Bernsten (Michigan State University)

C1.1 The simple econometrics of impact assessment: Theorywith an application to milk-market participation in theEthiopian highlandsGarth Holloway (ILRI) and Simeon Ehui (ILRI)

C1.2 Measurement and source of efficiency in Argentina’sagricultural science research system: A stochasticfrontier analysisDaniel Lema (INTA), Susana Mirassou (INTA), andMartin Guppy (INTA)

“Why has impact assessment research not made more of a difference?”

Monday, February 4, 2002

Opening SessionChairperson: Hans Gregersen (SPIA)

Opening address New science brings hope for people and the environmentTimothy Reeves (CIMMYT)

Keynote speech Impact assessment: Why should we care?Alex McCalla (University of California at Davis)

C1.3 The Tradeoff Analysis Approach: Lessons from thetradeoffs project in Ecuador and PeruJohn Antle (Montana State University), Jetse Stoorvogel(University of Wageningen), Walter Bowen (CIP, IFDC),Charles Crissman (CIP), and David Yanggen (CIP,Montana State University)

Contributed Papers Session 2Factors Affecting Adoption of Modern VarietiesChairperson: Erika Meng (CIMMYT)

C2.1 Factors affecting the adoption of maize technologies inthe hills of NepalJoel Ransom, Kamal Paudyal, and Krishna Adhikari(CIMMYT)

C2.2 Variety characteristics, transaction costs and maizeadoption in HondurasHernando Hintze and Mitch Renkow (North CarolinaState University); and Gustavo Saín (CIMMYT)

C2.3 Factors affecting the adoption of selected wheatproduction technologies by farmers in Njoro andRongai divisions of Nakuru District, KenyaAlice Ndiema and M.G. Kinyua (NPBRC); andW.N. Ongo’ndo (Egerton University)

2

Discussion Session 1: Donor Perspectives

One important group of users of impact assessment studies are the people who makedecisions about the allocation of research funding. Funding agencies increasingly demandaccountability from researchers, usually including concrete evidence that investments inresearch are paying off. In this session, a panel of donor agency representatives describesthe types of information that funding agencies need to make research-funding decisions.The panel will then engage the audience in a discussion designed to explore how results ofimpacts studies can be more effectively communicated to inform the various constituenciesthat influence funding for research and development.

Chairperson: Jim Ryan (Australian National University)Panelists: Stephan Krall (GTZ), Rodney Cooke (IFAD), Dana Dalrymple (USAID)

Tuesday, February 5, 2002

Plenary Paper Session 2Measuring Returns to Research InvestmentChairperson: Mahabub Hossain (IRRI)

Pl 2.1 Under investment in agricultural R&D revisitedJohannes Roseboom (ISNAR)

Pl 2.2 The impact of agricultural research in Bangladesh:Productivity, economic returns and varietalreplacement issuesJoseph Nagy (Consultant) and Ferdous Alam(Bangladesh Agricultural University)

Pl 2.3 Herding cats: Is impact assessment the ultimateexercise in futility?Patricia Kristjanson and Philip Thornton (ILRI)

C3.3 Demand driven technological change and thetraditional cereals in sub-Saharan Africa:The Malian caseJeff Vitale (Purdue University) and John Sanders (TexasA&M University)

C3.4 An assessment of IPGRI’s impact on internationalpolicy- making: A case study of the internationalundertaking on plant genetic resource (1996-2001)Raphaël Sauvé and Jamie Watts (IPGRI)

Contributed Papers Session 3Increasing Impacts of Agricultural R&D: Institutional andPolitical ConsiderationsChairperson: Patricia Kristjanson (ILRI)

C3.1 Expanding the use of impact assessment and otherevaluation research evidenceRon Mackay (Concordia University) andDoug Horton (ISNAR)

C3.2 Why does impact assessment continue to neglectinstitutional sustainability?Daniel Ticehurst, Simon Henderson, and AlistairSutherland (NRI)

Pl 2.3 Herding cats: Is impact assessment the ultimateexercise in futility?Patricia Kristjanson (ILRI) and Philip Thornton (ILRI)

Panel Session 1Crop Genetic Improvement ResearchChairperson: Robert Evenson (Yale University)

PA1.1 Estimating the benefits of plant breeding research:Methodological issues and practical challengesMichael Morris (CIMMYT) and Paul Heisey (USDA/ERS)

PA1.2 Returns to investment in maintenance research: Thecase of leaf rust resistance breeding in CIMMYT-related spring bread wheatCarissa Marasas (CIMMYT), Melinda Smale (IFPRI),and Ravi Singh (CIMMYT)

Conference Program

3

PA1.3 Economic impacts of post-harvest research for potatoand sweetpotato in developing countriesKeith Fuglie and Tom Walker (CIP)

Panel Session 2Natural Resource Management ResearchChairperson: Peter Hazell (IFPRI)

PA2.1 Impact assessment in natural resource managementresearchJohn Poulsen (INRM), Boru Douthwaite (IITA), andDouglas White (CIAT)

PA2.2 Environmental impacts of productivity-enhancing cropresearch: A critical reviewMywish Maredia (Michigan State University), PrabhuPingali (CIMMYT), and Michael Nelson (former WorldBank)

PA2.3 Measuring the impact of user participation in naturalresource management researchNancy Johnson and Jacqueline Ashby (CIAT); andNina Lilja (CGIAR)

Contributed Papers Session 4Impacts of Research Investment on Productivity GrowthChairperson: Mitch Renkow (North Carolina State University)

C4.1 International R&D spillovers and productivity growthin the agricultural sector: A panel co-integrationapproachLuciano Gutierrez (University of Sassari)

C4.2 Impact of modern technology adoption on outputgrowth and sustainability of major cereals productionin BangladeshFakhrul Islam (Bangabandhu Sheikh Mujibur RahmanAgricultural University)

C4.3 The contribution of different components of total factorproductivity in high potential rice-wheat systems inIndian PunjabJoginder Singh (Punjab Agricultural University)

Contributed Papers Session 5Impacts of Commodity ResearchChairperson: Robert Chambers (IDS)

C5.1 Effects of innovative wheat breeding in marginalenvironmentsMaximina Lantican, Prabhu Pingali, and SanjayaRajaram (CIMMYT)

C5.2 The impact of bean research in HondurasDavid Mather and Richard Bernsten (Michigan StateUniversity); and Juan Carlos Rosas, Aberlardo Viana,Danilo Escoto, and Julio Martinez (Escuela AgrícolaPanamericana)

C5.3 A study of Philippine peanut farming communities:Impacts of new peanut RSP technology and influenceson sustainabilityRobert Moxley (North Carolina State University), AidaLibrero (PCARRD), and Dave Alston (University ofMaryland-Eastern Shore)

C5.4 Impact of public sector versus private sector in R&Dand technology generation: The case of maize in AsiaRoberta Gerpacio (CIMMYT)

Contributed Papers Session 6Use of Impact assessment Research in NARSsChairperson: Nancy Johnson (CIAT)

C6.1 Winding up the impact pathway: An approach tostrengthening the impact orientation of nationalagricultural researchAndreas Springer-Heinze (GTZ); Frank Hartwich andDoug Horton (ISNAR); Simon Henderson (NRI); andIsaac Minde (ASARECA)

C6.2 The importance of impact assessment studies for theBrazilian agricultural research systemAntonio Flavio Dias Avila and Geraldo Da Silva eSouza (EMBRAPA)

C6.3 Potato production and pesticide use in Ecuador:Linking impact assessment research and ruraldevelopment intervention for greater ecosystem healthCharles Crissman, Steve Sherwood, and PatricioEspinosa (CIP); and Donald Cole (University of Toronto)

Conference Program

4

Discussion Session 2: Media Perspectives

Impact studies often reveal issues of great potential interest to the general public and topolicy makers in developing and developed countries, but these issues are rarelycommunicated in a compelling manner outside a narrow community of academics anddevelopment professionals. How can the results of impact studies be communicated morewidely? In this session, a panel of distinguished journalists will share their experiences andoffer insights into how stories related to impact assessment can be communicated effectivelyto policy makers, opinion leaders, and the general public.

Chairperson: Timothy Reeves (CIMMYT)Panelists: Barbara Rose (Future Harvest), Gideon Lichfield (The Economist), G. Venkataramani (The Hindu)

Wednesday, February 6, 2002

Panel Session 4Enhancing Research Impacts through GISChairperson: Jim Ryan (Australian National University)

PA4.1 What GIS can (and can’t) bring to impact assessment:Novel data, analysis, and insightsGerald Nelson (University of Illinois at Champaign-Urbana)

PA4.2 GIS tools: They’re not just for experts anymoreDave Hodson and Jeff White (CIMMYT)

PA4.3 Lost in space: Fulfilling the promise of spatial analysisin impact assessmentStan Wood and Jordan Chamberlin (IFPRI)

Contributed Papers Session 7Distributional Impacts of Technical Change: ModelingApproachesChairperson: Michael Morris (CIMMYT)

C7.1 The welfare effects of maize technologies in marginaland high potential regions in KenyaDaniel Karanja (Michigan State University) and MitchRenkow (North Carolina State University)

C7.2 How agricultural research affects urban poverty indeveloping countries: The case of ChinaShenggen Fan, Cheng Fang, and Xiaobo Zhang (IFPRI)

C7.3 Impact of the adoption of modern varieties of rice onproductivity gains and income distribution for theirrigated and rain-fed ecosystemMahabub Hossain, Manik Lal Bose, Tran Thi Ut, A.G.Agarwal, Jawhar Thakur, and Esther Marciano (IRRI)

Plenary Paper Session 3The Green Revolution: A Retrospective ViewChairperson: Peter Matlon (Rockefeller Foundation)

Pl 3.1 Resolving conflicting evidence about the impact of theGreen RevolutionPeter Hazell (IFPRI)

Pl 3.2 Professional error, critical awareness and good scienceRobert Chambers (IDS)

Pl 3.3 Why the Green Revolution failed in Africa and howthis impacted the poorSimeon Ehui (ILRI)

Panel Session 3Potential Impacts of Nutritional Improvement StrategiesChairperson: David Schimmelpfennig (USDA/ERS)

PA3.1 Golden rice: What role could it play in alleviation ofVitamin A deficiency?David Dawe (IRRI); and Richard Robertson and LaurianUnnevehr (University of Illinois at Champaign-Urbana)

PA3.2 Assessing the indirect impact of mungbeans onnutrition and productivity; new insights from casestudies in Pakistan and IndiaKatinka Weinberger (AVRDC)

PA3.3 Agriculture and nutrition: Adversaries, bedfellows orallies?Lawrence Haddad (IFPRI)

Conference Program

5

C7.4 Household resource endowments and impacts of soilfertility managementMeredith Soule (USAID) and Keith Shepherd (ICRAF)

Contributed Papers Session 8Impacts of Pest Control TechnologiesChairperson: Dave Watson (CIMMYT)

C8.1 Socio-economic, ecological, and policy impactassessment in the introduction of a transgenic staplecrop variety to the developing world: Insect resistantmaize for AfricaAdrian Ely (University of Sussex); and Hugo de Grooteand Steven Mugo (CIMMYT)

C8.2 A farm level evaluation of the impact of IPM onpesticide use: A comparative analysis of IPM trainedand ordinary farmers in Zimbabwe’s smallholder sectorShephard Siziba (University of Zimbabwe, CIMMYT)and Mulugetta Mekuria (CIMMYT)

C8.3 Impact assessment of biological control in Africa:Twenty years experience of the International Institute ofTropical AgriculturePeter Neuenschwander and Boru Douthwaite (IITA);and Hugo de Groote (CIMMYT)

C8.4 A socio-economic analysis of farmers’ field schoolsimplemented by the national program in integratedpest management of ThailandSuwanna Praneetvatakul and Hermann Waibel(Hannover University)

Contributed Papers Session 9Economics Benefits of Research CollaborationChairperson: Gustavo Saín (CIMMYT)

C9.1 Network approach in soil management research:IWMI’s experience in Southeast AsiaAmado Maglinao, Djoko Santoso, and Frits Penningde Vries (IWMI)

C9.2 The impact of Rockefeller funded Forum for NaturalResources Management Program in eastern andsouthern AfricaAlex Phiri (Bunda College of Agriculture)

C9.3 Impact of the Coordinated Rice Improvement Programon movement of improved germplasm and productivitygains across the Indian statesAldas Janaiah, Mahabub Hossain, and E. Cabrera(IRRI)

C9.4 Economic benefits of research cooperation: The case ofthe Regional Maize Program for Central America andthe CaribbeanMiguel I. Gomez (Cornell University) and PrabhuPingali (CIMMYT)

Thursday, February 7, 2002

Plenary Paper Session 4Impact assessment: Research and Policy MakingChairperson: Tim Kelley (CGIAR-TAC)

Pl 4.1 Measuring the benefits of international agriculturaleconomics researchDavid Schimmelpfennig (USDA/ERS) and GeorgeNorton (Virginia Tech University)

Pl 4.2 Evaluating the impact of economic policy research:Concepts and practicesJim Ryan (Australian National University)

Pl 4.3 Why has impact assessment research not made moreof a difference?Ponniah Anandajayasekeram (ISNAR) and MandiRukuni (FAO)

Panel Session 5Does Agricultural Research Alleviate Poverty?Chairperson: Anthony Bebbington (University of Colorado at Boulder)

PA5.1 Assessing the impacts of agricultural research onpoverty using the sustainable livelihoods framework:Concepts and methodsMichelle Adato and Ruth Meinzen-Dick (IFPRI)

PA5.2 Impact of rice research on poverty reduction: The caseof BangladeshMahabub Hossain (IRRI); and David Lewis, Manik LalBose, Alamgir Chowdhury, and Ruth Meinzen-Dick(IFPRI)

PA5.3 Improved vegetable and fishpond technology onpoverty in BangladeshKelly Hallman (PRDPC); and David Lewis, SuraiyaBegum, and Agnes Quisumbing (IFPRI)

PA5.4 The impact of improved maize germplasm on povertyalleviation: The case of Tuxpeño-derived material inMexicoMauricio Bellon and Javier Becerril (CIMMYT); andMichelle Adato (IFPRI)

Conference Program

6

Panel Session 6Benefits of Genetic Resources ConservationChairperson: Robert Chambers (IDS)

PA6.1 Economic costs and benefits of a participatory projectto conserve maize landraces on farmsMelinda Smale (IFPRI); and Mauricio Bellon, JavierAguirre, Jorge Mendoza, Ana Maria Solano, RafaelMartinez, and Alejandro Ramirez (CIMMYT)

PA6.2 The distribution of benefits from public internationalgermplasm banks: The case of beans in Latin AmericaOswaldo Voysest, Nancy Johnson, and Doug Pachico(CIAT)

PA6.3 Endowing future harvests: The long-term costs ofconserving genetic resources at the CGIAR centersPhilip Pardey (University of Minnesota), Brian Wright(University of California at Berkeley), and Bonwoo Koo(IFPRI)

Contributed Papers Session 10Non-Conventional Approaches to Impact assessment:Institutionalist PerspectivesChairperson: Simeon Ehui (ILRI)

C10.1 Can impact analysis be used for research evaluation?Javier Ekboir (CIMMYT)

C10.2 An evaluation approach for achieving and attributingimpact for INRM and IPMBoru Douthwaite (IITA); and Thomas Kuby and SteffenSchultz (GTZ)

C10.3 From measuring impact to learning institutional lessons:An innovation systems perspective on improving themanagement of international agricultural researchAndrew Hall (ICRISAT), Rasheed Sulaiman V. (NCAP),Norman Clark (University of Strathclyde), andYoganand B. (ICRISAT)

C10.4 Disciplines, institutions and organizations: Impactassessment in contextRajeswari Raina (NISTADS)

Contributed Papers Session 11Environmental Impacts of Agricultural R&DChairperson: Nina Lilja (CGIAR)

C11.1 Tradeoff analysis as a tool for assessment of economicand environmental impacts of agricultural researchDavid Yanggen (CIP, Montana State University), JohnAntle (Montana State University), Jetse Stoorvogel(University of Wageningen), Walter Bowen (CIP,IFDC), and Charles Crissman (CIP)

C11.2 Agricultural development and impacts on theenvironment: Experiences from IndiaD.D. Naik (D.D. & Associates) and Archana Godbole(AERF)

C11.3 Impact of salinity management research in NorthwestIndiaK.K. Datta (ICAR); and Laxmi Tewari and P.K. Joshi(NCAP)

C11.4 Adoption and impact of soil conservation technologiesin Central AmericaGustavo Saín and Monika Zurek (CIMMYT)

Contributed Papers Session 12Impacts of Investment in Training and ExtensionChairperson: Gerald Nelson (University of Illinois at Champaign-Urbana)

C12.1 Evaluating capacity development of the Plant GeneticResources Center in Bunso, GhanaSamuel Bennett-Lartey (PGRC, Ghana), RaymondVodouhe (GRENEWECA), and Jamie Watts (IPGRI)

C12.2 Evaluating capacity development in research anddevelopment organizationsDouglas Horton and Nancy Alexaki (ISNAR)

C12.3 The costs of transforming public extension servicestowards participatory approachesGerd Fleischer (World Bank), Hermann Waibel(University of Hannover), and Gerd Walter-Echols(Egyptian-German Integrated Pest ManagementProject)

Closing sessionWhat Next for Impact assessment?Chairperson: Prabhu Pingali (CIMMYT)

Panelists: Tim Kelley (SPIA), Peter Matlon(Rockefeller Foundation), andSimeon Ehui (ILRI)

Conference Program

7

Discussion Session Presentations

Discussion Session 1: Donor Perspectives

1. IFAD’s contribution

Rodney Cooke (IFAD)

However, on a less positive note, even though IFAD isan international financial institution with a portfolio ofaround US$ 500 million a year, grant aid (to allactivities) equates to a maximum of 7.5% of our totalexpenditure. Unfortunately, agricultural research is along-term investment with a risky reputation. Manydonors see agricultural research as a blunt instrumentfor change. For these reasons, IFAD has been a financialsupporter of the IAEG (now SPIA) in order to furtherclarify the advantages of targeted, effective R&Dprograms.

Agricultural research leading to the adoption ofimproved technology may reduce rural poverty inmany ways. We must get away from the too simplisticinternal rates of return based on adoption of high-yielding varieties. New approaches could include:• higher on-farm yields;• expansion of farm employment opportunities and

higher wages;• growth of non-farm activities;• lower food prices;• reduced vulnerability to crop and other risks; and• empowerment of the poor and of their

organizations.

The conference began with the observation that if wedon’t care about the impact of agricultural research anddevelopment, we are dead in the water. In other words,we must be concerned to demonstrate cost-effectiveinterventions arising from donor investments.Why should IFAD be interested? First, IFAD’s focus ison rural poverty alleviation through agricultural andrural development supported by loans and grants. Amajor problem facing the agricultural research anddevelopment sector is the declining financial supportfor rural development research projects from nationalgovernments and the donor community. IFAD’scommitment to reverse this trend relates to theobservation that 75% of the world’s poor people live inrural areas. Yet, despite this, the proportion of officialdevelopment assistance for agriculture fell from 20% inthe 1980s to 12% today. This is despite the fact that therural poor depend primarily on agriculture (directly orindirectly) for their livelihoods.

IFAD is attempting to reverse this decline by a series ofpolicy dialogues such as the production of IFAD’sRural Poverty Report in 2001, a partnership with theWorld Bank and others to reverse the declining supportfor rural development, and not least a continuedsupport for the CGIAR which equates to over US$ 100million in the last 20 years.

8

The last item, empowerment, has gained increasingattention in recent years. Unless the poor have thepower to participate in deciding which technology touse, they are unlikely to benefit from it. In other words,better farm technology will most benefit the farmerswho are active partners in setting priorities of R&D.

An analytical framework for approaching these sixissues should cover the following:

• What is the probability that applied research will besuccessful, and hence, what is a desirable measure ofsuccessful research in an R&D continuum?

• How soon will the results be available for adoption,how widely applicable will the results be, and whenwill they be adopted by various groups and for howlong?

• Once adopted, what is their contribution toproductivity and incomes of different groups ofpeople, especially smallholders?

In terms of screening of research and developmentconcepts, IFAD assesses such proposals according tothree types of criteria: consistency with the IFADmandate (targeting of the proposal); effectiveness of theproposal (institutions identified have competence andcomparative advantage), multiplier effects, feasibilityof the approach and potential to deliver medium-termbenefits to the rural poor etc.); efficiency (value formoney).

Discusson Session Presentations

9

It is a pleasure to participate in this panel. It will be, Ithink, informative from several points of view. First, itis seldom that practitioners get to hear directly frommore than one donor at a time. Second, it is perhapseven more seldom that donors get to hear what otherdonors think about this issue and get feedback frompractitioners. So I look forward to this session as alearning experience as well.

My message could be summarized in a few words: weneed all kinds of impact information in various forms.We need summary information for administrators andCongress, and we need more detailed information touse with other staff members and to use ourselves inpresenting the work of the international agriculturalresearch community and in making out own internalbudget decisions. Ours is a case where one size doesnot fit all.

Setting for Decision-Making in the U.S.Government

The setting for decision-making in the U.S. governmentmay share some similarities with other governmentaldonors, but also may have some distinct qualities.Clearly, a wide range of factors is involved, and theirrelative importance may change over time. The listincludes many factors external to USAID (exogenous):world and national events, politics at all levels(including both politicians and their constituencies,lobby groups, and special interests); the nationalbudget; and other governmental agencies (in our case,the Department of State). Internal (endogenous) factorsinclude; the individual beliefs of high-level politicalappointments in the agency made by the political partyin power, and the on-going nature of USAIDbureaucracy.

Obviously, not all decisions are made on the basis of ahighly analytic process and component factors may beweighted differently than they would be by a group of

economists. The challenge is to bring as high a degreeof analysis into the decision-making process aspossible. This cannot be done unless the analyses existand are available in a form that can be readilycommunicated.

Government Performance and Results Act(GPRA).

There are some measures in place, and others that arebeing stressed, which are meant to add to the analyticaldimension in the U.S. Government. One that has beenin place since 1993 is the GPRA. This is a well-intentioned Act that has triggered a lot of activity at theagency level. I have been involved in responding to itsince it was enacted. If the task were left to economistsand they were given sufficient time and resources torespond to it, I think that the outcome would have beenbetter. The development of quantitative measures for awide range of government activities can be difficult:those that seem important may be difficult to measure;those that can be more easily measured may be trivial.More generally, as a recent newspaper article noted:“There seems to be little consensus on how to definegood performance and how to measure it.”

The Bush administration has made its views clear withrespect to aid in general, the role of productivity, andthe need for evidence of impact. With respect to foreignaid in general, Paul O’Neill, Secretary of the Treasury,has stated that “Over the past 50 years, the world hasspent a lot of money in the name of developmentwithout a great deal of success.” Administrationconcerns have also been expressed. In recent years“tens of billions of dollars in aid for developingcountries have produced disappointing results.” Theemphasis of President Bush is clearly on “improving,rather than increasing foreign aid.” A key factor will bethe degree to which the effectiveness of the aid can bedemonstrated.

2. USAID’s contribution

Dana G. Dalrymple (USAID)


10

Currently, there is a strong Administration emphasis onraising productivity and measuring the results. Thishas been expressed most clearly in relation to U.S.expectations of the World Bank. Secretary O’Neillstated that, “Development banks must focus theirefforts of raising productivity growth in the developingworld.” Similarly, President Bush stated, at the WorldBank, that the World Bank and other developmentbanks must “…focus on raising productivity indeveloping nations.” Moreover, “President Bush willask Congress to base future increases in aid to poorcountries on evidence that the aid is actually bringingprogress.” Bank funding to IDA will be conditional ondevelopment indexes demonstrating that aid wasproductive. The Bank should be “…more rigorous inmeasuring the results of its aid.” A senior StateDepartment official said that there might be moredevelopment assistance available “…provided onecould be assured that it will be effective in projectingsustained growth and development.” It isacknowledged that, “…developing benchmarks toshow what kind of aid works best is not a simpleprocess.”

While the initial emphasis appears to be on health andeducation, agriculture has not been overlooked.Secretary O’Neill commented that “Economic historyhas taught us, for example, that investing inagriculture…is a key to development.” And AlanLarson, Undersecretary for Economics, Business andAgricultural Affairs at the State Department has noted,in the context of hunger, the need to “…increaseagricultural research in areas of importance todeveloping countries. The levels of support foragricultural research in developing countries havetended to decline. We need to find a way to build thoseup and support some of the institutions that have beena network for disseminating information aboutagriculture around the world.” USAID administratorAndrew Natsios has repeatedly expressed his supportof agricultural development.

Types of Information Needed by USAID

Now, I will turn more specifically to CGIAR matters, asviewed from where I sit at the staff level. Basically, ouroffice gets a certain amount of money that we can usefor unrestricted core funding of the CGIAR system. Thefactors that determine the amount we receive arelargely external (exogenous) to my office, but theallocation decisions are largely internal (endogenous).Our office makes recommendations on how they mightbe allocated and these are discussed with andsanctioned at higher administrative levels in ourbureau. On a voluntary basis, other offices provideadditional restricted funding for special projects andwe usually play no role in its allocation.

The exogenous factors that determine the amount offunding available to us are usually not very clear at thestaff level. They are partly determined by the overallagency funding request for development assistanceand the congressional response. Congress often willprovide earmarks for special activities or programswithout increasing overall funding. Biotechnology is arecent example. These earmarks range from mild tostrong. And the stronger they are, the more influencethey have on the amount of discretionary fundingremaining. CGIAR funding in the past has been in thediscretionary to mild earmark stage. It is not in thestrongest position.

The decisions concerning the actual allocation of theCGIAR funding have basically been made at the officelevel (principally Robert Bertram and myself) inconsultation with (1) the regional bureaus (the AfricaBureau has provided $2 million of funding in recentyears), (2) the environment office, (3) other members ofour office, and (4) those higher in the administrativestructure, especially those that serve as the officialUSAID representative to the CGIAR (Emmy Simmonsin recent years). We have drawn on our personalknowledge of the system, recent reports and activities,and USAID interests and priorities. Generally, year-to-year changes have been modest.


11

Information needs are different at the exogenous(external) and endogenous (internal) levels. In terms ofbuilding a case at the exogenous level, it is more a caseof selling the CGIAR system as a whole, drawing oncenter accomplishments and activities. It is partly amatter of building good will. At the endogenous level,allocating funding, it is more a question of individualcenter performance and changing needs and priorities.In terms of impact analyses, at the exogenous level weneed center-wide regional or global information onimpacts. This needs to be pulled together. TAC doessome of this in its thematic studies, but we need a moresystematic way of compiling and summarizingindividual center studies, such as is done in the IFPRI2020 briefs. At the endogenous level, we can makebetter use of what we have available. A moresystematic summarization, as noted above, would beuseful. It would also be useful to have a clearer idea ofthe strengths and weaknesses of the analyses,preferably done by a third party.

Summary Remarks

The process followed by the U.S. government inmaking funding allocation decisions, probably likemany others, may not seem entirely analytical andrational from some viewpoints. It depends on whereyou sit and how you weigh individual factors. But it isclear the Bush administration is generally puttingincreased emphasis on measuring the effectiveness andproductivity of government programs. This seems to beparticularly true in the area of foreign assistance, bothfor USAID and World Bank programs, where there theemphasis is on increasing productivity in developingnations.

Thus, there is need to strengthen the analyticalcomponent where we can, and impact information ispart of this process. What is being done by the centersis a vital component. But while necessary, it is notsufficient. More attention needs to be given by thecenters and the Science Council Secretariat to preparingreader-friendly summaries of impact studies. Thiscould be done in two ways at the System level: (1) aseries of fairly standardized summary briefs, much likethe IFPRI 2020 Vision Briefs just noted (which might beextended to electronic form), and (2) have someoneperiodically engaged to do an interpretative summaryof the evaluations. There is a need to close theinformation gap between the analysts and the donors.


12

Establishing Plausibility in Impact Assessment

This paper discusses findings, of the “Workgroup OnAssessing the Impact of Agricultural Research inDevelopment” (part of the Deutsche Gesellschaft fürTechnische Zusammenarbeit - GTZ), based on thedetailed analysis of an impact assessment exerciseconducted by the IITA1.

The workgroup contends that traditional approaches toimpact assessment tend to predispose impactevaluators to establish unrealistic and unjustifiablerelationships between development interventions andobserved impacts. Impact evaluators often try to proveor quantify impact on a highly aggregated level andgloss over the fact that innumerable factors play rolesthat are often un-quantifiable but even more significantthan the intervention itself. In most cases, this leads toattribution gaps. Attribution gaps are caused by theinteractions of numerous and significant variables thatare external to the experimental design. The effects ofexternal variables make it impossible to isolate theeffect(s) of a single development intervention, which,ultimately, leads to an over-interpretation of data.

The workgroup determined three fundamental goals ofimpact assessment:

1. Learning about more and less successful approachesto development and poverty reduction.

2. Steering projects, programs, and strategies, withintheir given dynamic settings, to maximizeeffectiveness and sustainability.

3. Improving accountability for investments indevelopment cooperation by trying to ensure thatthey truly effect changes in the lives of people,especially the poor.

To achieve the above goals and establish plausibility,the workgroup proposed impact assessment guidelinesand standards, namely:

• Identification of the source of the impact beinginvestigated.

• Presentation of the model or concept of impact usedby the impact evaluators and how it applies to thecase at hand.

• Statement of the objectives and limitations of theimpact assessment.

• Outline of specific theory of action on which theintervention or strategy has been based.

• Statement of the impact hypotheses that the impactassessment tests.

• Presentation of other factors that could have affectedthe observed changes and alternative impacthypotheses.

• Discussion of other informed opinions that supportand contest the study findings.

In conclusion, the paper emphasizes that establishingplausible links between development interventions andobserved impacts is the central task of impactassessment and that honesty, openness, andtransparency should be its cornerstones.

3. GTZ’s contribution

Stephan Krall (GTZ)

1 “Social impact of soybean in Nigeria’s southern Guinea savanna” by P.C. Sanginga, A.A. Adesina, V.M.Manyong, O. Otite, and K.E. Dashiell.


13

Future Harvest: Science for Food, theEnvironment, and the World’s Poor

Future Harvest was initiated four years ago to raiseawareness and support for international agriculturalresearch. We are a global public charity registered in theUnited States and the United Kingdom. Our principalfunding comes from the 16 food and environmentalresearch centers funded through the ConsultativeGroup on International Agricultural Research. We alsoreceive funding from foundations, governments, andprivate individuals. During our early history, we havebeen developing Future Harvest’s image, position,communication strategy, and approach to fund raising.

Future Harvest is issue-oriented, not focused onpromoting agricultural research per se. Rather wepromote the benefits of international agriculturalresearch to peace, the environment, economic growth,human health, and population—issues high on thepolicymaker’s and the public’s agenda.

Future Harvest commissions research on the linksbetween agriculture and these issues with organizationsoutside the agricultural community. We have sponsoredtwo studies: one on the links between agriculture andpeace with the International Peace Research Institute ofOslo, Norway, and one on the links between agricultureand wild biodiversity conservation with the WorldConservation Union (IUCN). We have another studyunder way with CARE.

Future Harvest has a cadre of international publicfigures who serve as ambassadors, including OscarArias, Queen Noor, Norman Borlaug, and DesmondTutu, among others.

Future Harvest promotes the studies it commissionsand the work of the 16 Future Harvest Centers. Newsand feature releases have appeared in many languagesin media outlets throughout the world. These include

the Economist, the Financial Times, the Hindu, the JakartaPost, the BBC, CNN International, the Asian Times, theUgandan New Vision, and the Wall Street Journal. FutureHarvest has built a cadre of journalists in theagricultural and science fields, but has augmented itthrough the development of contacts that cover theenvironment, business, health, and generaldevelopment. Through these “nontraditional”journalists, Future Harvest is expanding theunderstanding and coverage of linkages betweenagriculture and broader issues.

Future Harvest principally uses the Internet todisseminate its materials. Since the launch of itswebsite almost three years ago, some 6,000 individualsfrom a wide range of backgrounds have voluntarilysigned on to our email list at www.futureharvest.org toreceive regular updates.

The Centers assess the impact of their work for anumber of reasons: to determine if their science is ontarget, to see what results they have achieved, and toprovide donors and other interested audiences withinformation they need to justify investment. BecauseFuture Harvest works closely with the centers, I willrefer to “we” in my comments.

While scientists and donors have been discussed astargets for impact assessment research, there is anotheraudience that has not been addressed: thoseindividuals who pay the donor’s bill, the taxpayers.International taxpayers need to be part of our targetaudience for two reasons. One, we have an obligationto tell them how we are spending their money and thattheir tax dollars are making a difference. Two, theywant to know. Future Harvest is starting to positionitself and the centers to reach this important audience.

New science deserves new communications. In hispresentation yesterday, Tim Reeves noted four pointsabout the future of science.

Discussion Session 2: Media Perspectives

1. Barbara Alison Rose (Future Harvest, Washington, DC, USA)


14 Discusson Session Presentations

First, he said that the “Centers need to do morescience and to do it differently.” This is also trueabout how we communicate this new science. Weneed to do more of it and to do it differently.

Second, he said that “science is more complex andthat it will become more so.” This is true, but wehave to explain this science simply and sharply inorder to be heard over the many voices that arecompeting with us for attention.

Third, Tim said “the nature of the scientific tools weare using for research today are more advanced.”This is also true for our outreach methods. The toolsare more advanced. The Internet, including theWorldwide Web and email, and CD Rom and DVDtechnology have changed our outreach modes.

Fourth, Tim noted that the research undertaken bythe centers involves more partners and alliances inresearch efforts and that this often causes attributiondifficulties when reporting final results. Expandingalliances is a good development, one that we need tocommunicate as well. To do so, we need a unifiedframework, something that Future Harvest isworking to provide.

What we say is critically important, and it is alsochanging. Clearly, our message is about:• the environment, human health, political stability,

and so on; it is no longer just about a bigger pile offood;

• the critical role of agriculture in economic growthand poverty alleviation, and about framing theproblem and highlighting solutions;

• the complex relationships that define the research,what I call the three “Gs,” governance, gender,and genes;

• the people—particularly women, children, andfamilies who have better lives as a result of whatcenter science does. It’s about creating emotionalresonance as well as intellectual significance;

• long-term solutions as an alternative to short-termfixes.

These messages need to be delivered directly, withwell-explained technical concepts, and with the publicpersona well identified.

New Partnerships and Forward Focus

The issue of attribution has been discussed throughoutthis meeting. Some of the discussions have suggestedthat increasing the number of partnerships, andtherefore partners that share the credit for researchresults, can be a problem. I would like to suggest that itis an asset.

The journals Science and Nature have a strong followingamong the scientific and policy communities. Many ofthe main articles cover research in the biologicalsciences and are often authored by ten or morescientists. In many fields, this is not an uncommonpractice. There is no reason why center scientistscannot take up this example and share scientific creditmore broadly. In the past, agricultural research hasbeen viewed as too insular. We need change that image.Our new collaborative arrangements will help makethat happen.

We also need to focus on the future. There is apropensity, clearly in the area of impact assessment, tolook backward and to congratulate ourselves for thegood work of the past. We need to ask how can ourimpact assessments provide insights for the future.How can we better learn from the past to frame ourapproaches in the future? Our audiences want to knowwhat the critical issues of the 21st Century are and howwe are going to help solve them? Let me add that manyin the new generation of leaders, public officials, andgeneral public were born after the Green Revolution.They do not know the past. They want to know whatare you going to do for me next year and in the years tocome? What do our impact assessments say aboutwhere we are going?

15

What Can You Do?

I would like to end with some practical ideas aboutwhat center staff, and this may apply to others in theaudience as well, can do.

In his comments yesterday, Prabhu Pingali mentionedthat economists often feel that impact assessmentresearch is too important to allow the involvement ofcolleagues from other disciplines. He stressed the needfor more involvement of scientists from otherdisciplines. I agree and suggest that involvement needsto be as broad as possible, involving other FutureHarvest Centers and national research systems. It alsoshould include communications specialists who candetermine the type of information that will be mostuseful for message development.

Donors do not want the same kind of information.Currently, very few donors understand or use complexstudies, and factors other than impact data, such aspolitics and institutional needs, affect how they makefunding decisions. We need to find out what donorswant and how they operate. This means involving yourdonor relations staff when developing assessmentpriorities. They can help guide the direction of impactassessment work with an eye to donor needs.

We need to stop talking to ourselves and to starttalking to others outside of the agricultural researchcommunity. This will require that center scientistsreceive media training so they can better articulate thebenefits of research.

Although the research efforts of the future will includemore partners at many levels, we need to speak withone voice.

At every stage of the assessment and communicationsprocess, we need to ask how is what we are doinghelping the poor? How has it, or how will it, maketheir lives better? Thank you for inviting me toparticipate in this session.


16

Forging an alliance to win, win,win all the way

Respected Chairman and dear friends, let me thankthe organizers of this international conference forinviting me to participate in the session on MediaPerspectives.

My presentation may not be as well-structured as youwould expect. Quite often, representatives of themedia are not good public speakers. Public speakingis an art, while speech writing is a science. I belong tothe category of silent communication scientists,working behind the scenes. I would be morecomfortable in an interactive session than making apresentation like this one.

We all know that public awareness is critical forpublic-funded agricultural research anddevelopment. The public has the fundamental rightto know, and the policy-makers need to beenlightened to make well-informed investmentdecisions. Researchers are accountable to the public.The media is bound, by responsibility, to mouldpublic opinion and inform policy-makers and donorsin order to justify their decisions in supportingresearch programs.

Quite often, the role of media is either ignored orunder—rated. There seems to be some reluctanceamong many scientists to deal with the media. Theypresume the media is not relevant in publicizing theirwork. According to them, time spent with the mediais wasted. Some scientists are wary of journalists andavoid them like a plague. Scientists are quite contentto publish peer-reviewed articles in professionaljournals. Media stories are of no consequence tothem. However, there are exceptions to this generalobservation, and there are exceptionally gifted

media-friendly scientists, who provide us with “goodstories.” Our wish is that scientists should shed theirinhibitions and reach out to the media. I am sure themedia will only be too willing to join hands to ensurethat research findings reach the target audience.

Scientists should also make a distinction between thelay media and the science media. It is quite unfortunatethat, in the general media, inaccuracies creep in whenreporting agricultural research and development. Thislargely stems from the lack of understanding on thepart of the media, and a serious communication gapbetween the scientists and the media representatives.Scientists can help to avoid such problems by takingtime to explain complex scientific findings in simpleterms. Again, the communication gap could lead toadverse and negative stories. There have been manyexamples of such misinformed stories that caused greatdamage to the image of organizations such as ICRISATand IRRI in the mid-1980s. Such damage could havebeen easily averted, had these institutes adopted pro-active public awareness strategies.

The science media is different from the lay media. Thescience media is well trained and geared to accuratelyreport authentic advances. Agricultural research anddevelopments generally do not make headline stories;but they are excellent material for specialized columns,development features, and opinion-editorial pagearticles. The agricultural writers and mediaprofessionals are a step ahead of the general media ingetting their stories right. In addition to the usualjournalistic “mantra” of what? when? where? why?and how?, these professionals have their own formulaof what? and how is it done? how is it possible? howcan it help?, and how can one get such help?.Generally, the media relies heavily on media-friendlyand articulate scientists with good public relationsskills to get a good story supported with credibleinformation and authentic data.

2. G. Venkataramani (The Hindu Newspaper, Chennai, India)


17

I wish to share our experience as public awarenessconsultants at ICRISAT a few years ago. My wifeGemmarie, also a media specialist, and I produced atleast four issues of a glossy publication called “FoodFrom Thought.” This series was a brainchild of Dr. JimRyan, when he was the Director-General of ICRISAT.“Food from Thought” was based on a series of impactassessment studies initiated by Dr. Ryan. They areessentially 6 -to- 8 page executive-summary-likereports with lively quotes from farmers, NGOs, theindustry and other partners of ICRISAT research. Theywere uniformly well received by donors, media andpolicy-makers. These publications contained all thedesirable elements of a media feature, and were backedby intensive, peer-reviewed impact assessment studies.Interestingly, they turned out to be pithy human-anglestories any editor would only be too happy to publish.

I wish to mention here that, some of these publicationshelped ICRISAT to win the prestigious King BaudouinAward for two successive years. Scientists whocontributed to the publication tasted some media glare,and above all they had the sense of satisfaction ofeffectively communicating their research findings tothe end users (farmers and consumers). The societygratefully recognized their contributions, andrespected them.

Yesterday, Dr. Prabhu Pingali referred to the hard factof life that social scientists are under peer pressure,compelled to publish in reputed professional journalsand, while doing so, remain in a closed society. Thepublic awareness group is just one step behind thesocial scientists. The impact assessment group will dowell in forging a meaningful triple-alliance with them,the media and the donors in order to sustain theimpact assessment research efforts and the interest ofdonors. Social scientists are the best-trainedprofessionals who can provide the food for publicawareness. Be re-assured that you are the ones whocould build effective bridges between the scientists andother partners in the development process.

I have had the privilege of reporting on several impactassessment studies conducted by ICRISAT, IFPRI andseveral farmer-oriented success stories of IRRI since1979. I have written, and still write, about ICRISAT’sadopted village, Thadanapalle. I am sure that there areseveral good stories in the other CGIAR centers aswell. I must mention that the CGIAR has severalunique strengths, but they have to use them to theiradvantage. Research for public good, not for profit,and a transparent agenda are its strengths. Altruismand your role as the custodian of the largest collectionof germplasm of major crops add to your strength. TheCGIAR centers should take a pro-active initiative inreaching the media through concerted publicawareness exercises. The media have no reservationsin joining hands with these centers in disseminatingthe research findings for the benefit of the farmers,consumers and policy makers. We look forward tomore productive and mutually reinforcing media-related initiatives from the CGIAR centers.

Today, we are in a super-dynamic world. Things arechanging so fast. We are in an era of informationexplosion. New science and technologies, pointed outyesterday by Professor Tim Reeves, provide newopportunities. We know that a traditional approach isno longer adequate. Business-as-usual-approaches donot take us anywhere. We need innovative initiativesto build more partnerships and alliances to addresscomplex research and development challenges. Wehave to switch to consensus-building modes creating asituation where all the partners are winners. No onewants to be a loser. Let us make a quick SWOTanalysis of all the partners, and use our strengths toconvert the weaknesses and threats into newopportunities and create a situation of win, win, winall the way in our relentless battle against poverty,hunger and malnourishment.


18

I want to start with a saying of Richard Feynman’s:“Anything is interesting if you look at it closelyenough.”

Well, that’s nice and optimistic. Your first problem asscientists trying to get your stories into the press iswhat “interesting” means. I think Feynman could havemeant “interesting” in two ways. The first kind ofinteresting is that there is something unexplained,something mysterious about it. That’s what makes itinteresting to a scientist. The second kind is that thereis something new, or unexpected, or counter-intuitive.That makes it interesting to the general public. So yourfirst problem for making the story is to pull out of yourresearcher’s mindset and think like a non-scientist.

The second problem is that Richard Feynman was agreat physicist and a brilliant thinker, but he neverworked on a newspaper. My definition of a newspaperis this: a place where nobody looks at anything closelyenough. When I worked on the science desk of TheEconomist, each press release was looked at for anaverage of about 5 seconds. (That’s my unscientificestimate.) So, there is a third kind of interesting, whatmakes a story interesting to news editors. And that iswhat can catch their attention in five seconds with anangle that will interest their readers.

Let me give you a quick rundown of the average life ofa press release in a newspaper. It comes in and, if itdoesn’t get lost on the floor in a pile of papers, it will bescanned quickly. Whether or not it gets transformedinto an article depends on whether or not the paper hasrecently covered that general theme, whether thecompetition have done it, whether there is an anglethat will interest the readers, whether there is room forit among the other articles, and whether the journalistin question has time or is too busy writing other storiesor taking his kids to school, or whatever. So the storyhas to jump quite a few hurdles.

Now we come to the third problem.

When I was a science journalist, I covered physics,chemistry, materials, astronomy, space exploration and,new technology. When most people see words like thatthey automatically assume that they know nothingabout the subject at all, perhaps because they think thatthose areas don’t overlap with their daily lives. So, myjob was easy. Practically anything you tell somebodyabout those subjects will surprise them. “The electronsgo around the... really? Wow... the galaxy containsbillions of stars... no, seriously?”

The problem is that people think they know whatagriculture is. “Oh, yeah, we’ve been doing the samething for 12,000 years.” They don’t associate it withtechnology. And if they do, they automatically thinkgenetic modification. Basically, I’m saying that whenpeople hear “agricultural research” they either think“Frankenfood” or they think “rain and mud.” And asfor impact studies, well, those aren’t even agricultureper se, they’re numbers about agriculture. For thegeneral public, new means a new kind of wheat thatkills bugs by itself; new means a tomato that stays freshfor three times as long; new means a corn breed thatyields five times as much. But the discovery that thisnew corn breed reduces farmer’s costs by exactly, say,35% is not really big news.

That’s the prejudice of the general reader, and newseditors are at least as ignorant as their readers, or moreso—that’s what qualifies them to be news editors.Unfortunately, even specialized science editors share alot of the same prejudice. This is why you sometimeshave to rely on names like “Super wheat” or “Goldenrice,” even if not everyone in the research communitylikes them. People do catch on to them. In English, theprocess of making science accessible to the public iscalled “popularization.” In French they understandbetter: they call it “vulgarization.” Remember how wellthe church did when it switched from Latin to vulgateworship? Same thing.

3. Gideon Litchfield (The Economist, Mexico City, Mexico)


19

The result of all this is that the journalists who coveragricultural research stories aren’t necessarily the sameones who cover other kinds of science stories. Itdepends on what the news editor thinks the angle is. Itcould be the science reporter (if there is one), or anenvironment, economics, business or farmingcorrespondent. Each will pick a different focus.

Which means that it’s hard to get agricultural researchstories covered in the press. But you knew that already.Let me give you a more concrete illustration of justhow hard it is.

One of the many benefits of working for The Economistis that I have access to an online press databasethrough our website that covers more than 7,000periodicals, both specialist and general. Since this is aconference about impact assessment, I did my ownimpact assessment. This is the impact of your researchon the world’s media. Are you ready?

The first thing this tells me is that most of your stuffcomes up in the specialist journals. I found 185references to CIMMYT over the last five years, and ofthose 115 were in the journal Crop Science. Most of therest were in fairly specialized scientific and businessjournals too, hardly any in the general press.

Now comes the really interesting data. Let’s look at thetop stories in agricultural research, the ones that TimReeves mentioned in yesterday’s opening speech.

Drought-tolerant maize, the most excitingbreakthrough Tim has seen. About 150 stories over fiveyears — I searched for both drought-tolerant anddrought-resistant, and maize and corn, and that’sabout the number. I’d say that at least half of themwere in specialized or semi-specialized journals thatvery few people read. Remember, too, these are storiesmentioning drought-tolerant maize; they’re notnecessarily mainly about it.

“The story that everyone knows about,” Golden rice.Well, Tim, I’m sorry, but I don’t think everybody does.Over 300 mentions altogether. It got decent coveragewhen the Science paper about it was published inJanuary 2000 — a couple of dozen stories. But the

roughly half of those used the same report, from theAssociated Press news agency. I’ll have a bit more tosay about the news agencies later. That means perhapsa dozen original pieces were written about the storythat everyone knows about.

The World Food Prize, the equivalent of the NobelPrize for food research. Quality protein maize, whichtook 30 years develop, got the prize in February 2000.Now of course QPM had been written about before,but for most people the first they hear about a scientificdiscovery is when it wins a major award. OK, inFebruary 2000, there were just 13 stories about theWorld Food Prize, and none of them appeared in amajor newspaper; there was one piece in the HackensackRecord of New Jersey. Some of the news agenciescovered it but it seems like nobody took the story up.

By comparison, let’s look at the Nobel Physics Prizeawarded in October 2001. I chose this one rather than2000 just to make it a really brutal comparison becausethe prize in 2001 was for Bose-Einstein condensates. Asa physicist I can tell you that Bose-Einstein condensatesare very clever and very weird but so far, unlike qualityprotein maize, they haven’t done anybody any good.Well, 42 stories in October 2001, including some of thebiggest international papers. The Economist makes apoint of covering the Nobel Prizes every year. It doesn’tcover the World Food Prize. I don’t think anybodydoes. Why is that?

So the problem with getting your stories in the press, tosum up, is that news editors think they’re dull. Thebest way to get them in is to think of how they affectthe readership.

To do that, I want to give you some illustration of howdifferent media cover the same story. I’ve taken onethat got quite a lot of play last year; it was a study oforganic apple farms in the state of Washington,reported in the journal Nature. It compared organic,conventional and hybrid farming methods. The basicconclusions were that after six years, the three methodscame out equal on horticultural performance, but theorganic farms did better in soil quality, environmentalimpact, energy efficiency and the taste of the fruit.


20

Finally, let’s take a look at the agencies, or wireservices. Most local and regional papers use these forinternational news, science news and anything elsethey don’t have specialized staff to cover, so what theysay can have a huge impact. Remember that most ofthe coverage of Golden Rice in the general press camefrom Associated Press.

The Xinhua news agency has a short and very balancedreport; one of the few reports to make a point of thefact that the results may not apply to apple productionoutside Washington State. This was also recognized inthe Associated Press. Oddly, though, the United PressInternational’s Farming Today roundup simply clippedthe Los Angeles Times story, though it included theskepticism.

So, the conclusion from all this is that, if you want astory placed in the press, you have to think about:

• Which media you want it to appear in?• What their readership is?• What that means their likely angle is?• Why would your story interests them, remembering

that people instantly associate your research witheither Frankenfood or mud and wet weather?

•· Which journalist is likely to be assigned to it andwhat’s his or her speciality?

• What’s the best way to approach them: a pressrelease, or a direct contact?

How was it reported? Pretty much, as you wouldexpect. Journals with a special interest in organicfarming like the Natural Foods Merchandiser describedthe science carefully but had quotes from people in theorganic-food movement portraying it as good news.

Mother Earth News also said it was good news and toldits readers the best way to care for their organic apples.

The Independent of London had a strong pro-organicfarming line, and brought in the Friends of the Earthview. The paper clearly wanted the issue put on thegovernment’s policy agenda and used this report as away to push that.

The Economist, as you might expect, took an economicangle, even though this story appeared in the sciencesection. It described the science carefully but went intosome detail on the potential profitability of otherorganic crops, even though there was no furtherevidence about it.

The Seattle Times is the paper you would expect to givethe best coverage, because the story was in its ownback yard. However, it used a report from the LosAngeles Times, which presumably assigned the story toits Northwestern correspondent. Quite possibly theSeattle Times didn’t have a science journalist on thestaff. Naturally, the report focuses on the likely effectsfor the local apple-growing industry, though it’s alsomore balanced about the benefits.


21

Discussion Session 3: What Next for Impact Assessment?

1. Hans Gregersen and Tim Kelley

stage and at external fund raising stage, inaccountability needs vs. planning needs);

• Using such interactions as an opportunity to learnnot only about user needs, but also to informdecision makers (including investors in the CGIAR)about what is possible with different levels ofresources and different time constraints. We have toend up with realistic expectations for IA. TheCGIAR may not always be able to solve poverty andenvironmental degradation problems directly, butwe can develop technologies, inform policies andhelp build formal and informal institutions that can.As IA is often expensive, resources available need tomatch the increased demands for IA; often increasedsustainable resources do not accompany increaseddemands for IA results.

• Going beyond a cost-benefit framework to providericher information on the factors influencing impactsby spelling out and testing programs underlyingtheories and assumptions. This may entailbroadening the disciplinary base beyondagricultural economics and, when appropriate,using mixed-method approaches. The IFPRI-led‘impacts on poverty project’ is a good model.

• Remembering that major user groups consist notonly of the funders of research (an accountabilityfunction mainly), but also of planners and decisionmakers who are shaping future programs; i.e., theywant ex-post and ex-ante IAs as input in decision-making about such programs.

• Effectively publicizing and disseminating results ofIA. At the system level, there is a need to bringtogether the disparate information on IA beinggenerated by the centers. This could entail thedevelopment of a website “IA in the CGIAR,” whichcould serve as: (a) a central repository of credibleimpact information (peer reviewed plus others); (b)a channel for exchange of information, and; (c) ameans of reporting results to users, and otherfunctions as needed.

• Recognizing that, in many situations, attributionmay not be important. The more effective a researchpartnership is, the less desirable and feasible it is to

Where Do We Go From Here?

This conference has produced a number of insights forguiding the future direction of impact assessment (IA)research. We have heard about a variety ofmethodologies and analytical tools, both quantitativeand qualitative, which have been applied across a widerange of research areas, including some of the hard tomeasure areas like Natural Resource Management(NRM), policy and training. Some interaction has takenplace between those who use IA results and those whoproduce them. However, much more dialogue is inorder to better match what is needed and wanted bydecision makers with what reasonably can be producedby analysts and researchers, given data, resource andtime constraints.

The question addressed in this last session is: “Whatnext for impact assessment?”

From the interim Science Council (iSC) perspective,which is consistent with the main threads of thediscussions taking place at this conference, there arethree main areas in which progress is possible and inwhich the iSC would like to get involved:1. matching IA outputs to decision makers’ priority

needs;2. making IA outputs more credible, plausible and

understandable, without losing rigor in theprocess, and;

3. improving methods, particularly in terms ofdeveloping a set of impact indicators for a broaderarray of impacts beyond the traditional economicones.

In terms of each of these three areas, there exist thefollowing needs and opportunities.

Make IAs more relevant to decision makers needs, by:

• Interacting more frequently with the various users ofIA outputs, recognizing that their needs are varied,even for a given user (e.g., at internal allocation


22

Improve IA methods and develop new more integrativeapproaches to consider a wider set of impact indicators moreclosely linked to CGIAR goals by:

• Formulating a set of principles and strategicguidelines for future ex-post IA in the CGIAR,highlighting good, credible studies as models tofollow, and working towards “best practices,”addressing key issues such as plausibility, impactpathways, use and synthesis of evidence acrossmultiple sources and causality/linkages issues.

• Recognizing the many, mostly complementary,approaches that are used by different analysts whencarrying out IA and, hence, the opportunities formore interchange of ideas, methods and results.Mechanisms are needed to facilitate the exchange ofinformation. A web site can be one of thosemechanisms, but not the only one.

• Devoting more resources to developingmethodologies / procedures for:� Multidisciplinary IA based on a problem driven

approach,� Up-scaling and synthesizing (of cases studies,

smaller studies),� Rapid, low-cost data collection, with acceptable

levels of accuracy,� Modeling adequate counter-factual estimates,

and,� IA and evaluation methods for capacity building,

NRM and policy research that have provenelusive in the application of existing IA methods.

Many of these needs and opportunities were alsoidentified at the 2000 SPIA workshop on IA in theCGIAR2. Some progress has taken place in many ofthese areas since 2000, yet much remains to be done.The Interim Science Council welcomes comments andsuggestions from conference participants on which ofthese needs and opportunities are no longer prioritiestoday, and what additional conclusions andrecommendations flow out of the current conferencethat should be added to the list? Such suggestionswould be helpful as the Council and its CGIAR centercolleagues move ahead in formulating a program forthe future.

attempt to attribute impacts separately to eachpartner. It can be counterproductive and eventhreaten good working relations within thepartnership. This point needs to be understoodparticularly by funders and decision-makers whopromote partnership as a means of making researchmore effective and efficient.

•· Not limiting IA to success stories. Honest attributionof project shortcomings as well as benefits isrequired; recognition of risk and uncertaintyassociated with successes must be transparent inorder to gain confidence of those who use IAs. Thisrelates to sampling of case studies; i.e., purposive(cherry-picking) vs. random (including pickinglemons).

Make IAs more effective, plausible, credible andunderstandable by:

• Establishing credible counterfactuals fortransparency and plausibility purposes (e.g.,validate assessment using the double deltaapproach), especially critical for external audiences.Related to this is the importance of establishing andmaintaining baseline data on the livelihoods of thepoor.

• Based on the expressed needs of donors, producingand disseminating attractive, reader-friendly, shortCGIAR IA briefs based on peer reviewed studies.

• Not confusing IA with program performanceevaluation. These quite distinct functions belong indifferent places. Different types of assessment andevaluation have separate functions and should beexecuted by different actors.

• Fully institutionalizing IA as a management function(e.g. for priority setting, resource allocation,feedback to program planning). In many researchorganizations it is often carried out in response toexternal demands rather than as an integral part ofplanning, and for drawing out lessons and derivingimplications.

• Developing improved methods of IA throughcollaboration/division of labor among thoseinvolved in impact assessment. A more effectiveprocess for organizational collaboration is needed.

2 TAC. 2001. The future of impact assessment in the CGIAR: Needs, constraints and options. Proceedings of aworkshop, 3-5 May, 2000. Rome: FAO, TAC Secretariat.


23

Abstracts of conference papers

To obtain copies of the papers associated with these abstracts, contact either thejournals listed in the section “Conference papers published to date” or contact the authors directly,

using the information in the section “List of participants”

Milestones in Impact Assessment Research in the CGIAR, 1970-1999

Prabhu L. Pingali

applied the methodology developed in the study. In allcases, the milestone contributions were published inrefereed journal articles, books or official, peerreviewed center publications (not working papers andother non-refereed documents).

This review documents and highlights the importantmilestones in impact assessment research that can beattributed to current and past CG economists andsocial scientists. While the scope of impact assessmentresearch in the CGIAR has expanded over time,beyond production impacts, there continues to be aneed for and a pressure to further broaden the agendaof impact research. Evidence that modern technologyhas contributed to poverty alleviation is particularlysought after by the donor community. Similarly, studieson the impacts of agricultural modernization on thesustainable management of agricultural eco-systems, ingeneral, and biodiversity in particular, are also indemand. CG economists have also begun measuringthe impacts of policy research and advocacy, trainingand human capacity strengthening, and networks fortechnology generation and exchange. The impactassessment community faces enormous methodologicalchallenges as it gears up to measure impacts in theabove areas, but these are also the areas where futuremilestones can be achieved.

The Consultative Group for International AgriculturalResearch (CGIAR) has probably had a greater impacton agricultural production, productivity and thelivelihoods of the rural poor in the world than anyother agricultural research organization. This paperdoes not intend to document the substantial impacts ofthe CGIAR (subsequently referred to as CG) but ratherto dwell on the stock and state of impact assessmentresearch conducted by CG scientists, especiallyeconomists. What is not so widely realized is that overthe past three decades, contributions by CGeconomists and other scientists to the science of impactassessment have, in many cases, been path-breaking,and have created milestones in the ever increasingbody of technical literature on impact assessmenttheory and methods. In numerous instances, theCGIAR was a forerunner to a substantial body ofacademic research literature on particular themesrelated to impact assessment.

For the purposes of this paper, a milestone is definedas a research contribution that identified a new area ortheme of impact assessment research; it could alsoinclude methodological contributions. In all cases, theseminal contributions (milestones) were quicklyfollowed by other studies, both from within andoutside the CGIAR, that verified the findings or

24

The Green Revolution: An End-of-Century Perspective

Robert Evenson and Douglas Gollin

From a methodological standpoint, the study hasunderscored some of the difficulties inherent in carryingout impact assessment. Too often, data on productivitygains are non-existent or are simply inadequate forderiving convincing conclusions. Traditional approachesto measuring impact often overstate some benefits (suchas the added value of production within a crop) whileunderstating other benefits (such as spillovers into themacro economy). There is little nationally representativeevidence, and available data are often incompatible withaggregate measures of production such as thosereported by FAO.

In an effort to move beyond the traditional single-cropimpact estimates, the study included two novel features:first, a set of country studies, and second, a series of“meta-analyses” that sought to synthesize data frommany crops and regions. Both the country studies andthe meta-analyses involved methodological innovations.This paper will report on the results of these analyses,summarize the approaches taken, and briefly describethe methodological innovations required.

This paper reports on the empirical andmethodological lessons to be drawn from a multi-yearstudy of the productivity impact of crop researchwithin the CGIAR. This study was commissioned bythe Standing Panel on Impact Assessment of theTechnical Advisory Committee (SPIA -TAC) of theCGIAR. The overall goal was to document the impactof international research on crop genetic improvementin developing countries. The study focused on elevenmajor food crops: rice, wheat, maize, sorghum, millet,barley, beans, lentils, groundnuts, cassava, and potato.Eight CGIAR centers participated in the study, whichdrew together an unprecedented amount ofpreviously unpublished data.

From an empirical standpoint, the major lesson of thestudy was that productivity impacts have been largeand have occurred across crops and regions. Contraryto popular opinion, the so-called Green Revolutiondoes not appear to have been limited to rice andwheat in Asia and Latin America; instead, there isconvincing evidence of productivity gains in all majorcrops and all regions. For the most part, impacts havebeen related to research inputs. In general, whereimpact has been small, we can attribute this to modestresearch efforts or comparatively recent research.Where research efforts have been large and sustained,there have been demonstrable impacts.


25

Under-investment in Agricultural R&D Revisited

Johannes Roseboom

behavior (i.e., selecting the R&D project with thehighest ERR first), the under-investment gap can bedefined in strict economic terms as the differencebetween the ERR of the marginal R&D project (theactual cut-off rate) and the social rate.

Important findings of the study are:� Not the mean but the mode of the ex-post rate-of-

return distribution is the relevant variable forassessing under-investment in agricultural R&D.

� Under the assumption of full information andeconomic rationality, developed countries couldhave invested about 40% more in public agriculturalR&D and developing countries about 137% more. Interms of agricultural R&D intensity (i.e.,expenditures as a percentage of agricultural GDP),developed countries could have invested 2.8%rather than 2.0%, and developing countries 1.0%rather than 0.4% in the period 1981-85.

� Low investment in public agricultural R&D indeveloping countries is caused first and foremost bya relatively smaller portfolio of profitable R&Dprojects to invest in. Under-investment certainlyplays a role (the gap is bigger for developingcountries), but it explains only a small part of thedifference in agricultural R&D intensity betweendeveloped and developing countries.

� While efforts to reduce the under-investment gapshould continue (e.g., better information andpriority setting), more emphasis should be placedon designing policies that help to shift the portfolioof R&D projects higher up on the ERR scale, even atthe risk of increasing the under-investment gap.

Since the early work by Griliches on hybrid corn, rate-of-return studies have become standard practice indocumenting the economic impact of agriculturalresearch and development (R&D). Although theestimated ex-post rates vary quite substantially, theaverage tends to range in the order of 70-80%. Despitecriticisms on the accuracy of these rates, as well as howrepresentative the selected projects were, a widelyshared belief is that the estimated rates are robustenough to accommodate such criticisms and still be ina range that is substantially above the social rate. Basedon this evidence, Ruttan argued that there is seriousunder-investment in public agricultural R&D. Thisargument has become widely accepted amongagricultural economists and hence any slowdown orcontraction in the growth of public agricultural R&Dexpenditures is reason for serious concern.

What do we actually know about this under-investment? How real is it? Is under-investment greaterin developing than developed countries? Is it greaterfor some types of research than others? Has it increasedor decreased over time? And, how much more shouldhave been invested in agricultural R&D? In order toanswer these questions, the under-investmenthypothesis needs to be defined more clearly. This canbe done by means of the following simple model ofR&D project selection. The ex-ante distribution ofpossible R&D projects on an expected-rate-of-return(ERR) scale can be thought of as decliningasymptotically, which can be approximated by a semi-log function with a negative slope coefficient.Assuming full information and rational economic


26

Impact of Agricultural Research in Bangladesh: Productivity,Economic Returns, and Varietal Replacement Issues

Joseph G. Nagy and Md. Ferdous Alam

wheat (US$ 0.16 billion) and sugarcane (US$ 0.08billion) were more modest. Jute earned additionalforeign exchange of US$ 0.07 billion and (5) the rate ofvarietal adoption and timely replacement of oldervarieties is a cause for concern—the 1995-1999 meanaverage age of T. Aman rice varieties is 24.6 years,wheat 15.7 years, and sugarcane 18.9 years and thedominant variety in terms of area planted for each ofthese crops was released in the early 1980s.

The high rates of return, since 1980, of the commoditiesstudied are an indication that the Bangladesh researchand extension system has done reasonably well.However, there is no room for complacency. Because ofpoor adoption by farmers, many research projects andinterventions developed by the Bangladesh nationalagricultural research system have yet to show a returnon research and extension investment. This studydemonstrates that the rate of return on benefits fromthe aggregate of all research and extension benefits,relative to the total of all research and extensionexpenditures would be much lower for rice, wheat, andpotatoes. The immediate future economic returns toresearch and extension and future productivity gainsfrom varietal development and adoption is uncertain inlight of the low new varietal adoption rates and lowvarietal replacement rates of major commodities.Increasing research and extension funding is only partof the answer. To effectively utilize any increasedfunding, research and extension systems requirefurther institution building and strengthening, researchpriority setting, and increasing management andresearch capacity.

An impact study of the Bangladesh agriculturalresearch and extension system was undertaken forvarietal development interventions of rice, wheat,potatoes, sugarcane, and jute. The study focused oncontributions made by the research and extensionsystem to varietal development over the past ten totwenty years. The study looked at five impact areas:(1) trends in agricultural production and yield/hectare productivity growth; (2) percentageadoption of post-1980 varieties and incrementalproduction due to adoption; (3) economic rates ofreturn to research and extension expenditures onvarietal development, using the economic surplusmethodology; (4) foreign exchange earned or savedfrom new varietal development; and (5) the rate ofvarietal adoption and replacement.

The findings indicate: (1) that the growth inagricultural production and partial productivitygrowth rates (yield/ha) among the major cropsvaried; some crops exhibited highly positiveproduction and yield/ha growth rates (Boro rice,wheat, potatoes, jute) while some exhibitednegative growth rates (T. Aman and Aus rice, andsugarcane); (2) the highest adoption andincremental production rates were for rice, wheatand potatoes, while sugarcane and jute adoptionand incremental production rates were much lower;(3) there were high rates of return to research andextension for some major crops (rice, wheat,potatoes) but more modest returns for others(sugarcane, jute); (4) substantial foreign exchangesavings have been made for rice (US$ 3.8 billion in1997-98 prices), while foreign exchange savings for


27

Herding Cats: Is Impact Assessment theUltimate Exercise in Futility?

Patti Kristjanson and Philip Thornton

traditional top-down, uni-disciplinary approach toimpact assessment research that underlies its historicallack of effectiveness. We believe that it is the ongoingsea change in research approaches now beingundertaken within the CGIAR, coupled with recentadvances in technology (making high resolutionsatellite data cheaper and more accessible, forexample), that will in fact make impact assessment,including monitoring and evaluation activities, muchmore effective in the future.

In this paper, we look at some of the costs and benefitsof the systems approach we are advocating and themethods we have been using at ILRI for institutionaland donor priority setting as well as ex ante and expost assessments, in the hopes of shedding some lighton the following issues. First, what benefits do we seearising from our impact assessment research? Second,why has it not had more impact? Third, how can wemake it more effective without devoting anunserviceable level of capital expenditure in order todo so? It is hoped that our experiences at ILRI willbenefit our partners and other institutions grapplingwith similar issues.

The International Livestock Research Institute (ILRI)has invested a considerable amount of time, effort,and resources in a wide range of impact assessmentactivities in recent years. It is a good time to addressthe issue of how effective that effort has been. ILRIinitiated the adoption of a systems approach aroundsix years ago. However, it is only now that we havethe array of tools and databases needed to facilitatedeeper understanding of the various livestockproduction systems and examine impacts ofinterventions or changes within a particular system.Even so, each new assessment we undertakehighlights the insufficient resolution of critical data.However, as we develop a better understanding oflivestock systems, we are able to develop predictivemodels incorporating important driving andconditioning forces such as human populationgrowth, climate change, and policy changes.

Beyond endorsing a systems approach, we are alsowitnessing, and hopefully catalyzing, a shift towardsa multi-centered, multidisciplinary, and participatoryresearch approach based on strong teamwork andcollaborative databases. We propose that it is the


28

Resolving Conflicting Evidence About theImpact of the Green Revolution

Peter Hazell

The Green Revolution was launched in the late 1960s ata time when many developing countries faced analarming widening of their national food gaps andrapid population growth. Much of the initial focus wason growing more food, a tenable view at the time giventhe threat of famine. Green Revolution technologiesplayed a major role in increasing food supplies and inlowering food prices. They also increased farm incomesand generated powerful trickle down benefits in theform of additional income and employment in the non-farm economy. These impacts raised an enormous

number of poor people out of poverty and preventedmany more people from falling into poverty andhunger. Despite these successes, the Green Revolutionis widely criticized today for not having done more toeradicate poverty and food insecurity in thedeveloping world. This paper reviews the validity ofthis criticism, and asks whether Green Revolutiontechnologies might have been managed moreeffectively to achieve even greater poverty reduction.Important lessons are drawn for the design of futuretechnology transformations in developing countries.


29

Professional Error, Critical Awareness and Good Science

Robert Chambers

The history of development is littered with examples ofbeliefs which, though sincerely held by professionals inthe social and natural sciences, have later come to beseen as ill founded or wrong. Nine examples help toexplain the tendency for questionable and erroneousbeliefs and policies to be robustly resilient. Interactionsof power, interests and mindsets, and of behaviour andexperiences play a part in generating and maintaining

myth and error. Critical epistemological awareness tooffset and correct misleading influences of professional,institutional and personal interests and orientations isproposed for a more prominent role in good scienceand policy, and for enhancing the impact of impactevaluations. Questions for self-critical reflection areproposed. The reader is invited to improve on these.


30

Why has the Green Revolution Passed bySub-Saharan Africa?

Simeon Ehui, Samuel Benin, John Pender, and Mark Rosegrant

depressing interventions in food markets. Despite thevery high tax levels, there has been very littleinvestment of the surplus in rural public services andinfrastructure. Subsidies for fertilizer and credit usuallybenefited larger, export-oriented farmers who arecapable of exercising political power. Inefficient inputmarkets and weak property rights have been majorlimiting factors for long-term entrepreneurial planningby undermining both the will and ability of farmers toinvest. Unfortunately, poor domestic policies werereinforced by protectionist policies of the OECDcountries. In addition to poor policies, conflict has alsoinhibited growth. Many countries, and often entireregions, are ravaged by wars. Lack of vision or effectivegovernance by African leaders contributed to thedeleterious situation. Consequently, rural incomes percapita have declined with negative consequences onpoverty, food consumption and asset development. Theoutlook for the future is not bright; although totalconsumption of food will double by 2020, per capitaconsumption will only increase marginally. As withAsia, aggressive public investments in education,research and infrastructure (e.g. roads, irrigation) willbe needed to sustain growth. In addition, policies mustalso reach out directly to the poor, particularly throughinvestments in their human capital and health.Agricultural transformation in SSA will occur only ifthe countries in the region follow stable macroeconomicpolicies (market-friendly and open trade policies).

Agriculture accounts for 17% of Sub-Saharan Africa’s(SSA) gross domestic product (GDP), employs 67% ofthe total labor force and is the main source oflivelihood for the region’s poor. Therefore, agricultureremains important in rural SSA and indicators of ruralwell-being are closely related with agriculturalperformance. However, the Green Revolution thatswept through Asia over the past 30 years failed inAfrica. SSA’s agricultural performance during the1967-97 period was the worst in the developing world.Low productivity has seriously eroded thecompetitiveness of African agriculture in the worldmarket. For example, Africa’s share of total worldagricultural trade fell from 8 percent in 1965 to 3percent in 1996. The question asked by many,including policy makers, researchers, anddevelopment agencies, is why has agriculture in SSAperformed so poorly over the past three decades whileit has improved significantly in Asia, lifting manymillions of people out of poverty? In this paper, weargue that the root causes of the poor agriculturalperformance in SSA rests mainly with poordevelopment strategies and policy choices. Althoughadverse resource endowments have also played a role,the overall unwillingness of many leaders to recognizethe importance of agriculture to overall economicgrowth has been a major contributing factor. For manydecades, policies continued to heavily tax agriculturethrough over-valued exchange rates and price-


31

Measuring the Benefits of InternationalAgricultural Economics Research

David E. Schimmelpfennig and George W. Norton

groups that underpin policy decisions.The model builds on earlier work by Feeney andHillman, and considers how firms and consumers,who are also investors, influence government policiesthrough their lobbying efforts in the presence of AER.Strategic interactions between different governmentsin policy formation are captured in a non-cooperativegame through the ability of each government to usesubsidies to indirectly shift profits from foreign todomestic firms in imperfectly competitive markets athome. Risk-averse investors, who can share in theprofits of domestic or foreign firms and can act aslobbyists at home, are impacted by the policyoutcome through the return on their portfolioholdings. General scientific and other non-SSRreduces uncertainty in productivity and firm profits,subsequently shared with investors. SSR messagesreduce uncertainty in political-economic decision-making, but may not necessarily override the will oflobby groups, if the influence of those groups isgreater than the reduction in uncertainty.

The framework is empirically tested by implementinga Bayesian decision theory approach to updating theuncertainty in the model, and is applied to AER(research on producer and consumer subsidyequivalents) that supported international strategicinteractions during the Uruguay Round and otheroperations of the World Trade Organization. Itspotential role in evaluating research on UnitedNations agreements for the international preservationof biodiversity and the Kyoto Protocol for the controlof greenhouse gas emissions is briefly addressed.

Impact assessments of agricultural social scienceresearch (SSR), including agricultural economicsresearch (AER), have made relatively littledifference, probably because there have been fewSSR assessments. The reason for so few assessmentsis that the output of SSR is difficult to measure, isaimed at diverse objectives, and is often embeddedin recommendations, institutional changes, orquantitative methods. Causality between AER andspecific decisions or institutional changes is difficultto establish because the research information isoften just one of many inputs into a politicalprocess; and political decisions may influence AER,reversing the causality.

Quantitative evaluation of SSR (or AER) ischallenging, but can potentially be achievedthrough casting the problem in a political-economicframework, and addressing the causality issue byusing a method that captures the role of SSR inhelping decision makers update their priorprobabilities for specific events. The economic valueof AER arises primarily from its effect on economicefficiency through reductions in uncertainty aboutthe optimal (or a preferred) way to allocateresources, or the optimal (preferred) design of apolicy or institution. Therefore a framework forevaluating the benefits of AER can be built thatexplicitly models (a) the impacts of AER onpolicymakers and private decision-makers’subjective beliefs about the consequences of actions,(b) the economic efficiency or welfare effects ofthose actions, and (c) the political-economicinteractions between policymakers and interest


32

Evaluating the Impact of Economic Policy Research:Concepts, Practices and Lessons

Jim Ryan

Evaluating the impact of economic policy researchremains a challenging assignment, with few “bestpractices” to draw on. Case studies remain thepreferred way forward in the absence of an agreed-upon impact evaluation paradigm. The major problemsare those of measurement, sampling and attribution.The approach to impact evaluation is conditioned, to asignificant extent, by the primary purpose thatunderpins such studies. For a research institution, thisis to improve accountability and credibility; quality andrelevance; program/project design andimplementation; future planning and prioritizing.However, depending on the primary purpose, the unitof analysis and approach tend to differ.

The importance of understanding policy processeswhen undertaking impact evaluation is discussed in thepaper. This includes the pros and cons of adopting a“supply-side” versus “demand-side” perspective andrecognition of the policy formulation, decision-makingand implementation phases when judging the influenceof policy research. The products of economic policyresearch can be delineated into four elements: outputs,outcomes, policy responses and final environmentaland socio-economic impacts. The paper describesvarious indicators that can be used to articulate andmeasure these products and alternative qualitative andquantitative approaches that can be employed to elicitthem. Key issues that continue to arise in the actualconduct of impact evaluations of economic policyresearch are discussed under nine headings: scale andscope; time horizon; supply- versus demand-sideapproaches; the importance of surprise; attribution;choice of indicators; case studies; time lags; and ex anteand ex post assessments. Methods of enhancing impactare then discussed, including the importance ofeffective communications and an understanding of thepolicy environment and processes.

Lessons are drawn from a number of case studiescommissioned by the International Food PolicyResearch Institute (IFPRI) which enable an assessmentof the influence and impact of IFPRI’s research andrelated activities on a number of policy themes. Thelessons are intended to guide the conduct of futureimpact studies; assist in articulating, measuring, anddocumenting successes and failures; and help enhanceimpacts in the future. Aspects addressed include theimportance of research quality, timeliness, andcommunications; type of collaboration andcollaborators; the policy environment; data availability;time scale; consensus building; and the breadth ofexperience on which to draw.

The paper concludes that, in the last five years, progresshas been made in the conduct of case studies designedto assess the impact of economic policy research andlessons have been drawn for the future. There has alsobeen some progress in the development ofmethodologies for quantifying impact in economicterms. However, a number of issues remain. Theseinclude attribution, measurement and the enhancementof impact. As the need for more accountability hasdriven much of the work on impact assessment in recentyears, attribution has been, and is still is, a challengingquestion to which answers remain to be found.

It seems that it is preferable to begin from the “demand-side” in impact assessment. This entails using majorpolicy events as the starting point and workingretrospectively to establish the separate influences of themany research suppliers and other factors on policyresponses. When the interest is on the impact ofparticular institutions in a “supply-side” approach,whole bodies of work on topics rather than individualprojects of limited duration would appear to bepreferred. As far as possible, joint impacts of variousplayers should be measured rather than trying toseparate the contributions of individual institutions.


33

There is a need for more research on sampling andelicitation techniques to ensure objectivity and devisebest practices. In this quest, it seems that exploring theelements of surprise in research information that policymakers were exposed to offers considerable promise.This is especially so with Bayesian approaches.

The most appropriate approaches to impact assessmentshould involve a mixture of both qualitative andquantitative methods. Retrospective narratives are anessential component of the former and, indeed, providethe basis for quantitative estimates and the relatedissue of attribution.

In order to be effective, impact evaluation must beinstitutionalized and not simply an exercise inaccountability. Staff and management should haveresponsibility for recording outputs, outcomes andpolicy responses from their research. Independentevaluators can verify these and endeavor to translatethem into meaningful measures of their impact oneconomic welfare.

Researchers have a responsibility to ensure that theirresearch is disseminated to policy makers usingappropriate communications media. A degree ofadvocacy is also appropriate. However this should notbe taken to lengths that might compromise theindependence and credibility of the researcher. Withthe increased availability of IT and the growing role ofparticipatory democracy and good governance indeveloping countries, there is increased scope forcredible policy research to be accessed by disparategroups and thereby better inform the policy process.

A remaining difficulty is the attribution of economicvalue when economic policy research reinforces currentpolicy settings. Economic policy research thatilluminates the distributional consequences of currentand/or prospective policies seems to have the mostinfluence on policy makers. Estimates of deadweighteconomic losses do not seem to be nearly as influential.

In order to further refine approaches andmethodologies, we need to continue to undertake morecase studies and apply the lessons learnt to enhancefuture impact and help to define “best practices.”


34

Why Impact Assessment Research Has Not Made More of a Difference:Lessons From Eastern and Southern Africa

P. Anandajayasekeram and M. Rukuni

The paper begins by providing a conceptualframework to assess the impacts of agriculturalresearch and development (R&D) activities. Theresults of impact studies conducted in the regionand lessons learnt are presented. Although the ratesof returns estimates (ROR) for R&D investments arerelatively attractive, available evidence suggests thatthe initial efforts had little impact on resourcemobility, and did not significantly improve thereallocation of scarce research resources. In addition,very limited progress has been made toinstitutionalize impact assessment as a planning/management tool.

The results identify reasons for the limited influence ofimpact assessment research. This survey of researchersand research managers provides empirical evidence ofhow group perceptions differ due to the use of impactstudies. Results include an assessment of impactstudies in the processes of; resource mobilization andallocation, informing policy, feedback to researchers,extension staff and donors, as well as performanceassessment of staff and programs. Finally, somesuggestions are made to improve the utilization ofimpact assessment research results.


35

Estimating the Benefits of Plant Breeding Research:Methodological Issues and Practical Challenges

Michael L. Morris and Paul W. Heisey

Impact assessment studies consistently show that thebenefits of plant breeding research are large, positive,and widely distributed. Economic analysisconsistently indicates that investment in crop geneticimprovement generates attractive rates of returncompared to alternative investment opportunities.Similarly, social impact analysis consistently indicatesthat welfare gains resulting from the adoption ofmodern varieties (MVs) reach both favored andmarginal environments and are broadly shared byproducers and consumers. Partially because of thelarge body of empirical evidence that supports thesefindings, governments, international lendingagencies, philanthropic organizations, and privatecorporations have invested millions of dollars in plantbreeding research.

But how reliable are the results of studies thatestimate the benefits of plant breeding research? Arethe methods used to conduct such studiestheoretically sound? And are the data sufficientlycomplete and accurate?

This paper reviews methods specifically used toestimate the benefits of plant breeding research anddiscusses methodological issues and practicalchallenges that often receive inadequate attention inapplied impacts work. The idea is not to question thevalidity of the broad conceptual frameworks used toestimate the benefits of plant breeding research (e.g.,the economic surplus approach, the productionfunction approach) or to examine the major issues inresearch evaluation in general. Rather, the objective ofthe paper is to examine and propose workablesolutions to a number of problems that frequentlyarise when the widely accepted conceptualframeworks are used for empirical analysis of plantbreeding research.

Issues and challenges can be grouped into threegeneral categories:1. Measuring adoption of MVs:� Defining the terms “improved germplasm” and

“modern variety” in the presence of seedrecycling, farmer selection, and creolization.

� Measuring the area planted to MVs (given thedefinitional problem).

� Measuring the area planted to MVs based onsample surveys, seed sales, and expert opinion.

� Consistency; given the limitations of dataavailability and reliability of estimates over time.

� Accounting for external factors that affect the rateand extent of MV adoption (e.g., cost andavailability of improved seed and complementaryinputs, farmers’ knowledge and managementskills, economic policies that affect theprofitability of adopting MVs).

2. Estimating the benefits associated with adoptionof MVs:� Estimating farm-level productivity gains

attributable to adoption of MVs.� Consistency between estimates based on yield

trials and aggregate production data.� Dealing with different types of productivity gains

(e.g., yield increases vs. yield losses foregone).� Distinguishing between the effects of improved

germplasm and the effects of improved cropmanagement practices.

� Accounting for non-yield benefits (e.g., reducedgrowing cycle, improved grain quality, positiveenvironmental externalities).

� Estimating the longevity of benefits attributableto the adoption of MVs when the benefits erodethrough time.

� Anticipating what would have happened toproductivity in the absence of the plant breedingprogram (counterfactual scenario).


36

3. Assigning credit for plant breeding research:� Dealing with research spillovers (assigning credit

to different breeding programs when innovationresults from collaborative research).

� Feasibility of evaluating returns to alternativeplant breeding strategies.

CIMMYT researchers, in collaboration with colleaguesfrom national agricultural research organizations, haveconducted a series of studies designed to documentand quantify the impacts of international maize andwheat breeding research. Based on lessons learnedfrom the CIMMYT studies, each of these issues andchallenges is discussed in detail, and practicalguidelines are presented to help those interested inconducting applied impacts studies to avoid commonpitfalls, which, if ignored, may lead to incorrectempirical results. We conclude with a discussion of thepolicy significance of improved empirical estimates.


37

The Impact of Agricultural Maintenance Research: The Case of Leaf RustResistance Breeding in CIMMYT-Related Spring Bread Wheat

C. N. Marasas, M. Smale, and R.P. Singh

Leaf rust, caused by Puccinia triticina, is a wheatdisease of major historical and economic importanceworldwide. Genetic resistance, rather than the use offungicides, remains the principal means of diseasecontrol. We estimated the impact in developingcountry production of efforts by the InternationalMaize and Wheat Improvement Center (CIMMYT)to breed leaf rust resistant spring bread wheatvarieties since 1973.

The challenge in estimating these benefits is dealingwith the fact that rust pathogens are able to mutaterapidly and form new races that can infectpreviously resistant varieties. Various single genes orgene complexes determine the type, level, andlongevity of a variety’s resistance. Breeding for leafrust resistance is therefore an example of cropmaintenance research. Whereas productivityenhancement is measured in terms of positive yieldgains, productivity maintenance is estimated interms of the yield losses that would have occurred inthe absence of research investment. Although itsimportance has long been argued, there are relativelyfew economic analyses of maintenance research, and,in particular, breeding for crop disease resistance.

The benefits of CIMMYT’s investment in leaf rustresistance breeding were estimated using aneconomic surplus evaluation framework, adjustedfor maintenance research. Gross research benefitswere modeled as the surplus generated by avoidinga cost-increasing shift in the supply curve resultingfrom changes in the environment caused by evolvingleaf rust pathogens. A standard capital investmentanalysis was then applied to estimate the economicreturns on the research investment.

A sample of major spring bread wheat varieties grownin the developing world were classified by their typeand level of genetic resistance to leaf rust throughtrials conducted at CIMMYT. The estimated yieldlosses of these varieties were compared with those thatwould have occurred had all these varieties been fullysusceptible. The area to which yield savings appliedwas estimated by fitting historical logistic diffusioncurves to the global area potentially affected by leafrust. The analysis was further separated by wheatbreeding mega-environment, a classificationdeveloped by the CIMMYT Wheat Program to guideits germplasm enhancement activities. We includedthe full cost of CIMMYT’s wheat improvement effortsince 1973 to calculate the net present value andinternal rate of return on the research investment. Arange of investment values were elicited by varyingassumptions on percentage yield savings, researchcosts, and the discount rate.

The results of this study have two major policyimplications. Firstly, they demonstrate the economicreturns on CIMMYT’s investment in leaf rustresistance breeding since 1973. The global decline inresources for agricultural research increasinglynecessitates their efficient targeting. Secondly, theresults emphasize the importance of maintenanceresearch, which has often been undervalued ineconomic analyses. Studies at CIMMYT indicate thatpart of the progress in wheat yield gains through theyears has been achieved by protecting this yieldpotential through disease resistance breeding. Failureto account for the effects of maintenance researchcould therefore bias rate of return estimates.


38

Economic Impact of Post-Harvest Research in Potatoand Sweetpotato in Developing Countries

Keith O. Fuglie and Thomas S. Walker

Research to lower post-harvest processing costs andlosses and to expand market utilization of cropcommodities has important economic implications forfarmers, marketing agents, and consumers. However,the public sector has not invested heavily in post-harvest research, which continues to be the domain ofthe private sector. The lack of success in rate of returnsand difficulties encountered in assessing the impact ofthis kind of research has not strengthened the case forgreater public sector intervention in post-harvestresearch, vis-a-vis plant-oriented production research.Nonetheless, demand for public-sector investment inthis type of research is rapidly growing because ofperceptions driven by market competition andincreasing globalization. We critically review the

International Potato Center’s efforts in post-harvestresearch on potato and sweetpotato, drawing on casestudies in several regions of the world. The paper alsopresents a general framework for examiningeconomic impacts of post-harvest technical change.Three ways of modeling technical change in foodprocessing are discussed: reducing unit costs ofprocessing, reducing post-harvest crop losses, andexpanding market demand for processed foodproducts. The model is used to illustrate the effects ofvarious types of post-harvest technical change onconsumer and producer welfare. We conclude with anassessment of how post-harvest research compareswith other areas of research investment in theimprovement of potatoes and sweetpotatoes.


39

Impact Assessment in Natural ResourceManagement Research

John Poulsen, Boru Douthwaite, and Douglas White

Hence, monitoring and evaluation constitute a keyfeedback mechanism. Monitoring and evaluation aretherefore the basis of good, adaptive projectmanagement. It is also necessary to know what successlooks like before progress in achieving that success canbe evaluated. Changes in the five capitals (natural,social, physical, financial, human) would capturechanges in production functions, human well-beingpoverty measures, and ecosystem functions, but actualindicators chosen for a specific project should beguided by what people think and want. Participatoryevaluation is an essential part of INRM.

Evaluation of success must include scaling up. INRMresearch by CG centers to produce location-specificsolutions is not enough. Processes by which location-specific solutions can be scaled up both horizontallyand vertically must also be developed. Both horizontaland vertical scaling up are about changing people’sopinions, thinking and practices at different levels. Thisincludes farmers, national stakeholders, andinternational researchers. This vital impact of INRM isoften overlooked by CG center research. Innovation is asocial learning process, including the use of newknowledge. There is an attribution gap. Learning itselfis a social process because people construct newknowledge often in interaction with others. Peoplesocially construct technology and in the process adaptnew technologies and ideas to their systems.Innovation is therefore inherently a complex processwith high degrees of non-linearity. This would seem tomake conventional impact assessment for INRMactivities on highly aggregated development indices,such as poverty alleviation, almost impossible.

INRM views innovation as a social process, in whichpeople construct solutions to their problems. Once oneaccepts that users are modifying technologies and theirown systems to accommodate new technologies andthat these adaptations affect adoption rates and whobenefits and loses, then one must also accept thattechnological change is an immensely complex process,

This paper provides an overview of natural resourcemanagement impact assessment (IA), conducted by theTask Force on Integrated Natural ResourceManagement (INRM). It represents progress madeduring and following the INRM workshop in Cali, 28-31 August 2001. It also incorporates the results of anongoing literature review of impact assessment studiesrelated to natural resource management research.

A Community of Practice initiative on ImpactAssessment in NRM research was formally launched atthe INRM workshop in Cali. The ultimate purpose ofthe initiative was to develop approaches andmethodologies for INRM impact assessment and tointegrate them into the framework of the INRMresearch cycle. Such methods are intended to identifymore effective and efficient research and managementinterventions as well as enabling an assessment of theimpact of agricultural research on decision-makingstrategies regarding the adoption or non-adoption ofspecific practices. However, some intermediateconcerns need to be addressed before appropriatemethodologies can be developed for IA of INRM. Inparticular, we explore concepts and operationaldefinitions with respect to:

� what types of impacts are of interest to decisionmakers

� what impacts need to be measured� what types of indicators/measures are appropriate

Furthermore, we explore how results of the abovewould be used and operationalized.

Impact assessment monitoring and evaluation must bean integral part of the INRM process. While ex-postimpact assessment is still essential, a greater emphasisis needed on monitoring and evaluation. This isprimarily because INRM attempts to catalyze change incomplex environments using complex interventions.Perfect knowledge is not possible before a project startsand project coordinators need to learn as they go along.


40

dwellers and public institutions. Conventionalextension approaches, which often merelydisseminate technologies, are therefore not likely tobe successful. It is difficult to attribute impacts on theground to INRM (given the array of other factorsaffecting local livelihoods and resource management).

Scaling up/out is difficult and there are few examplesof where it has been done. The difficulty is in how toattribute the scaling up/out of INRM to our workand not other variables. Some work is ongoing in thisarea (e.g., rules of thumb on how to deal withattribution). The difficulty of scaling up/out are twoimportant but separate issues. Scaling out difficultiesstem from the heterogeneity of geographic areas. Notonly are biophysical conditions different (the sameones that affected Green Revolution efforts), but alsoINRM includes socio-economic-cultural issues. Theselatter subtleties are the crux of scaling out challenges.Areas and people are different. As for scaling up, theappropriate level of action/analysis is determined bythe specific challenge. Scaling up incurs a loss ofdetail. The goal is to determine the correct scale ofanalysis. Just as the Green Revolution did not scaleup all efforts (e.g. CGIAR in different parts of theglobe) nor should INRM. A balance betweenspecificity and generality is required. Collaborativeresearch networks offer a cost efficient approach. Oneexample would be networks for sharing germplasmand research results for a range of commodities.

A key question is whether or not participatoryapproaches comprising monitoring and evaluationare sufficient in terms of timeliness and scientificrigor. While participatory efforts are important, thereis a probability that they need to be supplementedwith more structured analyses. How this balance canand should be found is an important developmentresearch question.

with a high degree of non-linearity. Current bestpractice economic evaluation methods commonly usedin the CGIAR system, which attempt to establish alinear link between a project’s outputs and wider levelimpacts, struggle with this complexity. Hence IPM andINRM require different evaluation approaches that canbridge this “attribution gap.” We can learn from otherfields, for example evaluation of social programs.However, different skills are required, such as thoseoffered by anthropologists, ecologists and sociologists.

The central question is whether or not we should aimto bridge the attribution gap or simply accept that thereis a grey area where multiple forces come togetherbefore a final impact is reached and supposedlyquantified. Therefore the term “gap” may bemisleading. It is more a junction where the manyefforts meet. Our difficulty is attempting toquantitatively measure the (often) distinct efforts. Butgiven the complementarity of inputs/efforts (e.g.improved seeds and fertilizers, using the agricultureexample), it may be nearly impossible to distill duecredit for any one effort. Moreover, given the limitedfinancial resources with which we can analyze impact,those monies used to precisely attribute impact areprobably best directed elsewhere. Thus, plausiblearguments and qualitative measures of impact becomemore reasonable analytical approaches.

IA is also essential for establishing accountability andsecuring continued funding. Technologies of the GreenRevolution could be relatively easily scaled-up/out,demonstrating high rates of return. However, INRM isnot as easily scaled-up/out. There are problemsassociated with scaling-up/out of site-specifictechnologies. INRM involves far more complextechnologies than those coming out of the GreenRevolution. The benefits of improved INRM aretypically accrued in the medium and long term,requiring coordinated efforts between private rural


41

Assessing Environmental Impacts of Productivity-Enhancing CropResearch: Concepts, Evidence, and Challenges

Mywish K. Maredia, Prabhu Pingali, and Michael Nelson

The objectives of the paper are to present and discuss(1) an empirical estimation the environmental impactsassociated with productivity-enhancing technologies;and (2) the problems and possibilities of attributingenvironmental impacts to research. This is achieved byexamining the factors that contribute to positive andnegative externalities and the conceptual andmethodological issues related to impact assessment.

Evidence related to the nature and magnitude ofenvironmental impacts of productivity-enhancingtechnologies include: land-saving impacts, soildegradation impacts, human health impacts, andimpacts associated with loss of genetic diversity.Aggregate cost estimates of these externalities indeveloping and industrialized countries are presentedwhere appropriate to give an idea of the scale andmagnitude of environmental costs of modernagricultural technologies. The review of this evidenceindicates that: (1) with the exception of salinityproblems associated with irrigation, the loss of soilfertility associated with monoculture, and the healthimpacts of pesticides, the evidence on the extent of thenegative externality problems and their environmentalimpacts are not well documented; (2) it is difficult inmost cases to move from examples to aggregate globalestimates of externality impacts, although it is possiblein a few cases; (3) the conceptual and measurementissues in estimating the monetary values (associatedwith impacts) are too complicated to derive anymeaningful estimates of aggregate environmentalcosts and benefits and; (4) an appropriate measure ofsuch impacts is reduced “land savings” or thecounterpart to the positive environmental impactassociated with productivity enhancing technologies,namely, “land savings” achieved.

The challenge for the environmental impactassessment of productivity enhancing research is toanalyze and measure impacts beyond productivityeffects. This means quantifying the positive andnegative externalities and assessing theenvironmental impacts of these externality effects.The emerging conclusions from the review indicatethe difficulty and complexity involved indetermining the counterfactual and attributing theimpacts to agriculture research. Factors other thanresearch have played an important role in creatingthe environmental problems. Moreover, many of thenegative externalities observed today and discussedin the literature have nothing to do with newtechnologies that resulted from agricultural research(such as the Green Revolution technologies in the1960s and 1970s).

A possible option for attributing environmental coststo past research efforts is to consider the contributionof mainstream research as speeding the rate ofincrease in the intensive use of inputs such asfertilizers, pesticides and irrigation. Alternatively,estimates of counterfactuals can be derived using ageneral equilibrium framework and modeling inputuse (HYVs, irrigation, fertilizers, pesticides) asfunctions of different technical, economic and policyvariables, and relating them with an associatedmeasure of externalities at different levels of inputuse. Another possible approach towards integratingenvironmental impact assessment into economicassessment is including resource quality variablesand environmental externalities in econometricmodels using Total Social Factor Productivity orproduction function analysis. A more feasibleapproach, at least in the short-run from a NARS’ orresearch organization’s perspective, would be tocarry out site-specific case studies on environmentalimpacts of research. The purpose of such case studiesshould be to review the linkages betweenenvironmental degradation and crop technologies inspecific regions/sites, which are impacted fromproductivity-enhancing research technologies.


42

Measuring the Impact of User Participationin Natural Resource Management Research

Nancy Johnson, Nina Lilja and Jacqueline Ashby

Persistent poverty and environmental degradationdemand a constant effort to improve the effectivenessand impact of agricultural and natural resourcemanagement research. Participatory research methodshave been developed to enable researchers to bettertarget their work towards the needs and constraints ofspecific stakeholder groups. Beyond improving theefficiency of the research process, participatoryresearch may also strengthen the capacity ofparticipants to initiate a continuous process ofinnovation suited to their particular needs andconditions. The capacity of users to locally evaluateand adapt technologies may be particularly importantin natural research management.

This paper assesses the impact of using participatorymethods in three natural resource managementresearch projects. The analysis assesses thetechnological, economic, human, social, and costimplications of incorporating users into the researchprocess. Data for the analysis was gather using bothparticipatory and conventional survey methods.

The results suggest that user input early in the projectcan be critical for identifying relevant technologies.User input was also linked to higher levels ofadoptability and/or economic impact. While humancapital impacts were found to be very high whenfarmers worked intensely with researchers, nosignificant social capital impacts were found. Thiscould possibly be related to the plot-level nature of thetechnologies being developed.

Finally, participatory methods do appear to implyincreased short-run costs. However, these costs are notas high as might be expected and are often one-timecosts associated with building capacity to doparticipatory research.


43

Golden Rice: What Role Could it Play in Alleviationof Vitamin A Deficiency?

David Dawe, Richard Robertson and Laurian Unnevehr

Golden Rice is a new variety that has been geneticallymodified to contain beta-carotene, a source of vitaminA. This modification was undertaken as a strategy toaddress vitamin A deficiency, which is widespread inthe less developed countries of Asia. Children’s foodintake data from a poor rural region of the Philippinesis used to simulate the potential impact of Golden Riceon vitamin A intake. The potential impact, coverage ofdeficient subpopulations, and costs of Golden Rice are

compared to two other interventions that have beentried in Southeast Asia, fortification and supplement.Golden Rice could deliver amounts of vitamin A thatare modest but significant in the context of currentintake levels. Also, it compares favorably with otherinterventions in terms of costs and coverage.Understanding where it may have greatest impact onvitamin A deficient diets would help to target efforts toadapt this new variety for release and adoption.


44

The Impact of Iron Bio-availability-enhanced Diets on theHealth and Productivity of School Children: Evidence from

a Mungbean Feeding Trial in Tamil Nadu, India

Katinka Weinberger

Previous studies in Pakistan have shown thatimproved mungbean varieties provide US$8.8million annually to producers and consumers. Ofthis direct economic benefit, 62% goes to producers(5.5 million US$), and 38% go to consumers (3.3million US$). While these figures are impressive,indirect benefits due to technological improvements,which typically exceed those of direct benefits, arenot included in this estimation.

This study quantifies the indirect benefits ofimproved mungbean-related technologies inPakistan and India. A feeding trial in Indiameasured the impact of improved mungbean foodpreparation practices on the bioavailability of iron.Blood iron content, health, and cognitiveachievements were assessed in 225 adult womenand 225 children during a one-year time period.Preliminary results indicate that, compared to thecontrol group, the anemic status of participatingwomen improved considerably.

The study found that in Pakistan, the release of newmungbean varieties led to an increase inconsumption. The impact of increased micronutrientintake on productivity was estimated using thecombined results of a food consumption survey andhealth and wage information. Since wages andnutrient intake may be determined simultaneously,income was treated as an endogenous variable, anda two-stage least square approach was applied.Preliminary results indicate that productivity,measured in wages, is determined by micronutrientintake. Mungbeans were an important iron sourcefor the women sampled.

In conclusion, policy interventions aimed atenhancing micronutrient intake can be regarded asinvestments in improved health and productivityand higher household incomes.

Currently, there is no consensus regarding themeasurement of indirect impacts of improvedtechnologies for the poor, for instance on health andproductivity. The objective of this paper is tocontribute to the discussion by quantifying theimpact of improved mungbean related technologieson the micronutrient status and productivity ofwomen in Pakistan and India.

Micronutrient deficiencies remain a persistent healthproblem in South Asia. Among the different forms ofmicronutrient deficiencies, iron anemia is one of themost prominent. It is estimated that 60% of SouthAsian children under five suffer from some form ofiron deficiency. The principal cause of micronutrientmalnutrition is inadequate diets. Foods rich inmicronutrients, such as vegetables and pulses, areconsumed in insufficient quantities because they areeither unaffordable or undesirable. Promotingmicronutrient rich crops, such as pulses, can help toreduce micronutrient deficiencies. This can beachieved by reducing pulse prices, throughtechnological innovations, and by attempting tochange processing and preparation habits.

Mungbeans are a promising crop in this respectbecause they benefit both producers and consumers.For producers, mungbeans require low inputs, restoresoil fertility, are a short-duration crop, and can easilybe integrated into cereal-based crop rotations usingwheat, maize and rice. For consumers, mungbeanscontain high levels of protein and iron, henceproviding important nutrients for the cereal-baseddiets of the poor.


45

Agriculture and Nutrition: Adversaries,Bedfellows or Allies?

Lawrence Haddad

There is a growing consensus that the reluctance ofagriculture and nutrition communities to collaborateis undermining attempts of both to meet their statedobjectives. This paper briefly reviews the conceptuallinkages between the two sectors, summarizes currentimpact evidence on these pathways, outlines the


drivers that will force a closer collaboration in thefuture, and makes some suggestions about thestrategic choices that need to be made if the “forced”collaboration is to be between allies, not betweenbedfellows or adversaries.

46

What GIS Can (and Can’t) Bring To Impact Assessment:Novel Data, Analysis, and Insights

Gerald C. Nelson

A key characteristic of geographic systems analysis isthat location of an observation is known. The goal ofthis presentation is to indicate where the addition oflocation information adds value to the impactassessment of agricultural research. I adopt the 6


categories of impact assessment research identified byPingali as an organizing framework for assessing thevalue of spatially explicit information and analysis.Selected results from the literature are presented andareas of future use and research identified.

47

GIS Tools: They’re not just for Experts Anymore

D.P. Hodson and J.W. White

To encourage GIS data access, CIMMYT has promotedthe distribution of such tools, including data sets, toagricultural researchers. Climatology, soils,topography, infrastructure, demographics, and cropproduction data are typical inclusions. Additionally,these tools have considerable utility for thedissemination of data products arising from moreadvanced GIS analysis, namely, accessibility surfacesor poverty maps.

Several examples are given of real world applicationsof these tools, including how they can assess impact.The adoption of conservation tillage (CT) technologiesin the rice-wheat region of the Indo-Gangetic Plains ishighlighted. Nearly two decades of research on tillageoptions now appear to be driving a substantial increasein the adoption of reduced tillage practices. Throughthe Rice Wheat Consortium; researchers are usinghand-held GPS units to rapidly record field locationswhere farmers have adopted different CT technologies.These locations, and all associated attribute data, canthen be transferred directly into a GIS to produce anaccurate spatial and temporal record of technologyspread across the region. This is possible in near realtime, is undertaken entirely by researchers in theregion, and requires an investment of less than US$200in GIS/GPS hardware and software.

Despite advances in software usability, data access andanalytic approaches, GIS is still only utilized by afraction of potential users. This is particularly true indeveloping countries. However, an increasing numberof software options and data sources are becomingavailable, and the situation is rapidly changing. Thispaper highlights a selection of widely accessible GIStools, coupled to geographic datasets. These includethe Africa Maize Research Atlas (International Maizeand Wheat Improvement Center-CIMMYT,incorporating ArcExplorer from ESRI Inc.), DIVA(International Potato Center-CIP/International PlantGenetic Resources Institute-IPGRI), and the AlmanacCharacterization Tool (Mud Springs GeographersInc.). These tools represent increasing levels ofcomplexity in entry-level GIS systems, but all permitnon-GIS experts to analyze and interpret geospatialdata. The tools provide a means to improve spatialawareness amongst researchers that, in turn, can fostermore efficient use of higher-level GIS resources.Additionally, such tools typically require a minimalamount of training (a maximum of two days), andrepresent zero or minimal cost options.


48

Lost in Space: Enhancing the Role of Spatial Analysis inStrategic Impact Assessment

Stanley Wood and Jordan Chamberlin

Assessing the past or potential impact of agriculturalresearch and development at national and internationalscales is confounded by a great deal of real-worldheterogeneity. One attraction of spatial analysis is itscapacity to represent simultaneously the spatialvariability of many key factors influencing agriculturalproduction decisions and performance: climate, terrain,soils, water resources, and the physical accessibility ofinfrastructure and markets. In the same framework,these factors can be juxtaposed against the spatialdistribution of ecosystems, people, crops, livestock,and the threat of pests and diseases. In theory, thisanalytical capacity holds great promise for our abilityto better understand the ranges and combinations ofsocioeconomic and environmental conditions underwhich agriculture takes place and, therefore, to makemore informed assessments of the potential impacts ofspecific policy, institutional and technologicalinnovations. In practice, there are very few impactassessment applications, at the meso or macro scale, inwhich GIS has played a significant analytical role –rather than serving simply as a data mapping tool.

This paper reviews some of the conceptual andpractical difficulties in converting the promise ofspatial analysis into a reality for the type of strategic,

cross-country impact assessment that is (or should be)of high priority for the CGIAR. Many of thesedifficulties relate to differences in definitions, formats,resolution, and time periods among spatial datasources. But even more problematic is the fact thatmany critical datasets simply do not exist, or do notexist at resolutions that are sufficient or consistent.Three examples are: the spatial distribution ofproduction, the spatial distribution of people, and thespatial distribution of soil and water resources.Ironically, these are precisely the core datasets we needto trace through the potential chain of impact from, forexample, technology adoption to productivity changeand hence to markets, food availability and incomes,and to natural resource stocks and flows.

We describe work being undertaken by IFPRI toaddress several of these challenges. These include thedevelopment of a robust method to perform the crop-specific spatial allocation of production, and ofimproved population distribution maps that explicitlydistinguish between urban and rural areas andpopulations. We also describe progress on thedevelopment of a compatible economic model(DREAM) designed to assess the potential benefits oftechnical change in a multi-region setting.


49

Assessing the Impact of Agricultural Research on Poverty with theSustainable Livelihoods Framework: Concepts and Methods

Michelle Adato and Ruth Meinzen-Dick

framework provides a common conceptual approach toexamining the ways in which agricultural research andtechnologies fit into the livelihood strategies ofhouseholds with different types of assets and wealth,and the types of outcomes for reducing poverty andvulnerability. Applying this framework requiresinterdisciplinary research and a combination ofquantitative and qualitative methods.

This paper reports on the conceptual framework,methods, and findings to date of these studies. The firstpaper provides an overview of the approach, followedby more detailed methods and results from three of thecase studies: two in Bangladesh and one in Mexico.

As the Consultative Group on InternationalAgricultural Research (CGIAR) goes beyond the goal ofincreasing food production to the broader goal ofreducing poverty, both agricultural research andstudies of its impact become more complex. TheStanding Panel on Impact Assessment (SPIA) hascommissioned a multi-country study of the impact ofagricultural research on poverty, led by InternationalFood Policy Research Institute (IFPRI) with theparticipation of several other centers.

The studies use an expanded definition of poverty thatgoes beyond income or nutrition-based head counts, toconsider other aspects of well-being andempowerment. The Sustainable Livelihoods


50

The Impact of Rice Research on Poverty Alleviation:The Bangladesh Case

Mahabub Hossain, David Lewis, Manik Lal Bose, Alamgir Chowdhury, and Ruth Meinzen-Dick

districts. The 1999-2000 survey selected half of thosevillages, which represented different rice ecosystemsand infrastructure development, and drew a sample of30 households from each village using the wealth-ranking method. The surveys generated informationon sources of household vulnerability, asset bases,livelihood strategies, and conventional economicindicators of outcomes, including self-perceptions ofpoverty status. The paper analyzes the quantitativeinformation using the sustainable livelihoodframework and tests hypotheses through focus groupdiscussions conducted in a few selected villageswithin the sample.

The analysis covers the following issues: a) changes inthe asset base for socially differentiated groups intechnologically progressive and backward villages andhow they affect livelihood strategies; b) determinantsof the adoption of modern rice technology using aTobit model incorporating biophysical and socio-economic factors; c) the impact of technology adoptionon productivity of inputs, profitability, householdincomes, and asset accumulation for householdsstratified by livelihood strategies and; d) therelationship between growth in rice productivity andincome generated for the resource poor households inrural non-farm activities using the village level datawithin a multi-variate regression model.

Bangladesh has made notable progress in sustaining arespectable growth in food grain production over thelast three decades through large-scale adoption ofmodern rice varieties. This is despite the decliningavailability of arable land and predominance of smallfarmers and landless agricultural laborers. Recentstudies analyzing secondary data have indicatedmoderate improvements in poverty for both the ruraland urban population. In order to understand theimpact pathways of rice research on the welfare of thepoor, this paper analyzes the differential adoption ofimproved rice varieties across socially differentiatedgroups and the effect of adoption on productivity,employment, profitability, and asset accumulation. Italso attempts to explain how resource-poor householdshave gained or lost from changes in livelihoodstrategies and outcomes induced by productivitygrowth in rice cultivation, especially mediated by keyinstitutions and infrastructure development.

The paper utilizes two-point (1987-88 and 1999-2000)sample household survey data on the operation of therural economy for a subset of a nationallyrepresentative sample drawn by the BangladeshInstitute of Development Studies in 1987. The samplefor the 1987 survey was drawn using a multi-stagerandom sampling method and consisted of 1,240sample households from 62 villages belonging to 57


51

Impact of Improved Vegetable and Fish PondTechnologies on Poverty in Bangladesh

Kelly Hallman, David Lewis, Suraiya Begum, and Agnes Quisumbing

(2) Using quantitative data on myriad communitycharacteristics, the credibility of the counterfactualdesign of the quantitative survey is assessed. This is toensure that NGOs did not purposefully select villagesfor introduction of the technologies where they mayhave been more likely to be successful as such asituation would contaminate the validity of the controlsample. (3) Assessing how various livelihood assetsinfluence membership in NGOs and how they helpshape adoption of the agricultural technologies.Findings will be drawn from focus group discussionsand analysis of a census of households in each site andthe household survey data. (4) Describing thevulnerability of contexts of households as reported inthe focus groups. (5) Characterizing how thetechnologies fit into livelihood strategies ofhouseholds. This will include findings from the focusgroups and from the construction of income sourceportfolios for “adopter” versus “likely adopter”households using the survey data. (6) Documenting theimpact of agricultural extension programs andtechnologies on the empowerment of women and otherdisadvantaged groups, via focus group results andanalysis of the survey data. (7) Demonstrating howtechnologies were described by participants in focusgroups to contribute to their overall well being andcomparing the livelihood outcomes of “adopter”versus “likely adopter” households using the surveydata. In both instances, traditional and broaderconcepts of poverty are considered.

This case study examines the impacts of improvedvegetable and fishpond technologies on poverty inrural Bangladesh. It builds upon a broad quantitativesurvey undertaken by IFPRI in 1996/1997 thatexamined the impact of the delivery of agriculturaltechnologies by NGOs—with training and credit—onnutrition, micronutrients, and women’s status withinthe household. Careful attention was paid in the surveydesign to the construction of a set of appropriatecounterfactuals for measuring impact. Study sites werenot purposefully chosen as areas where the pro-poorimpact was most likely to be observed, thus providingan unbiased setting to begin to address “povertyimpact” questions. The present study undertakes anintegrated economic and social analysis to assess theimpacts of such agricultural programs on poverty.Poverty is conceived here to encompass not onlyincome and expenditure, but also vulnerability andcapabilities. Issues not easily addressed in aquantitative study—such as perceptions of poverty,livelihoods strategies, the institutional setting, andtechnology dissemination pathways—are informed byqualitative data collection.

In this paper we report on: (1) Mapping of “well-being”concepts derived from an independent countrywideparticipatory poverty assessment onto the existinghousehold survey data for the purpose of determiningwhere survey households fall along a participant-defined global “well-being” spectrum. These well-being categories form the basis for selection of 54 focusgroups: 3 male and 3 female groups from middle, poor,and “hard core” poor households, respectively, in 3villages in each of 3 study sites.


52

The Impact of Improved Maize Germplasm on Poverty Alleviation:The Case of Tuxpeño-Derived Material in Mexico

Mauricio R. Bellon, Javier Becerril, Michelle Adato, and José Alfonso Aguirre Gómez

and a collection of agronomic evaluation of maizesamples were used. Twelve communities, withdifferent degrees of marginality (very high, high, andmedium) were selected in two of the poorest states inMexico (Oaxaca and Chiapas). This was done with thehelp of a geographic information system (GIS) thatsynthesizes information on marginality, formaldistribution of improved germplasm, and theagroecological adaptation of improved Tuxpeñogermplasm. In the twelve communities, a series offocus group discussions were organized to assesstechnical issues regarding maize production andvariety adoption. In eight of them, additional focusgroups were organized with poor and relativelywealthy informants to understand their livelihoodstrategies. Finally, in four of them, in-depth casestudies on the local livelihoods were carried out withselected households. Based on the qualitativeinformation generated with the focus groups, a surveyhas been developed and a sample survey will beimplemented in the twelve communities. The surveyincludes a section on farmers’ evaluation of maizetypes grown in terms of traits previously identified asimportant using participatory methods. This shouldallow a comparison of the different maize types basedon farmers’ perceptions. Finally, grow-outs of thecollected maize samples, as well as sample fromavailable commercial varieties, will be carried out toquantify the agronomic differences among the samplesand assess how the adapted maize compares to theoriginal products of breeding.

Improved maize varieties have been available inMexico for more than 40 years. Unfortunately, diffusionof these varieties has been relatively limited, despiterepeated government campaigns to encourage use ofimproved seed. However, the relatively low rate ofdiffusion may give a misleading impression of the trueimpacts of improved germplasm on the welfare of ruralhouseholds. A growing body of evidence suggests thatmany small-scale, subsistence-oriented farmers havetaken up improved varieties and through theirmanagement transformed them to better suit theirneeds. This process can be seen as a “middle way” bywhich scientific technology is adapted to local farmerconditions, thus serving as a vehicle by which the poorbenefit from improved technology. The objective of thisstudy is to assess the impact of improved maizegermplasm on rural poverty in lowland tropicalMexico. Its focus is on the Tuxpeño germplasmcomplex, which has been subjected to an intensivebreeding effort by the International Maize and wheatImprovement Center (CIMMYT) and its nationalpartner INIFAP during the last forty years. The studyinvolved three separate but related activities: 1)measuring the diffusion, local adaptation, and use ofimproved maize germplasm; 2) understanding howchoices about adaptation are linked to the livelihoodstrategies and vulnerability context of ruralhouseholds, and; 3) assessing the impacts of adoptionon the welfare of rural households. Differentmethodologies such as participatory methods,ethnography, case studies, a household sample survey


53

The Economic Costs and Benefits of a Participatory Project toConserve Maize Landraces on-Farm

Smale, M, Bellon, M, Aguirre, A. Manuel, J. Mendoza, A.M. Solano, R. Martinez, and A. Ramirez

The aim of this paper was to assess whether farmerwelfare in the Central Valleys of the state of Oaxaca,Mexico, could be enhanced through participatory cropimprovement while maize genetic diversity ismaintained or increased. The underpinning projectrationale is conservation of potentially valuable allelesin landraces that have evolved on farms under naturaland human selection pressures. The papercharacterizes the unique structure of the economiccosts and benefits of the project, which include bothprivate and public benefits; estimates the private costsand benefits to farmers as well as from the perspectiveof a private investor; and describes the distribution of

benefits among social and economic groups. Farmersas a group obtained a high benefit-cost ratio fromparticipating. However, from a private investorperspective, benefits are not likely to justify the cost.There is a clear gender bias in both participation andbenefit distribution, though there is also evidence of apositive welfare transfer to maize deficit households.Many public benefits of the project would beexceedingly difficult and costly to document. Thestudy raises issues concerning the application ofstandard economic methods to impact analysis insimilar situations.


54

The Distribution of Benefits from Public InternationalGermplasm Banks: The Case of Beans in Latin America

Oswaldo Voysest, Nancy Johnson, and Douglas Pachico

The unrestricted international flow of geneticresources from international gene bank collections isperhaps one of the greatest impacts of internationalagricultural research. This paper examines thedistribution of benefits from bean genetic resourcesacross countries in Latin America.

The genealogy of commercial bean cultivars releasedsince 1976 was analyzed. This enabled the source ofcommercially grown genetic resources to becalculated for each country. All countries were shown

to be heavily dependent on imported geneticresources for their commercial cultivars. Informationon the economic impact of improved bean varietiesenables the share of economic productivity benefitsassociated with imported germplasm to be calculatedby country of origin. The benefits received by eachcountry from improved bean germplasm arecompared with the contribution of that country’sgermplasm to other countries. Patterns in the flowand use of genetic resources and associated benefitsare analyzed.


55

Endowing Future Harvests: The Long-Term Costs ofConserving Genetic Resources at the CGIAR Centers

Bonwoo Koo, Philip G. Pardey, and Brian D. Wright

The eleven Consultative Group for InternationalAgricultural Research (CGIAR) gene banks have grownconsiderably in size over the past few decades,currently holding over 660,000 accessions ofgermplasm. Conserving germplasm is a long-term, ifnot in perpetuity, proposition. Hence, the mismatchbetween the predominantly annual financial supportfor this conservation effort and its long-term aims is aserious concern. Using the results of five CGIAR genebank case studies, this study estimates the size of anendowment or stewardship fund, the earnings fromwhich would assure a funding stream to conserve thisgenetic material for future generations. It is estimated

that the annual cost of conserving and distributinggenetic material presently held in the CGIAR genebanks is US$5.7 million/year, and a commitment tounderwrite these core gene bank services for thebenefit of all future generations could be met by settingaside a fund of US$149 million (invested at a real rateof interest of 4%/yr). This would be sufficient tounderwrite the CGIAR’s current conservation activitiesin perpetuity (estimated to be US$61 million), as wellas maintain the distribution activities (US$88 million)that provide germplasm to breeders, scientists, farmers,and others worldwide.


56

The Simple Econometrics of Impact Assessment: Theory with anApplication to Milk-Market Participation in The Ethiopian Highlands

Garth Holloway and Simeon Ehui

investigations of agricultural productivity, equity,poverty, social health, and nutrition, especiallywhere the traditional household-production model isthe modus operandi. Surprisingly, there have beenno applications of these powerful techniques toimpact assessment. Hence, the study represents thefirst of its kind.

In our application, we present empirical resultsrelated to an emerging market in the Ethiopianhighlands. We explore the measurement of variouspolicies targeted toward promoting participationamong formerly subsistence milk producers at twoperi-urban sites close to the capital city, AddisAbaba. In this context, the policy questions includethe identification of the covariates and the levels thatare required to effect entry by non-participants in amarket; in other words, measures of the inputsneeded to create and sustain a market. Thisapplication is important in the context of developingeconomies, where the density of non-participants ishigh and is considered to be the main impediment toeconomic development.

Our main objective was to estimate the covariatelevels that prompt market entry of non-participatinghouseholds. The strategy was very simple. First, weestimated each non-participant’s distance from themarket in terms of the (latent) level of marketablesurplus that the household would wish to produce,given its covariates. These are the augmenting data.By definition, they are negative real numbers.

Subsequently, we computed the levels of thecovariates that make these negative numbers becomepositive, signifying positive marketable surplus and,presumably, entry into the market.

This paper presents novel econometric techniques forenhancing the robustness of impact assessments in thecontext of econometric investigations of researchproductivity, institutional innovation or povertyalleviation. The methodology has broad application totime series or cross-section data. However, it isespecially useful when applied to household paneldata (the focus of this study) when adoption andparticipation issues are in question.

In these circumstances, one problem that frequentlyarises is a policy choice between alternative means ofachieving stated objectives. Because the impacts ofvarious policies are often predicated on econometricresults, policy options are associated with a probabilitydistribution about their impacts. When thesedistributions emanate from different parametricfamilies, a problem of comparison arises. For example,the distributions of the precise impacts of twoalternative policies may have different locations (asindicated by their respective means estimates) and mayalso have different scales (as indicated by theirrespective variance estimates). Even when thedistribution of one policy stochastically dominatesanother, a problem of ranking arises when budgetaryconsiderations limit resources.

In this paper, we demonstrate how measurementproblems of stochastic impact assessment can beresolved tractably, intuitively and robustly as anoutcrop of most conventional econometric modelingexercises. The methodology exploits recent advances inMarkov Chain Monte Carlo (MCMC) methods. MCMCmethods are a collection of iterative techniques forestimating models in which discrete-choice, truncatedor censored-regression formulations are at issue. Thisgeneral class of models appears frequently in empirical


57

In 1999, the total science and technology budget inArgentina reached US$ 1, 517 million. Thiscorresponds to 0.54% of the national PBI andrepresents US$ 50,800 of annual investment per capitafor the 20,911 research staff. Agricultural, forestry andfisheries research receives 10.6% of the budget.

The focus of the present study is to evaluate theproductivity of this sector by analyzing the efficiencyof agricultural research scientists in the NationalResearch Council of Science and Technology ofArgentina (CONICET).

The data used in the study was obtained fromCONICET databases and covers 1996-2000. Thespecific discipline of Agricultural Sciences has 107researchers representing 3% of a total of 3,685individuals. There are also a significant number ofresearchers who belong to other disciplines but withresearch competences in agricultural subjects and whoare not included in this sample.

To evaluate efficiency, we use the approach proposedinitially by Farrell in 1957. It consists of the estimationof a production function that allows the calculation ofthe maximum output (y *) that can be obtained byeach production unit for an input combination. Thelevel of technical efficiency (TE) of each productionunit can be defined as the relationship observedamong the actual product (y) and this maximum (y *):0 £ ET = y/y* £ 1. Therefore, to evaluate TE, it isnecessary to know the underlying productionfunction; i.e., the production function that could beestimated from the available data. Then, TE can becalculated from the best practices observed.

The proposed model is: yi = f (xi, b) + ei

where yi is the output of unit i; xi is the input vector, bis a vector of parameters and ei is a composed randomerror: ei = vi + ui . vi is a symmetrical error thatrepresents the aleatory variations in production, takenas i.i.d. N(0, s2). The ui is an asymmetric term thatrepresents technical inefficiency and it is assumed thatit is distributed independently of vi and that it satisfiesthe condition ui £ 0.

The stochastic frontiers models used in the study arethose proposed by Battese and Coelli (1996). Thesemodels allow the use of panel data (a sample of Nindividuals in T periods), and have differentspecifications about the composition of uit (technicalinefficiency effects). To assess the determinants ofCONICET Agricultural Sciences Researchers’efficiency and productivity, research output will bemeasured in quantity and quality of publications,patents, theses, and advice given. Different inputmeasures will be considered such as project budgets,actual expenses, and wages. Specific variables will beincluded to assess efficiency effects and to identifyscale effects and size economies. These variables arenamely: type of research institution (university, publicresearch institute, private institution); geographicallocation; use of technological and scientific networks;individual characteristics (age, gender); andenvironmental aspects.

Measurement and Sources of Efficiency in Argentina’s AgriculturalSciences Research System: A Stochastic Frontier Analysis

Daniel Lema, Susana Mirassou, and Martin Guppy


58

This paper addresses the question raised in theconference title: why has impact assessment researchnot made more of a difference? Impact assessmentresearch has been carried out for a long time, but wesee that the research agenda of most institutions is notsignificantly affected by its results. What is goingwrong? In this paper we will draw from lessons learntfrom the development of Tradeoff Analysis (TOA) tohelp answer this question. Perhaps the most importantlesson is that the integration of decision makers, otherstakeholders and scientists into the process of impactassessment is a critical element in making an impactwith impact assessment. In a companion paper, weaddress another key issue, namely, the need forsuitable methods and tools to implement integratedassessments of agricultural production systems thataccount for both economic and environmental impacts.

The TOA approach is based on the concept that, inorder to be used by public policy decision makers andother stakeholders, impact assessment research mustbe integrated into a process that involves both thedecision makers and the scientists engaged in theresearch. The TOA approach to integrated assessmentbegins by bringing together decision makers andscientists to identify quantifiable sustainabilityindicators that are judged to be relevant to assessingthe impacts of a production system. This group alsoformulates hypotheses about relationships betweenthese indicators (which may be either tradeoffs or win-win outcomes), and determines how information aboutthese relationships can best be delivered to the users ofthe information (decision makers, interest groups, thegeneral public, etc.).

The next step in the TOA approach is to develop dataand models that are capable of quantifying thesustainability indicators on relevant spatial andtemporal scales. Among the critical methodologicalissues that must be addressed are the availability ofsuitable data, and methods for appropriately

integrating disciplinary data and models into acoherent representation of the agricultural productionsystem. The Tradeoff Analysis Model was developed tofacilitate the process of data organization and modelintegration for application of Tradeoff Analysis.

Does the use of TOA guarantee that impactassessments will have an impact? Our experience withdeveloping and applying this approach in LatinAmerica suggests that the involvement of stakeholdersand the development of an operational tool that canquantify indicators selected by users can significantlyincrease the impact. However, there remain importantbarriers to the use of integrated assessment findings bydecision makers and the general public. Scientific datawill always be only one of many factors influencingpolitical decisions. The political leadership and publicinstitutions, especially in developing countries, arehighly dynamic and the identification of sustainabilityindicators in the beginning of the research does notmean that those are the indicators the same institutionis interested in. Moreover, social concerns about atechnology’s impacts may be evolving rapidly (as inthe case of the use of genetically modified materials),so the relevant set of indicators may change over time.

To illustrate these issues, we review how TOA wasapplied in Ecuador. Key concerns addressed were theeconomic, environmental, and human health effects ofpesticide use, as well as longer-term soil conservationissues related to tillage and land use practices. Wedescribe how research was formulated and how resultsfrom the impact assessment were communicated tolocal, regional and national interest groups; assess thedegree of impact that this project had, and drawlessons from this experience for impact assessmentresearch. We conclude that the integration of decisionmakers, other stakeholders and scientists into theprocess of impact assessment is a critical element inmaking an impact with impact assessment.

Making an Impact with Impact Assessment: The Tradeoff AnalysisApproach and Lessons from the Tradeoffs Project in Ecuador and Peru

John Antle, Jetse Stoorvogel, Walter Bowen, Charles Crissman, and David Yanggen


59

Factors Affecting the Adoption of MaizeTechnologies in the Hills of Nepal

J. K. Ransom, K. Paudyal, and K. Adhikari

occurred five years earlier in Dolakha than in Dailekh.Only 32% of the farmers used fertilizers in the remoteVDC of Dailekh. Somewhat surprisingly, availabilitywas not mentioned as a serious constraint to the useof fertilizers in Dailekh district. Averaged across thetwo districts, the use of improved open pollinatedvarieties (OPVs) of maize was less than 30%. Thehighest use of OPVs (59%) was recorded in theaccessible VDC of Dolakha and the lowest use (13%)in the inaccessible VDC of Dailekh. The adoption ofimproved OPVs was consistently less than theadoption of fertilizers in all VDCs. The main reasonfor the non-use of OPVs in Dolakha was the lack ofavailability of seed, while in Dailekh it was the lack ofknowledge of the new varieties.

Data from this impact assessment exercise suggeststhat the strategy for improving the adoption of newtechnologies may be quite different for the twodistricts. Since the lack of information and experiencewith the new technologies is the major constraint totheir adoption in Dailekh, a program of moreintensive on-farm demonstrations and testing isprobably justifiable. In Dolakha, on the other hand,availability is more constraining than knowledge.Therefore, actions should focus on improving theavailability of inputs. Community-based seedproduction programs should be supported andpolicy/institutional adjustments should be made toensure the timely delivery of fertilizers to thelocations where they are needed.

Maize, the most important cereal crop in the hills ofNepal, is used as food and stover for animal fodder.Furthermore, it is a traditional cereal crop and farmershave developed farming systems that efficiently utilizethe resources that are available to them. Average maizeyields, over the period 1970 to 1984, declined from 1.9t/ha to 1.4 t/ha. Since 1984, however, yields haveimproved to around 1.8 t/ha. Maize farms are smalland population pressure necessitates the intensificationof existing farming systems. Research directed atdeveloping technologies for maize production began in1965. In 2000, a survey was carried out in two districtsof Nepal to determine the current level of adoption ofimproved maize production practices. In each of thesedistricts, remote and accessible Village DevelopmentCommittees (VDCs) were surveyed. Questionnaireswere administered to 54 randomly selected householdsin each of these VDCs.

Data analysis determined the socioeconomic, physical,and technological factors that influence the use oftechnology by farmers. All households use compostedfarm yard manure (FYM) as part of their fertilitymanagement strategy. On average, 75% of the farmersused inorganic fertilizers, primarily urea. However, usepatterns varied considerably depending on location.Urea use was considerably higher in Dolakha,regardless of the remoteness of the VDC, than inDailekh, and the adoption of the use of fertilizers


60

Maize is the main annual crop in Honduras, both interms of its share of total cropped area and its role indirect human consumption. Approximately 25% of allarable land is planted with maize and Hondurans’ percapita maize consumption is among the highest in theworld. Yet despite the potential yield gains from theadoption of improved varieties and the fact that seedprices (of hybrids) are relatively low compared to otherLatin American countries, the level of adoption ofimproved varieties of maize in Honduras is below 20%.

This paper summarizes research into the factors thatcontribute to these low levels of adoption of improvedmaize varieties in Honduras. The literature onadoption of agricultural innovations is vast. Commonlyidentified constraints to rapid adoption of newagricultural technologies include credit rationing,information asymmetries and/or differential access toinformation resources; risk aversion; small farm size;human capital deficiencies; disruptions in markets forlabor and complementary inputs; and poorinfrastructure. The low levels of adoption of improvedmaize varieties in Honduras are commonly ascribed tothree factors: (a) inappropriateness of improvedvarieties for the specific environments or needs offarmers; (b) incomplete or non-existent markets forcertain varieties due to high transactions costs; and (c)inadequate transmission of information on newvarieties to farmers.

Our empirical work is based on an agriculturalhousehold model that explicitly incorporates varietycharacteristics and transaction costs into thehousehold’s optimization process. We considered amultitude of production characteristics (e.g., yield,yield stability, duration, and plant height), andconsumption characteristics (including taste,

storability, and husk cover thickness) that are valuedby farmers. Distance from markets, road quality,vehicle ownership, and availability of marketingagents were used as proxies for household-specifictransaction costs.

We implemented our model, using data collected in asurvey of 167 farmers located across 34 villages, in twodistinctly different agro-ecological zones. In one ofthese zones, maize farming is highly commercialized,average farm sizes are comparatively large, andhybrids are planted by 60% of farmers. Near-subsistence farm households with smallerlandholdings and much lower levels of high yieldingvarieties (HYV) adoption dominate the other zone.

Non-parametric tests indicate that farmers perceivesignificant differences among varieties. In general,improved varieties dominate in terms of productioncharacteristics but are regarded as inferior with regardto consumption characteristics in both productionzones. Zone-specific adoption equations confirm thatin both areas production characteristics variables arejointly significant factors that influence variety choice,while consumption characteristics appear to be lessimportant. In addition, transactions costs are found tohave a significant positive impact on HYV adoption inthe commercialized zone. No detectable effect oftransactions costs on adoption is found in the near-subsistence zone. Instead, the dominant elementconditioning adoption appears to have been whetheror not farmers received free HYV seeds in theaftermath of Hurricane Mitch, a result indicating thatinformation deficits may be an important limitingfactor to adoption there.

Variety Characteristics, Transactions Costs andMaize Adoption in Honduras

Hernando Hintze, Mitch Renkow, and Gustavo Saín


61

In Kenya, wheat is usually grown on a commercial basisfor the domestic market. However, as wheat supply hascontinually failed to meet the country’s demand, theshortfall is supplemented by wheat imports.

Despite the continuous generation of new wheattechnologies via Kenyan agricultural research centers,adoption by Kenyan farming communities has beenvery poor. Plant breeding has increased wheatproductivity, but farmers continue to experience lowyields. The adoption of recommended wheatproduction technologies by Kenyan farmers couldsignificantly reduce the need for wheat imports.

The purpose of this study was to identify and describefactors affecting the adoption of selected wheatproduction technologies by farmers in Njoro andRongai, Nakuru District, Kenya.

The factors investigated were derived from previousstudies and observations. These included high yieldingvarieties, land preparation and planting time, fertilizerapplication rates, improved seed source, pest, disease,and weed control strategies. A sample size of 150farmers was selected using a stratified proportionrandom sampling technique. The data were obtainedusing a validated questionnaire and analyzed using thedescriptive and inferential statistics. Each hypothesiswas tested using Chi-square at P<0.05.

In conclusion, the study developed the followingspecific recommendations to facilitate the effectivetransfer of wheat production technologies:

• In-depth similar studies should be carried out in allwheat growing district to confirm if the trends arethe same.

• Education for both large- and small-scale wheatfarmers should be emphasized to allow easy dis-semination of knowledge and information aboutwheat production technologies. This will modify thefarmers farmers’ perception and help them to graspthe potential benefits of adopting new technologies.

• There is need for collaborative efforts in educatingthe farmers by all parties interested in agriculturaldevelopment so as to increase adoption levels.

Factors Affecting the Adoption of Selected WheatProduction Technologies by Farmers in Njoro and Rongai

Divisions of Nakaru District, Kenya

Alice Ndiema, W. Ongo’ndo, A. A. Abdallahi, and M. G. Kinyua


62

Expanding the Use of Impact Assessment and otherEvaluation Research Evidence

Ronald Mackay and Douglas Horton

The rationale behind impact assessments is thecommonsense notion that the evidence from these willbe used to bring about improvements in the policiesand programs evaluated and thereby contribute toeconomic and social betterment. As it becomes evidentthat impact assessment research is making less than thedesired difference to the alleviation of human distress,evaluation researchers seek to understand why, andwhat can be done about it.

This paper addresses the difference/lack of differencemade to social betterment (e.g. food sustainability,poverty reduction, and the sustainable environmentalmanagement) by impact assessment and by evaluationresearch more broadly. First, the paper examines theprincipal reasons why evidence from evaluationresearch is not more immediately and effectively used;then it explores some of the ways that practitioners inthe international evaluation community are working todesign and conduct evaluations that promoterelevance, credibility, and the practical use of findings.

The paper reviews the international evaluationcommunity’s current thinking about impact evaluationand other evaluation research. It also draws on theauthors’ direct experience conducting and promotingthe use of evaluation in research and developmentorganizations over the past 25 years.

Evaluation is a powerful, emerging “transdiscipline.” Itis also a highly sensitive field of practice in that it isvirtually never undertaken in a politically orideologically neutral environment and seldom within apolicy vacuum. However, it has been longacknowledged that its principal weakness is therelatively mild influence that evaluation evidenceexerts on individuals, organizations, and the broadercommunities charged with decision making. Manyexplanations have been put forward for this, inaddition to the fact that evaluation evidence is only onesource of information used to influence policy andoperations, others being the values, beliefs, preferences,prejudices, and needs of the decision makers and theirconstituencies. It is acknowledged that someevaluations are biased towards the perspective of asingle stakeholder group, some fail to establish anevaluation team with credibility in content ormethodology, some unjustifiably assume that theevaluation approach adopted is acceptable to theprinciple decision-makers.

To address the lack of influence of evaluations, thedifferent philosophies, perspectives, values andpractices that make up the evaluation discipline havebeen subjected to close scrutiny. The purpose of thisscrutiny is a better understanding of the construct“evaluation use” and the identification of practices thatpromote utilization. The paper reviews this work andmakes practical suggestions as to what evaluators cando in order to “make more of a difference” to thepolicies and programs they address.


63

This paper is concerned with developing a balancedapproach to impact assessment (IA), and one thatdistinguishes between the need for accountability onthe one hand and learning on the other. Drawingupon both programmatic and strategic experiencegained from working with the UK’s Department forInternational Development’s Renewable NaturalResources Strategy, it addresses the need for learningabout institutional impacts of this program onResearch and Development (R&D) systems inSouthern partner countries. A central thesis is that theInternational Development Targets and thesubsequent adoption of more holistic, poverty-focused approaches are beginning to enhanceorganizational capacity in these countries to developa variety of approaches and methods to IA. The paperargues that these are driven by a perceived need foraccountability to better document success as aresponse to external pressure for clear evidence ofimpact. This concern for accountability tends tosupport historic decisions. The paper highlights fourinterrelated areas of concern with regard to themanagement of IA by donors and the internationaland strategic agricultural research agencies:

(1) agencies are driven to manage IA to justifycontinuing what they do, as opposed to informingfuture strategies; (2) the emerging concern withframeworks and approaches to poverty reduction,including those that espouse embracing error, havefailed to build on past lessons about the use of IA; (3)the absence of a systematic and rigorous frameworkto assess institutional capacity and prospects forsustainability has marginalized the role of nationalR&D systems and reduced them to “risks andassumptions;” and consequently (4) the capacity andsustainability of local R&D systems is largelyoverlooked during collaborative research betweeninternational and national research systems.

The current preoccupation with assessingbeneficiary impact, as gauged by movements in therelative values of household assets, tends to maskthe relative lack of information and concern aboutthe capacity and capabilities of local R&D systemsbefore, during, and after investment periods. Thismakes it difficult to link any sustainable impactamong beneficiaries with information oninstitutional capacity at the time that researchproducts were being developed. This could be onereason why there remains a lack of confidence in theresults of impact studies showing high or low ratesof return— a coherent and believable story linkingimpacts to research capacity and process is notpresented, especially given that policy makers arewell aware of such capacity constraints. There isneed for guidance on what research processesworked and did not work in order to identify thestrategic implications for future investments in R&D.

The paper concludes by proposing a more balancedapproach to IA, one informed by two mainconsiderations. First, that it is not sufficient to simplyundertake an assessment of institutional impacts,rather that, it is a question of understanding howsuch changes complement socio-economic studies.Second, that there is a need to develop sustainabilityindicators that relate to capacity and development.These considerations will help to guide thedevelopment of a more balanced IA strategy, onethat reflects changes within R&D systems and whichprovides information to policy makers about thesustainability of impacts among end-user/beneficiary environments.

Why does Impact Assessment Continue toNeglect Institutional Sustainability?

Daniel Ticehurst, Simon Henderson and Alistair Sutherland


64

Demand-driven Technological Change and the TraditionalCereals in Sub-Saharan Africa: The Malian Case

Jeff Vitale and John H. Sanders

Programming models of farms were incorporatedinto a sector model. The combination of newtechnologies, different demand expansion scenarios,and increases in liquidity were considered. Thefeasibility of these different scenarios and the policyimplications were considered in detail. Finally,different development strategies for traditionalcereals were compared with other investmentoptions with the sector model.

This paper emphasizes empirical results and policyimplications. However, the combination ofprogramming models within a sector modelprovides a strong analytical tool for analyzing actualand potential impacts.

The objectives of this field research were to estimatethe potential returns and constraints to theintroduction of new sorghum and millettechnologies in Mali (West Africa), and to undertakean aggregate analysis of the effects on consumptionand prices from the combination of new technologiesand demand expansion.

A principal problem with traditional food productsis the inelasticity of price with respect to demand. Ingood rainfall years, cereal prices collapse andfarmers quickly lose interest in the introduction ofnew technologies requiring increased inputexpenditures. The new focus on demand driventechnology needs to re-examine traditional foods asnew food products are becoming available and asfeed demand increases for the cereals.


65

While political influence is often assumed forinternational organizations, it is rarely subjected to athorough assessment. This study addresses thisempirical gap by examining the political influenceexerted by the International Plant Genetic ResourcesInstitute (IPGRI) in the international negotiationspertaining to the revision of the InternationalUndertaking for Plant Genetic Resources (IU).

More specifically, the study aimed to assess the levelof political influence attained by IPGRI in therenegotiations of the IU, and to obtain knowledgeabout the general processes by which the Instituteexerts influence. The knowledge gained throughthis assessment would then allow the identificationof potential future courses of action for IPGRI ininternational decision-making.

The evaluation covered a five-year period (1996-2001), representative of the different phases thatmarked the negotiation process. It employed amethodological framework developed expressly toassess the political influence exerted byinternational non-governmental organizations in theframing of the Climate Change and BiologicalDiversity Conventions. This method drawsconclusions from different dimensions of evidence,and is well suited to assessing political influence incomplex decision-making.

This study obtained evidence from IPGRI staffmembers who were involved in the negotiationsabout the expected ways and means by whichIPGRI was thought to have been influential, —theso-called “ego-perception.” These perceptions werethen either validated or negated by the gathering ofother players’ perceptions of the Institute’s influence

(“alter-perception”). In this case, the alter-perceptionwas provided by national delegates and by membersof the secretariat of the FAO Commission on GeneticResources for Food and Agriculture, which managesthe negotiation process. Finally, the validity of theseperceptions was checked through document analysis(“researcher’s analysis”).

Empirical results relating to IPGRI’s politicalinfluence were subsequently linked to a theoreticalframework that provided a basis on which to explainIPGRI’s capacity to exert influence. This theoreticalframework was built on three criteria: (1) that itwould be consistent with the chosen methodology;(2) that it would provide explanatory knowledge onthe processes of influence and fill in the gaps wherethe empirical data was inconclusive; and (3) that itwould take into account the limited capacity of aninternational agricultural research center, as opposedto that of a state, to exert influence in internationalbargaining. A review of the pertinent internationalrelations literature and the framing of relevantvariables provided such a framework.

The results indicate the most effective ways by whichan international agricultural research center can exertinfluence in international policy-making. Theprovision of timely and relevant technical input (thatis, input directly linked to the organization’s domainof expertise) is seen as the prominent means ofachieving influence in such a context. Yet, the resultsalso show that other factors may enable or constrainorganizations like IPGRI in international decision-making. While political neutrality and reliability wereseen as enabling factors, it remained clear that allinternational organizations lack the resources orformal rights that endow states.

An Assessment of IPGRI’s Impact on InternationalPolicy-making: A Case Study of the International Undertaking on

Plant Genetic Resources (1996-2001)

Raphaël Sauvé and Jamie Watts


66

In recent years, much research has been done to assessthe importance of research and development (R&D)and trade in influencing output growth and totalfactor productivity. There is now a large body ofliterature that provides theoretical and empiricalmodels where cumulative R&D is the main engine oftechnological progress and productivity growth. Theempirical evidence has been provided by Coe andHelpman in a seminal paper where they find thataccumulated spending on R&D by a country and itstrade partners helps to explain the growth of totalfactor productivity. R&D investments are still centralto agricultural productivity growth. Alston et al., inthe introduction of their recent book on the theme,underline that “Throughout the twentieth century,improvements in agricultural productivity have beenclosely linked to investments in agricultural R&D andto policies that affect agricultural R&D.”

Given the importance of agricultural R&D to sectorgrowth, many works have been devoted to reportingmeasures of the returns to domestic agricultural R&D,but in a world where the international trade ofagricultural products and the dissemination ofknowledge are widespread, domestic agriculturalproductivity depends not only on domestic R&D butalso on foreign R&D efforts. This point has been fullyrecognized by Hayami and Ruttan, who emphasizedthat a country can acquire substantial gains inagricultural productivity by borrowing advancedtechnology existing in other countries.

Recent works analyze the effects of internationalpublic and/or private agricultural R&D on domesticagricultural productivity growth, finding the presenceof strong international spillovers in the agriculturalsector and that, without recognizing knowledgespillovers, researchers will end with biased estimatesof R&D elasticities. However, the international

transfer of agricultural technology is more difficult thanindustrial technology. Modern agricultural technologyhas mainly been improved in developed countrieslocated in temperate zones. Thus, without appropriateadaptive research that helps to assimilate and exploitexternally available information, countries located inother ecological zones, for example tropical zones, maynot benefit from technological spillovers.

In the next section, we present a theoretical model thatlinks total factor productivity to the cumulativespending on R&D. In the third section, we introduce andreview the recent results on estimation and inference inpanel co-integration. Co-integrating regression enablesus to exploit the relationship among the variables inlevels without transforming the data, such as bydifferencing, to avoid spurious regression problems. Insection four, we estimate a simple Cobb-Douglasproduction function for a sample of 47 countries, duringthe period 1970-1992, by using panel co-integration. Wealso split the sample and estimate two productionfunctions, one for the countries in the sample located intemperate zones, and one for the countries in the samplelocated in tropical zones. The results indicate that bothproduction functions show constant returns to scale butfactor elasticities are quite different. Using these resultsand following Coe and Helpman’s empirical model, weare able to utilize panel co-integrating regression toestimate the relationship between total factorproductivity and domestic as well as foreign R&Dcapital stocks. Using these estimates, we calculate theeffect of change in a country’s R&D spending on thechange of total factor productivity in that country, aswell as in partner countries. In summary, we find strongR&D spillovers between countries located in temperatezones and, inside this group, between EU countries.International spillovers are of less importance whenanalyzing tropical countries.

International R&D Spillovers and Productivity Growth in theAgricultural Sector: A Panel Co-integration Approach

Luciano Gutierrez


67

Bangladesh comprises a small area—147,570 km2 —with a large population (122 million). Rice and wheatare the staple foods of the Bangladeshi people. By theyear 2020, 37 million tons of food grain will berequired for a projected population of 172 million. CanBangladesh sustain the growth momentum, driven bythe development and application of new technologies,which is required to ensure future food security? Thisstudy examined the impact of modern technologyadoption on output growth and the sustainability ofmajor cereal production in Bangladesh. The study areaconsisted of 17 greater districts in Bangladesh,characterized by agro-ecology. A time series databaseon major cereal production was used. The growth rateof major cereal production was measured using acompound growth rate model, while the sustainabilityof modern food grain production technology wasmeasured by estimating total factor productivity (TFP),using the Tornqvist Theil (TT) index. A moment basedproduction function was used to estimate factorsaffecting the sustainability of modern food grainproduction technology. The model was estimated bythe generalized method of moment procedure.

Good progress in the adoption of modern variety (MV)rice was achieved during 1970s and 1980s because ofthe release of high-yielding varieties and governmentsubsidies on fertilizers, pesticides, and irrigation. Overthe last 25 years, since independence, the growth ratesin rice production, acreage and yield were 2.53%,0.33% and 1.20% per annum, respectively. In general,this was achieved via high growth rates in rice yield.However, over the same time period (25 years), the

growth rate of production, area, and yield of MV ricewere 7.99%, 8.72% and -0.73% per annum,respectively. During the 1980s and 1990s the highestgrowth in acreage (9.64%) and production (9.03%) wasachieved for MV wet season rice (Aman) followed byMV winter season (Boro) rice whose growth in areaand production were 8.84% and 8.83% per annum,respectively. Growth in acreage and production of MVsummer rice (Aus) were 7.93% and 5.46% per annum,respectively. In general, a positive growth rate in MVrice production was achieved due to the areaexpansion of MV rice. Yield for MV rice all threeseasons declined. This could be attributed to thedegradation of soil fertility due to intensivecultivation, inappropriate fertilizer application,deficiency of micronutrients in soils, and a generaldeterioration in varietal traits. Since independence, thearea and production of all cereal crops in Bangladeshhas increased. However, with the exception of wheat,yields have decreased. The estimates of TFP indices ofMV Aus, Aman, Boro and wheat indicated thatmodern food grain production technology becameunsustainable after the mid-1980s. It revealed that MVAus production technology was sustainable up to1983, MV Aman technology was sustainable until1985, and MV Boro was sustainable until 1989. MVwheat production technology was sustainable until1984, indicated by a declining trend in TFP indices.The study also identified and quantified the impact oftechnological (fertilizer, pesticides, etc.) andenvironmental factors (rainfall, humidity, etc.) on thesustainability of modern rice and wheat production inBangladesh, using production distribution moments.

Impact of Modern Technology Adoption on Output Growth andSustainability of Major Cereal Production in Bangladesh

S.M. Fakhrul Islam


68

After the mid-1960s throughout the Punjab, India, theGreen Revolution resulted in rapid growth of farmoutput. During the 1980s, growth in paddy fieldproductivity averaged 1.27%/ha, declining to -0.04%/ha in the 1990s. In the case of wheat, there was anincrease in average productivity from 2.73 t/ha in 1981-82 to 3.80 t/ha in 1991-92 and 4.56 t/ha in 2000-01. Thisdemonstrated a CGR of 3.00% in the 1980s, whichdecreased to 1.45% during 1990s. The decomposition ofproductivity growth, made by Kalirajan, Obwona andZhao, has demonstrated the components as technicalchange, technological improvement and input use. Thefirst two components account for total factorproductivity (TFP) growth. Using this model in theIndian Punjab, which has almost exploited theavailable potential of soil, water and other naturalresources required particularly for the rice crop,sustainability, as a component of TFP, was segregatedout. An effort to apportion the contribution ofsustainability as an important determinant of TFPgrowth was made. The decline in TFP, due to a fall inthe output index since early 1990s, has highlighted the

issue of the deterioration of ecological parameters. Inthis high potential rice-wheat belt, resourcedegradation attributed to the rice crop is more seriousthan wheat. Further, the contribution of sustainabilityto the TFP growth is increasing at an alarming rate.Based on the household data collected from 300farmers, over the 1981-1998 period, it is clear that theTFP growth is declining. Deducting their effect fromTFP growth, the negative unaccounted factor is termedas “sustainability”, the value of which came out to -1.11% and -4.90% in case of paddy fields and -0.81%and -3.07% in the case of wheat cropping in the 1980sand 1990s respectively. This calls for immediate policyoptions to arrest the trend by providing alternatives torice cropping in Punjab agriculture. Moreover, farmersengaged in environmentally degrading practices suchas the early transplanting of rice, indiscriminate use ofpesticides, burning of paddy straw, irrational use ofunderground water, and those cultivating paddy cropsin areas of brackish water need to be curtailed throughstrict policy measures.

Component Analysis of Total Factor Productivity in HighPotential Rice-Wheat Systems in the Indian Punjab

Joginder Singh and M. Hossain


69

Approximately one-third of the developing world’swheat area is located in environments that are regardedas marginal for wheat production because of drought,heat, and soil problems. Nearly one-third of the areaplanted to bread wheat and about three-fourths of thearea planted to durum wheat suffer from severedrought stress during the growing season.

Despite these limitations, the world’s dry and difficultcropping environments are increasingly crucial to foodsecurity in the developing world. Worldwide,investment in irrigation infrastructure continues to fall,while population growth and demand for wheat areincreasing. Gains in wheat productivity in marginalenvironments are important because it is unlikely thatincreased productivity in the favorable environmentswill be sufficient to meet the projected growth indemand for wheat from the present to 2020. Thedemand for wheat is projected to be 40% greater thanits current level of 552 million tons by 2020. Improvedproductivity in marginal areas would improve foodsecurity for the poorer populations that live there.

It is widely believed that the Green Revolution hadvery little effect in marginal environments, where theharsh agricultural conditions and slow spread of GreenRevolution technology resulted in very modest yieldgains. For some time, the development community hasbeen concerned about progress in marginal areas andthe level of research resources allocated to those areas.This paper provides new information to address theseissues by answering the question: Is growth in wheatyield potential in marginal environments approachingthe levels attained in favorable environments? Morespecifically, we:� describe breeding research that improved

productivity in marginal environments (with anemphasis on CIMMYT’s wheat breeding strategies);

� estimate rates of growth in wheat yield potential inmarginal and favorable environments;

� examine the crossover and spillover of wheatvarieties from favorable to marginal environments;

Effects of Innovative Wheat Breeding for Marginal Environments

Maximina A Lantican, Prabhu L. Pingali, and Sanjaya Rajaram

� identify implications for wheat productivity growthin marginal environments; and

� discuss future challenges for marginalenvironments.

Data for this study were obtained from CIMMYT’sElite Spring Wheat Yield Trial (ESWYT), grown in 246locations in 65 countries between 1979 and 1999, andfrom CIMMYT’s International Spring Wheat YieldNursery (ISWYN), grown in 411 locations in 82countries between 1964 and 1995. Nurseries such as theISWYN and ESWYT are one way in which breeders indeveloping countries regularly gain access to andexchange a large number of new wheat varieties bredby CIMMYT and partners in national researchprograms. This system of breeding, germplasm andinformation exchange is often referred to as “theinternational wheat research system.” Its impact onwheat yield trends in marginal environments will bediscussed later.

Data on spring wheat varieties planted in 1990 and1997, including their pedigrees, year of release, areaplanted to each variety, and target mega-environment,were obtained from the CIMMYT Wheat Impactsdatabase. A mega-environment (ME) is a broad,frequently transcontinental but not necessarilycontiguous area with similar biotic and abiotic stresses,cropping system requirements, consumer preferences,and potential volume of production. Mega-environments usually encompass more than onecountry and are useful for defining breeding objectives,because each ME comprises millions of hectares thatare relatively homogenous for wheat production.

The ISWYN data were grouped into two periods: theGreen Revolution period (1964-78) and the post-GreenRevolution period (1979-95). The ESWYT data (1979–99) was taken from the latter period. The average of thetop three wheat yields for each location per year wasused in the analysis. Locations were grouped by ME.


70

The general objectives of this study are to estimate therate of adoption of improved bean varieties released inHonduras since 1987, estimate the ex post economicrate of return to the development and adoption ofthese varieties, and investigate the agronomic, marketand consumption characteristics of traditional andimproved bean varieties that help to explain patternsof varietal use by Honduran bean farmers.

Collaborative research on beans in Honduras by theUSAID-funded Bean/Cowpea Collaborative ResearchSupport Project (CRSP), the bean program at EscuelaAgricola Panamericana (Zamorano), PROFRIJOL (aregional network funded by the Swiss Agency forDevelopment and Cooperation, SDC), and DICTA (thenational agricultural research program) has resulted inthe release of many improved varieties since 1987.Data collection for this study included a farmer survey(N=210) in the three main bean-producingdepartments of Honduras (El Paraiso, FranciscoMorazan and Olancho), implemented in January-February 2001. These departments account for over60% of national bean production. In addition, thesurvey team conducted “bean price experiments” withbean traders in the three regions to gain a roughmeasure of price premiums/discounts for beancoloring, size, shape, etc., of improved beans formerlyreleased, as well as some promising experimental lines.

Adoption rates and yields from this survey, as well asfrom a CRSP-funded 1994 survey in the same threedepartments, will be used along with experimentalyield data to estimate the financial and economic ratesof return to the CRSP/ Zamorano/PROFRIJOLinvestments in bean research in Honduras. In addition,sub-sector analysis, combined with farmer varietalpreferences, will help to explain patterns of varietaladoption, non-adoption, and dis-adoption, to suggestpriorities for future bean breeding activities, and torecommend the nature and implementation ofcontinued socio-economic research to support theHonduran bean breeding programs at Zamorano andDICTA.

Martel’s 1994 survey and analysis highlighted thedifferential agronomic and market characteristicsmost valued by different types of farmers in theirbean varieties. For example, Dorado, an improvedvariety with BGMV resistance, enjoyed reasonablepopularity among valley farmers, who saved morerevenue in stable yields (due to BGMV resistance)than they lost from Dorado’s 15% price discount bytraders due to its dark color. On the other hand,BGMV was not as serious a problem for mosthighland farmers, who thus adopted Catrachitainstead—a high-yielding variety with little diseaseresistance but better market acceptance.

In 1997, Zamorano released Tio Canela, an improvedvariety with the disease resistance beyond that ofDorado and an improved (lighter red) color, thusintended to be more desirable to both valley andhighland farmers. The recent survey will thereforegive the Zamorano and DICTA bean programs its firstfeedback on farmers’ reaction to this new release, andthe trader interviews will provide additionalinformation regarding the market acceptance of TioCanela.

Although this survey data is currently beinganalyzed, initial data analysis suggests severalfindings. As anticipated, adoption of Tio Canela is notvery high given that the Honduran seed system iswoefully underdeveloped. More surprisingly, manyfarmers who have experimented with Tio Canela (andeven some who previously grew Dorado) haveswitched back to traditional varieties. This impliesthat; a) the experimental yield advantage of TioCanela does not hold up under farmer conditions; b)Tio Canela receives a significant price discount; c)disease pressure in many areas has changedconsiderably, thus making the resistance embodied byTio Canela less valuable to farmers.

The Impact of Bean Research in Honduras

David Mather, Richard Bernsten, Juan Carlos Rosas, Abelardo Viana, Danilo Escoto, and Julio Martinez


71

This is an exploratory analysis of an extensive studyof attempts to introduce two new peanut varieties totwo farm communities in the Philippines over the1988-89 to 1994-95 (5-year) period. This analysisfocuses on the: (1) extent of impacts, (2)sustainability of impacts, (3) production impacts,and (4) influences on adoption (including possibleinfluences of the farmer, farm, contact network,community, and a national institution).Adoption and impact studies have almost

exclusively focused on adoption at the end of anextension effort and on the characteristics of the farmand farmer in the target communities. This focus hasled to the charge of blaming the victim if there is littleimpact. This long-term research project intends to drawon context variables in order to identify community,sub-community and contact network constraints andfacilitators, which theory suggests will have greatereffects on adoption, impacts and sustainability.

A Study of Philippine Peanut Farming Communities: Impacts of NewPeanut CRSP Technology and Influences on Sustainability

Robert L., Moxley, Aida Librero, and David Alston


72

The modern maize varieties which eventually maketheir way into farmers’ fields are products of aninternational breeding system that includes theInternational Maize and Wheat Improvement Center(CIMMYT), several national breeding programs, and amultitude of private, national and multinational, seedcompanies. During the period 1998/99, 179 public andprivate seed companies were interviewed. Thesecompanies were selected due to their involvement inmaize breeding research and the production anddistribution of maize seed in seven Asian countries:China, India, Indonesia, Nepal, Philippines, Thailandand Vietnam. This study evaluated four principal facetsof seed company performance, namely: (a) the level ofinvestment in maize breeding research; (b) germplasmoutputs; (c) the nature and extent of roles played in themaize seed industry; and (d) the rate of farm-leveladoption of improved maize germplasm.

During 1995-97, Asia produced 152 million tons ofmaize from 42 million ha, yielding an average 3.6 t/ha.Since the 1960s, increases in maize production havebeen fueled by yield gains rather than by areaexpansion. Yield gains, in turn, have been due to theshift in maize cultivation from predominantly openpollinated varieties to hybrids. From a base area of 2.4million ha in 1990, hybrid maize production grew to 8.7million ha, or 45% of the total maize area in Asia, in1998. The shift to hybrid maize cultivation alsotransferred the locus of modern maize breeding

research from government research organizations tonational and multinational private-sector seedcompanies. In countries where both public and privatesectors participated in maize research, private sectorresearch investment in maize research far exceededthat of the public sector. In 1998, primarily due to moreaggressive marketing programs, private sector maizeseed sales grew by 24%, while those of the publicsector grew by only 2%.

In conclusion, public seed research agencies (includinguniversities and cooperatives) have: tended to developmore open-pollinated maize varieties (OPVs) thanhybrids; mass-produced and distributed seed cheaply;addressed location-specific production problems and;provided agricultural extension services. Conversely,private seed companies have developed and marketedtheir own proprietary hybrids. In recent years, thepublic sector has faced two important challenges: anexpanding maize industry that demands improvedtechnologies; and inadequate and declining publicresources that, in turn, limit technology developmentand dissemination activities. Amidst these challenges,governments have encouraged the participation of theprivate sector by improving the business environment.At present, the public and private sectors in Asia’smaize seed industry are linked in knowledge andtechnology dissemination via human resourcedevelopment and cooperative experimental trials.

Impact of Public Sector and Private Sector in R&D andTechnology Generation: The Case of Maize in Asia

Roberta V. Gerpacio


73

Not many national agricultural research institutionsin developing countries are sufficiently orientedtowards generating development impacts. Observersexplain the problem by the way research institutesoperate, in particular by organizational variablessuch as the lack of capability in planning and M&E,and the feeble influence that farmers have onresearch programs. This paper argues that part of thesyndrome is the poor knowledge of publicresearchers about the actual significance of theirwork for agricultural development. Manyagricultural policy makers and researchersmisconceive the innovation processes in agricultureby assuming that research is the only source oftechnical change and that productivity change canalways be put down to preceeding research results.

Their assumption has consequences. One is that onlyfew research institutions make efforts to analyzetheir impact. If they do, most are content withmeasuring the change in technology use or in theproduction figures of the respective commodity,expecting that technical change will be favorable forrural development anyway. As a consequence,research institutes often do not see the necessity tostudy agricultural innovation and their own role in itmore thoroughly.

This is in sharp contrast to the widely sharedrecognition of the complexity of agriculturaldevelopment and the increasing demand for agreater accountability of public developmentagencies. In fact, the causal chain between aninvestment in Research and Development (R&D)projects and agricultural change is quite long. As theanalysis moves away from the origin—research—anddown along the assumed succession of effects, theobserved changes become less and less attributableto the initial research investment. The variety ofactors increases and hence the possibility ofconflicting views on the ongoing change and the

factors driving it. To this, add that we observednegative trends as well; e.g. the depletion of naturalresources or the dwindling biodiversity. Claimingpositive research “impacts” without a serious effortto understand the real change process is less and lesscredible—and it prevents research institutes learningfrom previous experience and becoming moredevelopment-minded.

The objective of impact assessment is a greaterimpact orientation of research. Research institutionsclearly need to strengthen their impact orientation ifthey are to reverse the downward trend in funding.The term “impact orientation” characterizes aresearch organization that is focused on impact,uniting the different approaches to improve researchunder the perspective of their contribution toachieving impact. Impact orientation expresses itselfin such features as a good comprehension of howimpact is achieved, sufficient organizational capacity,intensive communication with all stakeholders, clientand service orientation, and the use of planning andM&E techniques. Important aspects of an impactorientation are to better understand the processes ofagricultural change, to clarify hypotheses along theimpact pathway, and to critically assess them.

An alternative approach to understanding researchimpact is the pathway concept. The paper introducesthe concept of impact pathways in order to betterconceptualize the processes by which researchachieves development impacts. This is not anotherimpact assessment method but rather a means tointegrate and organize information so that the role ofresearch in development becomes more transparent.An impact pathway describes the chain of eventslinking research and agricultural development. Itestablishes a series of measurement points; e.g.,research input, research output, rates of technologyuse, production figures, indicators of social change. Itnames the change processes that are supposed to link

Winding up the Impact Pathway: An Approach to Strengthen theImpact Orientation of National Agricultural Research

Andreas Springer-Heinze, Frank Hartwich, Simon Henderson, Douglas Horton, and Isaac Minde


74

these points. At least three processes may bedistinguished: (1) the research process leading to theresearch outputs, (2) the agricultural innovationprocess leading to technical change, and (3) theagricultural (rural) development process leading toeconomic and social change. By combining measurableindicators with process analyses, pathway analysis canhelp to address the attribution problem and takeaccount of the many factors, other than research, thatmay be driving change. The three different processesmay be studied separately while maintaining acomprehensive view on the way impact is generated.Thus, pathway analysis can reveal opportunities forresearch as well as identify problem situations in which

it would be pointless to conduct research as a result of,say, the conditions in commodity markets andinfrastructure or the availability of complementaryservices. Pathway analyses can also help to generateplausible arguments demonstrating how researchactivities actually feed into the development process.The paper presents ideas and examples generatedduring a series of workshops where a regional project“Strengthening Impact Orientation in AgriculturalResearch in Eastern and Central Africa” was plannedby representatives of the “Association forStrengthening Agricultural Research in East andCentral Africa (ASARECA)” with support from theGerman Development Co-operation (GTZ).


75

A significant number of agricultural impactassessment studies have been conducted during thelast three decades. These studies have demonstratedthe considerable contributions of agriculturalresearch towards improved productivity, profitability,and the sustainability of agribusinesses. Brazil is oneof the countries where this type of study wasfrequently utilized. EMBRAPA, a governmentresearch company, conducts the majority of impactassessment studies in this country. The mostfrequently posed questions are, how useful is thistype of research and how does it impact onagricultural research funding and researchinstitutions themselves? The principal aim of thispaper is to present clear evidence that thedevelopment and use of impact assessment studies inBrazil predominantly served to strengthen theposition of EMBRAPA.

Brazilian impact studies began in the early 1980s inresponse to demands made by Brazilian society, whowere interested in the efficacy of significant stateinvestments in agricultural research. Initial studiesdemonstrated that funding from the national treasuryand international loans (IDB and World Bank)invested in infrastructure and human capital in theEMBRAPA’s research centers generated high returns.

During the next two decades, the Braziliangovernment continued to invest in EMBRAPA. Thisallowed the institution to develop a modern,nationwide infrastructure of agricultural research andto adjust the human capital profile dedicated toresearch activities. In 1974, only 17% of researchershad MSc and PhD degrees. In less than 10 years, thisfigure changed to 75%. In reality, less than 5% ofEMBRAPA’s staff lack postgraduate training.Calculated in 1993 dollars, resources allocated toEMBRAPA increased from an average of US$245million in 1976-80 to more than US$450 millionduring the 1990s.

Continuous public support allowed EMBRAPA tooffer competitive salaries, maintain a well-trainedresearch staff, and provide good conditions in whichstaff could develop their research activities. Thiscontext was fundamental for the institution. It enableda continuous flow of technological outputs that haveprogressively been integrated into Brazilianagriculture. EMBRAPA was continually asked todemonstrate the efficacy of high state investments.

The importance of EMBRAPA’s role in thedevelopment of agricultural technology has beensystematically demonstrated through impactassessment studies conducted by EMBRAPA and otherBrazilian research institutions. EMBRAPA hasunderwritten the validity of its impacts research bystressing the close collaboration of Yale University’sInternational Food Policy Research Institute and theUniversity of California-Davis in the development ofimpact assessments. The use of foreign researchers,with international experience and using diversifiedmethodological approaches, has played a vitalsupporting role in EMBRAPA’s accountability reports.

Evidence to support EMBRAPA’s important role in themodernization of the agribusiness sector can also befound in annual progress reports where the role ofcrop varieties in the seed market and othertechnologies are clearly shown. Aggregate studiesdemonstrating the link between agricultural research,increased exports, and agricultural technologicalchange have also been contracted to well knowneconomists with influence over the Brazilian decision-making process. A further kind of impact assessmentis the annual social balance. Here, EMBRAPAdemonstrates that its technological outputs are alsogenerating social benefits.

The Importance of Impact Assessment Studies for the BrazilianAgricultural Research System

Antonio Flavio Dias Avila and Geraldo da Silva e Souza


76

Another aspect of the EMBRAPA policy is the adoptionof an integrated evaluation and award system based onresults. All its research centers have been evaluatedsince 1996. Evaluations are based on economicefficiency (DEA – Data Envelopment Analysis),research funding grants, institutional image,technological impact and research quality. The bestteams and employees are awarded annually. As thisaward is not incorporated into staff salaries and variesaccording to performance, it has stimulated a generalincrease in institutional productivity and effectivenessof research centers.

Finally, it should be noted that impact assessmentstudies are still making a difference in EMBRAPA. Thisis due to the adoption, in the mid-1990s, of anaggressive policy of communication and marketing.Economic, social, and environmental impacts of

agricultural research outputs have been effectivelycommunicated to congress and the government and,more recently, to the general public. EMBRAPA’smarketing policy clearly recognizes the role ofpartnerships, state research organizations, universities,CGIAR centers, and the private sector in agriculturaltechnology development.

If impact assessment studies are being conducted butare not making a difference, there are two possibleexplanations—either impact assessments are not beingconducted effectively or assessment results are notbeing communicated effectively to decision-makers. AtEMBRAPA, there is strong evidence that impactassessments are important and are making a realdifference. They are also strengthening EMBRAPA’sinstitutional image within Brazilian society.


77

Most potato-farming families in the Carchi Province ofEcuador use the highly toxic insecticides carbofuranand methamidophos. Exposure to these neurotoxinscauses health problems and productivity losses. Anintegrated research and intervention project, financedby the innovative IDRC–Ecosystem Health Program,aimed to reduce that exposure, whilst maintainingand, in some cases, enhancing farm productivity. Thispaper analyzes the results of a follow-up interventionproject and survey that incorporates a longitudinalstudy to measure changes in neurobehavioralperformance of farm families before and afterexposure is reduced. The interventions are designed toinfluence social units, ranging from the farm family tothe international community. The integration ofparticipatory intervention techniques and sample-based quantitative research created tensions for themultidisciplinary team and demanded compromisesthat, on balance, have been positive.

Drawing on risk reduction principles of industrialhygiene, our intervention program targets variouslevels of social aggregation with diverse strategies. Atthe farm and family levels, we have utilized individualcounseling and farmer field schools for enhancingcapacities in integrated pest management (IPM) andpromoting safer pesticide use. Following trainingactivities, which contributed to pesticide reductions,productivity increases and heightened awareness ofthe negative consequences of pesticides, farmers haveshown a remarkable willingness to invest in personalprotective equipment. At the community level, weworked with women’s groups, primary schools, andcommunity education campaigns. One particularlyeffective intervention was an action-research tracerstudy using fluorescent dyes to identify pesticideexposure pathways. At the regional level, we heldpublic forums to inform and develop policyrecommendations and used radio spots to increase

public awareness. At the national level, together withthe ministries of agriculture, health, and theenvironment, we supported a policy formulation andlobbying process that included information sessionswith ministries and the pesticide industry, a publicforum, and media events. Furthermore, we arecoordinating with national programs throughout theAndes, as well as the Food and AgricultureOrganization and the Latin American Pesticide ActionNetwork, for regional-level advocacy. While the projecthas generated promising local results, the broadercontext will ultimately prevent lasting impact. Effortsto promote national-level policy intervention havearguably been futile due to internal political pressuresthat prevent rational policy formulation. As a result,the project has begun to use international support toprovide further pressure for change.

The combination of quantitative methods needed forrigorous economic and epidemiological analysis, andparticipatory methods needed for community-basedaction, created unique conflicts in disciplinary andmethodological design. The rhythms of research andparticipatory intervention move at different cadences,calling for adjustments in both. Other conflicts include;participant-beneficiary control over activities, farmfamily sample identification and maintenance,burdening participants with tedious interviews, andscheduling measurements, interventions, andcommunications.

Impact assessment research is often an ex-post,conducted long after the principals are dispersed.Consistent follow-up is key to obtaining impact fromimpact assessment. When integrated into an ongoingprogram, impact assessment provides mid-courseguidance that can justify follow-up by actors that arestill present.

Potato Production and Pesticide Use in Ecuador:Linking Impact Assessment Research and Rural Development

Intervention for Greater Eco-System Health

Charles Crissman, Donald Cole, Steve Sherwood, Patricio Espinosa and David Yanggen


78

A long-standing debate within the CGIAR systemconcerns the effects on various populations(particularly the poor) of different allocations ofresearch effort between marginal and favoredproduction environments. Some argue that there hasbeen systematic under-investment in marginalproduction environments, to the detriment of the largegroup of impoverished people within those areas.Others counter that historically: investment in marginalareas has been low precisely because the returns tothose investments are low, and that diverting researchresources away from favored production environmentswould, overall, do more harm than good.

In Kenya, this debate is critically important for severalreasons. First, agriculture is the dominant sector in theeconomy, accounting for 28-30% of GDP. Second, thecountry has one of the fastest growing populations inthe world, which puts considerable pressure on arableland (20% of total land area) to produce sufficient food.The consequence of this has been reduced fallowperiods, fewer crop rotation options, and loss of soilfertility and land productivity. Third, and partly due tothe rapidly growing population, there has been notableout-migration from high-potential to low-potentialagro-ecological zones, with an accompanying increasein the importance of agricultural production on less-favored lands. This transformation of the spatialdistribution of production has serious implications forboth agricultural research and the environment. Finally,Kenya’s economy has been on the downturn for thepast two decades, resulting in severe reduction inavailable resources for agriculture research.

Past increases in maize production were fueled by thedevelopment of high yielding varieties suitable for arange of agro-ecological zones, as well as an increase in

the area under maize cultivation. However, futureproductivity gains are likely to rely more on theformer than the later. While maize productivitygrowth has declined since the mid-1970s, a wide gapseparates experiment station yields from thoseachieved in farmers’ fields. This indicates thatsignificant productivity gains could be achievedthrough better targeting and promotion for adoptionof improved technologies.

This study uses a multi-market model of maizeproduction to assess the potential impact ofimproved maize technology on incomes and welfareof both rural and urban households in Kenya. Weanalyze the likely impacts on various householdtypes of the diffusion of improved maize varietiesand crop management technologies that are currently“on the shelf.” The direct effects of technical changeare based on assessments made by experts in theKenyan agricultural research system. The modelcomputes the indirect effects that are transmittedthrough product and factor markets via endogenouschanges in output prices and input prices.

The model is disaggregated into six distinct agro-ecological production zones. This allows us toinvestigate alternative technology adoption scenarios(e.g., a “balanced diffusion” scenario) in whichtechnical change occurs in all areas, versus scenariosin which diffusion is confined to either favored ormarginal production environments. Our resultsindicate that on both efficiency and equity grounds,the most desirable outcomes include targeting ofagricultural technologies toward more favored agro-ecological zones.

The Welfare Effects of Maize Technologies in Marginaland High-Potential Regions of Kenya

Daniel D. Karanja and Mitch Renkow


79

Agricultural Research and Urban Poverty: The Case of China

Shenggen Fan, Cheng Fang and Xiaobo Zhang

This paper develops a framework to measure theimpact of agricultural research on urban poverty.Increased investment in agricultural research anddevelopment (R&D) will lower food prices throughincreased food production. Lower food prices will, in


turn, help the urban poor, as they often spend morethan 60% of their income on food. Using China as acase, the study empirically estimated these effects.These effects are large, and are comparable to those onrural poverty.

80

Modern varieties (MV) cover approximately 75% ofAsian rice lands, with 55% of rice production occurringin irrigated ecosystems. However, a significant portionof MV rice is grown under rain-fed conditions withunfavorable growing environments. Good watercontrol is considered a precondition for realizing theproductivity and profitability potentials of modern ricevarieties. This paper tests the hypothesis that theadoption of modern varieties under rain-fedconditions would not increase yields enough tocompensate for additional input costs, and would onlyhave a marginal impact on profitability gains andhousehold incomes compared to substantial gains forthe irrigated ecosystem.

The data for this paper are drawn from householdlevel sample surveys conducted by the Social SciencesDivision of the International Rice Research Institute(IRRI) in collaboration with national agriculturalresearch system partners in Bangladesh, in the states ofBihar and Chhatisgarh in India, and in the Philippinesand Vietnam during the late 1990s. The villages wereselected to represent different rice growingenvironments. All households in selected villages werecovered by the survey except for Bangladesh andVietnam, where a sample of the households wasselected using stratified random sampling.

The primary data are used to estimate unit costs, laborproductivity, and household incomes, and the factorshares of land, labor, and capital in the production ofmodern varieties under irrigated and rain-fedconditions. A multivariate regression model ofdeterminants of income is estimated to identify theeffect of technology, factor endowments, andenvironmental factors on incomes from rice cultivation,as well as total household incomes. Two interactionterms of the coverage of modern varieties andirrigation are included in the model for rice incomes tomeasure the separate effects of the adoption of modernvarieties in rain-fed and irrigated areas. The model isestimated for total household incomes to assess the neteffect, as the adoption of modern varieties may lead todiversion of household resources from non-rice andnon-farm economic activities to rice cultivation. A GiniDecomposition Analysis is conducted to assess theincome distribution effect associated with the adoptionof MV rice for the irrigated and the rain-fedecosystems. Results are discussed in this paper.

Impact of the Adoption of Modern Varieties of Rice on Productivity Gainsand Income Distribution for the Irrigated and the Rain-fed Ecosystem

Mahabub Hossain, Manik Lal Bose, Tran Thi Ut, A.G. Agarwal, Jawhar Thakur, and Esther B. Marciano


81

Scientists are increasingly being challenged to providepolicy makers with information on the trade-offsbetween environmental and economic developmentgoals. There is a need for new approaches forassessing potential impacts of interventions thatcombine scientific knowledge with empiricalinformation on constraints of land users. In this paperwe use a custom-built, economic-ecological simulationmodel to examine the impact of existing and proposedsoil management strategies on farm productivity,profitability, and sustainability for smallholder farmersin western Kenya. The model is applied to threerepresentative farm types, which were developedusing wealth ranking and other participatorytechniques to reflect the differences in resourceendowments and constraints faced by farmers inVihiga District of western Kenya.

The model is first used to examine the impact ofexisting soil management strategies. Results show theimportance of differentiating farm types by importantcharacteristics that impact opportunities andmanagement. The current soil management practicesof high resource endowment farms are productive,profitable, and ecologically sustainable. Low andmedium resource endowment farms, on the otherhand, use techniques with much lower levels ofproductivity and profitability that result in a decliningsoil resource base.

The model is then applied to analyze the potentialimpacts of improved soil fertility management, usingvarious sources of phosphorous (P) and nitrogen (N).Results indicate that high resource endowment farmersare already using near optimal management practices,and new management techniques have little impact onyields or other indicators. Low and medium resourceendowment farmers, on the other hand, can greatlyimprove farm profitability and sustainability throughimproved management of P and N, but low land andcapital resources constrain the adoption of theimproved practices. Targeting interventions to low andmedium resource endowments will, therefore, havemuch larger payoffs.

Model results—as well as extensive discussions withfarmers during the participatory data collectionexercises—suggest that most farmers are aware ofbetter soil fertility management practices but are notable to apply them due to lack of resources. When cashis available, investment in soil fertility must competewith other needs and investments such as education. Itis likely that high resource endowment farmers onlybegan to invest in soil fertility when other needs andhigher-return investments (such as educating children)were already satisfied.

Household Resource Endowments and the Impactsof Soil Fertility Management

Meredith J. Soule and Keith D. Shepherd


82

In Kenya, as in most of East and Southern Africa, maizeis the primary staple. However, the country has yet tobecome self-sufficient for maize, and current increasesin productivity fall short of population growth.Throughout the region, pre-harvest losses due to stemborers are estimated by farmers to range around 15%.The identification of maize varieties with seed-basedinsect tolerance has been an ongoing focus of theKenyan Agricultural Research Institute (KARI) since itsinception in 1979. The Insect Resistant Maize for Africa(IRMA) Project, a partnership between CIMMYT andKARI funded by the Novartis Foundation forSustainable Development, was initiated in 1999. Its aimwas to increase maize production and food securitythrough the development and deployment of insectresistant maize, both through conventional breedingand through the use of lines transformed with toxingenes from the entomopathogenic bacteria Bacillusthuringiensis (Bt). Whereas genetically modified insectresistant maize has been grown widely in the US since1996, controversy, public opposition, and regulatoryconfusion have characterized the history of Bt maize inEurope. Although Kenya is ahead of most Africannations in the adoption of the technology (with thethird GM crop currently undergoing assessment priorto the import of germplasm), policies on biosafety andbiotechnology in general are still in the early stages ofdevelopment, and public awareness is minimal.Through the provision of practical experience,workshops, and continuous dialogue withstakeholders, the IRMA project aims to raise public

awareness of the issues surrounding the technologyand to build capacity among local institutions inbiosafety and biotechnology policy as encapsulated inArticle 22 of the Cartagena Protocol on Biosafety.

IRMA represents the first case in Kenya where non-target effects, genetic erosion, and insect resistancemanagement are all to be assessed prior to the releaseof the crop. The project is also unique in itsincorporation of socio-economic studies in theassessment of the technology, addressing equity,market demand, and intellectual property-rights issuesto equip Kenyan scientists, administrators, and policymakers with the full range of tools required fortechnology assessment. This paper analyses theapproaches used to assess the impacts of Bt maize inthe ecological and socio-economic realms, and IRMA’ssubsequent impacts on Kenyan biotechnology policyand public awareness, especially relating to GMOs.

Whereas previous attempts at project impactassessment may have been retrospective, coincidingwith or following interventions, IRMA’s prognosticimpact assessment work will determine whether arelease policy is to be pursued and, if so, will continueafter commercialization. This predictive andmonitoring approach is especially vital with theintroduction of a new technology that is under intensescrutiny from national and international observers.This case study provides a model for projects dealingwith the introduction of biotechnology products in thedeveloping world.

Socio-Economic, Ecological, and Policy Impact Assessment in theIntroduction of a Transgenic Staple Crop Variety to the Developing World:

The Insect Resistant Maize for Africa Project, Kenya

Adrian Ely, Hugo de Groote, Steven Mugo


83

Most farmers in Zimbabwe are smallholders whoseproduction systems are predominately subsistencebased and in which maize accounts for over 65% of thecultivated area. In Zimbabwe, horticultural productionhas grown steadily, becoming an important additionalsource of income for smallholders, and farmers arebeing encouraged to diversify to the production of highvalue crops. Tomatoes account for about 60% of theland area allocated to horticultural crops.

Tomato production is vulnerable to pests and diseaseoutbreaks. Currently, pest management in tomatoproduction is characterized by a heavy dependence onchemical pesticides. These are viewed as a quick andeasy solution to pest problems. In Zimbabwe and manydeveloping countries, chemical pesticides havereceived much government support as a means ofreducing crop losses. However, mounting evidencepoints up the negative effects of chemical pesticides onhuman health and the environment. Toxic substancescan accumulate in the ecosystem and have adetrimental effect on non-target organisms. Integratedpest management (IPM) promotes the use of all known(biological and cultural) environmentally benign pestcontrol measures. Farmers are encouraged to integratethe various methods so that chemicals are usedminimally and judiciously. In Zimbabwe, IPM has notbeen widely promoted. Given the increasing cost ofagricultural inputs in Zimbabwe, there could be scopefor wider adoption of IPM.

This study presents a preliminary assessment of theimpact of IPM technology on farmers’ pest controlpractices, perceptions of chemical pesticides, andknowledge of non-chemical pest control alternatives.

The study involved a comparative analysis of 80 non-IPM trained smallholder tomato growers and 50 IPMtrained smallholder tomato growers. The farmers weresurveyed in 1999 in Chinamora communal area, ahorticulture zone 50 km north of Harare.

The study findings indicate some promisingimplications for wider adoption of IPM by smallholderfarmers. IPM training had a positive influence onfarmers’ knowledge of pests and of the health hazardsof chemical pesticides, and led to a reduction inchemical pesticides use. About 92% of the non-IPMtrained farmers—compared to only 30% of the IPMtrained farmers—used chemical pesticides. All IPMtrained farmers knew of the five major pests oftomatoes, compared with only 75% of the non-IPMtrained farmers. More IPM trained farmers knew andused alternatives to chemical pesticides to controlpests. However, neither the IPM trained nor the non-IPM trained farmers knew how to estimate theeconomic threshold level of pest infestation. IPMtrained farmers anticipated yield losses of 60% due topest damage compared to 95% perceived by non-IPMtrained farmers. IPM trained farmers were more awareof both the acute and the chronic illnesses of chemicalpesticides than non-IPM trained farmers: 80% of IPMtrained farmers compared to 5% of the non-IPMtrained farmers were aware.

Given the findings, policy makers should encouragethe use of a pest management strategy like IPM that isinformation based. This will improve the smallholders’effective use of chemical pesticides, increasing theirprofitability, and will increase smallholder farmers’awareness of the health hazards of chemical pesticides.

A Farm-level Evaluation of the Impact of IPM on Pesticide Use:A Comparative Analysis of IPM and non-IPM Trained Farmers in

Zimbabwe’s Smallholder Sector

S. Siziba and M. Mekuria


84

IITA has a long experience with biological control inAfrica. Since the widely publicized success of theAfrica-wide biological control of the cassava mealybug,successful classical biological control projects havebeen undertaken with the same partners against themango mealy bug, the cassava green mite, and waterhyacinth. The four projects documented the impacts ofthe introduced, exotic natural enemies on the invadingpest species and the food web consisting of indigenousspecies. In a second step, the impact on cropproduction and the amount of crop loss avoidedthrough biological controls were quantified andtranslated into economic gains for farmers and/orconsumers. In a fifth project, an indigenousentomophagous fungus of locusts and grasshopperswas developed into a commercially availablemycopesticide, and the economic viability of theproduct was evaluated both in terms of avoidance ofcrop loss as well as environmental and health hazards,compared to the usual insecticide treatments. Theseprojects form the centerpieces of successful integratedpest management (IPM) projects.

The search for natural enemies in the original home ofthe pests, as well as their importation, rearing, releaseand monitoring, was financed by donor agencies. Allactivities in Africa were executed in collaboration withthe national quarantine and research organizations,which were prepared for releases and monitoringthrough workshops and conferences, and whoparticipated as full scientific collaborators. This left acadre of well-educated scientists capable ofcollaborating in subsequent projects, executing projectson their own, and influencing public opinion. In eachcase, collaboration between entomologists,pathologists, and economists proved highly fruitfuland the ex-post economic analysis revealed highreturns on investment by the donors, with directimpact at the farm level. Benefits to the environment

(compared to pesticide treatments) is usuallydocumented with anecdotal evidence, but notcalculated in monetary terms. Similarly, nationalprograms benefited in terms of human resources,capital, and capacity building, but these impacts werenot systematically analyzed.

This paper first presents an overview of the ex-posteconomic analyses of the different projects and asynthesis of the ecological impacts, and is a firstattempt to quantify the impact on national programs.It then analyses how the documentation of thoseimpacts influenced decision-making in research andthe development of pest management. It is arguedthat the importance of biological control is nowwidely felt within the national and internationalresearch community, but that it did not have the sameimpact on the general public, political decisionmakers and, especially, the donor community. Thoughthe documented biological control projects presentvery sustainable solutions and are clearly highlyprofitable, their profile with donor agencies is ratherlow. Some of the reasons presented here are the lackof guarantee for a sustainable solution from theoutset, the long timeframe required, the limitedpotential for collaboration with the private sector, andthe lack of visibility and related poor publicawareness. For these reasons, biological control andIPM are often not given their share of recognitionwithin the CGIAR system and the donor community.

It is therefore recommended that biological controland IPM projects include an impact assessmentcomponent from the beginning. This will help toguide the development of the technology anddocument the potential impacts from the start.Furthermore, effort is needed to bring the impact ofbiological control to the attention of the general publicand the donor community.

Impact Assessment of Biological Control in Africa—20 YearsExperience of the International Institute of Tropical Agriculture

Peter Neuenschwander, Hugo de Groote, and Boru Douthwaite


85

This paper analyses the impact of the popular FarmerField School approach implemented by Thailand’snational program on integrated pest managementapplied selectively in rice, in five provinces.

Data were collected through a standardizedquestionnaire following the classic “double deltaapproach,” whereby three groups of farmers wereinterviewed before and after training. The groups were(1) farmers participating in the training, (2) farmers notparticipating in the training but living in the samevillage (to measure spread effects), and (3) farmers in acontrol village. Farmers were interviewed during theirrigated rice-cropping season 1999-2000 for the firsttime and again in February 2001. All trainingparticipants were interviewed. In addition, ten non-participating farmers in the “training village” andfifteen non-participating farmers in the control villagewere also interviewed.

The analysis was performed in two steps. First, thefactors affecting dropout from training were analyzed,because not all farmers completed the season-longtraining sessions. The analysis was performed by

applying a multinomial logit model. It was found thatfarmers’ level of pesticide use, the quality of training,farmers’ a priori knowledge on pest and cropmanagement, and the opportunity cost of labor are themain factors determining dropout.

In a second step, the income effects of farmer trainingand of farmer-to-farmer knowledge transfer aremeasured by applying an econometric adoptionmodel. Although the analysis is not yet complete,results indicate that impact varies according to almostthe same factors as those explaining dropout.

The paper concludes that to make the concept ofFarmer Field Schools an effective and efficientextension tool, these schools have to be well targetedrather than randomly spread over a large number ofrural villages. Results also indicate that there are otherreasons besides economic ones that determinefarmers’ participation. This suggests that furtherstudies should investigate the existence of non-marketbenefits of participatory extension approaches.

A Socioeconomic Analysis of Farmer Field Schools Implemented by theNational Program in Integrated Pest Management in Thailand

Suwanna Praneetvatakul and Hermann Waibel


86

The International Water Management Institute(IWMI) established its Southeast Asia regional officein Bangkok, Thailand, in April 2001. It absorbed theprograms and staff of the International Board for SoilResearch and Management (IBSRAM), which ceasedto exist upon the merger. The merger carried with it awealth of experience on the networking approach,which was IBSRAM’s major strategy. Activities hadbeen based on joint identification of soil managementresearch priorities with the national research systemsand the establishment of regional networks toaddress problems identified.

The ASIALAND Network on the Management ofSloping Lands started in 1988 with initial fundingfrom the Asian Development Bank (ADB) andcontinued with support from the Swiss Agency forDevelopment Cooperation (SDC). The fourth phaseof the project was completed in August 2001. Thefifth and final three-year phase began immediatelyand is expected to end in 2004.

While it is difficult to clearly delineate the contributionof the network approach in affecting institutionalchange in the participating research systems, morethan 13 years of network experience have yieldedmany benefits. Through involvement in theimplementation of network research, attendingtraining programs and workshops, having betteraccess to information, improved facilities andadditional resources, many national systems haveimproved their capabilities and enhanced theirdecision-making. This has resulted in the reorientationof research priorities and funding allocations, theintegration of sustainable land management in nationalpolicies and research plans, the narrowing of the gapbetween research and extension, and more activeparticipation in regional and global research activities.In most instances, national systems have integrated theconcerns for environmental and natural resourcemanagement into their institutional work plans.

Network Approach in Soil Management Research:IWMI’s Experience in Southeast Asia

Amado R. Maglinao, Djoko Santoso, and Frits Penning


87

This paper reviews the impact of the RockefellerFunded Forum for Natural Resources ManagementProgram in East and Southern Africa since the early1990s. Since the early 1990s, the RockefellerFoundation has been funding Masters-level trainingthrough grants to principal investigators inuniversities in Kenya, Malawi, Uganda, Zimbabwe,and other nations of East and Southern Africa. Severalprincipal investigators, mainly from agriculturalcolleges in these countries, have benefited from thesegrants through a competitive grant scheme.

The major theme of the research work carried out bythese students, together with their principalinvestigators and collaborators in both East andSouthern African, has been household food security.

The major cause of household food insecurity in sub-Saharan Africa has been declining soil fertility.Following the effects of Structural AdjustmentPrograms (SAPs), which have among other thingsinvolved removal of fertilizer subsidies, the majorityof smallholder farmers are failing to manage thedeclining soil fertility through increased use ofinorganic fertilizers. In other words, the adoption of a‘Green Revolution’ package has been impossible forthe majority of smallholder farmers. This has led to aparallel decline in the productivity of the maize-based farming systems, the major staple for themajority of these countries. The RockefellerFoundation has therefore funded research in thesecountries that has tried to develop alternativetechnologies for managing soil fertility.

The Impact of Rockefeller Funded Forum for Natural ResourcesManagement Program in East and Southern Africa

Bharati K Patel, Paul Woomer and Alexander R. Phiri


88

In 1965, the Indian Council of Agricultural Research(ICAR) initiated the All India Coordinated RiceImprovement Program (AICRIP). Its principal aimwas to facilitate the free flow of information andbreeding material, and to organize multi-locationtesting of improved germplasm across Indian states.AICRIP involves 102 research centers belonging tostate agricultural universities (SAUs) and a few ICARinstitutions at the national level. AICRIP collaborateswith the International Rice Research Institute (IRRI)in the collection, evaluation, and utilization ofimproved germplasm.

The main objective of this paper is to assess adoptionand productivity impacts of genetically improved riceinitiated through the AICRIP in India. It alsoexamines farmers’ preferences for different traits byrelating the rate of adoption of modern varieties(MVs) with their characteristics using a multi-variateregression model.Data on released MVs (1965-2000), namely pedigree,source of origin and variety traits, were obtained fromresearch reports/bulletins of AICRIP and IRRI. Theyield and area planted for all MVs adopted during1998-2000 in 15 major rice-growing states (accountingfor 98% of India’s rice area and production) wereobtained from the state department of agriculture. Agenetic improvement index was computed and itssources were broken down into shares by the nationalsystem and IRRI by assigning weights for the sourcesof parent materials. The area planted to different MVsand their yields were used as weights to assess theadoption and productivity impacts.

Farmers in different states of India adopted 309 of thetotal 620 MVs released. These covered approximately81% of the total rice area during 1998-2000. Amongthem, 30 MVs were adopted in more than one state,occupying 61% of total MV area. About 64% of overallgenetic improvement for the released MVs in India wascontributed by improved germplasm of the nationalsystem, while the remaining 36% came frominternational spillovers.

Nearly 42% of adopted MVs originated from andmoved through the AICRIP network from one state toanother, covering 41% of the total MV area. The directreleases of IRRI’s material covered about 12% of totalrice area in India. The improved germplasm of thenational system contributed about 53% of overallproductivity gains while the remaining 47% derivedfrom research spillovers of international germsplasm(IRRI). The eastern Indian states, which are dominatedby rainfed ecosystems, are the major gainers of theAICRIP and IRRI-induced productivity gains. The stateof Andhra Pradesh is the major contributor with a shareof 50% of the AICRIP-induced productivity gains. Thedistribution pattern of these productivity gains amongthe ecosystems and states is discussed in this paper.

The estimates of the variety adoption model indicatethat yield, grain quality, and maturity period are majortraits significantly affecting the adoption of MVs. Hostplant resistance to insects and diseases—traits thatplant breeders consider important—are not statisticallysignificant in the model.

Impact of Coordinated Rice Improvement Program on Movement ofImproved Germplasm and Productivity Gains across Indian States

Aldas Janaiah, Mahabub Hossain and E. Cabrera


89

This study develops an empirical model to test thehypothesis that technology spillovers produced byregional research networks can reduce the cost ofconducting agricultural research in less developedcountries. Furthermore, it examines whether returns topublic research are overestimated if the spilloversgenerated by regional cooperation are not identified.The study focuses on the Regional Maize Program(Spanish acronym: PRM), a network of maize breedersin Central America and the Caribbean (CAC). Itclassifies commercial maize germplasm according towhich institutions produced the basic populations,managed the breeding process, undertook adaptiveresearch, and screened the populations. Next, adatabase on experimental yields was assembled andused in combination with data on technology adoptionto: (1) estimate a matrix of potential spillovers; (2)conduct a research impact assessment identifying theincidence of spillovers and; (3) estimate cost-benefitratios showing the incidence of regional networks.Cost-benefit ratio estimates (CBRs) indicate that

ignoring spillovers leads to overestimating nationalagricultural research systems (NARSs) impacts by afactor of three. The findings also suggest thatalthough NARSs research has had high payoffs,ignoring technology spillovers considerably inflatesthe benefits from public maize breeding research inCAC. CBR estimates also indicate that the incidenceof spillovers is large: only one-third of researchimpacts come from independent NARSs’ research,leaving the rest to spillover. Amongst all institutions,the PRM is the largest contributor of spillovers and itsimpacts are more important in countries where maizeis used for direct human consumption. The cost-benefit ratio indicates that the PRM contributes toreducing research costs and, therefore, to improvingresearch’s financial efficiency relative to NARSs’independent research. Finally, results demonstratethat CAC countries have exploited the advantages ofspecialization and economies of scale in research byparticipating in the PRM.

Impact of Regional Agricultural Research Networks: The Case of theRegional Maize Program for Central America and the Caribbean

Miguel I. Gómez and Prabhu L. Pingali


90

Can Impact Analysis be Used for Research Evaluation?

Javier M. Ekboir

The complexity framework has broad consequences foragricultural and research policies. Impacts result fromthe technology’s own evolution, from changes in thenetworks that generate it and from the actions ofcompeting and complementary networks. Sinceimpacts ensue from the ways the whole network reactsto internal and external signals, they cannot beallocated to individual agents. In evaluating networks,the relevant parameters to study are the mechanismsthat govern their reaction to new technological andmarket environments. These include rules forgenerating, collecting and sharing information,financing procedures, intellectual property rightregulations and availability of human and financialresources. What can be evaluated for individual agentsare their patterns of participation in particularnetworks and the factors that determine those patterns.These factors are (1) benefits and costs of participation,(2) criteria for promotions, (3) evaluation criteria, (4)financial arrangements, and (5) institutional cultures.

Since technology generation and adoption are randomprocesses, any impact is, in part, the result of chanceand could not have been be fully foreseen by anyagent. Because of this feature, impact analysis shouldconcentrate more on the analysis of development andadoption process and less on sophisticated rules toallocate the result to particular agents. Additionally,since research impacts cannot be predicted, an ex-anteestimation of impact is a poor instrument to allocateresources. Again, the emphasis should be put on theparticular mechanisms that enable networks to beefficient: collective learning routines and the inter- andintra-institutional incentives to participate ininnovation networks.

Many impact studies tend to relate changes in aparticular impact indicator (e.g., output) with ameasure of research investment. This requires theimplicit assumption that the link between researchoutputs and impact indicators dominates all otherrelationships. However, this is true only for minorimprovements along stable technological paradigms(like successive replacement of modern varieties). Formost technologies, such as new crops or cropmanagement, other factors, such as policies andmarkets, influence adoption and, consequently, impact.Allocation of impact (or lack of) to research assumesthat all other factors were unimportant in thegeneration and adoption processes.

Since many factors influence adoption, researchimpacts must be analyzed within a broader framework.Such a framework is provided by complexity theory,which posits that impacts depend on the technology’sown evolution, on external forces (e.g., markets andregulations), and the direct and indirect interactions ofnetworks of agents (e.g., researchers, input suppliers,policy makers and financial institutions). Some of theseinteractions are formal and some informal; some areplanned while others are random. These interactionsalso depend on the maturity and nature of thetechnology. Newer and/or more complex technologiesare more uncertain and, thus, require greatercollaboration. In the development of simpler or better-known technologies, each agent is more aware of therole of other agents and of technology standards; thus,direct relationships are less important than marketinteractions. The problem is occasionally compoundedbecause the impacts appear after many years and oftencannot be measured.


91

An Evaluation Approach for Achieving andAttributing Impact for INRM and IPM

Boru Douthwaite, Thomas Kuby, Elske van de Fliert, and Steffen Schulz

In the first stage of evaluation, the research projectbegins by developing a program theory for itself andthen conducts its own self-evaluation, guided by theprogram theory, to the point of establishing directbenefits of its pilot site(s) outputs. Self-evaluation andthe learning it engenders contribute to adaptiveproject management, which is crucial to successfulINRM and IPM. Based on this learning, the programtheory evolves to map out, in greater detail, how theproject’s direct benefits can later be scaled-up.

The second stage, conducted some time after theproject has finished, is an ex-post impact assessmentin which the project’s wider benefits areindependently measured. This begins by establishingthe extent to which the program theory was valid inthe pilot site(s) and the extent to which scaling up hasoccurred. It is the job of the impact assessor to build aplausible and persuasive case for a link between theproject outputs and general level developmentalchanges, using case study methodology.

We illustrate the usefulness of the 2-stage PETthrough two examples: an ongoing project to developintegrated management options for the control ofStriga spp (a parasitic weed) with farmers in northernNigeria and a completed, integrated cropmanagement project in Indonesia.

Integrated natural resource management (INRM) andintegrated pest management (IPM) take a holistic—rather than reductionist—approach to research. Bothapproaches see innovation as a social process in whichpeople construct solutions to their problems. Once oneaccepts that users modify technologies and their ownsystems to accommodate new technologies, and thatthese adaptations affect adoption rates and thedistribution of benefits, one must also accept thattechnological change is an immensely complex processwith a high degree of non-linearity. Currently, the ‘bestpractice’ economic evaluation methods, commonlyused in the CGIAR system, struggle as they attempt toestablish a linear link between a project’s outputs andwider level impacts. Hence IPM and INRM require adifferent type of evaluation approach that can bridgethis attribution gap.

In this paper we look outside the CGIAR system tolearn lessons from the broader field of social programevaluation and incorporate experience of the Germandevelopment organization GTZ (Deutsche Gesellschaftfür Technische Zusammenarbeit GmbH). We find that atwo-stage program theory evaluation (PTE), guided byan impact model developed by GTZ, is better matchedto the needs of IPM and INRM. This type of evaluationis guided by an explicit theory or model of how aresearch project achieves impact. In other words, PTE isbased on a map of the impact pathway(s).


92

From Measuring Impact to Learning Institutional Lessons: AnInnovation Systems Perspective on Improving the Management of

International Agricultural Research

Andrew Hall

This paper argues that efforts to measure the impact ofagricultural research have largely failed to improvedevelopmental impacts, because commonly usedassessment methods fail to provide research managerswith critical institutional lessons for improvinginnovation. Approaches such as measuring rates ofreturn to investment, while providing politicallyexpedient evidence of the value of public research,offer little help in understanding the complex processesand institutional factors that gave rise to the relativesuccess of different research initiatives. The reasonsbehind this problem relate to the disciplinary andconceptual conventions that underpin both impactassessment methods and agricultural research policygenerally. The former is based on linear assumptionsabout the relationship between research inputs andeconomic outputs. Methodological problems aside,economic analysis of this type is ill equipped to giveinsights into the evolution of agricultural systems.

This paper proposes a systems model of innovation asa complementary framework to understandingagricultural innovation (technical and economicchange) in its wider institutional context. In thissystems conceptualization, innovation performanceand impact is viewed as resulting from the existence

and ability of coalitions, or systems of public andprivate research and non-research institutions, tointeract, create, transfer, and apply economically usefulknowledge. In these systems, innovations are derivedfrom evolutionary combinations of technical andinstitutional change. Case studies from India andAfrica are used to demonstrate how thisconceptualization of the innovation process could beused to evaluate research and technology programsand inform planning and research management.

The paper concludes by recognizing that internationaland national research efforts have some way to gobefore better integration with a wider set of innovationsystem actors can be achieved. However, there isincreasing evidence that developmental impacts frominvestments in agricultural science could be improvedif policy was informed by the institutional lessonsprovided by an innovations systems perspective.Adopting the evolutionary institutional learningprocesses that innovation systems thinking identifies ascritically important could be part of a complementaryassessment approach to improve planning and researchmanagement practices. This would assist policy toaddress the efficiency of the research process ratherthan, as at present, only measuring its outputs.


93

Disciplines, Institutions, and Organizations:Impact Assessments in Context

Rajeswari S. Raina

The second section discusses this isolation, which isevident in the information base and methodologiesused for impact assessment. Literature discussing theemergence of ‘manual mentality’ and the over-preoccupation with methods (for impact assessmentand/or research management) points to this increasingisolation. Institutionally and organizationally, the socialsciences used for impact assessment are tuned tolooking for consequence indicators alone. How can weuse indicators of requirements, relationships, andmonitoring? Will the epistemological and disciplinaryconstraints of the social sciences within nationalsystems and international centers permit the use ofthese indicators, which draw from systems conceptsand democratic values? These indicators help impactassessments “…demonstrate how research leads totechnology and technology leads to development.”

Our commitment to an evaluation culture (Trochim1992) is discussed in the next section. How willscientific credibility (as in economics) andaccountability or democratic processes be features ofthe desired evaluation culture when they contradicteach other at a fundamental epistemological level inthe social sciences? It is important to challengedisciplinary constructs and develop interdisciplinaryframeworks and institutions to facilitate the evaluationculture. This paper will also present analysis ofevaluation methodologies and disciplinary handicapsfrom agricultural research systems in three leadingdemocracies: India, the USA, and Sweden. To make animpact, impact assessment studies must be cognizantof disciplinary and institutional contexts.

Impact assessment is designed to inform decision-makers and relevant stakeholders in nationalagricultural research systems and internationalagricultural research centers about the impacts ofagricultural technologies, research programs orpolicies. This paper contends that impact assessmentresearch is located within the uneasy relationshipbetween agricultural and social sciences. Thisrelationship is determined by disciplinary constructs(theoretical and empirical remits) of the social sciencesand the institutions (rules of the game) andorganizations (structures governed by these rules/norms) of agricultural science. Social sciences,especially the economics and sociology used in impactassessment studies, embody this instrumentalismwithin the agricultural sciences.

The first section discusses how this instrumentalism isoperationalized within the institutions andorganizations of agricultural research. Agriculturalsciences use the social sciences: (a) as a substitute forextension, or the transfer of technology; (b) forstatistical verification of experimental results, aneconomic viability certificate before the release oftechnology; and (c) for conditional priority setting,evaluation, and policy formulation. Within nationalsystems and international centers, we are nowwitnessing the ‘anomalous expansion’ of rigorousacademic work on impact assessment. However, theinstitutional and organizational context of this researchhas remained grounded in instrumentalism. Thisisolation of impact assessment studies from theinstitutional and organizational reality of agriculturalscience, and of agricultural and social sciences from theecological, economic, social and cultural realities of theworld (which adopts or utilizes the knowledge/technology) can be traced back to the clear distinctionmade in instrumentalism between reality and morality.


94

Tradeoff Analysis as a Tool for Assessment of Economic andEnvironmental Impacts of Agricultural Research

David Yanggen, John Antle, Jetse Stoorvogel, Walter Bowen, and Charles Crissman

The public demand for information about botheconomic and environmental impacts of agriculturalresearch has been widely recognized. Indeed, onecould argue that one explanation for “Why HasImpact Assessment Research Not Made More of aDifference?” is that until recently it has focused onlyon economic impacts. However, it is also true thatuntil recently, methods and tools (data, models, andmodel integration software) needed to conduct moreintegrated types of assessments have been lacking.In a companion paper, we address another key issue,namely the need to integrate public decision makersand stakeholders into the assessment process.

This paper introduces the reader to recentlydeveloped methods and tools for integratedassessment of agricultural production systemsknown as tradeoff analysis (TOA) and the TOAmodel, and discusses how these methods and toolscan be used to assess quantitatively the impacts ofagricultural research on the sustainability ofproduction systems. Sustainability is defined interms of economic, environmental, and otherquantifiable indicators of the system’s performance.

The paper begins with a discussion of a generalapproach to the integrated assessment of agriculturalproduction systems that is known as TOA. We showhow the conventional economic assessment ofresearch impacts can be embedded in this moregeneral framework, and how it can be used for bothex post and ex ante assessment of research impacts.TOA is based on the identification of keyquantifiable sustainability indicators (e.g., economicindicators, environmental and social impactindicators) by stakeholders. A key methodologicalquestion is how to assess the impacts of research in

these multiple dimensions and communicate thatinformation to decision makers and stakeholders. Inthis section, we discuss several approaches taken byeconomists to this problem, as well as the solutionprovided by the TOA approach.

The next section of the paper discusses issues relatedto the implementation of an integrated assessment ofan agricultural production system for research impactassessment. Various methodological issues ariserelated to the choice of spatial and temporal scales,data availability, the suitability of existing disciplinarymodels, and how these data and models can beintegrated to implement such an assessment. In thiscontext, we briefly introduce the TOA model softwarethat can be used to implement this type of integratedassessment. This software organizes data in a GISframework and links inputs and outputs fromdisciplinary simulation models (biophysical crop andlivestock models, economic models, andenvironmental process models) on a site-specific basisand aggregates results to a larger spatial unit (such as awatershed or political region).

The final section of the paper illustrates the use of thistype of integrated assessment tool to assess impacts ofagricultural research using recent case studies inEcuador and Peru. Examples of technologies beinginvestigated in these areas include late-blight resistantpotato varieties, integrated pest managementtechniques, and soil conservation technologies(terracing, agroforestry, and alternative tillagepractices). These examples illustrate how this type ofassessment tool can be used to assess impacts ofexisting and alternative technologies within the cropand livestock production systems.


95

Agricultural Development and Impacts on theEnvironment: Experiences from India

D.D. Naik and Archana Godbole

India is principally an agrarian country. However,over the last two decades, it has been rapidlyindustrializing. The principal aim of industrializationhas been agricultural self-sufficiency. Nevertheless,India’s population continues to grow, increasing theneed for agricultural land. Changing land-usepatterns and the increasing use of irrigation arehaving detrimental impacts on the environment.Agricultural improvements and the Green Revolutionhave also affected genetic diversity preserved in theform of local crop varieties. The agro-biodiversity ofthe crop ecosystems is declining rapidly. In 1994, theIndian government passed the EnvironmentalProtection Act. As a result, environmental impactassessments are now mandatory for alldevelopmental projects. This includes irrigationprojects designed for agricultural improvement.

Between 1998 and 2001, impact assessment studieswere conducted for twelve irrigation projects inMaharashtra, India, including the arid zone in thenorthwestern part of the state. Results indicate thatenvironmental impacts occurred outside pre-determined parameters. The process designed forimpact assessment is not infallible. The role of keyactors in the process of development has not beenproperly addressed; specifically the role of policymakers and project planners in the process of impactanalysis. The necessity of providing water foragriculture to achieve food security overrides the aimof sustainable development. Based on the resultsattained from analysis of the twelve irrigationprojects, particularly when considering long-termimpacts, it is evident that a detailed analysis of theactors involved is essential.

The studies highlighted the need to evaluate the actualcost-benefit ratios in terms of environmental coststogether with the cost of project implementation andconsequent increases in agro-productivity. It isnecessary to achieve a balance between therequirements of irrigated agriculture andenvironmental protection/management. Impactanalysis and environmental managementmethodologies are important considerations.Appropriate impact assessment methodologies havethe potential to establish linkages between the processof agro-development and impacts on the environment.

The paper illuminates the difficulties involved inaccommodating the various attitudes and agendas ofdevelopers, farmers, stakeholders, and governmental,non-governmental and local agencies when pursuingthe goal of sustainable development. Certaininitiatives, namely stakeholder training andconsultation, are also evaluated in this paper.In conclusion, this paper argues that there is an urgentneed to undertake comprehensive environmentalassessments/audits for all major agriculturaldevelopment projects. In many cases, this may involvethe need for the development of suitable, moresustainable, alternatives. There is also a pressing needto generate a deeper understanding of these issues atthe policy-making level.


96

Impact of Salinity Management Research in Northwest India

K.K. Datta, Laxmi Tewari, and P.K. Joshi

Soil salinity is one of the complex abiotic phenomenaadversely affecting agricultural production worldwide.Globally, salinity constrains agricultural production on45 million ha of irrigated land, and this equates toannual loses of approximately US$ 11.4 billion.India is no exception. Diverse statistics indicate that theproblem is threatening agricultural production onbetween 5.5 and 13 million ha, some 1 million of whichare seriously affected and where agriculture has beenabandoned. To manage the problem in highly fertileirrigated areas, the Indian Council of AgriculturalResearch invested in systematic research efforts at theCentral Soil Salinity Research Institute to rehabilitateand manage soil salinity for agricultural production.This concerted and multi-disciplinary effort led to therecommended provision of sub-surface drainage forsalinity control. After small-scale operational studies,large-scale pilot projects were launched to install sub-surface drainage in problem areas. One such attemptwas initiated in northwest India. This paper assessesthe impact of investments in sub-surface drainage forsalinity control. Specific objectives of the paper are: (1)to assess impact of sub-surface drainage on efficiency,equity, and sustainability, and (2) to examine factorsaffecting the sustainability of the technology.

The internal rate of return was computed to assess theefficiency indicator of sub-surface drainage for salinitymanagement. Gini concentration ratios were computedwith and without installing sub-surface drainage tomeasure changes in equity issues. The Radar Approachwas used to quantify sustainability in terms ofoptimizing economic gains and conserving orimproving the quality of soil and water resources.

The results showed several farm-level benefits as aresult of installing sub-surface drainage. Theseincluded (i) a substantial increase in farm income, (ii)crop intensification and diversification towards highvalue crops, and (iii) the generation of employmentopportunities. High internal rates of return justifiedinvestment in sub-surface drainage. The program alsoindicated reduced income inequalities across farmproducers. The Radar Approach demonstratedimproved sustainability in terms of economic gains andresource conservation.

Despite economic, social, and environmental benefits,the sustainability of the sub-surface drainagetechnology is always questioned. The specific reasonsdiscussed in the paper are: (1) the indivisible nature ofthe technology, (2) lukewarm collective action by thebeneficiaries, (3) conflicting objectives amongbeneficiaries, and (4) a growing number of free riders.These were due to the absence of appropriateinstitutional arrangements. To a large extent, thesewere addressed in the study area by forming villagecommittees.

The analysis noted that technology withoutinstitutional arrangements might not yield the desiredresults. A technology with high potential benefits maynot make a difference and could be abandoned in theabsence of essential institutional arrangements.


97

Due to the mountainous topography of CentralAmerica (CA), heavy rainfalls during the long-lastingrainy season, and the continuing use of soil-degrading farm management practices, soil erosion isone of the major threats to the natural resource base,agricultural productivity, and survival of small-scalefarmers in the region. Despite the promotion of manydifferent soil conservation techniques by manygovernmental and non-governmental organizationsin the last 20 years, the overall adoption of thesekinds of technologies has been relatively low,especially among small-scale farmers. This studyassesses the adoption and impact of conservationtechnologies at three different levels of aggregation.

First, at the regional level, fifteen widely promotedsoil conservation practices are assessed according tofour economic criteria and compared with the maineconomic circumstances of small farmers in CA.Results suggest that most techniques promoted in CAdo not match small farmers’ circumstances for one ormore of the criteria.

Second, we conducted a comparative analysis of thefactors affecting the adoption of differenttechnologies, using the results of adoption studiesperformed over the past ten years. The analysisexamines both the demand for new technologies andtheir supply in the region. Results from the supplyside show that institutions failed to recognize theparticular characteristics that differentiatedenvironmental innovations from commercialinnovations, suggesting a need for to understandfarmers’ demands for such technologies. In particular,three characteristics of environmental innovations

were found to be incompatible with small farmers’circumstances: (1) the need for farmers to make asignificant initial investment in terms of land, capitalor labor; (2) the existence of a time lag before farmersobtain benefits; and (3) complex managementrequirements in terms of land allocation andtechnology maintenance. On the demand side, threemain factors were identified as being responsible forthe lack of widespread adoption of soil conservingtechnologies: (1) farmers’ inability to capture long-term benefits caused mainly by insecure land tenureand low wealth levels; (2) high transaction costsassociated with adopting the innovations,particularly the costs of acquiring and processinginformation about the technology, as well as highland and labor opportunity costs; and (3) marketfailures related to the interaction between croppingand livestock systems in areas where commongrazing predominates.

Third, the impact of conservation tillage andmulching was assessed. Using the results of an in-depth case study in El Salvador, a simulation modelwas built to identify the effects of a set of alternativepolicy measured aimed to foster a more sustainableand productive maize production system in CA. Aswell as beneficial effects on soil characteristics andincreased land productivity, some negative effectswere identified in terms of the increased use ofpesticides that could be deleterious to theenvironment and farmers’ health. Diffusion of thetechnology to other regions could be restricted by theinteraction with the livestock system and themalfunction or non-existence of fodder markets.

Economic and Institutional Factors Affecting the Adoption of SoilConservation Technologies in Central America

Gustavo Saín and Monika Zurek


98

This study addressed two themes. First, thedevelopment of capacity within the Plant GeneticResources Center (PGRC), Ghana, to conserve,evaluate, and utilize Ghana’s plant genetic resources.Second, the capacity building contributions made toPGRC’s development by the International PlantGenetic Resources Institute (IPGRI) and the GeneticResources Network of West and Central Africa(GRENEWECA). The study aimed to (1) analyze thedevelopment of PGRC capacity in human resourcesand conservations facilities/methods; (2) illustrate andlearn from the capacity development experiences ofGhana, to develop other national programsthroughout West and Central Africa; and (3) promotethe analysis of and improve evaluation skills incapacity development. The evaluation analyzed datafrom 1980-1999, a time of major growth and change.

An organizational capacity development modeldeveloped by the International Development ResearchCenter (IDRC) served as the theoretical basis for thestudy. This model defines capacity development as theability of an organization to meet its goals on asustainable basis and consists of three inter-relatedcomponents, namely, motivation within theorganization, resources and equipment, and theexternal environment. Capacity developmentindicators were developed from an internationalagreement (The Global Plan of Action for theConservation and Sustainable Use of Plant GeneticResources for Food and Agriculture, or GPA), whichprovides a comprehensive framework for priorityareas within which capacity must be built.

The case study approach was used for analysis, with astrong emphasis on self-assessment by theorganizations involved. Multiple methods were used,including self-assessment workshops for PGRC staffgroups and stakeholders and for GRENEWECAsteering committee members, interviews of IPGRIsenior managers and high-level Ghanaian officials, apersonal history of the head of the PGRC, a survey ofIPGRI staff, and a review of archival records from allthree organizations.

The results demonstrated that capacity has beendeveloped within the PGRC in many areas, namely,staff number and qualifications, budget allocation, andoperational autonomy. This led to an increased numberof germplasm accessions conserved and thedevelopment of programs to promote the use ofconserved germplasm for research and in farmers’fields. IPGRI and GRENEWECA have made significantcontributions in the areas of training, infrastructure,information services, and more recently, policy andawareness raising. A major strength of IPGRI’scontribution has been the continuity of involvementover a prolonged time period. The study also foundthat contributions to capacity building have been madeby other organizations; for example, bilateral donors.The PGRC has also benefited from a strong network ofsupporters within Ghana, including other researchinstitutions and universities, who are partners in thework of germplasm conservation and use. The studyalso identified institutional weaknesses in PGRC,GRENEWECA and IPGRI. These related to a lack ofvision, policy and a strategy for capacity development.Building a capacity development culture into theorganizations has been identified as a priority.

Evaluating Capacity Development of the Plant GeneticResources Center in Bunso, Ghana

Samuel Bennett-Lartey, Raymond Vodouhe, and Jamie Watts


99

Capacity development has moved center-stage inagendas of development organizations. Strengtheningthe capabilities of individuals, organizations, andinstitutions is viewed as essential to ensure thesustainability of development efforts and theeradication of poverty. Substantial sums are beinginvested in the development of organizational andinstitutional capacities. Yet, the design, management,and evaluation of capacity-development efforts leavemuch to be desired. Results indicate that capacity-development efforts often fall short of their goals andexpectations. Relatively few of these efforts have beensystematically and thoroughly evaluated. Thecomplexity and dynamic nature of capacity-development processes complicates evaluation. Thefamiliar challenges facing evaluators of capacitydevelopment at the individual level are exacerbated atthe organizational level.

A number of organizational or institutional assessmentframeworks have been published. However, little isknown about their use in evaluating capacitydevelopment. Most of the information available in thisfield is in “gray” or “fugitive” literature.

This paper reports on a project that seeks to contributeto the effectiveness of capacity-development efforts inresearch and development organizations through theuse of evaluation. The project focuses on capacitydevelopment at the organizational level, rather than atthe micro level of individuals or the macro level ofnational institutions. It recognizes the multipleperspectives of key actors involved in capacity

Evaluating Capacity Development in Research and DevelopmentOrganizations: Emerging Results of a Global Study

Douglas Horton and Nancy Alexaki

development, rather than the single perspective of anexternally funded and directed capacity-developmentintervention. The impacts or contributions ofinternational organizations to capacity developmentin national organizations are being assessed “from thebottom up.” The project is implemented by capacity-development practitioners and evaluators based in 13organizations.

An action-research approach is being employed toadvance understanding and practice in the evaluationof capacity-development efforts in organizations thatconduct research or use research for development.Professionals involved in capacity development areevaluating their own efforts, and in the process, theyare developing and refining evaluation methods thatmay later be suitable for use on a broader scale.The project involved participants from six nationaland local organizations who are working to developtheir own capacity and six international organizationswho support these capacity-development efforts (fourof the six international organizations are CGIARcenters). Five donor agencies that provide resourcesfor organizational capacity development, alsoparticipated in the project.

The paper reviews the current state of practice inevaluating capacity development and presents aframework for evaluating organizational capacitydevelopment. Highlights of six evaluation casestudies are followed by general conclusions andlessons for improving capacity development effortsand their evaluation.


100

The notion that public extension services in agricultureare no longer adequate to meet the changing needs offarmers, in many developing countries, is becomingincreasingly emphasized in both research literature andthe strategies of development agencies. The problemsencountered with the T&V system have causedconsiderable frustration among donor agencies andreduced their willingness to invest in large-scale publicagricultural extension programs. For some, the solutionis delivering information and services through thecomplete or semi-privatization of public agencies.Others argue that the public sector has to maintain itsrole in knowledge and information transfer but see theneed to change the hierarchical structure and the topdown approaches of the public sector towardsadopting participatory methods of extension.

This paper compares and contrasts five differentagricultural extension projects in Egypt, which aresupported by different donor agencies, promotingvarious forms of participatory methods to agriculturalextension and farmer training. Comparisons are being

The Costs of Transforming Public Extension Services towardsParticipatory Approaches: The Case of Egypt

Gerd Fleischer, Hermann Waibel, and Gerd Walter-Echols

made on the basis of cost-effectiveness parameters.Furthermore, break-even benefits are calculated basedon the effects on farmer’s income. Results indicate thatusing unit costs of farmer training is not a valid basisfor comparison but that there is a need to take intoaccount the start-up conditions and the specifics of thetraining method, which result in different levels of basecosts, start-up costs, recurrent costs and farmers’ costs.It is also shown that depending on the situation, thebreak-even benefits are surprisingly low to justify evenconsiderable investments in participatory training andextension. However, in for example a crop like cotton,such approaches are unlikely to be economical viable.

The paper concludes that there is a need for highselectivity and careful planning in public sectoragricultural extension projects. However, the study alsoemphasizes the notion that from an economic point ofview, public investments in participatory agriculturalextension are likely to be superior to other approachesin extension, if the target is well chosen.


101

List of participants

Mauricio BellonEconomics ProgramCIMMYT, Int.Apartado Postal 6-641,México 06600MéxicoTel. +52 55 5804-2004Fax: +52 55 5804-7558/59Email: [email protected]

Samuel Bennett-LarteyPlant Genetic Resources Center (CSIR)P.O. Box 7BunsoGhanaTel. +233 81-24124/24138Fax:Email: [email protected]

Richard BernstenDepartment of Agricultural EconomicsMichigan State University211E Agriculture HallEast Lansing MI 48824U.S.A.Tel. +517 351-5063Fax: +517 432-1800Email: [email protected]

Robert ChambersUniversity of SussexBrightonFalmer Brighton BN19REUKTel. +44-0-1273-606261Fax: +44-0-1273-621-202Email: [email protected]

Rodney CookeInternational Fund for AgriculturalDevelopment-IFADDirector, Technical Advisory DivisionVia del Serafico, 107Italy 00142ItalyTel. +0039-06-5459-2450Fax: +0039-06-5459-2018Email: [email protected]

Sam CordesUniversity of NebraskaInternational Programs58 Filley Hall, UN-LLincoln NE 68583-0947U.S.A.Tel. +402 472-1772Fax: +402 472-0688Email: [email protected]

Arthur CosbySocial Science Research Center,Mississippi State UniversityDirector103 Research ParkMississippi State MS 39762U.S.A.Tel. 662-325-7127Fax: 662-325-7966Email: [email protected]

Dana DalrympleOffice of Agriculture and FoodSecurityResearch Advisor2.11-075 RRB, 14th St.Washington, D.C. 20523-2110U.S.A.Tel. (202) 712-5596Fax: (202) 216-3010Email: [email protected]

K.K. DattaICAR Research Complexfor NEH RegionDivision of Agricultural Economics &StatisticsUmroi RoadUmiam-793103 (India)Tel. 91-0364-570257Fax: 91-0364-570288Email: [email protected]

David DaweIRRIDAPO Box 7777Metro Manila 1301PhilippinesTel. +632 845-0563Fax:Email: [email protected]

Michele AdatoFood Consumption and NutritionDivisionIFPRI2033 K Street N.W.Washington, D.C. 20006-1002U.S.A.Tel. (202) 862-8195Fax: (202) 467-4439Email :[email protected]

Aden Aw-HassanSocio EconomicsICARDAP.O. Box 5466Aleppo, SyriaTel. 963-21-2213433Fax: 963-21-2213490Email: [email protected]

Henning BaurGTZ-BEAFP.O. Box 120 508Bonn 53047GermanyTel. +49 (0) 228-2434-867Fax: +49 (0) 228-2434-861Email: [email protected]

Anthony BebbingtonUniversity of ColoradoAssociate Professor, Department ofGeography and Co-Director, Programin Developing Areas ResearchUniversity of ColoradoBoulder Colorado 80309-0260U.S.A.Tel. 303-492-5438Fax: 303-492-7501Email: [email protected]

Jere BehrmanUniversity of PennsylvaniaDirector of Population Studies Centerand Professor of Economics3718 Locust WalkPhiladelphia PA 19104-6297U.S.A.Tel. 1-212-898-7704Fax: 1-215-898-2124Email: [email protected]

102

Hugo De GrooteCIMMYT - KenyaP.O. Box 25171NairobiKenyaTel. +254-2 524-600 / 524-608Fax: +254-2 522-879Email: [email protected]

Antonio Flavio Dias AvilaBrazilian Agricultural ResearchCorporation - EmbrapaParque Estaçao Biológica - Pq.EB,Via W3 NorteBrasília, DF 70770-901BrazilTel. +55 61-4484167Fax: +55 61-3474480Email: [email protected]

Boru DouthwaiteIITAIbadan OyoNigeriaTel. +00234 2-241-2626Fax: +00234 2-241-2221Email: [email protected]

Jose Rafael Esteban DuránInstituto Nacional de Investigación yTecnología Agraria y Alimentaria(INIA)Ctra. De la Coruña Km 7.5Madrid 28040SpainTel. 349-134-76812Fax: 349-135-72293Email: [email protected]

Simeon EhuiInternational Livestock ResearchInstituteLivestock Policy Analysis ProgramP.O. Box 5689Addis AbabaEthiopiaTel. +251-146-3495Fax: +251-146-1252Email: [email protected]

Javier EkboirCIMMYT, INT.Apartado Postal 6-641,Mexico 06600MexicoTel. +52 55 5804-2004Fax: +52 55 5804-7558/59Email: [email protected]

Adrian ElySchool of Biological Sciences,University of Sussex, UKMantell Building, University of SussexFalmer Brighton BN1 9RFUnited KingdomTel. +44 0-1273-877-301Fax: +44 0-1273-678-433EMail: [email protected]

Robert EvensonEconomic Growth Center - YaleUniversityProfessor of Economics and Director,International and DevelopmentEconomics program27 Hillhouse AvenueNew Haven Connecticut -6520-8269USATel. +203-432-3626Fax: +203-432-5591EMail: [email protected]

S.M. Fakhrul-IslamDepartment of AgriculturalEconomics, Bangabandhu SheikhMujibur Rahman AgriculturalUniversityDept. of Agricultural Economics,Bangabandhu Sheikh Mujibur RahmanAgricultural UniversitySalna, Gazipur 1701BangladeshTel. +8802-9334677 / 9339007EMail: [email protected]

Shenggen FanIFPRI - Environment and ProductionTechnology Division2033 K Street, NWWashington, D.C. 20006U.S.A.Tel. +202 862-5677Fax: +202 467-4439EMail: [email protected]

Ade FreemanICRISATP.O. Box 39063NairobiKenyaTel. +254 2-524553 / 521450Fax: +254 2-524001EMail: [email protected]

Roberta GerpacioCIMMYT - Philippinesc/o International Rice ResearchInstitute (IRRI)Economics ProgramDAPO 7777Los Banos LagunaPhilippinesTel. +63-2 762-0127 / 845-0563 local6827Fax: +63-2 761-2406 / 891-1292EMail: [email protected]

Archana GodboleApplied Environmental ResearchFoundation46/4/G2, Hill View Apartment, PaudRoadPune Maharashtra 411038IndiaTel. +0091 20-4004233Fax: +0091 20-4004244EMail: [email protected]

Miguel Ignacio GómezFood Industry Management Program -Dept. of Applied Economics &Management - Cornell University109 Warren HallIthaca NY 14853U.S.A.Tel. +607 255-8472Fax: +607 255-4776EMail: [email protected]

Hans GregersenCGIAR Standing Panel on ImpactAssessmentBox 1322Solvang, CA 93464U.S.A.Tel. +1 (805) 693-1148EMail: [email protected]

Bill GriffinCrop & Food ResearchPB 4704ChristchurchNew ZealandTel. +64 3-325-6400Fax: +64 3-325-2074EMail: [email protected]

Luciano GutierrezUniversity of Sassari. Italy.Dept. of Agricultural EconomicsVia De Nicola 1Sassari 07100ItalyTel. +39 079-229256Fax: +39 079-229356EMail: [email protected]


103

Lawrence HaddadInternational Food Policy ResearchInstitute (IFPRI)Food Consumption and NutritionDivision2033 K Street, N.W.Washington, D.C. 20006U.S.A.Tel. +202 862-8179 / 5686Fax: +202 467-4439EMail: [email protected]

Fred Hakon JohnsenNoragric, Agricultural University ofNorwayBox 5001N-1432 AsNorwayTel. +47-64949797Fax: +47-64940760EMail: [email protected]

Andrew HallSocio-Economics and PolicyProgramme, ICRISATPatancheru Andhra Pradesh 502 324IndiaTel. +91 403-296-161Fax: +91 403-241-239EMail: [email protected]

Kelly HallmanPolicy Research Division PopulationCouncil1 Dag Hammarskjold PlazaNew York, NY 10017U.S.A.Tel. +212 339-0687Fax: +212 755-6052EMail: [email protected]

Frank HartwichPublic-Private Partnerships forAgroindustrial Research ProjectISNARIICAApdo 55-2200Coronado, San JoseCosta RicaTel. (506) 216-0270Fax (506)216-0247Email: [email protected]|

Peter HazellInternational Food Policy ResearchInstitute (IFPRI)Environment and ProductionTechnology Division2033 K Street, NWWashington, D.C. 20006U.S.A.Tel. +1 (202) 862-8151Fax: +1 (202) 467-4439EMail: [email protected]

Paul HeiseyCIMMYT and Economic ResearchService, USDAUSDA Economic Research Service,1800 M St., NWWashington D.C. 20036-5831U.S.A.Tel. +1-202 694-5526Fax: +1-202 694-5774EMail: [email protected]

Dave HodsonCIMMYT, Int.Apdo. Postal 6-641,México, D.F. 06600MéxicoTel. +52 55 5804-2004Fax: +52 55 5804-7558/59EMail: [email protected]

Douglas HortonISNARP.O. Box 93375The Hague 2509AJNetherlandsTel. +3170 349-6197Fax: +3170 381-9677EMail: [email protected]

Mahabub HossainInternational Rice Research Institute(IRRI)Social Sciences DivisionDAPO # 77 77Metro Manila 1301PhilippinesTel. +63-2 845-0563 Ext. 309Fax: +63-2 845-0606 / 891-1292EMail: [email protected]

James HydeUniversity of ConnecticutAgr. Nutrient ManagementUnit 4067 Plant ScienceStors CT 06269U.S.A.Tel. +860-486-6385Fax: +860-486-0687EMail: [email protected]

Aldas JanaiahSocial Sciences Division - InternationalRice Research Institute (IRRI)Head and ResearcherDAPO # 77 77 ; Metro ManilaThe PhilippinesTel. +63-2 845-0563 Ext. 309Fax: +63-2 845-0606 / 891-1292EMail: [email protected]

Nancy JohnsonCIATResearch FellowAA 6713Cali Valle de CaucaColombiaTel. +572 445-0000Fax: +572 445-0073EMail: [email protected]

Satwant KaurCIMMYT, Int.EditorApdo. Postal 6-641México, D.F. 06600Tel. +52 55 5804-2004Fax: +52 55 5804-7558/59EMail: [email protected]

Timothy KelleyTAC/interim Science CouncilSecretariatFood and Agriculture Organization ofthe United NationsSenior Agricultural Research OfficerRoom C-632Viale delle Terme di CaracallaRome 00100ItalyTel. +39 (06) 570-54210Fax: +39 (06) 570-53298EMail: [email protected]

Bonwoo KooInternational Food Policy ResearchInstitute2033 K Street, NWWashington, D.C. 20006U.S.A.Tel. +202 862-8184Fax: +202 467-4439EMail: [email protected]

Stephan KrallGerman Agency for TechnicalCooperation (GTZ),Dept. 45, Rural Development, ProjectInt. Agricultural Research / BEAFInternational Agricultural ResearchP.O. Box 5180Eschbom 65726GermanyTel. +49 6196-79-1416Fax: +49 6196-79-7137EMail: [email protected]


104

Patti KristjansonInternational Livestock ResearchInstitute (ILRI)P.O. Box 30709NairobiKenyaTel. +254-2-630-743Fax: +254-2-631-499EMail: [email protected]

Daniel LemaINTA-IESCerviño 3101 2o. PisoCapital Federal 1425ArgentinaTel. +11 4-802-5155Fax: +11 4-954-7285EMail: [email protected]

Aida LibreroPhilippine Council for Agriculture,Forestry and Natural ResourcesResearch and DevelopmentInstitution Development and ResourceManagementLos Banos, LagunaLaguna 4030PhilippinesTel. +0063-49 536-0014 to 20Fax: +0063-49 536-0016/536-0132EMail: [email protected]

Gideon LichfieldThe EconomistAv. Mexico 117-10Col. Hipodromo Condesa, CP06100Mexico, D.F.Tel. +52 55 8596-4062EMail:[email protected]

Nina LiljaCIAT/PRGA Program78 N. Main Ave.Orono ME 04473U.S.A.Tel. +1-207-866-2093Fax: +1-207-581-4278EMail: [email protected]

Alex McCallaCIMMYT, Int.Board of TrusteesUniversity of CaliforniaOne Shields AvenueDavis, California 95616 8512U.S.A.Tel. (530)752-1529Fax: (530)752-5614EMail: [email protected]

Ronald MackayConcordia UniversityAnnex EN-409, 1455 de MaisonneuveBlvd. WestMontreal Quebec H3G 1M8CanadaTel. (514) 848-2451Fax: (514) 848-4295EMail: [email protected]

Amado MaglinaoInternational Water ManagementInstitute (IWMI)P.O. Box 1025, Kasetsart UniversityBangkok 10903ThailandTel. +66 0-2561-4433Fax: +66 0-2561-1230EMail: [email protected]

Carissa MarasasCIMMYT, INT.Apartado Postal 6-641,Mexico 06600MexicoTel. +52 55 5804-2004Fax: +52 55 5804-7558/59EMail: [email protected]

Mywish MarediaMichigan State UniversityBean/Cowpea CRSP200 International CenterEast Lansing MI 48824U.S.A.Tel. +517 3554693Fax: +517 4321073EMail: [email protected]

David MatherDepartment of Agricultural Economics- Michigan Stete University(Ph.D.) Student202 Agriculture Hall, East LansingMI 48824U.S.A.Tel. +517 353-7898 / 351-5063Fax: +517 432-1800EMail: [email protected]

Peter MatlonThe Rockefeller FoundationFood Security420 Fifth AvenueNew York NY 10018U.S.A.Tel. +212 852-8237Fax: +212 852-8442EMail: [email protected]

Héctor Medina CastroInter-american Institute forCooperation on AgricultureCarretera a Coronado, 600 mts.Noreste del cruce de IpisCoronado San José 552200Costa RicaTel. +506 216-0222Fax: +506 216-0233EMail: [email protected]

Ruth Meinzen-DickInternational Food Policy ResearchInstitute2009 K St NWWashington, D.C. 20006U.S.A.Tel. +636 405-1711Fax: +202 467-4439EMail: [email protected]

Mulugetta MekuriaCIMMYT - ZimbabweHarare Mount Pleasant MP 163ZimbabweTel. +263 4-301-945Fax: +263 4-301-327 / 301-945EMail: [email protected]

Orlando Meneses PeñaPRONATTACoordinador de DesarrolloTecnológicoAve. Jiménez No. 7-65 Piso 5BogotáColombiaTel. +57-1 2841815/2845670Fax: +57-1 284818/2845670EMail: [email protected]

Erika MengCIMMYTApdo. Postal 6-641México 06600MéxicoTel. +52 55 5804-2004Fax: +52 55 5804-7558/59EMail: [email protected]

Susana Beatriz MirassouCONICETSubgerencia de Evaluación en CienciasBásicas Apricadas y TRivadavia 1917 2o. PisoBuenos AiresArgentinaTel. +54 11-4-954-7472Fax: +54 11-4-954-7346EMail: [email protected]


105

Michael MorrisCIMMYTEconomics ProgramApartado Postal 6-641,Mexico 06600, MexicoTel. +52 55-5804-2004Fax: +52 55-5804-7558EMail: [email protected]

Carmen Nieves MortensenDanish Government Institute of SeedPathology for DevelopingThorvaldsensvej 57Frederiksberg DK-1871DenmarkTel. +45 35 28 37 22Fax: +45 35 28 37 01EMail: [email protected]

Edgardo Moscardi CarraraFontagro1300 N. York Av.Washington, D.C. 20577U.S.A.Tel. +202 623-2873Fax: +202 623-3968EMail: [email protected]

Robert MoxleyNorth Carolina State UniversityDept. of Soc. & Anthro., Box 8107Raleigh NC 27695-8107U.S.A.Tel. +919 515-9011Fax: +919 515-2610EMail: [email protected]

Joseph NagyNagy International407 Forsyth WaySaskatoon Saskatchewan S7N 4H7CanadaTel. +1 306-249-4196Fax: +1 306-249-5768EMail: [email protected]

D.D. NaikD.D. & Associates41, Ex Servicemen Coloney, Paud RoadPune Maharashtra 411038IndiaTel. +0091 20-546-3722Fax: +0091 20-546-1174EMail: [email protected]

Luis NavarroInternational Development ResearchCentre (IDRC)People, Land and Water ProgramP.O. Box 62084Nairobi NairobiKenyaTel. +254 2-713160Fax: +254 2-711063EMail: [email protected]

Alice NdiemaNational Plant Breeding ResearchCentrePrivate Bag NjoroKenyaEMail: [email protected] /[email protected]

Gerald NelsonUniversity of Illinois315 Mumford Hall1301 W. GregoryUrbana IL 61801-3608U.S.A.Tel. +217 333-6465Fax: +217 333-5538EMail: [email protected]

George NortonVirginia TechDept. of Agr. and Applied Economics,Virginia TechBlacksburg Virginia 24061-0401U.S.A.Tel. +540 231-7731Fax: +540 231-3318EMail: [email protected]

Kamal PaudyalCIMMYT-NepalP.O. Box 5186, LazimpatKathmanduNepalTel. +977 1-429612 to 14Fax: +977 1-419352EMail: [email protected]

Alexander PhiriBunda College of AgricultureP.O. Box 219Lilongwe, MalawiTel. 277-257 Ext. 114Fax 277-286/251/403E-mail:[email protected]

Prabhu PingaliCIMMYTEconomics ProgramApartado Postal 6-641, Col. JuárezMexico 06600MexicoTel. +52 55-5804-2004Fax: +52 55-5804-7558EMail: [email protected]

Consuelo QuirozUniversidad de los AndesApartado Postal #22Trujillo Trujillo 3150VenezuelaTel. +58 272-6721672Fax: +58 272-2360467EMail: [email protected]


Rajeswari Sarala RainaNational Institute of Science,Technology and Development Studies(NISTADS)NISTADS, Dr. K.S. Krishnan Road,PusaNew Delhi 110 012IndiaTel. +91 011-572-9151 / 572-6406 / 577-3179Fax: +91 011-575-4640EMail: [email protected] /[email protected]

Timothy ReevesCIMMYT, Int.Apdo. Postal 6-641,México, D.F. 06600MéxicoTel. +52 55 5804-2004Fax: +52 55 5804-7558/59EMail: [email protected]

Mitch RenkowNorth Carolina State University -Dept. of Agricultural and ResourceEconomicsBox 8109Raleigh, N.C. U.S.A. 27695-8109U.S.A.Tel. +919 515-5179Fax: +919 515-1824EMail: [email protected]

Jarryl RitchieMississippi State University, SocialScience Research CenterCoordinator, The Monitor LaboratoryP.O. Box 5287Mississippi 39762U.S.A.Tel. +662 325-7127Fax: +662 325-7966EMail: [email protected]

Carlos Alberto Robles CocuyamePRONATTAAve. Jiménez No. 7-65 Piso 5BogotáColombiaTel. +57-1 2841815 / 2845670Fax: +57-1 2841818 / 2845670EMail: [email protected]

Barbara Alison RoseFuture HarvestPMB 2382020 Pennsylvania Ave, NWWashington, D.C. 20006-1846U.S.A.Tel. +202 473-4734Fax: +202 614-1552EMail: [email protected]

106 List of participants

Johannes RoseboomISNARP.O. Box 93375The Hague 2509 AJNetherlandsTel. +31 70-3496-159Fax: +31 70-3819-677EMail: [email protected]

Jim RyanAustralian National University,Camberra ACT AustraliaEconomics Division, Research Schoolof Pacific and Asian Studies18 Nungara Place, ArandaCanberra ACT 2614AustraliaTel. +61 2-6251-1942Fax:EMail: [email protected]

Raphaël SauvéInternational Plant Genetic ResourcesInstitute (IPGRI)10 De la Fondrière app. 7Hull Québec J8Z 2X8CanadaTel. +819 776-1167Fax:EMail: [email protected]

David SchimmelpfennigU.S. Department of Agriculture -Economic Research Service1800 M Street, NW, Room 4195Washington, DC 20036-5831U.S.A.Tel. +202 694-5507Fax: +202 694-5774EMail: [email protected]

Joginder SinghPunjab Agricultural University - Dept.of Economics & Sociology, 141004LudhianaIndiaTel. +91 161-401-960 / 301EMail: [email protected]

Shephard SizibaUniversity of ZimbabweCIMMYT - ZimbabweHarare MP 163ZimbabweTel. +263 4-301-945Fax: +263 4-301-327EMail: [email protected]

Melinda SmaleCIMMYT and IPGRI2033 K Street, N.W.Washington, D.C. 20006U.S.A.Tel. +202 869- 8119EMail: [email protected]

Meredith SouleUSAID/Economic Growth,Agriculture, and Trade/Office of FoodSecurityRoom 2.11-085 RRB1300 Pennsylvania Ave., NWWashington, D.C. 20523-2110U.S.A.Tel. +202 712-1058Fax: +202 216-3010EMail: [email protected]

Andreas Springer-HeinzeGTZ Division Rural Development -Alliance Services for RuralDevelopmentProject on Knowledge Systems inRural DevelopmentGTZ-4556, Postfach 5180Eschborn 65760GermanyTel. +49 6196-791-441Fax: +49 6196-797-162EMail: [email protected]

J.L. Tadeo LluchINIACtra. De la Coruña pk 7MadridSpainTel. 34913476821Fax: 34913572293EMail: [email protected]

Deborah TempletonACIAR - Impact Assessment ProgramP.O. Box 1571Camberra ACT 2601AustraliaTel. +02 6217-0547Fax: +02 6217-0501EMail: [email protected]

Jawahar ThakurRajendra Agricultural UniversityDirectorate of Research, RajendraAgricultural University,Pusa-848125 Samastipur, BiharIndiaTel. (091)6274 74226(0)Fax: (091)6274 74255 or 6274 74266Email: [email protected]

Daniel TicehurstNatural Resource InstitutePerformance and Impact ProgrammeCentral AvenueChatham Maritime Kent ME4 4TBUKTel. +44 1634-883-062Fax: +44 1634-883-706EMail: [email protected]

Clesencio TizikaraNational Agricultural ResearchOrganisation (NARO)Monitoring, Evaluation and PlanningUnit (MEPU)P.O. Box 295Entebbe Wakiso District 256UgandaTel. +256 77-408-636Fax: +256 41-321-070EMail: [email protected]

Roque TomasiniEMBRAPABR 285, km 174CP 451-CEPPasso Fundo-RS 99001-970BrasilEMail: [email protected]

Enrique TorresFacultuad de Agronomía, UniversidadNacional de ColombiaAssociate ProfessorApartado Aereo 14490Bogota D.C.ColombiaTel. +57 1-2175031Fax: +57 1-3165146EMail: [email protected]

Leonardo Velásquez ArgüelloPRONATTACoordinador de Seguimiento yEvaluaciónAve. Jiménez No. 7-65 Piso 5BogotáColombiaTel. +57-1 2841815 / 2845670Fax: +57-1 2841818 / 2845670EMail: [email protected]

G. VenkataramaniThe Hindu, Indian’s NationalNewspaperKasturi Buildings, 859-860Anna salai, Chennai – 600 002IndiaTel.841-3942841-3344Fax: 841-5395Email:[email protected]

Jeff VitaleTexas A&M University - Dept. ofAgricultural Economics2124 TAMUCollege Station Texas 77843-2124U.S.A.Tel. +979 862-3342Fax: +979 845-7504EMail: [email protected]

107List of participants

Hermann WaibelHannover UniversityHerrenhauserstr. 2 Institut fürGartenbauökonomieHonnover 30419GermanyTel. +49 511-762-2666Fax: +49 511-762-2667EMail: [email protected]

Thomas WalkerInternational Potato Center (CIP)Social Science DepartmentAv. La Molina 1895Lima 12PeruTel. +51-1 3175335Fax: +51-1 317 5326EMail: [email protected]

Dave WatsonCIMMYTApartado Postal 6-641,Mexico 06600MexicoTel. +52 55-5804-2004Fax: +52 55-5804-7558EMail: [email protected]

Jamie WattsInternational Plant Genetic ResourcesInstituteIntern and Impact Assessment andEvaluation SpecialistVia dei Tre Denari, 472/aMaccarese (Fiumicino) Rome 00057ItalyTel. +39 06 611-8253Fax: +39 06 611-8405EMail: [email protected]

Katinka WeinbergerAsian Vegetable Research andDevelopment Center (AVRDC)Research Associate-SocioeconomistP.O. Box 42, ShanhuaTainan 741TaiwanTel. +886 6-583-7801 Ext. 463Fax: +886 6-583-0009EMail: [email protected]

Stanley WoodInternational Food Policy ResearchInstitute - IFPRI2003 K St.NWWashington, D.C. 20006U.S.A.Tel. +202 862-8122Fax: +202 467-4439EMail: [email protected]

David YanggenCentro Internacional de la PapaApartado 17 21 1977QuitoEcuadorTel. +593-2-269-0362Fax: +593-2-269-2604EMail: [email protected]

Monika ZurekCIMMYTApartado Postal 6-641,Mexico 06600MexicoTel. +52 55-5804-2004Fax: +52 55-5804-7558EMail: [email protected]

Research and Development - CAS · John Poulsen, Boru Douthwaite, and Douglas White 41 Assessing...

Documents

Transcript of Research and Development - CAS · John Poulsen, Boru Douthwaite, and Douglas White 41 Assessing...