-CG, g Consultative Group on International Agricultural Research -~~~~~~~~~~~~A o4 /9fr Study Paper Number 28 Social Science in the CGIAR Proceedings of a Meeting of CGIAR Social Scientists held at the International Service for National Agricultural Research (ISNAR), the Hague, the Netherlands, August 1992 Edited by Michael P. Collnson and Kerri Wright Platais Social Science in the CGIAR Proceedings of a Meeting of CGIAR Social Scientists held at the International Service for National Agricultural Research (ISNAR), the Hague, the Netherlands, August 1992 CGIAR Study Papers No. 1 Technlgical Innovation in Agriculture The Political Economy of Its Rate and Bias No. 2 Modem Varieties, Interational Agrcultural Research, and the Poor No. 3 Plant Gerneic Resources: The Impact of the International Agricultural Research Centers No. 4 Costa Rica and the CGIAR Centers: A Study of Their Collaboration m Agricultural Research No. 5 Guatemala and the CGIAR Centers: A Study of Their Collaboration in Agricultural Research No. 6 Zimbabwe and the CGLAR Centers A Study of Their Collaboration in Agncultual Research No 7 Nepal and the CGLAR Centers: A Study of Their Collaboration in Agricultural Research No. 8 Bangladesh and the CGLAR Centers A Study of Their Collaboration in Agricultural Research No. 9 Brazil and the CGIAR Centers: A Study of Their Collaboration in Agricultural Research No. 10 Indonesia and the CGLAR Centers: A Study of Their Collaboration in Agricultural Research No. 11 Ecuador and the CGIAR Centers: A Study of Their Collaboration in Agriculhural Research NO. 12 Peru and the CGIAR Centers: A Study of Their Collaboration in Agricultural Research No. 13 Syria and the CGIAR Centers: A Study of Their Collaboration in Agricultural Research No. 14 Cuba and the CGIAR Centers: A Study of Their Collaboration in Agricultural Research NO. 15 Philippines and the CGIAR Centers: A Study of Their Collaboration in Agricultural Research No. 16 Thailand and the CGIAR Centers: A Study of Their Collaboration in Agricultural Researc No. 17 Gender-Related Impacts and the Work of the Intermational Agricultural Research Centers No. 18 India and the International Crops Research Institute for the Semi-Arid Tropics A Study of Their Collaboration in Agicultural Rsearch No. 19 Burma and the CGIAR Centers: A Study of Their Collaboration in Agricultural Research No. 20 Chile and the CGlAR Centers: A Study of Their Collaboration in Agricultural Research No. 21 The Impact of Agricultural Research in Tropical Africa: A Study of the Collaboration between the International and National Research Systems No. 22 The Intemational Agricultural Research Centers: Their Impact on Spending for National Agricultural Research and Extension No. 23 Burkina Faso and the CGIAR Centers. A Study of Their Collaboration in Agricultural Research No. 24 Partnes in Research The CGIAR in Latin America (also available in Spanish-Socios en Ia Investigaci6n- El GC4AI en America Latina) No. 25 Plant Quarantine and the International Transfer of Germplasm No. 26 Intemrabo Agricultural Researc: A Database of Networks No. 27 Governance and Management of the CGIAR Centers Also of related interest Warren C Baum, Partns against Hunger The Consultative Group on Intrationd Agricultural Research Jock R. Anderson, Robert W. Herdt, and Grant M. Scobie, Science and Food.- Th CGIAR and Its Partner 1985 Amal Report of the ConsultativE Group on International Agricultural Research Consultative Group on International Agicultuml Research CGIAR Study Paper Number 28 Social Science in the CGIAR Proceedings of a Meeting of CGIAR Social Scientists held at the International Service for National Agricultural Research (ISNAR), the Hague, the Netherlands, August 1992 Edited by Michael P. Collinson and Kerr Wright Platais The World Bank Washington, D.C. Copyright 01994 The Interational Bank for Restruction and Development/NE WORLD 8ANIC 1818 H Street, N.W. Wasington, D.C 20433, USA. AU rights rserved Manufactured in the United States of America First printing May 1994 At its annual meeting in November 1983 the Conaultive Group on Intemational Agriultural Research (CGIAR) commissioned a wide-ranging impact study of the resuts of the activities of the international agriculhual research organizabons under its sponsorship. An Advisory Committee was appointed to oversee the study and to present the pincipal findings at the annual meetings of the CGIAR in October 1985. The impact study director was given responsibility for prepanng the main report and commissioning a seres of papers on particular research issues and on the work of the centers in selected countries. This paper is one of that series. Because of the inmoality and to present the results of wearch with the least possible delay, the manuscript has not been prepared in accordance with the procedures appropriate to formal printed texs, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be infomal documents that are not readily available. The judgments expressed in this paper are entirely those of the author(s) and should not be attnluted in any maunner to the World Bank, or to members of its Board of Executive Directors or the countis they represent, or to its affiliated organizations, including the CGIAR Secretariat, or the international agriculhual research centers supported by the CGIAR, or the donors to the CGIAR, or any individual acting on their behlf Staff of many national and international organizations provided valued information, but neither they nor their institutions are responsble for the views exvessed in this paper. Neither are the views necessarily consistent with those expressed in the main and summary reports, and they should not be attibuted to the Advisory Committee or the study director. The World Bank does not guarantee the accuacy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not inply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for ncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted though the Copynght Cearance Center, Inc, Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, US6A The complete bacldist of publications from the World Bank is shown in the anuwal Index offPtblkan, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from the Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, USA, or from Publications, The World Bank, 66, avenue d'Iena, 75116 Paris, France. 1SSN: 0257-3148 Michael P. Collinson is a science adviser and Kerri Wright Platais is a research analyst for the Secretariat of the Consultative Group on International Agrincultural Research at the World Bank. Libray of Congress Cataloging-in-Publication Data Social science in the CGIAR : proceedings of a meeting of CGIAR social scientists held at the Intenational Service for National Agricultural Research (ISNAR), The Hague, The Netherlands, August 1992 / edited by Michael P. Collison, Kerri Wright Platais. p. cm. - (Study paper, ISSN 0257-3148; no. 28) ISBN 0-8213-2798-4 1. Agricultur-Research-Sociological aspects-Congresses. 2. Consultative Group on Intemational Agricultural Researh- Congresses. L Collison, Michael P. IL Platais, Kerri Wright 111. Consultative Group on International Agricultural ResearcIh IV. Senes: Study paper (Consultative Group on International Agrultural Research) ; no. 28. S540S63S63 1994 338.1-dc2O 94-13344 CIP About the CGIAR Ie Consultafive o on Ag Resech (CGA) is an imfomal anociato of 40 public and prvate sector donor that suppnot a network of 18 intnatonl agrculual research cente. The Group was establshed m 1971. The Word Bank, fte Food and Agriculture Or n of the United Natio (FAO), and the United Nations Development Prognmme (UNDP) ae cosponsors of the CGIAR The Charman of the Group is a se official of te Wod Bank whih poides the CGIAR system with a Scrf aiat in Washington, D.C. TI CGAR is asdsted by a Technal Advsoy Committe, with a Secetiat at FAO, Rome. The United States, Japan, and Canada are the leadin don countries, followed closely by severa European counties. Develping oumtry members of the CGLOAc are ima, Btazi, Idia, Indones, Mexico, Nigeria, the Plines, and the Rpblic of Korea. The annal CGIAR budget is over SUS250 milion. Intenational centers supported by the CGLAR are part of a global agrcultural sarch system. The CGIAR functions as a guartor to develpting counties, ensuringat internional scientific capacity is brought to bear on the problems of the wd's dis nged people Food duvity in developing countries has increased through the combined efforts of CGIAR centers and their partners in developing countries. The same effrts have brught about a rae of other benefit, such as reduced prces of food, better food distrib systems, better nutrtion, more rational policies, and stnger institution CGIAR cenfte have otaied over 45,000 agiculural scietists from dcvoping countri over the past 20 years. Many of them form the nucleus of and provie leadesip to natia agrcultra reserch systems in their own countries. Progams canied out by inenational centers in the CG1AR system fa into six brmad categr Productivity Research, Maaget of Natual Rmurces, Improving the Policy Envronment, ILstuon Building, Germplasm Consrvatin, and Bling Linages. v Contents About fteCGIAR ............................................ v Foreword ................................................ I Abstct ................................................ Lis of Abra banbas andnsandAcroym ................................ xv oduc ................................................ 1 Part One: ThemePapadDuon .......PaperandD....................... 3 Toward Global Germpasm and Ecoregional nisms: Impica for Social Science in theCGIAR .............................................. 4 Socioecnmic Research Agenda for Natnra Rsurce Isse in Tropical America Douglas Pachc ad Guy Hen, CIT ..................... 5 Issues and Options for Socia Scietists m Global Germplsm Imprvet Derek B$yerke l Ce................................. 7 Prioriis for Socia Scientsts in an Ecoregional Center R-hard Tutwiler and Wlem Janssen ICARDA ................. 9 New Chadiges Cofonting Socida Scient in IARCs with Ecoregioa Mandates Joyotee Sih, A.................................. 11 Prity Iues for Social Sciec atIRRI MahabubHossaU,IJU .13............................ 13 LssusandDiscusons ................................ 17 nstituions, National Pogrms, and Policy Resarch ...................... 21 An Eoomist's Contbuton to an [RC's Plans and Stategies Non Kelky, ICOSAT .......... ....................... 22 Potential I4RC Contrbums to Buiding Socid Science Capacity in NARS Peter Mon and Akin Adesia, WARDA ..................... 24 vii The Role for Social Scices in M!rs Mandate to Strengthen Ilstitutons and Link Reseach to Poicy Douglas Merrey and lssa Bandaagoda, IIMI ................. 27 Food Policy Reserchm die CGIAR-IFPRI and Other Centers See Vost, IFPRP ................................... 29 Issues and Discussions ................................ 32 The Users' Perspcives: Factring Farmers' Needs into Research Agendas ........ 36 Research Program Planning and Priority-Setting: Stengthening Farmers' Inpu Marie-HeAe Collon, Thomas Eponou, and Deborah Merrill-Sands, ISNAR .......................................... 37 Issues Rdlated to the Users' P c s Robert 7)4p, CZMMYT ................................ 39 Users' Pepectiv for Par1icIory Research and Development Doug Vennrilion, IMI ........... ..................... 40 Institutionalizing Participatory, Clientdriven Research and Technology Devlo ment in Agriculftu Jacueine Ashby and Louise Sperling, CUAT .................. 42 The Social Sciences and Participator Research in Germplasm Management Greta Watson, CIP and Jorge Quiroga, JETA .................. 45 Issues and Discussions ................................ 46 Sustainability: su for Social Scientists in Agricultal Research .............. 50 Susainabiity, Eoonomics, and Ecology: Issues and Opportunities for IARC Social Scientists Larry Harrington, CMUYT .............................. S1 Integat of Economics and Ecology in Addressing Susility Issues in A _csem Arne-Marie zac, Michael Swift, and Karen Dvork, 7TA .... ...... 53 Institional Armgements for Community-based Coastal Fisheries Management- Common Property Resources Robert Pomery, ICL4RM .............................. 58 vmii Integrting Natual Resource Management Issues into ILCA's Research Agenda Brent Swaow ad SieonA DM, 4 ....ILC4.................. 60 Issues and Discussions ................................ 61 Socioeconomic Datas, Priority-setting, and Impact Assessment .............. 67 Ibnforation for Planning and Priority-setting fPp Pady, ISNAR ................................ 68 Development of a Geofenced, Socioeconomic Data Base for African Agricultre Karen Dvorat, IITA, in collaboranon with Si5m Carter, lSBF; Mlke CoUinson, CGIAR Secreariat, Barbara Grwfin, Rutgen Univmrty; John Lynm, Rockefler Foundw ion, and Steven Romanoff, independent consulant ................................ 71 Socioeconomic Data Bases: Prospects for Effective Intercenter Collaboration Edward Vander Velde, IM ............................ 74 Assessing Impacts of Improved Control of Livestock Diseases: Issues in CoLbction and Modeling of Field Data Adrian Mukhebi, John Cr7y, and Briane, ffBAD ...Per.. ........ 75 Soioeconomic Analysis of Agroforestry Adoption and Impact: Issues and Options Suan Mbin and Steven FrwnZ, ICRAF ..................... 77 Monioring and Evaluation in National and International Research: Issues for CGIAR Social Scientists Doug Horton, ISNAR ................................. 79 Working withi NARS in the PRAPACE Network to Develop an Informafion System for Monitoing and Evaluation PeterEwel, CIP .................................... 81 Lssues and Discussions ................................ 83 PartTwo:InvitedPapers ........................................ 87 Soial Scen in Agncultural Research: Implicaions and Issues for IARCs JockAnderson, World Bank ............................. 88 ix integrting Resarch Ping, Piority-setting, and Input Evaluaion Within the CGIAR John Lynnm, RoclefeJerFo wu on ...................... 105 TAC's Perspcives on Social Science Reserch in tie CGLAR Gwido Giyss, TAC CSecaria, CG4R, ad Food and Agricuwu Oygandadon, Untd Nadons ........................... 126 Anex A Meetingof theCGLARSocilScientiS Agenda .................. 134 Annex B lit of Partcipants .................................... 139 Annex C List of the CGLAR Centers that ..................... 142 Figur1 Overiew of hamework for fedback and thnologydeve t ..... 12 Figu 2a Continuum of polcy research: pohcy gemeaon and impact, poicy and 2b filters, and policy feedback and Continuum of cofaborative activities .... 30 Figure 3 Efficiency of resource use in a cropping system ................. 55 Frgure4 eachy ofa giutusystem s ........................ 57 Fiue 5 Risk and prducivit in favorable and r- ginal envionmments .65 Table 1 Ch istics of Technocratic Cenalized and Decend Negoat odels .47 Table 2 Analyzing Equity Objecives in Ag l Resech Pianning .109 Table 3 Steps in Reserch Pnning, Priority-setting, and Evaluation .119 Table 4 Intno Agricdture Resrch Data Base Needs .120 x Foreword The themes examined in this Study Paper emerged frm papes presented and discussed at a conference of social scientists working in CGIAR centers. The purpose of the meeting was to bring a social science perpective, based on concept and expeience, to bear on the dynamics of change in intmational agncultural esearch. The effort was timely. Environmental damage, including burgeong defoston and soil and water degadation, has persuaded the international community that natal resource management and productivity have to be melded together in the pursuit of agricultual sustainability. The fict tat degrdation stems from human pressures on the resouce base reinforces the need for a people-oriented approach to agricdtural development. The social sciences make it possible for that need to be fufilled. The broad role of social sciensts at CGLAR centers -s to bring the human ftor - from fmily, community and policy levels-into the researchagenda and into the design and testing of research products. Social scientists seek to undtand the cirums of resource poor farmers, to draw on the research carried out by farmers themselves, and to create new synergies by lining moder science with the indigenous knowledge and prces of rua communities. This ensures that research products ae approprate to farmers' needs and to the social and economic ircmsnces within which they opeate. The integation of social science profussionals into fonnal agricultural reseach establishments began in developing counties in the late 1950s. Even today the relationship between natural and social scientists is stfill evolving. Within the CGIAR, a number of center programs have promoted the use of social scientists. A 1991 suvey showed that social sientsts comprise close to 15 percent of the senir science staff at CGIAR centers. Promotion arose from a conviction that under ing the piities, ircumst and decision making processes of resouce poor farmers is vital to the development of innovative techniques atctve to them. In this, as in other facets of the research process, the CG[AR has seved both as role model and as a source of support to national agricultural research systems in developing countries. For many historical reasons, the formal research estalishmets in some demeloping countries have been isolated from the resource poor majority of farmers who need most help. The CGIAR has helped to build bndges between naional research systems and local communities as well as between natual and social scientists - nafionally, regionally, and intionaly. The ideas expressed and the aaches taken in this publication demonstrate that social scientists in the CGIAR centers are poised to confrnt chalenges and respond to opportunities; widening their roles, providing essential com plementaity to the contribution of natmual scientists, and re-emphaszing the piotal function of the farmer and farming communities as catalysts of development. Alexander von der Osten ExecuDve Sectary Id Abstract A meetng was held of socidal sciis frm tie tnanal Agriculture Research Centes (IARCs) of tie Consultative Group on nernatonal Agricultbl Research (CGIAR), at tie Inrnatonal Savice for Nadona Agrculural Research (ISNAR) in the Hague, Netherlans, in August 1992. be meeidng was prompted by changes taking place within the CGIAR system towards increased investment in natural resurce ement research and emerging needs for closer collboraton among the IARCs. The meeting addressed issues identified by the soc scientists concening thar roles in the CGIAR. It focused on the following thiemes: Global germplaSM and. regional research MechanismS Links with national systems and policy research The users' p: Factoring farmers' needs into the research agenda -Sstainaility issues in agriculuaal research Socioeconomic data bases, priority-setting, and impact assessment These five themes wer supported by short issueso papers, by presentations from center scientit and three invited speakers, and by workng groups with subsequent plenary discusson sessios. This procedings is divided into th summaries of the papers and the -iscussion gups' conclusions for each theme. Te text from the invited eakrs follows dte five them sections. miII List of Abbreviations and Acronyms ACIAR Austaln Centre for Inteaional Agricultural Research AEZ agroecological zone CBM oommunity-based managemert CGIAR Consultative Group on Inernaional Agricultunl Research CIAT Centro Intemacional de Agicultura Tropical CIMMYT Centro Internacional de Mejoramiento de Maiz y Trigo CEI Centro Internacional de la Papa COSCA Collaborative Study of Cassava in Africa (IITA project) FAO Food and Agriculture Organization (UN) FPR fanner participatory reseach FSR farning systems research GIS geographic information systems [ARCs intenational agricultural research centers IBTA Instituto Boliviano de Tecnologia Agropecuaia ICARDA Intenational Center for Agricultural Research in the Dry Areas ICLARM International Center for Living Aquatic Resources Management ICRAF Intenational Centre for Research in Agroforestry ICRISAT Interational Crops Research Institute for the Semi-Arid Tropics EFPRI International Food Policy Research Institute [M International Irrigation Management Institute IITA International Institute of Tropical Agriculture ELCA International Livestock Centre for Africa EILRAD International Laboratory for Research on Animal Disease IRRI Intenational Rice Research Institute ISNAR Intenational Service for National Agriculbtual Research M & E monitoring and evaluation MTP Medium-term Plan (ICRISAT) NARI national agricutural research institution NARS national agricultural research systems NGOs nongovemmental oganizations OFR on-farm research PRAPACE Programme Regional d'Amelioration de la Culture de la Pomme de Terre et de la Patate Douce en Afrique Centrale et de l'Est R & D research and development SANREM CRSP Susinable Agriclture and Natual Resources Management Collaborative Research Support Program (USAID) SEP epidemiology and socioeconomics program (ILRAD) SSA Sub-Saharan Africa TAC Technical Advisory Committee (CGIAR) TFP total ctor productivity TSBF Trpical Soils Biology and Ferdlity Programme WANA Western and Northern Africa WARDA West Africa Rice Development Association xv Introduction The Consultative Group on Iternaional Agricula Research (CGIAR) supports an expanding network of internan cultual research centers (1ARCs) whose staffs include representatives of many disciplines. In this arena, the role of biological scientisb is evident and well defined. The role of social scientists is less clear, especily as the centers move toward missions differntiated by idther a global or an ecoregional focus. New models for natnual resource management and the attendant new possibilities for collaboration among IARCs are rasng additional questions for CGIAR social scientists. To address their roles and responsibilities in the context of these changes, the CGIAR's social scientists requested a systemwide meeting. The Directors General of the centers agreed, and the meeting took place in August 1992. The social scientists appreciate this support and particularly wish to acknowledge the assistance of Christian Bonte- Freidheim, Director General of ISNAR, who offered conference facilities in the Hague as a convenient location, and Alexander von der Osten, Executive Secretary of the CGIAR, who agreed to the editors organizing the meeting. Before the meeting, LARC social scientists identified issues that appeared crical to the effectiveness of social science in the changing CGIAR. From these issues, an agenda was developed around the following themes: D Toward Global Germplasm and Eregional Mchanisms: mpcations for Social Science in the CGIAR Institutions, National Programs, and Policy Resarch - The Users' Perspectives: Factoring Farmers' Needs mto Research Agendas * SustainabilitT. Issues for Social Scientists in Agclturmal Research * Socioeconomic Data Bases, Priority-setting, and Impact Assessment The agenda's issues were suported by shotpapers and presenttios by center scienfists and three invited speakers, and by working groups and a subsequent plenary session. This publicafion presents summaries of the papers, the discussion groups' conclusions, and the text of the invited speakers' papers. Consensus at the meeting identified four broad roles for IARC socia scientists: * To conduct strategic social science research to improve understanding of human behavior at the fann, community, and policy levels. 1 To contnrbute to hrea that will produce technologies usefil to farmers. To participate in capacity-building in a range of national oions, with national agricultural research institutions as a dominant target. To assist management in priority-settng and planning in research institutions. Participants agr'4 that the integration of social science into agricultural research at the national level remams a fragile process, suggesting continuing role model and support responsibilities for IARCs. Other consensus views emerged, among them: * Social scientists in the centers can make a valuable contribution to strategic and program planning and impact assessment. However, managers should exerise caution in makdng social scientists into planners by removing them from the agricultual research process, where partnersips with technical scientists need nurturing. * It is urgent that centers pool knowledge and learn as cheaply and quickly as possible how to undertake new roles. This is especially true in the ecoregional context, but it also applies to building social science capacity in national agricultual research systems (NARS) and other activities. - Intercenter collaboration can be an effective means of using scarce social science capacity effectively. Participants identified a need for greater coherence: a coherent CGIAR-wide position on the roles of social science in agricultural research, a coherent approach to building capacity in NARS, a coherent view about how NARS should approach priority- setting, a coherent (and replicable) means for articulating farmers' needs and reconciling them with national priorities. Participants strongly supported the idea that enhanced communication between socil scientists across all IARCs would bnng rapid and significant benefits to the CGLAR and its constituent centers. New vision is needed to exploit synergies across the CGIAR system. The meeting itself demonstrated how communication reduces duplication, avoids rediscovery, and offers faster progress in methods development, application, and institutionaliion. The editors wish to thank Myra Green for her hard work and persverance in helping bring this document to life and Barbama Shapiro for her expert technical editing assistance. 2 PART ONE THEME PAPERS AND DISCUSSION Toward Global Germplasm and Ecoregional Mechanisms: Implications for Social Science in the CGIAR Centro tecional de Agriculta Tropical (CIT) Douglas Pachico Guy Hery Centro Itrnacional de Mejoramiento de Maiz y Trigo (CIMMY7) Derek Byerlee Internaional Center for Agicultual Research (CARDA) Richard Twwir Widem Janssen Internahional istitute of Tropical Agricultuie (IrrA) Joyotee Snih Intemational Rice Research histitute (IRRM) Mahabub Hossai Introduction The degradation of natual resources increasingly thetens food supplies for grwing pations. Since 1990 a series of papers from the Technical Advisory Committee (TAC) developed the concept of an inteonal agricultual research system that would evolve a dual structure duing the next decades. The concept distnguishes two sets of activities: mprovement, policy, and management of germtplasm, which can be aomplished most cost effctively at a global level Each stategic research trs offers potential benefits to a range of developing countries from spillovers. * Research and research support activities centered on natural resource management and national program capability for the special circumstances of designated ecoregions. Eventually, as the enhancement of local capacity allows, this second set of activities will devolve to reonal institutons or to collaborating national systems. TAC notes that such a two-level strucure would be possible only with unrestricted collaboration. This emphasis on collaboration acknowledges at least three inportant qualificaion the sepation of the two sets of activities. First, the smained t of pducity is a prerquisite for long-term success. More research on natural resource management will underpin continuing imov ents in productivity. Second, the global germplasm centers already orgnize much of their work in an agrecological framework. Third, global programs 4 are always heavily dependent on local collaboration to validate their material under differt conditions. CGIAR histy sugges that reinal intmational agricultual rearch cente (ARCs) whose mandate includes both commodities and natual resource management focused narowly on commodity management Historically, commodity, research offered a clear focus, and there was a strong tradition in the management of crop-specific research program. A more holistic approach was introduced with advent of systems thinking. Tiis brought concepts and methods beyond physics and chemistry to bear on natural resource management, emphasizing the interaction between physical resources and human activity. It was one factor that stimulated TAC to propose a dual structure for intnational agricultural research. The first day of the workshop was devoted to exploring the roles of the social sciences in this dual structme and identifying and discussing the issues raised for social science in the CGIAR. To open the session, a paper from Douglas Pachico and Guy Henry explored CIAT's onomic researh agenda for natural resource management research in tropical America. Derek Byerlee looked at social sience roles in CIMMYT as a global germplasm center. Other presentations by Richard Tutwiler and Willem Janssen from ICARDA and Joyotee Smith from IITA amined aspects of the dual structure for centers with both ecoregional and commodity mandates. Finally, Mahabub Hossain addressed the future social sience research priorities of IRRI, a germplasm center with programs based on agoecologies. tocioeconomic Research Agenda for Natural Resource Issues in Tropical America Dogiz Paclico and Guy Hewy, CI4T In their overview of CIAT's new Resource Managment Division, the authors explained the three priority agoecos : the firest margins, Central America and the Andean hillides, and the savannas. These three were chosen for their economic and poduction potental, the promnence of their natural resource problems, their relevance to oveC ng p y, available technological oppouities, and institutional considerations. Common to all tbree agr yss are acid infertile soils and the need to understnd decisionmaing processes from the individual farm level to the regonal and policy levels. Such sonom ic studies reesent 42 percent of the core resources of the Resource Management Division. A summary of issues that may become central to the sociconoc research agenda follows. 5 Land Use Progam In this umbrella program, servicing and synthesizing from the three agroecom programs, geographic information systems (GIS) will be used to identify land use systems, trends in land use, and the factors that drive change. Economic analysis will be an important tool for appraising land use and policy options. Understanding the private and social costs and returns of alternative systems of resource management is the key to establishing the agroecosystems' priorities for technology development. Policy will play an important role, determining access to natural resources and influencing returns to pnvate exploitation. CIAT will concentrate on methods to assist national policy analysts in presenting policymakers with a range of options and will work closely with IFPRI. Agroecosystems P ams Tle three agroecosystem-based programs will study the decisionmalkng of farmers, emphasizing analysis of farmers' strategies for resource management and choice of technologies. For the Forest Margins Program, this will mean intenfying and stabilizing shifting cultivation systems on already cleared land. Market channels and oppotnities will have ignificant influence in the adoption of new technologies. The role of women as resouce managers also will be examined. The Hillside Program will probably focus on encouagng fanners to invest in resource regeneon by increasing productivity and incomes. Analysis of linkages from agroindustry and markets will identify opportunities to introduce technologies that can increase incomes and capital accumulation, thus providing the incentive to invest in sustinable resource managent. An important effort will be strategic research on mstitutional innovation that links indigenous expertise with a more decetralized formal research system. The savannas provide unique opporatmities for inresed productivity. The Savannas Program will recognize that successful and sustainable development of the savannas demands new fiaming and soil management systems. The highly acidic soils require special management, and acid-tolerant germplasm is seen as the pnncipal technology option. The challenge to socioeconomics will be the profitability and sustinty of designed farming systems. Formal economic modeling will be a useful tool, but technologies will need to be tested on farm. 6 Pachico and Henry highlighted three factors that may affect successful implementation of CIAT's Land Use Program: * The complexity of the problems makes the setting of research priorities of tremendous importance. * Valuation problems, particularly extemalities, may not be settled easily. * Monitoring of research progress will be critical to maintaining financial support for research. Issues and Options for Social Scientists in Global Germplasm linprovenent Derek Byerlee, CIMMYT Several global centers-IFPRI, ISNAR, and IMI in particular,-are dominated by social science. Acknowledgng this fact, Byedee's presentaion and the discussion focused on the role of social science in global germplasm improveent The par highlighted new opportities for social science in germplasm research and defined organizational issues in exploiting these oprities. Byerlee outined three ars of socil science research historically addressed by global gemplasm centers: - Activities focused on the generation and adoption of technology. In these activities, ex-ante participation in design and evaluation is necessary to ensure that social and economic perspectives (from farmers or society) are addressed in the decisionmaking process. By feeding back information and anayzing input supply, ex post studies of adoption also may be usefil for the technical scientist in designing new technologies. * Assessment of research resource allocation and impact studies to docment reseach 'payoffs' for resech managers. Together, they can improve the efficiency of a research progrm and prvide infomation to donors. * Analysis of the commodity sector. This prv.ides a general overview of the circumstances surrounding research decisions, for example, the analysis of supply and demand trends for a specific commodity where the 'users" are individual researchers or managers. 7 Historically, social science has been oriented largely to crop and resource management reseh. Social scientists have conducted little interdisciplinary research in collaboration with plant breeders or representatives of other disciplines worldng on varietal development. Faming systems research (FSR) has helped in this area. Byerlee examined important implications of FSR findings for the centers' breeding priorities. Because FSR is not yet well linked to national commodity research programs, however, it is often unclear how site-specific information from FSR can be used consisntly in setting pnorities for centers with global mandates for germplasm research. Similarly, social scientists have done little work on breeding stratgies and on the efficiency of resource use in crop breeding programs. Byerlee cited two questions: * What is the cost-effectiveness of increasing productivity through breeding for low-input conditions compared with changing input levels? Few research data are available on this subject. * What are the costs and benefits of adaptation of center materials by national agricultural research systems (NARS) compared with the costs and benefits of screening imported varieties? Despite the fact that IARCs and NARS typically devote half of all research expnditures to germplasm inprovement, no effort has been made to model altenative plant breeding strategies in economic terms, ncluding a measurable set of inputs and outputs and an array of techniques from which to choose. New Opportunities Byerlee cited global trends in germplasm development and the new research oppumnities these bring. Social scientists should play an inportant role in setting priories for the use of biotechnology tools and techniques, paricularly in the ex-ante assessmnent of their social and economic consequences and in the monitonng and ex-post evaluation of their effects on small farmers and rual Laborers (e.g., transfer of herbicide resistance genes to cereal crops). The growing priva on of research and changing pespectves on intelectual property rights have increased research activities in the private sector in industrial counties. Many see the IARC global centers as a conduit for the transfer of these new biotechnologies to developing countries. Social science may be called on to place economic values on speci genetic traits (e.g., disease resistance) available from the private sector. The recent Earth Summit in Rio failed to find consensus on economic and equity issues asoated with genetic biodiversity. Several mainstm m economists have begun analyzing tradeoffs in the conservation of genetic resources; this work will have important repercussions for the questions facing IARCs. 8 These rnities Se a seies of issues. Social scientists at the global LARCs in the next fiw years will need to strengthen their capability in socal scence analysis and to balance their agendas among field-level activities promoting 'user' perspwective in technology design, activities involving the new biotechnologies and intellectual property rights at the global level, and serving the needs of management in an era of shrinldng budgets. Before this balance can be achieved, a number of organizational questions must be answered: * Will we need more specaliztion of social scientists by subdiscipline? * How can we improve linkages with plant breeders to provide perwectives in technology design? * How can we improve the collaboration among social scientsts working in germplasm-related activities at the various CGIAR centers? * Can we-and should we-ivorce the work of social scientists in global research centers from the work of those in ecoregional centers? Final questions remain. The slow development of social science capacity in NARS has been disintng. In many ways the CGIAR social scientists also serve the NARS. (CDIMMYT, for example, has had severl regional social science progams. For more than fifteen years, the one in Eastern and Southern Africa was devoted exclusively to building the NARS social science caacity.) What is the apppate division of labor for social scientists in NARS and in the CGIAR centers? How much CGIAR social science zapacity should be invested in building capacity in NARS? Priorities for Social Scientists in an Ecoregional Center Richad Ttwler and R em Ja,wsens ICARDA The authors see social scientists within the CGIAR system as part of multidisciplinary teams that use a systems approach to incorporate farmers and their concerns in the research process. They perceive the new ecoregional stategy as bringing increased emphasis on resource management as opposed to individual crop improvement This implies added weight on a holistic approach in which problems are identified, research topics are concdved, and priorities are set at the regional level. Because resource conditions are site specific, intensive interaction with NARS of the region will be essential. Tutwiler and Janssen emphasized the importance of setting priorities for social science activites. Li selectig priorty activities, ifluence within the center, influence with NARS partners, and the agricultural sector of the regon are always considerations. Although the authors appreciate the important contribution of germplasm improement, they noted that 9 without attention to resource n nt, it wil not sustain prductivity increases. It is necesary to ensue the quality, utility, and reevance to faners of center research products. Tutver and Janssen believe that resource management research must be given greatr prioity. They highlighted systms development, in which ICARDA social scientists wll focus on evaluation methods, especially ex-ante assessmenits of altenative rotations and cop-livestock inteactions. The success of newly developed systems in Westen and Northern Afica (WANA) often depends on the application of policy incentives, and policy reserc must have a higher priority. Common property manaent may be the highest priority policy issue because of its implications for resource managemet in WANA. Spealing fom experience, the authors stated that they are more effective when teamed with technical scientists than when they work in isolation, and that time is most usefully spent paticipating in the technology-generation activities of the center, rather than puung an independent policy research agenda. Where policy and technology interact, they recognize a need for their strong involvement. Filly, the authors regnize that through technology adoption studies, social science research can make a significant contribution to the image of the center and to its accountability in the eyes of donors, sponsors, and other stakeholders. They drew the following conclusions: i ICARDA's goal is to raise the state of the art in applying socioeconomic methods to concrete regional problems. This includes diagnostics and partcipatory work with resource users. - Technology adoption and impact field studies are the bread and butter of soci science work because of the importance of work showing impact of the center, impact on NARS, and the center's accountability to donors. e Support for resource management research becomes the new prority for the social sciences at the center. As more of the technology adoption and impact work is assumed by national scientists, ICARDA scietists will shift toward natual resource nagement and conservation. Emphasizing that this expected shift leaves unanswered questions, Tutwiler and Janssen identified four of these * How can we best decide the divson between centralized (headquarters) and decentralized (outrach) social science research in an ecoregional center? * Is the social science component of germplasm imprvement better conducted by an ecoregional center with a restricted commodity mandate or the relevant 10 global mandate center (e.g., wheat in WANA, cassava in Africa, rice in West Africa)? * What should be the balance between center social scientists' attention to NARS needs and those of exaregional development institutions with global concerns? * To what degme is it possible and desirable to sepaate policyr seach from support for technology generation rearch? New Challenges Confronting Social Scientists in IARCs with Ecoregional Mandates Joyotee Smith, IITA The ecoregional concept is still relatively new to the CGIAR. No centers are designated exclusively ecoregional; however, some centers that have assumed the responsbility for resource management research also will have mandates for developing germplasm for principal CGIAR crops, livestock, fish, and trees. An informal survey among scientists at IITA identified the most important function of CGIAR social scientists as providing guidance on technology development. Because the adoptabily of technology is a concern to many, ex-ante and ex-post impact studies also were consdered important in this survey. Socal scientists should take the lead in defining what is good and adoptable for ea resarch effort, according to resondents. All scientists suveye agreed that ex-post impact studies that emphasie esimating returns to research should be a low priority. Providing Guidance on Technology Development in an Ecoregional Context A dominauing question is how feedback from farmers in a highly heterogeneous mandate area (ecoregion) can be organizd and then synthesizd to allow identification of priority research themes for the center. Smith provided a framework for answering is question (see Figure 1). She idendfied three areas that require stategic systems research before the framework can be fiy operational: * Charactrization of the dynamics of major village systems in the mandate ar2a * Establishment of criteria for matching solutions to the characteriscs of tageted land use systems. 11 Figure 1. Overvlew of framework for feedback and technology development Broed4based ---afln Characterizatlon of Ruture Im oodcnctr oo l S viage systems and constraint In ptae agreoolsof village systseanmyses n In mandalt area agroecology Ust of control strategies by prorityf Strategic researh on cornUl F stabglby lARCs Generic technical optons for each type of viage system and agroecology -Selecdon by NARS of aPProprlat ognecthkl jopfions lor spedit village spsems opin_yNR Adaptive on-farm Oavebpment of flSshed research with acfive tednobgbls adapted to afarmer parfidpation specilc vfllage syslms 12 * Development of interactions among the constaint to the adoption and effctive use of technologies, agroecologi, and village systems to predict the future impoaince of the constaint in tie mandate area. Once limiting constraints, such as lack of soil nitrogen, are identified by mandate area, the franmwork can provide guidance on the allocation of research resurces among optional solutions, thus setting priorities for the research agenda. The need for ecoregional chacization that addres sustinability prests several implcations, which Smith cited: * A sbift in focus to the levels of village and land-use unit or watershed-the approriate scale for managing sustainable land use and for assessing the acceability of resource management technologies to local farmers and their community * Bamers to the shift toward sustainability presented by the fact that many of the biophysical processes are sill to be quantified * A move from the static to the dynamic, requiring study of the process of cdange in fning systems that collects longitudinal rather tan cross-section data, including data on the evolution of land distribution nghts Inplctu and 1aues Smith emphaized the need for longer-term investments in strategic social science reserch befor results can be obtained with a wider spectrum of discplines. Mechanisms for nterntitute collaboration may shorten this period; however, such collaboration requires coordnated data bases acrss the mandate area. As social science enters the eoregonal era, biological scientists are shownng a greater appreciation of its contribution. Effective methods and a collaboratve mode could hold the key to the IARCs' making an impact in risky, heterogeneous, dryland envioments. Priority Issues for Social Science at IRRI Maaibub Hossain, IRRI Hossain explained that IRRI follows a matrix approach in conductig research, identfying research problems by four major rice ecosystm: irrigated, rainfed lowland, upland, and deepwater and idal wedtands. The research problems are then approached throug integation of a wide range of disciplines. In addition, a broad cross-ecosysten research pwgram focuses interdisciplinary attention on problems not specfic to any one ecosystem. To promote interaction with national scientists and dissmination of rsearch 13 results, IRRI uses nation pogrms of network and consordia. The networks coordinate appLied and adaptive research, and the consortia advance strategic research at Irpresentative sites for differet eaostems. The Reappearig Food Probkm The food self-suffiiency atio in most Asian countries would have declined without the remarkable grwth in productivity that moden varietie brought to Asian rice frnning. The Asian share in total world rice imports doclined sharply from about 60 percent in the 1960s to 20 perent in the late 1980s. Hossain pointed out signs that the favorable production situafion of the two decades since 1965 may have begun to change. The growth in crop yield has started to decelerate in China, Indonesa, and the Philippines (and for Asia as a whole), and yilds have stagnated in Japan and Republic of Korea at a level that will soon be reached by China and Indonesia. The growth in crop yield has acceleated in Inda and Bangladesh, but within South Asia, the deceleration in yield growth is apparent for those reipons that have already achieved high levels (Punjab, Tamil Nadu, and Karnataka provinces hn India). There are also indications dtat the area under rice cultivation may decline in absolute terms in the near future. In China, annual increase in rice area declined from 354,000 haecares in 1965-78 to a negative 189,000 hecres during 1978-90. In India, area under rice cultivation continues to grow, but the rate of annual increase deceleted from 386,000 bectares to 110,000 hecrs during the same period. With apid growth in urbaniation, prme rice land is lost to housing and industry; salinization and degrdation of the irrigation systms also are reducing the volume of irrigated rice land. Thus, future growth in rice production must come from a reduced land base, yet the global rice-eating population is still growing at 2 percent per year. In view of these trends, it is pertinent to ask if Asia will be able to meet the future demand for rice of its fast-growing populaton, given the present level of technology and prices. Policymakers now face the choice of accepting higher relative prices of rice or developing and maintining low-cost production technology to provide production incentives to frmers at existing prices. A sigificant increase in the relative price of rice will have an adverse impact on landless laborers, marginal farmers, and urban consumers, and hence on human development and alleviation of poverty. Social scientiStS in geral, and those at IRRI in particular, must consider what combzination of resource and policy among the alternatives below will lead to efficient, eqitable, and sustainable growth in rice productivity: * E on of irrigation infrastucture or improved management of natual resources? 14 * Increase in allocation of researh resouirces to irrigated or to less-favorable nce ecosystems and within each ecosystem, to genetic improvements or natual resource management? * Corrections in price distortons in sectoal and macroecnomic levels or further investnent for pushing forward the technology frontier? Priority Remarch Areas The Social Sciences Division at IRRI has identified the priority rearch activities under the different rice ecosystem-based research programs for the 1994-98 period. Irrigated Ecosystem * Identify factors contributing to the reported yield decline of intensive rice cultivation systems. * Measure extemalities for technologies requiring intensive use of chemical fertlizrs and pesticides. * Evaluate on-fann experiments on hybrid rice and high-yield sustainability research. * Understand factors that affect the adoption of improved nutrient and pest management technologies. Develop methods to assess the impact of technology diffusion on the operation of commodity and factor markets. Rainfed Lowland Ecosystem * Charactrize different subecosystems to guide research design and idenify criteria for evaluation of research results. * Estimate production variability, assess farmers' response to risk, and draw implications for policy. * Undertake constraint analysis to determnine yield gaps and the potential for incraing productivity. * Study the operation of labor and tenancy markets for ex-ante evaluation of biological and mechanical technology. 15 Upland Ecoysm * Analyze drminants of low and variable rice production and study economics of altetive land-use systems. * Evaluate on-farm experiments to develop farer-apprpriate management alternatives for soil conservation. Deepwater and Tidal Wedands Ecosystem * C:haracterize production environments to identify opportunities for rice research. * Study cropping systems to identify constraints to optimum land use. * Evaluate on-farm experiments on rice and fish cropping patterns and on modifications of irrigated technology for use during dry seasons. AU Ecosstems * Develop socioeconomic data bases by ecosystem and by agroecological zones. * Undertake analysis of rice sectrs and draw unplications for policy. * Underake village studies in collaboration with NARS to understand how biophysical and socioeconomic factors affect farmers' decisionmaking in exploitaton and managemen t of natal resources. - Study ewpected pay-offs from different resarch activities to help set priorities for rice research. * Analyze rice research capacity in NARS to assess the comparate advantage in addressing vaious rsearch issues between IRRI and NARS. * Assess returns to the use of on-farm resurces, especially libor, and idenify constraints to adoption of knowledge-intensive technologies. * Evaluate the roles of disadvantaged groups, partcularly women, and assess the impact of existing emering technologies on their productivity and welfar Hossain noted that the matrix approach to research management followed in recent years at IRRI has helped focus socil science research issues. ChaLenges and problems are now idenified in ecosystem-based progams in which scientific inputs are provided by 16 discipline-bsed research dvsions. Program leaders now require he social sciene contributions. This fact has focused resources on what IRRI considers pnonty research and has facilihtted integration of scal sciences witi otier disciplines. Nevetheless, scil science resources remain scarce in the center. These resourc could be greatly augmented if IARC scientis allocate more time to strengthen NARS social scienists, allowing them to take up functions now prformed by IRRI. Issues and Discussions Four working groups covered several topics. Discussions are reported under two heads: * The evolving role of social science in global germplasm activities and how germplasm contributes to natural resource manaent research * Making social science effective in coregional research The Evolving Roles of Social Science in Global Gemplasm Activities and in Natural Resour Ma 1r The maintenance of biodiversity was a focal point of the group's disussions. It included inigenous conservion of germplasm, farmer evaluation of germplasm, and changing patterns of release of advanced mateials. Three challenges were identified for social science: * To develop cost-efftive methods for identifying farme-preferred varieta Ncharaeisics ealy in the breeding process To help detemine which varietl characteistics are best intmduced at the different stages of the breeding process, and the implications for feedback from social suience, * To provide input on many new issues, such as the importance of genetic engineering, including poperty rights and productivity, equity, food securty, and gendr implications of genetic material developed by pnvate industry The contnbution of germplasm to natual resource research was discussed. Clearly, host plant resistance is an important part of integrated pest management, and tolerance to unfavorable soil conditions such as acidity and salinity will continue to be important. 17 Two ohe questions were raised: * How are trade-offs weighed between natural resource managemit and genetically based strategies for meeting specific objectives, for example, breeding materials for low available inputs compared with resource management strates to increase inputs available to the system? * How do genetic improvemets affect cropping patterns, rotations, and other agricultural system charcteristics that help manage sustinability? Making Social Science Effective in Ecoregional Research The main functions of an ecoregional approach to research as perceived by the CGIAR are to address natumal resource and sustminability issues and to provide a framework for regional coordination and collaboration of IARCs with national programs. For social scientists, important aspects of sustinability include the ability to charactize systems at higher levels than the commodity, the development of a paadigm for natnul resource management research, and exploitation of synergies in the effective combination of crop gerpasm, crp management practices, livestock, and policy. rganizationally, the most obvious approach to ecoregionalism issues is to mandate ecoreonal responsibilties to IARCs. Altenative evolving mechanisms are the intercenter --o-nsrium and national groupings in an ecoregional context, for example, the Southern African Centre for Cooperation in Agricultmal Research (SACCAR). The working group suggested that further resolution is needed on the following points: * The cost in coordination and management of assgning ecoregional functions to centers may be considerable. * Global priority-setting and planning for CGLAR centers may not be consistent with the more decentalized pnority-setting and planning within an ecoregional fraework. * Eooregional orniton of center activities should not interfere with the centers' need to link basic, applied, and adaptive research to their mandates. * Centers that have mixed mandates (both ecoregional and commodity) may distort the commodity pniorities of NARS in their ecoregion by favoring IARCs as collaborators, and thus favoring IARC commodities. 18 Tbe group decded that an cregional center should first seek to understand the cosytem in its region and define teir priority needs. Within this foundation, an ecorgional center needs to match ills to the chalenges and identify research areas where outside intituions or NARS have a comparative advantage, including a valuable combination of people and skills. The ecoreional center would tak the lead on finding funding and organizing priority reseach activities for its region. Linkages might take many forms. Options include seconded staff to achieve cntical mass, separate but shared responsibilities in a consortium fraework, and the use of workshops, networks, and other common channels of communicftion. Joint training progams also could be implemented and further work encouraged though NARS and nongovernmental organizations. The group saw an important role for social scientists in ecoregional centers to help identify the social or economic causes of physical processes of resource degradation. Social science should takle the lead in regional data bases, GIS, agroecological charateriion, and definition of ecosystems in the mandated region. Social science also must address policy issues relevant to the mandate and consider the extemalities, particularly as they pertain to the suinabty of naural resources. Questions were rAised about the accountability of ecwregonal center research and how to measure its success. Social science has a role in developing new criteia for evaluation and impact assessment. The discussion acknowledged that the ecoregional approach had much in common with farming systems research. It noted that FSR had often been commodity-led and squarely focused on traditional productivity issues rather than on resource quality. In this tradition, FSR rarely consdered time periods of longer than one season, did not deal with extemahties, and, with recent exceptions, had accepted the policy environment as a given. It had evolved as an apprach to adaptive research. The strong strategic research dimension required to address the human/biophysical interface at the center of the suinability issue, and the linking of several levels in the decision hierrchy (farm, community, enabling institutions, and policy nisms). The queston of how to deal with the policy and technology inteface was raised. In ecoregioal research, progress wil be difficult to attain unless policy and institutions are maleable. There is a need for models that assist technology design by evaluating alternatives in contrastng policy envionments and idenfifying ways in which policies might best be amended. The group suggested that ecoregional centers may have to contact out policy work; centers will not be able to afford special policy staff to cover aU the situations that arise. IFPRI staff noted their desire to work closely with other IARCs in this area. Coclson The papers and discussions during this session saw a continuing role for social scientists in global activities and a widening role in ecoregional activities. The discussion of global germplasm ire t highlighted the need to improve interaction with breeders and identified a new agenda for social scientists. Ecoregional activities require social science 19 kiLls at severl lnd levels: the commodity, the housho, the commuity, institutonal opetion, and policy formulaon. The wider roles mus be recnd with te ned to build social science capacity in naional research sysem. 20 Institutions, National Programs, and Policy Resarch lCrps R arch Insfftt for the Semi-Arid Tropics (ICRISAl) Thn Kellq West Africa Rice Deveopment Astio n (WARDA) Peter Madon Akin Ades.a InternatioaInsDttu of Tropical Agrcultu (1TA) DougJas Merrey 7Bssa Bwidargoda Food Policy Research Institute (IFPRI) Se Vosa Intruction Managemen t of the nterface with NARS has been a coinuing deba within the CGIAR. NARS are the foundain of the globa agricultural research system through which the CGLAR crop and livestock imprmt cet hae engaged heir di to fulfll their mission. Without strng NARM to mobilize results, thee is no justifican for intemationa riultural esearch. BRecgng this, the CGLAR included a capacity-building role in the mandate of each IARC. In 1980 fte Group estblised ISNAR, whoe missi is to help build ma t skdlls and impove institutional capcity in NARS. EFPRI, wich joined the CGIAR in 1978, was the firt in th Syem to focus on naional isituons oe than NARS, often the minise of d o tplanning and finane whr policy is formulated. Several of the new censm join the CGIAR in 1991-92 (1M1, ICLARM, ICRAF, and CIFOR), in particular, have fiur widened the institutional inoerfice as they deal with those national bodies most approprat to the mandates. The progams of these new centers also have a greater polic contet than the traditioal enters. Together with te ecoregional approah to rsearch recety espoused by the Group, which emphasises the inter ce of technolo and polcy, this has reewed emphasis on IFPR's collabo nwith othe L4RCs. The past fifteen years have seen a gradual prolferation of nationa instions iteracting with the CGIAR and its IARCs. The definitin of NARS has ibelf widened. Historily, the memondums of ud ig with co for an LAlC presence have fead the natoal agricultal research institution (NARP) as the impe ing arm of govrnment. Recently it has been rcognized that universities in devepig cuties shold beencouraged in theirimportant, complemntary research rl It also has beenrogzed that nongovenmental ions (NGOs) are incrasingy acve in adaptive reearch at 21 the gassoots level. Both universities and NGOs are ineasingly seen as important potential parn, pardy oNas of the continuing constaints on NARI opating budgets in many countries and the weak prospects of increased political commitmet to effective NARIs. Within NARIls, social scientists are scarce and often inexperienced. Widening social science roles in agricultural research are confusing managers about priorities for their limited human resources in these disciplines. The resulting uncertainties adversely affect the morale of their social scientists. This session of the workshop was built around these institutional and policy issues. Tim Kelley's paper focused on the social scientist's role in assisting IARC management. It was followed by Peter Matlon and Akin Adesina, who addressed the traditional interface between NARS and IARCs from a social science perpective. Douglas Merrey and Tissa Bandaragoda widened the discussion to national institutions beyond NARS, and Stephen Vosti extended it to the policy arena. An Economist's Contribution to an IARC's Plans and Strategies Tim Kelfey, ICRISAT Keley deineated the functional domains of an economist at an LARC and discussed the comparative advantage that person may have in assisting management decisionmaking, particlarly in setting research priorities. He identified three functional domains: bMainstream economics studies carried out within the disciplinary group, for example, factor markets, commodity markets, risk, production relations, rural welfa, policy Assessment studies, for example, technology evaluation (ex ante), adoption studies, caaizon, and diagnostic analysis-often carried out in collaboration with other groups within the resource management and commodity improvement programs * Research management support, for exaample, priority-setting, resource allocation methods, impact appraisal, exploratory studies, information and analysis to support managemet decisionmakng Clients for the first function are disciplinary peers; the research builds on and contributes to the existing body of economic theory. Primary clients for the second function are IARC and NARS scientists. Much of the racterization analysis done by economists and other social scientists at IARCs helps to formulate the demand for new technology as well as to define future resrch challenges. The better scientists undernd the interaction between the agroclimatic and socioeconomic evirnment, the more likely they are to design and target tecnology effectively. 22 The third function has several dlients: first, managers of the rrch institutes (supoting thieir medium- and long-tm decisionmaing); second, donors (documenting LARC successes to demonstrate tOe soundness of their earlier inveshnents tirough impact evaluafion); and third, developing country governments (convincing them of the need to invest in research). Kiley showed that the economist has a comparative advantage over other scientis in this role. storical ive Dung the past twenty years of economics research at ICRISAT, the economics group has focused on mainstream production economics and technology evaluation studies; however, economists have often contributed-either as part of a team effort or individually-to management decisionmaking in the following areas: * Impact assessment (ex post). Quick analyses provide management and donors with information about the impact of ICRISAT research. * Priority-setting (medium term). In 1981 an innovative approach to . termining regional allocation of research resources to mandate crops of ICRISAT employed a combination of ten criteia chosen for efficiency and equity and was instumental in formulating a ten-year strategic plan for ICRISAT. Along similar lines, and in prepaation for the Strategic Plan to the Year 2000, an ICRISAT economist examined the cufrent status of research resource allocation and impact appraisal at ICRISAT. i Exploratory assessments of long-term strategic research proposals. One of the first assignments carried out by an economist on behalf of management was to consider whether ICRISAT should develop a research program on groundnuts. There have been others. Typically, these assessments have been driven by the needs of institute-level management. A Recent Example ICRISAT's 1994-98 Medium-term Plan (MTP) planning process began in September 1991 with the first meeting of the MTP WVorkxing Group, composed of senior scientists from each of the commodity groups and an agroclimatologist and economist from the Resource Management Program. Apart from the chair, only the economist was designated full time on the MTP. Some examples are given of the economist's critical input * HElping to translate the director general's guidelines (analytically rigorous, clear critea, transparent procedures) into a comprehensive approach and methodology and develop step-wise procedures for implementation. 23 * Clarifying terms and concepts, and in some cases, suggesting or developing new concepts to move the process along when it was stalled. * Developing pro foona data to elicit data from scicntists and ensure that a common approach for estimating data would be used. * Organizing, assembling, and (in some cases) generafing socioeconomic profiles of all crop and resource management (soil- and climate-based) resarch domains and ultimately deciding which variables would be used in the analysis. * Estiating the likely impact of each research activity in light of the goals ard objectives of the institute. * Translating the ICRISAT mandate and CGLAR Mission Statement into clear objectives for the MTP exercise, objectives into criteria, and criteria into measures for whi h data were available. Ultimately, an institute-wide ranking of research themes was generated, and prionties were selected accordingly. The role of the economist in providing information to IARC management has become nceasingly important. Evidence comes from the substantive input of economists at other lARCs (e.g., CIMMYT, CIAI) to their recent stategic plans and the recent recruitment of an -assessment economist' at ICRISAT, CIAT, and ICARDA. From the professional's point of view, this role of support for research management may be seen as a diversion from other, more rewarding, areas of research; however, by helping to build a strong case to donors for a greater commitment of funds for agicultumal esearch, the economist provides value to the center. Kelley noted that it is but a short step from donors toD xpayers, whose money should be put to the best possible use in assisting development that justifies the economist's role in management decisionmaking. Potential LARC Contributions to Building Social Science Capacity in NARS Peter Matlon and Akin Adesina WARDA Economists witiin NARS are often perceived as interlopers in an establishment taditioally 'owned' by technical sientsts. As such, they face a set of strucural and attibtdinal constraints that limit their potential contributions to meeting national research goals. 24 Problems of Economists in NARS Economists working within multidisciplinary teams are often looked upon with a degree of misunderstnding or outright suspicion by their technical scientist colleagues. In part, this can result from lack of recognition of the potential contributions of economists to technical research. Some technical scientists see an economist's role as limited to policy and market studies, or profit-and-loss "bean counting" at the plot level, or explaining superstitions and social curiosities of backward* peasant farmers as related to technical change. Because of their ability to synthesize, some economists are located at headquarters levels to prepare reports and give policy advice. Without the underpiming of sound research results to perform these tasks adequately, they can be perceived as public relations agents or worse. Suspicions are often exacerbated by the fact many economists have little understanding of even basic biophysical factors in agriculture and are thus il prepared to discuss technical elements with their colleagues. Another factor contributing to the uneasy relationship between economists and othe scietists within NARS is the perception that the priority given to economics research is the result of donor demands rather than national needs. During the fanning systems era from the late 1970s through the mid-1980s, a large number of well-funded donor enclave projects were planted acmss Africa. Most plans to restrcture NARS, developed by missions of expatiate donor experts, have included a prominent role for economists. A common perception is that economists try to have impact by discrediting the products of tehnical scientists through negative assessments of technologies tested on farm. When such results are presented in a critical rather an constctive fashion or obfuscated with unnecessary jargon, the potential usefulness of economists' findings is often never reazed. Few economics training programs in African, European, or North American universities provide guidance in techniques of on-farm research or experience in working in intdisciplinary teams. Too often young researchers who graduate from foreign universities join their national programs armed only with disciplinary knowledge set in alien envnments tat are poorly matched with the reality of the African farmers' situation. Even within Afican universities, thorough study of on-fam research methods is still generally laing. Given the budget situation in NARS, economists not supported by special enal funding face debilitating resource constraints that prevent effective on-fam research. The high costs of stationing staff in study villages and of internal transportation can be prohibitive. NARS economists respond by reducing the number of sites below optimal levels, cutting back on supervision to a dangerous degree, restricting study sites to unepresentative 25 locations (often nearest to research stations), or all of the above. Altmatively, they give up and concenate on library research. Finally, NARS economists are often isolated. Significant institutional barriers typically prevent coLaboration and often consultation between NARS and university economists. Poor library collections also isolate NARS economists from the current literature. Support from IARCs Most economists in IARCs have experienced similar problems in developing fruitful relationships witi colleagues in technical disciplines. We can help NARS economists understand the sources of these problems and share the methods we have found to manage them within our own research environments. This requires greatly improved communications among NARS economists and between NARS and LARC economists. IARCs can take the lead in establishing disciplinary groups similar to the scientific networks that are now ubiquitous in much of Africa. Annual meetings at rotating NARS and IARC sites with tours to visit ongoing farm-level research would be one format for such meetings. Their main roles would be to brng coherence to the use of economists and to satisfy taining needs. Networking provides a range of initiatives to improve NARS' access to key liteature. Current contents of all economics journals can be circultd periodically, with IARCs being responsible for sending copies of articles requested by NARS economists, or a selective dissemination of information scheme can be set up. Unpublished internal institutional reports can be identified, listed, and circulated on request with LARC support. [ARCs also can take the lead to help NARS economists publish research results through network ujounals' or newslete. One important key to strengthening the social science research capability in NARS is a uing of the linkages between university-based scientists and scientists based in the national commodity research centers. To think that IARCs can single-handedly strengthen NARS social science research capability without increased and well-directed participation of universty-based scientists is probably unrealistic. IARCs have a lead role to play in taining in on-farm research methods. With the advent of FSR, a series of easy-to-apply toolkits was made available for economic analysis of agronomic trials. Itinerant training courses, held in rotating NARS locations where facilities and basic equipment needs can be met, would be supeior to training solely at IARC locations. With a new emphasis on agricultural sustainability, the current core of eownomists must be strengthened with the new research tools necessary for state-of-the-art reseach. 26 Fortuately, LRCs are at thc forefront of the methodological developments in this area, but social scientists in NARS need to be more fily involved in the debate and to benefit more immediately from the disciplinary advances on sustainability. Support of thesis fieldwork at the graduate level offers excellent opportunities for IARCs to provide students with insight into the interface of economics and technical research. In return, research scholars not only funiish IARCs with high-quality, well- focused, and rapid research results on priority topics at reatively low cost, they also constitute a cohort that can form the base for subsequent research collaboration. The location specificity of most economics research makes it necessary for IARC economists to develop collaborative research with NARS colleagues. Moreover, conducted in the context of support networks, collaborative research can assist NARS economists in formulating well-defined and adequately funded research prqjects. Formal collaboration also can provide the critical review and peer recognition that NARS economists require. When initiated by IARCs, such collaboration can inadvettently divert national economists from priority national problems and further divorce economists from colleagues within the national program. IARCs must take care to ensure that collaborative initiatives are driven by NARS, in the sense that they correspond to national priorities, and that workplans are drawn up with NARS economists and technical scientists playing a ful and early role in the definition of problems. The Role for Social Sciences in MI's Mandate to Strengthen Institutions and Link Research to Policy Douglas Mefey and Thssa Bandamgoda, IIMI MIR's mission is 'to strengthen national efforts to improve and sustain the performance of irrigation systems through the development and dissemination of management innovations." Its main goals are to produce research results of regional or global significance, to strengthen national resarch capacities in irrigation management, and to strengthen national management organzations' capacity to improve performance through the adoption of research-based improvements. The social sciences ar- central to achieving IIM's mandate: more than 40 percent of its intenationally recruited staff members are social scientists. IMO's first strategy, published in 1988, emphasized the institution-building aspects of its work. This was consistent with the imperatives of IIMI's inc_ption stage. Because roughly 80 percent of its funds were obtained through restricted rrojects, mostly bilateral sources in specific countries, much work was driven by donor and host country priorities, mediated and shaped by lMI's own interests and capabilities. Dependence on this mode of 27 funding was n imporant facor in IMWI's inability to devdop a longer-term resarch agenda of its own that utscended individual countries' intests. Most of JIMI's work has addressed issues at the nmain system level" of irrigation syms, not the on-farm level. Much of that work is focued on questions of canal watr delivery pefo n; intio and manarial factors affecting the planning and lemntaton of thes deliveries; and farmer oranitions and thidr rlationship with extenal agencies, on both farmer-managed and government-managed sysms. Most NARS do not addres the lknds of manageme_ t and insituonal iss central to HIM's progrm. In geneal, MM's research addrs farm-level water maneent. In most counties few, if any, natonal resarch os address multidisciplinary irrigation management issues above the farm levd. MIO identified the irrigation m g t agecs, m whose systems it does research, as its primary cliets. R measued its success in terns of its bnpac on these client agencies, many of whom have no history of research. JIMI has no labomtory or e.xmmental falcilities and theefore cannot be effective without full collaboration from partner agencies and farmue. Depite this, an imporant lesson has been that the research process is itself a powerful institufion-building methodology. Thugh eir active paricipaon in research, management staff earn new modes and appraches. MC is experimenting with using aplied research itsef as an institution-building methodology that agency staff can continue to use afftr MM's witdrawal itally, MIO had explicitly excluded policy msearch. Early exp0ereces, however, demonstrated the interdePendenc of managemt and policy, and it is now approached by IMI in three ways: direcly, though local operatonal issues, and by assisting in developing more effective processes for linking research and policy. In Si TLanka this process of linking research and policies has been carried further trough a unique irrigation mepolicy support activity. Through this highly paricipatory project, IDM assistd SriFLanlAns to develop, elaborate, and make operatioal its p mt policy for irrigation t A decade of epeimentaton and rsach by varou oraizat in Sri Lanka provided a basis for these procedures. For a two-year period, MG worked with its Sri Iankan colleagues to develop ten policy papers, combining the analysis of results of past resarch, discussions of the lessons learned, and optio for the futur IThis apprach led to consensus and ofirmton of the recommendations by sior m . is participaty method of policy deveomet has potential for odher countries. EM! has concluded that sustinable institutions are as important to develpment as sustning natural resouc A prrequisite for this is al and institutional strengthenin and refom. This requires putting appropdate policies in place and gaining a 28 comi t to their lementafo. Policies too often reflect a balanc among various vested interests, with te estlishmnt sding tie stus quo. Food Pblicy Reseach in the CGLAR-IFPRI and Other Centers Steve Vosti, lFPRI Vosti offered a definition of food policy: 'Food policy mpsses the collective efforts of gov ts to influence the dsonmang virment of food producers, consumers, and food marketng agents in order to further social objectives." He used a policy research continuum with increasing levels of disag on from left to right (see Figure 2) and sought to locate the role of the CGIAR in it with the following questions: * W.hich research issues are most pressing? * What sorts of output are needed to resolve thes issues? * With the worishop objecies in mind, what sort of intercenter collaboration ae most lily to be successful in idtfying and tckling the most pressng iss, panicularly those least likely to be tackded by others in the reearc cnmmunity? Vosti sees policy research as more inportant and complex than ever in tackling food security, agricultural gowth, natural resource management, and poverty aeiaim. Proper pohcies will be needed to ensue that the greatest benefit posible is extacted from new technoloes, to assist in dibutn thes benefits equitably, and to aid in in alizing environmental and ote 'spillover' effe assoctd with the adoption and use of new tecnologies. Te CGIAR has been known pimarily for its products, such as new seed, but as it moves toward further integation of soal sciee and policy research, output will shift to a process orientation. The process of policy research shoud generate social sience 'germplasm, replicable methods widey Czvant to solving problems elsewhere Policy reserch wil also geneat "productso; spefic mmendations that can be of use to policymakers to achieve social objectives; with some adaptation these will also be relevant beyond the contexts for which they were identified. Diffusion of research results-getting tbem into policy dedsions-is a two- or thr-year process, eve in the countries in which the results were obtained. Tracking thwr impact is even more tenuous but deserves attention. Impact asessment for policy reseach (for both -products and -procses-) also is more difficult than for tradito lines of CGIAR reseach. The complemties of the real wodd, and the absence of dtiled time series data that might alow for the identification and 29 Figure 2a. Continuum of policy reearch: policy generation and Impact, policy filters, and policy feedback Polcy Generation and Implenlentadlon Policy ImpactI Intra- household Inter- Filter household lnput'Outut Riter Mafket Infrastnrcture filter Fdt o - Reagiojnal Nado|| Filter Filter ~~Market NJ=ofi Filter Prm e Individual Family Farm Local Regbinal National hItemulbonal Figure 2b. Continuum of collaborative activities- Zeo Logistcal Support Conceptualizaeon Concept blizaton @CIZu fl Sorne Fiel AcdvidesAnalyses Analyses Pubgcadon and Oubwach 30 quntification of policy impact, make it unlikely that the standard tools and metrics for impact assessment will ever be very useful. Vosti reported that a more detailed analysis may be required, along the lines of an anthropological study, charting the broad results of a given project as it unfolds: institutional change, cpapcity building, staffing charges, research results. Impact assessment would then subjectively attribute policy change to these (partially) project-induced institutional and other modifications. Aside from providing important donor feedback, this kind of assessment would provide additional insight into how, when, and where to modify the research process to improve chances of influencing policy for the project under study and more broadly. Collaboration: Who Should Do What? A move from the international side of the policy research continuum to the local side increases potential overlaps between IFPRI and other centers. Several centers are working on constraints to the adoption and efficient use of the technologies they are developing; this sometimes includes IFPRI in a collabortive mode. Among the issues arising along the continuum, which are most fertile for intercenter collaboration, and how do we proceed? Several important issues cut across the center mandates, especially those with an woregional role: exemalities, prperty rights, migration, and land degradation. Collaboration on these themes would generate findings to support increasingly robust generalizations. Conduits between centers will be vital to collaboamtion. Participaton by NARS will be criical, particularly to the use of results. Finally, the wide geographic distribution of social, agricultural, and environmental problems dictates a few critical locations where efforts should be concentrated. Collaborative mechanisms are a missing link. Experiences still lack evaluation. Often collaboration has been inhibited because of an IARC monopoly at the conceptual and planning stages with subsequent doubts on program ownership, and subsequent failure to commit to specific resarch agendas and outputs. Vosti set forth six principles for collaboration: * Centers should have a clear view of what poLicy research will best complement their research agendas in the next years, make explicit what aspects they would like to pursue both collaboratively and independently, and work together to achieve those goals. * When aU partners agree that collaborative research is the best mode, it should be based on clear projects withl terms of refernce and deliverable products. * National research institutes should be involved at every step of the intercenter collaboative process, including concepton and planning. 31 * Projects must have thc wholehearted sWort of all collaboradng institutions, from management of the institutional ovehead to publication and dissinaton of the research results. * The matching of individuals in research projects is just as important as the choice of the right projects. Individuals should complement each other in skldls and personality. * The method of collaboration should be endogenous, dictated by the type of problem being addressed. Vosti concluded by notng that IFPRI perceived collaboration as a potentially powerfil vehicle for bringing much wider resources to bear on impo.rant issues. Issues and Discussions The four worling groups covered a range of topics, which are reported under three headings: Roles for social scientists in IARCs and NARIs Technology, institutions, and policy * Methods of capacity-building Roles for Social Scientist in IARCs and NARIs One group identified policy, priority-setting, on-farm research, monitoring, evaluation, impact assessment, and human resource planning as the main topics for social science research. Across the working groups, four broad roles for social scientists were identfied: * Conducting strategic social science resarch at both microeconomic and macroeconomic levels * Contnbuting to the agricultual rsearch process, producing technologies usefu to frmers * Building capacity in social science in a range of national organizations, with NARIs a dominant target * Assisting management in planning and setting priorities in research institutions 32 For this session an appropniate focus was the final two of these four roles; however, the policy and technology dimension brought three of the working groups back to the balance of roles between global and eowregional centers. It was clear from the discussions that there could be no single conclusion on the appropriate balance across the four roles; it would vary with the type of center. The "social science centers," as IFPRI, IIMI, and ISNAR were labeled, were seen as having a strong lead in addressing strategic social science research questions, often through "research on research' to develop useful methods. It was agreed that both capacity-building and institutional planning were essential activities for IARCs and NARIs and that economists have a vital role to play in both. In some centers, a full-time economist has been devoted to planning, and ICRISAT and CIAT recently moved to institutionalize a position for an economist in priority setting and impact assessment. Some participants thought that too much time has been invested on these activities in some centers, that plans and priorities were sometimes window dressing for external use and rarely altered resource allocation in the center. Complex models combined with cruoe estimates of empirical parameters may lead to priorities that are neither robust nor transparent. Unless equal attention is given to monitoring and evaluation, ex-ante mleposts have little meaning. Many NARS have a generic problem of incentives and staff morale. Staff are often subject to pay delays, frequent reallocation, and low operating budgets, preventing strong program development. Tuonver is high among young scientists. Beyond these problems, agicultural research managers remain uncertain how to use social scientists. Participants strongly believed that the ad hoc actions of IARCs heightened this confusion; each IARC sought a social science commitmet from NARI managers to the particular role it was promoting. The division of "upstream and downstream," often used to descnrbe the role of LARCs and NARIs, was seen as simplistic. NARI capacity is often greater upstream than downstream; adaptive research is often the missing link in the national research process. A gap remains between small farmers and the research service. Because of its fledgling nature, capacity-building for social science in NARIs has been seen as a vital role for IARCs, yet there has been no cohesion or strategy in the way the CGIAR has encouraged the application of social science in agricultural reserch establishments. Participants questioned how the Technical Advisory Committee recommendation for a devolution of general training activities could be reconciled with the weak state of the social sciences in national agricultural resh establishments and the strong comparative advantage of LARCs. They thought there was a strong case for maling social science training an exception and saw few altematives to LARCs for providing the specalized expertise needed to integrate social science into the agricultmal research process and the institutions implementing it 33 Professionally, both planning and capacity-building have high opportniy costs in terms of research activities that win disciplinary recognition. IARC managers need to ensure that annual worcplans are balanced among research, institutional planning, and capacity- building and that reward systems recognize the planning and training roles. Technology, Institutions, and Policy The tradition of agricultural reserch is to mobilize policy research when a policy issue inhibits dissemination of a new technology. In the ecoregional approach, however, policies often may be the initial point of intervention. An example is West African countries pursuing rice self-sufficiency that want to open inland swamp areas for development This would be an appropriate issue for ecoregional researchers to address, in collaboration with others. The principal research product may be a modified policy rather than a technology. Only vastly increased collaboration can provide the pools of expertise required to address problems from the diverse perspectives that may be relevant to an appropiate answer, particularly for natural resource management and environmental impact. A broad consortium of institutions-CGLAR centers, NARS, universities, North and South research istitutions, NGOs-to address land use problems within an ecoregional context may be the most appropriate model. Yet uncertainty remained about just how wide a role the CGIAR itself should take. For example, in the West African rice case, the critical issue may be comparing the value of conserving the swamps with their value for rice growing. This was judged to be beyond the CGIAR mandate. Particular implications were highlighted for IFPRI and its work with other centers. It was hoped that IPPRI could assist the other centers in research on social and institutional mechanisms, especially at the local level, if not itself, then by helping to identify suitable collaborative partners that have local experience. More social science disciplines need representation in the CGLAR, including geographers, rural sociologists, anthropologists, and resource economists. Methods of Capacity-building Capacity-building methods were touched on only in the discussion. Two dimensions were identified: research collaboration and networking. Collaboration was perceived as a continuum of increasing intensity, ranging from logisfical support at the narrow end to long- term parteship in aU phases of the work at the other end. Networking intensified along the continuum from information exchange to fellowships to training workshops to graduate work to technical meetings and, again, in its most intensive form, to fully collaborative research programs. 34 The outanding condlumon from the discuions was an emphasis on tie theme of coherence-the need for a coherent CGIAR-wide position on the roles of social science in agricultural research, and a coherent approach to building capacity in NARS, not least the capacity to use socal science effectively. It was acknowledged that pressure was incresmng on NARS as well as IARCs to set prorties. Participants thought that consideable effort is being expended on pnrioty-sefting by both ISNAR and IFPRI without adequat examination of what is needed in NARS of diffrent sizes. A need was seen for a consensus about how NARS should approch priority- setting. This would require a replicable means for articuating farners' needs and reconcling those needs with priorities of national policy. A need was also seen for a coherent KARS institution- and capacity-building stategy acmss the CGIAR with the efforts of individual centers as key components. NARS themslves should be a party to the planning of such a stategy. Indnivdual LARCs have the challenging task of reconciling the needs of many counties. If countries collectively reconcled and artculated their priorities, perhaps on a subregional basis, IARCs could respond coherently. The impra of these articulated piorities, which would reflect frmers' needs, implied the need for country coverage by adapive srch capacty, including social scientists. An important subset of this need for greater coherence in capacity-buiing was coherence in priorities for the use of social science in NARS of different types, sizes, opertional strucs, and cicumstnces. Social science in many KARS is stll in its infany, but because adaptive research is lkcaton-specific, there is a false impression tat all is well. In fact, social science in NARS is so weak and managers are so confused about its roles that ther is a specal need to continue intensive traing for social scienss in naional agricultural research. 35 The Users' Perspectives: Factoring Farmers' Needs into Research Agendas Intemational Service for National Agricultual Research (ISNAR) Marie-HMi&Ae Collion Thomas EponoU Deborah Meril-Sands Centro Internacional de Mejoramiento de Maiz y Trigo (CIMMYT) Robert ipp International Irrigation Management Institute (III) Doug Vemulion Centro Internacional de Agricultura Tropical (CIAT) Jacquetine Ashby Louise Sperfing Centro Iteracional de la Papa (CIP) Greta Watson Instituto Boliviano de Tecnologfa Aoecuaria (IBTA) Jorge Quroga Introduction Identifying the real problems of developing country agricultr, and designing technologies to solve them that fit the resource-poor crum s of small farmers, have been recognized as chaLenges for twenty years. The need to articulat the users' perspectv on key constraints to farm imprvemet, and on appropriate means of relieving them, has spawned farming systms research and several tereinological equivalents. A more recent iteration is farmer participatory research (FPR), which has found favor in agriculture and beyond. With FPR, a sustained inteaction with farmers, who speak for themlvs and their community, has replaced the survey instruments of earlier years, when the social scientist stood as the farmers' pmxy in the design and evaluation of technologies. Farmers clearly are not the only stakeholders in the success of agricultual research; consumers are interested in the price of farm products, and policymakrs are interested in economic grwth. Although wider intests will weight resech priorities, the benefits for all interested parties remain dependent on innovation by farmers. Thus, undestanding the criteria they will apply in their choices is vital to the design and evaluation of acceptble and appropriate nologies. Processes for articulating users' persectives are now well known. IARCs have used methods for eliciting users' perspectives such as participatory rural ais; action research working groups involving users, researchers, officials, and NGOs; user technolog assessment techniques; farmer-tofamer extension; farmer paraprofessional researchers; and community-based infma networis. The IARCs also have imvested in biling capacity 36 in NARS. NARS have been Slow to adopt the process, often because of declining research budgets during the past two decades. The socal sciences are new to agricultural resarch, and for many NARS, budget shortages have meant that new disciplines can be introduced only at the expense of taditional ones. Additionaly, in some cases thc older technical scientists managing some NARS believe that the farmers or the extension staff remain the bottleneck. Both factors have contributed to inertia and low morale in resarch establishments. A rebalancing of both disciplines and the research process to marry the human and technical dimensions in the agricultual research process has been frustrated in most national institutions. In their paper Marie-Helehe Collion, Thomas Eponou, and Deborah Merrill-Sands from ISNAR provided a history of attempts to make this marriage and discussed continuing problems of institutional innovation to capture users' perspectives. This first paper offered a framework within which Robert Tripp explored definitions and tenninology of the user perpective, with reference to IARC activities. Doug VermiUion addressed the broader idea of stakeholders in research and examined six roles for them in participatory development. Jacqueline Ashby and Louise Sperling advanced several steps on the key issue of mobilizing farmers' perectives to influence research agendas. Finaly, Greta Watson and Jorge Quiroga discussed germplasm evaluation with special reference to farmers' perspectives and in siu conservation of germplasm. Research Program Planning and Priority-Setting: Strengthening Farmers' 1nput Maie-Helene CoUlion, Thomas Eponou, and Deborah Meri-Sands, ISNAR Tle benefits of user participation in research and development (R & D) have long been recognized by pnvate firms. In contrast, most agncultural rach systems in developing countries do not systeatically involve farmers. Typically, such fanners lack the political and economic power to ensure that their demands are met. The work of the past two decades to bring farning systems research, on-farm research (OFR), and farmer participaty reseh to the forefrot has helped increase farmers' direct contribution to the research agenda pniorities. The authors' conclude, however, that the impact of these new mechanisms in bringing farmers into the decisionmaking process has been limited. They assert that 'although these research methods have led in many cases to more client-responsve research scientists, they have not, in general, resulted in more client-resonsive research 37 The authors propose reasons for the lack of impact in reshaping research agdas: * Social scientists have not given sufficient emphasis to the docisionmaing structue and processes into which information must be channeled if it is to influence research agendas. * Social scientists have not adequately taed the issue of empowerment- farmers' ability to exert pressure on research organizations and hold them accountable. Like technology, the impact of research approaches is strongly shaped by the policy and institutional environment. FSR and FPR are merely sources of information on users' perspectives; nothing guarantees their use in decisionmaking. The authors identified two areas as pnonties for LARC socal scientists: including farm-level information decisionmaking and strengthening farmers' influence on research agendas. Iwluding Farm-level Iforation in Decisionmaking As research systems begin to adopt formal approaches to planning and pnority- setting, opportunities will arinse to integrate input from the users. The authors suggest an eight-step planning process and see the users' perspectives as contributing to five of these Diagnosis of the sector trgeted by the research program and reviw of development objectives for that sector (which may be a commodity or a production system) *0 Analysis of constraints that hinder the achievement of these objectives and the causes of these constraints Review of research results relevant to the causes of identified constraints *- Determination of research objectives and design of a stategy to achieve these objectives *- Evaluation of alternative solutions and identification of research projects *. Setting of priorities among research projects. * Human resource gap analysis by discipline across research projects 38 * Policy reommendations for identiication and implementation of measures to facilitate technology adoption Participatory planning sequences provide the oppormnity for systematically integrating input from users into the formulation of research agendas. The mechanisms for achieving this are an important focus for CGIAR social scientists. An important issue in devising such mechanisms is aggregation of local-level information to make it meaningful for decisionmakers, who will be represented in this process. A basic data set from all regions designated as a priority for development should provide a balanced expression of demand. Strengthening Farmers' Influence on Research Agendas Mechanisms to bnng the fanners' perspectives into policy formulation will not be enough. Farners must be able to exert pressure on research institutions and need empowernent to do this. Farmers' organizations and NGOs have great potential leverage. Social scientists in IARCs need to analyze experiences and define the conditions under which organimzations of resource-poor farmers can accrue power; rather than NARS, perhaps they should be primary clients for training in adaptive research methods. Until ways are found for resource-poor farmers to represent their interests more directly in the decisionmahdng process, on-farm researchers should repres t the farmers' interests in the institutional process of setting research priortties. These researchers will be most effective when the farmers also view them as their sepresentatives in dtrmining the rsearch agenda. Issues Related to the Users' Perspectives Robert Tnipp, CIMMYT This paper reviewed how the users' perspectives affect the actions of IARC social scientists and raised issues for IARCs in incorporating those perspectives. Facets of the Users' Perspectives It is important to precede research on users' perspectives by clearly defining which set of users is being targeted. Tripp provided several guidelines: lThe perspectives derive from socioeconomic information articulated directly by farmers, individually or as groups, or inferred from social science survey and analysis. * The perspectives are subject to change: seasonal variation and the dynamics of extemal cumstances, such as policy and market forces, may alter users' views. These sources of vanation should be reognized as relevant to an understanding of the users' perectives. 39 * Technology does not have to satisfy all users' criteria. To be acceptable to farmers, it does have to offer major gains on some criteria if it fails on others. * Research agendas are rarely developed anew. Farmers can influenoe priority-setting for research or program design or be involved in adjusting ongoing programs. IARC Contributions and Lssues Various L4RC activities contribute to an understanding of the perspectives of farmers. Sponsorship, networking, and the coordination of local-level studies on the users' perspeives, through NARS or NGOs, are the best ways to improve the design of a research effort. Often such studies will be conducted as part of a capacity-building program in which training in methods is important. IARCs have made strong contributions to development of methods in FSR, OFR, and FPR. IARCs and NARS often share a common interest in identifying target populations for research products. Zoning and the characterization of farming systems and prototype technologies allow better matches between what is needed and what is awvilable. The aggregation of the perspectives of users in a variety of local situations is an issue even at the national level. Breeding progrwams, for example, cannot respond to the needs of smail groups of farmers. This aggregation problem is even more evident at the international level. The question remains, can a mechanism be developed to synthesize local-level information to guide international breeding? If an effective mechanism is ever developed at the national level, canying those syntheses to international planning should be relatively easy. A second issue is the lack of coordination of efforts by individual IARCs. Each IARC does some diagnostic work and maintains its own data bases. When an IARC, or a NARS partner, mounts a field survey, small increases in cost would allow the collection of additional data on farming in that area to meet the needs of other centers. As yet there is no attempt to coordinate needs and methods to build a collective data base of on-farm information in developing countries. Users' Perspectives for Participatory Research and Development Doug Vemion, Mg Tle author sees the researcher's model as his best friend and wrst enemy. He makes the argument that drawing the users into research on natural resource ma ent is likely to lead to more flexibility in models and ftameworks and hence to greater creativity. He 40 emphasizes the changes noeded in scientists' attitudes before tiey will listen to farmers and users, rather than impose outside ideas. Vennillion suggested substituting the term 'stakeholders for 'users' to emphasize the need to include all who have an interst in the resources or the technology. He sees the set of stakeholders as diverse, ranging from the landless to landholders, male and female, to processors, traders, and consumers as well as researchers and policymakers. He then discussed six roles for stakeholders in research and development in natural resources: providers of information; partners in identifying R & D priorities; partners in research, management strategies, and technologies; partners in experimental management of resources and technologies; partners in assessing the performance of technologies and management strategies; and partners in training and institutional development. Providers of Information This is said to be the most common role for stakeholders in participatory research and development. Although there is enormous potential to develop methods to ob: iin information from local people, a two-way flow of information is needed; farmers are often unaware of laws, programs, and opportunities for obtaining support services. Key issues to be considered are the extent to which advanced-degree R & D professionals are willing to receive and respect information of this sort and the need to overcome the tendency merly to extract information and not use it. Partners in Idenfifying R & D Priorities The epresentation of farmers in decisionmaking forums is an increasingly common means of identifying research and development priorities. Two related issues are the willingness of government agencies to relate to farmers as partners in identifying priorities and the institutional arangemets needed to ensure adequate representation of key .staeolders. Partners in Research, Management Strategies, and Technologies In irrigation development, the participatory strategy is intended to instill a local sense of ownership and incentive to maintain the system with local resources. Participatory design of new plans has led to increased commitment and better agency and farmer cross- accountability in implementation and enforcement. Key questions are, is there a trade-off between the degree of stakeholders' participation in research and the "professional quality" of the research, and if so, to what extent are CGIAR and other researchers willing to sacrifice standards of conventional research to achieve the benefits of participatory R & D? 41 Partners In Eprhetal MLnagement of Resources and Technologies Farmers are lkown to experiment in a variety of ways. Such experimentation should be encouraged by both local people and agencies. Two key questions for this role are, how can agencies become flexible enough in timing and procedures and transparent enough with finances to encourage a joint experimental approach to natural resource management, and how can agencies reorient themselves toward providing technical services and encouraging participatory experimental strtegies? Paters in Assesing the Perfonnance of Technologies and Management Stntegies Self-assessment of management performance by organizations of irrigators has shown promise. Farmers identified their own performance measures and evaluated their own organizations over time. This exercise provided insightfil information, developed timely action-response arrangements, and strengthened local institutions. Relevant questions are, in what situations are either self-directed or external approaches to performance assessment needed, and how can self-assessment approaches be adapted to enable generalization? Partners in Training and Institutional Development Managers of natural resources, including farmers, can provide peer training and help in institutional development. Farmers from successfully managed irrigation systems have proved capable of extending effective management and organizational training to farmers in other systems. Engineers tended to dwell on the visible symptoms of problems, whereas farmers proved adept at analyzing the root causes of inefficiencies. Key issues are convincing government officials that farmers or community resource managers may be capable of augmenting extension and btaining programs and determining the extent to which CGIAR centers should engage in such activities. Institutionalizing Participatory, Client-driven Research and Technology Development in Agriculture Jacqueline Ashby and Louise Sperling, CIAT This paper outlined the main issues needing attention if farmers are to participate on a regular basis in the design and development of technical innovations in agriculture and highlighted areas where social science rsearch is needed in the process of institutionalization. Ashby and Sperling identified features of the participatory R & D approach that affect how it is institutionalized: 42 * Driven by clients. The farmers' criteria, indigenous knowledge, and subjective preferces have weight in piority-setting and reserh planning; farmers do not merely feed back information on finished and introduced technologies. This, in turn, involves a degree of deentral on that will require dose and constant inteaction with applied research and a flexibility to modify designs in response to farme' expaiences. * Options as the critical focus for applied research. Instead of producing finished technologies, applied reserch should emphasize identifying a menu of opdons, which can become prototypes for the solution of significant problems. * Farmer input to applied research. If farmers are brought into contact with optional prototypes at research stations at an early stage in research programs, a range of technologies potentially useful for local situations can be identified. Fhrmers identfy technologies most compatible with their circumstances, take resonsibility to test and adapt them, and, with the help of local farmer organiztions, dissemiate the usefil ones through their commuuities. - Accountability. Both researchers and farmers should share accountability and quality control for technologies produced through such a format. The authors suggest that the biggest obstacle to an effective public research process, and to institutionaliing client-driven partiipatory R & D, may be lack of accountability of most public-sector rsearch systms and their staff; they cannot be penalized for producing technologies that farmers cannot use. They identified two issues in setting a client-driven research agenda: reconciling the diversity of fanning situations and farmers' needs with limited applied public research capacity and reconciling farmers' short-tenn needs with other groups' priorites for long-tenn research, for example, on environmental conseation. The authors proposed the following mechanism for determining the weight given to each sukeholder's pionties in setting the research agenda: put a proportion of the resources available for research funding under the control of different client groups in relation to their importance to society. This would create the means for each group to contract research and to exert demand on the applied research system. Balncing Rearch Priofie The authors believe that socl scientists must do more than reconcile competing pnonties; they must answer two critica questions: which user groups are most important in policy terms, and how can one create the capacity and the oraniztion required for client groups to express demand in aggregate rather than as individuals? 43 On the second question, a numbcr of action programs are in place, mostly sponsored by NGOs, but very little is klown about their effectiveness in articulating farmers' demands for technology. There is a role here for social science research. Client-driven R & D may cause an imbalance among three categories of research: location-specific adaptive research (what farmers are most likely to demand), applied research geared toward understanding general principles (what scientists might like to pursue), and research oriented toward sustainability concerns (what special lobbyists might advocate). The appropriate balance is deterniined by a policy decision about the proportion of resources directed to long-term research rather than immediate pay-offs to farmers. In terms of accountability, those who control the financial resources have the power to ensure that the promised products are delivered. Client-driven agendas are realized only when clients have substantial leverage. Social science research has a role to play in clarifying these trade-offs. It can evaluate the effectiveness of various organizational forms for promoting both a client-driven and an applied research agenda. It also can assist in analyzing the effectiveness of resources allocated among different beneficiary groups, including future generations. Reorienting Applied and Adaptive Research Decentralizing adaptive technology development implies a reorentation in applied research in which researchers would develop a range of prototypes suitable for modificaton. On-station researchers need to understand the diversity of client needs and constraints at the beginning of the technology planning and development process; involving farmers in the early screening of prototypes can lead to earlier identification of shortcomings. Social scientists need to examine the drawbacks of exposing farmers to prototypes and assess how applied research planning might compensate. Institutionalizing farmers' participation involves developing a community-based adaptive research capacity, which can be achieved by working with groups of farmers (raer than individuals) and producer organizations. The authors raised the question of whether researchers have the skills to work directly with specific client groups. In delegating adaptive testng to farmers, are research systems committing themselves to worlkng principally through internediary organizations, such as NGOs, and if so, with what consequences? Another question addressed the quality of on-fann testing. One argument against devolution of adaptive research to farmers has been that, as their participation in managng on-farm experiments increases, the reliability of the data produced fls. Are conventional experimental methods, and therefore conventional standards of rigor, a =ppriate for adaptive work? 44 The cental issue rmains the costs and benefits of delegating adaptive research to fahmers: to what extent can the full costs of adaptive research be shifed outside the formal research system, and to what extent will retuns improve from expediting technical innovation on farms? The Social Sciences and Participatory Research in Germplasm Mnagement Greta Watson, CIP and Jorge Q&iroga, MBTA Watson and Quiroga noted that IARCs and NARS are best used to complement indigenous technical capacity in research for highly diverse areas. Through partcipatory research, social scientists and user clients have an important role as cultural and agricultimal brokers between local commuiities and the scientific establishment HIgliighdi'W in situ germplasm management, the authors compared the maintenance of indigenous landraces or cultivars by local populations with current stataegies in the more formal on sector. Their paper suggested that formal 'scientific intervention' may be substitutable for, or complemented by, the rich diversity used by farmers in local evaluation and experimentation. In pursuing this complementarity, Watson and Quiroga emnphaszed the importance of social science in identifying appropriate communities for in situ conservation, in understanding farmers' perspectives on conservation and use, and in initiating work wifth farmers at chosen sites. The authors then argued that farmers' participation can speed germplasm diffusion. Research into the decisionmakdng criteria of small farmers in varietal selection provides breeders with more accurate data on farmers' needs. Farmers' participaton in seed multiplication and exchange can diffuse information more rapidly and foster the spread of new varieties through the use of existing informal networks. For example, in the selection of tuber families resistant to pests, use of criteria from both scientists and farmers could reduce on-station maintenance costs for development. In terms of diffusion, channels of seedflow among farmers may depend on the ownership of new varieties and of landholdings by the community, by family, or by individual household members. Although informal systems to disseminate seed seem to function effectively, the distnbution of new seed varieties is likely to encounter botdenecks. Social scientfsts, through participatory research, should seek to understand the inherent inter- and intracultunal diversity of diffusion systems for new varieties through continued monitonng, evaluation, and impact assessment. In summary, participatory research provides a middle ground between formal research and extension. Social scientists should focus on the integration of formal and farmer research networks. 45 ILsues and Discussions Three woring groups wer formed to cover the follownng topics: * The client pepective in natural reurce research * The user persective in reseh priority-setting * User input in resarch and extension lhe Client Perspective in Natural Resource Research Four areas were identified as inportant to a better understanding of the client perspective in naural resource research: * The tme pevecti for issues of short-term secunity and long-term investments = The issue of tenure, which represents security, and the type of property held * The nonuser stkholders consisting of future generations and external rcpents e The heterogeneity of users and sholds Emphasis was placed on the need to consider the perspectives of several categories of stakeholders. Those listed included users, nonusing farmers, landless labor, dowosnmers, sumers, procsors, future stkholders, and society at large. Each ctegory n.flects unique dhractristics that require seaate approaches to technology development and different research methods. The Usw Perspecive in Resarch Priority-settg The group first identified the different stakeholders in the resrch priority-setting process: donors, managers, scientiss, national programs, and final technology users. Although the consderations of all stkeholders are now included in priority-setting, the perspective of resource-poor farmers is the least well intgated. Two models were presented (see Table 1), representing the extremes of priority- sewttng, ranging from technocratic centralization to decentralized negotiation. 46 Table 1. Characteristics of Technocratic Centralized and Decentraled Negotiation Models Ch;-arsactistic -:: Techcatic Centralized Dcentrized Negotiation .-u- .:- . s , _ _ Proaess .__ Decision tking Central Stakeholders Decision process Quantitative analysis Nege- .on Information Data base Stakeholder voiced Dominant discipline Economics Social science and politics The group recognized that neither of these two processes is ideal and that an optimal decisionmaking process would be a blend of both. The group considered the technocratic centralized process to be typical of current decisionmaldng in national and international agricultural research centers. To enlarge the role of the user perspective in priority-setting, two principal strategies were proposed: change the information on which decisionmaking is based and change the process of decisionmaking. At higher levels of aggregation (global, national), it is more feasible to improve information than to change the decisionmaking process itself. At lower levels (regional, local), revision of the decisionmaking process should be given serious consideration. A compiication of this second approach is that it requires current decisionmakers to accept reduced influence. Suggestions to deal with this problem were: - Improve the decisionmaking process so that decisionmakers are convinced that the gain in quality of decisions compensates for their loss of influence. * Improve the availability of data to make decisionmakers aware of the relevance of sharing decisions. * Incorporate market-like mechanisms for decisionmaking. A number of suggestions were made about how social scientists can contribute to an improved role for the user perspective in priority-setting at the program level: * Develop measures for stakeholders' criteria. * Experiment with and evaluate alternative decisionmaling pmcsses. 47 * Facilitate farmer organizations to bring farmen into the decisionmakdng process- * Drevelop management procedures that icrease the waght given to the user perspwctive in decisionmakdng. * Design a minimum required data set for priority-setting. * Document experiences and evaluate them for cost-effectiveness. * Collect ex-post data on the impact of research to improve undstanding of how users are affected by certain decisions. * Monitor and evaluate progress in implementing reseah priorities. Clearly there is much potential for interational agricultural research to adopt recent developments in naent information systems that involve assessing client perspectives hrough market research, client satisfaction, and needs assessment This is especilly relevant to the CGIAR as it considers the range of stakeholders with an interest in the processes and outcomes of research on agriculture and natual resources. User Input in Research and Extension The discussion on research agendas reinforced the importance of recognizing the role of institutions and policy in developing 'social terhnologies' as wel as the more traditional biological and physical technologies. The group considered two methods for increasing user contributions: farmer organization and privaition of research and etension. The group believed there is a need for empirical identificaton of the conditions that fcilitate or impede cooperation among farmers. Work in this area has already started. CIAT has conducted an expeiment comparing three types of intermediaies for organizing farmers: NGOs, extension agents, and self-selected farmers with rsach experience. The results show that the type of intermediary influences a trade-off between the reliability of the data coLected and the widespead tansmission of information. This finding implies that the research objective should guide the choice of internediaries. On the issue of equity in farmer organizons, the group emphasized the importance of ensuring that all categories of village residents were represented in such groups. Equity implications should be considered carefilly; for example, landless laborers could be affected by water management schemes, even though they do not own land. The group thought that the development of data bases categorizing village residents would facilitate the task of ensuring that membership in fanner organizations was representative. 48 Externality problems also can complicate farmer groups because their intewsts as prvate agents may diverge from the broader interests of the larger community. For example, there would be little demand among farmer groups for research and extension on resource degradation problems if the costs of such degradation were not bome by the farmers themseives. Field experience shows that one way to overcome this problem is to include all partes affected in the discussion groups, explore how actions by one party can affect others, and leave the group members to decide how they pay for the social costs incurred. To increase user input, an alternative to farmer organizations is to change the incentive structures within public institutions, for example, to pay researchers according to the extent to which their technologies are adopted. Through privatization, research and extension could be divided between the private sector and the public sector, witi possible contracting to pnvate companies, farner organiztions, and NGOs. Donors such as the World Bank have already made plans to privatize aspects of agncultural research and extension in some countnes. The group emphasized the importance of CGIAR centers contributing their experiences to the issue of privatization and agreed that the effectiveness of farmer-to-farmer extension needs furither investigation. Conclusions Demonstrating the diversity of experiences across IARCs, the papers and discussion shed new light on the search for a client-driven research process for resource-poor farmers. A strong body of literature now documents the effectiveness of research programs in which farmers participate in design, implementation, and evaluation. The dominant questions have become how to convince traditional research establishments of the importance of farmers' participation, how best to organize the participation of small farmers in the research process, and how to ensure that the needs of small farmers, once articulated, carry appropriate weight in research policy and priority-setting. 49 Sustainability: Issues for Social Scientists in Agricultural Research Centro Intenacional de Mejoraniento de Maiz y Trigo (CIMMYl) Lary Harrngton Intenational Institute of Tropical Agriculture (IITA) Anne-Marie Izac Midhael Swift Karen Dvorak International Center for Living Aquatic Resources Management (ICLAM Robert Pomeroy International Livestock Centre for Africa (ILCA) Brent Swallow Simeon Ehw Introduction There is a degree of overlap between the subjects of this session and that of the first. The session on global germplasm and ecoregional mechanisms focused on the ecoregional concept adopted by the CGIAR as a research approach to the challenge of sustinability as an issue. This session focused on the issue itself and the role for social science in addressing it. In introducing the topic of sustinabiliy, it is common to disparage the word as vague. Vague or not, it has unveiled neglected dimensions in development, in agricture, in the environment, and in the consistency of institutional performance, reorienting the thinking of academics and practitioners alike. Perhaps the vagueness should be seen as an asset. The impact made by the word deserves acknowledgment. Larry Harrington's paper looked at the definition and measurement of sustainability and the opportunities for social scientists at the IARCs in research on natural resource management. Anne-Marie Izac, Michael Swift, and Karen Dvorak sought to justify the integration of ecology and economics and outlined a research paradigm to address sustainability. Robert Pomeroy's focus on the institutional aspects of common property resources showed how the fishery is an example of the results of pressures on a resource, with clear relevance beyond fish to soil, water, and forests. Finally, Brent Swallow and Simeon Ehui's paper outlined experiences in integrating natural resource management issues into the research agenda at 1TCA, as an IARC. co Sustainability, Economics, and Ecology: Issues and Opportunities for IARC Social Scientists Lanry Ha=ington, CIMIIT Harringtn pointed out that the notion of sustainabiity encompasses population growth and pollution; deforestation and land degradation; agroecology and energy cycling; erosion and interenerational equity; biodiversity; global warming; and the ultimate fate of humanity. He began by outlining issues under active discussion. Issues Surrounding the Concept of Sustainability Harrington held that most differences in interpreting the meaning of sustainability stem from the diversity of answers to the question, what is it that we wish to sustain? People working at different scales and using different approaches have diverse perceptions, which are as far apart as people who emphasize the ethical duty of humans to steward nature and those who study energy flows in a single ecosystem. He expressed concern that the measurements of sustainability are fundamental to making the concept operational, but that how to go about measurement is less clear. Much of the recent work has focused on total factor productivity, an approach viewed as too narrow by those who interpret susaiability in terms of diversity, energy cycling, rreversible land degrdation, exhaustion of nonrenewable resources, or global climate change. During the next 60 to 100 years, by which time the global population is expected to stabilize, food security and sustinbility will become an indivisible issue. There is reason to believe that food security and agricultural sustainabilit are both at risk in Asia. Epansion of crop area has ceased, little new investment is being made in irnigation infrastructure, and farmers in advanced agricultural areas have already adopted most of the new productivity- increasng techniques that have been made available by science. At the same time, there is evidence of ongoing degadation of land and water resources. Stagnating productivity combined with land degradation will make it exceedingly difficult for Asian agriculture to continue to contribute to growth of income and employment and alleviation of poverty. Chronic poverty, in turn, can cause continued land degrdation and population growth. Land degradation in frmers' fields frequently is difficult to define and measure. Degradation caused by slow processes of nutrient mining and loss of organic matter may be consderably more difficult to detect than damage caused by erosion. Water-induced land degradation, especially salinization and sodification, is a particular problem in selected lowland irrigated areas of Asia. Change is also expected in farmers' extemal circumstances 51 though global warmig, depletion of nonrenewable reurces, and an irevesible loss of genetic diversity. Ecosystem resilienc is naced by system diverty. Modern monoculture is charactzed by low levels of diversity and is viewed as having a fragile ecological equilibnum controlled by exmernal inputs rather than mtenal feedbackc mechanisms. arrington noted that IARCs maintain at their core breeding progams fost system stability and resilience through the induction of disease resistace and tolerance to abiotic stress. Increased participation of famers in germplasm screning would help to balance the scientt' view. Resource degrdation is frequently the consequence of common property rights. Although privattion of common property rights is often suggested, it can place resources in the hands of the very wealthy. Actions to foster user groups and community management of common property rights can be preferable. Extemalities and the ability of certain farmers to impose off-site costs on others exacerbate sustinability issues. Finally, the public health problems linked to chemical use in agriculture are another dimension of the sinability question. The tade-off with food security brings these problems under the umbrela of "sustainable agriculture." Opportunities for IARC Social Sdentists A wide range of institutions are targeting the sustainbility issue; many are not involved in agricultural technology. Harrington asked how the CGIAR contribution is unique, how sustainability concens can be reflected in centers' programs, and again, what IARC socia scientsts can do to inform this decision. Direct contnbutions of agncultuwal research aim to halt or reverse processes of resource degradation through farm- or community-level interventions in thratened areas (e.g., research on alley cropping). Preventive contnbutions of research aim to avert or forestaUll resource degradaton indirectly through geneation of employment or income growth, even in areas where resources are not threatned as long as labor markets opate efficiently. Perhaps the most efficient route for IARCs is to help prevent problems before they occur. Harrington also addressed the extent to which center socal scientists should monitor external global level threats to farm situations from global warming or from an inreversible loss in genetic diversity for the centers' mandate crps. In monitoring internal threats to farm situations, Harrington noted the links with FSR. When FSR is combined with the dynamics of sustinability, an array of methodological questions arise, many of which are best addressed by social scientists. These include: 52 * How can social sentis foster farmes pticiation in diagnosis and experientatin in addring sinili tmes? * What ae cost-effective ways to montor farmers over tme, to track changes in resource quality, technology use, input levels, and factor prducii? * How can carryove effects be explicitly intmduced in the design and analys of on-farm trials? * When asesing altemative technologies, how should we adjust for xtemral effects and nonuse values such as off-site costs of aesion, on-site user costs, option values, and otiers? * How do social scientists oDnfront tie posibility that it may be essential to degrade one area to save another? Technology is clearly important to sustainaility, but poliy intvetions can be even more powerful. To what extent should IARC social scienists concentate on analysis of policy rather than generation of technology? Are there other policy reseach institutions to do this? Would a shift in emphasis sacrifice the IARC social scientists' compaive advantage in woddng partnerships with biological scientists? Srng alliances may be needed with other institutions. As IARCs pursue research on sustinability themes, there is a danger that they will move beyond their comparative advantage. Few of the CGIAR centers have adequate in- house crse to seie all the issues and opunities described ealier, particularly centers that focus on germplasm development. How should the boundaries be drawn on investments in research focusing on sustainability? What directions should these limited investments take in the different centers? And finally, how can social scientists contribute to these deciions? Integration of Economics and Ecology in Addressing Sustanabilit Issues in Agroecosysnms Anie-Marie Izc, Michael Swft, andKren Dvorak; ITA The goal of achieving suinable agriculural systems will remain elusive until some method for assessing susUinability in systms is developed. Sustainability in hoisfic nodons of ecological stability and reilience, economic viabfility, and human wmre. Scientific research requires a more specific and rigorous definition susceptible to measurement. Reconciliation of the broad holistic concept with a rigorous measuable one is the key to an opeational approach to research on suslainability. 53 To attempt this ncliton, a strategy was adopted at ITA based on a worldng definition at the scale of the cropping system that the authors consider uses a minimum number of components and assumptions. From this baseline, the authors developed a more complex definition at the agroecosystem scale, seeking a conceptual framework for research on sustainability that is both comprehensive and scientifically sound. They then summarized the salient points of conceptualization at the cropping system and the agroecosystem scales. Cropping System Scale The definition concentrated on the process of biological production and its economic output: *A cropping system is sustainable if it has an acceptable level of production of harvestable yield that shows a nondeclining trend from cropping cycle to cropping cycle which is resistant in terms of yield stability to normal fluctuations of stress and disturbance over the long term." With this as a baseline, a corollary proposition is made that sustainable production is dependent on the capacity to replenish resources so that stocks are maintained at constant or increasing levels from crop cycle to crop cycle. This approach to sustainable assessment necessitates study of the biophysical and economic processes that determine efficiency of resource use in cropping systems (see Figure 3). Susuinability requires conservation of biodiversity and resources vithin the cropping system and minimiion of the impact of the cropping system's activities on the external environment. his, in turn, necessitates widening the scale beyond the cropping system to the farning system and the small catchment. Agroecosytem Scale The cimumstances of the farming system become particularly relevant at the agroecosystem scale. The authors referred here to Sub-Saharan Africa (SSA). Key characteistics are that small-scale farmers in SSA live in village communities and that fanning activities are organized partly at the individual household (farming system level) and partly at the village level. Crop fields rarely more than 2 to 3 hectares in size and fields of different specialties coexist on the landscape with a variety of other land uses (e.g., grazing areas, areas of natural vegetation, and small plantations). Interactions within these different units in the landscape are regulated by both ecological and socioeconomic factors, with the direction and rate of the former largely dictatd by the hydrological cycle. The small, local catchment presents a natural boundary for such interactions. The majority of biotic dispersions of significance to agricultural sustainability-the movements of potential pests and their predators-probably occur within this spatial scale. 54 Figure 3. Efficiency of resource use In a cropping system lime Trarse In T1a out SpaW ounwy to sysem ham system Variable Inputs Harveste product e.g.eflizer-N - yield. N Nonharveted a~~~~ Iv E pln bms Transfer Renewable Resource Renewable Transfer phase - available - N .g..crop resoue - N (or cyde) (or cyde) - SOM - N ~~-SOM-N intemal /nd materps. net Lot Resource e source\ e.g.,N- lxaUon\ " ~~~Spada[ Botndr SOM =soD organic malter.E Lost resource e.g..gaseous -N 55 The authors proposid that the combination of ecological and social integration of resource maintenance creates a land unit of particular relevance for sustainability: *The village agroecosystem is the most relevant spatial scale for the analysis of agricultural sustainability for the small-scale farners of SSA and is defined as an agroecosystem managed at the social scale of the village community and at the ecological scale of the small catchment.' Interactions link systems together at different scales -- farming system, catchment, regional, and supraregional (see Figure 4). Decisions made at the farming system level have repercussions at the same scale, as well as at lower and higher scales in the hierarchy. These are mediated through various economic and biological processes such as nutrient cycling and the market mechanism. In decisions to manage their lands in a given way, fanners integrate a wide range of ecological and economic parameters emanating from both wider and smaller scales of the hierarchy. Because these processes transcend farm boundaries, the authors argue that the villagelcatchment system is a more appropriate scale of analysis of sustainability than the farming scale in the SSA. Propositions The authors offered four propositions: M The most appropriate scale for assessing the sustainability of small-scale agriculture in SSA is that of village agroecosystems. This agroecosystem is managed at the social scale of the village community and at the ecological scale of the small catchment (10 to 102 heclares). A sustainable village agroecosystem must be able to respond to normal changes in exogenous variables, that is, those that have a yearly probability of occurrence of at least 33 percent. Te system must have the capacity to respond to changes as well as to the disruptions caused by the by-products of agricultural production by maintaining a nondeclining trend in all its essential environmental (agricultral and ecological) outputs. A system becomes unsustainable when its assimilative capacity is surpassed and disruptions trigger downward trends in at least one of its output vectors. An appropriate period of time during which a sustainable agroecosysten' can be expected to maintain nondeclining trends in its output vectors is about one decade. The authors acknowledged the need for definition in some of the terms used. They believe the factors that determine sustainability have a number of implications for the 56 Figure 4. Hierarchy of agricultural systems Supraregional System dimate large landscape and national economic intemational geomorphology, etc. poieconomiceuns poicies and trade Regional System natural resources industrial and urban i institutions endowments, human rural regions regions nalines resources endowments Village/Catchment System toogapy.pstnofmmigsocial norms and populaon, etc. farmin l households ctustoms, roads, topography, pest nanfarrning ~~~~infastuctroas, Farming System variousinputs | decisionmaking [crops restocks agroforestry Cropping System weed inEpEthogens r]esidues Soil System l E Efauna Emicflora nutrients water l so Jds 57 research agenda, including strong focus on the ecosystem and a close integration between biological and social science. That research needs to be conducted at a number of spatial and temporal scales, some of which are wider and longer than the existing cropping and farming system, which is the traditional focus of IARCs. t also requires assessment of the processes that determine interactions on the landscape and understanding of the communal aspects of resource management. Beyond this still, the sustainability of agriculture is influenced by policy factors determined at the regional and national scales. There is a need to determine which components of such a sustainability agenda should be included in the research mandate of IARCs. Institutional Arrangements for Community-based Coastal Fisheries Management: Common Property Resources Robert Pomeroy, ICLARM This paper explored the concept of community-based coastal fisheries management as an alternative to government-centmalized management, within the framework of the fisheries resource as common property. Pomeroy presented a brief discussion of community-based management, examined institutional arrangements affecting the management of common property resources, and identified a research agenda to increase understanding of the portunities and issues. In recent times the transfer to centralized control over coastal fisheries management has often overestimated government skills and underestimated the capacities of the traditional management systems, which reflect the long and often difficult experience of people in managing local resources to meet their own needs. One result is that the management cultres of taditional common property resources, in which group size and behavioral rules are weli understood, have often been replaced by open access. This free-for-all creates the incentive for each person to exwtact as much of the resource as possible before others do and results in degradation of coastal fisheries resources. Rapidly increasing pressures on a finite coastl resource base suggest that the traditional management systems and local skills need to be partnered by the power of govemment to provide enabling policies and institutions and assistance. Such partnerships have become known as community-based management (CBM). CBM involves the decentralization of authority and functions to the resource users, community organizations, or both. It starts with the premise that people have the innate ability to improve their quality of life, and what is needed is support to organize and educate people to mobilize available resources to meet their needs. It strives to make maximum use of local knowledge and experise in developing management strategies and to have self- management, where the community takes responsibility for surveillance and enforcement. 58 The aim is to provide a sense of ownership over the resource and make the community responsible for its long-term sustainability. Property Rights and Resource Management Regimes With renewed emphasis on poverty alleviation and the acknowledgment that rights to common property are often all that separate the poor from destitution, renewed attention has been directed to the role of local, community-level institutions in the management of common property resources. These institutions are the foundation of CBM and share two key characteristics: exclusion (or control of access) of potential users is problematic, and limited supply means that consumption by one user reduces availability to others. The failure of new paradigms for managing development has renewed interest in overcoming 'the tragedy of the commons.' In this context open access, or unrestricted entry, has been confused with common property management, a distribution of property rights in which a number of owners are 'coequal in the use of the resource. Pomeroy argued that recent, more careful analyses of the foundations of common property, combined with better understanding of the management of collective goods in the developing world, strongly suggest that common property is the best management option in some circumstances. Resource Management and Institutions Fisheries management policies in developing countries are shaped through a convergence of institutional interests among resource users, resource stakeholders, communities, local government, national government, and international agencies. Pomeroy noted that those who are concemed with reorienting fisheries management to promote socialy sustainable policies need to recognize this convergence of interests. In seeldng management to replace open access, a precondition is a consensus that access be closed. A Research Agenda The geeral research question to be addressed is how various types of institutional anrangements function and perform under different political, social, cultural, resource, economic, and institutional conditions. Specific topics relevant to research on common property resources and CBM in coastal fisheries include: * How decision rules are developed and modifieci, if they are resilient enough to remain viable when challenged, and wI"t prevents them from breaking down * Clashes between national and local legal systems including the study of actual and potential conflicts * If users will invest in these arrangements when they do not have the legal rights to resources and recognized status as a resource user group 59 * The contents of traditional managenent systems and indigous lknowledge conerning resourc, regmes, and rles and the costs of enocng arrangements. Any resource management stegy should acknowledge institutional arangements hat are responsive to local conditions and variations in those conditions. This may reiquir the accptance of approaches that are ricipatory and more decntralized, such as CBM or comanagement. Governments need to recognize that smaller or onal uits, such as villages, are better equipped to manage their own resources an are large authorities and may be more effective for rural development and sustainable resowurce mana t. The success of these altenative situatons will depend on the legitimacy given to such institutions by national governments. Integrating Natural Resource Management Issues into ILCA's Research Agenda Brent Swalow and Simeon Ehuz, LZCA Swallow prested an ovemriew of the steps ILCA has undertakn to bring sstainabiity issues to the forefront of its resarch agenda by concentraing on resource management. The paper focused on the roles that socal scientists should play mi the development of technologies, institutional arrangemets, and policies to prmote sustinable icultura systems and impoved resource management Swallow outlined components of a suggested research framework to expose the viomnmental problems of taget groups, the constaints on and oppotunities for impwoved naural resource management, and the environmental inpacts of existing and proposed tedhnologies, policies, and institutions: * Identify the natural resources and ecological systems that are most vulnerable to degradation and crate oppotunities for susinable increases in production. * Study the dynamic processes associated with agricultural production and environmental change. * Within each agrecological zone, farmmg system, or both, identify goups that pose environmental rsk and would be most affected by technological or institutional change. * Identify the technical, economic, and institutional factors that are important for resource maagement practices. 60 * Evaluate the probable impact of alteanative inteventions for addressing resource management problems. Sepate issues rdated to institutions and policy from those rdated to available technology and oDnstraints to its adoption. When formulating a research framework, important groups to consider include those posing environmental risks, those exposed to environmental risks, and those exposed to risks of economic and social change. Swallow and Ehui then discussed technologies, changes in institutions, including community institutions, infrastructural development, and policy reform as the instruments available to governments for devising solutions to these types of problems. Swallow reported that a workshop hosted by ILCA on livestock and resource management policy, an international group of policy analysts and rearchers agreed that the most important research issues regarding the social and economic aspecu of natual resource management fall into three categories: * The effects of resource management institutions (rights, rles, and conventions) on resource use and conservation and the adoption of resource- improving technologies * T1he ways in which credit, commodity pricing, marketing, and selected macoeconomic policies shape individual production and resource management incentives = SThe environmental impacts of technologies for disease control, feed production, and animal-drawn cultivation In response, ILCA sees a need to balance the portfolio of stratgic compared with applied research, and continent-wide compared with ecoregional activities. The center plans to concentrate on the ecologically endangered areas of the East African highlands, the semi- arid zone of West Africa, and the high rainfall areas of West and Central Africa. In doing so, ILCA will explore both short-term changes in productivity and long-term changes in resource management institutions. ILCA will also investgate new resarch approaches to apply abstrt theories and concepts and seek wider collaboration from NARS and other CGIAR centers. lssues and Discussions Four working groups identified the following topics for further discussion: * Measurement issues in sustinability 61 * The new roles for social scientists, particularly economists, in susinability research * On-farm research and natural resource management * Sustainability research for high-potential and low-potential marginal areas Measurement Lssues The important issues identified are the basic questions surrounding measurement: what to measure; for what purpose; with what precision; with which instruments; when, and how frequently; and how to use the information? Within this framework, levels of aggregation were discussed as a basis for a decision on the geographic or systemic focus for research on natural resource management. When data are collected will depend in part on the speed of change in the characteristics of interest and the cost and precision of the instruments needed to measure them. The discussion highlighted several issues to be addressed: v Intergenerational issues. How can the preferences of fture generations be articulated? How can such preferences be effectively used? How many generations should be included? D Extnalities. Cross-sectionally at a point in time, who should be responsible for rating the needs of 'downstream" users and those in the community who wil not be benefited by the intervention-a social evaluation? Over time, should future generations of nonusers and downstream users be drawn into the evaluation? - Issues in setting research priorities. How can tade-offs be examined, ex ante, between reseach to better identify and diagnose sustainability problems and research to assess altemative solutions for well-understood problems? * Uncertainty. How can randomness in data collection and in subsequent analysis be accounted for? * IrIeversibility and system resilience. How can "kinks" in system supply functions be identified and the costs of going beyond those ldnks through technical change, policy, or both be assessed? The worldng group thought that use of external input was only a preiminary indicator of sustainability in a farming system. Total factor productivity (rFP) was seen as a valuable though still partial measurement. The participants believed that ecological, soocultural and human welfare (health and nutrition) were required dimensions beyond the TFP famework. 62 The group also saw a need to improve the quality of weights used in calculating TFP to capture both the social costs and the effects of imperfect markets. The group endorsed the importance of modeling but emphasized the need to concentrate on and manipulate only those parameters central to system improvement. The group saw variation in too many parameters as inimical to on-the-ground resolution of problems. In modeling work, the farmer's perspective is important but must be complemented by a social perspective that incorporates extemalities, intergenerational effects, and off-site influences. An even wider need is to research the influences of policy from the regional and intersectoral levels on production activities on the farm. The group thought that a better understanding is needed of the circumstances under which fanners invest in, rather than mine, natural resources and of when mining is the optimal short-term strategy, for example, when it allows producers to escape the poverty trap and take a sustainable perspective on future production activity. Social Scientists' Role in Sustainability Research The working group emphasized that sustainability will demand initiatives in all dimensions of CGIAR research: germplasm, natural resource management, institution- building, and management and policy. In addition, new areas, with little history in the CGIAR, will require new research to supplement ongoing activities: 0 Research existing, local institutional arrangements for managing resources as well as action-oiented research for enhancing land management. 3 Study existing land tenure patterns and how they affect natural resource management. * Address valiation, particularly of the 'intangibles.' * Define minimum standards for the quality of natural resources. The centers need to identify "threshold measures' (defined in economic terms) to evaluate the acceptable or unacceptable levels of a resource to determine if research should intervene. * Understand what long-term perspectives of farmers affect their current management of natural resources and how research might compensate. * Because of the long timeframe involved, in evaluation of natural resource management research, shift enmphasis from adoption to monitoring and use benchmarks to assess progress. 63 The workdng group specified new roles for economists in commodity research, particularly the interactions between improving productivity and sustainability. These include the impacts of commodity management on pollution, nutrient levels, physical soil erosion, quality of water, and incidence of pests and disease. The group posed the question, under what conditions do commodity gains contribute to sustainability? It proposed research on the relationship between generation of income from commodity gains and investments in sustainability issue. It also noted the high level of maintenance research required in crops such as wheat and asked how much investment in a permanent improvement in durability would pay off. On-farm Research and Natural Resource Management Several issues were seen as relevant for on-farm research. Prominent again was the measurement question: are there clear methods to measure whether experiments improve sustainability? Farmers modify continuously, and the approach taken to on-farm research and the methods used must be flexible enough to adjust. Some technology can be expensive, and poor farmers are not always able to participate. Some technologies also require group action, which poses problems when communities have no group action tadition. To test methods in on-farm research in natual resource management, IARCs need to collaborate wi grassroots institutions that are involved in the issue. Both government and NGO partnerhips will be important. Research in High-potential and Low-potential Marginal Areas This working group concluded that high-potential areas are targeted more strongly than marginal areas because of their role as the supplier of food for those in absolute poverty in the urban sector. Because degradation is less overt in high-potential agriculture, there may be a higher research overhead in diagnosis for articulating environmental problems. As Figure 5 indicates, sustainabirity issues are as important in the high-potential production zones as they are in the marginal zones-perhaps even more important. High- production environments are perceived to have the potential to meet the growing demand for food most efficiently. This may pose a greater risk as production nears capacity. Several questions were raised: How far can we push these systems without jeopardizing their long- term viability? Can we estimate (ex ante) the risks associated with expected productivity gains from research aimed at high- and low-production zones? Conclusions lltimately there is no dilemma of choice between productivity and maintenance of the resource base; the two are indivisible. Yet in the short term, under particular local crcumstances there may be a logic for depleting resources in order to survive or to shift capital assets from one space or time to another. Opportunities for alleviating poverty that use short-term exploitative strategies should be seized when they break the poverty cycle and 64 FIgure 5. Risk and productivity in favorable and marginal environments Productivity ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ;MdcIty Pmdue" ~~~~~~~~~~~~~~~~~~~~~~~gain t -- lPotentW | s.w. ~~~~~~~~~~~~~2010 Alsk Risk--------- - - ~~~~~~Potential ulsk The present 1992 65 then permit previously poor farmers the luxury of a long-term sustainable perspective. Extreme green"-conservation for its own sale-and even a conventional industrial country perspective on the environment run the risk of abusing this logic and thus the livelihoods of poor people. Figure 5 illustrates the high downside risk from failure in favorable environments. This, together with the dynamics of urbanization, gradually pulling population from the marginal lands as overall population growth rates slow, suggests a premium on research to keep the high-potential areas sustainably productive. It also suggests a research focus on those marginal areas that interact strongly with the high-potential lands, affecting their productivity, their sustainability, or both. 66 Socioeconomic Databases, Priority-setting, and Impact Assessment lnratonal Service for National Agncultural Research (ISNAR) Phil!p Pardey Doug Horton Intenional Institute of Tropical Agriculte (IITA) Karen Dvorak, in collaboration with Simon Carter, Tropical Soils Biology and Fernlity Programnme (7BF); Mike Collinson, CGL4R Secretariat; Barbara Granin, Rutgers University; John Lynam, Rockefeller Foundation, and Steven Romawnff independent consultant International Irrigation Management Institute (IEMI Edward Vander Velde International Laboratory for Research on Animal Diseases ([LRAD) Adrian Mukhebi John Cin Brian Perry nternational Centre for Researh in Agroforestry (ICRAF) Susan Minae Steven Franzel Centro Internacional de la Papa (CIP) Peter Ewell Introduction This session focused on dat requients for the resrch cycle eaborated by John Lynam in his overview paper. He called for integrated planning in three dimensions: the reerh cycle from priority-setting through achievement monitoring to impact assessment, spatially and geographically across the centers of the CGLAR system, and between the mandatrelated pesective of the individual centers and the global peVpective of TAC. He saw common data bases as the foundation for such integration. Historically, the predominance of physical and biological disciplines in agricultural research has promoted the development of physical and biological data bases, particularly of climate, soil, and vegetation. These have sometimes been abused, for example, zones of yield potential have been used as the basis for decisions on where crops are best grown, without consideration for the importance of socioeconomic parameters. Even a 1986 CGIAR publication on agricultural environments dismissed socioeconomic parameters as too transient for research planning. Although physical and biological conditions do limit the potential of each crop, socioeconomic data clearly are vital. Each set of conditions creates opportunities for a range of crops and options for the production of each crop; the highest yield potential is an eareme 67 among those options. Farmers' choices of crops, and options for growing each one, are dicUted by subsistence preferences, market conditions, interactions with other enterpnses in the farming system, and other nonfarm possibilities for cash and labor. Occasionaly the choice will be to grow the crop for its highest yield potential, but it may lie anywhere along the range of opportunities and options. This, the largest session, offered seven short papers. The first three covered information and its organiztion. Philip Pardey addressed information for planning and priority-setting, identifying gaps in theory and methods that needed attention. Leading a handful of co-authors, Karen Dvorak examined issues in the development of a socoeconomic data base for African agriculture. Edward Vander Velde complemented this paper with a commentary on the prospects for intercenter collaboration in the building of socioeconomic data bases. The next two contributions examined issues in data collection and analysis for impact assessment Adrian Mukhebi, representing an ILRAD trio, considered data needs in modeling strategies for control of livestock disease. Susan Minae and Steven Franzel looked at issues and options in adoption and impact analysis in agroforesty. Finally, Doug Horton moved to an institutional theme, reviewing monitoring and evaluation practice in IARCs and NARS and issues arsing for socal scientists. Peter Ewell extended this theme to the use of a network for monitoring and evaluation among NARS. Information for Planning and Priority-setting Philip Par*ey, ISNAR Pardey concentrated on the generation and use of information for planning and setting priorities for agriculturl research within the international agricultural research system and with the CGIAR's partners, the national agricultural research systems in developing countries. The Role of Research A prerequisite for developing information to guide the planning and priority-setting process is a clear undending of the objectives to which rsearch is asked to contribute. Objectives for such research must ensure that performance measures can be developed and that an informal means of weighting or ronalizng multiple objectives is possible. Pardey stated that even when policymakers agree on which (nonoverlapping) objectives to include, the weights given to those objectives differ among individuals. He noted the inmprtnce of maing a clear distinction between value judgments about weights for objectives and technical judgments about the effects of rserch on yield changes, adoption rates, and so forth. The task of identifying the trade-offs of the 'social welfare function' for a research system is endogenous to the planning and prioity-setting process. It involves ascertning 68 the oppounity costs of using research policy as an instrument of soci! policy. Information about how much growth must be forgone to direct benefits from research toward a specific target group can modify the expectations being placed on a research system. Pardey emphasized that such information cannot indicate the least-cost way of achieving social objectives. Combinations of the research portfolio and a second policy instrument (such as a tax and income transfer scheme) may constitute the best and least costly approach. These considerations strengthen the case for viewing policy formulatiun and portfolio selection in research in a holistic fashion with regard to other instruments of social policy. Modeing and Measuring Pardey noted that many priority-setting studies resort to crude, ad hoc attempts to weight combinations of objectives, criteria, and partial measures of research performance in ways that violate reasonable economic principles, perhaps reflecting the difficulty of deriving credible performance indicators. Socioeconomists have made progress in modeling the effects of research-induced technical change; markets have been horizontally disaggregated either spatidlly or across different socioeconomic groups. The same models make it possible to capture the spillover effects of research in price and technology. Vertically disaggregated market models provide the means of assessing the impact of postharvest or preproduction technical changes. As development proceeds and the value added in agriculture increasingly moves off farm, these models will permit assessment of the distributional consequences of processing, marketing, or input supply-oriented research. To the extent that decisions to invest in agricultural research and subsidize or tax agriculture are linked, these models lay the basis for thinkng about the 'political economy" aspects of agricultural research. They also provide information on the distribution of the costs as well as bnefits of research that can assist research managers in funding arrangements that minimize the policy distortions in the agricultural sector. Crow-commodity Effects Pardey highlighted the nature of cross-commodity effects. In the humid tropics, for example, muldiple cropping is common. Technical change that affects one crop directy may affect another through its impact on resource requirements and growing season, through effects on soil structure, fertility and pests and diseases, and as a substitute for domestic consumption. Some progress is being made in determining 'general equilibrium" measures of the benefits of research in these instances. Research Benefits i the Presence of Environmental Exteralities Although some work has been conducted on the general economic implications of environmental externalities in agricllture, very little information is available on the implicatons for returns to research. Issues such as sustainability, global warming, and food safety can be considered conceptually in a conventional supply-and-demand model that allows 69 for a divergence between private. and social costs or benefits from production. For example, when groundwater is polluted with agricultural chemicals or salt or surface water is polluted with eroded soil, salt, or agricultural chemicals, the social cost of agncultural production is greater than the private cost perceived by the farmers. In most cases the externality is borne by the country or region within which it is created, but not necessarily by consumers or producers of the commodity in question or even by the agricultural sector. In such cases, agricultural research may: * Reduce the social cost of production without offecting the extenaity. * Result in private cost savings while increasing the eiernaly without any reduction in social costs. * Reduce the externality without any effect on private costs. Pardey asserted that numerous alterative scenarios are possible. Although there are substantial difficulties in valuing external costs and the impact of research on these costs, in principle, externalities can be modeled like taxes or exchange rate distortions. By taldng into account the effects of the welfare of all affected groups, one can %ompute the effects of research-induced supply shifts on producer surplus, consumer surplus, government revenues, and those whio bear the costs of externalities. The Research Production Function Finally, Pardey addressed recent attempts to evaluate the effects of research on the measurement of the research production function. The 'K factor' (i.e., the potential reduction in unit cost arising from research), the probability of research success, and lags in R & D are all detrmined by the resources committed to research and the manner in which they are deployed. The sensitivity of these three technical parameters to changes in the nature and level of resources deployed is critical for assessing the margnal (expected) impact of research and with it, the marginal allocation of research resources. More information is needed about the extent to which changes in experimental yields give good measures of likly shifts in industry- level supply curves, particularly whe technical change is neither scale- nor factor-neutral. 70 Information Use Pardey concluded that the toolkit for evaluating the economic effects of research has expanded greatly in recent years. Many of the available modeling procedures are not widely known, however, and the data to apply these models often are not available in a form that is readily accessible for research evaluation. He agreed with John Lynamn that there are potentially large economies of scale and scope for an individual center, the CGIAR system, or NARS to collate these data for research evaluation and priority-setting purposes. Much remains to be done to estimate meaningful technological parameters such as unit cost reductions, probabilities of research success, and R & D lags. Sociocconomists can help structure and analyze such data to permit a constructive dialogue that adheres to basic economic principles. Development of a Georeferenced, Socioeconomic Data Base for African Agriculture Karen Dvorak, IITA, in collaboraton with Simon Carter, TSBF; Mike Collinson., CGIAR Secretariat; Barbara Grndin, Rutgers Universty; John Lynam, Rockefeller Foundation; and Steven Romanoff independent co=",'=-4&W This paper is a summary of the issues to be addressed in planning and developing a georeferenced, socioeconomic data base for African agriculture. The issues are germane to the development of other continental data bases and, with some modification, the linldng of continental data bases for systemwide applications. Dvorak discussed several broad areas for data base groupings: objectives of and rationale for data bases; the theoretical framework; scales of data collection and analysis; spatial referencing and projection; sampling procedures; selection of variables; resources for collection of field data; and data base management, housing, and distribution. Dvorak stated that a data base designed to be all things tn all people would serve no one well. The multiple objectives of the CGLAR system-productivity, equity, and sustainability-raise the issue of choice of roles for any CGIAR data base. The process of agricultural research and of related R & D activities also must be considered. The diverse social science programs that exist in concert with the CGIAR system, and their representatives (national agricultural researchers, extensionists, policymakers, donors, governmental and nongovemmental agencies, and university researchers), are all potential contributors to a socioeconomic data base on agriculture. 71 The authors recommended that the primary objective of a socioeconomic data base be development and transfer of agricultural technology within Africa, with research planning at the program and institute levels as a secondary objective. Theories and Data Bases Dvorak stated that development of a georeferenced, socioeconomic data base for Africa should be guided by theories of change in agricultural systems and successful paradigms for research planning and technology transfer. Although agricultural systems in Africa have changed rapidly and dramatically during the past century, with few exceptions, formal agricultural research has been absent as an impetus for change. Better research planning and implementation require characterization of agricultural systems and their dynamics. Such characterization depends on systematic data collection, which in turn depends on a theory of change in agricultural systems under African conditions. Current Western aggregate economic models are unsatisfactory in differentiating the structures of African agricultual systems. Literature on the dynamics of African systems dating back to the early 1950s characterized the systems, often in relation to reducing areas of fallow. A theoretical link to factor intensities was proposed by Boserup, who concentrated on increasing population and the consequences for labor productivitv. FarmiLg systems ais_5 znd household economnics have also miade contributions to theories of change. More appropriate theories and data bases need to be developed together. Validation and development of theory are hamstrung by the lack of data covering appropriate variables for a range of systems. Theory and empirical research must proceed interactively, but increasingly they do not. More comprehensive theories of agricultural change require correspondingly comprehensive empirical work. As the theoretical construct evolves, it provides a framework for analyzing disparate field data from a variety of sources. A standardized socioeconomic data base for African agriculture has the potential to inform the theoretical and empirical understanding of agricultural change. Selection of Variables for a Data Base Characterization is not unidimensional. Different systems will have different relative intensities of factors, including choices of crop, maintenance of soil fertlity and crop husbandry practces, land class for exploitation, and labor invested in capital improvements such as irrigation and terracing. These dimensions will define production techniques and trajectories of agriculturai change that underpin the adoption of new tec'mologies. It is important to understand the determinants of farmer choice in each dimension. Dvorak then discussed the scales of data collection, spatial referencing, and sampling. She highlighted aspects important to the construction of a socioeconomic data base for Afica: making choices between qualitative and quantitative data; detmining the number of 72 independent layers for application of GIS; asesing the feedback effects among the farmiing systems, household organization, and demography; and the extent of the trading sphere. The size of the trading sphere in which farmers sell produce, and that sphere's integration with larger spheres, is a key to understanding the response of farmers to market incentives. Data are required on differentials between export and import costs of staple commodities and market access. If the farming system is specified, a cross-section of cases could allow a sequencing of responses and the organizational changes that result from market penetration. Market systems probably will require a separate layer in a georeferenced data base because the structure of information from the household level and the village level differs. For a GIS. all data records will be spatially referenced to permit the use of several socioeconomic layers and an overlay with climatological and edaphic data. C, particular interest (and difficulty) will be methods of combining national-level data witt, data from the community, household, and farming systems levels. Resources for Collection of Field Data Two approaches were discussed for collection of field data: stand-alone data callection and 'add-ons' to continuing data collection efforts. Stand-alone field work will result in more consistent sampling procedures, a more even quality of data- and faster data base development but will be significantly more costly. The add-on approach offers economies of human and financial resources in getting to the field, but it requires an "economical' list of variables to avoid overburdening the field effort to which it has been appended. Data collection inevitably will proceed on an opportunistic, country-by-country basis- The most likely scenario is a pilot project in a limited number of countries, seiected to coincide with research projccts that are already planning to allocate human resources to colection of field data. nstitutional Responsibility and Data Base Management In establishing a data base, many practical matters would have to be resolved, including data "ownership," policies on data base access and citation, entry into the public domain, upkeep of the data, and policies for distributing duplicate data sets. The authors proposed that it would be most economical to align the data base with one or more institutions that have GIS capacity and very good external communications facilities, parts of which could be dedicated to the socioeconomic data base. Possible sites included the African CGIAR centers, the United Nations Environment Programme, the UN Food and Agriculture Organization, TAC, the World Bank, IFPRI, and universities. Currently, only the African CGIAR centers have the field presence to support the data collection activities. 73 Socioeconomic Data Bases: Prospects for Effective Intercenter Collaboration Edward Vander Velde, IIMI Vander Velde concentrated his discussion on the methodological issues of rationaliing socioeconomic data bases. He argued that the common mission of IARCs and similar perspective of their scientists should be underpinned by common data bases. Partcipation would encourage and strengthen NARS. Using the example of an IIMI irrigation program in Pakistan, he outlined an opportunity for intercenter collaboration with IFPRI on a socioeconomic survey of rural households, which each center had launched separately. Both projects began in 1986 and were combined as a result of whz.t Vander Velde described as a serendipitous opportunity in 1991. Because the centers had rot planned jointly from the beginning, the final results did not completely satisfy the needs of either institute. This experience served as the beginning of ,oint planning of a pooled data project and heightened awareness of the need to rationalize the use of socioeconomic data bases, particularly with diminishing IARC resources. Vander Velde made the argument for increased collaboration among IARCs, particularly in the areas of shared information and combined resources. The joint use and management of socioeconomic data bases, for example, would achieve economies in data collection by avoiding duplication of effort and establish, where possible, a basic data set to be enhanced for other research-specific purposes. Vander Velde cited the following questions in data base development: v What would be a common minimum set of variables to be included in an intercenter socioeconomic data base? * What would be suitable protocols to guide the collection, archivir.g, distribution, and multiple use of intercenter socioeconomic data bases? * How can the basic spatial framework or unit(s) of IARC socioeconomic data collection activities be made more appropriate for research objectives? * What are the opportunities for collaboration among IARC social scientists in GIS applications? * How can NARS be effectively included as participants in the development, management, and use of socioeconomic data bases? 74 Modia and dectronic developments have simplified rapid and reliable distribution of data to users in diverse locatons. A process to infonn IARC social scientists with up-to-date information on the data bases would be required and could be facilitated through networking. Most collection of socioeconomic data is organized by and subsequently made available for civil or administrative divisions. This, Vander Velde argued, is particularly true for data collected by public-sector national organizations. Much CGIAR research, however, uses other spatial units (e.g., agroecological zones, irrigation system service areas), and there is a critical need for socioeconomic data that can be related to such spatial units more accurately and efficiently. Geographers, in particular, would want as much socioeconomic information spatially referenced as possible. IIM's experience in Pakistan confirmed a need for capacity-building in the social sciences. National institutions commonly contrau their social science research needs to public or private oganizons, usually specialized in economics, or simply ignore the socioeconomic dimensions implicit in their work. Initiatives of CG[AR social scientists in socioeconomic data bases may provide useful standards suitable for use by NARS. Assessing hnpacts of xnproved Control of Livestock Diseases: Lssues in Collection and Modeling of Field Data Adrian Mukeebi, John Cuny, and Brian Perry, IERAD Poor animal health and nutrition remain constraints to livestock production in Sub- Saharan Africa. Standard control measures for diseases such as trypanosomiasis and East Coast fever are losing their effectiveness through poor delivery and drug resistance and their sustainability as rising costs of combative measures, such as sprays and elimination of vector habitat, become prohibitive for national veterinary services to underwrite. Livestock keepers have lost faith in the ability of government services to provide adequate animal health care, and numerous experiments in alteniative delivery systems have been tried. In this climate of change, it is impeative to assess the impacts on smallholder farmers and other clients of existing and potential unproved livestock disease control technologies and of their delivery systems. To determine the best ways to increase livestock productivity without causing deleterious effects on farm, national, or regional economies of Africa or the natural environment, ILRAD established an epidemiology and socioeconomics program (SEP) in 1987. The program has four main objectives: To quantify and predict the relative economic importance of the infectious and noninfectious diseases of livestock in different regions and livestock production systeas in Africa and elsewhere, and the justification for their control. 75 * To identify factors at both the national and the farm level that are essential for the successful production, delivery, and adoption of improved control measures for livestock diseases. * To determine the probable economic, social, and environmental consequences for the application of improved control measures for tick- borne diseases and trypanosomiasis in different locations, production systems, and agroecological zones of the world. * To support ILRAD's tick-borne diseases and trypanosom programs in assessing the cost-effectiveness of alternative research options for the development of disease control technologies. Although small in size (two social scientists led by a verinarian), SEP focuses on assessing the impact of ILRAD's strategic research, that is, the impacts of available diagnostic techniques and future vaccines, at the farm, national, and continental levels. SEP's social scientists are limited to asssing ex ante the probable impacts of methods to control immtinological diseases in the absence of field trials or adoption of a certain product The long-term and often uncertain nature of this research and the colhaboraing institutions involved make it difficult to establish cause-and-effect relationships and to attrbute the effects on end users to ILRAD's research. To evaluate experiences in disease control, SEP scientists use other technologies to understand the oganizaional and social behavior associated with control of the disease. They also have developed models to simulate the probable effects (economic and noneconomic) of disease control interventions in target livestock populations. Mukhebi noted that ILRAD's work is complementary with NARS work. SEP has the advantage of conducting impact assessment at broader (regional and continental) levels, whereas NARS have the advantage in more focused adaptive and applied research, which tends to be location specific. A major problem for SEP in its collaborations with NARS is the lack of trained social scientists working in national livestock and veterinary services. The veterinary services appear not to have been included in much of the farniing systems research training in Africa in the past decade; consequently, they lag behind crop research programs in integrating economists and other social scientists into research teams. This fact can cause difficulty in obtaining the necessary data, because SEP scientists prefer that NARS social scientists have primary responsibility for conducting impact assessment for the technologies in a host country. 76 Socioeconomic Analysis of Agroforeshy Adoption and hnpact: Issues and Options Sasan Minae and Steven Fraln, ICRAF MIeae and Franzel presented ICRAF's approach to analyzing adoption studies in agroforstry esearch. The center focuses on thfree ypes of studies. Analysis of the adoption of indigenous agroforestry technologies. These studies help identify farmers' needs and preferences (diagnosis); assess the potential from adaptation of existing technologies (design); and improve understanding of the characteistics of the technology and its users and assess its potential for areas where it is unknown (adoption). * Farmer-designed trials. Farmers select new species or technologies and design and conduct their own experiments. Researchers supply information to the farmers and monitor the farmers' experiments. Ihe adoption process is thus monitored from the introduction of the technology to uptake or rejection. These trials have both diagnostic functions (identifying farmers' needs and preferences) and design functions (monitoring farmers' problems in using the technology and the modifications they make). * Technology-testing trials. Researchers and farmers design and manage formal experiments, monitoring the adoption process. In addition t their study of the adoption process, these trials often have diapostic and design functions. The authors listed several questions arising from the present dominance of these three types of study: v What should be the balance among the types of adoption studies mentioned? What other types could be used? * Are the same tchniques (e.g., multiple regression models) that are used for understanding farmers' adoption of new technologies also usefil for examining the adoption of indigenous technologies? How can the techniques be adapted to handle the complexities of agroforestry (e.g., multiple products)? * What models are most appropriate for understanding adoption processes through on-farm trials? * How can adoption of an agrofory technology, which might be one tree w trees along a boundary, be defined and measured? 77 The authors also noted two important of studying adoption process in yAroforesMy on-firm trials: the limitations in sample size caused by the high cost per participant and the many years required to obtain results, and the biases caused by the contact between research staff and farmer. Variabs To Measure In Assessing hnpact Minae and Franzel reiterated tfie importance of socioeconomic analysis for assessig the productivity, profitability, sustainability, and acceptability of agroforestry technologies. Sustainability assessment must take into account users' perspectives and natural resource management. This calls for integrating analytical tools from many fields, including ecology, agricultural sciences, and socioeconomics. Economists often use profitability, with a measure of risk analysis, as a proxy for acceptability. In assessing the acceptability of new technologies, social scientists need to draw upon methods and experiences from such fields as cognitive psychology, political science, and consumer marketing. Variables to be monitored include: D Changes in the stock of natural resources such as maintenance of soil ferdlity and vegetation rehabilitation and improvement. 3 Outputs in terms of products (e.g., poles, timber, or fruits), intermediate inputs (fodder, green manure), or services (live fencing to protect crops). * Farm-level inputs (land, labor, capital, and management). These must be assessed in qualitative and quantitative terms from an understanding of their impact on the farming system as a whole. * Income generated from the sale of agroforestry products or from increased production resulting from the effect of agroforestry technologies. * Decisionmaking processes regarding how costs are allocated and the benefits shared within a household. * Externalities, both positive and negative. Even when based on results of on-farm experiments rather than adoption of new technology, impact assessment raises several issues: * Sustainability is not clearly defined and not easily measured. * Inputs and outputs are difficult to measure and value, particularly multiple products that are a mixture of nontraded goods and services. 78 * Data on inputs, especially labor, are difficult to collect. * In agroforestry, analyses must be conducted over a multiyear period, and the results are often highly sensitive to the discount rates. Finally, the authors said that ICRAF's experience points to the need for teamwork in assessing the adoption and impact of new technologies. This involves interactions between the biophysical and social sciences and a healthy inteiaction among the social sciences. Anthropologists and sociologists have a critical role to play; economists alone cannot do justice to the social perspectives involved. Monitoring and Evaluation in National and International Research: Issues for CGIAR Social Scientists Doug Horton, ISNAR Horton's paper dealt with three main topics: definitions of monitoring and evaluation (M & E), the status of M & E in the CGIAR and in NARS, and issues related to the roles and functions of social scientists. In agricultural research, Horton saw most M & E concepts borrowed from the development field and from economrics. He explained that development agencies generally view monitoring and evaluation as segments of a project cycle that begins with project identification, appraisal, and planning and continues with implementation, monitoring, and evaluation. He then offered a general definition more suitable for agricultural research: "MIonitoring" refers to observing or checkdng on research activities and on their context, results, and impact. Its goals are to enhance the performance of research, to document activities and results, and to warn scientists, managers, and research stakeholders of deviations from initial goals or expected outcomes. 'Evaluation' refers to judging, appraising, or determining the worth, value, or quality of proposed, ongoing, or completed research activities and of their results and impact. In reviewing the status of M & E, Horton said that its activities have not been systematically pursued in the CGIAR centers or NARS. He saw both the CGIAR and NARS sharing common activities in M & E, including day-to-day monitoring of research activities; progress reports, internal program reviews, and annual reports; project budgeting; external program and management reviews; and impact assessment. Horton presented eight sets of questions and issues that relate to the role of social scientists in M & E: 79 * Why M & Eli? Horton gav three reasons for moniborng and evaluating agriculturl research: to meet accountability requirements or gain support for research, to contribute to internal decisionmakdng and the research management process, and to contribute to knowledge in a more geeral sense. Insue: How might M & E contribute more effectively to decisionmaking and management within agricultural research organizations? * M&E for whom? M & E processes should generate useful information for specific groups including scientists, research managers, policymakers, development workers, farmers, and others. Issue: Who are the priority users of information from M & E, and how well are their needs being met? * What to monitor and evaluate? M & E is needed at different organizational levels including research projects, programs, institutes, and systems. For each level, research goals, plans, proposals, ongoing activities, results, and impacts may be assessed. Issue: To date, formal evaluation within CGIAR centers has focused on impact. Is this sufficient? - VWhat organizational options? M & E functions may be integrated into research program management, carried out by a specialized unit, or assigned to a multipurpose unit To date, most CGIAR centers have assigned M & E functions to line managers or social science programs, often on an ad hoc basis. The lack of clear responsibilities for M & E may be one cause of poor continuity and weak institutionalizion. Issue: Could changes in the locus of responsibility for M & E functions enhance their institutionalization and effectiveness? * WWhich techniques and processes? Most formal evaluations of agricultural research ha-re employed economic methods, often sophisticated. There has been notably less progress in establishing cost-effective institutional M & E processes that fully exploit the many techniques available. Issue: How could social scientist contribute to improving institutional processes to capitalize on available methods? * How to balance rigor and relevance? Horton argued there is a danger in a trade-off between rigor and relevance, erring on the side of rigor. ILsue: Could social scientists produce studies that are more useful for managers without compromising their professional standards? * How to use M & E results? Strategies that enhance the use of evaluation results include careful identification of decisionmakers and their information needs; involvement of potential users of results in planning the evaluation 80 process; rapid completion of studies; use of appropriate, transparent, credible methods; and presentation of brief, clearly written reports that focus on key issues and offe relatively few, feasible recommendations. Jaue: What could be done to enhance the use of M & E results? What roles for social scientlsts? Social scientists can contribute to improving M & E by participating as researchers, developers, impleme rs, and facilitators. Issue: What is the optimal combination of roles to ensure the contribution of social scientists to organizational performance and their job satisfacion? In general, Horton thought there was a move within the CGIAR toward improving M & E. To date, however, research activities have seldom been monitored or evaluated in relation to goals stated in center plans. Although extenal reviews are an established feature in the CGIAR, he argued that the link between review recommendations and future plans and operations remains weak. A great diversity of situations can be found at the national level. Some of the larger and more developed organizations are institutionalizing M & E processes for agricuItural research, but the norm is for formal M & E to be limited to activities that are needed to meet the requirements of ministries and donors for extemal accountability. Working with NARS in the PRAPACE Network to Develop an Information System for Monitoring and Evaluation Peter EweU, CIP Ewell's paper reviewed issues that have arisen in a recent collaborative project to develop a data base and documentation system for monitoring and evaluation within the PRAPACE network. In 1982 the Programme RMgional d'Amelioration de la Culture de la Pomme de Terre en Afrique Centrale (PRAPAC) was founded to link the potato research programs of Burundi, Rwanda, Uganda, Zaire, and CIP. The network has since expanded to include sweetpotato and two additional countries, Ethiopia and Kenya. The name was therefore modified to Programme Regional d'Amelioration de la Culture de la Pomme de Terre et de la Patate Douce en Afrique Centrale et de l'Est (PRAPACE). Its purpose is to provide a framework within which national programs can work together to improve their information base for prionty-setting, impact assessment, and research management. The primary goal is to find simple, flexible procedures that respond to the needs of NARS. PRAPACE is designed to strengthen the capacity of the national commodity programs to develop and transfer technologies. The countres have emphasized seed programs to produce high-quality, basic seed tubers on experiment stations for multiplication and distribution to farmers. The network sponsors both group and individual training in key areas of scientific knowledge and research skidlls. Courses, workhops, meetings, publications, and exchange visits provide opportunities for scientists to exchange informaion and develop working linkages between their programs. Since 1986 the network has been 81 funded by die U.S. Agency for Inional Development (USAID), and CIP has povded an interatonal scientist as coordinator. The goals of tie project are: * To compile existing information to document activities related to potato and sweelpotato research carried out within the network. * To implement rapid surveys in selected area to document the uptake of new technologies by farmers. * To collaborate with NARS to update and improve data in each country. The national scientists decide what information is needed to improve priority- setting, research management, and c ter potential applications. * To support the network's coordination office in maintaining a data base that will provide organized data for planning, reviews, and analytical studies. The intention is to collect, analyze, and present data with as much consultation as possible among the network partners. Ewell said that the challenge for the network has been to develop a "rapid monitoring and evaluation systemn which provides the collaborators with information that is immediately usefid to them, without taxing their resources. He noted that for any type of information, there is a tade-off between the degree of detail the most demanding user might like to have and the practical issues of what data are available, reiaiively easy and inexpensive to colle, or both. Information covered in the data base includes: v Training. Information is broken down by subject matter and type of training: regional, in-country, and individual. Training profiles of particular programs and individuals help both the network and NARS to ensure that sciendsts and technicians have the skills they need. * Research. Experiments and other research activities are routinely documented in the annual reports of each progran of the networL The information is shared through an open, consultative process. The network coordinator monitors the information and works with NARS to imove and standardize reporting in areas where clear benefit is expected. Ewel gave the example of using the data base to monitor germplasm and 'trackdng of the various stages in evaluation to allow comparisons between programs and agroecological subregions. 82 * Seed production. The national programs have placed considerable emphasis on the production of high-quality seed tuben for distribution to rgional development agencies, NGOs, and private farmers for multiplication and distribution to farmers. More effective feedback and a more cost-effective seed program are the desired outcomes. * New technologies and impact assessment. The PRAPACE network and member national programs are constantly faced with demands for information from the farm level. The network uses surveys conducted in a quick overview fashion to provide a relatively inexpensive means of collecting information. Data target important production trends, identify new problems as they appear, document the adoption of new varieties and other new technologies, and provide a mechanism for feedback into the research process. Ewell concluded that the success of the project will depend first on the institutionalization of monitoring (the routine collection of meaningful data in a consistent and useful format) and second, on the development of an analytical capacity within the network and the national programs to use the information for planning, priority-setting, and evaluation effectively. Issues and Discussions Compared with 1985, when an intercenter workshop on characterization was held, CGIAR centers today are devoting far more resources to charctizon. In this context, the group emphasized the importance of linldng data on socioeconomic characteizations with ongoing biophysical charactezation efforts and coordinating the large number of isolated data to prevent duplication. The role of CGIAR centers as collectors and organizes of census data was discussed. The group generally thought that it would be better for the CGIAR system to dedicate a group of people to compile and organize data for the whole system, enter them into GIS at the most disagd level available, and provide access to all CGIAR scientists than for individual centers to collect data related to their own area of focus. As a result of discussion, the participants requested that an informal working group be formed, with representatives named from each of the centers present to act as liaisons for future infonnation exchange. In addition, the formation of a steering committee was seen as desirable to address the following topics: * How to collaborate on planning data banks by providing 'irkputs' for other centers' planning processes. 83 * How to devlop a data base ownhip potcol to faciate saring of data banks. Working group pardcipants ited the work of lITA's Collaborative Study of Cassaa in Afica in this regard. * How to be documenng who is doing what and what softwa is being used at each of the centers, in the hope of sndaring the data base across the CGIAR system. The group also made the following rcommendations: * Socioeconomic data collected in the future should be georofencod to facilitate sharing of information across ceters. * IFPRI should provde a prce senes data base to other centers. * Te steering committee should nominate one or two persons to collaboxate with the biophysical scientists who are currently working on GIS and biophysical data banks. The difficultes of interpreting and usming data collected by others was cited, as was the need for better and stdardized documentation. It was noted that IFPRI is preparing trining manuals on this topic and that others might do so. The group addresing impact assessment examined the different experences of the CGIAR cents. In general, the range of metiods has been wide, and no effort has been made to sundardize the level of analysis, the type of data collected, or the resolution and aggregation. The group recognized that impact studies have been done at the technology program leve and even the national level and that the methodologies and emphasis vary; the grup als eognized, however, that comprehensive assessmt of impact is rlatively expensive, and in some cases, simple anecdotal information may be adequate. The types of stdies identified were adoption (long-trm studies), impact (evaluation of specific progams), and achievement (requiring systematic integration and docmentation). For issues of measurement, the group noted that * Assessment of impacts on germplasm is relatively easy. * With respect to natual resources, measumets of evironmental impacts may require development of new methodologies. * In policy impact assement, ex-ante analysis and modeling are important. Ex-ante inpact assessment methods evaluate what others have achived, examine histoncal trends and modeling, and establish benchmarks for measumt of surplus and 84 extenalities. The relibility of prdiction or probability of sestivity geneally varies. Ex- ante impact analysis is used in decisionmaking for priority-s g and raidng of issue, design of progams, prediction of outcomes, forecasting, and feedbacL Thc group noted that parameters wthin the CGIAR will vary among commodities and natual resource management, and recommended that for the fuure, the exsting models shud be exchanged among centers. Condusions Current ircumstances siould bring greater eaon of the value of socioeconomic data in research plamning; it is widely accepted that tie miraction of human decisions with physical and biological processes is central to research on naural resource management. There is a clearer understanding of the research cycle from farmer back to farmer in smaliholder agriculure and its potential as a grssroots information source. Fialy, the need for close linkages between technology development and policy formulation to speed adoption and diffusion of technology is increasing the value of information from grssroots sources. Although these ideas are gaining broader acceptance, integrted planning across the CGIAR system will m challenging. The histrical autonomy of the centers, ompetition among them, and their wanness of centmal command and control will inhibit efforts to move toward a coherent planning process. Old attitudes die hard, but change will be aided by pressure from donors and by the new dual structm of the CGLAR with its global and ecoregional entities. Incentives for intercenter programming and the need for cross-locational syntheses of reslts will move the centers toward collaborative planning and common data bases. 85 PART TWO INVITED PAPERS Social Science in Agricultural Research: Implications and Issues for IARCs Jock Anderson, Wodd Bank I take as my point of departur the assumption that the primary domain of social scientists is the undending of human factors in various contexts. The context here is international agricultual r , and the human factors are as follows: * Human beings as they work within agncultural systems that are the subject of research * Human capital as it enters the fabnc of decisionmaking -within agncultural systems * Human labor resources as they are applied within agncultural systems, especially faming activities * The products of human effort, such as capital items that come under the influence of the manipulations of managers Such a definitional context gives socal scientists a broad canvas on which to work, and it is thus not suprising hat they have contnrbuted so strongly-and perhaps increasingly-within the agricultural research community. The main surprise is that it took as long as it did for their potential contributions to be recognized and implemented (Hardin 1981). First a cautionary note about some of the limitations of this paper. Many issues are well represented on the program, for example, should the CGLAR rely more on policy work done in other institutions and buy particular sldlls and studies as they are perceived to be needed, as is done in some of the biological sciences from time to time? What are the implications for the organization of social science research of the move to ecoregionalism within the CGIAR system? How central is the role of social scientists in fostering partcipatory methods and more attention to fanners' perspectives within LARC work plans and the dmination of research pnonties? Are there more effective ways of fosterng such activities and addressng the expressed needs of NARS? In reflecting on the roles of social scientists, would it help us to think of different types of work along a continuum of basic to strategic to appled to adaptive research? I do not have good answers to aU these challenging questions but, in our collective wisdom, perhaps we can find worthy insights. Now let me consider some of the roles and functions of social scientists worling in IARCs. I*e vews expressed are the personal views of the author and de not represent the vews of the Wofid Bank, its m nt members, and affilfates. 88 Roles and Functions I rcall several convernations in the early 1970s on what the role of social scientis in IARCs might be. In that era, the only thing that some biologists knew of social scientit was that they were harsh critics of the distributional consequences of the first-round fruits of the 'green revolution'; therefore, in some minds, professionals of this persuasion were not to be trusted and surely were not safe to house under the same instutional roof. It is reassuring to observe that tfinking on this matter has come a long way. This progress is the result not only of changing evidence about the beneficiaries of the green revolution, as is encapsulated in the Hazell and Ramasany (1991) book, but of the more informed view of the role of agricultual technology in economic development, a role that social science research has helped to articulate and extend but that has yet to transate into more effective targeting of research for purposes of income redistibution. Special Roles for Economists Each of the social science disciplines brings its own special skills to the research tasks tackled in IARCs. The most central skill that economists offer is as an 'economizing perspective," namely, the efficient achievement of staed ends with limited means. The mechanics for implementing such a perspective vary widely from simple budgeting exerises designed to relate costs to retumns in an informative way, epitomized by Perrin and others (1976) and the modernized CIMMYT Economics Program (1988), through stat-of-the-art methods of research pnority-setting based, for instance, on expected economic surplus changes (Norton, Pardey, and Alston 1993). The methods that might be used in an economizing perspective for research planning and monitoring are highly diverse and feaure their own costs and benefits. In an attempt to review this field (now rather dated), Anderson and Parton (1983) argued that, under many public research cicmstnces, economists should be especially economizing, circumspect, and cautious. I do not wish to imply that the primary roles for economists amount only to the routine application of known principles to the economizing of phenomena studied by others. Many fields of application will require novel concepts and methods, making the contibution of economists genumely one of research rather than mere consultation. The proliferation of FSR methods and terminology in the 1970s and 1980s suggests that parallel development may need to be contained or coordinated in the current larger CGTAR system (see Collinson 1993). An example is the increasingly popular minefield labeled 'sustainability,' and much work will be required to clear the field and find useful operational paths through it (see, e.g., Ehui and Spencer 1990). Such issues return us to the appropriateness of IARC research workers delving into different crannies of the research continuum. They could also lead us to a reconsideron of 89 FSR work in the system and the historically and pvely) crucal roles of socal scientists in it. For brevity (cf. Anderson 1991b), I will not address FSR in detail her Spelal Roles for Anthropology- and Soiology-orented Profesdona Not all of the disiplines usually classfied as social sciences have been involved in IARC work and probably never will be. The other group that deserves singling out for its parficular contributions includes socia anthrologists and rural sociologists. Research workers in these disciplines bnng to agricultural researh a distnive perspectve. The work undertaken has spanned a wide spectrum of activities. At one end are long-tern sudies of life in villages undertaken to improve understanding of the attitudes of the villagers and the constaints under which they approach their agicultual work generally and their agricultural improvement efforts in particular. In the middle, perhaps, is active involvement in FSR teamwork (e.g., Tripp 1991a, b). At the other end is work that might be depicted as sociological pespectives on the adoption process (e.g., Ashby 1991), especialy as it relates to the uptake of innovations that are linked to the centers' technological research programs. Whate the scope of such work, the perspective is an important one in agiculural research. Some of my colleagues in the Bank would argue-and I think the case made is persuasive (Cernea 1991)-that the history of development interventions is so littered with damning failures arising from neglect of basic social issues that such neglect in any development work can no longer be excused. I am confident that anthropological aspects of aicultural and natual resource nagement are now unlilly to be underplayed in future IARC work, particularly with regard to gender, which deserves a separate secdon. (My perCption of anthoologists as "quiet achievers" has recently been jolted by a University of Georgia newsletter [Olszewsii 1992].) An Agenda for Gender The sociology of sexual differentiation in agricultural development in all its facets, from activity roles in traditional farming systems to staffing patterns in IARCs, has been the subject of increased attention from many interest groups concerned with agrcultual rserch. Of the conflicting lists of -must be dones- for the CGIAR hIpact Study, this area was a common and central concern. With considerable asistance from Janice Jiggins (e.g., rgins 1986), we in the core team grappled long and hard with the complex of contingent co ns id e ations. Pardy aided, we hope, by our efforts (with Janice's help scaed unobtvly through Anderson et al. 1987 and Anderson, Herdt, and Scobie 1988), the gender analysis activities in the CGIAR system (represented, e.g., by Poats 1991, and van Herpen and Ashby 1991) have greaty advancd the debate and the artcuation of research and develoent priorities that will be ongoing matters for agricultal research agencies, both national and inn. The ozation for which I work is alsD nizing the importance of gender-related policy in many dimensions, for example, through creaton of 90 the Women in Development Division, a crash program of rcruitment and a ent of women in the BanL This recognition is wdl illuated by the remt semnt by te Bank's chief economist that "investment in the education of girls [sic] may well be the highest return investment in the developing wodd" (Summers 1992). Mainstream Socal Science Research Wdhin the Centes Most of the preceding discussion has dealt with the role of social science in a transdisiplinary context. A significant amount of social sience work within IARCs can be described as 'mainsm in the sense that it does not differ greatly in form and function from what might be conducted in university departments of soci sience with an mterest in agriculture and rural development matters, or eve in other enai agencies that share similar concems, such as development bank and rlevant UJN agencis. The queston that must be asked then is, 'What is it about some agu ly oriented social science research that makes it best conducted within the IARC network generally and perhaps more specifically in non-biological research centers such as IFPRI and, possibly, ISNAR, for example?" The explicit mandates of IARCs have usualy provided a clear indication of the nature of the work to be addresed and a rationale for it. The research institute concerned can then implement the research progam in the conventional "IARC way," which features uninterrupted pursuit of research themes that are usually financed adequatey to achieve justified objectives, housed in a location that is convenient or reeant rative to aleative sources of supply, and provided with physical resources and workng arangemnt that are conducive to high research productivity. A second justification relates to filling important gaps in knowledge. if other instutions around the world were engaged tD a sufficient degee in the conduct of social science rearch oriented to all the important aspects of agrcultural devpment thae would be no case for institutionalizing altnative arangements in IARCs. The peristent lack of social science capacity in developing country NARS is a case in point, espeially when it comes to socil science analysis of sgic techncal questons. Derek Byeee, stuck by the absence of social scienists among woutide" partpnts at the 1992 Intemational Rice Research Conference, observed that this neglect stegthe the case for more IARC social scientists, at least in the short run. These explanations may be made more sophisticated but not necessarily more compelling by bringing up such matters as a wcritical mass' for moe effictive conduct Consider, for instance, the multitude of agricultmal social science research actvities under way at any given moment among thouands of instites of higher laning in many parts of the world. Consider, too, the ldndred activities that are analogously supported in public research agencies with broadly similar research agendas, albeit often focused on national research piorities, such as in the Economic Research Service of the U.S. Department of 91 Ag e The ovrwhming i n is of uncoordinadresearch agendas being tackled with highly dispare methods, data collections, and social objectives. The codai, cohert agenda, and developm ented priority-setting ihrent in, say, a CGIAR-sponsored endeavor pvides a firlher broad justification for suh work. Coming from an orgition for which similar aleate provider questions can be and are raised (particularly one that proclaims the joys of privatiaton), I am sympathetic to the cricism that is implcit in an examination of public-sctor undr ings and interventions, especialy when others on the public perphezy seem ready to do what is required. Some might argue that universities, in orienting doctoral studaents in their slection of research topic, are more driven by novelty and publishabilit than by social importance, but this argument may skate on thin ice, given the incentive stucture that supposedly accompanies the "invisble band in dissertation work. The question of efIctivness also can warm the ice. Can it be claimed, for instace, that IARC social science umts have been notably more productive in terms of upoaitant publisbed papers per research peraon or dollar spent than some of the leading unversity centers? Although I bave not done such an analysis (cf. Anderson 1978) for the CGIAR system as a whole or for individual centers, empiricism leads me to pose this question. Simiarly, one could exploe research products of different types such as artidcles in leading nrofessional journals compared with larger repor that might not be so readily produced and distbuted in the academic workLd Here is another empirical question that may be worth expration by those charged with independent evaluation of such IARC researh, peraps ing the ime yet unherled in an enal program review context to use analytical stucftu such as those employed by Thorpe and Pardey (1990). Let me clarify that these remarks about publication should not be interpreted as an opinion that published products constitute the only, or even the most, valuable things done by social scientists. Demonstbably they are not, but I suspect that their importn has been underplayed in the past When I discused some of these points with Peter Hzell, he thought that the preceding paragraphs downplay important aspects of IFPRr's mainsteam research rationale. Wben IFPRI was created and subsequently incorporatd into the CGLAR system, the case was made by Sir Jchn Crawford and others that there was a need for research on policy issues related to the production and distribution of food, with partcuIar cocern for the share of benefits of agriculual research going to the poor. The distictive fears of IFPRI's mainstream work-empirical, microeconomic-based studies that are not easily conducted by others because of their high cost, long duration, and thus limited atraciveness to academic social scientists-mean that somethng like IFPRI was required to fill this institutional gap. The point of having much of this work centralized in one institution rwather than spread among unnected elements of the LARC network is to reap the benefits of economies of size and scope. 92 It is also argued tha IPFR's politica neutrality as an independent institutio, unattache to the subsdiary objectives of lending and aid progrms, is a ageous m analying food policy issues in many developing counties. In fact, IFPRrs mandate is to train and bild up local research capacity, presumably one day putting itself out of a job. Althugh the food policy ent and FRIs work program have changed the l970s, the case for having a center such as IFPRIwith a mainstam program is still valid. Certainly, its ouputs are apprecated by other inton agencies that do not have the cmmitment and flexibility to addres such issues in depth. Would a fresh considerton of mainstream social scence research in an expanded CGIAR system give us much the same a ona partition that we now have? Would analogous size-and-scope argments lead to a concentraton of more stategic research efforts in a central International nstitute for Agricultua and Resource Policy Research-an unproonceale lIARPR-and analogously centrally coordinted vage studies? HDw would noncntraized [ARC social scientis relate to the center? I do not know if such questins have even been posed in the hals of visonary CGIAR thinkng, but it would be reassuing to hink that someone, somewhere, soeime wil struggle with them. Considable progress is being made to bring rmenal isses into the elanded agenda of the CGLAR and of the social science progams of its institu (the post-1992 IFPRI looks icreasinly li an HARPR), but other topics of importance have yet to surface in significant ways. I am thinkin partcurly of the political economy of agrcultur and agriclturl research, and cogent studies of the economics of institutions (and politics). The work of Roe and Pardey (1991) ilustrates activity that may facilitate undertnding of the possbiliiesfor agrclual research to contribute hr to of Ognizational theoy spefically, and manag t scince generaly, may have much to offer, especiay in ISNAR and thence to NARS, but leant disciplinary ills have yet to be hired into the sstm. Plaing ad Pioriy-sett Rols The econof ists in soca science IARC programs have a responsibility to be intersted in the economi efficiency of the research process, even if this means looking inwad to the Woines of the research portfolio of social science research progrms. In times of budgetary stess, it is an unfortunate fact of life that difficult judgments bave to be made about the efficiency of allocation of increasingly scarce resources. It behooves agdcultual conmstso devote some of their effort n tbe center work to com ding and documenting the opms for altentive research resouce allocations The o rnities are great, deie the in-house political problens, possibly strained intpswonal reliips, and difficutie that may have to be accommodated. In this regar, a good example of helpfiu policy-orented analysis is the recent work of Ben (1992), which guides i aonal resources toward natonal breeding prgrms 93 according to the coomics of particular breaeg siao. He shows that small coumtries cannot indulge in the compete specum of steps involved in, say, a wheat-breeding progam, and this fact provides a potentdal role for an international center to achieve something tat otherwise may never get done. Using a mode of the economics of the wheat- breeding process, he provides guidance about the size and nature of breeding progrms where particuar forms of assstance for plant-improvement prgrams may be contributed most effectvely. The same idea natually underies all social science effort directed toward detemining more efficient research resource allocation. Theme are tehical, as wel as economic, judgments to be made in this work, and one such endeavor was recently undertaken by TAC in its 'execisesw in assessing priorities (e.g., TAC 1992, Cummings 1992). Natonal programs, which always face such resource allocation asks, can prbably be helped greatly through the provson of analytical stuct=us, such as explored by the Austrlian Centre for International Agrcultual Research (ACIAR)- and IFPRI-suported wOrk of Davis, Oram, and Ryan (1987) and by the ISNAR-suppoRed efforts of Norton, Padey, and Alston (1992). Unforamately, the nature of these activities is such that pemanent and ideal answers are umachievable-o be Inder tm has been my recent inclination (Andes 1992). The work must thus be an ongoing task analogous to trying to keep one's sights on a moving tget. The CGIAR system should continue to look to ISNAR for gidance in mate of method, but IARC economists in gmenal will necesarily and prperly have specidal roles in working the sights.- OgazationAlng_ for pleeting Social Science Rah in LARCs Many different models have been implemented in setting social scienists to work in fte centes At the extemes are indpndent scia science progrms or even more scalized economics pgramsi the case of CIMM ) through to highly dispersed angements featuring social scientists woing within multidisciplinary teams of commodity research efforts (as was the case earlier at CIAI). All the different angements, perhaps for good structal reasons, seem to have been productie In searching for wisdom on this matter, I am tempted to encapsulate the experiences to date in an Anything Goes Principle: 1t doesn't matter much how one does it, just do it but engage good people. Clearly, organizational form cannot be considered adequately in a completely detached way, and other si s bear on the ultimate productivity of whatever model is implemented, such as the degree of specializd and dfentiated technical expertise that must be dmae by socia scientists workng m biological and tecnological research institutes. With the transfor on under way of most centers as etr global germplm centers or ecoregonal research centers, it will be usefu to consider the relevance of difernt modes of implementing social science research and pehaps to examine previous alternative modes. The results of such an investigation should also prvide beter guidance to NARS as they move tortoise-lile to establish social science research capacity. 94 in compre g thc success and failures of the past, it is pertinent to conider advioe and perceptions from the governing board as to the cogency of the roles of social scitiests, th personal pmesctive of the Director Genewal as to the virtue of such a role (and thus the institutional commitment to seeing that it is an effecdve one), and the percepions held by biologicl scientdsts working within an institute as to the legitimacy of the si science effort with which they are obliged to work coopratively. TMe latter point devolves to cnrter used in recruiting staff in the physical and biological sciences with an eye to their attitudes toward collaboaive work. Pehaps usnmela y, the growing proporton of Directors General with a social science badkgound is a positive indicator in this regard, although the scarcity of social scientis on CGIAR boards is less enco g. Non-social Senc Flements of the Skills Mix Research work within IARCs is often multidiscplinary. Insofar as social scientists are bringing social and economic persectves to bear on technologically oriented research efforts within an institute, however, they must exhibit an acceptance of work that might better be described as t plinary, a term first coined for me by the late John Flinn of IRR. Transdisciplinariness is probably as much an attitude as anything else, but it is aided by some multiplicity in the formal tuaning expenences of social scientists. I may be accused of personal bias but, as I have observed in LARCs and reflected on the matter, I-as did John Longworth in commenting Hardin (1981)-can see no better tmining than a broad agicultural sciences qualification as a basic tansdiscplinary peparation. If this is done as it sbould be, reflecting the many disciplines that contrbute to the comprehension of modern scientific agriculture, students are exposed to many different disciplines in their taining and can the integrate their more spealized social science training into a tndly transdisiplinaryfrmewor Of course, there are many ways of apprabing a ransdisciplinary perspective, and there are highly productive center staff from a diverity of mmother" disciplines. Role Models and Replication in NARS IARCs have increasngly tkn on roles that are variously described as twaining' or 'capacity-building' with regard to human capital resources in national systems. There ae several elements to fosteing relevant skills in national programs, including the general one of providing a model for the involvement of social science reseach within the broad naional agricultral research effort In too many nations, it is still exceptional to have any social science involvement, and what is done is typically restricted to the recruitment of a token economist who is gien pnsibilities in agricultura deveopment plaing and aommodity price and trade policy. As Brian Hardaker of the University of New England has wryly observed, such social scienists tend to become Jills and Jacks of all trades, workig mainly on administaive task such as acting as counterparts for visiting teams from development agencies, processmng aid proposals, and fielding queries from many sources 95 including local poiticians and introna bureaucrat. The centers can be important in influencing some desrable developments in national systems, just by demonstgating that the international system regards such research involvement and accompanyg staffing as critical To the extent that IARCs and NARS are complementy, thdr o tional arangements and work prorities will likely differ. Other contingent roles have developed. hrough their training and other progams, the centers have given formal and informal training eences to both social scientists and natal scientists working in national programs. Giving some social sence pspectives to, for example, agrnomists within broader production training pgrams has been usefid in the slow advance of acceptance of social science withiin NARS. This may be an appropriate place to raise the question of service in a center as a career prospect. I would vertue the opinon that LARC socia scientists have mostly done well out of the system in career terms, whether they have gone on to managemet positions in the system, continued analogous work in a devdopment bank, or retreated to the quiet life in academe. I think it is probably healthy for the relevance, adaptivness, and modernirty of IARC social science work that so few center social scientists have cracked the twenty-year service banier. Thaf After broaching the issue of sklls mix, let me address the related issue of formal prearon for such work in the centers. Several consideations are germane, and diverse approaches have been tried. For instance, one social science progam directr took the view that, in seekng out researchers who could submerge their disciplinary spectives to the greer common good, the best approach was to look for candidates who had already established themselves in a substantive field in socidal science. He reasoned that such people would no longer feel under pressure to prove themselves and coud thus adjust more readily to an interactive mode, even if this led to research activities that would be less publishable in maim joumals. At the time, this position may have been defensible, but it is less so in recet times grven the proliferation of journals that cater to a range of tansdisiplinary tastes, such as Agicdral Systms, Agiiutural Economics, and the Asian Journal Qf Famnwg Sysems Researh. Another approach that has been widely adopted is to seek young socal scientsts of the "politcally correct' mindset and to introduce them to work in the centers through postdotral progams. The Rockefeller Foundation has been successfil in these endeavors, and notably so in involving women in such social science worL Other attempts have been made to engage doctor candidates in work in the centers' multdisclinary tasks, including the novel entrpise betwoen Cornell University and CIMMYT involving doctoal candidates of one cohort in sevreal disciplines worldng in collegial fasbion (Contmras et aL 1977). This enteprise seems not to have. been successfu enough to have wan-anted a follow-on actvity. 96 In the domain of social science training, there are many ote possbilities, including the involvement of the centers in aduate education. This has been tred in several modes, but the one that seems to work best, as exemplified by the practices developed at ICRISAT, is where students spend a sustained period-often a year-within a doctoral program based at the center and supervised jointly by staff at the center and staff at the university. Positive exwtemalities abound for closer integration of resarch agenda between university and center staff, and for exposure of young social scientists at the impressionable dawn of their carers to center research work and lifestyle. It is all too easy, however, to overwhelm a center with gaduate students. An excessive number can tie up staff time and overburden the scarce research facilities available for tionary staff, even if such additional human resources come at low unit cost to the center. As Alex McCalla, current Chairman of TAC, has observd, the issue is the multiplier effect of 'Taining' activities, regardless of the degree of formality involved. In this sense the matter is a crucial element of the centers' contributions tr national capacity-building. Where Sociology and Economics Can Happily Meet In identifying elements of a social science research agenda that may give IARCs a comparative edge, at least in longevity of funding, one important candidate has been termed "village-level studies" (VLS), a tenn seemingly coined at ICRISAT but now widely used. The noteworthy character of such studies, apart from their being a natural meeting place for anthropological, sociological, and economics-related disciplinarians, is the commitment to continue them over a sufficient number of seasons and years to provide panel data on a consistent basis and with fine detail on transactions within the village. These studies provide data that are unique in richness and relevance for some of the investigations that must be undertaken by social science researchers, particularly those charged with ficilitating echnological change in village communities. Previously I alluded to some of the IFPRI experience, and elsewhere (Anderson 1991a) I have commented fiLvorably on the Indian VLS (finally distilled by Walker and Ryan 1990) of ICRISAT and the virtue of such data for "deep social science research, although I am not alone in appreciating such data (Ashenfelter, Deaton, and Solon 1986). At last count, approximately fort-three doctoral dissertations have been based on the seven-year VLS data set from the Indian seni-arid tropics. There was an almost annual battle between ICRISAT management and enthusiasts in favor of preserving the study for "yet another year." Fortunately, the enthusiasts won for at least these six or seven years, and the world is a richer place for it It is not easy to see how agencies other than those well supported through the CGIAR can sustain such research. I do not mean to say that IARCs have a monopoly on such studies (for one idiosyncratic set, see Raj and Baltagi 1992), but the others tend to be run more as snapshots at several points in time rather than the detailed, within- season perctive that is genemed through ICRISAT's studies. 97 Assessing the Consequences of Social Science Research I want to tun briefly to impact assessment. Detemining the impact of agriculual research is demanding, often controversial, and never easy. The problens are perap greatest, however, in dealing with the social science aspects of such work. I have labored on this theme-rather unsuccessfully, I tiink (Anderson 1985)-in the specific ontext of FSR, and am thus sympathetic to the conspicuous omission of any reference to the impact of 30cal science work at CIMMYT in the 1991 CIMMYrAnnual Report, which is deoted to the impact of the center's activities. Measring some features of the process is not too difficult. When social scientists interview farmers in surveys, panels, and so on, they record the results in various forms, such as completed questionnaires, neady tabulated and appropriately analyzed, ideally, and such activities can be documented. Unfortunately, these records can seldom be accessed, and one of the tragedies of latter-C v social science work, in developing countries agricultures in particular, is the lack of centrai repositories of such primary documents and of eased access to what should be a treasure trove for future scholars. The survey managers apparently regard such properties as their private domain, regardless of who sponsored the work and under what conditions. Social scientsts sometimes criticize coleagues in the physical and biological sciences for not always scrupulously retaining their data for subsequent analysis, but it is my observation, that they have much to answer for in this regard. the progress of social science research may be charted in other ways through intermediate products such as reports and journal articles. These product portray the evolution of thinkdng about social science topics and are informative of aspects of the work. They may, however, serve better as mileposts of the careers of the research worlkrs than of the real consequences of the insights distilled. Many social scientists would declare that their major contnbution to the affairs of the world are adjustments to or changes of policies. Here the going may be fairly straightforward, but the reality is confounded by an overlapping and possibly large number of 'players' involved in the complex web of policy formulation. With a few probably notorious exceptions, establishing just which bit of policy analysis proved critical in a policy determination is exceptionally difficult, which may explain why there are few atmpts to quantify such impacts. Within the IARC community, this is especially problematic for IFPRI, with its policy analysis obligations. Social scientists working in more biologically oriented centers in some ways have a simpler task, because they do not need to make a case to anyone, including outside review panels, that they have been unambiguously influential in policy frmulation. It may be easier to assess consequences in the case of helping ronoists formulte recommendations to farmers by taing into account economic considerations along with agronomic data Farm management economists worldng in the centers may thus be able to 98 identify v-ith less ambiguity than others how they have managed to exert their influence and apply their professional skills in a useful manner. Consider, however, the chalenge fiaced by social anthropologists, who are also engaged in sensitizing agricultural research managers and planners about useful directions for teehnologically orientod research. Anderson and Dillon (1968) have advanced a framework for considering the microeconomics of gaining better information about, say, the response of crops to fertlizer involving both agronomic and economic interpretations. This framework operated by adding systematically to new information and evaluating the improvanent in decisionmaking that comes from appropriate use of such information. In principle, such a framework could be used for ex-ante evaluation of the economics of further investigation in such response processes managed by farmers. The practical difficulties of implementing an approach of such statistical formality, however, have dissuaded all but the most dogged investigators. Perhaps there is scope to develop more pragmatic methods for dealing with this class of decision problems. A step in this direction has been taken at CIMMYT (Jauregui and Sain 1992), but much remains to be done. Consider, for one concrete example, how anyone is ever to make real sense of FAO's 25,000 (or is it 25 million?) fertilizer trials. I judge the state of the art of assessing and evaluating social science research to be still primitive and the incentives for developing it to be weak. If we fail to do it, who else will? Do it we must if we are not to be found arrogant and putting ourselves above the assessments we would make of others. Consider (a) the already-noted policy work and (b) the equally nebulous building of NARS capacity as two important sets of products that are difficult to assess. The difficulties are probably just as great in assessing another group of activities in (c) priority-setting and sategic planning, and (d) asing the impact of impact assessment has obvious formidability. Perhaps the most easily measured products are those that (e) are linked to the spread of usefully improved agricultural technologies and, more particularly, (f) represent novel methods in our profession (erroneously called 'new methodologies'). Finally, publications (g) are the preserved record of thinlkng in all these productive enterpises. I have ranked our products in more or less descending difficulty of measurement and assessmenL Any overall evaluation of social science output must also grapple with portfolio balance across this range of products and activities. I am unaware of an attempt to develop a framework that encompasses such a range and balance, and that is amenable to quanttaive measurement and comparative analysis. Until someone does what is needed, I fear our flank is highly exposed to those who may not have our professional best intrests at heart In so admonishing (but I trust encouraging) our own efforts in this regard, I am mindful that some usefiu stas have been made (e.g., Norton and Schuh 1981, Norton 1987, Lindner 1987), but there is a long way to go. 99 Closing Prejudices To cLy my position on some of the issues under diussion, I wil close withi some cryptic questions and tentative answers that we might return to m our deiatio. Has the morpoaion of social science in LARC programs boe worthy? I hiink the answer must be an unequivocal wyes. 0 Has the social science thrust been geneally worthwhile? And sufficient? Here tiings are not so clear, but I would say the review is probably a niixed one. Some efforts, regrdless of the size of the prgram, have boe very effective; others having been less fortunate. The question of sufficiency is a good one for research policy analysts to ponder, and I suspect the answer may be uno.' This seems to be the implicit conclusion from the most recent TAC priorities efforts (e.g., TACICGLAR 1992), although the logic for it is not fully evident to me. I would like to see attempts made to clarify the compaive advantage of IARCs vis-a-vis other agencies for particular research programs in the social sciences. Finally, has the social science agenda been dominated by economists? I am too prejudiced to offer a balanced opinion on this question, which some of my friends in sociology find to be a vexing one. Given the scace resources that have been available, I think the emphasis on economics has probably been appropriate in most of the centers, although the smal sociological effort in IFPRI is one that perhaps should exercise the attention of those charged with balancing the portfolio of that center. Doubtless many better questions than I have mustered will challege us duing the next few days, and I look forward pleasmuably to what is far too rere an opportuity for dialogue among the international agricultual research social science diaspora. Acknowledgements I have benefited through comments and discussions on drafts from several olleagues including Harold Alderman, Derek Byedee, Michael Collinson, John Dillon, Sume Gnaegy, Peter Hazell, Brian Hadaker, Alex McClla, Per PinstqPAndersen, and Donald Wikelnann. None of these friends, and especally none of those who have politely suggested that I should refrain from drftng papers on a dictaphone wile diving, is respnsible for the personal views exnessed here, and the same goes for the World Bank, its management, members and affiliates. Referens Andermn, L. R. 1978. 'An Analysis of Publication Perfbrmance in a Univsty Dqetmt of Agrcal Economics." Revew of Mareng and Ag4icuruW Economics 46(2):150-3. 100 Andes, J. R. 1985. 'Assesing the Impact of Farming Sysm Researdh: Framework and Problems.' Agricul l Aministaon 20(4):225-35. Ander, J. R 1991a. Review of T. S. Walk and J. G. Ryan, Viflage Household EFconomies in hIia's Semi-ard Tropics, Johns Hopkins University Press, Balfimor. Agcula ystems 36(4):495-6. Andern, J. R. 1991b. 'FSRE Impact Inqustion: Investor Issues. Jounal of dt Asian Fanming Systems Associataon l(l):55-68. Anderon, J. R. 1992. Research Priority Setting in Agricultue: Problems in Ex Ante Analysis. In D. Iee, S. Kearl, and N. Uphoff, eds., Assesing the bWIact Qflbn aonal Agrtblwd Researc for Sustinabk Developme. Cornell Ineational Insit for Food, Agrilure, and Development Workhop Proceedings. Idhca, N.Y.: Cornell Uniesty Press. Andermon, J. R., and J. L. Dillon. 1968. "Economic siderations, in Response Research.' Amercan Jourl of Agruhural Economics 50(1):130-42. Anderson, J. R., R. W. Herdt, and G. M. Scobie. 1988. Scence and Food: The CGMAR and Its Pannen. Washington, D.C.: World Bank- Andeson, L. R, R. W. Herdt, G. M. Scobie, C. E. Pray, and H. E. Jahnke. 1987. Itraona Agdaural Raearh Cen.m A Study of Ahevements and Potil. Eonomics Bullei No. 32, Department of Agicuual Economics and Busine Mam , Uniesty of New England, Armidale., N.S.W., Australia. (on microfiche) Andeso, J. R., and K. A. Parton. 1983. "Techniques for Guiding the Allocation of Resouc Among Rural Research Projects: The State of the Art." Promethu 1(1): 180-201. Ashby, J. A. 1991. 'Adopters and Adapters: The Parficipaion of Farmers in On-farm Reseach." In R. Tripp, ed., Plnned Change in Fwning Systems: Progress in On-farm Researh Cbkihester Wiley. Ashenfter, 0., A. Deaton, and G. Solon. 1986. Does It Make Sense? Collectng Panl Data in Delping C trie World Bank, Living Standards Me ment Study Working Paper No. 23. Washington, D.C. Ben, L. P. 1992. Economic Crieria for EFeablishing Plant Breding Pgrnms. Economics Woring Paper 92-01. CDIMYT, Lisboa, Mexico, D.F. Cernea, M. BE, ed. 1991. Puttig People Fbrst: Sodologcal Variabes in Rua Devwopmffe, 2nd ed. New Yorkc Oxford University Press for the World Bank. 101 CrJM Economics Program. 1988. From Agronomic Data to Fanner Riecommndation: An Econmis Tining Maual, completey revised ed. CM/T, Lisboa, Mexico, D.F. Collinson, M. 1993. 'Reflections on the CIMMYT On-farm Research Progm in Eastem and Southern Africa.- In P. Heisey and S. Waddington, eds., Predngs of a Networlehp on Impacts of On-farn Research in Eastern and Souther Africa. Harare, Zimbabwe, 23-26 June. Lilongwe and Harare: CIMMYT. Conteras, M. R., D. L. Gait, S. C. Muchena, K. M. Nor, F. B. Peain, and M. S. Rodriguez. 1977. An Interdisdplnary Approach to Intematonal Agdtural Training: The Cornell-CIMMYT Grduate Student Tewn Repon. Cornl tonal Agrictur Mimeograph 59, Cornell University, Itaca, N.Y. Cummings Jr., R. W. 1992. "CGL4R Agricultunal Research Priorities." Draf, Science & Technology, U.S. Agency for Intanal Development, Washington, D.C. Davis, J. S., P. A. Oram, and J. G. Ryan. 1987. Assessent of Agricultural Research Pror&ies: An Internationdl Perspectve. ACIAR Monograph No. 4. Austrlian Centr for Internaional Agricultual ResearchLnternational Food Policy Rearch Institte, Cabera. Ehui, S. K., and D. S. C. Spencer. 1990. Indicesfor Measuing dh S5utainabilty and Economic Wabitqy of Fanning Systns. Resource and Crop Managenent Program Reseach Monograph No. 3. lITA, Ibadan, Nigeia. Eardin, L. S. 1981. 'Emerging Roles of Agricultural Economists Worldng in Innational Research Institutions such as ERI and CIMMYT." In G. Johnson and A. Maumder, ds., Rural Change: The Chaenge for Agnculural Fcononts. Famborough, U.K.: terational Association of Agicultual Economists. Hazell, P. B. R., and C. Ramasamy. 1991. The Green Revolution Reco,widere& The Impnat of High-yielding Rice Varites in South India. Baltimore: Johns Hopkdns Univsty Press. Jaurgui, M. A., and G. E. Sain. 1992. Continous Economi Anaysis of Crop Response to Fertlizer in On-fann Research. Economics Paper No. 3. CIMMYT, isboa, Mexico, D.F. riggis, J. 1986. Gender-reated Imacts and the Work of the Intenwtonal Agricural Resarch Centers. CGIAR Study Paper No. 17. World Bank, Washington, D.C. Lindner, B. 1987. "Toward a Framework for Evaluating Agrictural Economics Research.' Aauraan Journa of Agricuur Economics 31(2):95-111. Norton, G. W. 1987. Evaluating Social-science Research in Agriculte." In W. B. Sundquist, ed., Evaluating Agricla Research and Prucivity. Miscellaneous 102 Publication 52-1987. NMnnesota Agricultural Experiment Station, University of Minnesota, Minneapolis. Norton, G. W., P. G. Pardey, and J. M. Alston. 1993. Sciece Under Scarciy: Theory and Practices for Agricullural Research Evaluation and Priority Seling. Itaa, N.Y.: Comell University Press, in preparation. Norton, G. W., and G. E. Schuh. 1981. 'Evaluating Returns to Social-science Research: Issues and Possible Methods." In W. L. Fishel, A. A. Paulsen, and W. B. Sundquist, eds., Evalaion of Agriculturol Research. Miscellaneous Publication 8-1981. Minnesota Agricultural Expeiment Station, University of Minnesota, Minneapolis. Olszewski, D. I., ed. 1992. Elmse: Newsktter of the Anthropology Facuty a-id Sudents, 9(2), University of Georgia, Athens. Perrin, R. K., E. R. Moscardi, D. L. Winkelmann, and J. R. Anderson. 1976. From Agronomic Data to Fanmer Recommendaions: An Economi Traiung Manual. CMYT Information Bulletin 27, Lisboa, Mexico, D.F. Poats, S. V. 1991. The Role of Gender in Agricultural Development. Issues in Agriculture No. 3. CGIAR Secretariat, World Bank, Washington, D.C. Raj, B., and B. H. Baltagi, eds. 1992. Panel Data Analysis. Heidelberg: Physics-Verlag. Roe, T. L., and P. G. Pardey. 1991. "Economic Policy and Investnent in Rural Public Goods: A Political Economy Perspective." In P. G. Pardey, J. Rosenboom, and J. R. Anderson, eds., Agriulural Research Policy: Interional Quantitative Perspeaiwvs. Cambridge: Cambridge University Press. Summers, L. H. 1992. Investing in All the People. Working Paper Series 905. Office of the Vice President for Development Economics, World Bank, Washington, D.C. TAC/CGIAR. 1992. 'Review of CGIAR Priorities and Strategies," Parts I and II, April 6 and 15, Rome: TAC Secretaiat, memo. Thorpe, P., and P. G. Pardey. 1990. -The Generation and Transfer of Agricultural Knowledge: A Bibliometric Study of a Researh Network. Journal of Infomnation Science 16(3):183-94. Tnipp, R. 1991. 'The Farming Systems Research Movement and On-farm Reseach.' in R. Tripp, ed., Plannd Chnge in Farming Systems: Progress in On-farn Resarch. Chichester. Wiley. 103 Tripp, R. 1991b. bThe Limitatons of On-farm Rsearch. In R. Tripp, ad., Plnned Change in Farming Systems: Progrss in On-farm Research. Chidiester Wiley. van Herpen, D., and J. A. Ashby, eds. 1991. Gender Analysis in Agrickuual Research. CIAT Publicaton No. 204, Cali, Colombia. Waler, T. S., and J. G. Ryan. 1990. Vilage and Household Economies in India's Senm- arid Thopics. Balimorc: Johns Hopiins University Press. 104 Integrating Research Planning, Priority-setling, and Input Evaluation Within the CGIAR John Lynwn, Rockefeler Foundaton In the nm up to the new millennium, interaional agtral rch as embodied in fte CGIAR faces the challenge of addressing an expanding set of objectives and research areas at the same time that overall funding is plateauing, if not acually declining. Natual resource management, trpical fiestry, bsuinble agriculture, and biotechnology have adde significantly to the researh agenda of the CGIAR sysm. These nww research ares will requ significant budgety resources dunng a period when aid budgets are being cut and demands for those aid funds are expanding to such areas as Easten Europe and envronmental concems. During periods of slow economic growth, developing countries' priorities shift to short-term development rirts and debt service, reducing the avaiability of bilatera fimds for investment in agriculUra research. These changes put greater pressure on the CGIAR system to justify continued investment in agricultural research, respond to shifting donor concerns, ensure the high permance of all programs in the system, and allocate limited rsarch funds effectively across the system. These are large hallenges that require a capacity to set priorities in relation to well-defined objectives and match limited budgetary resources to those priorities. Ihis paper addresses some institdonal and methodological issues in how that is done. Priionty-setting and inpact assemt at the level of IARCs and individual research progams are less than systematic and only episodic in application. During the past few years TAC has instituted its own printy-setting framework for the system, which was applied in the last review of CGLAR priorities (TAC 1992, McCalla and Ryan 1992). Imporantly, this initiative had limited links to IARC priority-setting activities. Moreover, the CGIAR periodically commissions impact assessments of the system (Anderson, Herdt, and Scobie 1988), but they the process can be impwved substantially, but that inprovement will require better integraion of research planning actvities across the centers and betwee the centers and TAC-a situation that some will dismiss as impracdtcal from the start but that neverheless better serves the system in a perod of financial stzingency. A larga investment, at least initially, in such a research planning and evaluation capacity will be motivated by the achivement of improved efficiencies in basic data collection, 7a vews expressed in this pper ar the author's own ad do not represent a position of the Rockefeller Poundion. 105 methodological integration between ex-ante priority-setting and ex-post impact assessment, more effective use of limited analytical capacity, and, most important, better deployment of limited research resources. Social scientists' time and skils are at the heart of such an enteprise. If more research time is to be devoted to resource allocation and impact assessment, the question arises of how seriously donors, TAC, and IARC management incorporate the results of these analyses in resource allocation decisions. The system is composed of relatively autonomous institutes, and the political economy of the CGIAR may demand that a little smoke and mirrors be applied in the resource allocation process. Research is an uncertain, long-term activity, and resource allocation models embody all those uncertainties while giving the impression of mathemalical exactitude. A large amount of expert, experienced judgment, one might say wisdom, is required in the allocation of resources within the CGIAR system. Moreover, tanslating research priorities into effective, productive research institutes is not a straightforward process. The CGLAR has taken on an immense task of sustainably increasing agricultural productivity in developing countries, and resources for the job are limited. The issue remains whether social scientists can bring more information, in a more systematic way, to bear on research planning, priority-setting, resource allocation, and impact assessment. The Research Investment Problem The ganization, staffing, and stategy of agricultual research are govemed by its problem orientation rather than its pursuit of pure science. The objective is not just new knowledge but the use of that knowledge in the production of new technologies that contribute to vanous societal goals. This goal definition lays out the benefit structure in the evaluation process, and, unlike pnvate research, bases resource investment on criteria other thn profitability. Moreover, because agricultural research must compete with other claimants for public funds, this priority-setting exercise is needed to justify those claims; the demonstration that current funds are being used as effectively as possible to achieve future benefits is probably a better argument than demonstrating the returns to past investments, although both e needed. Additionally, in the international arena the potential research problems are numerous. The problem for the CGIAR, therefore, becomes one of how to invest limited funds in a well-defined set of research areas evaluated by their contribution to societal goals. The decisionmaking process in priorty-setting and resource allocation is usually framed as an investment problem; in that sense it can be linked to what the World Bank refers to as the wproject cycle." A portfolio of projects is selected from all those for which ex-ante studies exist, projects are monitored and evaluated periodically during the project term for minor redesign or early termination, and projects are evaluated at completion, especally for rate of return on the investment. The project cycle makes an important point, which is that the research investment or portfolio problem is continuous, iterative, and management- and information-intensive. Ex-ante assessment, monitoring, and evaluation 106 dunng tie project and ex-post evaluation are all necessry components of portfolio magemot and should build on each other in the prcject cycle. Develping a capacity for continuous assessnent of research priorities and esource allocation wnthin the CGIAR is a complex matter owing to several charateistics of the CGLAL * First, the system has a multiple-objective function, and the number of objectves has expnded over time. How to specify all these objectives in income terms and then waght alternative objectives is a question that has been addressed by relatively ad hoc approaches. In essence the question is whether the CGIAR can correctly specfy a social welfare function, determine the arguments that enter into it, and speify the weights on those arguments. * Second, the system must not only consder current research activities included in the CGLAR portfolio but pernodically assess principal commodities and ecologies that lie outside its current purview. The CGIAR must consider all ential research activities that will best contnbute to its objectives. * Third and most important, the priority-setting and resource allocation process must take place at various levels; most of it is currently done independently and without integration. At the level of the CGIAR system, priorities are set across commodites and other research activities by the TAC Seriat There is little integation with prioty-setting within IARCs or within commodity programs, which is done at different levels of intensity, if at all, across LARCs. The research porffolio within the CGIAR is now managed by independent efforts across levels: TAC's priority-setting exercise at the system level, priority-setting by individual LARCs, and priority-setting within commodity and natual resource management progms. Also, there is litfle integration through the technology development cycle, in that ex-ante priority-setting does not feed into the monitoring and evaluation process, and neither feeds into ex-post impact studies. Most important, the relationship between the priority- setting exercises and the budgeting and resource allocation process is weak. The remainder of this paper addresses the extent to which current methodologies and data base systems would allow such integration, the institutional adjustments required, and the gains that might be exected. Integd ating Goals and the Research Agenda Within a Basic Model The argument is that an integrated pnonty-setting framework would ailow consistency across decision levels, achieve efficiencies in portfolio management and evaluation, and prvide a trans t and contnuous assessment of past and future rens to donor 107 investmet in the CGCAR. The develomet of such an integated stem invols four principal desizn caitemia: * The fiamework shod allow intmation trough time, that is, provide consistency betwen ex-ante resarch planning and ex-post impact asessment * There is a pronounced spatial dimension in agricultural research planing, made more explcit by the coregional fiamework (McCa_oa 1991), which should be clarified and integaed into the analysis. * Integration across decision levels should be possible wthin the frmework, at the reserch program, IARC, and CGIAR sytem ledL * The framework should be robust enough to address the Varous roles m portfolio manment tchnology desg and ageting, research program planning, resource allocaion, and donor justification. The fou design critia lead to four s8tuctual or o tional features in the development of an integrated priority-setting fiamework: * A consstent and robust methodology is needed to intga the four criteria. * Tbis methodology, in turn, needs to be suported by sndaized data bases tat are updated periodically. * The level of disggtion in the analysis should be set by the needs of the lowest oranizational level, that is, the indiidual commodity and resource manent program. * A mechanism is needed for intinst codinati. Choice of Ansltal Methodology The simple stement of the ex-ante problem is how to allocate limted budgetary resurces across reserh activities to best attain the socetal goals being addressed by the CGIAR. The quantitative assessment that measues each research actvity's contibution to the vanous goals is drmined by the choice of analytical methodology. The thee demands that detmine this choice of methodology are that it is consstent across exante and ex-post applications, that it must be able to accommodate the multiple objectves of the CGIAR, and that it is robust enough to address tenology design questions. Achieving some constenc in portfolio managemt requires a m e that allows consincy between ex-ante priority-setting, monitoring and evaluaton, and ex-post impact assessment (see Anderson and Pardey 1991 for a sifiction and review of 108 altente priority-setting approaches). Ex-post studies for individua c odities virally always employ a cost-benet/oomic surplus approah to impat asnment (ceverria 1990); coometic, production function approaches are the other method. Ex-ante priority- setting using economic surplus methods reMs on the assessment of benefit inheret in ex-ante impat assessment. ing and evaluation and ex-post asessment can be see as a continuous refinemt of the parameters in such an ex-ante model. At any point in time, an integrated systm based on economic surplus methods should be able to give the decisionmaler an esmate of ex-post benefits over the life of the insttution and the best possible esfimate of fubre impacL For donos this system provides not only te return on investment to date but a vision of fute impact, with the important notion that the resources are beng spet in as efficient a manner as possible. Other methodologes, such as scoring, do not allow such anessment; in them the ex-ante and ex-post phases are completely indqedent cise, with little capacty for continuous assesment. Prioity-setting based on economic surplus methods provides a raning across research altenativ on the bais of efficiency critria, that is, the increase in net benefits to all producers and consumers. This is a clear goal of the CGIAR-inceasing agWicultural produon at a delng cost to soety - but it is not the only goal. Other gois includ equity cs to(Le., how the net benefits are distriuted to disvataged grups) and envionmental and sustainabi iecives. Two appraches exist to how these other goals might be inprd in pority-setting. First, indendent masmres for the different goals ar developed, for example, the number or percentage of small fame growig a particular crop as a proxy for income distribution to small-scale farme In developing the ranking, te mea_su are used as modifiers to the net benefit estimate (Cessy et al. 1989). Tihe alternative is to measure each goal in net benefit terms thrugh modifications to the basic economic surpius methodology; altrnative sategies are then evaluated in terms of the weights givwn to the various goals. The latter approach allows one to ask the question of what it may cost, or may not cost, in pure efficiency terms to pursue equity and ssainability goas. It also maintains symmetry between e-ante and ex-post approaches. Mdeasurment of Equty Impact in ria Research Equity considraions have a range of dimensions, depending on the group to which the benefits should be tageted and the mechanism used to perform such targeting. These dimensions are presented in Table 2 in order of their analytical complexity. The distnbution of income between producers and conumers is an output of the standard economic surplus model. Producers are, however, connected to consumers by a marketing chain, and some complexity is introduced in demng how much of the ewonomic benefit is captured by markeing agents (Freebaim, Davis, and Edwards 1982). The distribution of benefits between consumer income classes and the impact on nutrition require further modifications to the sndard model and were first worked out by Pinstrup-Andersen, Ruiz de Londono, and Hoover (1976). This approach has since been applied in a number of er-ante and ex-post analyses. The methodology depends on the abiity to estmate budget shares and pice 109 Table 2. Analyng Equity Objeci in Agiculftual Research Plannig Bnftconsumers or producersa Choice of crop, eswy eiucity difi;aenfaLy ~~~~~~and can be txadedl Benefit (mcome and nutrition) low-income a. Choice of crop, weaght in diet, and consumers, especially urban demand elasticity b. Market segmentation based on quality c. Imprving nutritional value of crop through breeding or prossing Imprve income of low-income or small- a. Choice of crop; which farmers scale producers, especially in marginal grow the crop and where areas b. Targeting technology by scale, market access, or agroecological zone Benefit landle or rural labor a. Choice of crop depending on factor intensity and use of hired labor b. Shift in factor intensity within crop Improve women's welfare differenfaly a. Not yet well specified sicities by income group, which means having access to consumer expenditure and budget The complexity increases when the analyst attempts to estimate benefits across farm size or producer income strata. Farm size is generally used as a proxy for farm income, bewse farm size data are avaiable in agicultual censuses, but they need to be modified by such factors as access to igation, greological zone, and market access-the factors that also influence technology choice. The methodology is tatable by esimating a supply curve for each stratum of farm size and then aggegting Pacico, Lynam, and Jones 1986. The difficulty is spedfying the shifts for each strata. Analysis in this dimension requires a large degree of t chnological sicn. Technology development and research planning cr.n be very closely tied to producer equity considerons because of this specificity in the analysis, both withi and between ommodity progms. It is notable, nevertihess, how rarely this specicity is achieved, pardy because of data limitations and pardy because economists have devoted insffiient time to the topic. Finally, Pachico, Lynam, and Jones (1986) integrated 110 distributional issues in both consumer and producer dimensions into the same ex-ante, economic surplus analysis, that is, three equity dimensions can be integrated into the same analysis; however, the evaluation of the marketing chain has not been yet been incorporated into such an integrated analysis. The impact of new technology on landless labor and rural employment is methodologically tractable by working through an economic surplus calculation based on the factor market rather than the output market. Because all the other evaluations are made within the framework of the output market, however, the analysis moves to a general equilibrium approach to circumvent double-counting of benefits. General equilibrium approaches have been used to evaluate the income distributional effects of technical change in northwest India (Quizon, Binswanger, and Gupta 1991) and the wheat sector of Pakistan (Renkow 1991), the latter modeling the effects of differential technical change between favored and marginal areas. These models, however, are very data-intensive and are most feasibly applied at a national level. Tractability in such multimarket models put limits on the degree of spatial disagtion, of the level, for example, that Pardey and Wood (1991) used in analyzing research priorities in Indonesia, and on the degree of technological specificity of type that would be employed in determining a research strategy in a particular commodity program. Such models would best be applied to only certain research allocation questions, for example, the allocation of resources across rice ecologies in Asia. The gender issue is, in part, also a question of how to bias the benefits of technological change differentially to a target group through technology design decisions. Women are constituents of aU the target groups previously mentioned, which makes further specification necessary. All the previous equity dimensions narrow the universe of women that the CGIAR might want to target, but that fact is not sufficient to address the gender issue. The question is whether target groups based on household characteristics are sufficient to target women's income or whether women need to be differentially targeted within these groups. An answer to that question does not exist, but certain points can be made. First, income distribution by gender is principally determined by intrahousehold allocation mechanisms and societal norms on access to income-generating activities. Under such circumstances, there is no certainty that technology can be effectively targeted without institutional modifications. Second, the task is to identify those economic spheres over which women have some control and focus on those. These include subsistence food crop production, on-farm processing, and gender-specific labor activities. These are particularly difficult areas on which to target income streams from new technology. Moreover, the technological specificaion problem increases dramatically. For example, in maize those areas where subsistence production is controlled by women might be targeted, but where are those areas, what are the agrocimatic and farming system characteristics of those areas, and what are the technological requirements? In sum, substantial work is needed on how to target technological benefits to women. Detrminig how to incorporate this information into a priority-setting framework is even further down the road. 111 Priority-seftft and the Su bit Objctve ncorporating agdcultmal sustainabilit and environmental concens into the reserch agenda of tie CGIAR systen has led to its most dramatic restuctuing since the creation of the system. The susainability banner has been used to justify a significant expansion in the number of IARCs in the systen, redeployment of budgetary resrces out of commodity research and into natural resource management research, and an nitonal innovation defined as the 'ecoregional approach (TAC 1992). This stucta change has not yet been incorporated into a consistenty integrated framework for management of the complete research portfolio at either the IARC or system level. Analytically, sustainability probably should not enter as another argument in the CGIAR's objective function but rather should entail other modifications of the pionty-setting system that has been based on allocation of resources across commodity research programs. Four principal issues are involved in such an integration: * Defining clear targets for natural resource management programs that can be tranated into net benefit streams 3 Developing measureent and evaluation methods for the more complex benefit and cost streams arising from natural resource man ent research * Defining the terms on which the net benefit stream for naural resource management progams should be compared to that for commodity programs * Making priority-setting across research progams and across ecoregions-which are obviously overlapping-consistent Natural resource manent has been incorporated into the CGIAR system through addition of LARCs involved in water management, fish resources, forestry resouces, and agroforestry, and the development in existing IARCs of natural resource management programs, usually defined in terms of a ecoregion. The conceptual problem that this introduces in developnng a priority-setting framework is how to define discrete research programs in which inveshent will yield a definable net benefit stream. Fisheries fit within the standard commodity progam classificaon in that research targets can be identified across definable production systems and benefits can be measured in terms of increased fish production and productivity. For agroforestry research at ICRAF, for example, or research within CIAT's hillside, forest margin, and savannah programs, the target area, benefit stream, or both are difficult to atibute uniquely to a research line or program. Apart fron paddy land, there are no limits on where agroforestry can be appLied in its various system manifestations; however, the question remains where expected impact will occur as a result of the research investment To this must be added the multidimensional nature of the benefit stream. In agroforestry the resource dimension (soil erosion control, soil organic matter, and nitrogen fixation) is probably just as important as the output dimension (poles, fuel, forage, and fruit). The quesdon is, which benefits will be expected to ocur where and in relation to 112 which research invesmbent? The problen is magnified when the research program encompasses hillide agnculture or the forest margins. The potental number of baeefit streams is large and difficult to attribute to the research investment. It may be argued that such is the natre of systems research. Nevertheless, commodity resrch has boen successful partly because altemative research lines can be matched and evaluated against research targets; a structe is critically needed for defining targets within natual resource management research. Setting targets would also highlight what has been left out of the research portfolio. An example is soil management research. Because agriculture is impossible without soil, virtually all research progrms will have an element of soil management. Yet the CGLAR system cannot be said to have a clearly specified research stategy for management of developing countries' soils resources, other than that such issues should be dealt with within the ecoregional framework ther than at a global level. D-etrining the net benefit stream for natural resource management programs introduces several complexities in measurement and evaluation of benefits and costs. The problem here is that unlike commodity research, benefits are not easily expressed through output markets and there is a divergence between social and pnvate costs and benefits in much of natural resource management research. In exploring these issues, it is useful to distinguish between natural resource management practiced at the intensive and the extensive margin (Lynam 1992). Research on naural resource management at the intensive margin seeks to maintain the quality of soil and water resources within existing agricultual systems under increasing pressure from rising popuion, increasing commcalization, or rising output prices. The resarch focus here is on soil fertility and management-managing the crop, livestock, and agrofoy components to achieve enhanced soil quality; reducing agrochemicals in the water resources; preventing salinization of irrgated land; and using water resources in agricultural production efficiently. The analytical issue is how to evaluate the costs and benefits of this type of research. M3icrolevel benefits arising from soil erosion control, reduced salinization through improved drainage, or improved water management are usually evaluated in terms of changes in crop output. At a macro, piority-setting level, however, these benefits would be spread over a range of commodities, with no empirical basis for assessing the degree to which each of the supply curves would be shifted. This creates a problem in attributing benefits to the research investment. Alternatively, the benefit evaluation could be made in terms of changes in the stock of soil and water caital. Adoption of improved soil management practices in a deep and fertle soil will probably result in very little change in crop productivity for some time (and with techniques such as alley cropping, could result in a decline) but could reduce detrioration in the inherent stock of soil capital. The science to measure such stock changes and then to value the resource stock is not fully in place. Even more difficult measurement problems are on the cost side. Much of the justification for research in many of the areas cited arises from eernalities and the 113 associated social costs. Thbe social costs ansing from soi erosion, agmical use, or overgrazing are extremely difficult and costly to measure, difficult to value, and often very site specific; for example, they depend on downstream water control strucus or the pesticide-binding capacty of soils. The cost and sophistication of IRRI's (RRI 1992) and CIP's work in evaluating the costs and benefits of pesticide use should be highlighted in this regard. incorporating such sustainability esearch into priority-seffing and impact assessment will require traig effects though environmental or physical systems as well as economic systems. Cost-and-benefit assessment will have to be based on detailed microlevel studies, which will have to be aggegated within an appropriate frmnework to evaluate market impact. This has parTicular implications for data collection and maintenance. Natural resource management at the extensive margin involves the encroachment of agiculture into natual ecosystems, many of which are marginal for agncultural purposes. Managing ecologies such as tropical rainfowests, wedands, savannahs, and arid shrubland involves a continuum from conservation, through benign exploitation that does not change the structre of the ecosystem, to conversion of the ecology to agricultual or livestock systems. There is an implicit assumption that the agricultural frontier needs to be stabilized, and research and technology development have a role to play in that stabilizadon. Evaluation of such research rests on how a natual ecosystem is valued, in essence, the stock of natumal resource capital. In order of complexity, this valuation is based on sustinable yields in the action of economic products such as timber and fish, the value of unexploited genetic resources whose true value will be apparent only in the future, and the ecosysten's contribution to the regional water cycle and the global carbon dioxide budget. Research should concentrate on the natural ecosystems whose societal value is greater than their value in their next best use, usually agriculture or commercial forestry. The gain to research is the difference between the two values. Obviously, the CGIAR is not able to make such a determintion, which creates something of a lacuna. There is not yet a clear frmework withn which to evaluate such research and to monitor its progress other than its contribution to increased knowledge. (Pachico and Henry provide some suggestions in their contribution to this conference.) Another issue revolves around how net benefit streams from commodity and natural resource management programs should enter the objective function. Commodity research will be evaluated primarily within an economic surplus framework; natural resource management research will be evaluated in a social cost-benefit accounting framework that measures primary net benefits under constant prices but generally will not measure the Secondary benefits resulting from market effects. The latter will be based on microlevel case studies aggrgted witiin a spatial sampling frame (Graham-Tomasi 1991). Quantification of equity dimensions will be possible only to the extent that the relevant socioeconomic information on distribution of target groups can be mapped and benefit distribution can be defined in the microlevel case studies. The question is whether the addition of secondary market effects in the case of commodity research and social net benefits in the case of natural resource maagement research significantly biases the esimate of the net benefit stream of one group compared with the other. in this regard, the other issue is the choice of discount 114 rate on th net benefit stream. he poblem how to compare futue investment in a commoditq research proam that is already well established and in which any additional investent is expected to yield high retuns with a newer natural resource management program whose epected benefits wil not start untl far into the future and whose proWbability of success will be lower. (Pachico and Henry hihght this isse iD the case of CIAT's investment in natual resource man t elsewhe in this publcation.) The compaive advantage of the CGIAR is in investing in such programs. The question here is, what is the difference in discount rate that is acceptable in deciding between investnent in commodity and natural resource management programs? (Ihere are cen thrnier theoretical problems m selecting a discount rate for investments in natual resource management research, which are briefly reviewed in Graham-Tomasi 1991.) A final issue involves the compatibility between piiority-setting at the ecoregional and the global level. The problem is one of nesting within the resource allocation process, that is, are resources first allocated across ecoregions and then allocated across research programs within the ecoregion, or are resources first allocated across global commodity and naural resource management programs and then allocated across ecoregions within each commodity program envelope? Resource allocation crentdy tends to follow the latter procedure, but will this structure inhibit the development of ecoreional programs? The results wil differ not so much as a result of the analytics of the problem but as a restdt of organizational imperatives, because at each allocation point, deciions are made by different oganizations witi different research mandates. The TAC priorities document (TAC 1992) attempts to resolve this issue by distnging between strategic research where prioriies are set at a global level and applied research where priorities are set at an ecoregional level and most natal resource management resch is viewed to be of the applied type. The problem is that research is not easily divided either conceptually or organiztilly along such stategic or applied lines; most commodity research programs have major elements of applied and adaptive reserch, and many would argue that such elements are critical to mamntaing the relevance of the research progam. This issue highlights some of the 'political economy" problems at the levd of IARC decisionmaling in moving the ecoregional initiative tD operational stakus, for example: * Is it possible to make priority-setting witiin the CGIAR more 'bottom up' in approach? * How high will be the costs of co ion across IARC activities in implementation of the ecoregional concept? * Who resolves a situation in which an ecoregional commodity priority is not consitent with the regional priorities of the global commodity center? These questions lead directy into the next set of issues on disaggon and division of labor. 115 Level of Disaggregation and Center Division of Labor Th practicality of devdoping an integmted research pinning and priority-setting system rests on agment that such a systm s worable, usefu, and answers pertnent questions at various levels of research ement and on ollabotion in methodology and data base development. If the system is to be truly intgated, its deveoment should start at the bottom and build up; each stage in the develpment should stand on its own merits. The first stage star at the level of individual research programs and asks whether target area strtificion, technology tageting, and echnology design decisions can be better specfied, analyzed, and inWgated into research planning. Decisionmaling at this level detemines the level of disagg on required in the analytical system. Germplasm rserch in the mainfed sector (the major cost in CGLAR germplasm development), naturl resource ma nt, and production system reserch all require a significant dere of speification in eng that technology will be adapted to variation in the c and oeconomic envronment. This specification is termed atargeting. and is accomplished by decisions withn the research program. Thus, even if equity impact across the developing world is defined principally by across-progam consderations, e accomplishment of those objectives must be tanslated into intrnal program decisions. The CGIAR objectives are achieved by a close intgration at the two levels of decisionmaking. At the program level, goals are mathed to tehnology design decisions through the tageting process (Lynam 1989), which in turn becomes the bas on which priority-setting and impact analysis at the LARC and CGLAR system level are built Many resource allocation decisions at the CGIAR level are defined by within-program dcisionmaking. The division of resources among statec (imcluding biotechnology), appied, and adaptive research is determined essentially by an assessment of research needs within the commodity or natural resource angement progam. The spatial distribution of echnological benefits is only crdy defined by allocation decisions at the level of the CGIAR (although the ecoregial concept may change tht in the future) but is made specific by decsions within the rmsrch programs. The dirision between germplasm and production system research, the disciplinary mix in the system, and the deployment of training resources are all defined within the commodity and natral resource nm a t pams. TAC in essence provides the financial ceilings within which those decisions are made. Nevertheless, TAC tries to allocate CGIAR resources in all those dimensions and yet independent of the decisionmaking process at the research program level. The effectiveness of resource allotion in achieving stem goals is only as good as the resource allocation and planning process at the lowest level, that is, witiin the inidual commodity and natual resource management programs. It is this level that should be made more transent and then entered into priority-setting at higher levels. The dimensions in which technologies are specified include agmatic, biological (primarily pests and diseases), and sooenomic factors. Variations in aU of these dimensions are spatially defined, which allows a reserch progam's target area to be stratified, decisions to be made on how many separate rsearch prqects are required to meet 116 the technologicl needs of the target area, and piorities to be established on how many of tho research projects can be inopoae give dte size and resources of the progam. This stratification is clearest in tpecificati of gene poo, crossing and plant selection Pograms. Agcmaic fac have been used to define the number of en on which to trget separat breeding programs-Carter (1986) has done this for cassava at C1AT and Pollak and Corbett (1992) for mai at Cmm. Experience with such spatial tageting at a global or continental level has shown that the zonation is crop specific, the statificaion is patted at a regional and even a country level, but that such resolution is not inconistent with the establishment of priorities at a global level Tbis ion of crop ennrnments has shown that FAO's agroecological zones (AEZs), on whch TAC bases its definition of ecoregional zones, are insufficient to base within-commodity program priorities. The same conclusion will probably apply to global natural reource mat research. FAO's AEZs will be most useful in organizing inter-IARC collaboration and effcng tonaliztion of IARC-NARS program linkages. Resarch program priorities can be determined within an AEZ or aggregated up to the AEZ, but they have little utility in discriminag reserch needs; there is too much variation in ftothat influence crop and livestock productivity and choice of managemet systems within them. A socioeconomic dimension can be added to the statification and spatial targeting within a research program's target area. This was done in the development of the spatial swnpling frame for llTA's Collaborative Study of Cassava in Africa (COSCA) (Carter and Jones 1989). It is not well developed, however, in priority-setting-not because of analytical difficulty but because of data limitations (see the next section). Neverthdess, it should be emphasized that equity impacts in some aspects are spafally patterned and in other aspects are not, for example, farm size or farm wealth. This fact creates another analytical dimension in the targeting and is worthwhile if technologies can be differenally tgeted by scale or wealth factors. If goals are atained at this level of disaggregation in the analysis and decisionmking processes, the question then arises whether such disag is consistent with the priority-setting process at the higher level across researh programs and AEZs. Pardey and Wood (1991) have developed a priority-setting methodology based on economic surplus concepts that allow flexible agrecological zonation, specificity in within-program research options, and across-program priority-setting. Their approach is based on innovative aggreation procedures and the integration of GIS into the spatial tracng of the aggregation process. Although their methodology has been applied only at the level of NARS, the priority-setting problem is the same as that for the CGIAR, namely, how to allocate a limited budget over a multiplicity of commodities and natural resoure management rsearch progams while accommodating variation in agroclimatic conditions. The Pardey and Wood methodology also contains a framework for estimating spillover effects. Spillovers within international-as opposed to national-research are pardy a function of how the zonation is done and are necessary where the analysis relies on fixed AEZs (Davis, Oram, and Ryan 1987). More effective tageting reduces the probability of research spillovers, but spillovers 117 are important in the analysis when research resoures are not sufficient to cover tie whole target ai The vision of an integrated prority-setting and research evaluation system within hie CGIAR is based on an apprpriate level of disggregation in specifying technology options and a metiodology that allows consistent aggregation to dimensions critical to planning at the IARC or systen level. This framework shifts the onus of ex-ante impact assessment from TAC to IARCs and creates a more appropiate division of labor between leading and coordinaing ithe development of the system and its actual implementation. It also forces all the actors to agree on the methodological framework and to begin to standardize the data on which the analysis is based. Consistent Data Base Development in the CGIAR The methodology and the level of disaggregaton of the analysis have very heavy demands for data. As implied by the preceding discussion, such data for priority-setting complement, rather than substitute for, the emerging data requirements to plan and implement international agricultual research in general. Because data are costly to collect, maintain, and update, the development of the data bases that drive the priority-setting and research evaluation processes should be seen as part of a larger discussion on how to improve the efficiency and uility of the data-collection exercise within IARCs. Developments in data base technology, the declining cost of computer hardware, and the continued refinement of GIS and statistical support software all offer potential to improve the inforation and analysis that can be brought to research planing and evaluation in nterntional agricultural research. Because of the location of IARCs throughout the developing world, their mandates, and their extnsive research networks with NARS, the CCGIAR system has a real comparative advantage in collecting field data at a continental and global level that will improve the undending of agricultural systens, natural resource management, and some enonmental concerns within the trpical developing world. Achieving this will require developing some standards in the collection of basic data, a dision of labor in the collecdon and maintenance of the data, and agreement on exacdy what data are necessary to collect IARC's have always faced the challenge of how to undertake agricultural research at a global or continental scale, when most of agricultural research and technology development is suect to the influence of variation in agroclimatic, biological, and socioeconomic factors- this, in fact, was one of the original arguments for creation of the CGLAR and the location of IARCs in the rpical world. Resrchers must understand how plant and livestock technologies react to such variation and the distribution of those factrs through the crop- or livestock-growing area. Researchers have some undesanding of the first area and only limited understanding of the second. Bringing more and better information to the research enterprise is absolutely necessary for IARCs. 118 The first hee steps in Table 3 are as necssary to the day-to-day research proc as they are to the research planning, priority-setting, and evluaton processs. The point here is at both data collectdon and priority-setting are ftidamntlly fed to more efficient and effecfive planning, targeting, and specification of progam sarch. Diagnosis, characterizaion, and target area satification should be important to all researchers and incorporate their exprse. Table 3. Steps in Research Planning, Priority-setting, and Evaluation Exante I E-post Pohity-setting i Impact Aswwnnt 1. Diagnose constraint and charactri target area. 2. Zone and stratify target area with production and target population weights. 3. Define technology options, productivity incmase, and cost reduction by stratum: a. Delphi methods a. Adoption surveys b. Crop and livestock b. Secondary production statistics c. On-station and on-farm trials _ 4. Define probability of success and 4. Update adoption curve. adoption curve. 5. Evaluate possible spillover effects. 5. Evaluate actual spillover effects. 6. Develop country-speific prices, 6. Update prices, tansport costs, and tansport costs, supply and demand demand elasticities. elasticities. 7. Estimate benefits. 7. Estimate benefits. 8. Analyze in priority-setting 8. Evaluate rate of return and framework. distribution consequences. The preceding arguments define the need for a set of standardized, georeferenCed data bases at a relevant level of disag on or resolution. Aggregte, country-level data are of little value in developing such data bases. The types of data bases or layers that would be used to develop such a data base system are specified in Table 4. Spatial referencing links the individual data bases, but each is constructed independently. Such data base development builds on daa that are already collected for specific purposes, but these data collection 119 efforts can lead to the d opm of genic data bases that can feed into a wide range of potential applicafions. Partcully, they underlie GIS applications, especially target area stIatificatio, crop and other process modn, suey sampling designs (especially spatial design), and multisite trial evaluation. Thus, s n of taget area is an analyfical procdure that can be continually refined and updated from more basic data bases. Such data bases make the data colection effort more cost-effective by idenifying gaps in data, by developing sysmatic sampling frames, and by eiminating the need to resurvey for every specific reseach problem-a significant issue in socioeConomic and farming system surveys. Table 4. International Agriculture Rsearch Data Base Needs : >2~~~~"Oncio Agnc ins>.;CoUo ', '~r'-hw ;., f ;. ~~ .,. ... ;. Melhod: . .. Climate National metorological a. Monthly: CIAT departments b. Daily: under Development Soils, terrain, and National soil surveys a. Soil map 1: 5 million: FAO hydrology b. Soil map, terrain, and pedon Data base: World Soils and Terrain Digital Data base (SOTER), under development c. Pedon data base: Some regional data bases d. Topography: Austalian national university, Africa under development Vegetation Remote sensing a. Country level: nothing systematic Crop and livestock National agricultual a. Cassava, beans: CIAT systems census Crop and livestock Feld surveys a. Cassava: UTA for Africa sysUMs Population National population census a. Total populaion, Africa, Lain America CIAT Socioecowmic Field surveys a. Africa: committee discussion Elhnography Field intrviews a. Africa: discussion with Human Relations Area Files .____________ ._____________ C(Yale University) 120 Targeting of agricul technology, especially in th equity dimensions, depends critically on oooomic and frming system datL Many socioeconomic and farming system variables are spatially defined. In Latin America fann size and farm income are spaftally pattened with small-scale firmers clustered primarily in more marginal agricultural areas. The struc of fming systesn is determined by ardimatic and marketing fiators and is tfierefore also spatially defined. This, in tun, allows the spatial overlay or inWgration of agrodimatic and biological zonation with sociocoonomic zonation, as was done in the development of the COSCA sampling frame (Carter and Jones 1989). For other variables or in other areas, socioeconomic data may be defined in terms of distrbutions across spatially defined populations. For niost rural villages in Africa and Asia, there is a wealth of fann size distribution, but such distributions can be entered easily into the data base, as are temperature or rainfall distributions. Developing such a socioeconomic data base at a continental scale relies on some standardization in the set of vaiables and consistency in the definition. Agricultural and population censuses provide one source of such data, although they are usually collected at different times across countries; however, much of the detail on farming systems, household characteristics, natual resource use, and labor use is obtained from field survey data. These data are usually gathered for a specific purpose, analyzed with that purpose in mind (and often never analyzed), and then discarded. The major part of socioeconomic operational finds witLin the CGIAR is probably spent on such surveys, especially in the farming system area. Yet IARCs have little sysmatic understanding of these data for research planing; for example, what is known about the distibution of plantng dates and growig season for maize across Africa, other than what can be guessed at by delphi methods? It is expensive to get out into the field in Africa or Latin America; few oranizationshave the scope in survey cacity that exists in the CGIAR system to evaluate agricultural and natural resource management sems in the developing world, and CGIAR socal scientists have had difficulty placing dteir detailed microlevel studies into a larger macro context that contributes to a consideration of research alternatives. Seveal questions arise: * Can the limited field survey resources available to IARC social scientists be used more efficiendy and shategically? * Can more standardization be incorpora ed into field surveys? * Can sampling fiames be improved? * Can such valuable data be developed into data bases for a wider range of uses? An underlying argument in this paper is that data bases and GIS capacity provide an organizing and planning framework for agricultal research and provide the base on which the pnority-setting system in Table 3 rests. The other design element in this system is the macro-micro linkge. The diagnosis and chrateization at the macro scale provide the sampling frame or the criteria for selection of microlevel field research sites-for either 121 biologic or social science research. The microlevel research is focused on undsnding p s and quantifying relationdhips. The macroleve system, in tur, provides the ramework for ether aggregting or extapolating the microlevel studia to the lagr target area. Etrapolation may be based on simple intpolation procedures, etmated equations, or more sophisticated simulation models. The results in turn, provide a further itaion in technology planning and targetng. For reasons already dicussed, such a system will be absolutely necessary for organizing, planning, and evaluating natual resource management research. For example, a watershed is often used as the unit for detl field research on sustainable agriculture, for example, the sustainable agriculture ColLiborative Research Support Program (CRSP) (University of Georgia undated). The following anterior research questions often are not asked: * What criteria are used in selecting the watershed? * What is the population from which such sites might be drawn? * How r tative is the site or sites? * On what basis can the research results be extpolated? This ldnd of macro linkage has not existed in most of international agricultural eearch to date, primarily because the capacity to address those questions has not eisted. Such a situation need not, and should not, continue to exist for a research system that is proulgated on serving the developing world. Conclusions An argument has been sketched for an integated research planning and evaluation syste that meets the needs of individual lARCs and the CGIAR as whole and that provides a continuous snapshot of past benefits to investment in the CGIAR and the projected benefits into the futre. Moreover, all the infonnation in the system derives from and feeds into the research process. Most research decisionmraking in IARCs and the CGIAR is based on informed but qualitatie judgment. Becausz cassava mosaic virus is viewed to be more important than cassava bacteial blight in Africa, more resources are devoted to the former. Because maize diseases produce more damage than maize pests, a pathologist rather than an entomologist fonns part of the team. Because inland river ecologies have more potential than upland rice ecologies, a research team should be formed to address the constraints in that ecology. An integrated system makes such decisions more explicit and provides the best jutification for investment m the CGIAR, namely, that given what is currently known, this is the best allocation of research resources at all levels in the system. Such a planning and evaluation system does not yet exist in either individual IARCs or the CGIAR, nor will it exist without collaborative planning in the development of the systen. It may reasonably be asked whether development of such an elaborate planning 122 system is feasble, justifiable in terms of cost, or too inflexible and stfling of scientific initiative. There are swveral counter arguments: * The system is built from the individual research program upward, where the focus is on definition of refined target area, explicit technology speciation and targets, and assessment of potential yidd gains and the probability of success-that is, information that research progams need to plan effectively. * The system atempts to capture real ownomes of scale and cost efficiencs in the development of essential data bases, but this requires a reasonable division of labor across lARCs and agreement on data standards and variable requirements. * For the results to be aggrgated for use by TAC in CGLAR planning, some consistency in methodology is required, and LARCs will have to dedicate some resources to the -. stemwide exercise every five years. Moreover, cost savings can be achiwved by reducing the need for penodic, independent impact assessments. Mechanisms for coordinaton are already being developed for data bases, although the dwison of labor will require resource commitments by some centers (and reconition of thse req urements by TAC). An argument can be made for developing a capacity somewrhere in the system for centralized activities that meet needs across the system, for example, that develop administrative and topographical base maps. Methods for ar on and sttficaion of target areas are being employed in only a few centers and can be refined further. These methods can be extended throug the data base cooriating committee, with the link to economic surplus models reqWring at least part-time commitment of an economist in each of the centers. There is some scope for centraizing the demand elasticity estinates, most logicaly at IFPRI. rdinabng the aggrgtion procedures to the level of the CGIAR system, definng budget or program categories, ensuring consistency in equity estimates, and developing a derminant procedure for prionty-setting and budget allocation should all devolve to the TAC Secretiat or a steering committee chaired by the Secreriat. The time is disaring when LARCs could function as autonomous institutes. The ecoregional framework and activities such as those desribed in this paper challenge the system to be more innovative in defining areas in which intercenter collaboration is necessay and developing institutional arrangements that foster appiate divmsions of labor and resources th work toward agreed objectives. Collaboration should develop naturally from shared goals. If it does not occur, however, donors may decide that further budget stringency is needed to force such rationalization and collaborative link. 123 Acknowldgemet I am gratefil to Robert Herdt and Phil Pardey for constructve comments on an earlier drft. References Anderson, J. R., R. W. Herdt, and G. M. Scobie. 1988. cince and Food. The CGL4R and Its Paanecrs. Washington, D.C.: World Bank Andrsn, J. R., and P. G. Pardey. 1991. "Notes on Agricultural Research Evaluation and Priority Setting Methods." Paper presented at World Bank Agricultual Economics Program Workshop, Comnell University, Ithca, N.Y., April 28-May 3. Carter, S. 1986. Climatic and Edaphic Classification at a Continental Scale for Cassava in Latin America.' Cali, Colombia: CIAT, mimeo. Carter, S., and P. G. Jones. 1989. COSCA Site Selection Procedure.- COSCA Working Paper No. 2. UrA, Ibadan, Nigeria. Cessay, S, E. Gilbert, B. Mills, J. Rowe, and G. W. Norton. 1989. Analysis of Agriculual Research Priorities in the Gambia. ACIAR-ISNAR Prqject Paper No. 16. ISNAR, the Hague. Davis, J. S., P. A. Oram, and J. G. Ryan. 1987. Assessnt of4gricatul Research Priorities: .An Intonaional Perpectiv. Canberm ACIAR. EcheerF R. G. 1990. -Assessing the Impact of Agricultural Research. In R. G. Echeveria, ed., Methods for Diagnosing Research System Contraints and Assesing the ipact of Agricura Resarch. The H :ague ISNAR. Freebaim, J. W., J. S. Davis, and G. W. Edwards. 1982. -Distribution of Research Gains in Mltistage Production Systems.' American Journal of Agriuural Economics 64:39-46. Grham-Tomasi, T. 1991. 'Susainability: Concepts and Ilications for Agricultbl Research Policy. In P. G. Pardey, J. Roseboom, and J. R. Anderson, eds., Agrcu4 ural Researh Policy: Intenational Qanaitatve Perspecs. Cambridge: Cambridge Univsty Press. IRRL 1992. Enwvironmental and Health hnpact of Pesticides in the Philippine Rice Production.' Los Banos: IRRI, mimeo. 124 Lynam, 1. K. 1992. 'Sustinability: The Challenge for Agricultural Research." Par presented at Workbhop on Social Science Research and the CRSPs, University of Kentucky, Lexington. June 9-11. mimw. Lynam, J. K. 1989. wOn the Design of Commr'dity Reseach Progrms in the Intenational Centers." In D. Groenfeldt and J. L. Moock, eds., Soial Science Perspeves on Managing Agriculra Technology. Colombo, Sri Lanha: IIMI. McCalla, A. F. 1994. 'Ecoregional Basis for Interational Research Investment.w In J. R. Anderson, ed., Agrcudural Techology: Policy Issues for dte Intenaonal Communiy. Wallingford: CAB Internatonal, in press. McCafa, A. F., and J. G. Ryan. 1992. 'Setting Agricultural Research Priorities: Lessons from the CGIAR Study." Amerian Joumna of Agriculul Economics 74:1095-1100. Pachico, D., J. K. Lynam, and P. G. Jones. 1986. "T7he Distribution of Benefits from Technical Change Among Classes of Consumers and Producers: An Ex-ante Analysis of Bns in Brazil.' Research Policy 16:279-85. Pardey, P.G., and S. Wood. 1994. Targeting Research by Agriculual Environments.' In J. R. Andeson, ed., Agriutural Technology: Policy IssBus for the Internaional Communty. Wallingford: CAB Intnational, in press. Pinstrup-Andersen, P., N. Ruiz de Londono, and E. Hoover. 1976. 'Tbe Impact of Icreasing Food Supply on Malnutron: Implications for Commodity Priorities in Aicultural Research Policy.' Amenwran Jounal ofAgriuwlta Economis 58:132-42. Pollak, L. M., and J. D. Corbett. 1992. 'Using GIS Dataseft to Classify Mize-gowing Regions in Northem Latin America.' Lisboa, Mexico, D.F.: CIMMY, mimeo. Quizon, L., H. P. Binswanger, and D. Gupt. 1991. 'The Distribution of Income in India's Northern Wheat Region.' In R. E. Evenson and C. E. Pray, eds., Researc* and Axctviiy in Asian Agricultre. Ithac, N.Y.: Conell University Press. Rekow, M. 1991. 'Modeling the Aggete Effects of Technological Change on Income Distribution in Pakistan's Favored and Marginal Production Evnments." CIMMYT Economics Paper No. 4. Lisboa, Mxico, D.F. TAC/CGIAR 1992. 'Review of CGIAR Priorities and Strategies.' Rome: TAC Secretariat, mimeo. University of Georgia (undated). 'A Lndscape Approach to Sustnability in the Tropics: The Global Plan for the Sustainable Agricuture and Nahtal Resurce Managmt Collabortive Research Support Pmgam (SANREM CRSP).a mimeo. 125 TAC's Perspectives on Social Science Research in the CGIAR Guido Gryseels, TAC Secreuiat CGAR and Food and Agdcu rwe Organization, United Nations The Technical Advisory Committee to the CGIAR has four main funcdons: * To make recommendations on CGIAR priorities and strategies. * To ensure quality and relevance of CGIAR research. * To make recommendations on resource allocation among centers in the context of CGIAR priorities and strategies. * To provide intellctual leadership to the CGIAR. Tiis paper attempts to distill TAC's views on social science research by analyzing its recommendations on CGIAR priorities and strateges and its recommendations following extera reviews, which are the commitbee's main vehicle to assess the quality and relevance of center research. The analysis shows that TAC considers social science research to be a crucial and integral part of CGIAR activities. Increasing priority is being given to poLicy analysis, but the other major program areas of social science activities in the CGIAR are their contribution to prionty-setting, impact assessment, constaint analysis, and specification of desired technology characterics. With the rect epanon of the CGIAR and the need to give greater emphasis to environmental concerns, social science research will have to incaease attntion to natutal resources maement and on and to the needs of the new forestry and fisheries sectors. It will also need to play a leading role in guiding the implementation of the proposed ecoregional approach to research in the CGIAR. Social Science in CGIAR Priorities and Strategies The analyic framework for the most recent rew of CGIAR priorities and strategies (TAC/CGIAR 1992) had three dimensions: Ihe iews expressed in his paper are the personal ews of dt author and do not necessaiy represent those of the CGIAR Technical Advisory Comnmee or the Food axd Agicudtu O,ianition of the United Nations. 126 * An acdi'des dimension, with five categories of activities: conservation and management of natual resoures; germplasm enhancement and breeding; producdon systems development and management; socioeconomic, public policy, and public management research; and institution building. * A spatial dimension, with four geographical regions and nine broad agroecological zones. The overlay of these classifications led to a total of twenty-three regional agroecological zones. * A product dimension, with four main production sectors: crops, livestock, forestry, and fisheries, and their corresponding commodities. In the crop area alone, the CGIAR currently supports approximately twenty-five commodities. In discussing TAC's perectives on social science, particular attention should be given to the activities dimension. Social science research plays an important rc.e in all categories of activities, except perhaps in germplasm enhancement and breeding. It should also be noted at the outset that there is no clear dividing line between these categories of activities and that in many instances, activities may overlap in seveal categories. Furthermore, a number of CGIAR objectives cut across all these categories, such as the strengthening of national research systems and improvement of sustainability of farming systems. The categorization is needed to enable aggregation of center activities in systemwide terms. TAC recommeded a significant increase in the research efforts on natural resources conservaton and management and on socioeconomic, public policy, and public management. It recommended that the current prority regarding germplasm enhancement and breeding should be maintained. The committee also recommended that CGIAR investment in research on the development and management of production systems and on institution building be reduced because of increasing strength of national programs and lack of comparative advantage for the COR. In 1991 the expanded CGIAR allocated approximatey 9 percent of its core resources to socioeconomic, public policy, and public manageent research. TAC proposes to increase this to between 10 percent and 12 percent. The committee had made an analysis of the dcallenges facing reserch and development in agiculture, forestry, and fisheries untl the year 2010, globally, by region, and by agroecological zone. This analysis clearly showed that in all regions and most agroecological zones, tiere was a need for greater emphasis on policy research with respect to land-use and sustinability issues, poverty alleviaon and equity (particularly gender equity), irrigation maagement, and issues related to self-reliance. Even more emphasis on the new system level may be needed in Asia and in 127 Latin America and the Canbbean because of the pressing problems of sustility, deforesttion, and equity. Considerable through has been given to the role of the CGIAR, as distinct from the other actors on the scene, in policy research. New technology is widely perceived as the fuel for the engine of agricultural development. The CGIAR, as an international agricultural research organiztion, has an unnvaled overview of both the agricultural technology needs i developing countries and of the global technological opportunities in agriculture relevant to those needs. This overview, and the importance of equity in the mission and goals statement of the system, are primary components of the perspective which the CGLAR brings to policy research. At the country level, policy research institutions are primarily driven by national considerations. At the internadonal level, the UN agencies, the World Bank, the regional banks, and some developed country universities which do policy research, have missions and intended beneficiaries in common with the CGLAR. They lack, however, the CGIAR's capacity to identify and generate new agricultural technologies to meet developing country needs. It is this capacity, together with its political independence and its established reputation and track record, that allows the CGIAR a unique stance in policy research. The CGIAR recognizes that its capacity for policy research will of necessity remain modest. TAC has emphaszed that the CGLAR will remain no more than a catalyst in the field, because the resources invested will continue to be but a fiaction of the total national and intenional resources invested. Where there is no advantage from its umque perspective, the CGIAR should rely on other agencies. Its main in-house asks are to understand the inteactions between government action and human behavior in relation to agrculture, technology, natural resources, and consumption and to collaborate with national systems in identifying policy options that will improve the welfre of the system's beneficiaries. The CGLAR expansion into forestry and fisheries and the new priority given to research on natural resource managemet imply new priorities in policy research. Historically, the CGIAR has focused mainly on commodities, with some incursions into research on farming systems. In 1990, TA 7 concluded and the CGLAR accepted that effective research in natural resource management must address both the technical and the human sides of the problem at both the farm and community levels. It has proposed to restuce the CGIAR into global and ecoregional activities. This move will imply new responsibilities for social scientists in the CGIAR. An early pnorty is to understand more about how people degade the natural resource base-and especially the effects of poverty on the envimnment. Degradation in the marginal agricultural areas and at the forest edges highlights the need to reduce the human pressures on them. 128 TAC has also emphaszed the continuing importance of impact assesment and the role of s nomics in constraint analysis and specficaion of deired technology characteristics. However, the level of effort in the lattr role may be dcining becau of the reduced priority assigned to research on production systems and the increasing capacity of national programs to deal with this ye of activity. External Reviews TAC does not consider social science research as such in the process of external program and management reviews of CGLAR centers. The purpose of extnal reviews is to help to ensure that the centers continue to implement stategies and programs that are reevant to the CGIR goals, that they maintain or enhance their record of achievt, and that they are efficiently managed (TAC/CGIAR 1991d). As such, reviews assess a center's strategy and progams and program orga on and nmangement, rAtean individual research disciplines. In the terms of reference for extemal review, there is no speific mention of social science research, although occasionally issues are raised in the list of additional questions that are specific to the center under review. An assessment of the centers' mechanisms for priority-setting and impact assessment are, however, an integal part of the terms of referece for external reviews. An analysis of the most recent ernal reviews and the Cor ing TAC commentary of the Otraditional tirten CGIAR centers highights the importance TAC attaches to social science research. The following is a brief overview of some of the issues raised in these reviews, by social science program thrus and by center. Pliority-setting CAT: TAC concunred with the panel that ClArs strategy of inclusion of socioeconomic reseach within multidisciplinary commodity programs had worked very wdl. The committee was plased to note the strong interactions that took place between the commodity team workers and the contnbutions by social sciensts in te process of setting priorities for biological research within progms. TAC also noted that CIArs mant was giving greater attention to setting priorities across programs and ta its program economists were malkng a contribution to this process. The committee wished to encourage this ype of intisciplinary activity across programs (TACICGLAR 1990a). CIP: The panel recommended that a mechanim for both potato and sweepot be develqped to incorpora information derived from global containts stuies into the research planning process to effectively allocate resources. It detened tha,t a cear role should be devdoped for the Social Science Deparment to assist in both the asmet of interdisciplinary research needs and of CIP's impact (TACICGIAR 1990b). 129 ILRAD: In most CGIAR centers, sociomic and training activities play a crucial role in ensuing the deveopmental relevance of scientific progrms. The panel stated that each of these programs could play a greater role in ELRAD's research pii and stategy formulation CrACICGIAR 1992a). Poily Anaysi IBPGR: TAC agreed with the panel's recommendations dhat IBPGR should include issues relating to policy and to social aspec of the conseration of plant genetic resources in its research agenda, and should analyze ways in which this might be accomplished (TACICGIAR 1991b). IRRI: TAC agreed with the panel's recommendation that the Economics Department should place more emphasis on interdisciplinary research in areas such as resource management and the economics of sustainability in difet nce-growing envinments; the economics of pest management; and the impact of new rice technologies on the role of women in nce production and rice-farming families. TAC also agreed with SRI, however, that it is essential to maintain some capacity for the analysis of aggregate sectomal and policy issues. Both will be needed for several purposes, which include assessing the global rice situation, assessing the potential impact of IRRI technologies as an important contribution to the process of strtgc planning, and assisting national progms to analyze issues in rice reserch and policy (TAC/CGIAR 1987). CEMMYT: The panel recommended that the Economics Progam move expeditiously toward a reallocation of its resources by reducing its activities in on-farm research and aining and by increasing its research activities for geneating critical information for development of wheat and maize varietals for CIMMYT and commodity progam research resource allocation, and for wheat and maize market policy analysis (TAC/CGIAR 1989). Impact Assesrment IFPRI: TAC commended IFPRI for its stated impact, but noted that the assssm t of IFPRIrs impact had been somewhat anecdotal. It suggested that IFPRI incorporate into its review and monitoring processes an appropriate measure of the impact of its policy research, capacity-building in NARS, and contribution to science (TACICGIAR 1991a). ICRISAT: TAC noted that available information on ICRISAT's impact was largely anecdotl and concunred with the panel that ICRISAT should commission a study on an ex- post evaluation of the impact of a sample of its activities (TAC/CGIAR 1991c). 130 Constrant Analysi and Secf on of Dedred Te ology a ICRISAT: TAC sated its arion of die ineWation of biological and socal scientists within the framework of die resourcme managemet -progam. It noted, howeve, that this approach, which had been initiated in response to a recommendation of the previous external review, had not yet achieved its intended effect, and it was not clear how socia scientists contnrbuted to commodity resarch at ICRISAT. The committee concluded that greater attention to the role of social scientists might be warranted, particularly their role in identfication of cnstraints, specification of deired technology charactrsc, priority- setting, and assment of technology adoption. It further sugted that ICRISAT might conader allocating a greater share of resources to social science research, which at the tim rnepeeted no more than 3 prcent of ICRISArs total corm budget (TAC/CGIAR 1991c). Other Recommendatos CIMMYT: The pand recommended that the medium-term and annual plan of work of the socia and biological scientists in the regional progrms be prepared joindy, with the picipion of the directors of the respective prgams from headquartrs in regional planning meetings (TAC/CGIAR 1989). Conclusions TAC gives substantial importance to the role of social science rsech in CGIAR activities. Increasing priority is being given to policy analysis, but impact assessment, priority-setfing, constaint analysis, and speification of deired tchnology characteristics will remain important activities. Some centers also emphasize researh on marketng of their mandate commodities. It must be undesood, however, that these research areas are not the exclusive terr.iory of social scientists; other disciplines have an important role to play. TAC has said very little about the role of socal scientsts in tho develqpment of m n ic data bases, perhaps because this area is seen as an integal part of priority-setting. TAC does not want to become imnolved in setting research agendas of individual centers; nether does it want to second-guess the results of stategic and operaoinal planning at the center lvel. TAC will not, thereore, make ens on the level of involvement by centers in particular areas of social science research. Fm ore, it belieyes that evaluations of the quality and relevance of particular disciplines can best be done through peer reviews. In the future, incasing priority will be given to policy analysis, an area in which social scientist clearly can make an important contribution. Social scientists also wil need 131 to be closely involved in futu research on natural rsuces m r and fisheries. Additionally, the further of the ec g apprach to research wil also imply new challenges and repniiiisfor social scientists in the CGIAR. References TACICGIAR 1987. Report of the third extnal program review of IRRI. TAC Secraiat, FAO, Rome. TACICGIAR 1989. Report of the tiird extemal program review of CEMMYT. TAC Secrtriat, FAO, Rome. TACICGIAR 1990a. Report of the third external progam revw of CIAT. TAC Seretiat, FAO, Rome. TAClCGUAR 1990b. Report of the thrd extnal progam review of CIP. TAC Secretaiat, FAO, Rome. TACICGIAR 1991a. Report of the second extmal progam review of PRI. TAC Seicariat, FAO, Ronm. TACICGIAR 1991b. Report of the third extenal pwgram review of IBPGR. TAC Secretaiat, FAO, Rome. TACICGLAR 1991c. Report of h third extenal progr and scnd exteal m review of ICRISAT. TAC Secretariat, FAO, Romc. TAC/CGIAR 1991d. Terms of refrn for eal revi of CGAR centers, and guidelines to panels for the conduct of acl rviews. TAC Secraat, FAO, Rome. TACICGIAR 1992a Report of the third external progmam and met view of ILRAD. TAC Secraniat, FAO, Rome. TACICGLAR 1992b. Review of CGIAR prioties and atgies - parts I and IL TAC Secrtriat, FAO, Rome. 132 ANNEXES Annex A MEETING OF TEE CGIAR SOCIAL SCIENTISTS ISNAR, TBE HAGUE, August 17-20, 1992 Monday, 17th August 9:00 a.m. - 11:30 a.m. OVERVIEWS (30 minutes for each prsentation, with an hour for disussion at the end of the three contributions) Opening Remarks: Chrsta Bonte-Friedheim, Director General, ISNAR; Mke Collinson, Science Advisor, CGIAR Secretariat Jock Anderson: Social Science in Agricultual Research: Implicatons and Issues for lARCs Johm Lynam Ineating Research Planing, Prority-sting, and Input Evaluation Vithin the CGIAR Coffee break: 10:30 am. - 10:45 am. Guido Gryseel: TAC's Perspectie on Socal Science Reseach in the CGIAR 12:15 ainm - 3:45 p.m A NEW CGIAR? GLOBAL GERMPLASM AND ECOKEGIONAL MECHANIS: ROLE AND ORGANIZATIONAL IPOLICATIONS FOR THE SOCIAL SCENCES (20 minutes for each presentation, with 10 minutes for caficos and additions to the CIAT: Douglas Pwchico and Guy Henry: Socioeconomic Research Agenda for NaUral Rurce Issues in Tropical Ameica Lunch: 1:00 p.m. - 2:00 p.m. CIMMYT: Derek Byerlee: Issues and Options for Social Scienfs in Global Gamplas Improvement Ike sequece ofpeumatfons in the publishd preceediags dffers seaewhistfru the order of preseatdon. 134 ICARDA: Richard Ttwiler and Wilem Jansen: Priorities for Social Scientists in an Ecoregonal Center UTA: Joyotee Smnh: New Challenges Confronting Social Scientists in IARCs with Eooregional Mandates Coffee break: 3:45 p.m. - 4:00 p.m. 4:00 p. m. - 6:00 p.m. WORKNG GROUPS (break into two to four working groups, depending on the issues which anse; working groups to report back for an hour the next moniing) Tuesday, 18th August 9:00 a.m. - 10:00 a.m. WORKING GROUPS REPORT BACK 10:00 a.m. - 12:15 p.m. INSTITUTIONS,, POLICY, AND NATIONAL PROGRAMS (20 minutes for each presentation, with 10 minutes for clarifications and other issues) WARDA: Peter Matlon and Aldn Adesin Potential IARC Contributions to Building Social Science Cpacity in NARS Coffee break:: 10:30 a.m. - 10:45 am. IML- Douglas Merey and rla flandgodLl The Role of Sodal Sciences in IIM's Mandate to Strengthen Institutions and Link Research to Policy I`PRI: Steve Vosti Food Policy Reseach in the CGIAR - IFPRI and the other Centers ICRJSAT: Tim Keley: An Economist's Contribution to an lARC's Phans and Stategies 12:15 p.m - S:00 p.m. (Lunch: 1:00 p.m. - 2:00 p.m.) WORKING GROUPS (break intO two to thtree working groups, dependig on the issues which anse; work though lunch; working groups to report back for 20 minutes, including questions) 135 Informal dinner (optional): 7:30 p.m. (Indonesian food) Wednesday, 19t August THREE ISSUES FOR SOCIAL SCIENCE IN AGRICULTURAL RESEARCH 9:00 a.m. - 10:45 am. (1) THE USER PERSPECTIVE (15 minutes for each presentation, with 5 minutes for clarifications) CIP Greta Watson and Jorge Quiroga: The Social Sciences and Participatory Research in Gennplasm Management CIAT: Jacqueline Ashby and Louise Sperling: nstitutioig Participatory, Client-iven Research and Technology Development in Agriculture CIdMMYT: Robert Tripp: Issues Related to the Users' Perspectives ISNAR: Mrie-H&lne Coflion, Thomas Eponou and Deborah Merrill-Sds: Research Program Planning and Priority-Setting: Strengthening Farmers' Input IML Doug Venmlio: User's Perspectives for Paricipatory Research and Developmnent Coffee break: 10:45 am. - 11:00 am. 11.00 a.im - 2.00 p.m. (Lunch: 1:00 p.m. - 2:00 p.m.) WORKNG GROUPS (break into two to three working groups, work through lunch, with 2 hours for discussion and 15 minutes each for rports and questions) 2:00 p.m - 3:20 p.m. (2) SUSTAINABILITY (15 minutes for each presentation, with 5 minutes for clarifications) flTA: Anne-arie lzc, Miael Swift and Karen Dvorak: Integration of Economics and Ecology in Addrssing Suinability Issues in Agroecosystems ILCA: Brent Swallow and Simeon EhLu Integrting Naural Resourc Management Issues into ILCA's Research Agenda CIMMYT: I n : Susiablity, Economics and Ecology: Issues and Opporuties for IARC Social S s 136 ICLARMi Robert Pomeroy Institutional ngemets for Community-based Coastal Fisheries ManagnLt Common Prperty Resources 3:30 p.m. - 6:1S pm. (Coffee available: 3:45 p.m.) WORKIG GROUPS reak into two to three workng groups or more, depending on options aen, with 2 houn for discussion and 15 minutes for each for rports and questions) Thursday, 20t August 9.:00 a.m. - 12-.00 noon (3) THE CONTINUUM REVISiTED: SOCIOECONOMIC DATA BASES, PRIORITY SEWTING, AND DIPACT ASSESSMENT (15 minutes for each presentation, with 5 minutes for clarifications) ISNAR. Phiip Pardey: Information for Planning and Priority-setting FITA: Karen Dvorak. Development of a Georeferenced, Socioeconomic Data Base for African Agriculture IfqM Edward Vander Velde: Socioeconomic Data Bases: Prospects for Effective Interceter Collabtion IRAD: Adrian Mukhebi, John Curry, and Brin Pernr: Assesing Impacts of Improved Control of Livestock Diseases: Lssues in Collection and Modelng of Field Data IRRB Mababub Houair Priority Issues Social Science at IRRI Coffee break:- 10:45 am. - 11:00 am. ICRAF: Susan Minae and Steven Franze: Socioeownomic Analysis of Agroforestry Adoption and hnpact: Issues and Options ISNAR:. Doug Horton- Monitoring and Evaluation in National and International Research: Issues for CGIAR Social Scientists CIP: Peter Eweil: Working with NARS in the PRAPACE Network to Develop an Information System for Monitoring and Evaluation 12:00 noon - 3:00 p.m (lunch: 1:00 p.m. - 2:00 p.m.) 137 WORKING GROUPS ealk into two to hiree worling groups, work tirough lunch, with 2 hours for dcusion and 15 minutes each for reprt and questions) 3:00 pm. - 4:30 pm (Coffre available: 3:45 p.m.) PLENARY SESSION: MATTERS ARISING FROM THE MEETING Meet Adjourned 138 AnDne B Medfag f die CGLiR Socal Sdeat# ISNAR, he Hage Augt 7-20, 19.92 LLM of Iidk4,o CUT ICARDA Apartado AarOO 6713 P.O. Box 5466 Cali, Cooknbia AMeppo, Synan Arab Republc Phone: (57-23) 675050 Phone: (963-21) 213433 Fax: (57-23) 647243 Fax: (963-21) 225105 * Jacquline Ashby * Willen Janssen * Guy Henry * Richard Tutwiler 3 Douglas Pachico ICLARM MC P.O. Box 1501 3 Louise Spering Makati, Metro Mmanla 1299 IAT/Great T Las Philippines B.P. 259, Buae, Rwanda Phone: (63-2) 8180466 Phone: 250 30446 Fax: (63-2) 8163183 Fax: 250 30599 * Robert Pomeroy C_YT P.O. Box 6-641 ICRAF Mexico 06600, D.F. Mexico P.O. Box 30677 Phone, (52-595) 42100 Nairobi, Kenya Fax: (52-595) 41069 Phone: (254-2) 521450 Fax: (254-2) 521001 * Dkrek Brke * Steen Franzel * [ay FBr ngo * Susan Ma * Robert Tidpp hMlawi ICRAF AF Proect OEP P.O. Box 31188 Apartado 5969 Lilongwe 3, Malawi LinIa Peru Phone: (265) 767222 Phone: (51-14) 366920 Fax: (265) 731014 Fax: (51-14) 351570 * Dirk Hoeksta * Petxr Eweil P.O. Box 25171 Nabi, Kenya Phone 254-2-632-054 Fax: 254-2-631-499 139 * Greta Watson ICRISAT Casilla 4285, PROINPA Patadu P.O. Cochabamba, Bolivia Andhm Pradesh 50 324, Indi Phone: (591-42) 40929/49506 Phole: (91842) 22016 Fax: (59142) 45708 Fax: (91442) 241239 * h KTiey Email: 157:CG1272 gmu ERAD 1200 17th Stre, N.W. P.O. Box 30709 Washington, DC 20036-3099, USA Nairobi, Kenya Phone: (202) 862-5600 Phone: (254-2) 632311 Fax: (202) 457-4439 Fax: (254-2) 631499 * PeterHazli * Adrian Mukhebi * Steve Vosti IRIU P.O. Box 933 Ian Manila, PHip P.O. Box 2075 Phone (63-2) 818-1926 Colombo, Sri Tanka Fax: (63-2) 81748470 Phone: (94-1) 567404 Fax: (94-1) 56654 * Mahabub Hosain Email: MC 157:CGI129 ISNAR - Douglas Mre Law van Nieuw Oost Indie 133 2593 EM The Hague, Nehlands * Doug Vamion Phone: (31-70) 3496100 Fax: (31-7C) 3819677 * Edward Vander Vede MAI * Marie-Hlbn Coho 1-A/B Da Road GOR 1, Lahore, Pstan * Thomas Eponou Phone: 9242-368610 Fax: 9242-369194 * Peter Gddworthy Email: 157:CG1220 on IIMI-PAK * Doug Horton IITA PMB 5320 * Phlip Parf badan, Nigeria Phone: (234-22) 400300-317 * Deborah M USand. Fax: (234-1) 611896 RR1, Box 122B OHWdvme/Lyne Road • Karen Dvorak Orford, NH 03777, USA Phone: (60) 3534306 * Joyotee Smith Fax: (603) 353932 Email: 157:0CGIDK * Anne-Marie Izc 140 ILCA WARDA P.O. Box 5689 01 B.P. 2551 Addis Ab., Edhopia Bouake 01 Phone: (251-1) 613215 COte d'Ivoire Fax: (251-1) 611892 Phone 22 632395 Fax: (225) 634714 * Simwon Ehui *Akin Adesina * Brent SWalow P.O. Box 46847 Nairobi, Kenya Phone 254-2-632066 Fax: 254-24631481 AVRDC World Bank P.O. Box 42 1818 H Street, N.W. Shanhua, Tainan 741 Washingn, DC 20433, USA Taiwan, RO.C. Phone: (202) 473-0437 Phone: (886) 5837901 Fax: (8 6830009 * Jock Andeson Email: 157:CGI121 TAC Se8utariut 3 UHas Jansen FAO Via delle Terme di Camcala Rockeeler Foundation Rome 00100, Italy P.O. Box 47543 Phone: (39-6) 797-3306 Naimbi, Kenya Fax: (39-6) 5797-5731 Phone: (44-4) 228-061 (Naimbi) (212) 869-8500 (New York) * Guido Gryseds Fax: (212) 764-3468 CGLAR Secmtariat * John Lynam 1818 H Street, N.W. Washington, DC 20433, USA USAID Phone: (20) 473-8951 Washington, DC 20523-0057, USA Fax: (202) 473-8110 Phone: (703) 875-4285 Fax. (703) 875-4394 * bfike Comison * tRalph Cummings, Jr. * Kern Wright Platais 141 Annex C Lst of the CGL4R Ceatrs That Pa,*4wW CIAT UMI Apartado Aeroo 6713 P.O. Box 2075 Cali, Colombia Colombo, Sri anfa Phone: (57-23) 675050 Phone: (94-1) 567404 Fax: (57-23) 647243 Fax: (94-1) 5668S4 Email: IDMI 157:CG129 CIMMYT IITA P.O. Box 6-641 PMB 5320 Mexico 06600, D.F. Mexico Ibadmn, Nigeria Phone: (52-595) 42100 Phone: (234-22) 400300-317 Fax: (52-595) 41069 Fax: (234-1) 611896 CIP ILCA Apartado 5969 P.O. Box 5689 Limna, Pru Addis Ababa, Ethiopia Phone: (51-14) 366920 Phone: (251-1) 613215 Fax: (51-14) 351570 Fax: (251-1) 611892 ICARDA ELRAD) P.O. Box 5466 P.O. Box 30709 Aleppo, Syrian Arab Republic Nairobi, Kenya Phone: (963-21) 213433 Phone: (254-2) 632311 Fax: (963-21) 225105 Fax: (254-2) 631499 ICLARM IRRI MC P.O. Box 1501 P.O. Box 933 Maiaci, Metro Manila 1299 Manila, Philippines Philippines Phone: (63-2) 818-1926 Phone: (63-2) 8180466 Fax: (63-2) 817-8470 Fax: (63-2) 8163183 ICRAF ISNAR P.O. Box 30677 Laan van Nieuw Oost Iadie 133 Nairobi, Kenya 2593 BM The Hague, Netherlands Phone: (254-2)521450 Phone: (31-70) 3496100 Fax: (254-2) 21001 Fax: (31-70) 3819677 ICRISAT WARDA Pamancheru P.O. 01 B.P. 2551 Andbra Pradesh 502 324, Tndia Bouake 01 Phone: (9l-842)224016 C6te d'lvoire Fax: (91-842) 241239 Phone: (225) 632395 Fax: (225) 634714 1200 17th SUbn, N.W. Washiton, DC 20036-3099, USA Phone: (202) 862-5600 Fax: (202) 457-4439 142 Distributors of World Bank Publications ARGENTINA The Middle EntObw ITALY PORTUCAL Cab. HbadkifSL CDIhUStreUt SIna SPA Linalr PbaaglJ Clwisoa Case Via Duo Di Calabr. 1/1 RUA Os CAr. 70.74 Rlad. 165. 411 FbaaOfk. 49C/46 Cmits Paint 352 1200 Mb.. I= _t... Abe 5ND25 Fivma £gs.j, Kk* SAUDI ARANA. QATA. AUSTRALIA. PAPUA NfEW GUINEA . PAX Boxrr- JAPAN a RJILSOLOMON ISLANDS. S.eOIM HeldaI 10 L4ikm flak Sonc rP0. bs 3i96 VANUATU,ANDWSTENSAMOA Hongs Sarst. Bflunkysk 123 Riyadh 1101 DAIIania S 5m FRANCE Toky 6 WNOV. Read World flnk Plbhmdaon SINCIAPOILTAIWAN. bUIJUZ3132 66 vmvu' dl IKENYA MYANMALURIJN Vkie. 75116PU,b Aike fask Song (LU i d. GawwAdsFadErPb Ld. QOarant Mo t DapD Sir. Coldm Whil RUddW AUST1UA GEIMANW Pr.ao. 4524 41. Kdlang Puddkg OSC CIld and CL. UNO.vrlag Naebi SIngapore U134 A-l011 Wien DamDD A. KOREA. REWUUC OF SOUTH AFRICA, BOTSWANA Pan Kaim flask Caepsusde Fiw smgb ids UANGLADOEI nONC KONC. MACAO P.O. OaI iM. Kweawhmsun Odord Ua g uswoteyIe MIs ktdaDus Devabrma t AW- 2= Led. Sea SsUdwm AfIcI Asdta SuSdSyO.SDAS) 4&0 WJpaJMv SUM fle% 1 NAsaL RaiWXdh 2 P.O. Ibx 1141 HGOOSRAMNd 26 Wi Cn ent. Kaen SIOck Baok Cant. C pe Town 000 O2_urAwAl R/A_ 2nd Rasr P.O.Bwo34 Dk2d9I C-rr Honezg Ywa Forihw rm A 5eo Inentoa Subcpdo So 1ad aUgCeY P.O. Box 410 1Viewtir F otnfor Maketmmy MALAYSIA CMighd IIAgaubd CommJdal Arm DosbnvaI ut v-z9 Uuivuqof MaaaCa dve Johavoashurl2024 Ch-"54e01 H-t1V78udspt 500011IP a 4P.O. Sa11l 27Idesn Pant Nam SPAIN NELGIUM NI S910 Kuala Laimpu Mu.dI-PUUI. LElira SLA jeas De L 7 Allied Publishes Pilvab LdL Ctstls7 At d ald 202 751 Mouagt Raid MEICoM 1060NmRs _n - ra- 0t02 0EC Apasudo Pals 22-MO Lberfa Intenadia AUICO CANADA Psa spas: 1400 Tin, Meal DcF. Camel de Can. 391 LEa DOfimtN IS I.N. HIludla MMug OmRIAND DSIA BBuA de Mwbwu Est" NETHERLANS _S-Ab.Qrl. albp Sabh .y-400030 De Lin.dubsAla Pbllkad SRi LANKA AND TM MACADIV lUEa Pfa Om = Lake HKae BEkap 13/14 Am.! U Road 7OMAEH&sbsrP F. INb 244 : eailpowbhw"C& NeD W1-11002 lC0.S1rChftaps&nA. =IAlguafod NEW ZFALAND CGrdiltMawths ONEWS.Onfado, Vhitmad nAvause EMSCONZLNI Colornbsi KB3W11 Cikutia- 7w0a72 P1dvatMFaUBq94s NewMae SWEDDa anLE yda p ugding Auckland Ferasagklaad lhnCTSA. G Fnae.ordbk Fmclakore Av. SmstmmSl 6400 -B ob--Y60 NIGERIA Rhgmisptn . 3am i6 lEd.IRfi OL201 Unirvrty a. ILImld 5-10322Swckhalm Sa;e 3nflaKs&us ada ThmerCrowmuildlngsjeka c RodPate Mall Bag S9 Fwsrha oaEh. - CUNA Hyd d -500 027 adai Wannesgrean-WlrimWAS ah. FMmdi &tEaaaIc P. a afl 130 PubibbiHone ftlab= Fta hi RFowr NORWAY S-17155 Sa LDsFoSiDoaagJlr Nar himbere a Navumngpnr 6h InfolsilmarnCadar Wlas A hnwdabed - 38 g 1X19 Boa Engd -W snRAND P.O. Box 612 Eitsjfd &rm.eigfrdm CXXOLOIA PatialaHown N4e6m Oie6 Lie pa]tl lafaskCaLue. IWiA AshlkMar Cuat pof1 3212 ApfrdJAer34DO Ladnaw - 226 00 PAKISTAN H11002 Lunsaa flpro Mir£ flaok Agency CtBargoad 6Shhd-u-Ana Fanurba*anm w : CaM DIVOIRE 60 flajsager P.O. am No. 729 LteIsPrbt Canf dCEdion atde DlHfuon N"pur440O010 LeS401 ladeAbnmes AMkSN CEA) Cam posak 33 0c N5541 RPINDONESIA PERU CH 01M Lauauw Abidjan Maln Pft. ladl Limmiad Edaital DearolsA JabnB frbudnr2o Ap-do3124 T ALAD CYIRUS P.O. fla M Lima 1 CantI DepsmntStae Catle f ApplIed R h jakwt M2 3061CI IR0ad cprusCll4eg PIUIMNES Bangkok O _ ow Suet. EalmI WAN lrtmalaaaJ BOok CXAer P.O. Bot2 K1wbb Publi'wrs Suitw 17L Cityand 10 TRINIDAD & TOBAC4 ANTICUA NIMSl: P.O. BfX 1D75-51 Caadancg owe Ta I *ARDA, BAIRSADOS Tebn. Ayala Avem, HuV. dWs DOMINICA, GRENADA. GUYANA, DENMARK GCuExaLta JAMAICA. MONTSERRATS. SandaljlmbUr ItELAND MaktidMeuuManIsa IlOTs NEVIST.FLUCIA R_eaVtro AI1 11 Cavunmmnt Supple Agency POLAN SVINCLT4kCRENADfNES DK-1DFtsdwdSgC Harcou tRd POLAND SysaenvatcSeudlUistd Dubtln 2 lstIeafla PzblmIISoimS OS WO Strot DOMINICAN LPUIIUC UL Pt[=wa31/37 cGrp dllramTa1hr. pwrA. ISRAEL 0046l7Wamnwa TAidsd.We ladle Ranraciat l1ebelaCdus309 Yaste Lftara LW. ApembdodeCrOau2WZ-.1 PQlb ffUsgo Pwiido odte UNITEDKIGDOIM SaablAsDagw Tel Aviv S1SI0 Ipsjonsneb MicraInfo l.id. ULOtaS3 P.FOB 3 EGYPT. AAB RUUUUCOF 02-91o6Wamua Alm HampalI.U3Q 2PC AIs Enad Co St The World Bank Headquarters European Office Tokyo Office 1818 H Street. N.W. 66, avenue d'1ena Kokusai Building Washington, D.C. 204331 U.S.A. 75116 Paris. France 1-1 Marunouchi 3-chome Chivoda-ku. Tokyo IO(, Japan Telephone: (202) 477-1234 Telephone: (1) 40.69.30(.0 Facsimile: (202)477-6391 Facsimile: (1) 40.69.30.66 Telephone: (3) 3214-5001 Telex: wuli64I45w .Dn^LANK Telex: 640651 Facsimile: (3) 3214-3657 KCA 248423wOnxwLnK Telex: 26838 Cable Address: imi aKAu WASHINtflONDC ISBN 0-8213-27984