Planning and Priority Setting for Regional Research · methodology. But regional program formulation also requires a priority-setting method that recognizes that one research project

Planning and Priority Setting for

Regional Research

A Practical Approach to Combine Natural Resource Management

and Productivity Concerns

Willem Janssen and Ali Kissi

4RESEARCH MANAGEMENT GUIDELINES

I .N.R .A

Low income

Natural resource

degradation

Poor integration of

production systems

Poor management

of production

systems

Poor management

of natural resources

Poor management

and integration of

production systems

The International Service for National Agricultural Research (ISNAR) assists developing countries inbringing about lasting improvements in the performance of their national agricultural research systemsand organizations. It does this by promoting appropriate agricultural research policies, sustainableresearch institutions, and improved research management. ISNAR’s services to national research areultimately intended to benefit producers and consumers in developing countries and to safeguard thenatural environment for future generations.

ISNAR offers developing countries three types of service, supported by research and training:

■ For a limited number of countries, ISNAR establishes long-term, comprehensive partnerships tosupport the development of sustainable national agricultural research systems and institutions.

■ For a wider range of countries, ISNAR gives support for strengthening specific policy andmanagement components within the research system or constituent entities.

■ For all developing countries, as well as the international development community and otherinterested parties, ISNAR disseminates knowledge and information about national agriculturalresearch.

ISNAR was established in 1979 by the Consultative Group on International Agricultural Research(CGIAR), on the basis of recommendations from an international task force. It began operating at itsheadquarters in The Hague, the Netherlands, on September 1, 1980.

ISNAR is a nonprofit, autonomous institute, international in character and apolitical in itsmanagement, staffing, and operations. It is financially supported by a number of the members of theCGIAR, an informal group of donors that includes countries, development banks, internationalorganizations, and foundations. Of the 16 centers in the CGIAR system of international centers, ISNARis the only one that focuses specifically on institutional development within national agriculturalresearch systems (NARS).

Planning and Priority Setting forRegional Research

A Practical Approach to Combine

Natural Resource Management and

Productivity Concerns

byWillem Janssen

andAli Kissi

September 1997

I .N.R .A

Copyright © 1997 by the International Service for National Agricultural Research (ISNAR).ISNAR encourages the fair use of this material. Proper citation is requested.

About the Authors

Willem Janssen is an agricultural economist and senior officer in ISNAR’s Policy and SystemDevelopment Program. Since 1993 he has been involved in the design of research planning andpriority-setting tools at the national, institute, and program level. He has collaborated with nationalresearch institutes in Benin, Brazil, Cameroon, Chile, Kenya, Morocco, and Peru and has taughtplanning subjects at the anglophone ICRA (International Course for Research in Agriculture)course, ISNAR courses for research managers, and M.Sc. courses. Janssen has focused on theintegration of new research approaches, such as biotechnology, postharvest processing, and naturalresource management in planning and priority setting. Before joining ISNAR, he was employed byWageningen Agricultural University, and two other international agricultural research centers.

Ali Kissi is an agronomist and inspector general of Morocco’s INRA (Institut Nationale deRecherche Agronomique). At the same time he is an associate staff member of ISNAR’sManagement Program. Until 1994, Kissi led the program planning division of INRA. He has beeninvolved since 1990 in the development of program planning methods, as documented in the“Guide to Program Planning and Priority Setting” that he wrote in collaboration with Marie-HélèneCollion. Kissi has applied such planning methods in Morocco, Tunisia, Algeria, Benin, BurkinaFaso, and Senegal. He teaches research program planning in the francophone ICRA course. Kissihas also been involved in the development of project planning tools, monitoring, evaluation andauditing methods and the assessment of institutional structure.

Citation

Planning and Priority Setting for Regional Research: A Practical Approach to Combine NaturalResource Management and Productivity Concerns. Janssen, W. and A. Kissi. Research ManagementGuidelines No. 4. The Hague: International Service for National Agricultural Research.

AGROVOC Descriptors

management; organization of research; planning; public research; research policies; researchprojects.

CABI Descriptors

agricultural research; government research; management; planning; public research; research.

ISBN: 92–9118–034–3ISSN: 1022–9973

Table of Contents

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixExecutive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiAcronyms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv

Part 1. Rationale, Background, and Issues

1.1 Why Formulate Regional Research Programs? . . . . . . . . . . . . . . . . . . . . . . . 31.2 What is a region?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.3 Objectives and Features of Regional Research Programs. . . . . . . . . . . . . . . 5

NRM emphasis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Proximity to the user . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Problem orientation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Organization of regional programs . . . . . . . . . . . . . . . . . . . . . . . . 6

1.4 The Basic Methodology for Formulating Research Programs. . . . . . . . . . . 7Brief overview of the eight steps in program formulation . . . . . 8Participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Organization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.5 Preparing for Regional Research Program Planning . . . . . . . . . . . . . . . . . . 9Placement within the national agricultural research plan . . . . 10Defining the region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Participants in regional program formulation . . . . . . . . . . . . . . 15Sharing the same language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Part 2. A Stepwise Procedure for Crafting a Regional ResearchProgram

2.1 Regional Review and Analysis of Regional Development Objectives . . . 19Analysis of regional development strategies . . . . . . . . . . . . . . . 20

2.2 Constraints Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Categories of constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Exploiting regional potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.3 Evaluation of Existing Research Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.4 Defining Research Objectives and Strategy . . . . . . . . . . . . . . . . . . . . . . . . . 282.5 Identifying Research Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Classifying research projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.6 Choosing Priority Research Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

A methodological framework for priority setting . . . . . . . . . . . 35Defining benefit dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37Estimating the size of benefit dimensions . . . . . . . . . . . . . . . . . . 38Assigning projects to benefit dimensions and calculatingtheir expected impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Estimating other project parameters. . . . . . . . . . . . . . . . . . . . . . . 45

iii

Cost of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Defining and interpreting the final project scores . . . . . . . . . . . 47Organizing the priority-setting workshop . . . . . . . . . . . . . . . . . 48

2.7 Human Resource Gap Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502.8 Preparing for Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Part 3. Concluding Remarks on Regional Research Planning

3.1 Nature and Time Requirements of the Planning Process. . . . . . . . . . . . . . 573.2 Future Problems and Opportunities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Decentralization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58Questions for further development . . . . . . . . . . . . . . . . . . . . . . . 58

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Boxes

1. Program structure proposed in Benin’s agricultural research plan . . . . . 112. Importance of program types for various research themes in Benin . . . . 123. Defining target zones for Benin’s regional research programs . . . . . . . . . 144. Information to be collected in a regional review . . . . . . . . . . . . . . . . . . . . . 205. Summary of the regional review for Benin’s Central Region . . . . . . . . . . 226. Portions of the constraints tree developed for Benin’s Central

Region research program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 A selection of nonresearchable constraints for Benin’s Central Region. . 258. An overview of results from past NRM research in Benin . . . . . . . . . . . . 279. Selected researchable constraints, research objectives and proposed

project titles for Benin’s Central Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2910. Selected complementary development measures, by constraint,

for Benin’s Central Region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2911. Integrated research projects for Benin’s Central Region . . . . . . . . . . . . . . 3212. Calculation of potential research impact on crop, livestock, and

forestry production. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4013. Estimating the on-site benefits of soil quality research. . . . . . . . . . . . . . . . 4114. Estimating off-site impact of soil quality research . . . . . . . . . . . . . . . . . . . 4215. Importance of various types of research programs for the Central

Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4316. Estimating the size of the benefit dimensions . . . . . . . . . . . . . . . . . . . . . . . 4417. Calculating the impact of identified research projects in Benin’s

Central Region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4618. Calculating final project scores for Benin’s Central Region . . . . . . . . . . . . 4719. Defining and interpreting the final priority ranking. . . . . . . . . . . . . . . . . . 4820. Estimating human resource requirements for first-priority projects . . . . 5121. Suggested outline for a program formulation document. . . . . . . . . . . . . . 54

iv

Figures

1. Steps in research program formulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72. Identification of homogenous regions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Tables

1. Relative strengths of regional and commodity research programs . . . . . 102. Suitability of program types for selected research themes . . . . . . . . . . . . 123. Factors influencing the feasibility of research solutions for

different benefit dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434. Organization of the priority-setting exercise . . . . . . . . . . . . . . . . . . . . . . . . 49

v

About these Guidelines

These guidelines present a method for formulating regional research programs with a mixedperspective — one that combines productivity and natural resource management issues within aregional, decentralized context. Based on earlier work by Collion and Kissi (1994), it has beensubstantially modified to account for the peculiarities of regional research planning. The method waspilot tested in Benin. It has also been applied in Senegal and Morocco.

Part 1 presents the rationale and background of regional research planning. This section is particularlyrelevant for institute directors, policymakers reviewing research options, and, of course, for researchprogram leaders. The method for regional research planning is treated in Part 2. In the first five steps ofthe planning approach, research constraints, objectives, and projects are identified in a structured,logical, and participatory fashion. Step 6 of the research planning approach examines how priorities areestablished among the research projects. This presentation is aimed specifically at economists involvedin research planning and research program leaders. In steps 7 and 8 of the program planning process,the results of the planning exercise are translated into implementable recommendations. Part 2 isfocused specifically for program leaders and research institute directors. Part 3 reviews the overallprocedures and poses some questions and issues that should be resolved in the future. It may be ofinterest to all persons involved in regional research planning.

Throughout the document, a series of boxes illustrates the experience of Benin with regional researchplanning. These boxes aim to provide a clear example of the sequence, organization, and execution ofthe regional research planning process.

vi

Foreword

Demands on agricultural research are becoming more complex. Productivity concerns are nowcomplemented by food security, equity, and sustainability issues. There is a growing appeal foragricultural research not only to lead to higher yields, but also to ensure that the benefits of researchaccrue to target groups and maintain the natural resource base. To integrate these trends in nationalagricultural research programs is a major challenge for research planners. New organizational modelsare needed—ones that allow participation of users in defining and executing the research agenda. Suchmodels should make use of all available information, from both within and outside the nationaldomain. They should also recognize distributional issues of gender, class, and producer and consumerrelationships.

Addressing all these challenges in the traditional framework of the commodity research program ispractically impossible. Rather, innovative ways must be sought for organizing research. This bookoutlines one such way: regional research programs. Regional programs stand close to the user, allowfor a systematic diagnosis and planning, can integrate productivity and sustainability issues, and makeoptimum use of the sometimes limited information available.

Regional research programs provide a comprehensive approach for solving agricultural andenvironmental problems. Constraints to cropping, livestock production, soil management, forestry,and water management are first identified and their links established. This requires a comprehensive,but concise diagnostic tool. The authors present one such tool based on the constraints treemethodology. But regional program formulation also requires a priority-setting method thatrecognizes that one research project may have many types of impact. For example, it may contribute tosolving cropping as well as livestock and soil management problems. The authors suggest a simplemethod, low in data requirements, to set priorities research projects in the regional context.Nevertheless, there is no substitute for data. Regional programs should be based on solidunderstanding of the issues in the region. This knowledge is best obtained by analysis of available dataand by involving users in the planning exercise.

These guidelines are the product of fruitful collaboration between Morocco’s INRA, Benin’s INRAB,and ISNAR. They reflect the fact that research management tools are best developed in practice, indirect interaction with researchers, and in a situation where the results matter. We hope that theseguidelines will be of use to many research institutes in defining their response to the increased needsfor environmentally sound, user-based research.

Abdelaziz ArifiDirector, INRA

Stein W. BieDirector General, ISNAR

Moïse HoussouFormer Director, INRAB

vii

Acknowledgments

Developing a set of practical guidelines requires the interaction and collaboration of many people. Thisdocument is no exception. The authors therefore acknowledge the contributions of the manyparticipants that were involved in planning the regional research programs in Benin and in othercountries. Researchers, farmers, extension agents, lecturers, employees of NGOs, and personnel fromministries of agriculture have all been represented. We express our thanks to Dr. Moïse Houssou,Director General of INRAB (Institut National de la Recherche Agricole au Benin), and Mr. AngeAclinou, Head of INRAB’s Planning Unit. Our collaboration with Dr. Raymond Voudouhé, Dr.Marcellin Ehouinsou, and Mr. Alphonse Yehouenou, who led the formulation of the regional programsin Benin, was essential. Ms. Clemencia Acacha and Mr. Anicet Mouzourri provided excellent logisticalsupport. Finally, we thank the members of the “Programme Aridoculture” in Settat, Morocco, whohelped us develop the ideas that were later elaborated with the team from Benin, and the participantsin the program formulation process at ISRA (Institut Senegalais de la Recherche Agricole), where themethod was further refined and validated.

In the writing of the manuscript we benefited from the input of Gerry Toomey, who simplified ourEnglish and translated some parts from the original French. In addition, Gerry became our firstreviewer, producing critical questions on the logic and the flow of argument. Melina Tensen providedextensive secretarial support, drawing the original figures, and managing the master file. In ISNAR’sPublications Unit, Michelle Luijben, Fionnuala Hawes, and Richard Claase transformed the manuscriptinto a finished book. The document benefited extensively from comments, made by Govert Gijsbers,Peter Goldsworthy, Matthias Hitzel, Brad Mills, Paul Perrault, Willem Stoop, and four anonymousreviewers.

Willem JanssenAli Kissi

ix

Executive Summary

The rationale for regional research planning

Regional research programs can be part of ananswer to trends that many agricultural researchsystems are facing:

■ Concerns over the management of naturalresources. Agricultural research cannot take amere productivity perspective. It has to ensurethat the quality of the natural resource base isnot affected in the process of increasingproductivity, and it also has a responsibility toenhance the quality of natural resources forfuture generations.

■ Decentralization of government to the regionallevel. As part of an effort to improve publicaccountability, countries are increasinglybringing government closer to citizens.Agricultural research has to respond todecentralized public decision making by alsogetting closer to the user.

■ Emphasis on adaptive research. With theimprovement of communication andinformation mechanisms, strategic, applied,and adaptive research results from many partsof the world may be available and applicablein the local context. To resolve production andresource management problems, the testingand adaptation of solutions developedelsewhere becomes an increasingly feasibleand cost-effective strategy.

What is a region? A region is a contiguous areawithin a country, characterized by a certainhomogeneity in agroecological, socioeconomic,and administrative characteristics.

The objectives of regional research programs. Regionalresearch programs provide improved technologyfor a specific region. They aim to optimize currentagricultural production and natural resourcemanagement systems or exchange these systemsfor more advanced ones. The overall objective isto help improve people’s living standards in theregion, while maintaining or enhancing thequality of natural resources.

The place of regional research programs in the nationalresearch plan. Research on agronomy, animalhusbandry, resource management, diagnostics,and institutional strengthening is very suitablyplaced in regional research programs. Geneticimprovement and postharvest research are betterundertaken in commodity research programs(e.g., a maize program) or disciplinary researchprograms (e.g., a food technology program).Other research activities, such as crop protection,policy research, or resource classification studiescan be feasibly undertaken within a regionalresearch program, but may be more suitable toother program types, such as production-factorprograms or disciplinary programs.

Regional research should complement other typesof research programs within the national plan.The objectives of regional research are thereforedefined in relation to expectations of otherresearch programs. Where overlaps exist,responsibilities and interactions should be asclearly defined as possible. Ideally this is donebefore the regional research programs areplanned. If it is done afterwards, research projectsidentified can be transferred to the most suitableprogram type.

The boundaries of regional research programs. Threesets of criteria can be used to define theboundaries of a regional research program. Theseare agroecological, socioeconomic, andadministrative criteria. Many countries havedivided themselves for development strategy andadministrative purposes into a small number ofregions, often taking into account socioeconomiccriteria. If these divisions make any sense foragricultural research, they should be followed toallow for the best interaction between researchand other development activities. Theagroecological criteria can then be used to checkthe relevancy of the divisions, or to suggestsubregions, in case of very large areas. Thisprocedure is preferred over the development of aresearch-specific zonation, because of its thenlimited relevancy to other development activities.Geographic information systems can help in

xi

determining the possible boundaries of a regionalprogram.

Participation in regional program planning. Regionalresearch programs have the advantage of beingclose to the user. It is essential that this advantagebe used from the start, that is, in the planningprocess. Representatives of farmer organizations,

staff of the extension service, personnel fromdevelopment projects, NGO members, regionalgovernment officials, and of course, theresearchers themselves should participate in theformulation of the regional program. Ideally, thegroup should not have more than 25 participantswith a balanced distribution of disciplines andperspectives.

Stepwise planning of regional research programs

Regional research programs can be planned ineight sequential steps. These steps combineanalytical processes based on knowledge fromvarious disciplines with more creative processesthat require the input of people with differentperspectives. The steps can be organized (but notnecessarily) around two or three workshops.These promote interaction among the differentparticipants.

Step 1: Regional review and analysis of regionaldevelopment objectives. This step provides acomprehensive overview of the region and itstechnological demands. The analysis normallyleads to a document that provides knowledge onthe development goals of the region. Thisdocument can be presented to the participants atthe start of the first workshop. Responsibility forthe regional analysis lays normally with asocioeconomist, who may also be a programmember. Though the analysis is normallyprepared by the program planning committee, itis shared and discussed by all participants in theplanning process. The regional review is usuallybased on secondary information. But it may becomplemented with rapid rural appraisals tocollect first-hand knowledge. The review shouldbe started three months before the first workshop,and will occupy a socioeconomist full time.

Step 2: Constraints analysis. In this step theinformation from the regional review and theknowledge of the participants in the planningprocess is combined to define the technologicalconstraints and opportunities for agriculturaldevelopment and sustainable resource use in theregion. This step ensures that the regionalprogram is based upon clearly defined andstrongly felt problems within the region, ratherthan on the scientific ideas of the researchers. Theresearch program leader normally takesresponsibility for this step. But its successfulcompletion is critically dependent on full-fledgedparticipation of the whole planning group. It isespecially important that the nonresearchersspeak out at this stage. Constraints analysis can

be enhanced by using constraints treemethodology. This is especially useful forregional research programs, since the constraintstree becomes a sort of holistic cross-section ofregional problems. Developing a completeconstraints tree takes one to one-and-a-half daysof the first workshop.

Step 3: Evaluating past research. Before anydecisions on research projects are taken, it isuseful to know what research has already beendone in the region, and what research has beendone elsewhere that is applicable to constraintsfaced within the region. Step 3 helps the researchprogram avoid duplication of previous work andincreases the effectiveness of research projects.The output of this step may be a document thatcan be presented in the first workshop. Normallya member of the program committee with atechnical background is responsible for the pastresearch evaluation, but this person is supportedby other researchers, both from within andoutside the regional program. Annual reports,scientific publications, and bibliographic searchesare useful sources of information. If there is amanagement information system, this may alsobe explored. Normally the evaluation of researchshould start three months before the firstworkshop, but it can only be finished after thefirst workshop.

Step 4: Defining research objectives. Once theconstraints facing the region’s agriculturaldevelopment are defined and an inventory hasbeen made of previous research achievements,the research program may define what it wishesto achieve during the planning period. Objectivesare most usefully defined at two levels: at thelevel of the overall regional research program andat the level of specific technological constraints.Research objectives provide answers to theconstraints identified earlier. In defining theseanswers, both researchers and other stakeholdersplay a major role. Researchers assess the technicalfeasibility of the objective (for example, is varietalresistance a more relevant objective than chemical

xii

control?). The other stakeholders assess theadoption potential of the technology that will beobtained if the objective is achieved (for example,will resistant varieties be adopted by farmersmore easily than chemical control methods?).Defining research objectives is the processthrough which the research program sets out itsmajor directions.

Step 5: Identifying research projects. In this step theresearchable constraints and the researchobjectives are matched with a research approach.A coherent set of indispensable research activitiesis defined. Each activity is required to attain theobjective. For each activity the required humanresources, the location, and the duration aredefined. Research projects do not need to bespelled out in great detail, but their overall shapeshould be determined at this stage. The programresearchers, with the guidance of the programleader, are principally responsible for identifyingresearch projects. They may do this in the timebetween the first and the second workshop. Twotypes of projects may be identified: technicalprojects that help to overcome specific constraintsin the region, be they in cropping, livestock,forestry, soil management, or any other field; andsupport projects that improve the scientific baseof the technical projects and thereby increase theirchance of success. Once projects are identified afirst screening may be made to define whether acertain project is best undertaken in the regionalprogram or in an other program type.

Step 6: Choosing priority research projects. Theresources available for the program do notusually allow for the implementation of all theprojects identified. Thus priorities forimplementation need to be defined. Theresponsibility for this step is with thesocioeconomist, who with the support of theprogram leader designs the approach for prioritysetting and obtains the required data. All theplanning group members contribute theirknowledge and judgment to the exercise. Whilemany methodologies are available for settingpriorities, in these guidelines we propose asimplified cost-benefit analysis, which integratessome elements of trend analysis and resourcevaluation. The socioeconomist requires at leastone month to prepare background data, and theworkshop participants will need one toone-and-a-half days in the second workshop toachieve results.

In priority setting for a regional program, thechallenge is to combine and compare research

benefits in many different dimensions. Forexample, one project (developing productivecombinations of yam and tree species) may affectcrop production and forestry production. Asecond project (introduction of forage species incrop rotations) may affect crop production,livestock production, and long-term soil quality.A third project (identification of improved peppervarieties) may only affect crop production. Apriority-setting method is required that combinesthe benefits to these different dimensions in asimple and straightforward manner. To do this,crude assumptions are made to estimate theexpected value of technological improvement bytrend analysis for crop, livestock, and forestryproduction and by resource valuation for soilquality. Workshop participants afterwards assessthe validity of these assumptions, and definewhat share of the expected technological progresswould arise from the regional program (asopposed to other types of research programs).

Once there is agreement on the benefits that theregional research program can expect to achievein each of the benefit dimensions identified asimportant, each proposed research project isassessed for its contribution to the total. Thefeasibility of research and the expected rate ofadoption for the results of each project areestimated as well. With these data the expectedimpact of each research project can be calculated.By comparing the expected impact with a costindication of each project, we can calculate whichprojects have the highest impact per unit of cost.The ranking that results is assessed and discussedby workshop participants in a plenary session.After modifications are made to account forconsiderations not included in the procedure, afinal rank order is accepted.

Step 7: Human resource gap analysis. Once thepriority research projects are defined, theprogram should assess whether it is equipped toimplement them. The main concern is that theprogram has the right disciplinary mix of people.The human resource requirements are aggregatedand compared with the human resourceavailability. Gaps are identified where staffshould be hired or contracted from other units.This process is led by the program committee,with support of the human resource department.The required data are available from the earliersteps (5 and 6, in particular). A matrix can bedrawn up that shows human resourcerequirements by projects in descending order ofpriority. This step does not require much time. It

xiii

can be finished in one to two weeks after thesecond workshop.

Step 8: Preparing for implementation. Finally, theconditions for successful implementation of theprogram need to be established. Key is that thesupport of senior management is obtained, andthat stakeholders endorse the program formally

and meaningfully. This is best done through ahalf-day workshop in which the finalizedprogram plan is presented to senior managementand other stakeholders. The plan should be madeavailable to them, in written form, two weeksbefore the workshop. This validation workshopcan be held four to six weeks after the secondworkshop.

Some final observations

Regional research programs are one possibleanswer to the changing needs of the presenttimes, in which stakeholder participation,problem solving, and resource managementconcerns are taking on major importance inagricultural research. In these guidelines weexpose a method for planning regional researchprograms. The method is not so much a finalproduct as starting point for thinking aboutregional research. Many questions still need to beresolved. The practice of regional research

programs will certainly help to provide answers.Some of these questions are: How can the resultsof regional research planning be used forpurposes other than research? How can the use ofgeographic information systems be improved?How can the time dimension be adequatelyaccounted for? How can the factual database forregional programs be improved? How shouldregional programs be linked with internationalecoregional research efforts?

xiv

Acronyms

GIS geographic information systems

IFDC International Fertilizer Development Center

INFORM Information for Agricultural Research Managers

INRA Institut National de Recherche Agricole (Morocco)

INRAB Institut National de la Recherche Agricole au Benin

IPM integrated pest management

ISNAR International Service for National Agricultural Research

KARI Kenya Agricultural Research Institute

NGO nongovernmental organization

NRM natural resource management

xv

PARTRationale, Background, and Issues 1

1.1 Why Formulate Regional Research Programs?

Three types of concerns favor the develop-ment of regional research programs. Firstly,

concerns over natural resource management havegained attention over the last decade in both de-veloped and developing countries. Many peopleworry that agriculture is exploiting resources inan unsustainable manner, or at the cost of nonag-ricultural users. These concerns have several com-mon factors (Crosson and Anderson, 1993;Janssen, 1995a). Many practitioners think the po-tential contribution of natural resources to devel-opment is not sufficiently understood, and thatavailable resources could be used more effec-tively. Also, there are concerns that the impact ofeconomic development on the natural resourcebase isn’t properly reflected in decision making.Here the worry is that present economic growthis based on resource degradation, which will con-strain future growth. In addition, there are con-cerns that the impact of resource use by one useron another (the question of so-called externalities)are not taken into account.

Agricultural research is reacting to these concernsby putting more emphasis on issues of natural re-source management (NRM). However, payingmore attention to NRM in an isolated mannerisn’t enough. Negative impacts on the natural re-source base may be the unintended effects ofwell-intentioned technological change in com-modity production. For example, high-yieldingvarieties may be developed with large nutrientabsorption needs. Also, NRM research cannot beisolated from other factors governing resourceuse. Products of isolated NRM research on hill-side soil erosion, for example, may not be effec-tive if they don’t provide short-term benefit to thefarmer. Such research might also neglect to ac-count for the policy and market context in whichthe farmer operates.

The successful integration of NRM issues in agri-cultural research thus requires a new researchperspective. A more holistic view is needed, onethat allows us to understand the potential for im-proving both the quality and the productivity ofagricultural resources within the context of non-agricultural resource use, the socioeconomic set-

ting, and long-term trends in production and re-source management.

Secondly, many countries are trying to increase thedemocratic content of their governments by de-centralizing public structures to the regional level.In these decentralized systems, citizens have in-fluence in deciding what the government under-takes in their region or city. People are closer totheir representatives and are more directly in-volved in public debates. They thereby obtain anincreased awareness of the utility and relevanceof the public sector. Agricultural research mustfollow this trend by allowing technology users alarger voice in the planning and execution of theresearch program.

A third development that favors regional researchprograms is the increased emphasis on adaptiveresearch. With the advent of improved communi-cation and information systems in the 1990s,countries’ ability to import and test new agricul-tural technologies has increased rapidly. As a re-sult, the relative attractiveness of adaptiveresearch over more strategic research has in-creased, certainly for those research areas inwhich the country does not choose to make a ma-jor investment. In addition, many developmentthinkers believe that strong adaptive researchprograms are the only way to guarantee that re-search results are effectively translated in usefultechnologies for end users.

At the level of the individual research project, sys-tems approaches and participatory technologydevelopment have become popular ways to pro-vide a decentralized, adaptive NRM perspective(van Duivenbooden, 1995). At the program level,the strategy has been to de-emphasize commod-ity research in favor of programs with a geo-graphic focus. Such programs are known underdifferent names. In Morocco, for example, theprogram for the semiarid zone is called an “agro-systems” program; similarities among productionsystems provide the common denominator(INRA, 1993). Among international agriculturalresearch centers, the term “ecoregional” is used,emphasizing that the boundaries of a region are

3

usefully defined by ecological characteristics. Be-nin, Burkina Faso, and Mali have set up “re-gional” programs, based on a mixture ofecological and administrative criteria.

These guidelines present a method for formulat-ing regional research programs with a mixed per-spective—one that combines productivity and

natural resource management issues within a re-gional, decentralized context. Based on earlierwork by Collion and Kissi (1994), it has been sub-stantially modified to account for the peculiaritiesof regional research planning. The method waspilot tested in Benin. It has also been applied inSenegal and Morocco.

1.2 What is a Region?

A region is defined here as a contiguous areawithin a country characterized by a certain

homogeneity in its agroecological, socioeconomic,and administrative characteristics. Elements thatmay define the homogeneity of a region are dis-cussed in section 1.5. It is important here to ob-serve that a country is normally made up ofseveral regions. This contrasts with the use of theword “region” in the international context, whereit normally refers to a grouping of countries in acertain part of the world (e.g., the Southern Coneof Latin America).

A country may be divided into a small number ofregions for the purpose of agricultural researchplanning. A large country with a large agricul-tural research system may be able to support asubstantial number of regional research pro-grams. In a small country, there will likely befewer regional programs. For example, in Beninthree regions were defined for the purposes of ag-ricultural research, whereas in Senegal there areseven.

The regional concept can be most easily applied ifagroecological, administrative, and socioeco-nomic conditions are highly associated, as is thecase in Benin. In other countries, regional pro-

grams may need to be defined in a different man-ner. For example, in mountainous countries theconcept of a region is more difficult to apply be-cause agroecological conditions change markedlywith the altitude. Across a transect of 50 kilome-ters three or four completely different ecosystemsmay be found. In this situation it may be difficultto reconcile administrative and agroecologicalboundaries in one regional classification. Theremay be a need to consider distinctive ecosystemswithin the region, which could lead to the plan-ning of sub-programs.

If several sub-programs were concerned withvery similar agroecologies, the decentralizationargument in favor of regional programs might beundone in favor of an ecosystem approach. Basedon this reasoning, Morocco used the methodologypresented within these guidelines to plan pro-grams in which the agroecological, and not theadministrative, dimension was dominant for de-fining program boundaries. The interaction of so-cioeconomic, administrative, and agroecologicalconstraints should be, again, emphasized. Manyresearch projects have failed precisely becausethey did not consider socioeconomic and admin-istrative conditions.

Guide to Program Planning and Priority Setting

4

1.3 Objectives and Features of Regional Research Programs

Regional research provides improved tech-nologies for a specific region. It aims to opti-

mize agricultural production and naturalresource management systems or exchange themfor more advanced ones. The overarching goal isto help improve people’s living standards in theregion, while ensuring maintenance of the naturalresource base.

Regional research programs contribute to agricul-tural development in a number of fields:

■ the rational exploitation of natural resources(water, soil, vegetation, genetic resources);

■ the integration of cropping, livestock, forestry,and possibly fisheries activities, and of the re-gion’s agricultural production systems;

■ the elimination of specific constraints in the re-gion’s cropping, livestock, forestry, and fisher-ies systems;

■ the institutional aspects of agricultural pro-duction and use of resources such as credit,marketing, extension, and producer organiza-tions.

Regional research programs take a different per-spective from commodity-oriented or production-factor research programs. Each of theseapproaches has its advantages and disadvan-tages. The challenge for national agricultural re-search systems is to find the optimum mix ofresearch approaches and program types. For re-gional research programs three features meritspecial mention: their NRM emphasis, proximityto users, and their problem-orientation.

NRM emphasis

In many parts of the world, agricultural produc-tivity and profitability are constrained more bythe quality of the available natural resources thanby the production potential of the crops beinggrown. Soil quality, water availability, and vegeta-tive cover strongly affect productive potential.Moreover, for many of the world’s populations,especially the poor, survival depends on exploit-ing natural resources. This is true whether forfarmers, fisherpeople, or cattle-herding nomads.

There is often competition for resources betweenagriculture and other uses. Examples of poten-tially competitive uses are water for irrigationversus human consumption, and land for forestcover versus cropping. Developing rural areasoften depends on the rational exploitation ofavailable resources. But this can only be achievedby adopting a holistic perspective on the variouskinds of resource use. Regional programs can pro-vide research with that perspective.

Production-factor research programs (such as asoil management research program) also have aresource emphasis. But since the resource’s role

within the overall production system is less ex-plicitly studied, such programs may be less effec-tive than regional programs in helping bringabout changes in resource management.

Regional research programs take into account thedifferent ways in which natural resources areused, by farmers and others, and the links be-tween resource use and resource quality. This is apotential strength, but also a potential weakness.Collaboration and integration of researcherswithin the program should be high, but that de-mands careful management of program re-sources. The integrated nature of regionalprograms can easily result in a lack of focus in re-search projects. In this case, the risk exists that theprogram may concentrate too much on diagnosisand understanding without arriving at testablesolutions.

To minimize problems of management and focus,it is vital that regional programs, even more sothan commodity and production-factor programs,be carefully formulated.

Rationale, Background, and Issues

5

Proximity to the user

Since regional research focuses on a well definedgeographic area, the problems of the area’s in-habitants can be clearly analyzed. The researchprogram should produce concrete, specific solu-tions to these problems. This is made easier be-cause the regional context offers enhancedopportunity to build participatory on-farm ap-proaches into the research program. Working atthe regional level, researchers can achieve moreintensive contact with farmers, other resource us-ers, government institutions, and developmentorganizations than they could at the nationallevel. Interaction on the problems to be addressedis therefore more straightforward.

In the past, such proximity to users often resultedin researchers being somewhat isolated from theinternational scientific community. There wassome uneasiness among researchers about the sci-

entific quality of their work and professional rec-ognition among peers. With present communica-tion and information media, especially fax,e-mail, CD-ROM, and the World Wide Web, andwith greater emphasis on NRM issues in the in-ternational community, this concern no longer ex-ists. Regional research should now be able toproduce results of similar scientific quality to thatof commodity programs, and with widely recog-nized relevance and value.

Whereas regional research programs may haveexcellent contact with resource users, commodityprograms will normally have better links withother agents in the commodity marketing chain,such as processors and traders. Once again, thisunderlines the fact that regional researchshouldn’t replace commodity research, but be ra-tionally combined with it.

Problem orientation

Regional research programs tend to be problemoriented. In this respect, they are somewhat of anextension of more traditional on-farm research ap-proaches. For example, the planning approach foron-farm research by Tripp and Woolley (1989)takes a stepwise approach comparable to the onepresented in these guidelines. But the frameworkof analysis in their approach focuses on the farm-ing instead of the regional system. For regional re-search programs the different farming systems ofthe region must be put in the context of the overalluse of natural resources. For solving the uniqueproblems that are then found, the adaptation oftechnologies available from other parts of thecountry or the world may be sufficient, in which

case there is no justification for duplicating stra-tegic research.

Nevertheless, regional problems may also requiremore sophisticated technological innovation meth-ods. This is the case when there are no adaptablesolutions available, or when it concerns a solutionto a particularly complex problem. In the secondcase, the best approach would not necessarily beto do upstream research relying on very advancedscientific methodologies, but to develop an inte-grated approach that allows for continuous andcomprehensive evaluation of the different impactsof the innovation. Such innovation is, again, bestmanaged in close interaction with users.

Organization of regional programs

Regional research can be organized in specificprograms, as will be elaborated in the remainderof this document. However, in cases where re-search is strictly organized along other lines, forexample, commodity programs or disciplinaryprograms, this regional approach to researchplanning could still be effective. In such a situ-ation, representatives of different programswould sit together to jointly review problems andopportunities in the region, and jointly plan andprioritize research interventions.

The main difference between this and a strict re-gional program would be that the control over thehuman resources, the equipment, and the moneywould not be with a regional program leader, butwould remain with the other program adminis-trators. Such a situation represents a compromisein which NRM, proximity to the user, and adap-tive concerns are taken seriously by the researchsystem leaders, but have not led to the decentrali-zation of decision making on research. The pre-sent guidelines can be used to plan such research,if some additional thought is given to the coordi-nation of the resulting research program.


6

1.4 The Basic Methodology for FormulatingResearch Programs

Our point of departure in laying out amethod for planning regional research is the

sequence of program-formulation steps describedby Collion and Kissi (1994). Readers’ knowledgeof that approach is, therefore, clearly an advan-tage. For those unfamiliar with the method, inthis section we provide an overview and a briefexplanation of how the method can be adapted toaccount for the integrated nature of regional pro-grams and the various types of benefits these pro-grams may produce.

Contributions of skills and knowledge from manydifferent people and disciplines are needed in or-der to formulate an agricultural research pro-

gram. However, combining these various types ofinformation, scientific disciplines, and humanskills can be a daunting task. Use of a stepwiseprocedure helps guide this process. In each step,new information is incorporated. The balance be-tween creative and analytical thinking maychange from step to step. In some steps, econom-ics skills are key to furthering the procedure; inother steps, technical skills from, say, agronomyor animal sciences are essential. Figure 1 presentsthe eight steps in agricultural research programformulation as initially outlined by Collion andKissi (1994) along with a way in which the stepscan be organized in a set of meetings.

1Review ofresearchdomain

2Constraints

analysis

4Determination

of researchobjectives

and strategy

5Identificationof research

projects

6Prioritysetting

7Human

resourcegap analysis

8Recommen-dations for

implementation

3Evaluationof existing

results

Workshop 1: Workshop 2: Workshop 3:Share and exchange information from theregional analysis and review results of pastresearch. Formulate a common problem andstrategy perspective

Source: Collion and Kissi, 1994, with modifications by the authors

Evaluate identified projectsaccording to priority

Internal and external validationsin order to arrive at broadagreement on the programto be implemented

Figure 1. Steps in research program formulation


7

Brief overview of the eight steps in program formulation

Step 1: Review of the research “domain.” Whilecertainly in the past, the research domain wasoften a commodity, in regional research programsit is a region within a country. The review of theresearch domain, therefore, consists of analyzingthe agricultural production and NRM situationwithin the region. It results in an overview of theconstraints on, and opportunities for, agriculturaldevelopment and improved management of natu-ral resources. It should also clarify the objectivesfor regional development, because it is to theseobjectives that agricultural research should re-spond.

Step 2: Constraints analysis. In this analysis, thereview of the region is further elaborated andconstraints to development are identified. For aregional research program, constraints relate notonly to production limitations, but also to market-ing problems, deficient social and institutional or-ganization, and poor management of water andland resources.

Step 3: Evaluation of existing research results. Toformulate an efficient and effective research pro-gram, it its important to know which problems orconstraints have already been researched andwhat the outcomes of that research were. Thatmeans not only successful research. Also unsuc-cessful efforts to solve a constraint through re-search are good to know, in order to avoidrepeating the failure.

Step 4: Determination of research objectives andstrategies. Once constraints are identified andpast research evaluated, an initial effort can bemade to define the objectives of the regional re-search program. These objectives guide decisionslater in the program formulation process.

Step 5: Identification of research projects. Re-search projects are the building blocks of the re-search program. In this step, projects are

identified to help overcome the constraints de-fined in step 2. Projects don’t have to be spelledout in great detail at this stage, but for each one,an objective should be formulated. Major researchactivities and their location also should be identi-fied in this step, and the human resources neededto do the work should be estimated.

Step 6: Priority setting among research projects.Typically, more projects are identified than can beimplemented. So the most important ones forachieving the research program’s objectives mustbe selected. To do so, evaluation criteria that takeinto account the special nature of a regional re-search program are established and elaboratedon.

Step 7: Human resource gap analysis. Once pri-ority projects are identified, the aggregate humanresources required to implement these projectsare estimated. By comparing required human re-sources with available human resources, it be-comes clear where the program needs to bestrengthened.

Step 8: Recommendations for implementation.Once project priorities have been set and humanresource requirements identified, the program isready for implementation. In this step, othermeasures needed to make the program opera-tional are defined and the conditions needed forthe program to succeed are described. Measuresto stimulate adoption of research results are alsodiscussed.

Stepwise program planning is strongly recom-mended at the start of a program. Typically, everyfive to seven years an update should be made.Updates can be done in a more continuous man-ner if the program is able to sustain a clear recordof how and why it has been changing its objec-tives and activities and if stakeholders are in-volved.

Participation

Formulating a research program is not only ananalytical task but also a creative one. The out-comes are strongly determined by who partici-pates. Therefore, program formulation shouldbenefit from a variety of perspectives and re-spond to the concerns of farmers and other natu-

ral resource users. The best way to bring togetherthese different perspectives is through workshopswhere representatives of all stakeholder groupsare present. Careful selection of participants,however, is crucial to the quality and relevance ofsuch events.


8

Wide participation brings another advantage.Once people have been involved in developing aplan, their willingness to support its implementa-

tion is higher. Participation, thus, improves boththe outcome of the program on paper and thechance of its being successfully executed.

Organization

The program formulation procedure requires aprogram committee to organize meetings andprepare the necessary background information.The committee is normally headed by the re-search program leader. Two other research pro-gram members should be on the committee, onewith a background in socioeconomics, the othertrained in agronomy. The fourth committee mem-ber should be a representative from the unit re-sponsible for research planning. This personcontributes knowledge of program formulationprocedures and ensures that the program underformulation is consistent with the overall direc-tion of the national agricultural research system.

The program planning exercise can be organizedaround three workshops. The first, over a periodof about three days, covers steps 1 to 4 of the pro-gram formulation process. It requires consider-able preparation, mainly because of the largevolume of information to be collected for steps 1and 3. The first workshop produces a set of re-search objectives formulated in the light of the re-gional environment, constraints on productionand NRM, and past research results.

Between the first and second workshops, researchprojects are identified by program committeemembers (step 5). Background data required toset priorities are collected and analyzed. The pro-

gram committee should also use this time to re-cord the preliminary results of the first four steps.

In the second workshop, lasting about two days,the research projects are prioritized. The overallpriority rankings of the various projects are dis-cussed and modifications are made to account forany considerations that could not be included inthe previous steps. Recommendations for imple-mentation are formulated.

After the second workshop, the program commit-tee estimates human resource requirements andassesses the gap between what is required andavailable. The committee writes the programdocument, which is then submitted to externalvalidation in a third workshop of approximatelyone day.

Essentially, the regional program planningmethod analyzes what needs to be done (steps 1through 4), formulates creative solutions (step 5),then evaluates, rationalizes, and finalizes the pro-gram (steps 6 through 8). This is an over-simplifi-cation of the procedure, but it underlines the“analyze-formulate-evaluate” sequence necessaryto plan any major human undertaking. Before wedescribe the method step by step, we should dis-cuss the necessary preparations and issues tokeep in mind before starting the planning process.

1.5 Preparing for Regional Research Program Planning

Four very different issues must be resolved be-fore regional programs can be adequately for-

mulated. The first is strategic in nature. Itconcerns the role of regional research programswithin the national research plan. The planning of

regional programs depends explicitly on the na-tional agricultural research strategy.

The second issue is the geographic focus of the re-gional research program. It is vital to defineclearly the boundaries of the zone being served.


9

The third issue is participation in the programformulation exercise. Although the proposed pro-cedure tries to maximize objectivity, we should beaware that the more creative elements of programformulation, such as designing research projectsin response to constraints, are highly subjective.The outcomes and credibility of the program for-mulated, as well as the program’s prospects forbeing implemented, are strongly influenced by

which stakeholders have participated and whichresearch disciplines have been represented.

The fourth issue relates to ease of communication.In a participatory procedure, some people inter-pret the same words differently. To avoid misun-derstandings, it’s useful to define and adopt acommon vocabulary at the outset. One guide toappropriate definitions of important terms is theglossary starting on page 62 of this book.

Placement within the national agricultural research plan

Regional research programs cannot substitute forcommodity or production-factor programs; theycomplement them. They do provide an effectiveNRM perspective and may be very close to theuser. However, their orientation is less discipli-nary and they cannot take into account consump-tion and marketing issues as well as commodityprograms (table 1). The challenge is to combineseveral research themes in the national researchplan and build on the respective strengths of each.

Regional research programs may well exist along-side commodity or production-factor programs.In such cases the objectives and responsibilities ofeach type of program should be clearly defined.For example, in the development of Benin’s agri-cultural research plan, although regional researchprograms were considered an important answerto concerns about NRM and user participation,they were still combined with programs that weremore discipline oriented or commodity based.Box 1 provides a summary of the program struc-ture proposed in Benin.

After the strategic orientation of various researchprograms within the agricultural research planhas been defined, more specific interactions be-tween the regional program and other programscan be identified. Here, a key point is to decide onwhich research each program will concentrate.This decision is based on two factors: the suitabil-ity of program types for certain research hemes,and the relative importance assigned to differentprogram types. Table 2 indicates the suitability ofprogram types for different research themes.

Before program formulation starts, it’s importantto consider how research themes will be distrib-uted over program types. There are three mainreasons for doing this. First, it avoids duplicationof effort in developing the program. Second, ithelps managers decide which program type ismost appropriate for a particular research projectand transfer projects to the most appropriate“home”. Third, clearly defining at the outset theresearch themes to be used in each program typecan reduce the need for coordination once theprogram starts. Box 2 shows the initial distribu-tion of research themes adopted in Benin.

highly user oriented

highly perceptive to NRM issues

fairly well recognized disciplinary quality

low market orientation

fairly user oriented

not perceptive to NRM issues

well recognized disciplinary quality

highly market oriented

Regional Programs Commodity Programs

Table 1. Relative strengths of regional and commodity research programs

Defining the region

Three sets of criteria—agroecological, socioeco-nomic, and administrative—can be used to set theboundaries of the region covered by the regionalresearch program. Agroecological criteria define a

target area that is relatively homogeneous in itsagricultural and resource management patterns.Socioeconomic criteria define homogeneity ineven more detail (for example, by incorporating


10

Box 1: Program structure proposed in Benin’s agricultural research plan

Regional research programs

Southern RegionConstraints: Low fertility, sedimentation, limited diversificationObjective: Increase income-generating potential of natural resourcesRelative importance: High

Central RegionConstraints: Low fertility, deforestation, poor cropping practicesObjective: Consolidate productive potentialRelative importance: Medium

Northern RegionConstraints: Deforestation, erosion, poor cropping practices, but ample land availableObjective: Develop socially and ecologically sustainable production systemsRelative importance: High

Commodity research programs

Staple foodsConstraints: Breeding and plant protection problemsObjective: Make improved varieties available to different production regionsRelative importance: Medium

CottonConstraints: Poor cropping practices but high export potentialObjective: Improve long-term profitabilityRelative importance: Low

Oil palmConstraints: Poor integration in production systems, few traditional usesObjective: Strengthen role in village economyRelative importance: Low

Nontraditional exportsConstraints: Low exploitation of favorable exchange rate and emerging commercial sectorObjective: Support development of a trading sectorRelative importance: Medium

Disciplinary research programs

Postharvest technologyConstraints: Postharvest lossesObjective: Increase value added to agricultural productionRelative importance: High

Policy analysisConstraints: Funding shortages brought about by structural adjustmentObjective: Contribute to improved policy environmentRelative importance: Medium

Source: Plan Directeur de la Recherche Agricole au Bénin, Vol. 1.

factors such as market access or farm size, or bydistinguishing ethnic groups). Administrative cri-teria ensure that the area of intervention for re-search is congruent with that of other public andprivate interventions, such as extension, credit,and rural development programs. It is difficult tojudge the relative importance of these three sets ofcriteria, but it appears that administrativeboundaries have often received insufficient atten-tion (Janssen, 1995b).

Agroecological criteria

The target area of a regional program should berelatively homogeneous with regard to climate,

soil, natural resource endowment, and land use.Climatic uniformity can be defined in terms ofmean temperature (or average minimum andmaximum temperatures) and rainfall pattern(millimeters per year and distribution). However,other parameters may also be relevant, such asevapotranspiration, insulation, hydric deficit, oraverage difference between day and night tem-perature. The choice of the key parameters shouldbe left to local experts.

Soil uniformity can be defined in terms of soiltype (clay or sand), topography (flat, gently slop-ing, mountainous), soil depth (shallow, intermedi-


11

ate, deep), and soil fertility. In special cases othersoil factors may be included.

Natural resource endowment concerns the avail-ability of water resources, the proportion of thearea covered by primary or secondary forest, thediversity of plant species, and the amount of landavailable per inhabitant.

Land-use criteria incorporate the human element.They show how people respond to climate, soil,and natural resource endowments within the ex-isting institutional and socioeconomic environ-ment. Land-use criteria include variables such as

production systems, the role of livestock, crop-ping and forestry activities, rotation and shiftingcultivation practices, land management practices(such as fertilization and erosion control).

Because land-use criteria reflect human reactionsto existing agroecological conditions, they are agood way to understand which agroecological cri-teria most strongly influence de facto variabilitybetween regions. Where land-use data do exist,they may supply the most suitable criteria for set-ting agroecological boundaries. Unfortunately,such information is often not available in suffi-cient detail to serve this purpose.

Plant genetic improvement or animal breedingCrop protection or animal healthAgronomy or animal husbandryPostharvest researchDiagnostic studiesPolicy researchInstitutional strengtheningResource managementResource classification

−±

−

±

±

+

+

++

Research Theme Program Type

RegionalProgram

CommodityProgram

Production-FactorProgram 1

DisciplinaryProgram 2

+±±±±±±−−

−−±−±−−±+

−−−+±+±−−

Key: + very suitable, somewhat suitable, not suitable.

1A production-factor program works on the different dimensions of a production factor, for example, soil improvement, watermanagement, or machinery use.2A disciplinary program is made up of staff with the same educational background. For example, most policy programs mainlyemploy economists; most postharvest research programs mainly employ food technologists.

± −

Table 2. Suitability of program types for selected research themes

Box 2: Importance of program types for various research themes in Benin

The West African country of Benin, with apopulation of five million, has a relatively smallagricultural research system. It has no produc-tion-factor research programs (such as soilquality or water management) and livestock re-search was recently absorbed into regional re-search programs. As we saw in box 1, Beninuses three research program types: regionalresearch programs (south center, and north),commodity research programs (staple foods,cotton, oil palm, and nontraditional exports),and disciplinary programs (postharvest technol-ogy and policy analysis programs). The discipli-nary programs are dominated, though notexclusively, by a certain discipline (in this ex-ample, food technologists and economists re-spectively). The regional programs were givenhigh or medium priority.

As the table shows, the regional research programs are of principal importance for agronomy, animal health and hus-bandry, diagnostic studies, and resource management. It is of secondary importance for work in animal breeding, insti-tutional strengthening, and resource classification. For plant genetic improvement, crop protection, and policy researchit is of only occasional importance.

Plant genetic improvementCrop protectionAgronomyAnimal breedingAnimal healthAnimal husbandryPostharvest researchDiagnostic studiesPolicy researchInstitutional strengtheningResource managementResource classification

Research Theme Program Type

RegionalProgram

DisciplinaryProgram

CommodityProgram

**

***********

***********

*******

***********

More asterisks indicate increasing importance.


12

Socioeconomic criteria

Population density and farm size, income level,market access, urbanization, ethnic composition,and gender are important socioeconomic criteriafor defining regions. In densely populated regionsthe demand for agricultural technology is nor-mally geared toward innovations that improveland productivity. In sparsely populated areas theemphasis is normally on technologies that im-prove labor productivity. The effect of farm sizeon technology demand tends to be similar to thatof population density.

Income levels also create large differences in tech-nology demand. With increasing income levels,consumption tends to shift from cereals to animalproducts and other “high-quality” food. Also, theminimum increase in labor productivity that anew technology must deliver in order to be con-sidered by farmers for adoption increases with in-come.

Market access defines the extent to which farmersare interested in new commercial options. Al-though nearly every farmer in the world is now incontact with the market, some get their produceto market readily at low cost, while others cannot.For farmers with poor market access, technolo-gies that reduce input requirements or focus oncommodities with a high value-to-volume ratioare most attractive. Farmers with good market ac-cess are likely to be interested in technologies thatimprove product quality and, hence, price.

Urbanization is closely linked to market access. Inhighly urbanized areas, agriculture tends to bemore commercialized. At the same time, farmersmay have access to alternative income sources,which improves their capacity to finance agricul-tural investments. Urbanization also changestheir long-term perspective in that they may notfeel as dependent on the farm for their future live-lihood as do people in largely rural areas.

Ethnic divisions and gender are important be-cause they may be associated with traditional dif-ferences in land use and labor distribution. Theymay therefore require alternative approaches toon-farm research and diffusion strategies (for in-stance, related to language, sex of the on-farm re-searcher, or timing of visits).

Administrative criteria

Most countries have different layers of politicaladministration. Large countries such as India, Ni-geria, Brazil, Mexico, and China are first divided

into states. Most countries (or states within bigcountries) are divided into provinces or depart-ments, and those are further divided into munici-palities. Since these divisions normally form theframework for implementing public policies, de-veloping private business strategies, or passinglaws, the effectiveness of a regional research pro-gram is enhanced by respecting administrativeborders.

Combining the three sets of criteria

Geographic information systems (GIS) can beused to overlay geographically referenced data onsuch criteria as those just described. GIS is a use-ful tool for animating a discussion of the criteriafor defining a region (Wood and Pardey, 1994).With GIS, it’s possible to distinguish, for instance,low rainfall areas (say below 600 millimeters peryear) from high rainfall areas (more than 600 mil-limeters per year) and areas of low populationdensity (e.g., less than 50 persons per square kil-liometer) from areas of high density (more than50 persons per square killometer). Figure 2 is anexample of a data overlay. It is based on work bythe priority-setting working group of the KenyaAgricultural Research Institute (KARI, 1995).

Overlays allow relatively homogeneous regionsto be defined. The quality of the resulting regionaldelineation depends, of course, on the relevanceof the criteria. For example, is 600 millimeters peryear the amount of rainfall at which cropping pat-terns tend to change? Is 50 persons per square kil-lometer the threshold at which productionsystems become more intensive?

In defining target zones for regional programs,the challenge is to arrive at a reasonable numberof zones that together cover the whole country.Even for a medium-sized country like Morocco, itwould be difficult to envisage more than eight re-gional programs. But GIS makes it possible tooverlay so many types of data that thousands ofsupposedly “homogeneous” microzones are de-fined. Only the most important criteria from thelist given above are selected. And these are care-fully interpreted and applied. The principal crite-ria for zonification may be defined by identifyingthose criteria that change simultaneously acrossthe country. For example, when density of popu-lation, soil fertility, rainfall regimes, market ac-cess, and income levels all change along the samegradient, this may be an important division line.Information from GIS should, therefore, be com-bined with the assessment of a group of experts,who judge actual similarity across criteria. In ourjudgment, it is also wise to review what zonation


13

has been successfully used in the past or is in useby other organizations before pursuing a newone.

Clear definition of regional boundaries is a stepthat must precede regional program formulation.Questions should be asked as to whether a coun-try’s research system can adequately cover all the

different regions. Box 3 describes the zonationprocess carried out in Benin.

The regional boundaries that are finally acceptedshould also make intuitive sense. Otherwise therewill be lingering doubts and unfocused discus-sion among participants during the program for-mulation exercise. Hence, we re-emphasize howimportant it is to spell out clearly the regional

Box 3: Defining target zones for Benin’s regional research programs

In Benin, five agroecological zones were defined by thenational institute for agricultural research (INRAB). Theprincipal criterion for zonation was rainfall. Whereas thesouthern zone (zone 1) is characterized by abundant rain-fall distributed over two seasons per year, the extremenorth (zone 4) has a semiarid Sahelian climate, with onewet season. The Atacora region (zone 5) was separatedfrom the extreme north because the presence of a moun-tain ridge changes rainfall and soil conditions. In the Zouregion (zone 2), the bimodal pattern changes to a unimo-dal pattern, and the south of Borgou (zone 3) has a uni-modal pattern, like that in the north but less dry.

The initial agroecological classification in Benin also tookinto account population density and land availability be-cause the south is almost 10 times as densely populatedas the north. The borders of the agroecological zones andof the administrative departments did not differ greatly.The five agroecological zones were thus good targets forregional programs. However, for practical reasons (par-ticularly the absence of research infrastructure in thenorth), INRAB decided to initially merge zones 3, 4 and 5.

Departmental limitsNatural zones:

BORGOU

ATACORA

ZOU

ATLANTIQ

UEMONO OUÉMÉ

5

4

3

2

1

1 South2 Center3 South Borgou and Atacora4 North Borgou5 Atacora

B

A

1

2

B1

A1 A2

B2

Population density

A = high population densityB = low population density

Climate

1 = low rainfall2 = High rainfall

Homogeneous regions

Figure 2. Identification of homogenous regions


14

boundaries before program formulation begins.Misunderstandings arising from different inter-pretations about the regional program’s coverage

could seriously delay progress or render other-wise fruitful discussions useless.

Participants in regional program formulation

A regional program, by its very nature, must takeregional development needs into account. As alsopointed out by Mills and Karanja (1995), partici-pants in program planning must not only beaware of such needs. They must be able to articu-late them. It is therefore prudent to select repre-sentatives from the following groups:

■ Farmers. In cases where producers are repre-sented by farmer organizations, leaders ofthese groups should be invited to participate.But it shouldn’t be forgotten that the charac-teristics and needs of farmers can vary greatlyfrom region to region. In the case of southernBenin, invitations were extended to organiza-tions representing the interests of farmers andfisherpeople. If no farmer organizations oper-ate in a region, technical staff from the exten-sion service can be asked to participate.

■ Extension services. These are the main part-ners of the regional research program, espe-cially when the aim is to ensure the potentialimpact of new technologies is maximized. Fur-thermore, extensionists usually have a goodunderstanding of their region’s problems.

■ Development agencies or projects. Regionalresearch programs provide a suitable frame-work for directly integrating research intoother development efforts. The presence of on-going development projects can help identifythe most appropriate research activities,avoiding ones that may have been motivatedby other considerations.

■ NGOs. Nongovernmental organizations oftenwork on highly practical projects and encour-age direct contact with end users. So they arekey potential allies for research programmingpurposes.

■ Regional government. Representatives fromregional government play a major role in de-fining research objectives and help ensure thatresearch harmonizes with other developmentwork. They can also advise on regulatory mat-ters, explaining the potential legal conse-quences of certain activities in NRM.

■ Researchers. The success of a research pro-gram depends on how motivated researchersare and on their ability to compromise. Identi-fying and developing projects always de-mands the input of experienced researchers.

The participants should also represent a cross sec-tion of disciplines. Agricultural production sci-ences, veterinary sciences, forestry, soil sciences,agricultural and resource economics, sociology ororganization sciences, and possibly, law are alldisciplines that might usefully contribute to for-mulating the research plan. Normally it isn’t pos-sible to combine all these, but for the sake of thequality of the exercise it is important that there besubstantial disciplinary diversity. This prevents asingle way of thinking from dominating discus-sions and ensures that those people without auniversity background are not silenced by thecommon jargon of one or a few key disciplines.(Where many disciplines are present, there is stilla chance of nonacademics being overwhelmed bya more generic scientific vocabulary, but it is sub-stantially reduced.)

The number of participants in the three programformulation workshops should be kept between15 and 25. With more than 25, communication isdifficult and costs escalate. With less than 15, thediversity of opinion is probably less than optimal.The same people should be invited to the first andsecond workshops.

Sharing the same language

For most participants, formulating a regional re-search program is an unfamiliar task. Yet theirknowledge and experience is the foundation of agood program. If everyone is to communicate ef-fectively and make clear, concise contributions to

the discussions, the meaning of a few basic termsmust be uniformly agreed on. This makes the jobof the program formulation leader much easier,allowing the group to be guided smoothlythrough the different steps.


15

Here we describe only a few key terms. A moreextensive glossary is provided at the end of thebook.

Program: The set of research projects to be under-taken in response to regional development objec-tives and technology demand by user groups. Theaim of the research program is to improve the useof resources available. A program is normallymultidisciplinary. The composition is flexible andresponds to observed technological demand.

Constraints: The improved use of the region’s re-sources is normally obstructed by a set of con-straints. These can be technical (for example,cereal varieties not resistant to a certain disease),physical (low soil fertility), socioeconomic (smallfarm size), or institutional (poor access to credit).The research program normally aims to eliminateor reduce the effects of such constraints. Con-straints may usefully be merged with opportuni-ties. In the case of an opportunity, there is not somuch a limitation to be removed as an advantageto be exploited.

Research project: A coherent set of experimentsand studies necessary to accomplish a goal, whichis normally to remove one of the identified con-straints. A research project may aim to develop atechnology or methodology and is often executed

by a group of researchers with different discipli-nary backgrounds. Research projects are thebuilding blocks of the program and provide theoccasion for multidisciplinary integration. Thenumber of research projects in a program is nor-mally limited to two or three projects per re-searcher. At the level of program formulation, aproject is characterized by a goal, a set of researchactivities, a timetable, an expected result, and anestimate of time requirements by discipline. Oncea project is selected for implementation, it can bedescribed in more detail. Then, monitoring andevaluation parameters should be defined, budgetsdrawn up, and annual workloads planned.

Activity: A project is made up of activities—indi-vidual experiments and studies that aim to gener-ate part of the knowledge required to solve theconstraint being addressed. Normally the activi-ties are monodisciplinary and are undertaken byone researcher. A project normally has more thantwo but less than eight activities.

Participants in program formulation should un-derstand that a program is made up of projects,that projects are composed of activities, and thatprojects respond to constraints. Once they areclear on this, they have the basic tools to contrib-ute efficiently to discussions and decision making.


16

PARTA Stepwise Procedure for Crafting aRegional Research Program

2

Detail of Step 1 of the Program Formulation Process, ‘Review of the Research Domain’

RationaleFormulation of a coherent research plan requires a com-prehensive overview of the region

OutputWritten and oral presentation on the region and its devel-opment objectives

ResponsibilitySocioeconomist from program or a social sciences unit

ParticipantsA program committee prepares the information, which islater discussed by a broad group of workshop participants

Information needsSocioeconomic and biophysical variables about the region are collected, as well as descriptions of production andresource management systems

MethodsSecondary data analysis, principally from statistical sources and earlier diagnostic work. This may be complementedby rapid participatory appraisals on specific issues

Time and resourcesReview should start roughly three months before first workshop. It will be a full-time job for the person responsible.Additionally, a small budget to allow data collection may be required


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results

Workshop 1: Workshop 2: Workshop 3:


2.1 Regional Review and Analysis of RegionalDevelopment Objectives

Step 1 of the program formulation process isthe ‘review of the research domain’. For a re-

gional research program, the domain is the re-gion. Most participants in program planning willhave only a partial understanding of the regionand its problems. They may know certain munici-palities much better than others, or certain activi-ties (e.g., cropping or cattle production) betterthan others. Although their personal knowledgeis key to an insightful research plan, it must beput in perspective by a “cool” assessment of re-gional characteristics.

A list of key variables that may be assembled in aregional review is provided in box 4. As discussedin Part 1, the boundaries of the region should bynow be clearly established.

The regional review also presents opportunity toensure that research is integrated with other de-velopment activities. It may show, for example,

that local government wants to develop irrigationin a set of municipalities. Or that the road systemis being upgraded, thereby creating opportunitiesfor marketable commodities. The review shouldprovide a wider context for analysis of specificproblems to be identified later. For example, itmay identify drought or land-tenure systems askey constraints to agricultural development, ormake it clear that diversification and addingvalue are key development strategies of the min-istry of agriculture.

Further, the regional review provides institutionalinformation. It may, for example, highlight theimpact of new legislation concerning markets orland ownership. It may identify constraints in thecredit system. And it may analyze the institu-tional setting, identifying potential collaboratorsfor agricultural research.

19

The regional review is a key way to ensure thatopportunities aren’t overlooked. Most of the dataneeded can be collected before the first workshop

by the program committee members. Informationmay come from regional strategy papers, nationalstatistics, or other sources.

Analysis of regional development strategies

Regional development objectives and strategiesare often derived from national objectives. If theregion is a homogenous political unit, like a prov-ince, regional objectives and strategies may al-ready exist as part of a regional developmentplan. In this case, they can be used directly in theresearch planning exercise. Where the region con-sists of several political units, the objectives andstrategies of each one should be compiled. Thismay require some preparation.

If no regional development objectives and strate-gies have been explicitly defined, a dialoguewithin the planning group may be initiated to de-fine them. The following checklist of national ob-jectives, and how they can be translated intoregional development objectives, strategies, and

actions, may then be helpful (see also Boughton etal. 1995):

■ Improving rural incomes. National develop-ment objectives such as “improving rural in-comes” are often translated into regionalstrategies like agricultural intensification, di-versification, enhanced postharvest process-ing, or market liberalization.

■ Contributing to balance of payments. Contri-butions to balance of payments may be ex-pressed at the regional level in strategies tofocus extension services, credit facilities, or in-put supply on export crops (existing or new),or as strategies that aim to reduce input use.

■ Improving food security. At the regionallevel, food security objectives are often ad-dressed by improving the potential yield of

Box 4: Information to be collected in a regional review

SocioeconomicPopulation size and rate of growth, possibly by ethnic groupDegree of urbanization and rate of urban growthIncome per capita and distribution, by income strata and sectorContribution of region to national economyContribution of agriculture to regional economyRelative importance of cropping, livestock, fisheries, forestry, and other subsectorsEmployment situation, by sex and sector (primary, secondary, tertiary)Farm size and distributionPrincipal agricultural commodities and production systemsLand and water usePossible target groups for agricultural researchPrincipal constraints on, and opportunities for, regional development

BiophysicalGeographic boundaries of the regionArea and topographySoil types, plus water, forest, and wildlife resourcesClimateVegetative cover

InstitutionalRegional development objectives and strategiesCompetency of regional government bodiesFunding availability for research and development at the regional levelCharacterization of agricultural support services (credit, extension)Market structureInput availability and costsDevelopment projectsDevelopment organizations (public agencies and NGOs)Tenure arrangements


20

staple foods, reducing production losses inyears with poor production conditions, devel-oping better storage capacity, or improvingsales opportunities for a possible marketablesurplus. In regions with a land surplus, devel-oping the frontier may also be identified as aregional food security objective.

■ Improving living conditions of the poor. Byfocusing development projects or infrastruc-ture development on regions with many poorpeople, an equity objective can be pursued.The equity objective may also lead to nonre-search solutions, such as employment pro-grams (like food for work).

■ Managing natural resources. Sustainablemanagement of natural resources may trans-late into reforestation projects, anti-erosion

legislation and enforcement, watershed man-agement, and the establishment of nature re-serves.

In the south of Benin, regional objectives had notbeen clearly defined in a national developmentplan. The research planning group thus formu-lated the following regional development strate-gies:

■ intensify agricultural production

■ generate employment in agriculture-relatedactivities

■ reduce degradation of the natural resourcebase

A summary of the regional review for Benin’sCentral Region is presented in box 5.

2.2 Constraints Analysis

In step 2 of the program formulation exercise, abroad-based group of research stakeholders

identifies the central constraints or problems fac-ing the region’s sustainable agricultural develop-

Detail of Step 2 of the Program Formulation Process, ‘Constraints Analysis’

RationaleRegional research programs should respond to problemsand opportunities clearly felt within the region

OutputCoherent set of principal constraints and their interactions,as well as a definition of the central constraint to be ad-dressed

ResponsibilityProgram leader and planning facilitator

ParticipantsBroad selection of research stakeholders. Special attentionshould be paid to “nonresearch” partners, for example, farmers, other resource users, and development projectpersonnel

Information needsSector review upon which to base initial analysis, as well as the common judgment of participants to identify and re-late constraints to agricultural development

Time and resourcesFree brainstorming followed by construction of a constraints tree. The workshop should last one to one-and-a-halfdays for all workshop participants. Time is spent partly in working groups


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results



Identifying Constraints, Objectives, and Research Projects

21

ment and resource management. Much of thiswork is accomplished at a workshop. There, theparticipants construct a ‘constraints tree’, whichtakes the form of a flowchart. All participants areasked to write a statement of what they considerto be the central constraint or problem facing theregion. The statements are then compared andsynthesized into one central constraint. Here arethree examples:

■ limited capacity to generate income with theexisting natural resource base

■ absence of sustainable and profitable farmingsystems

■ low incomes among agricultural producers

The participants’ initial statements are used tostructure the top branches of the constraints tree.

Box 5: Summary of the regional review for Benin’s Central Region

Benin’s Central Region is made up of the municipalitiesOuèssè, Savalou, Banté, Djidja, Dassa, Glazoué, Savè,Kétou, Aplahoué, Bassila, and the south of Tchaourou(see map). It covers 15,215 square kilometers and repre-sents a zone of transition from two rainy seasons to one.As a result, there is a season of stable rainfall betweenApril and July and a season of unstable rainfall betweenOctober and January. Average rainfall is 1100 millime-ters, but may vary from year to year between 700 and1400 millimeters. The average temperature is 27° Cel-sius.

The region has a modestly undulating topography, withsome hilltops rising above the rest. Several soil types arefound, including ferralitic soil, black clay soils, and hydro-morphous soils. All are losing their natural fertility. Some2000 square kilometers are forested, much of it withplanted forests. Wildlife diversity has greatly sufferedover the years. Water resources are relatively abundant,since the region is crossed by the Ouémé river and two ofits tributaries. Because of extensive slash and burn, theSavannah-type vegetative cover that has dominatedmost of the region is gradually disappearing. More andmore land is found in poorly recovering fallow systems.

The Central Region has a population of 400,000, dividedamong three ethnic groups. Eighty-five percent is ruraland almost all depend on agriculture for their livelihood.Rural population growth is minimal. Growth is concen-trated in the urban centers, but also in the large cities ofthe adjacent Southern Region. In turn, the region re-ceives migrants from the densely populated rural areas inthe south of the country. The rural population is poor, withincomes less than half the national average.

The region’s contribution to the national economy is basi-cally through its agricultural production. Cotton productionshould be singled out because the crop accounts for thelion’s share of international trade earnings. Within the ag-ricultural and natural resource management sector, cropsaccount for 72 percent, livestock for seven percent, andforestry products for 21 percent of total value. Maize andyam are the principal food crops.

Many of Benin’s farmers are women. At three hectaresper agricultural worker, there is ample land, but labor pro-ductivity is low. The principal issue is not where to findemployment for people in agriculture but how to increaseproductivity. In the existing land-tenure situation, it is diffi-cult and most probably irrelevant to distinguish individual

farms and to calculate their size. Open-access land is stillvery important and is often a source of conflict betweenseminomadic cattle herders and farmers. Agricultural re-search has to consider the village as the focus for cropresearch and development activities, but livestock re-search should target nomadic groups as well as villages.

Until recently, agricultural support services were largelygeared to cotton production. There is an unstructuredmarket, but nevertheless, price arbitration is very effi-cient.

Several development projects are active in the region, in-itiated by the regional agricultural office or by NGOs andbilateral donors. The key objective expressed is to exploitthe region’s large agricultural potential in a sustainablemanner. Industrial and urban development do not yetplay a major role in the region. A more complete descrip-tion of the region appears in the Plan Directeur de la Re-

Djidja

Savalou

Kétou

Dassa

Bantè

Glazoué

Savè

Bassila

Sud Tchaourou

Ouèssè

Apl

ahou

é

Togo

Nigeria

Central Region limitDepartmental limit


22

Working groups are then formed, each to developone of the top “branches” further. Slowly a chainof causality is built. At each stage participants askthemselves what are the causes of a particularconstraint, then, what is causing the causes of theconstraint, and so on. Finally, they have brokendown the central constraint into problems that areconcrete enough to be translated into unambigu-ous objectives for a research or development pro-ject (e.g., inappropriate fertilization or absence ofa road). Constraints identified may also be com-pletely exogenous (e.g., falling world marketprices). Once the working groups have finalizedtheir branches the whole constraints tree is assem-bled and checked for internal logic and consis-tency.

The same constraints often appear at differentplaces in the overall tree or are associated withdifferent causes and effects. By checking the con-sistency of the tree, some of the duplicates can beremoved, though some will remain simply be-cause one problem may have many effects. This isa reflection of the highly complex and integrated

nature of agriculture and natural resource man-agement.

The use of a constraints tree promotes an inte-grated perspective on the region’s problems.However, analysis using such a flowchart is not afull substitute for systems analysis because it con-centrates only on constraints and not on other in-teractions. But because it maps outcause-and-effect relationships, the tree serves as a“constraint cross section” of the region’s agricul-tural and natural resource management system.

For further details on how to develop a con-straints tree, see Collion and Kissi (1994) or GTZ(1988).

The first constraints tree in box 6 shows the high-est levels of constraints identified for Benin’s Cen-tral Region. Below, the branch on ‘poormanagement of production systems’ is further de-veloped. The constraints at the lowest level of theconstraints tree are shaded.

Categories of constraints

Constraints on the sustainable development of aregion’s agriculture and on the management of itsnatural resources generally fall into six categories:

Degradation of natural resources. That is, water,soil, vegetative cover, plant genetic base, andwildlife due to production practices and the so-ciocultural milieu.

Land management. As it relates to type of exploi-tation, land-tenure policy, water control, crop-ping, livestock rearing, fishing, forestry practices,empowerment of local people, and the influenceof cultural traditions on people’s behavior.

Production systems. In relation to species, varie-ties, and breeds; crop pests; contagious animaldiseases; cropping, herding, fishing, and forestrypractices; and processing and marketing of prod-ucts.

Integration of production systems. Specifically,poor integration of crops and livestock eitherwithin a single farming operation or among dif-ferent types of farms, the lack of interaction be-tween perennial and annual crops and amongannual crops, poor links between components ofthe regional agricultural system.

Organization of commodity subsectors. Relatedto the problems of poorly developed markets andinsufficient information on market conditions,storage facilities, and the availability of qualityseed.

Integration of agriculture within regional socio-economic institutions. This constraint categoryincludes problems related to credit, farmer or-ganizations, coordination of various developmentagents, input supply, and the role of women in re-gional agricultural development.

Many of the problems within these categoriescan’t be solved through agricultural research.They depend on other development initiatives,such as creating infrastructure or improvingcredit, input supplies, education, and health serv-ices. There may actually be more nonresearchableconstraints than researchable ones, which under-lines the close connections between research anddevelopment at the regional level. Box 7 lists thenonresearchable constraints for livestock produc-tion and subsector development identified in theresearch plan for Benin’s Central Region.

It is important to share such information with allparties able to address the constraints. Many such


23

Box 6: Portions of the constraints tree developed for Benin’s Central Region research program

In the analysis of Benin’s Central Region, two principal constraints were identified as underlying the central constraintof low income. These are low profitability of agricultural activities and natural resource degradation. The low profitabilityproblem was, in turn, attributed to poorly developed commodity subsectors, low crop productivity, and low livestock andfisheries productivity. Similarly, natural resource degradation was attributed to poor management of natural resourcesand poor management and integration of production systems. The underlying causes of poor management of produc-tion systems are developed below. Lowest level constraints are those in the shaded boxes below.

Poor managementof production

systems

Poor managementof soil fertility

Improper crop rotation

Low usage of mineralfertilizers

Low usage of improvedvarieties

Lack of agriculturalknowledge

Land availability notproperly takenadvantage of

Recommended inputlevels not adhered to

Poor rainfalldistribution

Animal tractionequipment not

profitable

Low usage of organicfertilizers

Lack of training andinformation on

organic fertilizertechnology

Lack of awareness ofthe benefits of

organic fertilizermanagement

Lack of informationon optimization of

animal tractionequipment

Lack of varietiesadapted to

rainfall patterns

High cost ofchemical fertilizers

Lack of adaptedvarieties

Cropping schedule notgeared to rainfall

patterns

Lack of credit

Lack of credit Weak network forcollecting agricultural

data

Production factorsnot optimized

Poor managementof water

Surface-watercatchment methods

not in use

Straw mulchingnot practiced

Improper plantingdensity

Furrowing notpracticed

Absence of croppingmethods that conserve

rainwater

No irrigation orwater-reserve

policies in place

Low income

Low profitability ofagricultural activities

Natural resourcedegradation

Poor integration ofproduction systems

Poor managementof production

systems

Poorly developedcommoditysubsectors

Low livestock andfisheries productivity

Low cropproductivity

Poor managementof natural resources

Poor managementand integration of

production systems

people will be among the invited participants inprogram planning exercise. For nonresearchableconstraints that concern the work of institutionsnot represented in the exercise, the programleader should take the initiative to communicate

the information to them afterwards. This will helpthem to improve their own development plansand, by extension, will improve the prospects fora successful research program.


24

Exploiting regional potential

Research programs should not only aim to re-move constraints, even though this is a safe wayto ensure impact. They should also exploit oppor-tunities that arise from specific advantages or un-derutilized potential enjoyed by the region. Theseadvantages fall into several categories:

Geographic location. Proximity to seaports, air-ports, road networks, or navigable waterwaysthat can be used for trade.

Physical environment. Sufficient and regularrainfall, warm temperatures, moderate humidity,good weather, even terrain (flat or not too hilly).

Natural resources. Fertile, deep, and homoge-nous soil; abundant and diverse vegetative cover;water resources such as rivers, streams, a shallowwater table or a good quality of water; the pres-ence of useful ecotypes.

Human resources. Young population, satisfactoryemployment rate, good level of technical exper-tise among farmers, large and qualified work-force.

Organization and administration. A high degreeof organization among professions, strong inte-gration of the region’s various economic activi-ties, and good production support services suchas credit, research, and extension.

Opportunities may be included in the constraintstree by expressing them in negative form. If, forinstance, the presence of an international airportwould allow for the export of vegetables and thisis not done at present, the constraint may be for-mulated as “limited recognition of export possi-bilities.”

Box 7: A selection of nonresearchable constraints for Benin’s Central Region

Constraints related to low livestock productivityinadequacy of watering stations during the dry seasonlack of demarcation of natural pasture land zonesanimal health not monitored by farmersdistribution and use of pharmaceutical and veterinary products that are fake, expired, or ineffectiveveterinary products inappropriately packaged for village conditionslivestock producers’ lack of informationlow investment by livestock producers in their operationslow availability of immediately useful technologies

Constraints related to the underdevelopment of commodity subsectorslack of confidence in formal credit institutionsproduction based on usuryuse of loan money to cover religious expenses or for purchases that enhance prestigevagaries of the weatherlack of credit to grow crops other than cotton and lack of means for ensuring regular supply of maize seedno competition among agricultural credit institutionslack of information on the use of maize seedimproved technology not available locally for postharvest storage of cropslack of access to marketing information


25

Detail of Step 3 of the Program Formulation Process, ‘Evaluation of Existing Results ’

RationaleEfficient research programs do not inadvertently repeat re-search already done within the country or elsewhere

OutputDocuments that help the program committee properly de-fine the scope of research projects and avoid duplication ofpast work

ResponsibilityProgram committee member with background in agronomy

ParticipantsResearchers who belong to regional program or collabo-rate with it

Information needsAnnual reports and research publications concerning the region or comparable regions

MethodsBibliographic review, interviews with extension staff and farmers and data from research management informationsystems (e.g., INFORM)

Time and resourcesReview should be started roughly three months before the first planning workshop. It will occupy most of assignedpersons’ time


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results



2.3 Evaluation of Existing Research Results

The evaluation of existing research results,step 3 of program formulation, builds an in-

ventory of technological progress alreadyachieved and applicable to the region. It identifiesresearch results already transferred or ready to betransferred, preliminary results in need of furthertesting, and partial results requiring further in-vestigation. It also takes note of problems not yetresearched. The review paves the way for new re-search projects to begin at the most advantageouspoint by taking account of previous knowledge.At the same time, it helps weed out projectswhich, in fact, are duplications of work alreadydone.

The review looks at the past work not only of theprogram and the institute, but also of other re-search programs covering similar areas and prob-lems. Computerized searches help identify resultsboth internal and external to the national researchsystem.

For internal results, the research institute’s man-agement information system should be consultedwhere one exists. Systems like INFORM, devel-oped by ISNAR, provide data on research objec-tives, approaches, and results. They are apowerful tool for avoiding duplication in researchprojects (Vernon, 1995).

External searches are also important. As computeraccess improves, searches of international infor-mation bases are becoming the norm for research.Since regional research programs usually face awide range of constraints, the external review isparticularly important. It may find that identifiedresearch issues have already been successfully ad-dressed elsewhere.

But the research review poses a potential timingconflict. On the one hand—and this needs empha-sizing—the review can only be completed oncethe constraints have been laid out. This is becausethe constraints analysis pinpoints those research


26

themes for which previous research results needto be compiled. On the other hand, because the re-search review is a lengthy process, the interests ofthe overall program formulation are served if thereview is conducted beforehand. In this case itmust anticipate as many constraints as possible.The conflict can be resolved by doing a second re-search review, after the first programming work-shop, to cover those constraints not anticipated inthe review conducted prior to the workshop.

The pool of past research results and knowledgevaries from country to country and, within acountry, from one region to another. In most de-veloping countries, it is small because little re-search has been done or because the results havenot be archived. Available results are most oftensectoral in nature and concerned more with cashcrops than food crops. There are usually few re-sults in the area of natural resource managementas well. This reflects what has, in the past, beenthe fundamental aim of research in most develop-ing countries: to increase crop productivity, espe-cially for export crops, without much worryingabout management of natural resources. Further-more, the emphasis has been on varietal improve-ment and plant protection. There are exceptions,of course. Early on some countries built nationalresearch systems with the explicit goal of promot-ing sustainable agricultural development.

By way of example, Box 8 highlights some of Be-nin’s past research results in the area of natural re-source management.

It should be noted that, at this level of planning,such results are to be presented as a synthesis.They are examined in more detail later, during theformulation of research projects.

Sometimes research results that could eliminate aconstraint have already been obtained in anotherregion of the country or abroad in agroecologicalconditions similar to those of the region con-cerned. In such cases, regional research programsare called on to formulate projects to confirm oradapt these results. But identifying such resultsrequires complete familiarity with the state ofboth national and international research. Thispoints to the necessity of involving people fullyinformed of the latest innovations in the planningprocess, be they national or foreign resource per-sons.

Experience shows that analysis of the existingpool of research results, while it is a pivotal stepin planning, doesn’t always elicit the interest itdeserves. It is wrongly neglected and, most often,incomplete and limited to the specialization of theperson assigned to conduct it. For this reason werecommend entrusting this detailed work of col-lection and analysis to a multidisciplinary team ofmotivated, veteran researchers under the leader-ship of one of the members of the program com-mittee. In any case, the first stage of the reviewshould be ready for discussion and validation atthe first planning workshop, in the presence offarmers and development authorities.

Box 8: An overview of results from past NRM research in Benin

Soil fertility❖ Ferralitic and ferruginous soils: Progressive degradation of these soils’ chemical fertility has been demon-

strated—lowered pH, reduction of organic matter, low content of exchangeable bases, and N, P, and K deficien-cies.

❖ A “sandwich” composting method that uses crop and fallow residues, as well as industrial wastes, has been devel-oped. It yields sufficiently decomposed matter in 90 days.

❖ Anaerobic composting accompanied by biogas production has been largely mastered.❖ Use of natural phosphates: At 50 percent acidification, yields on ferruginous soils are very close to those obtained

with ordinary phosphate.

Soil conservation❖ On ferralitic soils, runoff reaches 60-80 percent and erosion causes soil losses of 20-30 tons per hectare per year.

On the edges of plateaux with ferralitic soils where slope exceeds four percent, losses reach 70-80 tons per year.

Conservation and promotion of woody plant resources❖ Fast-growing species have been selected for each soil type.❖ The required periods of growth in bags in the tree nursery have been determined for the main species.


27

Detail of Step 4 of the Program Formulation Process, ‘Defining Research Objectives and Strategy ’

RationaleThe research program must focus on unambiguous objec-tives. Its projects, in turn, must respond to identified con-straints

OutputConsensus on overall program direction and on objectivesto be addressed to solve identified constraints

ResponsibilityProgram leader supported by facilitator

ParticipantsAll workshop participants in close interaction with scien-tists and others to ensure relevant, feasible objectives

Information needsRegional review, constraints analysis, and past research review

MethodsDiscussion in plenary workshop and working groups

Time and resourcesBetween half and one day


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results



2.4 Defining Research Objectives and Strategy

Step 4 of program formulation, defining re-search objectives and strategies for the re-

gional program, is a highly synthetic step.Information from steps 1, 2, and 3 is combined.Use of the results of the regional domain reviewand constraints analysis ensure that program ob-jectives are relevant. Information from the reviewof past research helps guarantee feasibility of ac-tivities and prevent duplication of research effort.

In earlier days it was recommended that the re-searchers take exclusive responsibility for settingobjectives. We find, though, that for a regionalprogram, which is meant to involve users directly,research objectives are best defined jointly by re-searchers and users. While researchers assess ap-proaches on their scientific merit, users and otherstakeholders may assess them on their adoptionpotential.

Consider the example of research on bollworm, apest that causes major damage to cotton. The re-search objective corresponding to this problemmay be formulated in many ways: to develop re-

sistant varieties, to identify an effective pesticide,to develop an integrated pest management (IPM)scheme, or to search for a biological enemy. Re-searchers may limit the range of research objec-tives by arguing that genetic resistance to the pesthas never been found. Other stakeholders maysimilarly limit the range by pointing out the prob-lems encountered in using IPM techniques onscattered plots. The group may thus agree that theobjective should be to identify a chemical controlagent (in the short run) and a biological enemy (inthe long run).

Once a research objective is set, project titles canbe formulated almost immediately. For example,if the objective is to identify a chemical controlagent, the project may be called “evaluation ofpesticides for controlling bollworm damage”.Thus, defining research approaches and strategiesfollows directly and logically from setting re-search objectives.

As shown in this example, one constraint doesn’tnecessarily result in a single research objective.


28

Several objectives may be formulated, reflectingthe complexity of the constraint or the uncertaintyof finding a solution using a particular approach.By defining several objectives, the strategy onhow to tackle a certain constraint is further re-fined.

Research objectives should be defined with par-ticular care for the constraints at the bottom andtop of the constraints tree (formulated during step2). The highest level objective is the binding ele-ment of the program. It gives a sense of directionand strategy to program members, both in formu-lating the program and in executing it.

The research objectives formulated to correspondto constraints at the bottom of the tree are closelytied to possible research responses. They are thestarting point for designing research projects, thebuilding blocks of the research program.

Specific research objectives can be formulatedonly for researchable constraints. For nonre-searchable constraints, complementary develop-ment measures may be formulated, but these arenot addressed in the research program. Box 9shows the link between researchable constraints,research objectives, and proposed research projecttitles for selected constraints in Benin’s CentralRegion. Box 10 presents a selection of suggestedcomplementary development measures.

Box 9: Selected researchable constraints, research objectives and proposed project titlesfor Benin’s Central Region

Low crop productivityResearchable constraint: Shifting cultivation of yam damages environmentResearch objective: Develop yam production systems that preserve environmentResearch projects: Develop a yam-based production system that preserves soil fertility and environment; developmanure mix for yam that doesn’t affect tubers’ eating characteristics; develop productive combination of yam andtree species

Low livestock productivityResearchable constraint: Low productivity of natural pasture land during dry seasonResearch objective: Improve pasture land productivity during dry seasonResearch projects: Enhance natural pasture land through use of leguminous shrubs; develop natural pasture man-agement techniques

Poorly developed commodity subsectorsResearchable constraint: Credit not matched to financing requirementsResearch objective: Adapt credit to financing requirementsResearch project: Study financing requirements and credit supply over time

Poor management of natural resourcesResearchable constraint: Lack of knowledge about lumber speciesResearch objective: Achieve expertise in silviculture for most important local lumber species (Isoberlinia, Afzelia,Anogeissus)Research projects: Develop seedling multiplication techniques for lumber species, study of planting methods andtree growth

Poor management of production systemsResearchable constraint: Absence of cropping methods that help conserve rainwaterResearch objective: Develop cropping methods wherein plants make better use of soil moistureResearch project: Use agronomic measures to conserve soil moisture

Box 10: Selected complementary development measures, by constraint, for Benin’s Central Region

Low livestock productivityNonresearchable constraint: Lack of demarcation of natural pasture zonesComplementary development measure: Improve land use regulation

Poorly developed commodity subsectorsNonresearchable constraint: Lack of confidence in formal credit institutionsComplementary development measure: Improve targeting and sensitization of farmers. Improve structure of institu-tions to bring them closer to farmers


29

2.5 Identifying Research Projects

Detail of Step 5 of the Program Formulation Process, ‘Identification of Research Projects ’

RationaleTo attain defined research objectives, research projectsneed to be specified

OutputList of research projects with their objective, correspond-ing constraint, project activities, and required human re-sources

ResponsibilityProgram leader supported by the program committee

ParticipantsResearchers associated with the program

Information needsResearch methodologies and, if possible, data on human resource requirements from earlier projects

MethodsReview of earlier projects, for example, through a management information system such as INFORM and consult-ation with possible collaborators

Time and resourcesUp to two months


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results



Identification of research projects is step 5 in theresearch program formulation process. Re-

search projects follow from the constraints al-ready identified. Each project consists of acoherent set of research activities. Every activity isrequired to attain the research objective; one can-not be removed without jeopardizing the project’ssuccess. At this stage, research projects don’t needto be spelled out in great detail, but their overallshape should be determined. This takes the formof a title, the constraint to be addressed, the re-search objective, the component activities, the re-quired human resources, and the location andduration of the project.

Research projects normally correspond to the con-straints and objectives identified at the lowestlevel of the constraints tree. Occasionally a con-straint at a higher level may be identified as thestarting point for a research project. This occurs incases where it is clear that all underlying con-straints must be simultaneously removed if re-search is to successfully meet its objectives.

Research projects often involve various disci-plines, but are made up of activities, each ofwhich is normally undertaken by scientists fromone discipline. Once the research project activitiesare defined, the time required of the different dis-ciplines for each activity can be estimated. This ismost easily done in terms of months of researchertime per year. In addition, the duration of the pro-ject (in number of years) has to be assessed.

When the time requirements per disciplinechange considerably during the course of theproject, another approach can be followed. Then itis best to assess for each activity the number ofyears required and the number of months thatneed to be spent in each year. In this manner, aprofile of time allocations across the length of theproject can be established.

Time is most easily estimated in actual months ofresearch time, but it should be recognized that aperson will not spend 12 months per year on re-search. Even a person with a full-time research


30

dedication spends time maintaining skills andknowledge, participates in conferences, has ad-ministrative duties, and takes vacation. Rarelywill people spend more than eight months peryear effectively on research projects.

Time is, of course, not the only cost of researchprojects, but on average it provides a high correla-tion with the actual total costs of a project for tworeasons. First, the cost of human resources makesup the lion’s share of most research institutionbudgets (Brush, 1996). Second, many of the opera-tional costs (e.g., transport, fertilizers) are linkedwith the time commitment that researchers aremaking. For these reasons, at this stage of projectidentification, using approximations of human re-source requirements to estimate total project costsis acceptable.

The program’s research staff is normally responsi-ble for formulating projects. In the first workshop,they may come up with preliminary project titles.But laying out project activities and estimating

the required human resources are tasks best doneafter the workshop. Between the first and secondworkshops researchers may need up to twomonths to identify research projects in sufficientdetail. The time is needed mainly to conclude bib-liographic studies, reliably estimate the timeframe of projects, and consult with potential col-laborators.

Identifying research projects and their key pa-rameters certainly isn’t easy. Researchers tend tounderestimate the time needed for successfulcompletion of a project and overestimate the re-sults that can be achieved. These weaknesses injudgment can be mitigated by consulting data onearlier projects as they become available in man-agement information systems (Vernon, 1995).Projects documented in the data base will obvi-ously be somewhat different from the ones beingplanned. But they may give a good idea of thetime needed for various types of experiments andother activities.

Classifying research projects

Research projects can be usefully divided into twocategories. Most fall under the heading of “techni-cal”, that is, they resolve a constraint to increasedproduction or improved resource management.Once the results of such projects are dissemi-nated, they should have a direct impact on the re-gion. The second category is “support” projects.These eliminate scientific constraints to the pro-gress of technical projects, thereby improving theoverall effectiveness of the research effort. For ex-ample, without a proper assessment of the loca-tion and gravity of soil degradation, it may beimpossible to make progress in other types ofNRM research projects.

Another way to classify research projects is bytheir expected benefits. Some projects will haveonly one type of benefit. They may influencecropping, livestock, water management, forestry,or soil quality, but never more than one of those.These can be referred to as “focused” projects.Other projects may have an impact on several di-mensions. Agroforestry work, for example, oftenaffects cropping, forestry, and soil quality. Such re-search efforts can be referred to as ”integrated”projects. These surpass traditional researchboundaries and are a distinctive feature of re-gional programs. Integrated projects are the glueof regional programs, allowing them to handle

the complex problems that arise in the context ofresource use at the regional level. In Benin’s re-gional research programs, integrated projectsmake up 36 percent of projects identified for theCentral Region, 25 percent in the Southern Re-gion, and 37 percent in the Northern Region. Box11 lists a number of integrated projects identifiesfor Benin’s Central Region.

Project screening, transfer, and removal

As we saw in step 2 of program formulation, agri-cultural and NRM constraints in a region andtheir relationships can be highly complex. So it’sunderstandable that the same constraint may beidentified at two or more places in the constraintstree or that similar research projects may be iden-tified by different researchers. “Screening” refersto the elimination of unnecessary project duplica-tion within the program and other programs.Management information systems, which com-bine information on all projects in the institution,may be used as a source for checking for duplica-tion across programs.

It may also happen that some of the proposedprojects are more suited to research programsother than the regional one—whether they becommodity, disciplinary, or production-factor


31

programs. Passing such projects to those pro-grams better suited to deal with them is referredto as “transfer.”

Finally, in the creative atmosphere of researchproject identification, some researchers may pro-pose projects that, in fact, don’t require research atall, but only extension, credit, or input support.“Removal” is the identification and elimination ofthese types of projects.

The screening, transfer, and removal procedurediffers from priority setting (step 6 in the programformulation process) in that it analyzes projectspurely on their technical merit. If technical con-tent does not justify the duplication of a project orits presence within the regional research program,the project is removed from the program orshifted to a more appropriate one. It can thereforebe assumed that in step 6, all the proposed pro-jects are appropriate for the regional program

Project title Activity Subsector Human resource requirementsin first three years

Development ofmethod for

Development ofimproved

adapted todifferent types ofterrain

Development ofefficient methods of

Development ofagronomically feasible

Development ofmethods for

Development of

methods

Development ofproductive


Development ofmethods for using

Introduction of

bushfire control

anti-erosionmethods

fallow management

crop rotation system

soilmoistureconservation

low-cost fertilization

combinations of yamand tree species

managing naturalpasture land

agroindustrialby-products andcrop residues inanimal feed

forage species incrop rotations

Diagnose indigenous methods for bush fire controlScreen species for use as “green” firebreaksExperiment with setting fire during different periods

Take inventory of different types of erosionDiagnose traditional anti-erosion methodsScreen shrubby and herbaceous species for use inerosion controlConduct comparative study of mechanical methodsof erosion control

Develope techniques for enhancing natural fallow landDevelope different techniques for shrub-based fallowDevelope ways to recycle animal manure and cropresidues to fertilizer soil

Diagnose traditional crop rotation systemsIdentify best cropping practicesDevelop optimal rotation systems

Compare various soil cultivation practices that allow forbetter use of moistureCompare soil-cover practices that allow for better use ofmoisture

Characterize soilsEstablish nutrient budgets for specific cropsEstablish fertilizer formulas according to crop and zoneDevelope appropriate mixes of organic and mineralfertilizers

Screen tree species that can serve as living stakesfor yam

Study resistance of pasture forage species to tramplingStudy stocking potential of natural pasturesDevelop an appropriate schedule for pastureexploitation (periods for grazing, hay making and silaging)

Compile an inventory of locally available agro-industrialand food processing by-productsDevelop methods for storage and preservation ofby-productsDevelop methods for use of by-products in animal feed

Test the performance of forage species in the regionDevelop pasture-establishment methods compatiblewith regional production systems

ForestrySoil quality



Soil qualityCrops

Soil qualityCrops

Soil qualityCrops

ForestryCrops

Soil qualityLivestock

CropsLivestock

Soil qualityCropsLivestock

Forester: 7 monthsSoil scientists: 4 monthsSocioeconomist: 3 monthsTotal: 14 months

Soil scientists: 8 monthsSocioeconomist: 1 monthAgroforester: 4 monthsTotal: 13 months

Agroforester: 3 monthsZootechnician: 3 monthsAgronomist: 7 monthsTotal: 13 months

Socioeconomist: 1 monthAgronomist: 7 monthsTotal: 8 months

Soil scientist: 3 monthsAgronomist: 7 monthsTotal: 10 months


Agronomist: 3 monthsTotal: 3 months

Pasture scientist: 10 monthsTotal: 10 months

Zootechnician: 6 monthsNutritionist: 3 monthsTotal: 9 months

Pasture scientist: 4 monthsAgronomist: 3 monthsTotal: 7 months

Box 11: Integrated research projects for Benin’s Central Region

Project title Activity Subsector Human resource requirementsin first three years

Development ofmethod for

Development ofimproved

adapted todifferent types ofterrain

Development ofefficient methods of

Development ofagronomically feasible


Development of

methods

Development ofproductive


Development ofmethods for using

Introduction of

bushfire control

anti-erosionmethods

fallow management

crop rotation system

soilmoistureconservation

low-cost fertilization

combinations of yamand tree species

managing naturalpasture land

agroindustrialby-products andcrop residues inanimal feed

forage species incrop rotations

Diagnose indigenous methods for bush fire controlScreen species for use as “green” firebreaksExperiment with setting fire during different periods

Take inventory of different types of erosionDiagnose traditional anti-erosion methodsScreen shrubby and herbaceous species for use inerosion controlConduct comparative study of mechanical methodsof erosion control

Develope techniques for enhancing natural fallow landDevelope different techniques for shrub-based fallowDevelope ways to recycle animal manure and cropresidues to fertilizer soil

Diagnose traditional crop rotation systemsIdentify best cropping practicesDevelop optimal rotation systems

Compare various soil cultivation practices that allow forbetter use of moistureCompare soil-cover practices that allow for better use ofmoisture

Characterize soilsEstablish nutrient budgets for specific cropsEstablish fertilizer formulas according to crop and zoneDevelope appropriate mixes of organic and mineralfertilizers

Screen tree species that can serve as living stakesfor yam

Study resistance of pasture forage species to tramplingStudy stocking potential of natural pasturesDevelop an appropriate schedule for pastureexploitation (periods for grazing, hay making and silaging)

Compile an inventory of locally available agro-industrialand food processing by-productsDevelop methods for storage and preservation ofby-productsDevelop methods for use of by-products in animal feed

Test the performance of forage species in the regionDevelop pasture-establishment methods compatiblewith regional production systems




Soil qualityCrops

Soil qualityCrops

Soil qualityCrops

ForestryCrops

Soil qualityLivestock

CropsLivestock

Soil qualityCropsLivestock

Forester: 7 monthsSoil scientists: 4 monthsSocioeconomist: 3 monthsTotal: 14 months

Soil scientists: 8 monthsSocioeconomist: 1 monthAgroforester: 4 monthsTotal: 13 months

Agroforester: 3 monthsZootechnician: 3 monthsAgronomist: 7 monthsTotal: 13 months

Socioeconomist: 1 monthAgronomist: 7 monthsTotal: 8 months



Agronomist: 3 monthsTotal: 3 months

Pasture scientist: 10 monthsTotal: 10 months

Zootechnician: 6 monthsNutritionist: 3 monthsTotal: 9 months

Pasture scientist: 4 monthsAgronomist: 3 monthsTotal: 7 months


32

from a technical standpoint. They can then beanalyzed on their expected benefits.

In Benin, the screening, transfer, and removal pro-cedure resulted in a major streamlining of theprojects identified for the regional research pro-grams. In the Central Region’s program thenumber of projects fell from 90 to 48, in the South-

ern Region from 110 to 53, and in the NorthernRegion program from 125 to 67. The procedure isthe responsibility of the program committee,headed by the program leader. The representativeof the planning unit should collaborate inten-sively with the program committee to ensure thatprojects to be transferred to another program areactually being considered there.


33

Detail of Step 6 of the Program Formulation Process, ‘Priority Setting ’

RationaleFinite resources available to the program usually meansthat only a limited number of the identified research pro-jects can be implemented

OutputA list of projects ranked from highest to lowest priority

ResponsibilitySocioeconomist supported by the program leader and afacilitator

ParticipantsAll workshop participants

Information needsEconomic data on agricultural production and resource values plus information on potential spread of tech-nology and adoption behavior

MethodsSimplified cost-benefit analysis, trend analysis, and natural resource valuation

Time and resourcesOne month to prepare background information and one to one-and-a-half days during the second workshop


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results



2.6 Choosing Priority Research Projects

Invariably, more research projects are identifiedthan can be implemented with the available re-

sources. Hard decisions must therefore be madeabout the relative importance of projects.

Step 6 in the program formulation process is pri-ority setting. Here, the analytical approach ofsteps 1, 2, and 3 and the creative approach ofsteps 4 and 5 are confronted with the managerialimperative to ground the overall programmingprocess in reality. Projects can be ranked using anumber of methodologies. In these guidelines, weuse a scoring method to decide which projects toimplement first and which to put off until later ifadditional resources are found.

Choosing priority projects for a regional programis complicated by the fact that the different re-search projects may provide different kinds ofbenefits. For example, it’s not easy to compare aproject for improving natural pastures with onefor eliminating sorghum losses due to a pest, orwith one for reducing erosion through agrofore-stry techniques.

In commodity research programs, all projects ulti-mately aim to increase commodity subsector prof-itability. Comparison of projects isstraightforward because the relative benefit is asufficient criterion for evaluation. If project “A”contributes more to maize productivity, andhence to profitability, than project “B”, it’s not im-portant for the purpose of comparison to knowthe absolute values of the expected benefits.

In regional research programs, the situation is dif-ferent. Projects may contribute to the profitabilityof the cereal subsector, or to the value of dairyproduction, or to the quality of soils. Therefore, asreasoned by Springer-Heinze (1994), comparingthe likely benefits of different projects requires amore elaborate method. The priority-setting pro-cedure presented here allows comparisons of thedifferent types of benefits that could be attainedthrough research projects. The procedure com-bines elements of a number of priority-setting ap-proaches (multiple criteria, cost-benefit analysis,research production functions) with productionand natural resource valuation techniques. Forfurther description of priority-setting methods,


34

readers are referred to Alston, Norton, andPardey (1995) and Contant and Bottomly (1988).

Other procedures, of course, may be used thanthe one presented in this chapter. The choice de-pends largely on the time and data available andon the economics skills of the program commit-tee. If time allows, a team of economists couldtheoretically develop analytical models for esti-mating the specific benefits of each and every re-search project proposed for the regional program.This might be to great advantage. But, in practice,such an all-encompassing procedure would be ex-tremely data intensive and could take years giventhe number of projects normally considered in aprogram formulation exercise. Just as bad, suchan approach would probably be so complex as tobe unintelligible to outsiders, who would losesight of why one research project might be chosenover another. Alternatively, a simple assessmenton a few key criteria can be used. This approach is

rather straightforward but often confirms ratherthan challenges the established thinking about theproblems of the region. Readers who would wantto set priorities by using a few key criteria with-out an economic evaluation framework shouldconsult Tripp and Woolley (1989).

The priority-setting challenge, then, is to use anapproach that effectively deals with the differenttypes of benefits coming from a variety of re-gional research projects, but that can also be ap-plied relatively quickly. The method presentedhere requires a modest amount of data collection,manipulation, and interpretation, and allows allprojects to be evaluated using the same methodol-ogy. We first explain the methodological back-ground of the approach and guide you throughits application. Then, in the final section of thischapter, we discuss how the method can be feasi-bly applied in the context of a planning work-shop.

A methodological framework for priority setting

The basic mathematics of the methodology forsetting regional research priorities have beenadapted from the approach developed by Collionand Kissi (1994) for commodity programs. In theirapproach, a project score is calculated as follows:

Ni = (V × pi × si × ai) / Ci (1)

where

Ni = the score for project i.

V = the total potential impact of the research pro-gram on improving the value added to the sub-sector under consideration (e.g., cotton).

pi = the contribution of research project i to realiz-ing V. This can also be expressed as the relativeseverity of the constraint that the project ad-dresses. The sum of p for all projects is one.

si = the chance of success of research project i (be-tween zero and one).

ai = the expected rate of adoption of the resultsforthcoming from project i if it is successfullycompleted (between zero and one).

Ci = the cost of project i (normally approximatedby the required researcher time).

Projects that receive the highest scores have thehighest priority for implementation. The equationadmittedly suffers from several theoretical short-comings, the main ones being that time lags in thegeneration and diffusion of new technologies arenot sufficiently accounted for and cost estimateshave been greatly simplified (see Collion andKissi, 1994). However, the formula has been suc-cessfully applied in numerous planning exercises.The high level of stakeholder participation,mechanisms for fine-tuning participants’ judg-ments and assumptions, and the transparency ofthe results compensate largely for the theoreticalshortcomings.

The adaptation of this priority-setting approachto regional program planning is based on threeconsiderations which bear repeating:

1. Research projects in a regional program maycontribute to different types of benefits andobjectives. To account for such differences be-tween projects, the priority-setting approachhas to distinguish between various benefit di-mensions, such as the benefits of increasedcrop productivity, the benefits of increasedlivestock productivity, the benefits of im-proved soil quality, and so on.

2. Because one project may contribute to onebenefit dimension and another project to a dif-ferent dimension, it is not sufficient to know


35

the relative contributions of each project to thebenefit dimensions on which they impact. Theabsolute sizes of each of the benefit dimen-sions need to be estimated in order to com-pare, for example, the benefit of a five percentincrease in crop production with a 20 percentincrease in livestock production.

3. Regional research programs take into accountthe integrated nature of agricultural produc-tion and resource management. Many projectswill therefore contribute to more than onebenefit dimension.

Based on these considerations, the score of projecti can now be calculated by extending equation (1)into the following:

Ni = [ (Vc × pci + Vl × pli + Vf × pfi +Vq × pqi +Vw × pwi) × si × ai ] / Ci (2)

where all previously used symbols have the samemeaning as in equation (1), and V is the total im-pact of the regional research program on improv-ing the value added or quality of the benefitdimension indicated by the subscript. The sub-script letters have the following meanings:

c = cropping activitiesl = livestock activitiesf = forestry activitiesq = soil qualityw = water management

Equation (2) states that the total potential benefitof project i is equal to the sum of its potential ef-fects on improving the value added to crop pro-duction, livestock production, and forestryproduction, plus its effect on soil quality andwater management. This sum is then corrected totake into account project i‘s chance of success (si),expected adoption rate (ai), and project costs (Ci).

Many projects contribute to only one of these fivebenefit dimensions, which simplifies calculations.And not all the benefit dimensions may be rele-vant to the planning of every regional program.In Benin, for instance, only the first four wereused. Other benefit dimensions may be added tothe equation as well. In Senegal, for example, fish-eries was included.

Finally, it is important not to double count projectcontributions, for example, in both the crop pro-duction and soil quality benefit dimensions. Toavoid this, the types of benefits to be assigned tothe different dimensions should be clearlydefined. For crop production, to continue the ex-ample, one should estimate the short-term pro-duction effects on the assumption that soil qualitydoesn’t change. For soil quality, one should con-sider the long-term effects of soil improvement, asmeasured in land prices, production potential, ornutrient value.

Once the relevant benefit dimensions are speci-fied, the identified research projects are assignedto the different dimensions and the potentialbenefits of each of the projects are calculated.Then the other project parameters are re-viewed—chance of success, potential for adop-tion, and research cost—and a final project scoreis calculated. Finally, the resulting priority rank-ing is discussed and evaluated.

The sequence for priority setting in Benin’s Cen-tral Region research program is illustrated in theboxes throughout this chapter. Benin used simpleestimation procedures and relied on the collectivejudgment of workshop participants to arrive atreasonable assessments. The estimation proce-dures were kept simple for three main reasons:

1. In many situations sophisticated procedurescannot be used because the necessary data arenot available.

2. Simple procedures keep the priority-settingmethod transparent for the workshop partici-pants. This promotes acceptance of themethod, and allows the assumptions and out-comes to be improved on by all.

3. At the level of program planning, the informa-tion on the projects being prioritized is in out-line form only. There is no reason tocomplement such sketchy project informationwith highly detailed benefit estimation proce-dures.

However, in cases where detailed data are avail-able or where more precise methods can be easilyimplemented, their use is certainly recommended.


36

Defining benefit dimensions

In the area of agricultural productivity, there aretwo main benefit dimensions that will define pro-ject scores:

■ The benefits of increased crop productivity.That is, the productivity of both annual andperennial crops. In regions where perennialcrop production is especially important, thismay be split off into a separate benefit dimen-sion.

■ The benefits of increased livestock produc-tivity. This concerns large and small rumi-nants, pigs, poultry, and other animals. Wherefisheries are of limited importance, they can beincluded in this benefit dimension. Otherwisea separate fisheries benefit dimension may beneeded.

The benefits of improved pasture productivityshould be treated with special care. To the extentthat improved pastures lead to better animal nu-trition, their benefits fall under the livestock bene-fit dimension. To the extent that they lead to moreproductive crop rotations, their benefits fall underthe soil quality benefit dimension of improvednatural resource management (see below).

For improved natural resource management,three benefit dimensions are most significant:

■ The benefits of increased sustainable for-estry production. By supplying wood, fruit,medicine, and other products, forests canmake a major contribution to the regionaleconomy (Franzel, Jaenicke, and Janssen,1996). By improving forest management, theflow of such products may be maintained orincreased.

■ The benefits of improved soil quality. Soildegradation is a serious threat to long-termagricultural productivity and may, througherosion, have major off-site effects. Siltation ofdams may undermine the potential for irriga-tion and energy generation. Siltation of riversand reservoirs may reduce fish catch and in-crease the threat of flooding.

■ The benefits of improved water manage-ment. Water, not land, may be the most limit-ing resource in a region. An example is thesemiarid regions of Morocco. In such areas,improved use of water may have a high im-pact. This is especially true where competitionexists among different uses of water—for hu-man consumption, energy production, and ir-rigation, for example.

The benefit dimensions related to natural resourcemanagement have clear economic value to the re-gion, though they tend to be long term ratherthan short term. Incorporating resource manage-ment concerns into project priority setting, there-fore, does not move us away from maximizingthe economic benefits of research. Rather, it im-plies an explicit effort to put a value on such long-term effects.

Some benefit dimensions are difficult to assess be-cause the major share of the benefit will fall toother regions and, possibly, other countries. Also,the effect of research within the region may not beclearly traced. This may be the case, for example,of research to maintain genetic diversity. The pri-ority of such research to a (strictly) regional pro-gram may be low because the users within theregion obtain only a small part of total benefits,and because for maintaining genetic diversityglobally, it may not be clear how important workin the specific region is. This is not a reflection onthe importance of genetic diversity, but on thegeographic boundaries used to delimit researchbenefits. The importance of genetic diversitywork is better assessed at a national, or possiblyinternational level. If from such assessments theconclusion follows that work in a certain region iskey to maintaining diversity, resources should bemade available and genetic diversity work shouldbe integrated into the regional program.

As we shall see, the next task in applying this pri-ority-setting approach is to estimate the potentialimpact of the regional research program on eachof the selected benefit dimensions. This requiressubstantial work. Indeed, including additionalbenefit dimensions makes the exercise more com-plex. The caution here is that benefit dimensionsshould not be added without good reason. In thecase of Benin, it was decided not to include an ex-plicit water management benefit dimension, be-cause it would complicate the calculationsunduly.

Finally, it should be noted that benefit dimensionsdon’t directly correspond to the principalbranches of the constraints tree. This is simply be-cause identified constraints often have repercus-sions in several dimensions. So research aimed ateliminating a constraint such as “poor integrationof farming system components” may bring bene-fits to crop and livestock production as well as tosoil quality.


37

Estimating the size of benefit dimensions

The “size of a benefit dimension” is defined as thedifference between, on the one hand, the expectedvalue of a region’s agricultural production andnatural resources and, on the other hand, their po-tential value if all constraints that are researchablefor that benefit dimension are removed.

Estimation principles

What are the potential benefits of removing all re-gional research constraints to, say, livestock pro-duction or soil quality? For livestock production,it’s a matter of estimating what portion of pro-duction growth is attributable to removing theseconstraints. This is a rather straightforward exer-cise, as we will see below. The same goes for cropproduction (see also Collion and Kissi, 1994).

It is less obvious how soil quality relates to agri-cultural supply (Norse and Saigal, 1993; Mendel-sohn et al., 1994). Here it may be easiest to turnthe question around and ask what is the value ofsoil quality losses and what share of those lossescould be prevented through regional research. Wealso know that soil degradation often causestraceable problems downstream (off-site effects).So we should try to estimate the extent of down-stream problems and the potential of the regionalresearch program to solve them.

For forestry, both the quality and supply ap-proaches can be attempted. The supply approachfollows the livestock example, estimating theadded increase in the supply of forest productsachieved by removing constraints. The quality ap-proach is similar to the soil example in that it esti-mates the value of the forest cover that disappearsannually because of poor management and thenassesses the extent to which this could bestopped.

For water management, an efficiency improve-ment approach may be used for valuation pur-poses. Current water use in the region iscompared with likely use following the introduc-tion of water-saving measures. The value of suchsavings can be calculated by multiplying theamount by the price of water. The price of watercan be obtained by estimating the cost of supply-ing water (production and transport costs) (Arn-tzen, 1995). This approach will not be elaboratedfurther here.

Estimation assumptions

To estimate the size of the benefit dimensionsthree steps are taken:

1. Initially a crude estimation is made of the ex-pected benefits of research. This estimationmay be that research contributes one percentadditional production growth (or value), orthat it stops all soil quality losses. The prelimi-nary nature of these assumptions allows theprogram planning team to prepare initial esti-mates of the size of the benefit dimensions be-fore the second workshop.

2. Second, the initial, crude assumptions are re-viewed. Is it probable that research could in-crease both crop and livestock production byone percent? Is it realistic to presume that re-search can stop all soil quality losses? If not,what share of the initial assumption can be re-alized? This review is done by the participantsin the second workshop.

3. Finally, the share of the regional program inthe total benefits derived from the whole re-search effort is estimated. The regional re-search program may be more important forone benefit dimension than for another. For ex-ample, this may be because of significant re-search efforts in other programs. Theestimation of the share of the regional pro-gram of total research benefit is also madeduring the second workshop.

Initial estimates of the size of the benefitdimensions

Initial, crude estimates of the size of the benefitdimensions are based on the supply impact of re-search, the on-site impact of agricultural research,and the off-site impact of research. As we haveseen, for crops, livestock, or forestry, crude esti-mates of potential research impact can be basedon the supply impact of research. For the soilquality dimension, the initial estimation is basedon an assessment of the on-site and the off-site ef-fects of research.

Supply impact of research. For production bene-fit dimensions, the first step is to assess the addi-tional growth in supply resulting from successfulresearch. Consider an agricultural sector at sub-sistence level, with a static policy framework (itdoesn’t get better or worse) and with no efforts toimprove the technological base of production.


38

Under these conditions, one may assume thatsupply growth will be strongly defined by the de-mand induced by population increases. Existingproduction constraints would cause supplygrowth to lag somewhat behind populationgrowth, but the lag would be shortened by thedissemination of technologies from other parts ofthe world. With population growth in Africa attwo to three percent per year, we can assume thatwithout any domestic efforts to make new tech-nologies available, the rate of supply growth inagriculture would be two percent. Probably suchgrowth would occur largely through increased in-put use and area expansion (where possible).

Technological progress, including agricultural re-search, helps increase supply growth. The initialquestion is by how much. Aggregate data on thisquestion are sparse. Research by Block (1995)showed that roughly one-third of productiongrowth in African agriculture could be attributedto technological change. With this estimate inmind we make the initial assumption that a suc-cessful research program could increase produc-tion growth by one percent—that is, raise it fromthe reference level of two percent to a total ofthree percent.

The impact of research started today will be feltonly after the several years needed for technologydevelopment and dissemination. But impact willbuild rapidly once research results are adopted.Rather than trying to construct benefit profilesover time, we propose to estimate the research-in-duced benefits by using the projected differencein supply in 10 years’ time. The projected supplyvalue in 10 years assuming two percent growth issubtracted from the value obtained assuming athree percent growth rate. This estimation methodis a compromise between two observations. First,the expected benefits in 10 years’ time will prob-ably still be below the projected difference, be-cause of the lag between technology developmentand diffusion. Second, benefits may build rapidlyin the years afterward, in which case their abso-lute value will be higher than the projected differ-ence for year 10.

Box 12 shows the estimates made in Benin forthree benefit dimensions: crop production, live-stock production, and forestry production.

Research impact on on-site resource quality. Theimpact of research on soil quality can be assessedin various ways. Norse and Saigal (1993) proposefour alternative approaches:

1. The difference between the market value ofquality land and that of degraded land ismeasured. The feasibility of this approach de-pends on the presence of a functioning landmarket and a sufficient number of transactionsto allow analysts to isolate the quality effectsfrom other price-shifting variables (populationpressure, accessibility, urbanization, and soon). To operationalize this approach, referencequalities must first be defined for differentplots.

2. The difference between the productivity ofquality land and that of degraded land ismeasured. To apply this approach, the interac-tion between land quality and productivitymust be clear. As with the method above, it isimportant to define standard reference quali-ties, which may be quite difficult.

3. Expenditures to prevent soil degradation—forexample on anti-erosion windbreaks—aremeasured.

4. Costs for replacing the elements that embodythe fundamental quality of the resource aremeasured. In the case of soil degradation,these are the nutrients being lost, as well assoil texture and structure. Van der Pol (1992)elaborated this approach for Mali.

For practical purposes, the choice of approach isdetermined largely by the availability of data. Letus look at each of the four approaches in thatlight.

First, in many developing countries, land marketsdo not function well enough to observe differencesthat can be interpreted. Second, the quality-pro-ductivity relationship is not disputed, but a factualestimation is very difficult, to say the least. Third,expenditures to prevent soil degradation are diffi-cult to measure, because of their nonlinearity. Toprevent 80 percent of soil degradation, a certain in-vestment may be sufficient. But to prevent the next10 percent may cost just as much. Last, the replace-ment approach is highly sensitive to the nutrientprices that are imputed and the effect of nutrientlosses is not yet fully understood (Smaling, Fresco,and de Jager, 1996). The replacement approachgives equal value to each cubic meter of soil lost,whereas it is well known that the initial losses donot have as large an impact as later losses.

In the case of Benin, market data, data on thequality-productivity interaction, and preventiondata were not available. However, there were dataon nutrient losses from an Africa-wide study. Fur-


39

thermore, the replacement approach (under 4above) had been applied before in the region (vander Pol, 1992). The replacement approach wastherefore used to estimate the size of the soil-qual-ity dimension. Box 13 illustrates this for the on-site benefits and costs.

Off-site impact of agricultural research. In thefirst, second, and fourth approach above, differentmethods are applied to measure the on-site effectof soil degradation. However, soil degradationmay also have off-site effects, when sedimenta-tion causes losses in fish catch, increased risks offlooding, or the shortening of the economic life ofhydroelectric plants. These losses should be meas-ured as well. For the approach under three above,the prevention of soil degradation stops the off-site damage as well, and there is no further needto measure off-site effects.

Estimating the off-site impact of research in agri-culture is certainly difficult. Soil quality research,

for example, may produce technologies to reducesedimentation, which in turn, benefits the fishcatch, reduces the chance of flooding, and im-proves the life span of dams. If soil quality re-search leads to improved water retention,flooding risks may be further reduced—a benefitto which forestry research and cropping technol-ogy development may simultaneously contribute.Similarly, agricultural research to increase wateruse efficiency may have the off-site effect of im-proving the quality of drinking water in urban ar-eas.

The off-site impact of soil degradation by erosionis considerable. Pimentel et al. (1995) estimatesuch impact to be on the order of 60 percent of on-site effects for the USA, but they do not draw con-clusions for other countries. Ashby et al. (1994)estimate the off-site effects of erosion of Colom-bian hillsides to be greater than erosion’s on-siteimpact on soil quality.

Box 12: Calculation of potential research impact on crop, livestock, and forestry production

The figures in the table at the rightshow 1992-94 values for crops, live-stock, and forestry in Benin’s CentralRegion (expressed as millions of dol-lars). We see, for example, that crop-ping is clearly more important thanlivestock production. It is also worthnoting that the value of the region’sforest products is more than threetimes that of its livestock production.To a more modest extent this canalso be seen at the national level: for-estry production is more valuablethan livestock production, and cropsare most important. The numbers re-flect low income levels in the country.Animal protein is too expensive formany people and firewood is the pri-mary source of energy.

To estimate of the potential value ofresearch, the difference is calculatedbetween a two percent and three per-cent increase in the value of produc-tion in the region (in the table below).

Here we see that if research resultedin a one percent supply growth (inaddition to the two percent growthexpected to result from populationincreases), the expected benefits ofresearch in the region would beclose to $30 million for the year2005. Benefits would obviously beskewed towards cropping, and awayfrom livestock.

CommodityNational

Production ValueRegion's Share

of Value (percent)Regional

Production Value

Crops

Total crops

Livestock

Total livestock

ForestryForest products

MaizeCassavaYamGroundnutSorghumCowpeaCottonRiceTomatoPepperSweet PotatoCashew nutOil palmCoconutPineapple

BeefSmall ruminantsMilkPorkPoultryEggsFish

99.671.1

115.424.929.129.664.64.9

18.66.71.91.5

78.19.11.2

35.27.9

34.48.9

15.65.0

70.0

556.3

177.0

227.6

16323753144332257

181540000

5125

1710105

25

6

17

15.922.842.713.24.1

12.720.71.21.31.20.30.60.00.00.0

1.80.91.71.51.60.53.5

136.7

11.5

38.7

Value 1992-94Category Value in 2005 Difference

2 Percent Growth(a) (b) (a)-

3 Percent Growth(b)

CropsLivestockForestry

All values in millions of dollars.

136.711.538.7

173.414.649.1

194.916.455.2

21.51.86.1


40

Identifying the principal off-site processes thatagricultural research may affect is the first step tovaluing such potential benefits. Off-site effectsmay be estimated following the same principlesas those presented for estimating the on-site im-pact of research on soil quality. That is, the valueof assets affected may be estimated, the produc-tion loss of the assets may be estimated, or thecost of compensating for the damage may be cal-culated. Box 14 provides an example from Benin’sCentral Region.

Valuing natural resources, both on-site and off-site effects, has become an important field ofstudy. Advanced methodologies, requiring so-phisticated data are available, and wherever pos-sible, these methodologies should replace thecrude calculations made here. We refer the reader

to Munasinghe (1993), Pearce (1994), Serageldinand Steer (1994), Bartelmus et al. (1994), and Wil-lis and Corkindale (1995) for further theory andexamples.

Estimating the feasibility of research for thedifferent benefit dimensions

Thus far, we have assumed that, on average, asuccessful research program could increase agri-cultural production growth rates by one percent.But the feasibility of achieving one percent pro-duction growth through research is not necessar-ily the same for cropping, livestock, and forestry.Whereas for one subsector, the total growth rateas a result of successful technology developmentmay be sizably higher than one percent, for othersubsectors it may be equal or lower. Moreover,

Box 13: Estimating the on-site benefits of soil quality research

The research program for Be-nin’s Central Region used a re-placement approach toestimate the direct benefits ofsoil-quality improvement re-search. Stoorvogel and Smal-ing (1990) estimated nutrientbalances for different land usesystems in Benin for the year2000. They distinguished sixprincipal land categories ac-cording to moisture availability:low rainfall, uncertain rainfall, good rainfall, problem area,naturally flooded, and irrigated. A seventh land categorywas added for permanent pasture land. By overlaying themap of the land categories with the map of the regionaldivisions, the share of the Central Region in the differentland categories was defined.

The Central Region has 35 percent of the area in Beninwith good rainfall and 19 percent of the permanent pas-ture land. The region also has part of the irrigated and theflooded areas—13 percent and 28 percent respectively ofBenin’s total. The other categories are not present in theCentral Region. By multiplying the total nutrient losses ineach land category by the share of the Central Region,the total nutrient losses in Central Region can be calcu-lated (see table above).

Valuing nutrient losses for the year 2005. To attach avalue to these losses, nutrient prices have to be known.Nutrient prices were obtained from IFDC statistics on fer-tilizer prices in West Af-rica (IFDC, 1995). Priceswere available for variousfertilizer sources, and thecheapest source for eachelement was chosen. Inthis example, these wereurea for N, TSP for P2O5,and KCl for K2O. Beforemultiplying nutrient losses

by fertilizer prices, a correction was made for the avail-ability of the element in the fertilizer (Agricultural Com-pendium, 1981) (see table below).

In the year 2005, expected losses would be larger be-cause the agricultural area is expected to grow. At agrowth rate of two percent per year, the total value of nu-trient losses in the year 2005 would be: 1.025 × $10.6 mil-lion = $11.7 million.

The annual value of nutrient losses should be interpretedcarefully. On the one hand, minor elements are not in-cluded in the calculations. Neither are losses of organicmatter. This suggests that the present value underesti-mates nutrient losses. On the other hand, not all nutrientsthat are lost need to be replaced and fertilizers may be acostly way to replace nutrients. This suggests that thepresent value overestimates losses.

Total Nutrient Losses (tons per year)

ElementGood Rainfall

AreasNatural

PasturesIrrigated

AreasFlooded

AreasLosses in

Central Region

NP OK OCentral Region'sshare of areas

2 5

2

33,74910,87629,662

35%

5,3041,3261,874

19%

141100274

13%

———

28%

12,8384,072

10,773

ElementNutrient Loss(tons per year)

Fertilizer Price(dollars per ton)

NutrientAvailability

(correction factor)

Value ofNutrient Loss

(millions of dollars)

NP OK O

2 5

2

Total value

12,8384,07210,773

230170112

42% (2.38)49% (2.04)55% (1.82)

7.01.42.2

10.6


41

the chance of resolving, for example, all soil deg-radation problems through research is certainlynot similar to that of achieving one percentgrowth in crop yield.

The share of benefits that can be realisticallyachieved through research may be larger than,equal to, or smaller than the one percent referenceestimate, depending on the amenability of thebenefit dimension to solution by research. Thisdepends on two factors:

1. The extent to which the constraints related toa benefit dimension are researchable in na-ture. Constraints to increased production orimproved resource management may becaused by factors outside the research domain,for example, by poor infrastructure, deficientlegislation, inadequate institutional develop-ment, or a poor education system. If develop-ment is principally constrained by such issues,agricultural research will not have major im-pact. In most circumstances, researchable con-straints affect crop production strongly. Forlivestock, the institutional environment (credit,extension) appears to be very significant. Forforestry, research solutions are important, butlegal elements and education strategies mayalso play a large role. For soil quality andwater management, nonresearchable con-straints such as poor infrastructure, absent leg-islation, and poor education and institutionalarrangements tend to dominate. By reviewing

the relative importance of the researchable andnonresearchable constraints identified in step3, the planning group may assess this factor.

2. The ability to deliver new technologies to us-ers. Knowledge-intensive technologies arenormally difficult to transfer, whereas physi-cally embodied technologies (seeds for exam-ple) can be transferred more easily. Capitalintensive technologies (e.g., new livestock, ero-sion control investments, or irrigationschemes) can often only be offered on a lim-ited scale and require extensive organization.If there is a strong extension system, it in-creases the probability of obtaining high bene-fits through such research solutions. Table 3summarizes the degree to which these factorswill normally affect the feasibility of research.

Ranking these factors, one would expect that theimpact of research would be, on average, greatestin cropping activities, followed by livestock, for-estry, water management, and soil quality. Thesedifferences suggest the need to weight the calcu-lated values of the size of the different benefit di-mensions. For example, the value for the livestockbenefit dimension (based on the one percent refer-ence value) could be multiplied by 1.5 because ofthe strong amenability to technological progress.For the soil quality dimension, the value might bemultiplied by a much smaller correction factor,say 0.5, because of lower research feasibility.

Box 14: Estimating off-site impact of soil quality research

Two principal off-site effects of soil degradation wereidentified in Benin:

1. River and reservoir sedimentation reduces fish habi-tat and therefore the catch principally in the south ofthe country. Since fish are the principal source of ani-mal protein, this effect is strongly felt.

2. Sedimentation causes frequent flooding in the south-ern departments.

Between 1989 and 1994, the average fish catch in thesouth of Benin dropped by 2680 tons per year. At a priceof $1.8 per kilogram (the local fish price), this works outto a loss of $4.8 million per year. Half the annual loss($2.4 million) was attributed to sedimentation, the otherhalf to overfishing.

The expected costs of increased flooding could not beestimated as easily. It was decided to double the costs ofsedimentation to account for the losses due to flooding.The total costs of sedimentation were therefore estimatedat $4.8 million per year. Sedimentation is expected to in-crease with the area under agriculture. The growth rate of

the agricultural area was estimated at two percent. Forthe year 2005 the expected losses were thus calculatedas 1.0210 × $4.8 million = $5.8 million.

Losses to fishery and losses due to flooding are indirectlycaused by agricultural activities on all land that drains tothe south of the country. This includes all of the southernand central regions and 30 percent of the land in theNorthern Region. Taking into account the surface areasof the respective regions, it was estimated that 38 per-cent of sedimentation losses should be attributed to agri-cultural activities in the south, 28 percent to the center,and 34 percent to the north.

The cost of sedimentation in the year 2005 attributed toagricultural activities in the Central Region was thus esti-mated as $5.8 million × 0.28 = $1.6 million.

Using the replacement approach, the total size of the soilquality benefit dimension (BD) is then estimated at $1.6million per year (off-site impact) plus $11.7 million peryear (on-site benefits), for a total value of $13.3 millionper year.


42

Estimating the share of the regional programin total research benefits

Once we know the feasibility of research for eachbenefit dimension, we still need to define the re-gional program’s share in overall research bene-fits. As explained in Part 1, commodity anddiscipline-based research programs may also playa role in the region, and they may be more or lessimportant than the regional program. Therefore,only part of the total future research benefits thatwill accrue to the region can be attributed to theregional program.

The relative importance of the regional programdepends strongly on the strategic considerationsdiscussed earlier. If, at the national level, a deci-sion was taken to share research between a re-gional research program and a commodityresearch program, then the regional program’srole is obviously smaller than if the work hadbeen concentrated only in it. It is difficult to pro-vide generic guidelines on the role of a regionalprogram. In box 15, we show, by benefit dimen-sion, the relative importance of different kinds of

programs in the overall research effort for Benin’sCentral Region.

To sum up, the share of a regional program in thetotal research effort is always between zero andone. For Benin’s Central Region research pro-gram, as implied in box 15, one would expect thefigure for soil quality and livestock research to behigher than that for crops.

Assigning values to the feasibility of research andthe share of the regional program in total re-search-derived benefits is done in plenary by theparticipants of the second workshop. It is a highlysubjective exercise based on their collective wis-dom. At the same time, it is truly a very impor-tant activity, since it defines the relativeimportance that the program will attach to thedifferent types of benefits that it can pursue.Alertness of the workshop participants is critical,since the data base assembled to estimate thebenefit dimensions often contains gaps. At thisstage, workshop participants should considerhow to correct these deficiencies.

Benefit Dimension

Factor Cropping Livestock Forestry Soil QualityWater

Management

1. The extent to which constraints are researchable in nature

2. The ability to deliver new technologies to users

high

high

medium

medium

medium

medium

low

low

low

low

Amenability of benefit dimensions to solution through research is classified as low, medium, and high.

Table 3. Factors influencing the feasibility of research solutions for different benefit dimensions

Box 15: Importance of various types of research programs for the Central Region

Various types of researchprograms play a role in Be-nin’s Central Region. A com-modity research programfocuses on the genetic im-provement of staple foods,cotton and oil palm. There areno separate commodity pro-grams for livestock and for-estry. Some research on soilquality (characterization andmapping) is done by a centrallaboratory, and irrigation re-search is undertaken by thetechnology program. A disci-pline-based program doespostharvest research in both the crops and livestock sectors. Some policy research is done across the agricultural andNRM sector. The international agricultural research centers contribute to some extent to crops and livestock research.Finally, NGOs do some adaptive research in forestry, livestock, and water management. The share of the regional re-search program in the total research effort for the Central Region is smaller for crops than for the other benefit dimen-sions.

Benefit Dimensions

Research Program Types Cropping Livestock Forestry Soil QualityWater

Management 1

Regional programCommodity programsPostharvest programsPolicy research programInternational centersNGOs

mediummedium

lowlowlowlow

high—lowlowlow—

high—lowlow—low

high——low—low

high——low—low

1Water management was not distinguished as a separate benefit dimension in Benin. The column hasbeen added for purposes of illustration.

Importance of Research Program Type for Benefit Dimensions


43

Final estimates of the size of each benefitdimension

Once estimations have been made for the threevariables—initial estimates of the size of eachbenefit dimension, the feasibility of research, andthe share of the regional program in total researchbenefits—a final estimate of the size of each bene-fit dimension can be calculated. This is done bymultiplying the initial estimate of the size of thebenefit dimension by the feasibility of researchand the share of the regional research program intotal research benefits. Thus:

V = BD × fr × rp (3)

where:

V = size of the benefit dimension

BD = initial estimate of the benefit dimension

fr = correction factor for the feasibility of research

rp = correction factor for the share of the regionalprogram in the total research effort that will bene-fit the region

Box 16 shows the sequence for Benin’s CentralRegion.

In equation (3), subscripts have been omitted.Nonetheless, the formula has to be applied toeach benefit dimension selected (c for cropping, lfor livestock, f for forestry, q for soil quality, and wfor water management). Care should be taken toinclude both on-site and off-site benefits in esti-mating the size of BD.

Box 16: Estimating the size of the benefit dimensions

In Benin’s Central Region, initial estimates of the benefitdimensions were made for cropping, livestock, forestry,and soil quality (on-site and off-site benefits). Dollar fig-ures (expressed in millions) are given in the accompany-ing table. In the second workshop the participantsdiscuss the feasibility of research and the share of the re-gional program in total research benefits.

Probability of research success (fr) . For cropping,workshop participants agreed that the benefit dimensionwas quite amenable to research solutions. An additional0.8 percent growth would be feasible.Since the reference value for benefit di-mensions measured in terms of supplywas one percent, the calibration factorwas set at 0.8. For livestock, they foundthis to be overly ambitious, and agreedon a factor value of 0.6. For forestry, theparticipants thought the amenability wasmuch lower. The factor value was set at0.3. Also for soil-quality matters, partici-pants agreed it would never be possibleto solve all degradation problems withresearch results. They agreed on 0.3 as a reasonablecorrection factor.

Share of the regional program (rp). For crop productionit was recognized that a sizable amount of research

would be undertaken by commodity programs and inter-national centers. The contribution of the regional re-search program was therefore set at a modest 0.4. Forlivestock, some university research, some postharvest re-search, and some research efforts by international cen-ters were taken into account (rp = 0.7). For forestryresearch, the participants concluded that, apart from theregional program, there were no major contributors. Theforestry value was therefore closest to one (rp = 0.9). Forsoil quality, the strength of the national soil laboratorywas recognized (rp = 0.75). The resulting potential bene-

fits by benefit dimension are shown in the table. The val-ues shown in the last column were then used to estimatethe potential contribution of identified research projects.

Correction Factor

fr rpFinal Estimate of

Benefit Dimension 1Initial Estimate of

Benefit Dimension 1Benefit Dimension

Crop productionLivestock productionForestry productionSoil quality

21.51.86.1

13.3

0.80.60.30.3

0.40.70.90.75

6.90.81.73.0

1Millions of dollars.

Assigning projects to benefit dimensions and calculating their expected impact

Research projects may contribute to one or morebenefit dimensions. For example, an agroforestryproject may improve crop production, forestryproduction, and soil quality. In principle, all pro-

jects could be evaluated for each benefit dimen-sion, but this would require more time than isavailable to workshop participants. It is more effi-cient for the program committee to identify, be-


44

fore the priority-setting workshop begins, thebenefit dimensions to which research projectscontribute.

Then, in the workshop, working groups areformed — one for each benefit dimension — toassess the likely impact of each project. Thoseprojects that contribute to more than one benefitdomain are assigned to more than one workinggroup for analysis. Following the approach usedby Collion and Kissi (1994), each working groupdistributes 100 points among the various projectsthat impact their assigned benefit dimension. Inthis way, each project’s share of total projectedbenefits for that dimension is expressed as a per-centage.

In its preliminary work, if the program committeehas doubts about a project’s impact on a particu-

lar benefit dimension, it is best for it to go aheadand assign it to that dimension anyway, as well asto those dimensions which the project clearly im-pacts. After all, the responsible working groupcan always decide to set the value of the project’simpact at zero if its members agree the project re-ally does not affect that particular dimension.

Once each project’s impact on one or more benefitdimension has been estimated as a percentage orpercentages, their total impact in dollars can becalculated. Percentages are simply multiplied bythe potential benefit (V expressed in dollars) foreach benefit dimension and the products areadded. Box 17 shows the calculations for a set ofprojects in the regional program for Benin’s Cen-tral Region. The projects are the same as thosehighlighted in Box 11.

Estimating other project parameters

Once the value of each project’s potential contri-bution is known, we can continue the priority-set-ting exercise following the Collion and Kissiapproach. Now each project’s chance of successand the expected rate of adoption of the resultshave to be defined. (See equation 2, page 36.) Toget the final project score, the overall product isdivided by the expected cost of the project.

Chance of success “s”

For any project, the value of s, the chance of suc-cess, lays between zero and one. Many factors in-fluence the determination of the value:

Complexity of the research project. If a projectrequires the collaboration of many disciplines,and the sequencing of many steps, it is more com-plex and this reduces its chance of success.

Potential for building on research results fromother regions. If similar problems have been re-solved elsewhere, the chance of success increases.

Progress on the project. If certain elements re-quired to solve the problem have already beenidentified, the chance of success increases.

Available human resources. Competence andmotivation of the researchers involved increasesthe project’s chance of success.

Techniques and methodologies required. If aproject is to utilize advanced techniques not yetroutinely used, the chance of success declines.

Project duration. The longer the project’s dura-tion, the greater the chance that it will be discon-tinued before its successful completion.

Rate of adoption “a”

The value of a, the rate of adoption, also varies be-tween zero and one. Factors influencing adoptionrates for different technologies are as follows:

Profitability of a technology. If users have tomake only a small investment for a large return,they will be inclined to adopt the technology.

Risk. If the payoff of a technology varies widelyfrom year to year, and if users run a big risk oftaking a loss in the year of adoption, then thechance of adoption declines substantially.

Complexity of a technology. If users have tomake a major effort to master the technology, thechance of adoption drops.

Capacity to deliver. If the extension service caneasily deliver the technology or customize it forthe user group the adoptablity increases. Espe-cially for complex technologies, the presence of acapable extension service helps improve theadoption rate.


45

Availability of accompanying inputs. If, for acertain technology, required inputs are not easilyfound, the rate of adoption declines.

Infrastructure. In areas with poor roads and otherproblems of infrastructure, new technologies tendto diffuse more slowly.

Estimates of the chance of research success andthe potential rate of adoption of research-derivedtechnologies are made during the second work-shop, by the working groups, based on the abovecriteria. If a project is being evaluated by twogroups (because it contributes to two benefit di-mensions), both groups give estimates of chanceof success and the potential rate of adoption,based on the factors discussed above. If the re-sults are similar, the average can be taken. If not,the estimates can be discussed and compared in aplenary session to confirm the accuracy of the fi-nal value.

Obtaining good estimates of chance of successand rate of adoption is very important. Both fac-tors are included in a multiplicative fashion in theequation to measure benefits. Their estimatestherefore, strongly affect the final outcome. Twotools may be applied to obtain estimates that areas reliable as possible:

1. Rather than making estimates as a group,working groups may split up in couples. Cou-ples discuss the projects and provide initial es-timates. The results of the different couples arecompared afterwards, and for those projectswhere major differences are found, a discus-sion is held and new estimates are suppliedthat satisfy all people present. For the projectswhere the differences are small, the average istaken.

2. Once all projects are assessed, estimates for asubsample of projects (e.g., five projects) are

Box 17: Calculating the impact of identified research projects in Benin’s Central Region

To arrive at the expected value of an identified researchproject’s impact (expressed in dollars), the project’s im-pact on each benefit dimension (expressed as a percent-age) is multiplied by the contribution of that dimension tothe total potential benefits. The results are then added.For example, the contribution of the project on bush firecontrol to the various benefit dimensions is 0.05 x 1.7(for forestry) plus 07 x 3.0 (for soil quality), for a

(rounded) total of 0.30. Expressed in dollars, this projecthas an expected annual value of $300,000.

Because the potential benefits (V) in the crops and soilquality dimensions are highest, projects that contribute tothese are more apt to have high values of total expectedimpact.

Benefit Dimension

Cropping Livestock Forestry Soil Quality

Vc

6.9

Vl

0.8

Vf

1.7

Vq

3.0

Project

Expected Value ofProject's Impact

(millions ofdollars per year)pc pl pf pq

Bush fire controlAnti-erosion methodsFallow managementCrop rotation systemsSoil mositure conservationLow-cost fertilizationCombining yam and tree speciesManaging natural pasture landsAgroindustrial by-products and

crop residues in animal feedsForage species in crop rotation

SubtotalOther projectsTotal

———6784

—

34

3268

100

———————6

44

1486

100

5108

———6

—

——

2971

100

778967

—5

—4

5347

100

0.300.370.370.680.660.760.370.19

0.240.42

4.47.9

12.2

Symbols are defined in equation 1, page XX. More detail on projects in Box 11, page XX.

Potential Benefits (millions of dollars per year)

Project Contribution (percent)


46

compared. The group now assesses whetherthe rank order of the estimates (from lowest tohighest) is correct. If not, the estimates for theunder- or overrated projects are redone. Afterconsensus is reached for the subsample of five,the place of each other project in comparisonwith the first five is evaluated, and adaptedwhen necessary. In this manner, estimates aremade consistent across the total pool of possi-ble projects.

Cost of the project

Research costs normally consist of personnel,equipment, transport, inputs, land, and infra-structure. At the level of program planning, pro-jects are not sufficiently detailed to make preciseestimates of the different cost components. It is

assumed, therefore, that total costs are propor-tional to personnel costs (which are normally thebiggest cost component). Personnel costs were al-ready estimated during step 5 of program formu-lation, project identification.

Once the chance of success, potential rate of adop-tion, and cost of research projects are estimated,the score for each project can be finalized usingequation (2).

N = [ (Vc × pc + Vl × pl + Vf × pf + Vq × pq +Vw × pw ) × s × a ] / C (2)

The part within parentheses was calculated in theprevious sections and was defined as the ex-pected value of the project’s impact. Box 18 pur-sues the example from Benin’s Central Region.

Box 18: Calculating final project scores for Benin’s Central Region

The final project scores vary widely, the range being 1.3to 32.1. Among the projects for Benin’s Central Region, afew ranked higher, close to 600, and some ranked lower,as low as 0.35. The variability shows how important it isto calculate the scores, rather than rely on only scientists’

spur of the moment judgment. The program would bemaking a grave error if it adopted the project scoring 354instead of one with a score 100 times larger, simply be-cause of the personal interest of one of the people con-cerned.

Project

Final ProjectScore 2

(N)

Expected Value ofProject's Impact

(Vc...pw)1

Rate ofAdoption

(a)

ResearcherRequirements

(months)(C)

Bush fire controlAnti-erosion methodsFallow managementCrop rotation systemsSoil moisture conservationLow-cost fertilizationCombining yam and tree speciesManaging natural pasture landsAgroindustrial by-products and

crop residues in animal feedsForage species in crop rotation

290370370680660760370190240

420

Chance ofResearchSuccess

(s)

0.280.810.690.740.620.420.200.300.47

0.50

0.220.460.540.510.600.890.310.300.53

0.25

1413138

10183

109

7

1.310.610.632.124.615.87.61.76.6

7.5

12Values in thousands of dollars per year.Expected benefits per year for each month of researcher time invested in thousands of dollars.

Defining and interpreting the final project scores

Once project scores have been calculated, theyneed to be put in perspective—that is, interpreted.The final score indicates the expected benefit peryear that would be produced for each month ofresearcher time invested. In the present approachwe have not made any explicit assumptions abouthow long project benefits will continue to be gen-erated. However, let us assume that projects pro-duce a benefit stream over 20 years, which wouldonly begin in 10 years’ time. Let us also assumethat the real interest rate is seven percent and that

all research costs are concentrated at project start-up. Under these conditions, to be profitable a pro-ject should have an annual benefit above 18percent of the project cost.

Based on the project scores, a priority ranking cannow be established. The higher the score, thehigher the priority for implementation. Certainly,however, the ranking has to be reviewed by work-shop participants for its internal consistency andoverall logic in a coherence test.


47

For this purpose, rather than focusing on minutedifferences in project scores, the programmingteam divides the projects into large clusters offirst, second, third, and fourth priority accordingto their relative scores. Dividing lines betweenclusters can be set using two criteria. The first isthe presence of large differences in scores betweentwo projects next to each other in the ranking.Where these differences exist, they may form aneasy criterion to distinguish first from second pri-ority projects, and so on.

The second criteria considers the resources avail-able to the program. A natural distinction is be-tween those projects that can be implementedwith the available resources, and those that can-not be implemented. Projects of first priority willbe implemented under any conditions. These arethe core of the program, the bare minimumneeded for program viability. Second-priorityprojects are ones that program managers willwant to implement if additional resources aremade available. Third-priority projects are ones

that will be implemented only if financed fromoutside sources. If fourth-priority projects havebeen identified, they will, alas, probably never beimplemented!

The ranking is then checked for internal consis-tency. Perhaps workshop participants erred intheir initial judgments. Or maybe certain specialconsiderations were not taken into account by thescoring method and analysis. To correct for suchflaws, the ranking and clusters resulting from theinitial analysis are presented in the workshop anddiscussed among the participants. If good argu-ments are presented, acceptable to all workshopparticipants, a project may be moved up or downin the ranking or from one cluster to another.However, if the ranking is completely overhauled,this raises the issue of whether the method hasbeen applied carefully enough or even whether itwas appropriate for the program concerned. Box19 shows the ranking for 10 projects in Benin’sCentral Region and explains why some projectswere afterwards moved up in the ranking.

Box 19: Defining and interpreting the final priority ranking

The group consensus was that livestock pro-jects were undervalued because the scoringapproach didn’t sufficiently take into accountthe potential for fast growth of the livestocksubsector in the region due to urbanizationpressures in the south. Within the ranking forthe 10 projects listed here, therefore, two live-stock projects—introduction of forage speciesin crop rotations and techniques for usingagroindustrial by-products in animal feed(ranked 4 and 5 at right)—were moved abovethe project on improved anti-erosion methods.

In Benin, costs per researcher (including sal-ary and overheads) were estimated at$60,000 per year, or $5000 per month. At these cost levels, a project should have a score above 900 (18 percent of5000) to be considered for implementation. Since working groups’ estimates of the chance of success, the rate ofadoption, and the costs in research time, tend to be rather optimistic, the limit may be doubled to 1800. In this case,projects ranked as 9 and 10 should never be implemented as they represent poor investment opportunities. The othereight projects appear to be good investment opportunities.

PriorityRanking Score Project Title

12345

6

78910

32.124.615.87.56.6

10.6

10.67.61.71.3

Agronomically feasibleMethods for

methodsIntroduction ofTechniques for using

Improved for differenttypes of terrain

Efficient methodsProductive

techniquesMethods for

crop rotation systemssoil moisture conservation

Low-cost fertilizationforage species in crop rotations

agroindustrial by-productsand crop residues in animal feed

anti-erosion methods

fallow managementcombinations of yam and tree species

Natural pasture managementbush fire control

Organizing the priority-setting workshop

The procedure for setting priorities among pro-jects has now been fully presented. In this finalsection we make some practical recommendationson how to make the exercise go smoothly. We alsoinvite you to consult ISNAR’s Research Manage-ment Guidelines No. 2, titled Guide to ProgramPlanning and Priority Setting. Chapter 5, Approachto Priority Setting, describes techniques for tap-

ping the combined experience of planning groupmembers.

As we have shown, project scoring hinges onthree main factors:

1. Assigning values to various factors that affectthe potential impact of regional research onproduction and resource quality (crops, live-


48

stock, fisheries, forestry, soil, and so on). Thefactors are

● initial estimate of the benefit dimension(BD)

● feasibility of research solutions (fr)

● regional program’s share of total researcheffort (rp).

2. For each project, determining the contributionit will make to each benefit dimension (p).(The contribution of any given project willvary from one benefit dimension to another.)

3. For each project, determining the expectedadoption rate (a), the expected chance of suc-cess (s), and the cost (C).

To make the workshop run smoothly and to takemaximum advantage of the participants’ knowl-edge and skills, we propose the procedure and di-vision of work shown in table 4. Activities aredivided into two stages: those before the work-shop and those during it.

The pre-workshop stage is devoted to estimating,with the participation of development organiza-tions, the size of the benefit dimension (BD) foreach component (crops, forestry, etc.). This stageis also used to make an initial estimate of the costof research (C), in terms of months of researchertime required. The exercise is coordinated, ofcourse, by the program leader. If for some reasonthese tasks have not been carried out before theworkshop begins, researchers should be calledout of the workshop for the time needed to com-plete them.

In the second stage—the actual workshop—par-ticipants work together to validate the BD valuesthat have already been calculated, to quantify thecorrection parameters (fr and rp), and to estimatethe chance of success (s), the expected rate ofadoption (a), and the contribution of each projectto the applicable benefit domains (p).

The validation exercise and determination of val-ues for parameters fr and rp take place in plenary.Estimating the value of parameter a and s is bestdone in working groups, and afterwards dis-cussed and reconciled in plenary. As for quantify-ing the contribution that the project makes to thedifferent benefit dimensions (the p for crops, live-stock, fisheries, forestry, and soils), this task is alsodone in the working groups formed to analyzethe contributions of projects to the benefit dimen-sions.

Before setting up and assigning the groups theirtasks, there is good reason to classify the projectsaccording to the benefit dimensions that they willlikely affect. This may be done by the programleader or by the program committee.

Once the projects have been divided up, thegroups are formed. One group is assigned to eachbenefit dimension. For Benin’s Central Region,four groups were formed: crops, livestock, for-estry, and soils.

It’s important to ensure that developmentauthorities and farmers are well represented inthese groups. Their participation is indispensablein setting the value of the parameters that comeinto play in the project scoring.

Table 4. Organization of the priority-setting exercise

Pre-workshop

During the workshop

1.

2.

3.4.

5.6.7.

A multidiscliplinary group of researchers and development officers estimate the size of the benefitdimension for each of the components or benefit domains, for example, crops, livestock,fisheries, forestry, and soil quality.The same group estimates the cost of each project.

In plenary, workshop participants validate values and estimate, by consensus, the values of and .Working groups estimate the expected value of project impact of and Groups are formed by

dividing participants along the lines of their knowledge and interest.In plenary, participants validate the values of and estimated in the groups.Program committee calculates project scores using equation (2).In plenary, workshop participants conduct the coherence test.

(BD)

(C)

BD fr rp(p), (a) (s).

a s


49

Support projects are dealt with in the same wayas technical projects, with one exception: thevalue of the contribution of support projects to agiven benefit component will be the sum of thevalues assigned to the technical projects they sup-port. Sometimes participants may decide that allsupport projects are high priority. In such cases,

these projects are not subjected to the priority-set-ting procedure.

Finally, the values of all the parameters are com-bined to produce the final project scores and todevelop the project ranking. The ranking is thensubmitted, in a plenary session, to a coherencetest.

2.7 Human Resource Gap Analysis

In step 7, the human resource gap analysis, theprogram committee defines, by discipline, the

number of researchers needed to implement thepriority projects. By subtracting the number of re-searchers available in each discipline from thenumber needed, the gap can be determined. Thegap should be filled by recruitment, by transfer-ring researchers from other programs, or by con-tracting research out to other organizations. Box20 provides estimates of the human resource re-quirements and gaps for top-priority projects inthe program for Benin’s Central Region.

The spectrum of scientific specializations neededin a regional program is wide. This reflects the in-

tegrated nature of such programs. Specialistsfrom the agricultural sciences, animal sciences,and forestry and fisheries are needed, as well aseconomists and sociologists. On the one hand,certain disciplines can be anticipated to have spe-cial importance within a regional program—forexample, crop management, soil science, engi-neering, and ecology. On the other hand, the roleof plant breeding, plant protection, and posthar-vest technology is relatively less. These differ-ences are more marked when regional programsrun side by side with commodity programs,which take care of breeding and plant protectionresearch. The range of specializations within theregional program must, of course, be kept within

Detail of Step 7 of the Program Formulation Process, ‘Human Resource Gap Analysis’

RationaleThe program should be equipped to implement priorityprojects and human resources are central to this

OutputRequired number of researchers by discipline and staff re-quirements per year for priority projects, identification ofgaps between requirements and availability

ResponsibilityProgram committee led by the program leader

Participants

Program committee, human resources department

MethodsMatrix of priority projects by disciplinary requirements and sensitivity analysis

Time and resourcesOne to two weeks after the second workshop


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results




50

reasonable limits. Otherwise, it becomes impossi-ble to recruit the necessary personnel. For exam-ple, when a pasture production specialist, apasture ecologist, a pasture botanist, and a pas-

ture taxonomist are “required”, serious considera-tion should be given to having the assigned tasksdone by a single pasture scientist.

2.8 Preparing for Implementation

Once projects are identified and prioritized,and human resource implications have been

assessed, the regional program is ready to be putin place. Implementation may occur under one oftwo scenarios:

First, if the program focuses on a new area of re-search, the activities to be carried out under thepriority projects are implemented gradually as re-searchers are recruited or reassigned to the newprogram and as financial resources are mobilized.

Second, if the program already exists and theplanning exercise has served principally to reori-ent research activities, the priorities establishedby consensus during the planning exercise mustbe reconciled with activities already in progress. Ifresearch in progress matches an established prior-ity, it must continue within the framework of thereoriented program and receive the necessarysupport. However, if projects in progress do notmatch an established priority, but have beengoing on for quite some time and are close tocompletion, it is best to complete the work—espe-

Box 20: Estimating human resource requirements for first-priority projects

Benin’s Central Region research program currently em-ploys an agronomist and a soil scientist. Scientists canbe expected to dedicate 70 percent of their time to re-search (eight months). The remainder will be spent ontraining, administration, organizational duties, leave, etc.Using these estimates, the program will have to recruit

an economist and the soil scientist’s time should beshared with another program, possibly in exchange forpart of a forestry scientist’s or a livestock scientist’s time.The pasture scientist and engineer could be brought inthrough special contracts.

Project Title Agronomy EconomicsSoil

Science Engineering(Agro)

ForestryPastureScience

AnimalScience

Total requirementAvailabilityGapGap in person-years

per annum

2.3

2.02.3

2.5

8.88.00.8

0.1

1.31.01.0

1.0

2.02.01.0

9.3

9.3

1.16

1.0

1.0

2.08.0

-6.0

-0.75

1.0

1.0

1.0

0.13

1.0

1.0

2.0

2.0

0.25

1.0

1.0

1.0

0.13

2.5

2.5

2.5

0.31

Bush fire controlRural credit needs and availabilityImproved use of in-land valley areasSustainable cropping systemsPlant multiplication techniquesYam-based production systemsSociological study of land useAnimal husbandryCommodity system analysisFactors affecting technology adoptionAnimal traction technology use


51

Detail of Step 8 of the Program Formulation Process, ‘Recommendations for Implementation’

RationaleConditions for program implementation need to be cre-ated. All stakeholders should have access to programplanning outcomes

OutputA program document and validation of program by stake-holders and research managers

ResponsibilityProgram leader, program committee, and possibly, pro-gram members

ParticipantsAll those involved in the previous steps

MethodsConsultation and report writing

Time and resourcesFour to six weeks after second workshop


2Constraints

analysis

4Determination


and strategy


projects

6Prioritysetting

7Human



implementation


results



cially if the project has already consumed a largepart of the resources set aside for its implementa-tion. The same holds true for activities that a re-searcher is doing as part of a thesis in preparationfor an academic degree.

With the exception of these two cases, it’s prefer-able to stop on-going research and assign newpriority project activities to the affected re-searchers.

Major shifts in the direction of research don’t oc-cur without difficulties of course. Some scientistswho have experienced such a reorientation haveexpressed frustration because they consideredtheir research work important to the country’s ag-ricultural development. Managers must explainto them that, yes, the activity may be importantbut it is not an immediate priority given morepressing needs of users and the extent of availableresources. Sometimes training is needed to pre-pare researchers to carry out their new duties un-der the best possible conditions.

To ensure successful implementation, it is crucialto have both internal and external support for theprogram. This will already be largely achievedthrough the participation of various stakeholdersin the planning process. But it needs to be rein-forced once the program has been finalized. Theavailability and communication of clear andtransparent information on the program are keyto obtaining this support. The program leadershould therefore take the following steps:

■ Write a program document explaining howthe program came about. Its outline can

roughly follow the eight steps of the planningprocess. In the preface, it may be useful to ac-knowledge the participants since they give theplan a good measure of its credibility. Thedocument should also provide specific recom-mendations on filling the human resource gap,pay special attention to the nonresearchableconstraints identified, and suggest develop-ment measures. Box 21 suggests an outline fora program document with 10 chapters.

■ Stage an internal validation workshop withinthe research institute. This identifies potentialcollaborators and possible overlaps with otherprograms and may also serve to assess theprogram’s scientific merit. The internal valida-tion shouldn’t be too concerned with the pro-gram’s relevance, as the different stakeholderspresent in the planning workshops will havealready established this.

■ Present the program to the various stakehold-ers (external validation). In this presentation,encourage questions on the program’s rele-vance. If the program has been prepared withcare, the validation workshop will only help tostrengthen it, in part by building commitmentamong possible research partners in the stake-holder community. Highlight the nonre-searchable constraints since these demonstratethe importance of integrating research withother development activities. An extensivediscussion on the complementary develop-ment measures needed and on their feasibilitywill greatly enhance the viability of a regionalprogram. Are the legal arrangements or thepolicy guidelines suggested by the research


52

planning team feasible? Should the researchteam anticipate other changes in the environ-ment? If the complementary measures are fea-sible, try to obtain a commitment from otherorganizations to try to implement such meas-ures. Similarly, discuss the usefulness of regu-lar contact between research and developmentorganizations and whether certain activitiescould be undertaken together. A discussion ofthe role of development will lead many stake-holders to appreciate that the program’s archi-tects recognize the limits of research; and theywill leave the meeting with new ideas for theirown development work.

■ Finally, once the program has been validatedinternally and externally, the program leadershould ask program staff members to start de-veloping the research projects in more detail.This means identifying research methodolo-gies, resource requirements beyond scientificstaff, budgets, and monitoring and evaluationmechanisms. Once available, data on the pro-jects should be entered in the institute’s man-agement information system to avoidduplication in future planning exercises.


53

Table of Contents

10. Conclusion (1-2 pages)

References

highlights of principal outcomes

Preface

Executive summary

1. Introduction (1-2 pages)purpose of documentinstitutional contextplanning context (long term, medium term, annual plan)participantsacknowledgments

2. Sector review (7-10 pages)production and consumption information, resource use, production and marketing structure, etc.regional development objectives

3. Evaluation of past research (5-7 pages)within the institutewithin the countryin other countries

4. Constraints analysis (10 pages with constraints tree, 4 pages of text)presentation of constraints treeexplanation of constraints tree (concise)identification of constraints that are amenable to research solution

5. Determination of research objectives (3-5 pages)description of principal objectives for research program as derived from the constraints treedescription of derived objectives (principal research fields that need to be addressed)

6. Identification of research projects (10 pages)description of research projects to address the constraints, expected research results,activities to execute these projectsclassification of research projects into technical projects and support projectsestimation of project duration (in years)estimates of human resources needed by discipline for each project

7. Priority setting among research projects (10 pages)explanation of method used in priority settingestimates of parameters required in priority-setting exerciseshowing and explaining the results of priority-setting exercise (ranking)description and explanation of changes made to initial rankingjudgments on support projects

9. Policy recommendations (4 pages)review nonresearchable constraints and make recommendationssuggest institutions best placed to pursue recommendations

8. Resource gap analysis (4-6 pages)on the basis of the list of priority projects and indications by management on the size

of the program, identify:– human resource requirements– physical resource requirements– location– comparison of resource requirements with resource availability and identifications of gaps– proposals for filling resource gaps (recruitment, training, building, purchasing, etc.)

Box 21: Suggested outline for a program formulation document


54

PARTConcluding Remarks on Regional Research Planning 3

3.1 Nature and Time Requirements of the Planning Process

Regional research programs can be an effec-tive tool for responding to the growing de-

mand for better integration of NRM concerns inthe agricultural research agenda. By developing ashared perspective on the characteristics, con-straints, and potential of a region, regional plan-ning combines a holistic diagnostic perspectivewith a focused research plan.

The regional planning approach examined in thisdocument combines subjective and objective in-formation from different sources. It uses a quanti-tative, economics-based procedure to prioritizeresearch projects. But it does not pursue very de-tailed and precise information on the differentprojects. Neither does it apply advanced and de-tailed project evaluation methods. Rather, it as-sumes that the confrontation of these roughresults with the subjective judgments of the mem-bers of the planning group is the best way to im-prove relevancy and orientation of the program.Thus the type of estimation procedures proposedand the extent of participation required arestrongly intertwined. Without participation, theestimation procedures do not hold sufficientguarantee for reliable results. But with well man-aged participation, further sophistication of theprocedures would not necessarily do justice to thecombined expertise of the planning group. The fi-nal procedure is a mix between analysis, expert

knowledge, economic estimations, creativity, andteamwork.

Whether the amount of detail provided in theseguidelines is sufficient under all conditions is dif-ficult to assess. When regional research programsdeal with larger mandates (bigger regions, moreproduction), one might expect that the level of in-vestment in the exercise might be higher than thatsuggested in this document, both in terms of thenumber of people that participate, and in terms ofthe amount of data collection and processing. Inthe suggested form, the time requirements areroughly as follows: the planning group of some20 people would be involved for roughly fivedays, for a total of 100 days. The program com-mittee of three members would be involved for 12weeks each (60 days), for a total of 180 days. Re-searchers would spend two days to provide infor-mation for each project, or approximately 100days. And a stakeholder group of about 25 per-sons would assess the plan during one day, total-ing approximately 25 days. The timeconsumption would then be 405 days, or 81weeks, roughly comparable to two person years.If the exercise is used to plan a program of, let ussay, 10 full-time researchers for a period of eightyears, the total costs of planning the programwould be roughly 2.5 percent of the resources thatthe program will spend.

3.2 Future Problems and Opportunities

Regional programs may start absorbing tradi-tional farming systems units (or on-farm re-

search units). They will have a mix of on-station

and on-farm research components, depending onthe specific constraints addressed by each re-search project, but not depending on the organ-

Concluding Remarks on Regional Research Planning

57

izational unit in which the research is done. Re-gional programs tend to soften the rigid linear se-quence of basic, strategic, applied, and adaptiveresearch, especially in the applied and adaptivestages. The feedback loop between farmers andresearch is shortened. Most researchers in a re-gional program will likely have experience withon-farm participatory methods, though not all tothe same degree.

Regional research programs must be accountableto more than farmers. They should pay attentionto the needs of other resource users as well. Someof these are water management authorities andnonagricultural households. They should also beaware that their support will have to come fromregional authorities and other representatives ofregional interests. The program may feel pres-sures from many sides but should be well placedto address them.

Decentralization

Regional programs normally require a measure ofdecentralization and often they will be based out-side the capital city. This is justified because inmany countries the concentration of research injust a few central locations results in scientists be-ing unaware of the wide variety of conditionsfaced by farmers elsewhere in the country. Thisundermines their effectiveness as problemsolvers. Nevertheless, decentralization must notbe taken to extremes. For researchers to maintaintheir skills and qualifications, they need to belongto a peer group. Decentralization can also reduceliving standards for researchers and their families.The difference between daily life in the capitalcity and daily life in the countryside can be enor-mous, especially in developing countries. If de-centralization is pushed beyond, say, the secondtier of cities (e.g., provincial or departmental capi-tals), researchers may start looking for alternativeemployment. Of course, the best researchers willbe the ones to find it most easily.

Effective implementation of regional programsdemands adequate resources. Transportation is akey factor. It may be as important as having a

well-equipped research station—perhaps evenmore important.

Human resources are the heart of any researchprogram. Regional programs demand a mix ofmany disciplines. This, in turn, means payingspecial attention to people management, espe-cially taking measures to focus different kinds ofscientists on shared objectives and outputs(Janssen and Goldsworthy, 1995).

What about program management? The regionalapproach to research affects it in two very differ-ent ways. On the one hand, the wide spectrum ofscientific subjects and issues to be dealt with com-plicates management. On the other hand, theproximity to users of research results facilitates it.If researchers can’t clearly explain to farmers andprogram leaders why certain problems are ap-proached in certain ways, it’s probably becausethey have not defined them properly. The varietyof issues that come to play in a regional programprovides another argument for developing suchprograms in a systematic fashion.

Questions for further development

The procedure described in these guidelines isstill open to improvement. We would thereforelike to end with some of the questions that con-fronted us while we were developing and testingthe approach.

Is it possible to make the program planningprocess more participatory? Rapid rural apprais-als could be used in the regional analysis (step 1)and in the identification of researchable con-straints (step 2). The challenge in strengtheningparticipation is to maintain a representative over-

view of the region and not be biased toward theneeds of small pockets of farmers.

Is workshop participants’ knowledge adequatefor effective regional program planning? Wefound that even in carefully selected groups onlya few people could speak with some depth ofknowledge about the range of issues in a particu-lar region. If such people are absent, participatoryregional planning may be biased toward the inter-ests of over-represented groups or it may be diffi-cult to achieve consensus. Our solution was to


58

emphasize a rigorous regional review by meansof secondary data and the development of a re-gional profile. But the extent to which partici-pants are able to absorb this information at thebeginning of the planning process should not beexaggerated.

How can the results of the planning exercise beexploited for purposes other than research? Theplanning method does make provision for formu-lating development (i.e., nonresearch) measuresto deal with nonresearchable constraints (part ofstep 4). However, it may be useful to go beyondthis step by organizing a workshop for the maindevelopment organizations working in the re-gion. For the sake of credibility, such an initiativewould have to be supported by the regional gov-ernment.

What is the value of exploiting GIS beyond itsrole as an aid in setting the overall boundariesof a region? The program formulation method as-sumes that regions are relatively homogenous.On this assumption, one could argue that furtherzonation through GIS may not have much to of-fer. Nevertheless, we are aware that the impact ofidentified constraints on agricultural develop-ment varies among different groups within a re-gion. GIS could be used to differentiate in moredetail the likely beneficiaries of research. Thismight improve the quality of priority settingamong projects.

Should the priority-setting framework (step 6)not be more concerned with questions of timing,especially since NRM is often concerned withlong-term effects? Developing procedures that doa better job of incorporating discounting princi-ples is not really a problem, especially whenspreadsheets are used. In fact, we first developeda priority-setting approach that accounted forbenefit streams over 50 years. To streamline theprocess, we went back to the present nondis-

counted method. But we do encourage users ofthis manual to pursue more sophisticated dis-counting procedures if they believe that such a re-finement would be welcomed by their programformulation group.

Is there a way to improve experience-based as-sessments of the research project’s chance forsuccess (part of step 6)? Under the current plan-ning method, many assessments reflect the edu-cated guesses of working-group members simplybecause information on the feasibility of futuretechnology development is so scant. ISNAR islooking into ways to fill part of this gap. But ourexpectation is that project feasibility assessmentswill remain largely subjective, even if they are fedwith better background information. For the mo-ment, we believe the best way to improve assess-ments is by identifying unambiguous researchprojects and specific research activities. Validatingjudgments on these activities is easier than onoverall projects and this improves the quality ofthe resulting program.

Is it realistic to expect regional programs to befully implemented? This depends largely on howflexible the programs are and on the availabilityof resources. New or revised research programsare constrained by previous commitments and bystaffing patterns that cannot be changed immedi-ately. However, if a solid program is presented,the chances of finding the necessary resources in-crease markedly. It should be obvious that the ex-pected success and vigor of a program increase ifit has been properly planned.

We hope the planning approach presented inthese guidelines will be useful to you. If you havethe occasion to apply them in whole or in part,please share your experience with us. This willcontribute significantly to future refinements ofthe method.

Concluding Remarks on Regional Research Planning

59

References

References

Agricultural Compendium for Rural Develop-ment in the Tropics and Subtropics. 1981. Pro-duced and edited by ILACO B.V., InternationalLand Development Consultants, Arnhem, TheNetherlands. Amsterdam, The Netherlands:Elsevier Scientific Publishing Company.

Alston, J.M., G.W. Norton and P.G. Pardey. 1995.Science under Scarcity. The Hague, The Neth-erlands: International Service for National Ag-ricultural Research (ISNAR), and Ithaca NY,U.S.A.: Cornell University Press.

Arntzen, J. 1995. Economic instruments for sus-tainable resource management: the case of Bot-swana’s water resources. Ambio 24 (6): 335-342.

Ashby, J.A., R.D. Estrada, and D. Pachico. 1994.An evaluation of strategies for reducing natu-ral resource degradation in the hillsides oftropical America. Cali, Colombia: Centro Inter-nacional de Agricultura Tropical (CIAT).

Bartelmus, P., E. Lutz, and J. van Tongeren. 1994.Environmental accounting: an operational per-spective. In Valuing the Environment. Proceed-ings of the First Annual InternationalConference on Environmentally SustainableDevelopment, held at the World Bank, Wash-ington, D.C. September 30 - October 1, 1993.Pp. 47-57. Eds. Serageldin and A. Steer. Wash-ington, D.C.: The World Bank.

Block, S.A. 1995. The recovery of agricultural pro-ductivity in sub-Saharan Africa. Food Policy 20(5): 385-405.

Boughton, D., E. Crawford, J. Howard, J. Oehmke,J. Shaffer and J. Staatz. 1995. A strategic ap-proach to agricultural research planning inSub-Saharan Africa. Mimeograph. EastLansing, Michigan: Department of Agricul-tural Economics, Michigan State University.

Brush, E. 1995 Research Remuneration Policy.Draft Chapter sumbitted to: Tabor, S., Janssen,W. and Bruneau, H. 1997. Sourcebook on Fi-

nancing Agricultural Research. ISNAR, theHague, in preparation.

Collion, M.-H. and A. Kissi. 1994. Guide d’élabora-tion de programmes et d’établissement de priorités.Research Management Guidelines n° 2. TheHague: International Service for National Ag-ricultural Research (ISNAR).

Contant, R.B. and A. Bottomley. 1988. Priority set-ting in agricultural research. Working Paperno.10. The Hague, The Netherlands: Interna-tional Service for National Agricultural Re-search (ISNAR).

Corkindale, J. and K. Willis (eds). 1995. Environ-mental valuation. New perspectives. Walling-ford, U.K.: CAB International.

Crosson, P. and J.R. Anderson. 1994. Concerns forsustainability: integration of natural resourceand environmental issues in the research agen-das of NARS. ISNAR Research Report No. 4. TheHague, The Netherlands: International Servicefor National Agricultural Research (ISNAR).

Dagg, M. and F. Haworth. 1988. Program formu-lation in national agricultural research. Work-ing Paper no.17. The Hague, The Netherlands:International Service for National AgriculturalResearch (ISNAR).

Duivenbooden, N. van. 1995. Land use systemsanalysis as a tool in land use planning, withspecial reference to North and West Africanagro-ecosystems. Doctoral thesis. Wageningen,The Netherlands: Wageningen AgriculturalUniversity.

Franzel, S., H. Jaenicke and W. Janssen. 1996.Choosing the right trees: Setting priorities formultipurpose tree improvement. ISNAR Re-search Report no.8. The Hague, The Nether-lands: International Service for NationalAgricultural Research (ISNAR).

References

63

GTZ. 1988. ZOPP: An introduction to the method.Eschborn: Deutsche Gesellschaft für Technis-che Zusammenarbeit GmbH (GTZ).

INRAB. 1995. Plan Directeur de la RechercheAgricole du Bénin. Volume I. Politique Nation-ale de Recherche Agricole. Cotonou, Bénin: In-stitut National des Recherches Agricoles duBénin (INRAB).

INRAB. 1996. Plan Directeur de la RechercheAgricole du Bénin. Volume II. Plan deDéveloppement à Long Terme. Cotonou,Bénin: Institut National des Recherches Agri-coles du Bénin (INRAB).

Janssen, W. 1995a. Characteristics of NRM re-search: institutional and management implica-tions. IRDCurrents No. 9. Uppsala, Sweden:International Rural Development Centre(IRDC), Swedish University of AgriculturalSciences.

Janssen, W. 1995b. Aggregating economic knowl-edge for use at national, regional, sector, andfarm level. In Eco-Regional Approaches for Sus-tainable Land Use and Food Production. Pp. 143-163. Eds. J. Bouma, A. Kuyvenhoven, B.A.M.Bouman, J.C. Luyten, and H.G. Zandstra. Dor-drecht, The Netherlands: Kluwer AcademicPublishers.

Janssen, W. and P. Goldsworthy. 1995. Integratingand managing multidisciplinary teams in agri-cultural and NRM research. Briefing Paper No.23. The Hague: International Service for Na-tional Agricultural Research (ISNAR).

KARI. 1995. Development of a zonation schemefor the Katumani Regional Research Priority-Setting Exercise. Nairobi: Kenya AgriculturalResearch Institute (KARI).

Lundgren, A.L. et al. 1994. Planning and manag-ing forestry research: A self-learning course(module 5) developing the research program.Minnesota, U.S.A.: IUFRO, University of Min-nesota.

Mendelsohn, R., W.D. Nordhaus and D. Shaw.1994. The impact of global warming on agri-culture: a Ricardian analysis. The American Eco-nomic Review 84 (4): 753-771.

Mills, B.F. and D.D. Karanja. 1997. Processes andmethods for research programme priority set-ting: the experience of the Kenya Agricultural

Research Institute Wheat Programme. FoodPolicy 22: 63-79.

Norse, D. and R. Saigal. 1993. National economiccost of soil erosion in Zimbabwe. In Environ-mental Economics and Natural Resource Manage-ment in Developing Countries, p. 229-240. Ed. M.Munasinghe. Washington, D.C.: Committee ofInternational Development Institutions on theEnvironment (CIDIE).

Pearce, D.W. 1994. Valuing the environment: pastpractice, future prospect. In Valuing the Envi-ronment. Proceedings of the First Annual Inter-national Conference on EnvironmentallySustainable Development held at the WorldBank, Washington, D.C. September 30 - Octo-ber 1, 1993. Pp. 47-57. Eds. Serageldin and A.Steer. Washington, D.C.: The World Bank.

Pimentel, D., C. Harvey, P. Resosudarmo, K. Sin-clair, D. Kurz, M. McNair, S. Crist, L. Shpritz,L. Fitton, R. Saffouri, and R. Blair. 1995. Envi-ronmental and economic costs of soil erosionand conservation benefits. Science 267: 1117-1123.

Schubert, B., U.J. Nagel, G.L. Denning and P.L.Pingali. 1991. A logical framework for plan-ning agricultural research programs. LosBaños, The Philippines: IRRI.

Serageldin, I. and A. Steer (eds.). Valuing the En-vironment. Proceedings of the First Annual In-ternational Conference on EnvironmentallySustainable Development held at The WorldBank, Washington, D.C., September 30 - Octo-ber 1, 1993. Environmentally Sustainable De-velopment Proceedings Series No. 2.Washington, D.C.: The World Bank.

Smaling, E.M.A., L.O. Fresco and A. de Jager.1996. Classifying, monitoring and improvingsoil nutrient stocks and flows in African agri-culture. Ambio 25: 492-496.

Springer-Heinze, A. 1994. Addressing agriculturalsustainability through research: Planning anddecision making issues at the national level.PhD dissertation. Stuttgart, Germany: Hohen-heim University.

Stoorvogel, J.J. and E.M.A. Smaling. 1990. Assess-ment of soil nutrient depletion in sub-SaharanAfrica: 1983-2000. Volume I: Main Report (sec-ond edition). Report 28. Wageningen, TheNetherlands: The Winand Staring Centre forIntegrated Land, Soil and Water Research.


64

Tripp, R. and J. Woolley. 1989. The planning stageof on-farm research: Identifying factors for ex-perimentation. Mexico D.F., Mexico: CIMMYT,and Cali, Colombia: CIAT.

Van der Pol, F. 1992. Soil mining: an unseen con-tributor to farm income in Southern Mali. Am-sterdam, the Netherlands: Royal TropicalInstitute.

Vernon, R. 1995. INFORM: An information systemfor agricultural research management. BriefingPaper No. 21. The Hague: International Servicefor National Agricultural Research (ISNAR).

Wood, S. and P.G. Pardey. 1995. Supporting agri-cultural research policy and priority decisions: aneconomic-ecologic systems approach. In Opportu-nities, Use, and Transfer of Systems Research Methodsin Agriculture to Developing Countries. Proceedingsof an international workshop on systems researchmethods in agriculture in developing countries, 22-24November 1993, ISNAR, The Hague. Pp. 45-81. Eds.P. Goldsworthy and F. Penning de Vries. Dor-drecht, The Netherlands: Kluwer Academic Pub-lishers.

References

65

Glossary

Glossary

Formulating a research program involves a vari-ety of partners, usually from different institutions.These program architects need to speak the samelanguage. That is, they must agree on the mean-ing of the words they use. This isn’t always easy.Experience has shown that terminology differsgreatly from one country to another, between in-stitutions, and even within a single institution. Itdoesn’t really matter what terms are used. Theimportant thing is for members of a single pro-gramming group to have a common vocabulary.So at the start of a programming exercise theyshould pay attention to standardizing the termi-nology.

To help with this, some definitions are suggestedbelow. They are organized alphabetically. The 12terms bearing an asterisk (*) comprise the mini-mum vocabulary for which group membersshould have a shared understanding of meanings.Italicized terms within definitions are ones thatare themselves defined elsewhere in the glossary.

Agroecological zone: Contiguous geographicalarea that is homogenous in climate and soils.Agroecological zones can be useful for identifyingtarget areas for technology generation.

Benefit dimension: The main economic activitiesor resources that are affected by the regional re-search program. Cropping, livestock, forestry, soilquality, water management, and fishery are exam-ples of possible research benefit dimensions. The“size of a benefit dimension” is defined as the dif-ference between, on the one hand, the expectedvalue of a region’s agricultural production and re-sources and, on the other hand, their total poten-tial value if all research constraints are removed.Signified by the variable “BD.”

Central constraint on a subsector: From a devel-opment perspective, this constraint embodies allthe problems blocking the development of an ag-ricultural domain, region or subsector’s potential.Poor income-generating capacity is an example ofa central constraint.

Coherence test: The final step in setting prioritiesamong projects. The test consists of examining theranking of research projects based on their scores toensure it is logical and coherent. It serves to iden-tify differences between the ranking that resultsfrom applying the priority-setting equation andthe subjective or intuitive ranking by planninggroup members. Any divergence is discussed. Thiscan lead to a change in the weighting of the crite-ria used to evaluate the project whose score is be-ing questioned. On the other hand, the discussionmay simply demonstrate deficiencies in groupmembers’ intuitive evaluation.

The coherence test, then, allows the group to fine-tune the results of the priority-setting exerciseand to correct for weaknesses in the method beingused and in its application. It often stimulatesfruitful debate and improves the degree of con-sensus.

Complementary development measure*: A rec-ommended action, aimed at policymakers, that islikely to promote the adoption of new technolo-gies. Such a measure can be institutional or eco-nomic.

Institutional measures include organizing profes-sions, creating associations, or restructuring pub-lic services.

Economic measures include building infrastruc-ture, importing inputs or equipment, supportingexport incentives, price interventions, and tax re-form.

In Morocco, for example, three complementarydevelopment measures were advocated to pro-mote the adoption of technologies coming out ofthe country’s olive research program:

■ the creation of “tree centers” for distributingplant material

■ organization of the tree nursery profession

■ a public information campaign to avoid “sell-ing on the tree.”

Glossary

69

Constraint*: A situation or factor that preventsproduction potential from being fully achieved.This potential may be based on extending the areaunder cultivation, increasing yields, cutting pro-duction costs and losses, or raising value addedby processing and packaging.

There are four main types of constraints:

■ physical, that is related to climate (rainfall, tem-perature, wind, air humidity), terrain, soil(depth, texture, composition, etc.)

■ socioeconomic, that is related to the land tenuresystem, market structure, and pricing andcredit policies

■ institutional, that is related to marketing prac-tices and input distribution

■ technical, that is related to production, process-ing, packaging, or storage

Constraints tree: A hierarchy of research and de-velopment (R&D) problems originating from thecentral constraint. It is a tool for systematicallyanalyzing such problems, allowing for a partici-patory approach to formulating research projectsand programs.

The starting point in building a constraints treefor a particular subsector is to identify the centralconstraint. The causes and effects of all factors un-derlying the constraint are then analyzed, alongwith their interdependency. The constraints treetakes the form of a flow chart composed of vari-ous boxes at different levels. Each box representsa separate constraint that is further broken downinto subconstraints at the levels below it.

Development objectives*: Economic and socialaims that a government seeks to achieve for a pro-duction sector, production system, or agroecologi-cal zone. Realistic objectives are defined in light ofcomparative strengths and weaknesses, existingconstraints, and potential technological break-throughs.

Here are some examples of development objec-tives:

■ increasing the level of self-sufficiency in a cer-tain commodity

■ developing a marginal area

■ maintaining or increasing production

■ increasing the value added to production inthe postharvest stage

■ boosting exports

■ increasing income or employment.

Duration of an activity or project: The periodfrom the beginning of the research to achievingthe final results. For a research activity, it is the in-terval between the starting date of the first specificaction and completion of the last. For a researchproject, it is the interval between the date onwhich the first activity begins and the completiondate of the last.

Institutional or interinstitutional program: A re-search program can be conducted by a single in-stitution or by several bodies within the nationalresearch system. In the first instance, the programis institutional; in the second, it is inter-institu-tional.

Location of a research activity: Where the mainactivities are carried out. This can be an experi-ment site, an institute’s or teaching estab-lishment’s laboratory, a private farm, a unit in anagrifood industry. A research activity may haveseveral locations if its specific actions are spatiallyduplicated. This is often the case when the sameresearch objective is being pursued in several agroe-cological zones.

Nonresearchable constraint*: In most cases, suchconstraints have to do with the physical environ-ment. Agricultural research doesn’t provide directsolutions to these problems but it can suggestways to eliminate or reduce their effects. It isn’tpossible to change the climate, the slope of theland, or soil depth. However, research needs todescribe such factors and study their interactionswith those factors that are under the producer’scontrol, such as choice of varieties, technical cal-endar, management.

Socioeconomic and institutional constraints canoften be removed by development work or by so-cial science research. Some of these are beyondthe scope of the research program under formula-tion and, for its purposes, are therefore consid-ered nonresearchable.

Planning by objective: A systematic method forplanning research activities for a given domain orsubsector. Specific intermediate research objec-tives are determined as a means of achieving adefined overall objective. Intermediate objectivesare then translated into research projects composedof research activities, the results of which contrib-ute to the achievement of the research project’sspecific objective.


70

Planning committee: The committee responsiblefor organizing meetings and preparing the neces-sary background information. It is normallyheaded by the research program leader. Two otherresearch program members should be on thecommittee, one with a background in socioe-conomics, the other trained in agronomy. Thefourth committee member should be a repre-sentative from the unit responsible for researchplanning. This person contributes knowledge ofprogram formulation procedures and ensures thatthe program under formulation is consistent withthe overall direction of the national agriculturalresearch system.

Planning group: An analytical, coordinating, andadvisory body whose job is to plan, monitor,evaluate, and adjust research activities for a givendomain or subsector. It is a structure within whicha dialogue is established among researchers andthe users of research results for the purposes ofprogram formulation. The researchers come fromvarious disciplines and belong to different re-search bodies. The users are selected according tothe type of research program to be formulated.

Priority projects within a program: The mini-mum set of research projects needed for a programto have the intended impact on the developmentof the domain or subsector concerned. Setting pri-orities among projects is done in light of theirrelevance and the resources (particularly humanresources) expected to be available to the researchprogram over the medium term.

Program as structural unit: The word programcan refer either to the actual set of research projectsfor a particular domain or to the organizationalunit that brings together researchers from variousdisciplines to carry out the projects. When severalresearch organizations are involved, the programconstitutes a national research network for the do-main or subsector in question.

The composition of a program makes it a flexibleunit. The participation of individual researchersisn’t indefinite. They cease to be members of theprogram once the intended results of their re-search have been achieved—unless, of course,they initiate other research activities that are alsopart of the program.

Program planning workshop: A working meet-ing in which a group of researchers and users ofresearch results—here called the program planninggroup—formulates or adjusts a research program fora given domain or subsector.

Project research results: Results expected from aresearch project can take the form of new informa-tion, the development of a new technology, or theimprovement of an existing technology.

Technologies may have to do with

■ selecting a variety or clone resistant to a dis-ease or insect pest adapted to specific agroe-cological conditions or having well-definedquality characteristics

■ developing new practices for cropping, seedmultiplication, plant care, harvesting, soilmanagement, storage, or adding valuethrough processing and packaging

■ designing a crop rotation scheme

■ developing a method for crossbreeding ani-mals

■ determining feed levels for livestock

■ designing new agricultural or agro-industrialequipment

■ developing or improving a laboratory instru-ment or experimental methodology.

Research results in the form of information can beagroecological or socioeconomic studies and pro-jections.

Researchable constraint*: A constraint likely tobe removed by agricultural research (includingsocial science research). Adoption of the results al-lows the causes of the constraint to be eliminatedor at least their effects to be reduced.

Research activity*: A component of a research pro-ject. A research activity is a coherent set of specificactions to be carried out in a given period, all ofwhich are necessary to attain the desired result ofthe activity. It is generally monodisciplinary andcarried out by one researcher, sometimes with theassistance of technicians, according to an experi-ment protocol. It has an objective, location, sched-ule, and cost. A research activity can have severallocations if the desired objective applies to severalagroecological zones. The number of activitiesvaries from project to project.

Research objective*: An innovation that scientistshope to make through research in order to re-move a researchable constraint. This can be in anyof several areas:

■ plant and animal materials

■ farming techniques

Glossary

71

■ crop rotation

■ herd management

■ natural resource management

■ product storage, processing, and packaging

■ understanding the physical or socioeconomicenvironment

■ research methods in the agricultural sciences

Research objectives tree: A hierarchy of researchobjectives that must be achieved to eliminate thecentral constraint or problem faced by a particulararea or subsector of agriculture.

The objectives tree is constructed using the con-straints tree. First, an overall objective is assignedto the central constraint. Next, the specific re-search objectives required to achieve the overallobjective are determined. Each specific objectivecorresponds to a research opportunity and servesas the basis for identifying a research project.

The cause and effect relationships set out in theconstraints tree thus become the links betweenthe intermediate objectives and the overall objec-tive. Development objectives do not come intoplay in this tree.

Building an objectives tree is not as mechanical asit may seem. It isn’t simply a matter of replacing anegative phrase describing a constraint by a posi-tive one defining the objective. Rather, the processfirst requires an analysis of the constraint. It’s im-portant to ask, for example, whether the con-straint is researchable. There is no point inassigning research objectives to constraints suchas “poor rainfall distribution” or “uneven ter-rain”. But these factors should be taken into ac-count in the analysis that leads to programformulation. Similarly, results from earlier re-search should be reviewed since a solution or par-tial solution to the constraint may already exist.

Like the constraints tree, the research objectivestree takes the form of a flow chart with variousboxes organized at several levels.

Research program*: A set of research projects car-ried out to address development objectives and us-ers’ needs in a particular domain or subsector.The domain or subsector covered by the programcan be a commodity (such as olives), a group ofcommodities (such as food legumes), an agroe-cological zone (such as Central Benin), or a pro-duction factor (such as farm mechanization).

Research project*: A component of a research pro-gram. The combined results of a program’s con-stituent research projects allow the programobjectives to be met. A project is a coherent set ofresearch activities all of which are necessary tomeet the project’s research objective and whichare to be completed in a given time frame. A re-search project generally produces a technology.

A project is carried out by several researchers andoften provides a framework within which severaldisciplines can interact to achieve the common re-search objective. A research project can involveseveral disciplines of the agricultural sciences;this is called a multidisciplinary project. A re-search project limited to one discipline is calledmonodisciplinary.

In Benin, a research initiative titled “Developmentof Millet Varieties Adapted to Northern Benin” isan example of a multidisciplinary project. Itbrings together agronomists, weed scientists,plant pathologists, and agricultural economists.An example of a monodisciplinary project, in Mo-rocco, is “Improved Extraction of Olive Oil in Ar-tisanal Pressing Operations.” In this researchproject, only technologists are involved.

The number of projects varies from program toprogram. In Morocco, for example, the Date PalmProgram has 17 projects, the Oilseeds Programhas 26, the Dryland Crops Program has 48, andthe Food Legume Program has 52.

A project has several defining characteristics: anobjective, component research activities, a projectleader, a timetable, a budget, expected results,and target indicators for the purposes of monitor-ing and evaluation.

Research-year*: Twelve months of research activi-ties. The time a researcher devotes to “research”over the course of a calendar year is, in the strictsense of the word, less than 12 months. This is be-cause of other duties such as supervising studentsor junior researchers, teaching, and participatingin technology transfer and scientific events. And,of course, there is time off for annual leave. A re-search year therefore covers more than one calen-dar year of a researcher’s time.

In Morocco, for example, INRA estimates that re-searchers devote an average of eight months ofthe calendar year to actual research.

Specific action*: A basic component of a researchactivity. Specific actions, when combined, make itpossible to achieve the intended result of the ac-


72

tivity. The specific actions of a research activitycan be

■ manipulation of plant or animal materi-als—such as hybridization, crossing, castra-tion, grafting, or isotopic marking

■ agricultural interventions such as plowing,sowing, maintenance work, spraying, or har-vesting

■ testing, analyses, observations, measurementsor weighings in the field, on station, in a foodprocessing plant, or in the laboratory

■ censuses, canvassing, surveys, data collection,etc.

In principle, all specific actions of a research activ-ity are carried out by the researcher in charge ofthe activity. However, in practice one or more spe-cific actions may be handled by a technician un-der the researcher’s supervision or by anotherresearcher.

Subprogram: A research program is sometimes di-vided into subprograms. When a program coversseveral products or commodities, there is often asubprogram for each. For example, a cereals pro-gram may have subprograms for wheat and bar-ley. Alternatively, subprograms may correspondto agroecological zones if there is sufficient diver-sity to justify the pursuit of separate researchstrategies.

Support project*: During program formulation,the planning group may identify certain projects tobe carried out in addition to those research projectsthat are supposed to lead directly to new or im-proved technologies. A support project is one thatgenerates information useful to researchers or de-velopment authorities. For example, the informa-tion may improve researchers’ understanding of

the physical or socioeconomic environment, al-lowing for better targeting of technologies. Or itmay increase the probability of success of anotherresearch project by enhancing control of interven-ing factors.

In the case of development authorities, informa-tion from a support project may help them to im-prove the chances of new technologies beingadopted.

Here are examples of support projects identifiedduring the formulation of a food legume pro-gram:

■ study of the socioeconomic impact of mecha-nization on small farms

■ study of food legume supply and demand

■ taxonomy and mapping of diseases and pests

■ study of parasitism mechanisms and host-parasite relationships.

Users of research results: Target groups for thetechnologies developed by research. Direct usersinclude the various types of crop and livestockproducers, agro-industries, traders, and scientists.Indirect users are those whose activities shouldtake research results into account: decision mak-ers, development authorities, extension manag-ers, and so on.

Identifying future users allows for better target-ing of research. The nature of the technologies tobe developed should be determined in light of theneeds of the user groups for whom they are in-tended. For example, in the area of farm mechani-zation, tools for cultivation, crop maintenance,and harvesting need to be adapted to the type offarm where they are to be used.

Glossary

73

Index

Index

AActivity 16, 30, ?Adaptive research 3Administrative criteria 4, 11, 13Adoption 28, 36, 49Agricultural criteria 11Agricultural development 52Agricultural productivity 37Agricultural research agenda 57Agricultural research planning 4Agroecological

conditions 4criteria 11-12zone 69

Aims of research programs 5

BBasic methodology 7-15Benefit dimension 35, 39, 43-44,49-50, 69Bibliographic studies 31Boundaries xi, 3-4, 10-15, 19, 59Budgets 53

CCategories of constraints 23Central constraint 24, 69Chances of research success 36, 45-46, 47, 49Coherence test 47, 50, 69Commodity programs 3, 6, 34- 35Communication 6, 10Complementary development measures 29, 52, 69Complexity 45Constraints 5, 16, 24, 29, 57, 70Constraints analysis xii, 7, 8, 21-25, 28Constraints tree 22-23, 24, 25, 37, 70, 72Correction factor 42Cost xiii, 31, 36, 47, 49Cost-benefit analysis xiii, 34Credibility 52, 59Crop productivity 37

DDecentralization 3-4, 58

Development agencies 15Development authorities 49Development measures 52Development objectives 19-21, 70Development strategies 19-20Disciplinary program 6Domain review 28Duration 30, 70Duplication 28

EEconomic skills 35Economic value 37Ecoregional 3Ecosystem 4Estimation assumptions 38Estimation principles 38Estimation procedures 57Evaluation xii, 8, 26-27Existing research results 26Exogenous constraints 23Experience-based assessment 59Exploitation of natural resources 5External validation 52Externalities 3Extension services 15Extension system 42

FFarmers 6, 15, 49Feasibility 41-43, 52, 59Final project scores 47First workshop 7, 9, 27, 31Focused projects 31Forestry production 37

GGeographic focus 9Geographic information systems xii, 13-15, 59Government 3

local 19regional 15, 59

Index

77

HHolistic 3, 5Homogeneity 4Homogenous regions 14Human resources 6, 25, 30-31, 45, 57

gap xiii, 7, 8, 50-51, 52requirements xiii, 31, 50-51

IImpact 39-40, 42, 44, 46Implementation xiv, 7, 8, 47, 51-53Importance of the regional program 43INFORM 26Infrastructure 46Integrated pest management 28Integrated projects 31-32Integration 5Institutional or interinstitutional program 70Institutional aspects 5

LLand-use criteria 12Language 15Livestock productivity 37Location 30Location of a research activity 70

MManagement information systems 31, 53Methodological framework 35Monitoring and evaluation 53Multiple criteria 34

NNational agricultural research plan xi, 9-10Natural resource

competition 5endowment 12management 3, 4-5, 27, 37, 59valuation 34

Nongovernmental organizations 15Nonresearchable constraints 23-24, 52, 59, 70

OObjectives 28-29On-going research 52Off-site impact 40, 42On-farm research 6On-site benefits 41, 44

On-site resource quality 39Opportunities 57Organization 6, 9, 25, 50Overlays 13

PParameters 50, 52Participants xii, 7-8, 10, 15, 24, 44, 47Participatory technology development 3, 6Past research results 9, 26-27Pasture productivity 37Personnel costs 47Planning by objective 70Planning committee xii, 9, 44-45, 71Planning group 71Planning workshop 52Potential 57Potential impact 37, 48Potential rate of adoption 46Preparation 9-16Pre-workshop stage 49Priority projects 71Priority ranking 47-48Priority setting xiii, 7, 8, 34-50

exercise 49framework 59methods 34procedures 34workshop 45, 48

Problems 57Problem orientation 6Production factor 5Productivity concerns ixProfitability 45Program 16Program as structural unit 71Program committee 9, 33, 49Program document 52, 54Program leader 24, 33, 49, 52-53Program planning exercise 9Program planning workshop 71Program research results 71Program structure 10-11Project 34, 53

classification 31-33identification 7, 30-33parameters 45ranking 36, 50removal 31, 32-33score 35-36, 49, 50screening 31titles 28transfer 31

Proximity to users 6


78

QQuality approach 38

RRanking 48Rapid rural appraisals 58Rate of adoption 45, 47Recruitment 50Region xi, 4, 10Regional analysis 19-21, 58Regional context 6Regional development objective xii, 19Regional planning approach xi, 57Regional potential 25Regional problems 6Regional research programs ix, xi, 3, 4, 5, 34Regional review xii, 7, 19, 22Research activity 71Research domain 19Research feasibility 42Research methodologies 53Research objectives xii, 8, 28-30, 71Research objectives tree 72Research production functions 34Research program 72Research projects xiii, 8, 16, 30-31, 44, 46, 72Research themes 10, 12Research-year 72Researchable constraints 23, 29, 38, 42, 58, 71Researchers 15, 57Resource requirements 53Resources 52Review 8, 26Risk 45

SScreening 31, 33Searches 26

external 26computerized 26internal 26

Second workshop 7, 9, 43, 46, 48-49Share of the regional program 43Size of benefit dimensions 38, 44Socioeconomic criteria 13Soil quality 37Specialization 50-51Specific action 72

Stakeholders 28, 52Stakeholder group 57Stakeholder participation xiv, 35Strategic considerations 43Strategic research 3,6Strategy 28-29Sub-programs 4, 73Supply approach 38Supply impact 38Support projects 31, 50, 73Systems approaches 3

TTechnical progress 39Technical projects 31Time lags 35, 39Time requirements 30, 57Timing 59Third workshop 7, 9, 52-53Training 52, 53Transfer 32-33

UUsers 6, 28Users of research results 73

VValidation exercise 49Validation workshop 52

WWater management 37

Index

79

Produced by ISNAR Publication Services

Text editor: A. Michelle Luijben and Gerry ToomeyLayout: A. Michelle Luijben and Fionnuala HawesCover design: Ron AugustinPrinter: Rapporten Service Drukkerij B.V., Rijswijk, The Netherlands

Laan van Nieuw Oost Indië 133, 2593 BM La Haye, Pays-Bas

Boîte Postale 93375, 2509 AJ La Haye, Pays-Bas

Téléphone : (31) (70) 349-6100, Télécopie : (31) (70) 381-9677, Télex : 33746 Email : Internet [email protected]

Service International pour la Recherche Agricole Nationale

Siège

Correspondance

Communications

ISBN 92-9118-017-3ISSN 1022-9973

About these Guidelines

These guidelines present a method for formulating regional research programs with amixed perspective - one that combines productivity and natural resource managementissues within a regional, decentralized context. Based on earlier work by Collion andKissi (1994), it has been substantially modified to account for the peculiarities of re-gional research planning. The method was pilot tested in Benin. It has also been ap-

plied in Senegal and Morocco.

Part 1 presents the rationale and background of regional research planning. This sec-tion is particularly relevant for institute directors, policymakers reviewing research op-

tions, and, of course, for research program leaders. The method for regional researchplanning is treated in Part 2. In the first five steps of the planning approach, research

constraints, objectives, and projects are identified in a structured, logical, and participa-tory fashion. Step 6 of the research planning approach examines how priorities are es-

tablished among the research projects. This presentation is aimed specifically ateconomists involved in research planning and research program leaders. In steps 7 and

8 of the program planning process, the results of the planning exercise are translatedinto implementable recommendations. Part 2 is focused specifically for program lead-

ers and research institute directors. Part 3 reviews the overall procedures and posessome questions and issues that should be resolved in the future. It may be of interest to

all persons involved in regional research planning.

Throughout the document, a series of boxes illustrates the experience of Benin with re-gional research planning. These boxes aim to provide a clear example of the sequence,

organization, and execution of the regional research planning process.