STRENGTHENING POVERTY REDUCTION PROGRAMMES USING … · social and political context of their...

Agricultural Research& Extension Network

Network Paper No.134

January 2004

ISBN 0 85003 704 2

The Agricultural Research and Extension Network is sponsored by the UK Department for International Development (DFID)The opinions expressed in this paper do not necessarily reflect those of DFID.

We are happy for this material to be reproduced on a not-for-profit basis. The Network Coordinator would appreciate receiving details ofany use of this material in training, research or programme design, implementation or evaluation.

Network Coordinator: Robert Tripp Administrative Editor: Alana Coyle

AbstractThis paper explores the use of actor-oriented approaches in natural resource-based development. It begins byreviewing the need to bring an analysis of actor linkages, coalitions and information flows higher on the agendain planning, implementation, monitoring and evaluation. Various tools which could assist in doing this areintroduced and their use is illustrated in case studies of natural resource-based research and development (R&D)projects in Nepal and Bangladesh.

Research findings• Use of actor-oriented tools can change perceptions of development actors, encouraging them to engage with the

social and political context of their activities in a productive way.• Actor-oriented tools provide practical ways to monitor, document, and assess and thus legitimise crucial

institutional strengthening activities.

Policy implications• Actor linkage analysis and coalition building for effective and sustainable development should be legitimised

and rewarded.• Development interventions should include actor-oriented tools in development planning, implementation,

monitoring and evaluation.• Development agencies should employ and integrate professional staff with actor-oriented social science skills

(e.g. applied anthropologists, evaluation specialists, applied ethnographers) into their mainstream activities.

STRENGTHENING POVERTY REDUCTION PROGRAMMES USINGAN ACTOR-ORIENTED APPROACH: EXAMPLES FROM

NATURAL RESOURCES INNOVATION SYSTEMS

Stephen Biggs and Harriet Matsaert

Contact Details

Stephen Biggs can be contacted at the International Centre for Integrated Mountain Development (ICIMOD), GPO Box 3226, Kathmandu,NEPAL. Email: [email protected]

Harriet Matsaert can be contacted at 188 Gulshan Avenue, Dhaka, BANGLADESH.Email: [email protected]

Acknowledgements

We would like to thank Don Messerschmidt, Scott Justice, Devendra Gauchan and Rob Tripp for helpful comments on an earlier draft ofthis paper.

iii

Strengthening poverty reduction programmes …

CONTENTS

Page

Abstract i

Contact details i

Acknowledgements i

Acronyms iv

1 INTRODUCTION 1An actor-oriented approachSome actor-oriented toolsThe actor linkage mapActor linkage matrix (ALM)Actor determinants diagramActor time linesActor learning and response analysis

2 CASE STUDIES 4Pre-project activity in a crop post-harvest coalition project in BangladeshRestructuring the national agricultural research system in NepalChanging power tiller innovation systems in Nepal

3 DISCUSSION AND REFLECTIONS 11Political issues in using actor-oriented toolsImportance of events and key locationsImportance of individual actorsActor linkage maps versus actor linkage matricesA role for quantification?Cultural dimensions of transaction costsKeep it simple

4 CONCLUSIONS 14

REFERENCES 14

ENDNOTES 16

Figures and tablesFigure 1 Illustrative actor linkage map 2Figure 2 Map showing key actors in a Bangladesh chilli innovation system 5Figure 3 Example of an actor derminants diagram 6Figure 4 Suggested framework for NARC’s PTD and multiple linkages programme 8Figure 5 Example of an actor linkage map for NARC 9Figure 6 Time line of major phases in the spread of power tillers in Nepal 10Table 1 Illustration of an actor linkage matrix 2Table 2 Actor linkage matrix used to monitor partnership building 7

Agricultural Research and Extension Network Paper No.134

iv

AcronymsALM Actor Linkage MatrixAREP Agricultural Research and Extension ProjectCPHP Crop Post-Harvest Research ProgrammeDFID Department for International Development (UK)ISNAR International Service for National Agricultural ResearchMOU Memorandum of UnderstandingNR Natural ResourcesNARC Nepal Agricultural Research CouncilNGO Non-governmental OrganisationOR OutreachPTD Participatory Technology DevelopmentRCT Resource Conservation TechnologyR&D Research and DevelopmentS&T Science and Technology

1 INTRODUCTIONThis paper focuses on development interventions innatural resources-based innovation systems. We presenta number of tools which we have found useful inallowing us to focus more closely on the actor linkagesfound in innovation systems. The paper suggests thatthese social science and more qualitative tools shouldbe seen as parallel and complementary to the analyticalapproaches of natural resources research (experiments,surveys, etc.) and the tools of quantitative economists(rates of return studies, resource allocation prioritysetting exercises).1 We start by introducing these toolsand go on to illustrate their use through a number ofcase studies from recent work we have been involvedin. Finally we reflect on our experiences and makesuggestions for others who are interested in developingactor-oriented tools to suit the context of their own work.

By innovation system we mean the system of allmajor social actors affecting the revealing,acknowledgment, generation and diffusion of technicaland institutional knowledge over time (see Hall et al.,2001; Nelson and Winter, 1977; Freeman, 1988; Ekboir,2002; Clark et al., 2003). We are working from thepremise that a strong, effective and sustainableinnovation system is one where institutions2 facilitateflows of information and good partnership coalitionsbetween key actors over time. Powerful support forthis view can be found in Douthwaite’s recent analysisof a selection of innovation systems (ranging from cropvarietal developments to computer softwareinnovations). One of his findings was that successfuland sustainable innovations are invariably thosedeveloped in a system that can be characterised as a‘bazaar approach’. This is where users andmanufacturers of technologies are always interactivewith ‘researchers’ and fully involved as equal partners,especially in adaptive research (Douthwaite, 2002).3

While most of us acknowledge the importance oflinkages between actors, coalitions, alliances and flowsof information to successful innovation and to thedevelopment of sustainable innovation systems, theseaspects are often not addressed systematically andexplicitly in the management of natural resource (NR)-based development activities. All too often this resultsin the development of technologies which sit inresearch stations, replication of effort, waste ofresources, unproductive rivalry between differentactors, etc.

The need to address actor linkages and coalitions isbecoming increasingly important for NR developmentactors today. Research funders and governments areactively encouraging new, pluralistic models of researchand development (R&D) and extension which bringtogether actors in the private, public and civil societysectors and reduce transaction costs. (Byerlee, 1998;Kidd, 2002). Alongside this there is no lack ofdocumented evidence of the difficulties and problemsencountered by those who try to go forward in formingnew partnerships (International Service for NationalAgricultural Research (ISNAR), 2001).

Despite the need to look more closely at theseaspects of NR development activities, there is a dearthof practical and user-friendly techniques available toproject managers, which address these institutionaldimensions of innovation systems. Mainstreamplanning, implementation, monitoring and evaluationtools such as the log framework tend to emphasiseactivities and products, which do not relate to theseactor linkage and process issues.4

An actor-oriented approachThis approach is concerned principally with mappingrelationships and flows of information to provide abasis for reflection and action. These ideas and toolsare not new. Their parents are many and includeanthropological and social network research techniques(see Long and Long, 1992; Long and Van der Ploeg,1989; Lewis, 1998; Davies, 2002); stakeholder analysis(Ramirez, 1999; Grimble and Wellard, 1997; ODA,1995); economic input and output models (Falcon,1967); agricultural information knowledge systems(Roling and Jiggens, 1998; and Berdegue and Escobar,2002); processes monitoring and documentation (Mosseet al., 1998); graphic theoretical techniques (Temel etal., 2003); communications systems (Mundy, 2003); andthe analysis of the behaviour of disciplines inagricultural sciences (Raina, 2002).5 However thesystematic application of these techniques bydevelopment actors within NR innovation systems isstill not common.6

In our recent work we have been using anddeveloping these tools in a range of NR contexts. Wehave found them to be a very useful complement toother planning tools such as formal surveys andexperiments, log frames, conventional monitoring and

STRENGTHENING POVERTY REDUCTION PROGRAMMES USING ANACTOR-ORIENTED APPROACH: EXAMPLES FROM NATURAL RESOURCES

INNOVATION SYSTEMS

Agricultural Research and Extension Network Paper No. 134

2

evaluation mechanisms, and more quantitative typesof research priority setting studies.

Some actor-oriented toolsThe first stage in all these exercises is to identify thekey actors who bring about or prevent change in aninnovation system, i.e. identification of the actors whoare the actual drivers or preventers of change. Thebreadth of the analysis can vary. One can look at anational system, a particular region, or at a particulargroup of actors, e.g. farmers. One can disaggregatemore or less depending on the breadth of the study. Anational analysis might put researchers in one box. Ina separate analysis one might want to set up an actormap or matrix just to look at the interactions betweendifferent types of researchers in the public and privatesectors. On other occasions one might want to separateactors into those who are in the public, civil and privatesectors. Increasingly actor analysis is being used toanalyse the role of aid donors, international researchorganisations, international non-governmentalorganisations (NGOs), etc. in the same framework aslooking at actors at the village and national levels. Theframework can be used in an analysis of genderrelationships. It should be pointed out that the emphasisis on identifying specific social groups or actors in aspecific location at a given point in time. Consequentlythe actor approach differs from some economicframeworks where ‘sectors’ of the economy are definedby what is produced: the agricultural sector, themanufacturing sector, etc. In actor analysis it is thepeople who make decisions which defines the groups.One would not then have a group called ‘economicforces’, or a category called ‘research’. Research doesnot just happen; it is people who do research, so thecategory would be ‘researchers’.

The actor linkage mapThe actor linkage map is a useful starting point fordiscussing relationships and flows of information inan innovation system. Key actors are shown on a mapwith arrows between them indicating flows ofinformation. In actor linkage analysis there is always

an arrow going in each direction. Single two-headedarrows are never used, as one of the main points ofthe mapping is to examine power relationships in thecontrol of flows of information in different directions.The intensity of these flows can be illustrated by thewidth of the arrows. In Figure 1 the thick arrow goingfrom farmers to researchers illustrates a strong flow ofinformation. The fairly weak flow of information fromresearchers to farmers is indicated by a thinner arrow.It should be noted that these maps need to representactual flows of information, etc. rather than officialorganisational charts. The map gives rise to discussionsof formal and informal mechanisms used to transmitand control information. It also highlights the issue ofwhich actors and linkages are going to be in theanalysis. In the past many actor linkage maps used inagricultural research and extension discussions haverestricted themselves to public sector actors (e.g.government research institutions, governmentextension organisations and ‘beneficiaries’ (e.g. ‘passive’farmers). In addition few maps included such actorsas ‘funders of research’ or an analysis of how thesefunding actors interacted with other actors, oftendetermining research agendas and research processes.7

The actor linkage maps are particularly useful whenfocusing on one actor and his or her linkages withother groups.

As the number of actors increases, however, the mapcan become too complex. At this point it may be usefulto work with maps of part of the system or move to anactor linkage matrix.

Actor linkage matrix (ALM)The matrix is similar to a map in that it identifies allthe actors and shows the links between major actorsin an innovation system. In the matrix this is representedby listing actors along the vertical and horizontal axes.The cells in the matrix represent flows of informationfrom the actors in the rows to actors in the columns.8

(See Table 1 and example in the Bangladesh casestudy).

In Table 1 cell 1B refers to information flows fromresearchers to farmers. In Figure 1 this was arrow 1.Cell 3C refers to information flows betweenmanufacturers and other manufacturers, illustrated inthe actor linkage map by arrow number 2. In the matrixall cells can be identified by their coordinates (numbersfor rows and letters for columns).Figure 1 Illustrative actor linkage map*

*See also example in Bangladesh case study

farmers manufacturers

researchers

2

1Table 1 Illustration of an actor linkage matrix (ALM)

A B CActors Researchers Farmers Manufacturers

1 Researchers 1

2 Farmers

3 Manufacturers 2


3

This tool does not lend itself as easily as the mapto group work. However it has a number of advantages:• It can deal with more complex situations and more

actors (maps can get very complex and web-like, asmore and more arrows are added).

• It has a cell for every possible linkage, and soencourages one to explore all possibilities, to thinkcreatively and innovate! It helps to keep a ‘holistic’perspective on which key actors really do determinewhat happens in a specific innovation system. Thisdoes not mean that all actors and linkages have tobe looked at all the time. Quite the contrary, as itforces a realisation that only certain linkages can beanalysed and worked on at any one time.

• It is a useful tool in helping to pinpoint particularlysignificant links, e.g. strong links, coalition groups,weak links or opportunities. This makes it moreuseful than the map for planning, implementation,monitoring and evaluating change.

• It enables users to quantify the strength of linkagesusing symbols in each cell. e.g. plusses and minuses,or letters such as s (strong), m (medium), w (weak),dn (don’t know).

• It enables users to condense and store a lot ofinformation about linkages in the spreadsheet ALM(each cell reference can be linked to a text).Consequently it is a useful tool for documenting agiven situation or the outcomes of an event.The actor linkage matrix is best used with a small

group, with people familiar with the technique, or aftera discussion to summarise findings that are thencirculated. For those familiar with the technique, as isthe case of an ongoing research project in Bangladeshthe team uses for group discussions a ‘matrix board’,which hangs on the wall as an alternative to a printoutfrom a computer spread sheet. Here linkages arerepresented by tokens placed on small hooks.

Actor determinants diagramThis tool, an example of which can be seen in theBangladesh case study, is similar to the ParticipatoryRural Appraisal (PRA) problem tree. It is intended as agroup discussion (or individual thinking) tool to analysethe nature of a particular linkage.

The starting point is a cell of the actor linkage matrixor a linkage on the map. Normally this would be onethat is particularly significant (and might need to bestrengthened, weakened or learnt from). The diagrammaps weakening and strengthening forces on thelinkage and helps a group to identify possible areas ofintervention.

This tool can be used in a brainstorming exerciseand obviously some ‘areas for intervention’ will be morepossible to implement than others. However, this isone of the important reasons for using the tool. It helpsopen up a discussion about the feasibility of differentactions within the current social and political context.It is a useful tool for building an action plan from theanalysis of a particular situation. For this reason it is

most usefully carried out with the key actors who wouldbe involved in any future ‘implementation’ of suggestedactions.

Actor time linesCoalitions, relationships and narratives of changeprocesses change over time. Getting a group of keyactors to construct an actor time line of key past eventsfor a particular innovation system can build a morecomprehensive understanding of past change processesand a better understanding of the current situation (seeexample in Nepal power tiller case study).

An actor time line is a listing of key events in theevolution of an innovation system. The events are ‘actor’events. Which actor made key important decisions atwhat time in the past? As in other parts of actor analysisthe emphasis is on human actions. For example, theplanning commission abolished restrictions on theimports of two-wheeled tractors. This is different fromsaying import restrictions were abolished, or structuraladjustment policies were implemented. Whereverpossible one has to be as specific as possible regardingwho took what decisions, when and where. This helpsto take discussions out of the realm of generalities intothe specifics of understanding the actual causationalprocesses in a particular innovation system.

It also raises awareness in the group of the differentperceptions amongst people of what caused things tohappen in the past. It is sometimes difficult to getpeople who have strong views about past events(especially as regards what caused what to happen) tosee those events in a different way. Even whensomeone has been ‘convinced’ that there are differentand legitimate alternative narratives about past events,one can still find ‘old’ views jumping out unexpectedlyand completely undermining an agreed way forwardfor a coalition. The group’s construction of actor timelines is designed to help address this problem. Whenprojects and development activities have become ‘pathdependent’ it is sometimes because old uncontestednarratives about past events have been used to maintaina ‘business as usual’ control over decision-making.Helping people to drop old ways of thinking and seethings in new ways is one of the major challenges thatthe actor approach takes on.

Again we recognise that this is not a particularly newidea. We can all think of occasions when we have seentime lines in a publication or a list of key events in thehistory of (or a plan for) a project or a programme.9

However, the way we suggest actor time lines are usedhere is more as a learning and reflection tool, a way toestablish new common ground in a coalition ofpartners, and as a tool to guide future action. The timeline can either be given as a list of events, with datesalongside, or as a figure with a sequenced bar chart ofactor events over time. The figure helps to reinforcethe notion of time, sequencing and the path of causationof past events. See Figure 6 in the Nepal power tillercase study for an example of a time line.


4

Actor learning and response analysisThe last set of tools concerns learning and responseanalysis on the part of coalition partners. We do nothave a specific tool as such. What we have found,though, in the projects we have been working on isthat explicit attention needs to be given to ways inwhich partners can systematically collect informationfrom different sources, analyse it and draw up localaction plans as they go along. The existence of papersdocumenting this analysis and the planned/actualoutcomes can be used to monitor the innovativebehaviour of partners in the coalition. In principle inall projects information can come from three mainsources:• First, from planned activities, which may be planned

experiments, development interventions, surveys ormeetings. Often in conventional projects theinformation from surveys, experiments and meetingsis not acted on locally. This is especially the casewhen academic publications, and ‘projectrequirements’ are the primary reasons for theplanned data collection activities.

• The second source of information is from‘unexpected sources’ and is revealed in the processof collecting planned information or conductingother planned activities. This kind of information isalways coming up in projects. For example, inconducting a survey it is found that there is anotherproject in the same region doing similar work.

• The third source of information is from ‘unexpectedchanges’ in the context of the project.We have found that explicit attention to the ways

information from these three sources is analysed andused to draw up short-term action plans has become amajor component in the actor-oriented approach. Inthe Bangladesh case study the actor linkage matrix isused to formulate quarterly plans of action to addressinstitutional linkage capacity-building issues. In theNepal Agricultural Research Council case study the six-monthly agreed plan of action against ‘mid-term review’indicators served a similar purpose. In the power tillercase study, the six-monthly Learning and Responsetables provided a similar framework. What is significantis that in all cases it was the partners themselves whojointly agreed what they would do over the succeedingmonths. The onus was on self-learning and appropriateactions on the part of the group itself, rather thanmaking ‘recommendations’ to other actors on what theyshould do.

2 CASE STUDIESIn this section we describe various contexts in whichwe have used and are using actor-oriented tools.10

Pre-project activity in a crop post-harvest coalition project in BangladeshThe islands formed by siltation in the river deltas ofBangladesh are known as chars, and the people wholive on them are amongst the most vulnerable in

Bangladesh. In Jamalpur, the district in which thisresearch is focused, 80% of char-based householdsare estimated to be in the ‘extremely poor’ category.Because the islands are often temporary islands in theshifting river bed, they have little infrastructure (i.e.they fall in the category which is 40% below the nationalincome poverty line). Most families are forced byerosion to move residence several times in their lives,as well as frequently migrating to the mainland in timesof flood.

Char-based livelihoods are very insecure, but thesilt deposited annually during the flood period createshighly fertile pockets of land. When the floods recede,char dwellers farm their land intensively and marketthe produce to the mainland. The Jamalpur chars areparticularly well known for their high-quality chilli,vegetable production and livestock. Land is morereadily available and the poorest households haveaccess to more land for cultivation and grazing on thechars than their counterparts on the mainland.

While many char dwellers believe natural resource-based production potential is their key relativeadvantage over mainlanders, they are relativelydisadvantaged in their access to sources of informationand markets. Because they are often impermanent,chars tend to have little infrastructure (roads, electricityand government offices are rare); transport can presenta problem (dangerous boat crossings in the rainyseason, long walks through sand in the dry season);and most development actors are reluctant to visit thearea. So while potential for developing char-basedproduction exists, weak linkages with key externalactors (extension, research, NGOs, private sector, etc.)prevent char dwellers from participating in innovationsystems which would allow them to develop newtechnologies and market opportunities.

The crop post-harvest research programme (CPHP)funded by the UK Department for InternationalDevelopment (DFID) has recently been focusing onstrengthening sustainable innovation systems and onthe importance of partnerships and coalitions in thiswork (Hall et al., 2001; Biggs and Underwood, 2001).In the Jamalpur chars, the CPHP has funded a researchproject to examine and strengthen char-basedinnovation systems for two key enterprises: chilli andlivestock. The research asks: What is the status of thechar-based innovation systems? What linkages are madewith other key actors at national and internationallevels? And what opportunities exist to strengthen char-based innovation systems through building linkagesand coalitions?

The study is being carried out by a research coalitioncomprising a local NGO (Development Wheel), anational business advisory services centre (BASC) whohave an interest in building up a farmers’ businessassociation, an anthropologist with experience ofknowledge systems in the chars, and an expatriateanthropologist/agricultural engineer with experienceof developing and using actor-oriented tools.


5

This core research team is using actor maps andmatrices with other key actors in chilli and livestockinnovation systems to map out the current reality,identifying strengths, opportunities and weak linkages.Through working together with other key actors theproject team expects to build coalitions to enhancethe focus and sustained capacity of local innovationsystems. The project team does not see itself outsideof the process, and uses actor maps and matrices tomonitor its own relationships and successes (andfailures) in building partnerships with other key actorsthroughout the life of the project.

In Figure 2 arrows refer to flows of goods andknowledge. The map shows that the strongest linksbetween the char dwellers and the mainland actorsare through the private sector. Key information fromprivate sector companies and government extensionservices is passed on to char dwellers by local inputdealers. Local middlemen play the key role in providingmarket access. However national-level processors andretailers are making efforts to develop direct links with‘contract farmers’ (shown by dashed line).

The research team has found actor linkage mapseasier to use in meetings with potential coalitionpartners than the matrix which is initially too complexfor people to grasp. However, for our internal teamworkand compiling the information we are collecting oninnovation systems we have found the actor linkagematrix very useful. Team members have observed thatthe ALM makes things visible and helps them to beaware of gaps in their knowledge and identify linkagesthey have not considered.

The team has experimented with different types ofmatrix, beginning with a simple quantification, then

developing a more complex system of quantifyinglinkages and moving towards a more useful ‘qualitative’approach where key linkages are highlighted anddescribed in the attached text. Current uses of the matrixinclude:• Monitoring the team’s progress in building

relationships with other key actors. A matrix is drawnup quarterly, highlighting useful linkages made, andpinpointing linkages we want to develop further inthe next quarter.

• Illustrating the expected impact (on buildinglinkages) of a forthcoming workshop.

• Documenting changes in significant linkages andcoalitions observed in the innovation system throughcase study monitoring.The research is ongoing but to date the use of actor-

oriented tools has resulted in a number of importantoutcomes. Using the tools with research and extensionstaff has helped discussion reach beyond the formalstructures – the organograms of formal relationshipsand the way things ‘should happen’ – to the reality ofwhat is actually going on. For example while officiallythe Bangladesh Agricultural Development Board bulkschilli seed and provides it to farmers, a time linerevealed that the uncomfortable reality was thatalthough the ‘release and distribution’ of seeds by theBoard was planned to have happened during the lastten years, it still has not taken place. In the meantimethe regular and effective introduction of new seedvarieties is being carried out by a national seedcompany (which currently has no links with the chilliresearch institute). Using the actor tools has encouragedthe chilli researchers to confront the reality of thesituation and consider its implications. Should they,and how can they form linkages with this dynamic privatesector actor (see Determinants diagram, Figure 3)?

By using the tools to discuss the current status ofinnovation systems, we can see how they help to revealand legitimise previously unacknowledged but vitalactivities by individuals. For example, in the above-mentioned discussion at the chilli research institute,the director told us that they had no links with theprivate sector, farmers or NGOs. However one of hisjunior scientists reminded him that he had recently,on his own initiative, begun to work with farmers onthe chars and had already formed links with a localNGO there, inviting both members of the organisationand farmers to visit the research station. When thisactivity was marked on the linkage map it emphasisedhow important the previously unacknowledged workof this scientist was in bridging the divide between theresearch institute and other actors.

There has been considerable interest in actor-oriented tools from research and extension staff, towhich the team has responded by providing briefingson the tools and ‘learning by action’. They expect inthis way to provide research and extension managerswith the means, once they have recognised the needto focus more on building linkages, partnerships and

NGOs

Femalechardwellers

Middlemen

Chilliprocessorsand retailers

Mainlandfarmers

Researchteam

Male chardwellers

Inputdealers

Extensionstaff

Private sectorinput supplyreps.

Figure 2 Map showing key actors in a Bangladeshchilli innovation system


6

coalitions, to draw up action plans and also monitortheir own progress in this area.

The matrix has helped the research team to analyseits own relationships with project partners, for instancewhen monitoring their activities in the first quarter. Amatrix was used in this process to show the linkagesmade with key actors, highlighting the fact that mostlinkages had been between the project team and asingle actor. However, to make a sustainable impactlinkages had to be expanded to include other key actorsin the system. The fact that the team was one of theactors being analysed helped to achieve this in apractical way. The matrix in Table 2, shows how ameeting was planned which would build relationshipsnot only between the project and key actors but alsobetween other actors who are rarely engaged.

It should be said that using these tools is not withouthazard. On one occasion the research team was highlycriticised by agricultural extension staff when theypresented an actor linkage matrix indicating that theextension field staff never visited the research area.Since the extension service takes great pride in workingin every area of the country, contradicting this essentialpart of their identity was unwise and unproductive.The team later became aware that the farmers’ groupthey were working with was keen to represent its

members as isolated and without any services fromthe government, as a means of increasing their chancesof receiving inputs from the researchers. The realitylies somewhere between the two. Since the difficultmeeting with the extension team, researchers havenoticed that the extension ‘block supervisor’ now makesregular visits to the focus chars. To some extent thenin this case the project has helped to bring about achange in the culture of the local extension staff.Regular actor-oriented monitoring would acknowledgeand reward the supervisor for strengthening theselinkages, and also note the mechanisms being usedwhich are within the current budgets and rewardsystems of the public sector extension service.

David Lewis (1998) and Brigitta Bode (2002) haverecorded similar experiences in Bangladesh of revealinginformation that contradicts the image projected by agovernment service. Lewis describes how hisinvolvement as an ‘outside’ process monitor in aresearch project came to a premature end when theorganisations he was partnering began to find theinformation he was uncovering about themuncomfortable. In the actor approach we are suggestinghere there is no ‘outsider’ process monitoring. All actorsin the coalition (including the research team) are onthe ‘inside’.11

Create formal linkthrough signingMOU

Other private-sectorcompanies havesigned MOUs withthe research institutefor joint projects

Seed companiesdon’t see whatresearchers canoffer

Seed companieswork near the stationand have their ownresearch plot

Make informal locallinks through on-farmwork etc.

researchers/SeedCompanies link.(currently non-existent)

Formal seed distributionprocedure in govt./ parastatalcontrol

Strengthening factors

Possible interventions

Invite seed co. staffto visit researchstation and to keymeetings

Possible interventions

Weakening factors

Figure 3 Example of an actor determinants diagram


7

Restructuring the national agriculturalresearch system in NepalThe AREP (Agricultural Research and Extension Projectin Nepal) was a World Bank-funded project first mootedin the early 1990s. In the mid-1990s it was designedalong fairly conventional lines with a strong emphasison MSc and PhD training and restructuring work. Oneof the major organisations in the project was the NepalAgricultural Research Council (NARC). Two of theproject’s key restructuring goals were: (i) theencouragement of more participatory technologydevelopment (PTD) in NARC, and (ii) the promotionof linkages and partnerships between NARC and awhole range of government and non-governmentpartners.

Actor linkage maps were used extensively at thenational and regional levels and with some commodityprogrammes to address these issues. The maps helpedto examine and understand existing relationships/partnerships and focus attention on linkages thatneeded strengthening.

An early project document had seen the lack of PTDactivities in NARC as a major problem. However, asNARC staff actively searched for examples of linkagesand working relationships, they found many examplesthat had not been reported through the usual researchmonitoring and documentation processes. One of thereasons for this was the official perception of whatconstituted PTD in NARC, which was that ‘on-farm’and PTD research only took place in their Outreach(OR) division. This division was involved in a range ofstandard activities arising out of the conventionalfarming systems transfer of technology approach(benchmark surveys, village meetings, final stages ofvarietal screening, the management of a number of‘representative’ outreach sites, formal impact/adoptionstudies, evaluations, etc.).

However, on closer analysis of who was actuallydoing PTD in NARC, it was revealed that a far greaternumber of scientists than those in the OR division wereinvolved in a large range of innovative PTD activities.This information had not been reported in the past

1. Male char dwellers

2. Female char dwellers

3. Local leaders

6. Local middlemen

8. Dealers

9. Local government staff

11. Bank staff

12. Local NGO staff

13. National government staff

14. Researchers

16. National middlemen

18. Chilli processors and retailers

19. Input suppliers

20. Media

21. Project team

Mal

e ch

ar d

wel

lers

Fem

ale

char

dw

elle

rs

Loca

l lea

ders

Loca

l mid

dlem

en

Dea

lers

Loca

l gov

ernm

ent s

taff

Ban

k st

aff

Loca

l NG

O s

taff

Nat

iona

l gov

ernm

ent s

taff

Res

earc

hers

Nat

iona

l mid

dlem

en

Chi

lli p

roce

ssor

s an

d re

taile

rs

Inpu

t sup

plie

rs

Med

ia

Proj

ect t

eam

A B C F H I K L M N P Q R S UCHILLI/SOJ

Table 2 Actor linkage matrix used to monitor partnership building

linkages made by the team this quarter

linkages to be developed through activities planned next quarter

d d a a a a 11

d d a a a a 11

c

13

14

a a a 3

a a a 3

a a 2

4

a a a a a a a 1

5

10

8

6

11a 11a c 13 14 3a 3a 2ab 4a 1ab 5 10 8 6


8

because much of this type of PTD came under specialprojects, often conducted with a large range ofinternational and government partners and with anumber of local R&D NGOs. For years these specialprojects had made up a very substantial part of NARC’swork.12 The inclusion of these data gave a very differentperception of the PTD situation in NARC. The actorlinkage analysis helped ‘reveal’ a whole range ofactivities and linkages that were generally notacknowledged in official documents, or seen by‘outsiders’ who did not know of this work. In someways many of the PTD prescriptions for the projectwere already in place, but were not ‘seen’ or‘acknowledged’. From a new start based on what wasactually happening in NARC discussions could takeplace on how to encourage and facilitate new PTDactivities. This focused on how future PTD activitiescould be managed primarily by other actors in theoverall national agricultural research system. Figure 4is one of the actor linkage maps used to direct attentionto these issues.

A second goal of the AREP project was to promotelinkages with a whole range of non-governmentpartners, such as local and international NGOs,international agricultural research centres, the privatesector, etc. Again it was found that there already existeda great number of partnerships/linkages that were notrevealed or acknowledged. However, in this case theuse of actor linkage maps, where NARC was placed inthe centre of the page and circles around NARC

represented other existing and potential partners,helped focus attention on the need for new types ofmechanisms for linkages with different categories ofpartners. One of these actors’ maps for NARC is givenas Figure 5. To go forward with the institutional reformprogramme NARC organised workshops at the nationaland regional levels to address these issues. One set ofworkshops looked at NARC/NGO linkages and anotherset discussed NARC/private sector linkages (Gauchanand Joshi, 2000).

Actor linkage maps were used in a similar way toencourage regional stations to think about how tochange their role from being conventional public sectorresearch providers, to being promoters and facilitatorsof a strong regional agricultural and natural resourcesinnovation system. Regional technical working groupswere established to foster partnerships between a widerange of private, government and NGO actors. Regionalstation chiefs found that keeping updated inventoriesof all R&D and development actors in each region,and promoting/facilitating regional networks of R&Dactors a very different type of work from being in chargeof a conventional public sector regional researchstation.

The actor linkage maps helped to introduce a changein institutional behaviour on the part of NARC towardsold and new partners.

A further way actor maps were used in therestructuring of NARC was to help raise awarenessabout the diversity of actors and linkages in different

Figure 4 Suggested framework for NARC PTDand multiple linkages programme

Source: Gauchan, Joshi and Biggs (2003)

NARC(a) On-station Research(b) Policy Analysis

DOA/DLS-managedPTD:PPP, Minikits,IPM/FFS, etc.

NARC-managed(i) Outreach activities(ii) Other NARCmanaged PTDactivities

Farmers groups/Individually managedPTDPrivate sector

(Agribusiness firm)-managed PTD activities

PTD activitiesmanaged by otheractors

PTD activities indevelopment projcets.e.g. Third LivestockDev. Project, IrrigationProjects

Others Sources ofInformationrelevant to PTD I/NGOs-managed

PTD. E.g. LIBIRD,Care Int., ICRISAT

University-managedPTD

Other HMGministries (MoF, DoI,etc.)-managed PTD


9

technology innovation systems. The overall Nepalagricultural research and extension system is dominatedby the conventional crops-oriented ‘transfer oftechnology’ conceptualisation of R&D processes. Thisis partly due to the long-term connections with plantbreeders from the international centres for the majorfood crops grown in Nepal: rice, wheat, and maize.This mainstream plant breeder paradigm is verypersistent and is often unthinkingly applied in policydiscussions to other technologies, such as livestock,horticultural crops, agricultural engineeringtechnologies, and even to applied social science actionresearch. Actor linkage maps were used in a series ofworkshops to bring out the diverse nature of differentinnovations systems in Nepal. For example the actorlinkage map of the Nepal horticultural innovationsystem diagrammatically showed that NARC was a fairlyminor actor and the private sector actors bringing seedsfrom India and elsewhere were major actors in theexisting system. In the case of the livestock innovationsystem, the maps revealed that the Department ofLivestock had a significant research capability as didsome livestock development projects. The actor linkagemaps helped reveal these different institutional realitiesin various parts of the overall agricultural and naturalresources innovation system in Nepal.13 While therewere often lively discussions about what the futurerole of NARC in each of these innovation systemsshould be, the actor linkage maps helped keep thereality of the current situation prominent and provideda framework for thinking about future policies andprogrammes in a more open and outward lookingway.14

As regards monitoring in the AREP project, aninteresting development occurred in the mid-termreview. Up to that point the project had been goingvery badly and there was even talk of terminating theloan, the regular monitoring missions from the Bankhaving resulted in a series of uncomplimentary reports.Members of the review missions had often changed,they sometimes knew little about working in Asia andthey usually kept close to the original project blueprintdrawn up many years previously. However, for themid-term review the Bank team included two memberswho were very experienced in the analysis of research

and extension issues and practice in Asia in generaland in Nepal in particular. They were up to date oncontemporary thinking on pluralistic approaches toresearch and extension practice. The head of the teamalso had extensive experience in Asia. As regardsmonitoring and change in the project, one of theimportant outcomes of the review for NARC was thejoint drawing up of a limited number of action plansto be monitored and reviewed every six months. Theold confrontational culture between the Bank and theproject changed to being one of a supportivepartnership in addressing the dif ficult job ofrestructuring the Nepal agricultural research andextension innovation system.15 For its part the Bankagreed to keep the same reviewers who could comeregularly every six months to discuss how NARC wasprogressing in implementing its own plans and whatnew actions needed to be included. For the Bank tokeep the same reviewers, who knew the context inwhich understandings had been drawn up, wasapparently a novel idea. However, even here therewere funding problems within the Bank and it tookthe actions of another donor to provide funds, so thatthe Bank could keep to its commitment of regularmonitoring by the same Bank team.

The data from these six-monthly reviews, mainlyempirical evidence of institutional change taking placein the innovation system, has been summarised byGhimire et al., 2003.16 It is an excellent example of‘process documentation’ by ‘insiders’ who were creatingthe information as they went along for projectmanagement purposes. In a sense the mid-term reviewchanged monitoring from being an ‘outsider’confrontational evaluation exercise to a more usefulactivity, which resulted in both the Bank and NARCplaying a more constructive role in a difficult task.

In summary the use of actor linkage maps made anumber of important contributions to the project’s aims.They provided a way to investigate, document andlegitimise existing linkages, e.g. the PTD work. Theyprovided a framework that encouraged NARC scientiststo think in new ways and develop long-term changesin organisational structure and institutional behaviour.The six-monthly reviews based around the changinglocal action plan resulted in substantial changes in thedirection and content of the project. As in all innovationsystems the processes of institutional change never end.At the present time it is hard to forecast what will bethe institutional characteristics of the overall Nepalagricultural and natural resources innovation systemin a few years time. However, it can be predicted withconfidence that a return to old government research/extension institutional models is highly unlikely, as isa return to the expatriate/international science-ledinstitutional models of earlier years. Pluralisticinstitutional models are more likely to emerge, andthe actor linkage maps, used so far in NARC and otheragencies in Nepal, will probably continue to provide auseful framework for institutional analysis and action.

Figure 5 Example of an actor linkagemap for NARC

Private SectorActors (Farmers,agrovets, labourers,manufacturers etc.)

NARC

Civil SocietyActors (E.g.INGOs, localNGOs, advocacygroups, etc.)


10

Changing power tiller innovationsystems in NepalAn actor time line has been used recently in anotherproject in Nepal. In the early 1990s a conventionaltransfer of technology farming systems project wasstarted on the Terai (plains), where farming isdominated by rice/wheat cropping systems. The projectconcentrated on the introduction and development ofresource conservation technologies (RCTs). One of theprinciple technologies introduced was the Chinesepower tiller (PT) (a two-wheeled tractor/walkingtractor/mobile power unit) which can be used for,amongst other things minimum and zero tillageoperations. The project has changed over the yearsand now has far more of an interactive participatorytechnology development orientation, concentratingmore on poverty reduction and gender equity issues.However, one of its interesting features was that it wasnot linked in any systematic way with the strong, robustpower tiller innovation systems that have existed inthe Kathmandu and Pokhara valleys for many years.To some extent the work in the Terai power tillerproject had proceeded as if the previous power tillerinnovation systems had little to offer it. Some of thisattitude was due to a perception that the power tillersin the Kathmandu and Pokhara valleys were used onlyfor hauling construction goods and not for agriculturalpurposes. One of the ways of getting this projectassumption (narrative) questioned was to involve coremembers of the Terai team in a discussion and writinga paper about the changing overall PT tiller innovationsystem in Nepal (Biggs et al., 2003). In this exercise atime line was constructed (see Figure 6).

Each of these phases is associated with a particularcoalition of donor agencies and local and internationalactors. For example, from the end of the farmingsystems phase up to the current phase developmentswere facilitated to a major extent by a large coalition,comprising the United States Agency for InternationalDevelopment (USAID), the Department forInternational Development (UK) (DFID), the AsianDevelopment Bank (ADB), The New Zealand overseasaid programme, the International Fund for AgriculturalDevelopment (IFAD), the International Maize andWheat Improvement Center (CIMMYT) and the Nepal

Agricultural Research Council (NARC). The time linehas helped the current promoters of power tillers toinvestigate the outcomes of earlier projects and puttheir own work into an overall contextual perspective.One of the activities this led to was a brief exploratorysurvey by the current team of the power tillerinnovation system in the Kathmandu Valley. Theyinvestigated ownership patterns, usage patterns, andservice rental arrangements, etc. This has already ledto a major change in the team’s perceptions of theKathmandu power tiller system. They found that powertiller use for periods of peak agricultural demand washighly integrated with use for construction haulagework. This helped the Terai project staff to change theemphasis of their work and encourage the use of powertillers for both transport and agricultural purposes. TheTerai group also learnt that power tillers in theKathmandu Valley are owned by entrepreneurs (oftenrural entrepreneurs) who sometimes have some landof their own, in which case the power tillers are firstused on their own land at times of peak agriculturaldemand, before being hired out to others. Generallypower tiller operators were hired to operate the tractors.This quick analysis of the Kathmandu system led theTerai project staff to investigate more fully the ruralentrepreneur/service provider dimensions of powertillers in their own work. They have quickly learnt alot more from the Kathmandu system. This includesinformation on how power tillers have been maintainedover many years with little access to internationalmarkets for spare parts, and how local industries havedeveloped and manufactured locally relevantequipment. Getting the project staff to stand back frombusy day-to-day activities has been a challenge.However, investigating the history of the spread ofpower tiller technology and learning from these otherongoing and changing innovation systems in Nepalhas resulted in the project making a better use ofrelevant, available local knowledge. The time line helpedplay a role in this reflection and learning process.

One of the features of the power tiller projects isthat it has partners from a wide range of differentinstitutions; also power tillers are part of other projectswith different management, monitoring and reportingstructures. In order to manage the poverty reduction

1970 1980 1990 2000 2010Japanese phase First Chinese phase

Korean phase Farming systems research & extension (FSR&E) phase

Participatory technology development (PTD) phase Second Chinese phase Equitable access & gender phase

Poverty reduction and innovation systems phase

Figure 6 Time line of major phases in the spread of power tillers in Nepal


11

component of the power tiller project, the ‘coalition’around this set of interests has established for itself asimple ‘learning and response’ mechanism. Each sixmonths they meet and decide what they have learnedfrom planned sources (surveys, experiments and otherplanned activities), from other information picked upwhile doing their planned work, and from unexpectedchanges in the contextual environment. They thendecide what the implications of this information are,and draw up a tabular plan of action for themselvesfor the next six months. The team has found that thissimple table helps them to focus on what changes areneeded. The action plan is for them, and does notinclude recommendations concerning what ‘policymakers should do’ or what others ‘should do’. As anexample, some of the agricultural engineers wanted tocontinue with the on-farm experiments to show thatzero and minimum tillage was a good resourceconservation technology. Others in the groupcontended that this was already known from previouson-farm trials and that the technology was rapidlyspreading in the areas where the trials had been carriedout, meaning that further experiments of the same typewere unnecessary, even if they were budgeted for.Different experiments might be justified, but theythought it would be better at this stage to concentrateon methods of disseminating the ‘proven technology’.In consequence, plans for the next six months includedemailing the relevant public sector extension agenciesand making contacts with major donor and NGOdevelopment projects.

An example of picking up ‘unexpected’ informationwas the team’s discovery, while engaged in on-farmactivities, that power tillers were spreading withouttheir knowledge in adjoining areas and districts. Thiswas taking place outside of any planned activities. Theoutcome of this unexpected information was that aquick exploratory survey was planned for the wholeteam. There was a great deal of interest in what wasgoing on, and what could be learnt by the projectfrom these adjoining areas. Significantly the group didnot send ‘a socio-economic team’ off to investigateand write an adoption report!

Another example of how the project responded tototally unexpected changes in the context of the projectis shown by their response to talk in some governmentquarters of introducing a tax on the import of farmmachinery, which would be a total change ingovernment policy. In order to inform the debate onsuch possible changes the team decided to divertresources to work with some existing and new partnersto produce a policy discussion paper on ruralmechanisation in Nepal and to distribute earlier papersand reports to policy makers. The distribution will bequick and widespread amongst those who influencepolicy processes and practice. The team is finding thatthe learning and project response table is helping themto focus on changing priority issues, and manage theirwork accordingly. An analysis of the content of the

tables at some future date will enable an assessmentof how innovative the project was in learning andresponding to new information and opportunities as itwent along.

In summary, this case study has illustrated how actortime lines and a learning and response table havehelped the project to: change the perceptions of itswork, start learning from local knowledge which earlierhad not been recognised as important, and how toincrease the learning and responsiveness of the projectto information as it becomes available. Some mightwell argue that we are taking about nothing new here.True, but we argue that the actor-oriented toolsdescribed here have shown themselves to be useful,and might be useful to others.

3 DISCUSSION AND REFLECTIONSActor-oriented tools have helped us to:• map a given innovation system visually and analyse

strengths, weaknesses and opportunities in thesystem;

• encourage technology users to look at existing (oftenunexpected) strengths in an innovation system andanalyse the institutional implications of this;

• provide a framework whereby actors in a specificinnovation system have been able to change theirperceptions of their role and relationships to otheractors in the system;

• provide tools for planning, monitoring andevaluating coalition building and information flows;

• provide tools which are appropriate for use bygroups (as part of coalition building).We have found the tools to be valuable in keeping

partnerships, relationships and sharing information highon the research agenda. Their use helps to legitimiseand reward actors who actively build linkages in theirwork. They often provide a more structured way ofstrengthening institutional innovations that are alreadytaking place, but the importance of which have oftennot been acknowledged.

As the case studies show, the approach we areadvocating is not about observing and analysinginnovation systems from the outside. We are concernedwith developing approaches which can help actors andourselves as non-neutral actors, to reflect and learn aswe act on the inside.

Our experiences of working with these tools haveraised a number of issues, which are important toconsider when using the tools in future.

Political issues in using actor-orientedtoolsThese tools are all about human relationships andtherefore cannot help but be political. They need tobe used with sensitivity, awareness and with anacknowledgement that the user is never neutral. Actor-oriented tools may reveal information which someactors may not find easy to accept. It is also importantto realise that different actors may have different


12

interpretations of reality and that these interpretationsmay be politically motivated. The actor time lines helpto reveal some of these orientations in perceptions. Tosome extent the actor approach is enabling some topicsthat used to be seen as ‘academic’ political economysubjects to be brought out into the open and analysedwithin the framework of development activities.17 Inthe past the analysis of natural resources andagricultural innovations systems was often‘deinstitutionalised’ and ‘depoliticised’ by using actorlinkage maps which only had on them ‘Farmers(beneficiaries)’ , ‘Researchers’ and ‘ExtensionDepartment’, with two-headed arrows between them.Often the funders of research were not on the mapand the motives and reward structures within thoseand other organisations were not systematicallyanalysed. Another way of ‘depoliticising’ and‘deinstitutionalising’ the analysis of innovation systemswas to restrict planning and evaluation exercises tonarrow types of financial and economic analysis. Theactor approach we are suggesting here enables one tobreak out of these depersonalised, depoliticised, anddeinstitutionalised frameworks of analysis.

Because of the political nature of this approach,when planning the use of these tools it is important tobegin by being aware of your own aims and use thetools accordingly. These are not instruments to beadded to the tool bag of PRA, etc., to be taught in ashort-term training.18 If the tools are being used forproject and programme planning, preparation andbuilding coalitions, it is particularly important that theyare used in a constructive way. Like all tools, they canbe used for a wide range of purposes.

For example, we found in some situations thatquantifying linkages (something which is very tempting)can be unproductive as the value given to a linkage issomehow ‘set in stone’. Its seems to be less controversialand so more productive to identify strengths and ‘areaswhere there are further opportunities for intervention’rather than give quantitative weights to strong or weaklinkages. In addition, this is not ‘just a matter ofsemantics’. The way things are spoken about and usedis important. The quantification of some linkages in anobjective way can also lead sometimes to an unjustifiedconfidence in the figures produced. As with allanalytical approaches and tools there is always thepolitical economy question for those concerned. Thisconcerns whether members of development coalitionswish to be effective in policy processes anddevelopment practice that results in effectively bringingabout such things as poverty reduction and the socialinclusion of marginalised groups.

Importance of events and key locationsAn event such as a fair or a seminar or locations, e.g.marketplaces, can be critical in developing linkagesand coalitions. We have found it useful to identify keyevents and locations at the same time as doing our‘actor analysis’ to identify key actors. The actor event

time line also helps to pinpoint why key events in thepast were effective in some way. It also makes theteam more aware that ‘meetings’ ‘platforms’,‘workshops’, ‘seed fairs’, to be effective in bringingabout change in the innovations system, have to beplanned with great thought. We are not talking hereabout ‘ritualistic’ meetings, which are planned andorganised for a whole range of other reasons.

Importance of individual actorsIn her study of power systems in rural Bangladesh,Brigitta Bode (2002) recognises the enormous powerof local elites and recommends that NGOs identify the‘good kings’ amongst the elite who can work with themand patronise their activities. The work of Tendler(1997) in analysing ‘success stories’ in technologydiffusion in Brazil also points to the importance of keyindividuals in some processes. The actor approachenables us to move beyond structural linkages tounique opportunities, which may depend on aparticularly innovative or dynamic personality. Theactor event time line exercises often bring out theimportant role that a key individual played in pastinnovation processes. In the actor linkage matrix keyindividuals can be given a cell of their own.

Actor linkage maps versus actor linkagematricesJudgement has to be used as regards when to use oneor the other of these tools. In the Nepal situation, theactor linkage maps were the most useful way to getpeople to think about ways to strengthen linkages withnew actors and develop new mechanisms to facilitatethese linkages. In a number of meetings where actorlinkage matrices were discussed the tool did not appearto be of use. In the Bangladesh case the situation isdifferent. Here a small, stable group uses the matrixfrequently. As the group now understands and feelscomfortable with the ALM, it has become a usefulthinking tool: a way to visualise the institutional context,to monitor the impact of activities and to plan futureactivities. To some extent, all the tools discussed hereare time- and location-specific. Professionals withexperience in the use of these tools have to takeopportunities as they arise to use them as and whenappropriate. In a development situation the contextdetermines what is useful to be used when. In a moreacademic context one might attempt to be more‘rigorous’ in the pre-planned research design. However,this is not too different from any creative researchprocess. As those who are experienced in creativeresearch processes (rather than repetitious technical/social science research) know, analytical frameworksand tools are generally adapted and changed as thework proceeds.

A role for quantification?In some cases a more critical use of the matrix, orquantifying linkages, might be appropriate, for example


13

where team members are setting themselves linkagegoals and monitoring their performance. Here theymight find it useful to set criteria for ranking the strengthof a linkage. Methods for assessing the strength of alinkage will be highly location- and time-specificbecause of the institutional culture in different settings.For example, in one setting, having a meeting of someactors who normally never meet could be a majorpositive achievement. In another situation, having ameeting of those actors might be just a continuation ofmechanical or ritualistic meetings and be more anindicator of ‘business as usual’, rather than of significantchange.19 Similarly quantification might be useful for abaseline assessment of an innovation system. However,even here one has to avoid the pitfalls of oldapproaches where people thought ‘baselines’ couldbe established against which progress could be‘impartially’ monitored and evaluated. The case studyof the Nepal Agriculture Research Council clearlyshowed that the project ‘benchmark’ assertions thatthere was little PTD research taking place and thatNARC had few non-government R&D partners weremisleading. In this case this was partly due to thefact that no social science professionals withinstitutional analytical skills had been included inthe project preparation team, although almost all theproject was about restructuring and institutionalchange. It was not until the mid-term review thatthese issues were addressed in a more substantialway and the overall culture of the project changed(Biggs and Smith, 2003).

Cultural dimensions of transaction costsOne of the advocacies of many of today’s aid projectsis to reduce transaction costs. Often this has come inresponse to the fact that some government proceduresare inordinately slow and involve files being movedand signed many times. The actor tools presented herehelp to address transaction cost issues in a number ofways. First, they give a framework in which transactionsbetween actors can be conceptualised andsystematically analysed. Second, they provide aframework in which new types of transactionmechanisms can be explored. Third, they keep issuessuch as perceptions and culture to the forefront of theanalysis. For example, in the Nepal agricultural researchrestructuring example, no amount of training in newstakeholder collaboration methods and accountingprocedures to reduce formal ‘transaction costs’ wouldhave produced results while there was a culture ofconfrontation and little trust between the World Bankand the AREP project. In the Bangladesh case, theculture of the extension service (and the perceptionsof its staff) was that it worked in all areas of the country.Until this culture was challenged and a different realityacknowledged, there would have been little use intalking about minimising transactions costs in thecontext of any of the actors involved, let alone talkabout ways to develop new long-lasting partnership

relationships. The use of the actor linkage maps andthe actor time lines in the historical analysis of powertillers in Nepal has helped change the inward-looking‘special project’ culture that is so prevalent in Nepal.What in the past was seen as a negative transition cost(i.e. the negative costs of having to go and contactand work with extension and other developmentagencies) is now being seen as a worthwhile‘investment’ to keep the research more focused andhave partners who run with the new technology.

Finally, the tools help to remind us that any analysisof transaction cost will be very different in a highlydemocratic social/political system from that of anauthoritarian system.

Keep it simpleIt is tempting to be complex. But in order to use thesetools productively, i.e. interactively, they must be keptsimple and specific to the political and socio-economiccultures in which they are used. Its better to haveseveral maps, several matrices, several time lines ratherthan trying to describe too much in one place.Membership of meetings to use the tools needs to bethought out very carefully. Trying to use the actorlinkage matrix with a wide range of actors can resultin difficulties in convincing some people present of itsusefulness.

The actor-oriented approach to innovation systemsencourages the user to look at the whole range ofactors involved in an innovation system, including theusers of these techniques and their roles. In theBangladeshi case our research key actors and potentialfuture coalition members include farmers, Bangla-speaking government field staff, businessmen and abusiness advisory service. Some of these actors areunfamiliar with research and have difficultyunderstanding the point of these techniques. We foundthat non-researchers, and even some technicalresearchers, tend to take relationships as ‘obvious’ andare eager to move on to the action stage immediately!In the case of public sector researchers they oftenunderestimate the challenges involved when workingwith the private sector and NGOs. While public sectorpersonnel have rules and procedures about the wayinvitations for meetings are made and transferred, someof these mechanisms (some of the substance oflinkages) do not work when inter-acting with the privateand NGO sectors. For example, while a seniorbureaucrat in a Ministry might send an invitation for aworkshop to person in a lower level department andexpect the person to attend, in the case of NGOpersonnel, farmer groups, etc., they may have otherimportant competing schedules and may not be ableto attend without prior consultation as regards thetiming and location.

Involving people in the mapping and analysis seemsto help build some awareness of the usefulness of thetechnique. This is another reason to keep the toolssimple and strive to make them user friendly.


14

4 CONCLUSIONSWe hope we have shown that the actor approach andthe tools presented here can be useful in understandinginnovation systems and as a basis for planned actionand change. We feel that the techniques are relevantto addressing many of the issues which actors incontemporary innovation systems are now facing. Thetechniques are complementary to other research andplanning methods. They have their strengths andweaknesses. Like all theories, methods and tools theyhave to be handled with care, with experience and ina responsible way.

REFERENCESAshford, T., and Biggs, S. (1992) ‘Dynamics of rural and

agricultural mechanisation: The role of different actorsin technical and institutional change’. Journal ofInternational Development, Vol. 4, No. 4, pp. 349–74.

Bennett, L. (2003) ‘Towards an inclusive society:Agency, structure and diversity in Nepal’. Paperpresented at ‘The agenda of transformation:Inclusion in Nepali democracy’ organised by theSocial Science Baha, Kathmandu, 24–26 April 2003.

Berdegue, J.A. and Escobar, G. (2002) ‘Rural diversity,agricultural innovation policies and povertyreduction’. Agtricultural Research and ExtensionNetwork Paper No. 122. London: OverseasDevelopment Institute.

Biggs, S., (1990) ‘A multiple source of innovation modelof agricultural research and technology promotion’,World Development, Vol.18, No.11, pp. 1481–99.

Biggs, S., Justice, S., Gurung, C., Tripathi, J. and Sah,G. (2003) ‘The changing power tiller innovationsystem in Nepal: An actor oriented analysis’. Draft.Paper prepared for Workshop on Agricultural andRural Mechanisation, Bangladesh AgriculturalUniversity, Mymensingh, 2 November 2002 (revised4 March 2003).

Biggs, S. and Matsaert, H. (1999) ‘An actor orientedapproach for strengthening research anddevelopment capabilities in natural resourcessystems’. Public Administration and Development.Vol. 19, pp. 231–62.

Biggs, S. and Matsaert, F. (2000) ‘An actor orientedapproach to micro and small enterprisedevelopment: A Namibian case study’ in D. Lewisand T. Wallace (eds) New roles and relevance:development NGOs and the challenge of change.Bloomfield, CT: Kumarian Press.

Biggs, S. and Smith, G. (l998) ‘Beyond methodologies:Coalition-building for participatory technologydevelopment’. World Development, Vol. 26, No. 2,pp. 239–48.

Biggs, S. and Smith, S. (2003) ‘Paradox of learning inproject cycle management and the role oforganizational culture.’ World Development, Vol. 31,No. 10, pp. 1743–1757.

Biggs, S. and Underwood, M. (2001) Review of the croppost harvest research programme: Partnerships and

innovation systems. Crop Post Harvest Programme(CPHP). Chatham, UK: NRI.

Blumenthal, D. and Jannink, J.L. (2000) ‘A classificationof collaborative management methods’.Conservation Ecology, Vol. 4, No. 2, p. 13.

Bode, B. (2002) ‘In pursuit of power: Local elites andunion level governance in rural north-westernBangladesh’. Dhaka: CARE.

Bromley, D.W. (1989) Economic interests andinstitutions. Oxford: Basil Blackwell.

Byerlee, D. (1998) ‘The search for a new paradigm forthe development of national agricultural researchsystems’. World Development, Vol. 26, No. 6, pp.1049–55.

Clark, N. (l995) ‘Knowledge systems: Interactive natureof knowledge systems: Some implications for theThird World’, in Science and Public Policy, Vol. 22,No. 4.

Clark, Norman, A. Hall, R. Sulaiman and G. Naik, 2003,‘Research as Capacity Building: The Case of an NGOFacilitated Post-Harvest Innovation System for theHimalayan Hills,’ World Development, Vol. 31, No.11, pp. 1845–1863

Crew, E. and Harrison, E. (1998) Whose development.An ethnography of aid. London: Zed Books.

Davies, R. (2002) ‘Improved representations of changeprocesses: Improved theories of change’. Paperprepared for the 5th Biennial Conference of theEuropean Evaluation Society: ‘Three movements incontemporary evaluation: Learning, theory andchange’, 18 October 2002. www.mande.co.uk

Douthwaite, B. (2002) Enabling innovation: A practicalguide to understanding and fostering technologicalchange. New York and London: Zed books.

Ekboir, L. (ed.) (2002) CIMMYT 2000–2001 world wheatoverview and outlook: Developing no-till packagesfor small-scale farmers. Mexico, DF: CIMMYT.

Ellis, F., and Biggs, S. (2001) ‘Evolving Themes in ruraldevelopment 1950s–2000s,’ Development PolicyReview, Vol. 19, No. 4, pp. 437–448.

Eyben, R. (2003) ‘Relationships matter for influencingpro-poor change in Asia’. Report commissioned byDepartment for International Development, UK.Institute of Development Studies, University ofSussex, UK.

Falcon, W.P. (1967) ‘Agricultural and industrialinter-relationships in West Pakistan’, AmericanJournal of Agricultural Economics, Vol. 49, No. 5,December, pp. 1139–54.

Freeman, C. (1988) ‘Japan: a new national system ofinnovation?’ in G.Dosi et al., (eds), TechnicalChange and economic theory. London: PinterPublishers.

Gasper, D. (2000) ‘Evaluating the “Logical frameworkapproach”– towards learning-oriented developmentevaluation’ Public Administration and Development,Vol. 20, No. 1, pp. 17–28.

Gauchan, D. and Joshi, M. (2000) ‘NARC roles andlinkages with private sectors in participatory


15

agricultural technology development: Issues andconcerns’. Special Publication Series ORD 5/2057–58. Paper presented in the NARC – Private SectorConsultation Meeting, 27 July 2000. Kathmandu:Nepal Agricultural Research Council.

Gauchan, D., Joshi, M. and Biggs, S. (2000a) ‘A newNARC strategy for participatory technologydevelopment and linkages with multiple actors: Adiscussion paper’. Special Publication Series ORD3/2057-58. Kathmandu: Nepal AgriculturalResearch Council.

Gauchan, D. and Joshi, M. (2000b) NARC roles andlinkages with private sectors in participatoryagricultural technology development: Issues andconcerns. Special Publication Series ORD 5/2057-58. Kathmandu: Nepal Agricultural ResearchCouncil.

Gauchan, D., Joshi, M. and Biggs, S. (2000c) ‘Thehorticultural research and development system inNepal: A new strategy for Nepal’. Special PublicationSeries ORD 6/2057–58. Kathmandu: NepalAgricultural Research Council.

Gauchan, D., Joshi, M. and Biggs, S. (2003), ‘A strategyfor strengthening participatory technologydevelopment in agricultural and natural resourcesinnovations systems. The case of Nepal’. InternationalJournal of Technology Management and SustainableDevelopment, Vol. 2, No. 1, pp. 39–52.

Gellner, D. Hirsch, E. (eds) (2001) Inside organizations:anthropologist at work. Oxford: Berg.

Ghimire, Y. N., Pokhrel, T. P., Gauchan, D., Joshi, M.,Thakur, N. S. and Khadka, R. (2003) ‘Recent spreadof improved R and D linkages in national researchsystem of Nepal’. in M. Joshi, and N.S. Thakur, (eds),(2003) Recent spread and impact of agriculturaltechnologies in Nepal. Proceedings of the SixthNational Outreach Workshop, Khumaltar, Lalitpur,4th July 2002. pp. 6–13. Kathmandu: NepalAgricultural Research Council.

Grimble, R., and Wellard, K. (1997) ‘Stakeholdermethodologies in natural resource management: Areview of principles, contexts, experiences andopportunities’ Agricultural Systems, Vol. 55, No. 2 pp.173–193

Hall, A., Bockett, G. Taylor, S., Sivamahan, M.V.K. andClark, N. (2001) ‘Why research partnerships reallymatter: Innovation theory, institutional arrangementsand implications for developing new technologyfor the poor’. World Development, Vol. 29, No. 5,pp. 783–98.

Hogg, D. (2000) Technological change in agriculture:locking in to genetic uniformity. Basingstoke, UK,and New York: Macmillan Press (UK) and St. Martin’sPress (USA)

ISNAR (2001) ‘Meeting the challenge of ecoregionalresearch’. International Workshop on Organizingand Managing Ecoregional Programmes.International Agricultural Centre, Wageningen, TheNetherlands, 26–28 March 2001. The Hague:

International Service for National AgriculturalResearch.

Kayastha, B.N., Mathema, S.B. and Rood, P. (1989)‘Nepal: Organisation and management of on-farmresearch in the national agricultural research system’,OFCOR Case Study No. 4, The Hague: InternationalService for National Agricultural Research (ISNAR).

Kidd, J. (2002) ‘Agricultural and rural livelihoods: Isglobalisation opening or blocking paths out of ruralpoverty?’ Agricultural Research and DevelopmentNetwork Paper No. 121. London: OverseasDevelopment Institute.

Krishna, A., Uphoff, N. and Esman, M.J. (eds) (1998)Reasons for hope: Instructive experiences in ruraldevelopment. West Hartford, CT: Kumarian Press.

Lewis, D.J., (1998) ‘Partnership as process: Buildingan institutional ethnography of an inter-agencyaquaculture project in Bangladesh’, in D. Mosse, J.Farrington and A. Drew (eds). Development as aprocess: Concepts and methods for working withcomplexity. London: Overseas DevelopmentInstitute and New York: Routledge.

Long, N. and Long, A. (1992) Battlefields of Knowledge.London: Routledge.

Long, N. and Van Der Ploeg, J.D. (1989)‘Demythologizing planned intervention: An actorperspective’ Sociologia Ruralis, Vol. 29, No. 3/4,pp. 226–49.

Matsaert, H. (2003) An introduction to actor orientedtools, available at www.developmentwheel.org

Messerschmidt, D. (ed.) (1981) Anthropologists at homein North America: Methods and issues in the studyof one’s own society. New York: CambridgeUniversity Press.

Mosse, D., (2001) ‘Social research in rural developmentprojects’ in D. Gellner and E. Hirsch (eds). Insideorganizations: Anthropologist at work. Oxford: Berg.

Mosse, D., Farrington, J. and Drew, A. (eds) (l998)Development as a process: Concepts and methodsfor working with complexity. London: OverseasDevelopment Institute and New York: Routledge.

Mundy, P. (2003) ‘Knowledge systems matrix’. (http://www.mamud.com/knowledgesystems.htm)

Nelson, R.R. and Winter, S.G. (1977) ‘In search of usefultheory of innovation’. Research Policy, Vol. 6, No.1, pp. 36-76.

ODA (1995) Guidance note on how to do stakeholderanalysis of aid projects and programmes. London:Social Development Department, OverseasDevelopment Administration, HMG.

Pandey, S.B., Kuwar, B.S. and Biggs, S. (2002) ‘Changinggoat technology innovation system in Nepal: Pastoutcomes and future directions’. Draft. Kathmandu:Nepal Agricultural Research Council.

Raina, R.S. (2002) ‘Disciplines, institutions andorganisations: Impact assessments in context’. Paperfor the international conference ‘Why has impactassessment research not made more of a difference?’,4–7 February 2002, San Jose, Costa Rica.


16

Ramirez, R., (1999) ‘Stakeholder analysis and conflictmanagement’ in Buckles, D. (ed) (1999) Cultivatingpeace: Conflict and collaboration in naturalresources management. Chapter 5, pp. 101–126Ottawa: International Development ResearchCentre.

Roling, N. and Jiggins, J. (1998) ‘The ecologicalknowledge system’ in N. Roling and A. Wagemakers(eds). Facilitating sustainable agriculture.Participatory learning and adaptive managementin times of environmental uncertainty. Cambridge:Cambridge University Press.

Squires, S. (1999) American plant improvement forAfrican farmers? An anthropology of technologydevelopment. Thesis submitted for the degree ofPhD, Department of Anthropology, University ofLondon.

Subedi, A. (2000) ‘Linkages and partnership inagricultural research: Experience of a NepaleseNGO, LI-Bird’ in M. Joshi, D. Gauchan and N.Thakur Proceedings of the 5th National OutreachResearch Workshop 30–31 May 2000. Kathmandu:NARC.

Temel, T., Janssen, W., and Karimov, F. (2003). ‘Systemsanalysis by graph-theoretic techniques: Assessmentof the agricultural innovations system of Azerbaijan’.Agricultural Systems, Vol. 77, No. 2, pp. 91–116.

Tendler, J. (1997) Good government in the tropics.Baltimore MD: Johns Hopkins University Press.

Westendorp, A. and Biggs, S. (2003) ‘Strengtheningsocial capital in agricultural and natural resourcesinnovation systems: The case of the community IPMfarmer field schools (FFSs) Programme in Nepal’.School of Development Studies, University of EastAnglia, Norwich, UK. (March 2003, Draft Paper).

ENDNOTES1 For example, if an international crop research

programme were to conduct a systematic technical/economic analysis to establish priority regions towork in, the tools of this actor analysis could beused to help systematically establish how suchtechnical priorities might be ‘implemented’. Thetools then help address the issues of which actorswould play which roles and in what way, indifferent situations, in the ever-changing political,cultural, economic and institutional context inwhich S&T takes place.

2 Institutions are the formal and informal ‘rules ofthe game’, while organisations are the formalinstitutes that make up the system, e.g. researchinstitutes, private and public sector extensionagencies, membership organisations, registeredNGOs, etc. Following Bromley, we use the terminstitutions in a general way to mean both the rulesof the game and the formal institutions.

3 This is similar to the findings of Norman Clark whostresses the importance of interactiveness indynamic Science and Technology (S&T) knowledge

systems (Clark, 1995). Biggs and Smith (1998) intheir analysis of natural resource research systemsalso emphasise the importance of coalitions in R&Dactivities: ‘the effectiveness of coalitions will oftenbe a key determinant of long term impacts oftechnical innovations’. These findings are supportedin a recent review of innovation systems byBlumenthal and Jannink (2000) who observe that‘collaboration among multiple stakeholders can becrucial to the success of natural resourcesmanagement’.

4 Of course, if institutional strengthening ofinnovation systems is a goal of projects/programmes these topics can be systematicallyaddressed and brought into log frames, andindicators developed to monitor the strengthening(or weakening) of the institutions (Gasper, 2000).

5 For a recent brief and practical introduction to theactor oriented tools presented here see Matsaert,2003. For a review of literature on planning,monitoring and evaluation, and a description ofthe actor approach as it might be applied to naturalresources policy and management processes seeour earlier paper, Biggs and Matsaert, 1999.

6 For some early systematic attempts to use actorlinkage matrices see the publications of the ISNARstudy on the On-farm client oriented research(OFCOR) project (e.g. Kayastha et al., 1989).However in that study they were used in anexternal, ex post evaluation mode. In this paperwe are looking at the way these tools can be usedwithin planned policy and development activities.A useful new addition to the literature onethnographic research methods is by Gellner andHirsch, 2001. Readers interested in contemporaryethnographic studies of development actors shouldsee the papers prepared for the workshop on Orderand Disjuncture: The Organisation of Aid andDevelopment organised by David Lewis and DavidMosse, 26/27 September 2003, at the School ofOriental and Asian studies (SOAS), University ofLondon. The papers are available at the web site:h t t p : / / w w w . s o a s . a c . u k / d e p a r t m e n t s /departmentinfo.cfm?navid=460

7 For an example of where actor linkage maps wereused to highlight the need to bring ‘researchfunders’ into the analysis of innovation systemssee Gauchan, Joshi and Biggs (2003).

8 This can easily be set up on Microsoft Excel.9 Ellis and Biggs (2001) used a time line to map major

changes in the mainstream agricultural and ruraldevelopment discourse from 1950 to 2000. Thearticle was written from an agricultural economist’sperspective. To some extent that articledocumented the product of using this tool ofanalysis in rural development postgraduate teachingfor many years in the School of DevelopmentStudies, University of East Anglia, UK. It isinteresting to note that David Mosse uses a similar


17

tool to encourage reflection and learning in auniversity teaching context in London (Mosse,2001). At the International Center for TropicalAgriculture (CIAT), Colombo, Boru Douthwaite isdeveloping a tool called The Innovation LifeHistories and How to Construct Them. This isanother example of attempts being made tostrengthen empirically based reflective learningwithin agricultural research systems, and highlightthat there are often many contending andcompeting historical narratives about cause andeffects relationships in past innovation systems.

10 In our earlier paper we illustrate how these toolscan be used in an ex post project evaluation of afarming systems research and extension project inNamibia (Biggs and Matsaert, 1999). A similar expost analysis was carried out on a micro enterpriseproject in Namibia (Biggs and Matsaert, 2000).

11 In a recent discussion of social science researchmethods that have largely evolved from withindevelopment situations Mosse (2001) defines twomain streams of methods: (i) Participatory Learning(also known as Participatory Rural Appraisal (PRA)),and (ii) Process Documentation Research or processmonitoring. In the first category it is unusual forthe ‘researcher or development team’ to placethemselves in the actor linkage map andsystematically analyse their relationships withdifferent actors. In process monitoring it more usualfor the researcher or development team to bereflective and consciously analyse their ownbehaviour and its effects on other actors. Thereforethey are more likely to include themselves in theactor linkage map. In the actor approach we areinvestigating here the researchers (or team) arealways included in the actor linkage map. As muchof the analysis is about their motives, roles andbehaviour with respect to other partners, as it isabout looking at the relationships between otheractors in that specific context. As part of the presentresearch activity one of the authors (SB) is followinganother type of insider/outsider researchmethodology. In this work the ‘outsider’ (SB) isworking very closely with ‘insiders’ to write upactor-oriented contemporary ethnographies of theinnovation systems of which they are a part. Forexample see Westendorp and Biggs, 2003; Biggset al., 2003; and Pandey, Kuwar and Biggs, 2002 .While there are predictable problems as regards‘the objectivity’ of the analysis there are greatadvantages in that the documents carry with thema degree of ‘insider’ authenticity and in-depthanalysis and insights that studies by ‘outsiders’ donot carry. Tendler’s perceptive studies of ruraldevelopment ‘successes’ in Brazil were conductedby an ‘outsider’. The autobiographical studies ofsuccesses in Krishna, Uphoff and Esman, 1998, aretotally written by the insiders themselves.Messerschmidt’s book, Anthropologists at home in

North America: Methods and issues in the study ofone’s own society, takes up the theme of lookingat the problems faced by anthropologists who workas ‘outsiders’ but study their own home cultures.

12 There was also a great deal more PTD taking placein NARC if one took into account the informalpersonal contacts scientists had with farmers,NGOs, farmers’ associations, consultancycompanies, etc. In addition some NARC researcherswere farmers themselves and some, in their privatecapacity, had seed multiplication farms. However,these informal linkages were not investigated atthe time.

13 The NARC outreach division published the papersdescribing these different major innovation systemsin the agricultural sector. They were all written forplanning workshops attended by major actors ineach of the innovation systems. The papers coveredcrops and soil fertility (Gauchan, Joshi and Biggs,2000 a), livestock (Gauchan, Joshi, and Biggs, 2000b) and horticultural crops (Gauchan, Joshi, andBiggs, 2000 c). An excellent paper by Subedi at anoutreach workshop in July 2000 showed how actorlinkage maps could be used to represent the waysa major local NGO (LIBIRD), made different typesof partnership linkages in different projects withgovernment, private and NDO actors (Subedi,2000).

14 Gauchan (pers. comm.) observes that the actorapproach has helped NARC staff to envision thenew NARC mandate and thrust areas in the recentlydeveloped and published NARC long-term visionfor 2021.

15 For a more detailed analysis of the Nepal case andthe importance of addressing in a substantial,analytical way cultural issues within projects,programmes, development coalitions, etc. see Biggsand Smith (2003).

16 For readers interested in how to define, recordand measure changes in ‘social capital’ this papermakes good reading. It uses similar indicators tothose used in Lewis (1998) and Westendorp andBiggs, 2003.

17 For an ethnography of aid which has a strongorientation towards natural resources innovationsystems see Crew and Harrison, 1998. Forethnographically oriented studies of international/national natural resources and agricultural researchsystems see Squires (1999) and Hogg (2000).

18 Although, of course, we recognise there will bereflective, open-minded people who may beexposed briefly to the tools and integrate theminto their personal and professional life accordingly.

19 Because of the diversity of political, cultural andinstitutional contexts in which actor approachesare used it would be unwise to try and come upwith international or global indicators to assesschanges in linkages and other institutionalcharacteristics of innovations systems.


18


19

Network Papers cost £3.00 sterling each (plus postage & packing).This paper may be downloaded free of charge from the AgREN website at: www.odi.org.uk/agren

Please contact the Network Administrator at:The Overseas Development Institute, 111 Westminster Bridge Road, London SE1 7JD, UK

Telephone: +44 (0)20 7922 0300 Fax: +44 (0)20 7922 0399 Email: [email protected]

Information about ODI and its publications can be found on our website at:http://www.odi.org.uk/

STRENGTHENING POVERTY REDUCTION PROGRAMMES USING … · social and political context of their...

Documents

Transcript of STRENGTHENING POVERTY REDUCTION PROGRAMMES USING … · social and political context of their...