process approaches to development: theory and sri lankan practice

20
Process Approaches to Development: Theory and Sri Lankan Practice RICHARD BOND and DAVID HULME * University of Manchester, Manchester, UK Summary. — Calls for ‘‘process projects’’ in the 1980s led to the development of a body of knowledge about process approaches and the implementation of innovative programs and projects. During the 1990s the focus has moved on to the narrower subject of beneficiary participation and much potential learning about how to promote develop- ment is being lost. This paper reviews the notion of process approaches and produces a conceptual framework that synthesizes these ideas. It then uses this framework to analyze 12 years of experience of a NORAD-financed IRDP in Sri Lanka. The findings point to the continued relevance of deepening the theory and practice of process approaches in development. Ó 1999 Elsevier Science Ltd. All rights reserved. 1. INTRODUCTION Disillusionment with the conventional approach to planning and managing develop- ment projects led to widespread calls for the adoption of ‘‘process’’ approaches throughout the 1980s (Gow and Morss, 1988; Honadle and Van Sant, 1985; Korten, 1980; Rondinelli, 1983) 1 . In contrast to rigid ‘‘blueprint’’ approaches these alternatives emphasized experimentation, learning, adaptation, partici- pation, flexibility, building local capacities and organic expansion. During the 1990s nongov- ernment organizations (NGOs), donors 2 and developing country governments have taken up such ideas (Eyben, 1991; Mosse, Farrington and Rew, 1998; Edwards, forthcoming) but there remains a lack of clarity over exactly what a process approach to projects means. Surprisingly, given the large volume of work done in the 1980s, there have been few attempts to explore the empirical experience of ‘‘process projects.’’ 3 Most discussion and study has focused on the important, but narrower, issue of participation. This paper seeks to partially fill these gaps in the contemporary literature. In its first part it explores the ideas that underpin process approaches and produces a conceptual frame- work that synthesizes these ideas. Subse- quently, it uses this framework to analyze 12 years experience of a process project in Sri Lanka. While there is much evidence that this project can be considered successful there have been significant problems in managing a process approach that generate potential lessons. In its conclusion the paper highlights a number of important findings about how process approaches might be understood and operationalized more eectively in the future. 2. PROCESS APPROACHES: A LONG VOYAGE OF DISCOVERY The term ‘‘implementation’’ understates the complexity of the task of carrying out projects that are aected by a high degree of initial ignorance and uncertainty. Here ‘‘project implementation’’ may often mean in fact a long voyage of discovery in the most varied domains, from technology to politics (Hirschman, 1967, p. 35). World Development Vol. 27, No. 8, pp. 1339–1358, 1999 Ó 1999 Elsevier Science Ltd. All rights reserved. Printed in Great Britain 0305-750X/99/$-see front matter PII: S0305-750X(99)00060-1 www.elsevier.com/locate/worlddev * We would like to acknowledge with gratitude the contributions support, and encouragement of the follow- ing; Mr. Arild Sk ara, Mr. S. Amerasekara, Mr. H. Gunadasa, Mr. P. Amarasinghe, Mr. G. Batuwitage, the Regional Development Division of the Sri Lankan Ministry of Plan Implementation and Parliamentary Aairs and the Norwegian Agency for Development Cooperation. Last, but hopefully coming first in the future, the people of Moneragala District for their time, patience and good humor. Final revision accepted: 25 January 1999. 1339

Upload: richard-bond

Post on 16-Sep-2016

218 views

Category:

Documents


0 download

TRANSCRIPT

Process Approaches to Development: Theory and

Sri Lankan Practice

RICHARD BOND and DAVID HULME *

University of Manchester, Manchester, UK

Summary. Ð Calls for ``process projects'' in the 1980s led to the development of a body ofknowledge about process approaches and the implementation of innovative programsand projects. During the 1990s the focus has moved on to the narrower subject ofbene®ciary participation and much potential learning about how to promote develop-ment is being lost. This paper reviews the notion of process approaches and produces aconceptual framework that synthesizes these ideas. It then uses this framework toanalyze 12 years of experience of a NORAD-®nanced IRDP in Sri Lanka. The ®ndingspoint to the continued relevance of deepening the theory and practice of processapproaches in development. Ó 1999 Elsevier Science Ltd. All rights reserved.

1. INTRODUCTION

Disillusionment with the conventionalapproach to planning and managing develop-ment projects led to widespread calls for theadoption of ``process'' approaches throughoutthe 1980s (Gow and Morss, 1988; Honadle andVan Sant, 1985; Korten, 1980; Rondinelli,1983)1. In contrast to rigid ``blueprint''approaches these alternatives emphasizedexperimentation, learning, adaptation, partici-pation, ¯exibility, building local capacities andorganic expansion. During the 1990s nongov-ernment organizations (NGOs), donors2 anddeveloping country governments have taken upsuch ideas (Eyben, 1991; Mosse, Farringtonand Rew, 1998; Edwards, forthcoming) butthere remains a lack of clarity over exactly whata process approach to projects means.Surprisingly, given the large volume of workdone in the 1980s, there have been few attemptsto explore the empirical experience of ``processprojects.''3 Most discussion and study hasfocused on the important, but narrower, issueof participation.

This paper seeks to partially ®ll these gaps inthe contemporary literature. In its ®rst part itexplores the ideas that underpin processapproaches and produces a conceptual frame-work that synthesizes these ideas. Subse-quently, it uses this framework to analyze 12years experience of a process project in SriLanka. While there is much evidence that this

project can be considered successful there havebeen signi®cant problems in managing aprocess approach that generate potentiallessons. In its conclusion the paper highlights anumber of important ®ndings about howprocess approaches might be understood andoperationalized more e�ectively in the future.

2. PROCESS APPROACHES: A LONGVOYAGE OF DISCOVERY

The term ``implementation'' understates thecomplexity of the task of carrying out projects that area�ected by a high degree of initial ignorance anduncertainty. Here ``project implementation'' mayoften mean in fact a long voyage of discovery in themost varied domains, from technology to politics(Hirschman, 1967, p. 35).

World Development Vol. 27, No. 8, pp. 1339±1358, 1999Ó 1999 Elsevier Science Ltd. All rights reserved.

Printed in Great Britain0305-750X/99/$-see front matter

PII: S0305-750X(99)00060-1www.elsevier.com/locate/worlddev

* We would like to acknowledge with gratitude the

contributions support, and encouragement of the follow-

ing; Mr. Arild Sk�ara, Mr. S. Amerasekara, Mr. H.

Gunadasa, Mr. P. Amarasinghe, Mr. G. Batuwitage, the

Regional Development Division of the Sri Lankan

Ministry of Plan Implementation and Parliamentary

A�airs and the Norwegian Agency for Development

Cooperation. Last, but hopefully coming ®rst in the

future, the people of Moneragala District for their time,

patience and good humor. Final revision accepted: 25

January 1999.

1339

The process approach to development projectshas several di�ering origins. All stemmedhowever from concern about the poor perfor-mance of development projects which took ablueprint or top-down approach to develop-ment. The blueprint approach is based withinthe concepts of objective rationality andreductionism and has roots in the ®elds ofengineering and construction. Prescribed stepslead through the stages of the project cycle;experts design and control activities; detailedplanning at the beginning speci®es objectives,targets, outputs, resources and schedules; localinstitutions are bypassed if they have inade-quate capacity and the job of management is toimplement as closely as possible to the plan.

Process approaches have evolved as a reac-tion to the ine�ectiveness of such a method toachieve development goals in diverse andcomplex environments. At their heart is therecognition that the challenges of developmentare not well-structured problems that can be``thought through'' by clever people. Rather,they are ``messes'' that have to be ``acted out''by social experimentation and interaction(Johnston and Clark, 1982, pp. 23±28). Theyaccept that many things are unknown at thestart of a project; that problem-solving capac-ities and local institutions must be built-up notsimply ``bought in''; that much activity isexperimental and error must be embraced; thatthe intended bene®ciaries of projects are actorsin their own right and not mere recipients; andthat large-scale initiatives should start smalland grow organically at rates dependent ontheir achievements.

Proponents of process approaches fall intotwo main schools of thought. Purists (such asChambers, 1997; and Korten, 1980) emphasizebene®ciary participation and learning and, ine�ect, argue for the abandonment of theconcept of ``project.'' Process is seen assynonymous with local institutional develop-ment in which the role of external agents andresources should be minimized. Where externalinterventions occur then these should not beprojects but ``para projects'' (Upho�, 1990)that help poor and disadvantaged people tohelp themselves, through raising their capaci-ties, rather than attempting to achieve exoge-nously de®ned ``impacts.'' In contrast,managerialists (such as Brinkerho� and Ingle,1989; Brinkerho�, 1992; Rondinelli, 1993; andSweet and Weisel, 1979) still see a signi®cantrole for external actors but argue that projects,managers and management systems must be

more ¯exible and adaptive. In practice, theseschools of thought can be seen as points on acontinuum that ranges from minimalistapproaches to support pre-existing, endoge-nous development initiatives and organizations,to relatively conventional project approachesthat see a limited need for ``process'' tostrengthen their problem-solving abilities. Theformer position is one that relatively idealisticNGOs seek to achieve. The latter, is the onewhich many aid donors, particularly thoseunder pressure to account for performance onshort time horizons4, have found it expedient toadopt.

In the paragraphs that follow we identify the®ve main elements of a process approach drawnfrom a comprehensive review of the literature(Bond, 1997). This creates a conceptualframework that permits speci®c examples ofprocess projects to be analyzed and comparedwith each other (While this paper focuses onrural development, Moser, 1998, p. 16 arguesfor the process approach to poverty reductionin an urban setting). The relationships betweenthe major elements of a process approach areillustrated in Figure 1. In this diagram, processde®nes relationships between the bene®ciariesand the management, not in an abstract,conceptual way, but through the activities ofspeci®c interventions. The diagram thereforelinks bene®ciaries to management by interven-tions. The spectrum of characteristics of theinterventions connects the three main elementsof process directly relevant to that relationship;participation, learning and ¯exibility. Thesecharacteristics overlap; participation is requiredof the bene®ciaries, ¯exibility is required of themanagement and learning links the two, over-lapping as shown. This gives a spectrum of fourdistinct characteristics; bene®ciary participa-tion, learning from the people, iterative learn-ing and managerial ¯exibility. Institutionalsupport relates to the environment containingboth management, bene®ciaries and the inter-ventions. Internal managerial aspects ofprocess lie within the circle of management.

(a) Flexible and phased implementation

Within the framework of a clear vision ofgoals, deliberate phasing and ¯exibility arerequired (Sweet and Weisel, 1979; Brinkerho�and Ingle, 1989; Conyers and Kaul, 1990)

ÐStart small and expand: piloting to buildcapacities and making inevitable mistakes only

1340 WORLD DEVELOPMENT

on a small-scale (Brinkerho� and Ingle, 1989;Conyers and Kaul, 1990; Cernea, 1991; Picciot-to and Weaving, 1994).ÐLong time frames: poverty reduction and pro-poor economic growth are unlikely to beachieved in conventional three-year or ®ve-yearprojects ± 10 to 20-year timeframes are morerealistic (Morss et al., 1976; ODA, 1995;Thompson, 1995).ÐExperimentation: ¯exibility (in methods, ®-nances and human resources) and careful phas-ing are necessary for experiments to be trialedand to be carried into further action (Rondinel-li, 1983, 1993; Brinkerho� and Ingle, 1989)ÐAction Learning Cycles: activities need to bephased according to the experiences of earlierlearning cyclesÐthey cannot be programmedat project initiation, thus rolling planning be-comes necessary (Lewin, 1946; Honadle andVan Sant, 1985; Brinkerho� and Ingle, 1989;Zuber-Skerritt, 1991).

(b) Learning from experience

Learning from experience means recognizingand managing ``the unknown'' in developmentthrough systematized and institutionalizedlearning. It also implies seeing development asa product of learning (Acko�, 1984; Pretty andChambers, 1993).

ÐEmbracing error: attitudes and organizationalcultures must accept error as an aid to learning(Korten, 1980). Without this mistakes may behidden and repeated (Michael, 1973; Korten,1980; Rondinelli, 1993; Chambers, 1997;Hulme, 1989).ÐLinks between implementation and planning:deliberate links between planning and imple-mentation should be established to ensure les-sons learned feed into new plans (Korten,

1980; Rondinelli, 1983, 1993; Honadle andVan Sant, 1985; Brinkerho� and Ingle, 1989).ÐIterative improvement of small interventions:the chief mode of learning is through the re-peated application of lessons learned to ongoingor new initiatives often on a small, but expand-ing scale (Brinkerho� and Klauss, 1985; Upho�,1992).ÐBe e�ective, become e�cient and then expand:a ``learning process approach'' is needed whereone ®rst learns how to do something e�ectively,then how to do it e�ciently and ®nally how toexpand its impact (Korten, 1980).ÐAppropriate technologies: this representslearning from indigenous knowledge and the lo-cal environments only to use technologies whichare appropriate to the local context (Sweet andWeisel, 1979; Korten, 1980).

(c) Bene®ciary participation

Project managers have to understand thecomplex social realities of local situations ifthey are to be successful (Korten and Alfonso,1983).

ÐIn problem analysis: local people are the oneswho best know their situation and problems(Cernea, 1985).ÐIn planning and decision-making: bene®ciaryinvolvement in the planning process ensuresrelevance, support and some measure of controlover their futures (Sweet and Weisel, 1979; Kor-ten, 1980; Rondinelli, 1993).ÐIn resource mobilization and implementation:local people can often mobilize many resourcesthus encouraging sustainability and lower exter-nal costs. Their involvement in implementationincreases a sense of ownership and subsequentmaintenance of assets created (Sweet and Wei-sel, 1979; Honadle and Van Sant, 1985).

Figure 1. Interrelationship of the main characteristics of intervention in a process project.

PROCESS APPROACHES TO DEVELOPMENT 1341

ÐIn monitoring and evaluation: an important as-pect of local control, participation in monitor-ing can ensure timely and quality work; inevaluation it can ensure ultimate acceptabilityof the project interventions (Upho�, 1985;Mosse, 1995; Mosse, Farrington and Rew,1998).ÐEmpowerment of bene®ciaries: ®nally partici-pation throughout the process of change hasits own bene®ts in terms of strengthening localorganizations and increasing con®dence toundertake self-initiated development (Fried-mann, 1984; Brinkerho� and Ingle, 1989; Con-yers and Kaul, 1990).

(d) Institutional support

Avoiding bypass and helping to build-uplocal institutional capacities within animproved operational environment are keyelements of institutional support (Dale, 1988):

ÐPolitical support: is often essential to thesuccessful outcome of projects and programs,this includes both national-level and local-levelsupport (Ickis, 1983; Conyers and Kaul, 1990;Rondinelli, 1993).ÐDevolved authority: the ¯exibility of a processapproach requires some measure of devolvedauthority. Frequent decision-making and themeans to carry them out, both at a locality closeto the activities being undertaken is the natureof process management (Sweet and Weisel,1979; BirgegaÊrd, 1991; Rondinelli, 1993).ÐUse of permanent institutions: temporary,well-resourced project management units canimplement projects quickly and e�ciently butdo not contribute to sustainable institutionbuilding. A process project will support theuse of existing agencies according to theircapacities (Honadle and Van Sant, 1985; Gold-man, Mellors and Pudsey, 1989).ÐLocal capacity building: development is notjust about improved facilities but also aboutstronger, more capable institutions and individ-uals (Korten, 1980; Brinkerho� and Ingle, 1989;BirgegaÊrd, 1991; Rondinelli, 1993).ÐOrganizational change: sometimes quitefundamental re-orientation is required withinkey institutions for development to proceed(Korten, 1980; Korten and Alfonso, 1983; Kor-ten and Siy, 1988; Rondinelli, 1983, 1993).ÐFacilitating bene®ciary organization: whilemany project activities may be carried out usingexisting government, private commercial andvoluntary agencies, membership organizations

of the bene®ciaries should be encouraged andsupported to provide a strong civil base foron-going development (Sweet and Weisel,1979; Honadle and Van Sant, 1985; Friedmann,1992).

(e) Program management

Management is central to process approacheswhich should not be used to justify manage-ment by abandonment (Honadle and Van Sant,1985; Roe, 1993)

ÐWell-quali®ed and motivated leadership: animportant element in successful projects gener-ally and more so in process projects. Flexibilityand the management of complexity requirestrong leadership (Korten, 1988; Conyers andKaul, 1990; Brinkerho�, 1992; Thompson,1995).ÐNew professionalism: Chambers (1993) con-cept of the reversals of attitudes typical of nor-mal professionalism have a special place in themanagement of process approaches where thequality of intervention by key facilitators is soimportant.ÐRetention of key sta�: stability and continuityof sta� so that complex lessons are not lost andso that there is a good corporate understandingof the ``di�erent'' approaches being undertaken(Chambers, 1988, 1993; Rondinelli, 1993).ÐVariety of short-term technical assistance: re-sponse to complex and diverse problems oftenneeds external advice, an ability to provideshort-term consultation to often remote areasis needed (Honadle and Van Sant, 1985).Long-term technical assistance in a facilitatingrole: the role of long-term technical assistance(TA) support should be restricted to facilitation,advice and training, and should exclude execu-tive roles where dependency may be created(Honadle and Van Sant, 1985; Brinkerho� andIngle, 1989; BirgegaÊrd, 1991).ÐProject management unit (PMU) with ¯exi-ble, informal approach: whatever institution ischosen to host the project management unit, itmust be free to develop a ¯exible, innovativeand informal approach in order to avoidbureaucratization, to initiate relevant and fastresponses and to encourage action by otheragencies (Honadle and Van Sant, 1985).ÐCreative management: the fast-moving worldof a process project needs imaginative ap-proaches to problems and project management(Brinkerho� and Ingle, 1989; Rondinelli,1983, 1993; Thompson, 1995).

1342 WORLD DEVELOPMENT

ÐInterorganizational coordination: many agen-cies and groups will come together in a processproject or program, especially in large-scaleintegrated programs (Korten and Alfonso,1983). This requires the establishment of goodcoordination, independent of the PMU (Sweetand Weisel, 1979; Rondinelli, 1993).

3. THE MONERAGALA INTEGRATEDRURAL DEVELOPMENT PROGRAM

(MONDEP): BACKGROUND ANDPERFORMANCE5

Sri Lanka's Moneragala IRDP (MONDEP)started in 1984, the 11th Integrated RuralDevelopment program (IRDP) under theNational Program of the then Ministry ofPolicy, Planning and Implementation (MPPI).The prime objective of the National Programwas to raise levels of social and economicdevelopment in districts by-passed by earliernational lead projects. It was supported in twophases by some US$ 28m grant assistance fromthe Norwegian Agency for Development Co-operation (NORAD), a commitment which hasbeen long-term, non-interfering, ¯exible, andsupportive. MPPI has encouraged a signi®cantdegree of experimentation and autonomywithin the district programs that comprise theNational Program. While it allows great free-dom in terms of operational activity it hasmaintained control of budgets and personnel.MONDEP covers the second largest district inSri Lanka and probably the most severelydisadvantaged outside of the civil war zone. Itcovers some 5,600 sq.km, is landlocked and ispredominantly ¯at to gently undulating land oflow altitude with a mountainous boundary inthe north-west. The district spans the ``inter-mediate'' and ``dry'' zones with annual rainfallvarying between 1,000 and 2,500 mm which isseasonal to bi-modal.

Until recent times much of the dry zone wascovered in natural forest but heavy in-migra-tion has seen a substantial reduction in forestas pioneer farmers establish slash and burnagriculture (chena). By contrast the higher andwetter zones have an established smallholderagriculture contributing to the 16% of landunder homesteads. Total population is some365,000 giving population densities of some 65persons per sq.km, well below the nationalaverage. Ethnically the population is predom-inantly Singalese Buddhist with some 7%Tamil (Hindu) and Muslim. Literacy is high at

75% (MONDEP, 1997). The poverty of thearea is illustrated by the 62% of the districtpopulation in 1993 on means-tested welfare;the 60±65% of households with no access toclean water; the 40±50% of households withinadequate sanitation (1993 and 1995 di�erentsources); the poor nutritional status (1993)with 32% of infants with weight for age belowthe 3rd. centile and the 42% of 1±2 year oldswith weight for age below the 3rd. centile(MONDEP, various years and Bond et al.,1996). Although these statistics should beviewed with caution, a general picture ofresilient poverty some 10 years into theprogram emerges.

The program was based on a carefullythought out process approach that showedconsiderable innovation and an unusuallyhonest analysis of the di�culties of improvingliving standards in remote, rural areas withdi�cult agro-ecological conditions6. It had a20-year time frame with rolling planning andbuilt-in review mechanisms. A direct focus onpoverty was envisaged by targeting projects onmarginal farmers, encroachers, plantationworkers, unemployed youth and rural women.Participatory approaches were to be utilizedand, where possible, community-based organi-zations strengthened or established. Programsta� had to focus on planning projects as allimplementation was to be done by locally basedpublic, private or voluntary organizations.MONDEP supported such organizations indeveloping plans, provided them with ®nancialsupport to implement plans and providedassistance to them in training sta� and buildingorganizational capacity.

Precisely measuring the performance of sucha long-term program with broad develop-mental goals is infeasible, if not impossible7.Over the years, however an increasing volumeof evidence has been built up leading both theGovernment of Sri Lanka and NORADjudging MONDEP to be ``successful.'' Thisevidence (Bond, 1997) comes from a mid-termreview (Moore, Rasanayagam and Tilakar-atne, 1995), 22 reviews of projects within theprogram, the opinions of bene®ciaries donorsand o�cials and an independent reportprepared by a development journalist (Wani-gasundara, 1995). (See Table 1.) While noneof these sources proves that the project hasbeen successful, they corroborate each others®ndings and make a strong case that theproject has performed well over its ®rstdecade.

PROCESS APPROACHES TO DEVELOPMENT 1343

4. MONDEP: AN ANALYSIS OF PROCESSIN PRACTICE

This section examines MONDEP's experi-ence during 1984±95 in terms of the conceptualframework outlined in Section 2. WhileMONDEP's attempt to operationalize aprocess approach led to it recognizing the ®vemajor elements that ``process'' entailsЯexibleand phased implementation; learning fromexperience; bene®ciary participation; institu-tional support; and, program managementÐthe ``®t'' it achieved varied. The best ``®t'' withthe framework was in terms of bene®ciaryparticipation; the weakest ``®t'' was withsystematically learning from experience (Bond,1997).

(a) Flexibility and phasing of MONDEP

MONDEP has been able to operate ¯exibly,but at the same time have a clear plan of actionfor programming human, ®nancial and otherresources and co-ordinating with other agen-cies, because of its long time frame and itsrolling plan mechanism. At the outset it had a20-year time scale so that there was norequirement to plunge into a rigid plan thatpretended a ``solution'' to Moneragala's many

problems had been found and was beingrapidly implemented (as is the norm withdonor-®nanced programs). The program wasseen as having two 10-year phases. At the endof Phase I a mid-term review would make ajudgment on whether Phase II should proceed.It was also spelled out in the initial plans that amajor aspect of Phase II would be a systematicapproach to phase out, so that MONDEPwould disappear but bene®ts should continueto ¯ow.

The rolling plan is at the heart ofMONDEP's activity. The Project Directorpresents an annual program for the followingyear at the end of the ®rst quarter of the eachyear. This includes a review of progress so far, awork program and budget for the followingyear, plus sketches8 of any new projectsproposed. These sketches and budgets areapproved in principle by the donor and relevantministry, or removed from the plan afternegotiations. Thus MONDEP moves forward ayear at a time within the overall objectives ofthe program. The approved sketches can beplanned in detail and ®nally approved at anytime. Some sketches wait years to be plannedwhile others are dropped altogether if notdeemed appropriate. The project can imple-ment sketches that have received detailedapproval at any time. This is the basic rollingplanning framework which enables the processapproach. It does however create particularbudgeting and management problems.

For example, the rolling planning systemmeans that all stages of the project cycle will beactive simultaneously among many smallprojects. This complex situation requiresdi�erent accounting systems from moreconventional programs. In the latter, plans andbudgets are estimated for the (relatively short)lifetime of the project, and expenditure ismonitored as the commitment is steadilydischarged. For MONDEP, auditing systemsneed to be able to cope with the fact thatactivities are continually planned, imple-mented, reviewed and modi®ed throughout thelife of the project. Thus compatibility must beensured between capacity, available ®nance andduration of planned activities and to ensurethat spare ®nancial capacity is reserved forfuture plans. This point was missing in the ®rstphase of MONDEP. The enthusiasm withwhich new commitments and plans were made,in the absence of a de®nition of the acceptablelimit of commitments which could be made,caused many problems. In particular, after

Table 1. The performance of MONDEP in its ®rstdecade a

Ð``The governments of Sri Lanka and Norway may bejusti®ably proud of MONDEP. . . the evidence availableon the distribution and the e�ectiveness of MONDEPinvestments indicates that the project has had a positiveimpact in reducing poverty and increasing welfare in thedistrict'' (Moore, Rasanayagam and Tilakaratne, 1995).Ð``The programme has been able to bring direct bene®tsto large numbers of bene®ciaries in the form ofcommunity facilities . . . as well as household facilities. . .which provides new opportunities for local populationsand which assist the development and strengthening ofhuman resources in the district. . . Women and childrenhave been important and designated bene®ciaries ofseveral interventions,. . .'' (Sùrbù and Zackariya, 1995).Ð``In 1987 all of us were desperately poor (shiftingcultivation) farmers. . . we were dirt poor and we did nothave Rs 10 between us. We had no houses, no savings,no credit, no subsidies. In fact we did not have a plate ora cup. . . All of us have built houses now, and wecultivate our plots with highland crops and grow somerice collectively on ®elds we lease out from others''Quote from some social mobilizer group members(Wanigasundara, 1995).

a Source: Bond (1997) pp. 13±15.

1344 WORLD DEVELOPMENT

seven years of broadly balancing new plansand commitments with the implementation ofearlier plans, planning activity substantiallyovertook the capacity to implement leading to asnowballing of undischarged commitments.The value of annual in¯ation (around 12%)presumably acting upon these undischargedcommitments began to approach the value ofthe annual budget9 which itself was in step withdistrict implementation capacity. This impliedthat a point might soon be reached where theprogram ``had to run to stand still.'' Moredirectly and disregarding in¯ation e�ects, thevalue of undischarged commitments alreadymade meant that there would be very littleresidual ®nance during the second phase of theproject for the application of lessons learnedand new innovations. In other words, theProcess program would have become a Blue-print. The ``overcommitment'' issue manifesteditself in a number of tangible problems. Onewas general portfolio overload of the keyprogram sta� with broad responsibility forplanning, implementation management and themonitoring and reviewing in partnership with avariety of line agencies. This led to a reductionin the quality of work as urgent and routinetasks pushed aside more re¯ective and timeconsuming responsibilities of reviewing andplanning. Another problem created by the largenumber of active projects was the small share ofannual budget available to each. In some casesthe letting of contracts had to be deferred dueto inadequate funds; when savings becameavailable toward the end of the year it was thenprocedurally too late to let contracts or withthe onset of rainy season, conditions becameunsuitable for works. These are also partlyexamples of a mismatch between the proce-dures of conventional blueprint and the needsof process approaches.

Given these limitations, a long-term processprogram with a rolling planning system shouldideally undergo three phases of implementa-tion. These are based on the experience of theMONDEP program and broadly re¯ect whathappened or is planned to happen. Problemstended to occur where MONDEP deviatedfrom this model.

(i) An initial capacity build-upThe agencies of the district are introduced to

the approaches and objectives of the program,and their planning and implementation capac-ities are slowly built-up, through a series ofsmall experimental projects. This phase also

requires special projects to establish andsupport the functioning of the PMU and togive material support to the key agencies.Social mobilization should be introduced earlyinto areas of concentration of target groupswhere projects are active.

(ii) Sustainable process phaseThis is the main phase of the program and

should maintain a stable level of commitments,made up of projects of increasing size, whichcan be discharged in a reasonably short period.Special characteristics of subphases are givenbelow;

ÐEarly; This would correspond broadly toKorten's ``e�ciency'' phase where small experi-mental projects from the ®rst phase have beenreviewed and best options and features selectedfor new initiatives but which are still learning toimprove systems/optimum scale of operationsetc. It is also likely that many projects wouldhave an emphasis on provision of social infra-structure. While various partner agencies wouldbe responsible for projects the PMU would havea major role of innovation, coordination, andthe introduction of and training for improvedmethodologies.ÐLate; In some sectors this might correspondto Korten's ``expansion'' phase where successfulinnovations are repeated on a more extensivescale with considerable co-®nancing andincreasing transfer of responsibility and func-tions to permanent institutions. Other projectsmight continue to improve e�ciency or experi-ment with e�ectiveness where new problemshave been identi®ed, or more likely, where ashift of emphasis from provision of infrastruc-ture to Human Resources Development, e�ec-tive management of services, and thedevelopment of productive capacity occurs.During this phase a mid-term review would beparticularly valuable to assess whether theobjectives are being ful®lled and to re-adjustthe strategies to ensure that the project and planportfolios re¯ect the stage of the program.

(iii) The Phase-outIn this phase, no new initiatives are taken on

except for a few essential small-scale specializedprojects to ease the transfer of all programfunctions to permanent institutions or provide®nal consolidation to former investments. Awell-planned phase-out should allow for theremoval of the PMU and special donor supportwithout harming the continual development of

PROCESS APPROACHES TO DEVELOPMENT 1345

the area by (now improved) state, commercialand civil institutions.

In a 20-year program, phase 1 might takefour years, phase 2, twelve years and phase 3,another four years. The boundary betweenearly and late phase 2 would be di�use.

The rolling planning structures need to bere¯ected in the budgetary procedures, sincebudgets are ®nancial plans. This often meanseducating higher-level ®nancial managers(generally accountants rather than planners)about the objectives and consequences of roll-ing plans.

Much of the above was not clearly under-stood at the beginning of MONDEP and sothere was only a moderate ®t in terms ofexperimental approach and phasing. A strong®t may yet occur in terms of the ``phase-out,''however, as planning for this started nine yearsbefore project closure. The rolling planningstructures and long time frame provided amuch better ®t with process theory. Overall,¯exible phased implementation at MONDEPrated a ``good'' ®t with the analytical frame-work.

(b) Learning from experience at MONDEP

The ability to review experience and build inthe learning into future plans is an essentialpart of a process approach. Learning in aprocess program has to be more than just anacademic exercise; it must directly feed backinto the planning of future activity.

The MONDEP program is clearly justi®ed inclaiming that it has followed a learning processapproach: ``the project management hasdisplayed considerable ¯exibility and innova-tiveness, and changed programs, policies andstrategies in response to experience'' (Moore,Rasanayagam and Tilakaratne, 1995). Learn-ing is not simply about formal systems anddocumentation. It occurs in both informal andformal ways. Informal learning tends to bebuilt up among sta� members as their experi-ence with the project grows and can be afeature of organizational culture. This type oflearning therefore depends in part on the build-up of institutional memory through the conti-nuity of sta�. E�orts to encourage MONDEPsta� to remain in post for several years(through inspirational leadership, on the jobtraining and promises of overseas training)helped to keep sta� turnover rates at relativelylow levels (especially considering that for many

Sri Lankans Moneragala is seen as a punish-ment posting) and supported informal learningprocesses.

Informal learning is certainly not foolproof,nor su�cient, so formal learning needs also tobe incorporated. Small project review10 teams,comprising a mix of internal and externalpersonnel but including someone from each ofthe involved agencies, has been a key deviceused by MONDEP. These teams have usedparticipatory rural appraisal techniques, trian-gulated with a minimum of formal methods, tolearn quickly and cost-e�ectively at relativelyhigh levels of reliability.

MONDEP depended on informal learning inits early years with a stable sta� ensuring bothcontinuity and institutional memory. In lateryears formal reviews have become moreimportant as sta� continuity declined. Thisenabled a comparison of the two methods.Evidence was found for the necessity of formalreviews where ¯awed approaches were beingperpetuated; the formal review helped to revealthe weaknesses. Examples include;

Ða survey of a politically driven agro-well pro-gram where grant assistance was being abusedand going to inappropriate areas;Ða review of highland farming stabilizationprojects where soil conservation measures werenot e�ective and inappropriateÐa review of past education projects done intime to stop major investments in unnecessaryclassrooms and which exposed serious teachingquality concernsÐa review of primary health care projectswhich showed the impact of a PHC strategyand the waste of substantial investments in aparticular kind of health center building.

Evidence was also found for the need to involvesta� of implementing agencies in the reviewprocess in order to ensure the incorporation of®ndings into their policy and planning proces-ses. Sending them a copy of a review report haslittle e�ect! Table 2 provides an illustrationbased on three of the above reviews.

These cases imply that greater involvement ofpartner agencies in the review process hasbrought about increasing adoption and use ofthe lessons learned. But, without the initiativefrom the MONDEP Project O�ce, backed bybudgets and sta�, such reviews would not havetaken place at all. Partner agencies have learnedfrom reviews but ``double loop'' learning(Argyris and SchoÈn, 1978) has not been

1346 WORLD DEVELOPMENT

achieved, i.e. partners have not started toconduct reviews on their own initiative.

To ensure the adoption and application oflessons learned has required not only proce-dures but also the scheduling of learningactivities, so that lessons come at the right timeto be absorbed into the next planning stage ofMONDEP or the partner agency. It has notproved easy to ensure that the o�cer respon-sible for planning has access to and has read therelevant review reports. Even when lessonslearned have been read in time for planning,action had depended on the courage andimagination of the individuals who are requiredto make the, often radical, changes needed.Even when lessons learned were not contro-versial or challenging institutional inertiaproved a powerful force at MONDEP. Adeliberately fostered organizational culture thatopenly embraces error, to take risks and toreward imagination and innovation is needed,but this requires great skill as it entails humanresource management approaches based on

enabling, rather than controlling, professionalsta�. For example, at MONDEP the evolutionof the participatory planning model (outlinedin the next section) involved a year-long processof critical re¯ection, innovation and consensusamong project and other agency sta�. Althoughthe previous model was popular, well knownand had been used for eight years on nineprojects, the project o�ce was able to carrythrough this necessary process of change, butonly after great and skillful e�orts.

Learning from experience was only rated a``moderate'' ®t for MONDEP (Bond, 1997).The Korten model of achieving e�ectiveness,e�ciency and expansion was not formallyincorporated and the program had too manylarge interventions which did not undergoiterative improvements from small experiments.There was however, a good ®t with other ele-ments of process theory including self-criticalreviews, links between learning and planning,and the use of appropriate technologies (Bond,1997).

Table 2. Reviews and learning outcomes at MONDEP: Three examples a

Review Partner Agency Participationin the Review Team

Communication of Findingsto Partner Agency

Results observed

Rain-fed HighlandFarming StabilisationProjects (1994)

No partner agencyrepresentation on the team.

There was a pre-publicationcheck of the draft by the headof the partner agency. Therewas no presentation of thereview ®ndings to partneragency sta�. Copies of thereport were sent to seniorsta�.

Two years later, theweaknesses identi®ed inthe review were still beingpracticed by the partneragency under programfunded projects. Thereview had failed to assistthe agency's learning.

District EducationalServices Projects(1995)

Retiring o�cer of partneragency on the team.

Copies of the report were sentand a presentation of review®ndings were made to seniorsta� of the partner agency.

Findings incorporated inthe joint planning of thenext project phase and inprovincial policy formula-tion. Substantial learningoccurred that has led toimproved activity.

Primary Health CareProjects, (1996)

Two senior sta� of partneragency on the team asconsultant and ®eldworker.

Team member from thepartner agency distributedcopies of the report;presented ®ndings toagency sta�; noted points`for action'

Top Ministry o�cialappreciates the study, hassent copies to WHO andput it forward as a model.Provincial government in-formed and action initiatedon three important points.The review appears to havecontributed to policydevelopment.

a Source: Based on Bond (1997), p. 66.

PROCESS APPROACHES TO DEVELOPMENT 1347

(c) Bene®ciary participation at MONDEP

At its outset MONDEP identi®ed bene®ciaryparticipation as central to its approach.Participation was seen to have many potentialadvantages in terms of project outcomes,sustainability and the empowerment of disad-vantaged people. It took several years howeverfor MONDEP to develop a satisfactoryapproach to participation and di�cult issuesrequired careful analysis. In particular issues ofscale, the attitudes and behaviors of sta� andlogistics have demanded resolution. While ahigh degree of direct participation can beachieved at a local level, and it is possible toinclude poor and marginalized people's voicesand energies, at higher levels (division anddistrict) representation becomes necessary withthe likely loss of minority voices and the like-lihood of elite capture. Although most sta�welcomed the idea of participation the neces-sary changes in behavior and attitudes (listen-ing not talking, facilitating not doing, beingequal rather than being superior) did not comeautomatically but required, and continue torequire, actions that encourage critical self-analysis. Finally, thought had to be given to thecosts, as well as the bene®ts, of participation.Frequent local-level meetings with communitiesspread over a wide area with poor transportinfrastructure and virtually no telecommuni-cations led at times to inordinate amounts ofsta� time being allocated to traveling11 andadministering logistics. Developing cost-e�ec-tive and time-e�ective participatory approacheswas essential.

Initial e�orts at MONDEP focused oninvolving villagers in the planning process.These entailed village-level consultative discus-sions followed by large planning workshops,sometimes of several hundred people, engagingin a brief consultation with the program'scharismatic director, line agency sta� and localpoliticians. Fortunately, rapid learning tookplace and this approach became less consulta-tive and top-down. Over a decade it evolvedinto a fully participatory demand-led model forlocal area planning (on average, units of 600families). This model, drawing on the growingbody of knowledge on participatory ruralappraisal (PRA) and participatory learning andaction (PLA)12, takes longer than simple groupdiscussions but can be facilitated by mixedgroups of trained government employees andvolunteers from the local community, thusreducing demands on MONDEP sta� time and

improving e�ectiveness through the admixtureof professional and local knowledges. In addi-tion, this PRA-based model has fostered local-level ``ownership'' of initiatives and is a vehiclefor behavioral change in MONDEP and othergovernment sta� that has longer term implica-tions for organizational culture13. While highlevels of PRA-methodological ``purity'' havenot been attained, the primary objectives ofbroad-based participation and a degree ofanalysis and democratically-voiced preferencesfor joint community/state action have beenachieved. Some relevant indicators from aseries of 23 PRA village planning sessions wereaverage daily attendance of 72 (about a quarterof the adult population of the surroundingarea); of which 45% were men and 55% women.Typical outputs were two group discussions,one map, 15 thematic charts, male and femaleseasonal activity calendars, a village wealthranking, a transect walk, two Venn diagramsand a ranked list of community-identi®edinitiatives that require some external support.These are indirect indicators of the signi®cantdegrees of joint analysis and negotiated futureaction that has occurred.

While PRA approaches to planning havebeen added to the ``armory'' of techniques usedby MONDEP they have not become theexclusive approach. They are seen as havingrelevance at certain times and places. At othertimes MONDEP sta� use ``hard'' techniquesmore typical of engineering projects (such ascritical path analysis), when complex sequencesof activities need coordinating.

A second major focus on bene®ciary partic-ipation at MONDEP has been through thesocial mobilization activities. The initialapproach as set out in the pre-feasibility study(Whist et al., 1982) was to rely primarily on theexisting people's organizations but to considerthe use of change agents trained by the Minis-try of Rural Development. The ®rst majorintegrated project under MONDEP made noprovision for social mobilization or changeagents but relied on the existing rural devel-opment and other societies for implementationof several components. Early experience wasthat existing peoples' organizations were not agood vehicle for social mobilization or distri-bution of bene®ts14. Meanwhile the introduc-tion of social mobilizers15 into other similarprojects 2.5±3 years later had the additionalobjective of improving coordination andmanagement of project activities. Thusalthough the training of Social Mobilizers was

1348 WORLD DEVELOPMENT

as change agents, their role was partly as``project agents'' from the start. By 1995 therewere 174 active Social Mobilizers, half male,half female, they were running 990 groups withover 5,500 members, 54% male and 46%female. Furthermore these groups, fromamongst the poorest in poor areas, had groupsavings funds totaling some US$ 40,000 anotable achievement which later led to theirincorporation into formal thrift and creditsocieties. The successful evolution of socialmobilization thus provides an example oflearning process in action.

Overall, bene®ciary participation inMONDEP had a good ®t with theory especiallyin terms of ensuring that action is based on anunderstanding of the social context. There wasa good ®t also with participation in problemanalysis and planning as well as with empow-erment of bene®ciary groups. Participation inimplementation activities was variablehowever, and at the monitoring stage was onlymoderate.

(d) Institutional support at MONDEP

A complex network of MONDEP seniorsta�, sta� in the Ministry of Policy, Planningand Implementation and NORAD sta� andconsultants have managed to ensure thatMONDEP has received the support it neededto function e�ectively. Donor support has beencrucial in both quantity, providing 95% ofMONDEP costs over a 20-year period, andquality: NORAD's enlightened ``recipientresponsibility'' philosophy means that it hastaken a back seat and helped build up SriLankan institutions rather than create depen-dency. While NORAD has remained commit-ted to MONDEP it too has needed to managepolitical relationships, particularly whenNorway's sophisticated electorate challengedits leadership on development assistance issues.For the Norwegian Embassy in Colombo thistranslated into requests for clear evidence ofMONDEP's achievements.

Managing relationships with national andlocal political leaders and parties has been apriority task for MONDEP's directors. Inevi-tably, this time-consuming work is largelycovert and there is little information availableon exactly how domestic politics has beenhandled. The steady progress of the program,and the lack of any program ``meltdown''situations, indicates the relative e�ectiveness ofMONDEP's directors in a di�cult environ-

ment. The Sri Lankan bureaucracy has beenincreasingly politicized along party lines inrecent times with great pressures on the publicsector to respond to the partisan allocativewishes of politicians rather than stated publicpolicies (Hulme and Sanderatne, 1996); the JVPinsurgents controlled much village level activityduring 1988±1990 while a heavy militaryresponse (and the operation of death squads)made ®eld operations di�cult; and, in the early1990s, the personal interests of the Minister ofAgriculture (a Moneragala District MP) led tomany pressures being brought on MONDEPand its sta�. It would be fascinating to knowmore about the ways in which such problemswere handled. It is nonetheless clear thatpolitical analysis and personal political net-works were essential to MONDEP continuingto function well.

The decision of the MPPI to take a decen-tralized approach to the country's IRDPsÐandgive them considerable autonomyÐcontributedto MONDEP's e�ectiveness. This meant thatmost operational decisions could be madelocally, raising the likelihood of their beingboth relevant and rapid. While MONDEP itselfis a ``temporary'' institution (albeit with a 20-year lifespan) its ``no direct implementation''policy has led to a focus on strengtheningpermanent institutions in the District. ItsDivisional-Level Planning Support Project hasfocused on raising the planning capacitieswithin Moneragala's divisional secretariats sothat program phase-out should not lead to aplanning ``vacuum'' in the District. Unfortu-nately, these secretariats are basically decon-centrated arms of central government and thismay ultimately limit their ability to prepareplans relevant to local needs.

In terms of institutional support the programhad a very good ®t with Process theory in termsof devolved authority, the involvement ofpermanent institutions and building localcapacities. There was only a moderate ®t withother elements including winning the supportof local politicians and using existing organi-zations neither of which had the needs of thepoorest at heart. The initiation of internalorganizational change among local agenciesrated a poor ®t and was considered beyond theremit of the program (Bond, 1997).

(e) Program management at MONDEP

Managing Process projects and programsis about removing the ``hiding hand'' of

PROCESS APPROACHES TO DEVELOPMENT 1349

Hirschman (1967) so that the unknown andunpredictable is acknowledged and faced up toin a responsible and manageable way. This, ofcourse, is done through all the aspects ofProcess described above but what does it meanin terms of day-to-day management and thechoice and organization of management system?

First, concerning the locus and responsibilityof the PMU, MONDEP was set up so as toavoid bypass of local institutions while stillretaining a distinct identity and the capacity toact as a catalyst within the district. This wasachieved largely through special funds and thepolitical and administrative support discussedabove and the teambuilding and inspirationalleadership discussed below, but also throughthe sensitive choice of an institutional base.There was an existing district planning unitunder the District Secretary which was absor-bed into the Project O�ce. Although thissubverted other non-IRDP planning for thedistrict, a small planning unit was later rein-stated and under Phase 2 is being supported forother central government and provincialgovernment initiatives. This arrangement wassuccessful to the extent that Project O�cebecame the major source of development andinstitution-building funds, the only institutionthrough which an integrated or co-ordinatedapproach to the solution of development prob-lems could take place and the source of muchinnovation in participatory and learningprocess methodology. It was also able torespond to the administrative decentralizationby decentralizing experienced planning sta�,providing direct support and initiating a newapproach to local area planning backed bydonor funds.

The PMU then had a major task in coordi-nating planning, implementation and monitor-ing activity as few components were directlyimplemented by the project. Within the zonesof control, in¯uence and appreciation (Smith,Lethem and Thoolen, 1981), process projectshave most activity taking place within the zoneof in¯uence. In the case of MONDEP thismostly took the form of contractual partner-ships with the line agencies involved in plan-ning or implementing projects. Other agenciesand individuals important to the outcome ofprogram interventions also fall within the zoneof in¯uence imparting a special character to therole of project management, di�erent fromorthodox blueprint approaches.

In terms of personnel there is no doubt thatthe quality of program leadership has played a

major part, particularly in the critical earlyyears, in maintaining the special role andcharacter of the program. As a consequence ofthis and the support to local and overseastraining there was an unusual degree of sta�stability and continuity within the ®rst eightyears, itself of enormous bene®t to the opera-tion of the program. This situation changedafter the decentralization of planning sta�mentioned above. The following years werecharacterized by a higher sta� turnover anddecline in the number of sta�. The need formore structured information systems becameclear at this point.

The use of Technical Assistance (TA) ®ttedwell with Process theory. In comparison withdonor policy and use of TA in other IRDPswhere TA personnel were often used by donorsto maintain a level of direct control, seniorministry o�cials often voiced the opinion thatNORAD was good to work with. Not only didthe donor not interfere with local decision-making but also ensured that long-term TAwere kept to a minimum16. The TA ``senioradvisors'' were given a strictly nonexecutivebrief and although they were quite in¯uentialon the program their in¯uence was achievedthrough advice, sta� training and back-roomwork for the Project Director. The early real-ization that a sta� predominantly composed ofsociologists, economists and geographers weresusceptible to technical naivety17 in a programwhere there were substantial infrastructurecomponents led to the early appointment of anengineer under local technical assistance.Similarly other short-term needs were metlocally with very little di�culty, facilitated by¯exible donor policy and a supportive ministry.

Chambers (1993) ethos of new professional-ism ®ts well with Process theory but is indi-vidually challenging within bureaucracies in ahierarchical culture. There were, however, somerecognizable aspects of new professionalismemerging within the MONDEP experience. Theteam spirit established early on within ProjectO�ce fostered an ethos of service where o�cerswere willing to serve for long periods in aremote, poorly serviced area18. Their predomi-nantly social science backgrounds mentionedabove meant that they had a perceptiveappreciation of socially constructed multiplerealities and an interest in ``who gets what.''They became expert in the use and training ofPRA so Chambers (1995) ``benign virus'' tooke�ect amongst project sta� and within lineagencies. Many projects are innovative and

1350 WORLD DEVELOPMENT

thus challenge personnel to consider approachesand techniques other than the conventional.Finally, personnel learn to be self-criticalthrough participation in reviews of theirprojects (or learn to accept the criticism fromsuch reviews, to change and improve workingmethods and approaches). On the other hand,there are many deeply entrenched aspects of``normal professionalism'' which showed nosigns of change, so the ®t of this aspect must beconsidered only partial.

The ®nal area where operational lessons werelearned from MONDEP was that of project®nance. The need to manage strategically thelevel of ®nancial commitments has already beenmentioned; coping with the additionalcomplexities of the annual budgetary process isanother. Since annual budgets in a processproject are no longer just adjustments to theoriginal ®nancial plan according to pastperformance, then a number of ``unknown''elements have to be included to ensure smoothrunning of the project. Typically, for a ``nextyear's budget'' these might include the follow-ing in decreasing order of certainty;

ÐOngoing activities carrying-over from theprevious year.ÐNewly approved activities starting during theyear.ÐNewly (in principle) approved sketches cur-rently undergoing detailed planning which areexpected to be approved during the year.ÐOther new ideas/phases which will have asketch presented in the coming year for ``in-principle'' approval and which will have a shortenough planning period to be approved andstarted during the year.ÐOn the negative side, there may also be ap-proved sketches which may be abandoned/com-bined with other sketches or just not approvedin their ®nal form.

The ®rst requirement for the above is authorityto include ``as-yet non-existent items'' in aformal budget, sure to cause the most accom-modating accountants and auditors to balk! Itwill also be appreciated that when the aboveare combined with the general uncertainties ofscheduling, especially for interventions with ahigh degree of participation and dependencyon partner agencies, the planning of ®nancial¯ows becomes an inexact procedure. At thesame time a program involving a large numberof inexperienced agencies cannot a�ord toabandon tight ®nancial management and

accountability. From the MONDEP experi-ence, an imprest account system with many``short ®nancial leashes'' combined with a greatdeal of ¯exibility in the annual budget andauthority to over-budget seemed to be anoptimum system.

Overall the internal management aspects ofthe program rated a moderate to good ®twith the theory, strongest ®t being in terms ofleadership and use of technical assistance andthe ¯exibility of the PMU. The weakest ®tconcerned the loss of sta� stability after the®rst eight years and the slow uptake of theconcepts of ``new professionalism'' in the vari-ous line agencies that MONDEP was part-nering.

5. PROCESS AND POWER

Much of the argument for a processapproach is based on the dictum, ``...it ain'twhat you do it's the way that you do it.'' TheMONDEP experience shows the need to extendthis, ``...it ain't only the way you do it, it's alsowho you do it with.'' A serious attempt at aprocess approach to intervention does notmerely incorporate bene®ciary participation;rather, it entails a fundamental recon®gurationof the involvement of stakeholders in programobjective-setting, design, implementation andmonitoring. This means a redistribution ofpower and in¯uence over decision-making. Theway that this has worked out in Moneragalaare revealed by comparing an ex post stake-holder analysis for MONDEP (Figure 2) withan ex post stakeholder analysis of the MahaweliDevelopment Program Downstream Develop-ments (Figure 3). This analysis follows themethodology of Montgomery (1995) wherestakeholders are grouped into ``primary''(target bene®ciaries), ``secondary'' (intermedi-aries in the aid delivery process), and ``exter-nal'' (other interested parties who may gain orlose from the intervention). The diagramsclassify the stakeholders according to ``impor-tance'' to the objectives of the projectand ``power'' over the outcome. These dimen-sions are plotted against each other in asubjective way giving four quadrants A±D. Thesigni®cance of the quadrants is in terms oftreatment; Those in ``A'' require special atten-tion, those in ``B'' need harmonious relations,those in ``C'' require good PR and carefulmonitoring and those in ``D'' can be safelyignored.

PROCESS APPROACHES TO DEVELOPMENT 1351

The Mahaweli Development Programmedownstream development is, in essence, ablueprint IRDP that was initiated downstream

of Victoria Dam and hydro-electric scheme inthe same part of the country. The main changesthat this comparison reveals are summarized inTable 3. These include a reduction in the powerand in¯uence of the donor agency; the greaterautonomy of the ®eld-level project manage-ment unit vis �a vis its parent ministry; anenhanced role for the local administration andan emphasis on working with national lineagencies at the local, rather than the head-quarters level; an increased role and access toresources for local NGOs and CBOs; theopening up of a space for the involvement oflocal politicians (albeit with careful oversight);and, a signi®cant increase in the power andin¯uence of intended bene®ciaries about whatto do and how to do it.

The basic thrust behind all of these moves isone of decentralization.19 The donor andrecipient agency (Ministry of Policy, Planningand Implementation) oversee a small numberof key functionsÐbudgets, the approval ofannual plans, senior appointments and exter-nal evaluationÐwith the vast majority ofprogram functions being delegated to theproject management unit based at the kachc-heri (the district o�ces) in Moneragala. Therole of MONDEP's project management unitwas structured, however, so that it is notplaced in a monopolistic situation where it hasthe bulk of power. As it is required to imple-ment all activities with partners, and not

# Stakeholder h sConventional

projectProcessproject

Quadrantlocated

Primary Stakeholders12 Primary (target group)

stakeholdersA B

SecondaryStakeholders

1 Donor Agency B A2 Recipient (National Line

Agency)B C

3 RegionalAdministration

D D

4 Local Administration D B5 PMU B B6 National Line

AgenciesB D

7 Local Line Agencies A A8 Local NGOs D A9 Local Contractors A A

10 National Political Leaders B B11 Local Political

LeadersC B

13 Non-target groupstakeholders

D D

Figure 3. A stakeholder analysis of the Mahaweli Devel-opment Programme downstream developments. Source:Bond's judgement based on a 15-year knowledge of theprogram (and six years as an in-country employee of the

program).

Figure 2. A stakeholder analysis of MONDEP. Source:Bond's judgement based on a six-year knowledge of theProgramme (and four-years as an in-country employee of

the programme), for key to numbers see text table.

1352 WORLD DEVELOPMENT

directly, it has to focus on building relation-ships and negotiating plans with the localadministration, the local level o�ces ofnational line agencies, NGOs and businesses.Finally, the adoption of a number of partici-patory techniques (and a gradual di�usion ofrelated attitudes and behaviors by MONDEPsta� and others) opened up increasing oppor-tunities for intended bene®ciaries to shape anddetermine what happens and how. This chainof delegations was by no means automatic, ithad to be fought for at all stages, neither did itrepresent an abrogation of responsibility: ateach link in the chain the ``higher'' leveloversees what is going on below, but this is interms of the main parameters of activity anddoes not involve attempts at micro-manage-ment. In essence, this is a means of dealingwith the complexity of rural development bycomplexity: not, attempting to produce asimple chain of command management modelwhich is simple in terms of both structure andintellectual content! Such decentralization ismuch more than a technical matter and it

required the political will of the donor and ofsenior Sri Lankan o�cials to achieve the activedelegation of responsibility and authority,rather than the more usual rhetoric andretention of the status quo. MONDEP wasfortunate in this and, while it cannot beguaranteed as something that will happen inall projects, the MONDEP experience doesprovide a model for donors and nationalgovernments: decentralization starts at the topor not at all.

6. CONCLUSIONS

The MONDEP experience provides evi-dence that process approaches to developmentprojects can be e�ective for relatively large-scale initiatives by the public sector. This isnot to argue that process approaches arealways optimal. Elsewhere (Bond, 1998a) wehave discussed the multiple ways in whichprocess and blueprint approaches can be

Table 3. A comparison of changes in stakeholder roles in MONDEP and the Mahaweli Development Programmedownstream a

Change from Blueprint to Process Comments

Reduction in power and in¯uenceof donor agency

The donor policy of recipient responsibility in MONDEP ®tted well withprocess theory and allowed control of the development process with notable¯exibility by the recipient. In a blueprint the donor retains in¯uence throughthe agreed plan (which allows for little ¯exibility) and the reimbursementmechanism which acts as a sanction.

Reduction in power and in¯uenceof the recipient agency.

The relative autonomy of MONDEP's project management unit from therecipient agency, over planning and methodology compared to the highlycentralized structures of the Mahaweli Program.

Increase in power and importanceof the local administration

The Mahaweli Program bypassed the local administration. MONDEP used adecentralized strategy to build up the local administration and to hand overresponsibility

Decrease in power and importanceof national line agencies

National line agencies had a key role in the Mahaweli program butMONDEP worked with their local o�ces (under a devolved administration)rather than linkages with the HQ.

Increase in the importance of localNGOs

NGOs were not encouraged in the Mahaweli program. They were indirectlysupported under MONDEP.

Increased recognition of the impor-tance of local political leadership

In the Mahaweli program local political leaders were subservient to nationalleadership and had no say in running the program beyond low-levelappointments. In MONDEP they are involved in the planning process onmany projects and are kept informed of program objectives. They also seek(and sometimes succeed) in in¯uencing investments and contracts

Increase in power and in¯uence oftarget group bene®ciaries

Through a high degree of participation in many of the MONDEP projectsbene®ciaries have a considerable in¯uence on investments. They are moreimportant because most investments depend on their co-operation andparticipation. In Mahaweli the settlers were seen as an input in infrastructuraldevelopments and had little voice.

a Source: Bond (1998a).

PROCESS APPROACHES TO DEVELOPMENT 1353

sequenced and related.20 The approach bestsuited to any particular circumstance isdependent on the objectives of the interven-tion and the speci®c context. Unfortunately,most national and international developmentagencies assume that there is one approach(their existing policy) which is best and omitthe essential ®rst stage of the project cycleÐ``what type of intervention approach is bestsuited to this type of issue in this context?.''21

It is to argue, however, that Developmentagencies should however, think through howprocess approaches ®t into their interventionstrategies.

The learning process approaches that char-acterized the 1980s have not continued in the1990s.22 Instead, the mantra of ``participa-tion''23 has dominated debates about how tomake projects and programs more e�ectivewithout a clear analysis of how participationrelates to other critical elements of interven-tionЯexibility and phasing, systematic learn-ing from experience, the development ofinstitutional supports and program manage-ment. The conceptual framework outlined inthe ®rst part of this paper provides a means ofplacing participation in perspective: whilebene®ciary participation is essential for e�ectivedevelopment interventions it is only one ele-ment of a systematic approach that builds onempirical experience. This point is of particularrelevance when the World Bank (1997) hasrelaunched its e�orts to promote rural devel-opment. The Bank's talk of ``...address[ing] oldissues in new ways...'' through, ``revitaliz[ing]rural development at local and communitylevels...involv[ing] stakeholders...deliver[ing]rural ®nancial services...involv[ing] the privatesector...[and] promot[ing] sustainable resourceuse through community-based management''(World Bank, 1997, p. 18) is no substitute forits need to re-engineer its processes for inter-vention. Ad hoc lists of ``new ways'' and ``les-sons of experience'' from internationaldevelopment agencies and national lead agen-cies will not contribute to the strengthening ofthe conceptual framework that underpinsdevelopment action.

In particular, it is hard to see how acommitment to decentralization can be fosteredin donor agencies that insist on short time-frames, resist ¯exibility because of the problemsit causes them with monitoring, insist thatinterventions are assessed in terms of theirability to achieve outputs rapidly and punishpartners for their ``errors'' rather than scruti-nizing how e�ectively they are learning. As wepointed out earlier, decentralization starts atthe top or not at all.

The process approaches that will be called forin the future will be management intensive.24

The move toward ¯exible, phased, knowledge-creating, participatory interventions may meanless time spent on rigid planning, schedulingand monitoring but it does not mean that thereis less to do. Negotiating changes in plans,monitoring resource use in such circumstances,ensuring appropriate techniques and behaviorsfor participation, accommodating the di�erentpreferences of di�erent stakeholders, managingaction learning cycles and reporting place apremium on recruiting, motivation and retain-ing high-quality sta�.

Our ®nal point relates to learning anddevelopment action. While the literature ondevelopment is replete with lists of ``lessons''and examples of ``learning from experience'' thetruth is that ``not learning from experience''characterizes the knowledge-creating dynamicsof the development endeavor. This need not bethe case. At the present time we have anopportunity to link the conceptual ideas of the1980s with the empirical experiences of theprocess projects that they stimulated in the late1980s and 1990s. If we seize this opportunity wecan push forward the understanding of devel-opment interventions and potentially makethem more e�ective. The alternative is tobumble along talking of participation, NGOs,stakeholders, decentralization and, mostrecently social capital without systematicallyoverhauling our conceptualization of develop-ment projects and programs. We shall not be somuch on Hirschman's ``long voyage of discov-ery'' as endlessly going round in circles on aboating pond!

NOTES

1. For a fuller discussion of disillusionment with

conventional project approaches see Hulme (1995).

2. Even World Bank sta� have pushed the idea

forward (Picciotto and Weaving, 1994), although in

1354 WORLD DEVELOPMENT

contemporary practice the Bank remains wedded to

conventional approaches to projects.

3. One possible reason for this is the withdrawal of

support for North American scholar-practitioners work-

ing in this ®eld by USAID in the 1990s.

4. The UK's Department for International Develop-

ment would ®t into this category.

5. For a detailed account of MONDEP see Bond

(1997).

6. When one of us asked the First Secretary for

Development Cooperation at the Norwegian Agency

why they had a 20-year time-frame, compared to World

Bank and IFAD time-frames of ®ve years the response

was: ``Does anyone seriously claim that they know how

to reduce poverty in remote, semi-arid areas in a few

years? The ®rst years of this project are trying to ®nd out

what the development opportunities are: we don't have

any answers yet'' (Hulme ®eld notes, 1987).

7. Sri Lanka has not had a population census for 16

years and there is wide agreement that pre-program

district-level aggregate indicators of economic and social

development are unreliable.)

8. A ``project sketch'' is an outline plan, generally

covering 2±3 pages but containing only brief details of

the problem, proposed bene®ciaries, objectives, planned

outputs, costs and justi®cation. These details were

considered adequate by the parent ministry and donor

for ``in principle'' agreement in order to allow detailed

planning to proceed, thus avoiding wasted resources on

unsuccessful detailed proposals and indirectly encour-

aging an innovative approach to planning.

9. There was also an appreciation of the rupee against

the donor currency of some 5% but whereas this

appreciation might increase the annual rupee budget

(and the foreign currency reserved for the program), the

e�ect of in¯ation on the whole commitment led to a

lengthening of implementation times. This plan was

modeled on computer to de®ne the sustainable level of

commitment that could be held without lengthening

implementation times (Bond, 1997).

10. Within MONDEP, reviews ®lled a perceived gap

between the monitoring of processes and outputs, and

evaluation which was seen as a future formal assess-

ment of the achievement of objectives. From the start

of the program budgetary provision was made for

``Studies and Reviews'' and the donors repeatedly

requested their commencement. After periods of no

reviews, and then donor initiated reviews, eventually

the program management started to initiate their own

reviews.

11. When one of us ®rst visited the program in 1987

he was amazed at the capacity of MONDEP's assistant

directors to cope with travelingÐcommonly, four or

®ve hours a day thundering along rough roads and, not

infrequently, a meeting in Colombo would require 14 to

18 hours on the road for the round trip.

12. For review on the origins, concepts and practice of

PRA and PLA see Chambers (1994a±c).

13. A former project director undertaking research for

his PhD found that line agency sta� not only valued

these approaches but were beginning to use them in their

own projects.

14. This was con®rmed by the ultimate failure of the

``Revolving Fund for Rural Credit'' project which had a

promising start (Hulme, 1988) and was based on

peoples' organizations for disbursement, supervision

and recovery, one of the objectives of which was to

promote peoples' organizations through thrift and credit

supply. One of the causes of failure was the use of

peoples' organizations which were not representative

and not capable of managing credit.

15. Social mobilizers were largely educated young

people from the poorest communities in the district,

given three sessions of 3-day training interspersed with

community action, then maintained on a very small

allowance with (eventually) a bicycle and support from

monthly review sessions. They were trained and paid by

the Project O�ce, but administered under the divisional

administration. Social mobilization programs have

been introduced into all integrated projects under

MONDEP.

16. For most of the life of the program there was only

one senior advisor at the ®eld level. For a short period

in the early years there were two and one foreign

volunteer.

17. The ``plurality'' concept of Chambers (1983) may

be di�cult to achieve individually, but there is no more

excuse for a team of ``political economists'' than for one

of ``physical ecologists.''

18. This may now have changed following transfers

of two Directors and loss of senior sta� (personal

communication, B. Baklien, NIBR, Oslo). Of all the

PROCESS APPROACHES TO DEVELOPMENT 1355

requirements for a process approach, perhaps team spirit

is the most impermanent and fragile.

19. This is decentralization in practice rather than the

much more common decentralization in rhetoric, Turner

and Hulme (1997), pp. 151±174).

20. These include blueprint then process, process then

blueprint, blueprint-process continuum, blueprint in

process and process in blueprint (Bond, 1998a, pp. 13±14).

21. For a discussion of the ``menu'' of di�erent appro-

aches to development interventions see Hulme (1995).

22. With the possible exception of the Process Moni-

toring and Documentation (PMD) group of approaches

(Mosse, Farrington and Rew, 1998). Whether PMD will

ever be operationally acceptable however is open to

question (Bond, 1998b).

23. The popular participation and community partic-

ipation of the 1980s has now been replaced with

participatory rural appraisal (PRA) and participatory

learning and action in which people are being trained by

the tens (if not hundreds) of thousands. These can be

excellent tools and ideas but not the panacea that is

pretended.

24. There is an irony here. It was the intention of David

Korten in his contribution to the 1980s debate to bring

his experience of business management, where the

concern for managing against results through rapid

learning and feedback, to development bureaucracies (D.

Korten, personal communication, 1998). That debate

has been interpreted by many of the new development

professionals, anthropologists and sociologists, as free-

ing-up systems from the rigid con®nes of management

techniques, the ``management by abandonment'' warned

against by Honadle and Van Sant (1985), p. 92).

REFERENCES

Acko�, R. L. (1984) On the nature of development &planning. In People Centered Development: Contri-butions Toward Theory & Planning Frameworks, ed.D.C. Korten and R. Klauss. Kumarian, WestHarford, CT.

Argyris, C. and Sch�on, D. (1978) Organisational Learn-ing: A Theory of Action Perspective. Addison-Wes-ley, Bostan, MA.

Birgeg�ard, L. E. (1991) A process approach to policyand project analysis. Issue Paper # 13, SwedishUniversity of Agricultural Sciences, Uppsala.

Bond, R. Dayananda, D. P., Jinadasa, K. A. D.,Karandagoda, K. K. W. and Ratnayake, U. I.(1996) A review of primary health care projectsunder MONDEP. Project O�ce, Moneragala Minis-try of Plan, Implementation and ParliamentaryA�airs, Colombo, Sri Lanka.

Bond, R. (1997) Operationalising process: The experi-ence of the ®rst decade of the Moneragala IRDP inSri Lanka. IDPM Discussion Paper Series 50,IDPM, Manchester.

Bond, R. (1998a) Lessons for the large-scale applicationof process approaches from Sri Lanka. SustainableAgriculture & Rural Livelihoods Programme Gate-keeper Series # 75, IIED, London.

Bond, R. (1998b) Reviews in process projects: The thirdleg of learning. Paper Presented at Monitoring andEvaluation Mini Conference Overseas DevelopmentGroup, University of East Anglia, August 1998.

Brinkerho�, D. W. (1992) Looking out, looking in,looking ahead: Guidelines for managing develop-ment programmes. International Review of Adminis-trative Sciences, 58 (4).

Brinkerho�, D. W. and Klauss, R. (1985) Managerialroles for social development management. PublicAdministration and Development 5.

Brinkerho�, D. W. and Ingle, M. D. (1989) Integratingblueprint and process: A structured ¯exibility ap-proach to development management. Public Admin-istration & Development 9.

Cernea, M. M. ed. (1985) Putting People First: Socio-logical Variables in Rural Development. World Bank /Oxford University Press, New York.

Cernea, M. (1991) Putting People First: SociologicalVariables in Rural Development. 2nd edn., OxfordUnversity Press, New York.

Chambers, R. (1983) Rural Development: Putting theLast First. Longman, Harlow.

Chambers, R. (1988) Normal professionalism and theearly project process: Problems and solutions. IDSDiscussion Paper #247, IDS, Brighton, UK.

Chambers, R. (1993) Challenging the Professions. ITPublications, London.

Chambers, R. (1994a) The origins and practice ofparticipatory rural appraisal. World Development22 (7).

Chambers, R. (1994b) Participatory rural appraisal(PRA): Analysis of experience. World Development22 (9).

Chambers, R. (1994c) Participatory rural appraisal(PRA): Challenges, potentials and paradigm. WorldDevelopment 22 (10).

Chambers, R. (1995) Making the best of going to scale.PLA Notes 24.

Chambers, R. (1997) Whose Reality Counts? Putting theFirst Last. IT Publications, London.

Conyers, D. and Kaul, M. (1990) Strategic issues indevelopment management: Learning from successfulexperience. Part I. Public Administration & Develop-ment 10.

Dale, R. (1988) A donor perspective on the Hambantotadistrict IRDP ± with emphasis on the promotion of

1356 WORLD DEVELOPMENT

institutional capacity. Manchester Papers on Devel-opment IV/1, University of Manchester, Manchester,UK.

Edwards, M. (forthcoming) Future Positive. Earthscan.Eyben, R. (1991) The process approach. Report of the

natural resources advisor's Conference; ODA, Lon-don.

Friedmann, J. (1984) Planning as social learning. InPeople-Centred Development: Contributions TowardsTheory and Planning Frameworks, ed. D. C. Kortenand R. Klauss. Kumarian, West Hartford, CT.

Friedmann, J. (1992) Empowerment: The Politics ofAlternative Development. Blackwell, Oxford.

Goldman, I., Mellors, R. and Pudsey, D. (1989) Facil-itating sustainable rural development: An experiencefrom Zambia. Journal of International Development1/2.

Gow, D. D. and Morss, E. R. (1988) The notorious nine:Critical problems in project implementation. WorldDevelopment 16 (12).

Hirschman, A. (1967) Development Projects Observed.The Brookings Institution, Washington, DC.

Honadle, G. and Van Sant, J. (1985) Implementation forSustainability: Lessons From Integrated Rural Devel-opmen. Kumarian, Westport, CT.

Hulme, D. (1988) Revolving Fund for Rural Credit;Moneragala IRDP. IDPM, Manchester University,Manchester.

Hulme, D. (1989) Learning and not learning fromexperience in rural project planning. Public Admin-istration and Development 9 (1).

Hulme, D. (1995) Projects, politics and professionals:Alternative approaches for project identi®cation andproject planning, Agricultural Systems 47.

Hulme, D. and Sanderatne, N. (1996) Public account-ability, Public expenditure, management and gover-nance in Sri Lanka, 1948±1993. Institute of PolicyStudies, Colombo, Sri Lanka.

Ickis, J. C. (1983) Structural Responses to New RuralDevelopment Strategies. In Bureaucracy and thePoor, ed. D. Korten and F. B. Alfonso. Kumarian,Westport, CT.

Johnston, B. and Clark, W. (1982) Redesigning RuralDevelopment, Johns Hopkins Press, Baltimore.

Korten, D. C. (1980) Community organization andrural development: A learning process approach.Public Administration Review (September/October).

Korten, D. C. (1988) From Bureaucratic to StrategicOrganisation. In Transforming a Bureaucracy, ed. F.F. Korten and R. Y. Siy. Kumarian, Westport, CT.

Korten, D. C. and Alfonso, F. B. (1983) Bureaucracyand the Poor: Closing the Gap. Kumarian, Westport,CT.

Korten, F. F. and Siy, R. Y. (1988) Transforming aBureaucracy. Kumarian, Westport, CT.

Lewin, K. (1946) Action research and minority prob-lems. Journal of Social Issues 2.

Michael, D.N. (1973) On learning to plan ± andplanning to learn. In Redesigning Rural Develop-ment, ed. B. Johnston and W. Clark. Jossey BassPublishers, San Francisco.

MONDEP (1985±97). Annual Programmes 1985±1997.UPC. Project O�ce Moneragala, Colombo.

Montgomery, R. (1995) Short guidance notes on how todo stakeholder analysis of aid projects and pro-grammes. Centre for Development Studies, Univer-sity of Swansea, Swansea.

Moore, M., Rasanayagam, Y. and Tilakaratne, K. W.(1995). Moneragala Integrated Development Pro-gramme Mid-Term Programme Review. Report toNORAD and Ministry of Finance, Planning, EthnicA�airs & National Integration, Colombo, Sri Lanka.

Morss, E., Hatch, J., Mickelwait, D. and Sweet, C. eds.(1976) Strategies for Small Farmer Development: AnEmpirical Study of Rural Development Projects.Westview Press, Boulder, CO.

Moser, C. O. N. (1998) The asset vulnerability frame-work: Reassessing urban poverty reduction strate-gies. World Development 26 (1).

Mosse, D. (1995) ``People's Knowledge'' in projectplanning: The limits and social conditions of partic-ipation in planning agricultural development. ODIAgricultural Research & Extension Network, Paper#58, ODI, London.

Mosse, D., Farrington, J. and Rew, A. eds. (1998)Development as a Process: Concepts and Methods forWorking Complexity. Routledge, London.

ODA (1995) A Guide to Social Analysis for Projects inDeveloping Countries. HMSO, London.

Picciotto, R. and Weaving, R. A. (1994) New projectcycle for the World Bank? Finance & Development.

Pretty, J. N. and Chambers, R. (1993) Towards alearning paradigm: New professionalism and institu-tions for agriculture. IDS Discussion Paper #334,Institute of Development Studies, Brighton.

Roe, E. M. (1993) Public service, rural development andcareers in public management: A case study ofexpatriate advising and African land reform. WorldDevelopment 21 (3).

Rondinelli, D. A. (1993) Development Projects as PolicyExperiments: An Adaptive Approach to DevelopmentAdministration. 2nd edn., Routledge, London.

Rondinell, D. A. (1983) Development Projects as PolicyExperiments: An Adaptive Approach to DevelopmentAdministration. Methuen, London.

Sùrbù, G. and Zackariya, F. (1995) Moneragala IRDP:Review of Socio-Cultural Aspects. Chr. MichelsenInstitute, Bergen.

Smith, W. E, Lethem, F. J. and Thoolen, B. A. (1981)The design of organisations for rural developmentprojects: A progress report. World Bank Sta�Working Paper 375. World Bank, Washington, DC.

Sweet, C. F. and Weisel, P. F. (1979) Process versusblueprint models for designing rural developmentprojects. In International Development Administra-tion, Implementation Analysis for Development Pro-jects. ed. G. Honadle and R. Klauss. Praeger, NewYork.

Thompson, J. (1995) Participatory approaches ingovernment bureaucracies: Facilitating the processof institutional change. World Development 9 (3).

Turner, M. and Hulme, D. (1997) Governance, Administ-ration and Development: Making the State Work.Macmillan, London.

Upho�, N. (1985) Fitting Projects to People. In Putt-ing People First: Sociological Variables in Rural

PROCESS APPROACHES TO DEVELOPMENT 1357

Development, ed. M. Cernea, Oxford UniversityPress, New York.

Upho�, N. (1990) Paraprojects as new modes ofinternational development assistance. World Devel-opment 18 (10).

Upho�, N. (1992) Participatory evaluation of ruraldevelopment projects. IFAD, M&E Division, Rome.

Wanigasundara, M. (1995) The MONDEP decade: Tothe needy a helping hand. MONDEP, Colombo, SriLanka.

World Bank (1997) Rural Development: From Vision toAction, a Sector Strategy. World Bank, Washing-ton, DC.

Whist, E., Hassendeen, S., Sandberg A. and Jessen, S.A. (1982) Moneragala District Development Pro-gramme: a Prefeasibility Study. Report to NOR-AD.

Zuber-Skerritt, O. (1991) Action Research for Change &Development. Avebury.

1358 WORLD DEVELOPMENT