Report No. Agricultural Extension: Lessons from Completed...

Report No. 13000

Agricultural Extension:Lessons from Completed ProjectsApril 29, 1994

Operations Evaluation Department

FOR OFFICIAL USE ONLY

Document of the World Bank

This document has as restricted distribution and may be used by recipientsonly in the performance of their official duties. Its contents may not otherwisebe disclosed without World Bank authorization

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

ACRONYMS

ADP Area Development Project (which equates to rural development projectsin this report)

AFR Africa Region of the World BankEA Village-level extension agentEAP East Asia and Pacific Region of the World BankECA Europe and Central Asia Region of the World BankHYV High Yielding VarietiesLAC Latin America and the Caribbean Region of the World BankMNA Middle East and North Africa Region of the World BankNGO Non-government organizationOED Operations Evaluation Department of the World BankPCR Project Completion ReportPAR Performance Audit ReportSAR Staff Appraisal ReportSAS South Asia Region of the World BankT&V Training and Visit Extension Model

DEFINITIONS

face-to-face This relates to extension in which there is personal contact between theextension extension agent and the farmer

mass media This refers to communication which makes information availableextension to a large audience through media channels which may include the

written word, radio, video film, television or other forms of masscommunication.

FOR OFFICIAL USE ONLYTHE WORLD BANK

Washington, D.C. 20433U.S.A.

Office of Director-GeneralOperations Evaluation

April 29, 1994

MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT

SUBJECT: Aaricultural Extension: Lessons from Comoleted Proiects

Attached, for information, is a copy of a report entitled "Agricultural Extension:Lessons from Completed Projects" prepared by the Operations Evaluation Department.

Considerable institutional development has been triggered by T&V, the majorextension model promoted in Bank-financed extension projects. However, the lessons ofexperience suggest that greater differentiation in design of extension systems is required toenhance the sustainability and development impact of agriculture support services, takingaccount of specific institutional, farming and technology constraints.

While transfer of technology is the dominant objective of extension-and the mainfocus of the attached report-the approach to agricultural extension should be seen againstthe broader context of human resource development in the rural sector and the need topromote effective natural resource management and sustainable agriculture.

Further evaluation work is needed to explore these dimensions, especially thelinkage between education, research and extension and the promotion of adapted ruralinstitutions in developing countries.

Attachment

This document has a restricted distribution and may be used by recipients only in the performance oftheir official duties. Its contents may not otherwise be disclosed without World Bank authorization.

FOR OFFICIAL USE ONLY

AGRICULTURAL EXTENSION: LESSONS FROM COMPLETED PROJECTS

Table of Contents

Page

PREFACE ..................................................... E

EXECUTIVE SUMMARY ...........................................

1. INTRODUCTION ..............................................

2. THE EXTENSION SUBSECTOR AND BANK LENDING OVERVIEW ........... 2

The Development of Extension Programs ............................... 2The Bank's Role .............................................. 3Bank Lending for Extension ....................................... 4Benchmark for Extension Performance in the 1980s ........................ 5

3. IMPACTOEXTENSIONINVESTMENTS ............................. 10

TheExtensionProjectSample ...................................... 10Difficulties in Measuring Outcomes .................................. 10Performance Ratings of Completed Projects ............................. 14

4. PERFORMANCE IN EACHREGION ................................. 17

S. FINDINGS .................................................. 30

The Ex-post Findings ........................................... 30More Recent Developments ....................................... 41

6. RECOMMENDATIONS AND ISSUES ................................ 44

Recommendations ............................................. 44Issues .................................................... 48

This document has a restricted distribution and may be used by recipients only in the performance of theirofficial duties. Its contents may not otherwise be disclosed without World Bank authorization.

Table of Contents (continued)

Page

BOXES

1 Agricultural Extension Has an Important but Limited Function .................. 82 Differences Between Extension Investments in Free-standing and Component Projects ... 113 Applied Econometrics and Evaluation of Extension ......................... 134 Reasons for Widespread Use of T&V in Bank Projects ...................... 355 Differences Between OED and AFR on Interpretation of Findings ............... 50

Table: Frequency of Problems Raised in Ex-post Evaluation Reviews of Free-standing Projects 18

FIGURES

1 Total Bank Commitments toAgriculture ............................... 62 Total Research and Extension Commitments ............................. 63 Research and Extension Commitments as a Percentage of Agricultural Commitments .... 64 Free-standing Extension Lending Relative to Total Extension Lending ............. 75 Regional Distribution of Commitments to Free-standing Extension Projects Over Time 7

ANNEXES

1 Table 1 Free-standing Extension Projects approved after June 1977which have ex-post evaluations.................................. 53

Table 2 Agricultural projects approved after June 1977 which have ex-post evaluations . .. 542 Review of Component Projects byRegion .............................. 563 Recent Reviews of Extension Experience in AFR .......................... 644 A Process to Prepare an Extension Strategy and Investment Program .............. 675 Bibliography ................................................. 72

-i-


PREFACE

This report is part of a comprehensive review by the Operations Evaluation Department(OED) of the Bank's assistance for agricultural research and extension development in the 1980s. Itpertains only to extension and focusses on completed and evaluated projects, although a literature survey(including self-evaluations by the Bank's operating regions) has been used to take account of recent andcurrent developments. OED's comprehensive review of completed and ongoing extension projects, aswell as the agricultural research portfolio, will be produced in 1995.

The report reviews the performance of projects which have used a number of different exten-sion methods and analyses each method with respect to its relevance, efficiency, and efficacy in meetingproject objectives. For extension investments with an institution-building character (the majority), theissue of sustainability under existing and anticipated circumstances is addressed. This approach allows(a) identification of constraints and benefits of particular approaches to extension under specificcircumstances, (b) verification of key extension principles which can be applied to advantage in mostsituations, and (c) elaboration of a suggested direction for the Bank in managing its current extensionportfolio and in designing and managing its future extension interventions.

In general, the analysis of completed projects as presented in this report has been supportedby the operational regions and divisions in the Bank. However, although in agreement with therecommended extension principles emanating from the analysis, the Africa (AFR) region did not fullyagree with some aspects of the recommendations concerning the Bank's future involvement in extensionprojects. Positive and negative experiences in past projects were used to derive recommendations, butthere isn't an operating example of a state/national project where all recommendations have been put intopractice to verify their legitimacy. Consequently, there must still be some room for debate on how thefindings should be interpreted in specific country extension strategies. The report refers to the issuesraised by AFR in Box 5.

Some commentators from the Latin America and Caribbean (LAC) and Europe and CentralAsia (ECA) regions would have preferred more emphasis on how agricultural information systems couldbe improved, how beneficiary ownership could be enhanced, and how the private sector could take moreresponsibility for technical assistance to farmers. These concerns reflect regional characteristics in thatmuch of agriculture is more commercially developed than in many countries in other regions. However,the scarcity of experience in completed projects on these topics precludes any in-depth treatment.

The heavy reliance on ex-post evaluations also limits the extent to which some of the issuesthat are currently of particular interest to the Bank could be addressed in the analysis. The benefit offarmer participation was not always considered in project design, but receives a lot of attention in theanalysis of extension outcomes in the report. However, NGO involvement received little attention.Poverty alleviation effects can be associated with the extent to which the programs succeeded in obtainingimproved production in the smallholder subsector (although quantitative attribution of these effects to

- ii -

extension is not feasible). Environmental concerns at the time of appraisal of projects were usuallyimplicit in an objective to support sustainable development, but were generally not specifically referredto in ex-post evaluation. The role of women in agricultural development began to receive its legitimatefocus only towards the end of the 1980s.

The comprehensive review due in 1995 will attempt to address some of the extension issueswhich could not be treated in this report. These will include, inter alia: efficient and effective meansto provide technical assistance for livestock enterprises and for the integration of productive livestock withagriculture in smallholder systems; the use of performance indicators to assist in the ongoing evaluationand management of extension systems; staff attitudes to the design and performance of ongoing projects;the extent to which gender issues are receiving the required attention in service organization, staffing andprogram content; and private sector and NGO involvement in extension programs.

The author of this report is Dennis Purcell, Principal Evaluation Officer OED. ElizabethRobinson and Charles Sheerin provided research assistance, and Helen Claverie provided administrativesupport.

- iii -


EXECUTIVE SUMMARY

1. OED is undertaking a review of the T&V was first introduced in Turkey by D. BenorBank's agricultural research and extension invest- in 1967 and subsequently spread to South Asiaments in the 1980s. This will be finalized in (SAS) region in the early 1970s.FY95. The current report focuses largely oncompleted extension projects which have had ex- 4. T&V has dominated Bank projects directedpost evaluations. at improving public sector extension since that

time, although not to the exclusion of support forExENSION DEVELOPMENT commodity-specific extension services, technical

AND BANK INVOLVEMENT assistance aligned with credit institutions, andother intensive services associated with specific

2. At the time of independence in most development projects.developing countries, extension services wereusually linked to export commodities without BANK LENDINGattention to foodcrops. Newly independentgovernments continued this emphasis, but in the 5. The Bank's commitments to extension1950s national extension services with a wider from FY77 to FY92 was about US$3,000 millioncrop coverage began to emerge when it became in nominal terms. Although this is substantial, itobvious that the majority of farmers had little represents only about five percent of the Bank'saccess to improved technology. These services lending to the agricultural sector in the samewere mostly multifunctional (technology transfer, period. This is partly because extension projectsinputs, credit). In the 1960s and 1970s an are usually not capital-intensive, but alsoexpansion of the rural development project underscores the Bank's appreciation thatconcept reaffirmed and even extended the wide extension is only one of a large number of factorsresponsibilities of extension staff. affecting agricultural productivity, and that its

potential impact is limited by deficiencies in any3. The mode of Bank interventions in exten- of the other key components of the productionsion support followed this general trend. How- change equation-availability of improvedever, by the late 1960s the Bank had come to the technology, access to resources and inputs to useview that, although multifunctional extension staff technology, and profitability in resource andserved a useful development purpose, their many technology use at acceptable risk.duties detracted from the potential for technologytransfer. A hierarchically organized method of 6. The Bank has committed about 55 percentextension management was designed to of its extension investments through free-standingexclusively focus on technology and to deliver projects in which a primary objective was toselected and timely messages to farmers with build effective extension institutions andstrict regularity. The concept also aimed to have programs. The remainder consist of componentsfeedback of farmers' needs through extension to in projects in which extension was included tothe research system, and to maintain close help realize the potential benefits of other projectlinkages between research and extension. It was investments. Free-standing extension projectscalled the Training and Visit (T&V) model. were initially dominant in SAS and East Asia and

- iv -

Pacific (EAP) regions, but towards the end of the 9. In the absence of a practical method for1980s Africa (AFR) region became the major quantitative assessment, an extension programvehicle for this type of project. Indeed, national can be judged, albeit subjectively, on: (a) itsextension improvement projects became an relevance to the circumstances of the borrowerimportant plank in AFR's strategy to increase and the intended extension clientele, (b) itsagricultural productivity and decrease rural efficiency in using available resources, and (c) itspoverty in Africa. apparent efficacy in meeting (or contributing to)

the stated objectives. Due to the institution-IMPACT MEASUREMENT ISSUES building objective in most projects, the sustain-

ability of the institutional arrangements must also7. Quantification of the impact of extension be taken into account. The overall performanceinvestment in economic terms is normally not ratings of the extension portfolio in this reviewundertaken in either ex-ante or ex-post evaluation are based on the ex-post evaluation assessmentsof extension projects. Collection of compre- of Project Completion Reports (PCRs) by thehensive data is a constraint, but the main Bank's operating regions and of Performancedifficulty lies in linking cause and effect in Audit Reports (PARs) by OED. The abovequantitative terms. Awareness of technology can elements have been considered in the Bank'scome from any one or more of various sources, ratings, but not always with the same weightsand numerous factors apart from awareness (para 12 below).impinge on farmers' decisions in the use ofavailable resources-changes in profitability in PERFORMANCE OF EXTENSION INVE'MENTSproducing particular commodities or in usinginputs, changes in rural infrastructure or in the 10. All 33 of the free-standing projectsaccessibility to inputs, credit or markets, or completed in the review period have beenchange in the risk environment. included, and 30 of these were modeled on T&V.

They have usually covered a substantial area of a8. Applied econometric techniques have been state or country (apart from two pilot projects),attempted to link extension investment to a and all have had a strong institution-buildingspecific output. However, it has proven character. Of 74 component projects in theextremely difficult to include all the relevant extension component sample, 26 used T&V. Thecontributing variables in production equations. component sample was selected on the basis ofHence, the conclusions derived from these statis- size of extension component and the need fortical exercises must be treated in a circumspect representation in all Bank regions.manner. Nevertheless, published estimates ofeconomic benefits have in most cases been 11. Seventy percent of the 107 projectssignificantly positive. A study comparing two reviewed in this report had a satisfactory ratingsimilar districts in India, one of which was for overall outcome in ex-post evaluations. Thissubjected to T&V under a Bank-funded project, applied equally to the free-standing andindicated a high probability of at least 15 percent component investments, and is higher thanreturn on the incremental investment in T&V. average for agricultural sector projects in theTwo ambitious econometric studies on T&V 1980s (63 percent). However, the "softness" ofrecently undertaken by the Bank (in Kenya and the assessment method, and the tendency forBurkina Faso), while subject to baseline data and evaluators to accept substantial achievement ofanalytical constraints, did conclude that the quantitative output by the extension institution asextension investments were worthwhile in both an indication of success, reduces the weightcases and estimated very high rates of return on which can be given to this relatively favorablefurther marginal investments in the T&V overall outcome from extension projects.systems.

-v-

12. In nearly all cases, the nature of the FINDINGS

investments ensured increased institutional outputthrough programmed services, improved trans- 14. Detailed analysis of project ratings forport to enable more staff time with farmers, each of the 107 projects in the sample reveals theconcentration on technology transfer, and following:increased and/or reorganized staffing in a super-vised system. This almost universally meant that General extension portfoliosubstantially improved farmer coverage occurred.In addition, increased training of staff and use of Positive outcomes:a large number of on-farm technology demon-strations were regular features, so that some * the large Bank extension portfolio hasmechanisms which had the potential to improve performed at least as well as the agriculturalthe quality of institutional output were also sector portfolio as a whole.widely implemented. Under these circumstances,it was not appropriate to rate a project as unsatis- * it has increased the attention given byfactory, even if there were serious concerns about borrowers to the role of improved technologythe quality of institutional output. Concerns for in increasing agricultural productivity andefficiency, relevance and sustainability were often smallholder welfare.outweighed by the desired achievement of oneimportant objective-better use of the pre-project * the focus of most projects on improvedinstitutional resources to get improved farmer organization and management of staff-intensivecoverage with staff concentrating on technology services has increased coverage of thetransfer during the project period. smallholder subsector by public extension

services.13. A closer examination of the project evalua-tions demonstrates that there were serious Negative observations:problems not only in unsatisfactory projects butalso in projects- rated as satisfactory. The 70 * funding shortfalls were common phenomena,percent rating should not be allowed to generate so that the sustainability of over 70 percent ofcomplacency with respect to the extension port- free-standing projects was rated as eitherfolio. The need for concern is underscored by uncertain or unlikely.the serious recurrent cost funding constraintsfaced by most of the national extension systems. * effectiveness was reduced when recurrent costEven if a favorable marginal economic rate of funding proved inadequate.return on an extension investment could beestablished, most governments would be unlikely * insufficiency of relevant technology wasto continuously allocate substantial amounts of frequently a problem, and a major constraint inscarce recurrent budget resources to a service resource-poor environments.which is not generally recognized as having highpriority and which is not perceived as being * linkages with research were generally weak,efficient and relevant. Apparent deficiencies in and neither research nor extension werethe qualitative aspects of extension provide sufficiently conscious of the need to understandadditional reasons for poor funding, and poor the constraints and potentials of the differentfunding of personnel-intensive services reduces farming systems as a basis for determiningtheir effectiveness. Thus, evaluation of extension relevant technology and technologyinterventions needs to take increasing account of development requirements.qualitative aspects and sustainability.

- vi -

* with few exceptions, little attention was given management system used. In addition, the(in project design or implementation) to hierarchical structure and "message-centered"systematic participation of the farming delivery in T&V has done little to change thecommunity in problem definition, problem "top-down" characteristic in formulatingsolving, and extension programming; where recommendations which has been traditional inparticipation was emphasized, however, (cases most extension bureaucracies, despite T&Vsin Thailand and Mexico) results were very objectives of using farmer feedback to ensurepositive. relevance in recommendations.

* a "top-down" culture is traditional in the * the foregoing implies that the Bank can bepublic sector institutions in most developing faulted in the extent to which it has promotedcountries; this persisted in most Bank projects the T&V model (especially in earlier projects)and was contrary to the development of in a relatively uniform package of investmentsresponsive services. and extension practices applied in large state

and national programs. Adjustments to fiscal,* staff quality was a major constraint in nearly institutional, human resource, farming system

all free-standing projects where large staffs and technology stock circumstances shouldwere required, both at the level of farmer have been made in project design to increasecontact and in the technical support staff; the probability of developing services whichdeficiencies in the latter category limited the would be perceived as being efficient and cost-potential impact of continuous training effective, and which would continue to attractprograms in large state programs. the required financial and political support.

* lack of commitment by senior government * despite the above criticism of projects based onofficials was noted in nearly half of free- the T&V system, many of the organizationalstanding projects and affected implementation principles included in the T&V model areand funding support, and monitoring to internalized in the majority of "good"support management and justify funding was extension services and are unquestionablyweakly developed in nearly all projects. sound: programming of activities, the

technology focus, continuous training of staff,T&V free-standing projects program supervision, close research-extension

linkage and farmer feedback to allow* T&V, as the major extension model in Bank adaptation of technology to farmer circum-

projects, must be given a lot of credit for the stances. Unfortunately, many of thesepositive achievements attained, but cannot be principles were not adequately developed indivorced from many of the shortcomings of the projects. As noted above, this was commonlyfree-standing projects for which it was the linked to the persistence of the "top-down"extension system used in 90 percent of cases. culture and the large scale of most programs in

circumstances where there was a serious* recurrent cost funding and sustainability con- shortage of qualified and experienced

cerns, lack of appropriate technology, def- personnel. An entrenched research cultureiciencies in staff quality, and poor development inadequately responsive to farmers' needs wasof a responsive attitude amongst staff can all an additional contributing factor to the weakbe partly associated with the large scale research-extension linkage.application of the T&V system across states.These problems could be expected with most * the T&V concept of a contact farmer as thenation-wide staff-intensive extension program primary recipient of extension visits (for subse-in developing countries, regardless of the quent transfer of technology to other farmers

- vii -

in the vicinity) was not very effective and was other end of the spectrum, results confirmedoften modified to a farmer group focus, partly that the highly professional services needed forbecause contact farmer selection yielded specialized higher value crops (e.g. vegetables,farmers who were unrepresentative of the export fruits and flowers) should be providedlimited resource circumstances of those to be directly by the private sector.influenced. Even then, insufficient attentionwas given to the use of farmer groups with * extension supplied as an adjunct to a creditsimilar resources and constraints, so that the delivery program was usually effective inpotential for increased effectiveness through supporting technology adoption by thegroup problem definition and ownership of clientele. Deficiencies sometimes occurred inproblem solutions could not be fully realized the use of credit in inflexible production(nor were staff trained in how to facilitate this packages which were not always relevant toprocess). farmers' circumstances. But where these

services were provided at public cost, anOther projects equity issue arose in that an intensive and

costly service was being provided to a very* in integrated rural development projects the small proportion of the smallholder subsector,

outcome of extension services, whether T&V and usually only to those with a betteror other even more intensive service models, resource base, implying reduced resources forwas related to the extent to which basic extension to the majority.requirements were met: availability ofrelevant technology for the targeted farming * periodic extension campaigns which respond tosystems and the efficiency with which the a widely acknowledged crisis were usuallycomplementary services in input supply and effective as they were able to avoid many ofcredit were delivered. the constraints faced in developing an efficient

"permanent" service.* commodity-specific extension programs

demonstrated their ability to provide all the 15. The findings do not support a contentioningredients needed for technology adoption to that there is a single extension model which hasenhance the production of the targeted sufficient superior features to justify its uniformcommodity. As these programs usually adoption in an extension service in all small-involve commercial crops, however, evalua- holder farming circumstances. Rather, theytions noted that cost recovery in some form for demonstrate that there are some key principlesthese specialized services should normally be which determine the effectiveness of theanticipated. Specialized public services with extension services offered. Paramount amongstcost recovery and private sector services can these is the principle that extension recommenda-be complementary to general public extension tions must be relevant to the conditions of theservices. client, regardless of the extension method used.

This, in turn, necessitates that extension must* there was only one case of a project understand the characteristics of the targeted

sponsoring (successful) private sector extension farming systems and the factors which impingeservices (Chile). The preponderance of small- on these systems, and must be underpinned by anholders at or near subsistence level in many effective technology generation/adaptationborrowers' rural sectors limits the extent to capacity.which payment for services can be instituted,although some form of farmer ownership of 16. Having met this basic tenet,services should be the ultimate objective, methodological innovations such as programmingregardless of resource circumstances. At the and scheduling activities can enhance service

- viii -

efficiency and its efficacy in meeting objectives, interventions, a poor credit delivery system toas long as a primary consideration in their design smallholders, and often political and/or economicand implementation is to be responsive to instability-as the Bank is relying on extensionfarmers' needs. The desired impact cannot be investment as an important component of itsobtained, however, unless a public sector sectoral support strategy. Such constraintsextension investment is designed with full severely limit the potential for national programsrecognition of other potentially limiting factors in to be efficient users of public funds in attainingfiscal, institutional and human resource technological change in the targeted smallholderconstraints. Although investment in extension is communities unless their scale, type andusually justified as a public good in most composition are adjusted to individual countrydeveloping countries, fiscal realities dictate that circumstances.an extension program must be as cost-effective aspossible so that demands on the recurrent budget RECOMMENDATIONS AND ISSUESare minimized and prospects for sustainability ofthe service are enhanced. Just as the unavail- 18. Recommendations are elaborated for futureability of a supply of new relevant technology and ongoing Bank interventions in agriculturalwill lead to diminishing returns on continuation extension (Annex 4). Underlying theseof an extension service, so too will a scarcity of recommendations is an assumption that publicfunds which limits the field activities in staff- sector extension investment can give favorableintensive, face-to-face services. Apparent economic returns, provided (a) the elements ofefficiency and effectiveness, of course, are relevance and efficiency are adhered to, (b) thereconducive to funding support. is a significant technology gap between current

practices and available, more profitable17. The report also briefly reviews recent technology which is relevant to, and sustainabledevelopments in each of the Bank's regions. in, the targeted farming systems, and (c) alter-LAC was notable in that it is placing greater native, effective forms of technology transfer areemphasis on encouraging higher farmer partici- not able to be provided from private sources.pation in ownership of services through direct The author believes that there is ample evidencepayment or through a local taxation base. In in the literature to support this contention.general, Bank staff and borrowers appear to beaware of the major constraints to increasing 19. The major thrust of the recommendationsadoption of improved technology through sustain- is that considerably more attention must be givenable public sector extension services, and most to pre-project analysis and adequate projectare trying to make adjustments in their extension preparation assistance furnished to borrowers.systems to reflect at least some of these Only with this emphasis and an analyticalconstraints. Plausible solutions, however, in the approach to defining institutional and farmerlarge national T&V programs (which are clientele circumstances, can a responsive exten-dominant in Bank-supported project) are difficult sion strategy and an investment program repro-to realize. Deficiencies and inefficiencies in the senting a current phase in the strategy beextension systems themselves, as well as factors developed. Rapid rural appraisal techniques areexogenous to the extension systems, limit the rate likely to be necessary to analyze the existingat which the ultimate extension objectives of farming systems. Heavy borrower participationincreasing productivity and alleviating poverty in this process is essential to ensure ownershipcan be achieved. This is of special significance and commitment.in the case of AFR-scarcity of relevanttechnology, major limitations on staff quality, 20. The design would, inter alia: prioritizesevere fiscal limitations, weak development of the target groups; take account of public and privateprivate sector in providing complementary sectoral services; incorporate traditional massservices with a reliance on inefficient government media and modem information technology to the

- ix -

extent that it is applicable; define the scale, type * The limited capacity of most borrowers, andand intensity of face-to-face services in particular even of staff in some operational units in theareas in accordance with needs and resources; Bank, to undertake the required analysis andconsider the needs of all socioeconomic groups, extension design must be faced. Few Bankincluding women in farm households; and staff have had significant experience in exten-develop a needs-based staff training program sion design and operations outside the Bank,which focuses not only on technology but also on and most staff dealing with extension projectsmethods of interacting with farmer groups to have had little exposure to systems other thanmaximize their participation in problem definition T&V due to its dominance in Bank projectsand resolution and their support for the extension over the last 15 years. Bank supervisionprocess. capacity may also be constrained by limited

divisional expertise. This implies extensive21. An appropriate program design, however, use of experienced consultants in projectwould have to be complemented by responsive design, but also suggests in-house trainingactions by the borrower and the Bank during within the Bank assisted by external experts.implementation. Borrowers should appreciatethat success is not measured by having the * Although in many countries there is a largeplanned service organization in place, but by the proportion of smallholders near subsistenceextent to which the service is providing relevant production levels, rapidly changing economic,technology as defined by farmers in the targeted trade and sectoral conditions will impact on thefarming systems. This would require the institu- type of extension services that will be needed,tionalizing of a capacity for ongoing evaluation to and the respective roles of private and publicpermit responsive decisions by extension manage- sector technical assistance will have to adjustment. Adherence would also be necessary to the to these demands. Not only must the nationaldesign principles that aim to ensure that appro- strategy to encourage technological improve-priate technology is generated and extended. ments be responsive to changing circum-

stances, but any program to develop a cadre of22. For the Bank, more emphasis would be extension staff should be cognizant of the needgiven to monitoring the effectiveness of the for staff to be capable of meeting requirementssystem in providing efficient and relevant for a more responsive and analytical serviceservices to each of the farming systems in the (whether this be private or public).program, and the responsiveness of the service toevolving circumstances and new knowledge. 24. AFR supports the report recommendation

that the location and rate of introduction of23. Acceptance of these recommendations, extension investments should be in accordancehowever, would mean that a number of issues be with local circumstances. It differs, however, inconfronted: how this should be interpreted in practice, and

these differences are elaborated in Box 5 of the* Fiscal, human resource, complementary report. AFR has no disagreement with the basic

service and technology stock constraints may extension principles highlighted in this report,dictate that investment in state-wide, staff- including the need to give priority to theintensive services is inappropriate. This may development of extension services which are fullynecessitate a departure from the current responsive to the needs of farmers in the targetedapproach in some regions of installing (or farming systems, and the need for a close linkagecontinuing to support) relatively uniform with a research service which is itself organizednational programs regardless of local circum- to be similarly responsive.stances.


1. INTRODUCTION

1.1 The Operations Evaluation Department (OED) of the World Bank is undertaking areview of the Bank's agricultural research and extension investments made from the late 1970s to1992.1 The review (to be finalized in FY95) will analyze performance of these investments, andcompare this with performance of the research and extension portfolio implemented during the 1970s,which was the subject of an earlier OED study.2

1.2 This report is designed to disseminate early findings concerning the extension componentof the review. It pertains to an analysis of extension projects which have been completed and beensubject to ex- Tost evaluation as Project Completion Reports (PCR) by the Bank's operationaldepartments. The final study will also take into account on-going extension projects, cover theresearch portfolio, and encompass a broader analytical framework.

1.3 The report tries to determine (a) whether the extension investments made in the 1980stook into account the concerns and recommendations of the earlier OED review;4 (b) whether all theOED recommendations continue to be valid; (c) which measures need to be taken to improve theoutcome of extension investments; and (d) which issues have the potential to continue to impedesuccessful investment in extension services. This analysis is done in the absence of a specific Bankpolicy to guide extension investment.5

1.4 Extension projects are classified as free-standing or as extension components of morecomprehensive agricultural development projects. Free-standing extension projects are characterizedby their institution-building nature.6 The development of an effective extension delivery system isusually the primary objective and is expected to result in more efficient use of resources to attain ahigher level of sustainable agricultural production than would occur without the investment inextension services. The component projects normally involve extension services which are designed

1. Projects initiated after FY92 are considered to be too early in their implementation to offer any reliable lessons.

2. World Bank (1983).

3. Fifty seven percent of the projects have also been subject to independent review by OED as Performance AuditReports (PAR).

4. World Bank (1983) op. cit.

5. This does not suggest there should have been an accepted method to be used in extension projects; it is simply astatement which rationalizes the use of the OED review as a yardstick for performance.

6. Free-standing extension projects are titled as "extension", *extension and research" or, less commonly, as'agricultural support services" (para. 3.1).

-2-

to complement and enhance the benefits to be obtained from other major development investments inthe project.

1.5 All the free-standing extension projects approved from FY77 and which have had ex-postevaluations are studied, as well as a representative sample of agricultural development projects with asignificant extension component in each of the Bank's six operational regions. Although most of thefree-standing projects have a common extension model, the inclusion of component projects providesa variety of extension management systems.

1.6 Chapter 4 provides a summary of the performance of each of the free-standing projectsand of the lessons from component projects in each of the Bank's regions. This detail may not berequired by some readers, and the findings from the performance review of the projects is given inthe following Chapter 5.

2. THE EXTENSION SUBSECTOR AND OVERVIEW OF BANK LENDING

THE DEVEIwPMENT OF EXTENSION PROGRAMS

2.1 The term agricultural extension can mean different things to different people. Thecommon elements of most of the Bank's extension project portfolio would suggest that extensionmight be described as the process of assisting farmers to become aware of, and adopt, improved tech-nology from any source to enhance production efficiency, income and welfare. Such a simpledefinition, however, belies the complexity of the process that involves changing farmer behavior and,often, institutional behavior. Nor does it reflect the changing approaches which have been and stillcontinue to be used by governments and donors in attempts to provide or improve extension services.

2.2 An informative review of the historical approaches to extension and of donor involve-ment in these developments is provided in the 1990 Bank publication "Agricultural Extension: TheNext Steps". Before independence in most developing countries, the colonial powers and dominantprivate firms initially provided advice to farmers, tenants or workers in the production of specificcommodities in which the colonial power or firm had a particular commercial interest. As thecolonial systems were taken over by national interests after independence (often as parastatals) theyusually retained the strong hierarchical commodity orientation of their predecessors. Thesecommodity-specific programs, which often provide an array of services including technical advice,input supply, credit-in-kind and market arrangements, still exist (or have been created subsequent toindependence) in all regions.8

2.3 The commodity-specific services had the advantages of a relatively intensive, compre-hensive service which has a high potential for increasing the output and quality of the particular

7. World Bank (1990). Chapter 2 for approaches and Chapter 3 paras. 3.01 to 3.48 for donor involvement.

8. This type of organization is especially common with traditional estate crops represented as industrial tree crops,beverage crops, sugarcane and cotton.

-3-

commodity. The commercial nature of the service also allowed cost recovery through deductions inprices paid for the commodity or other form of payment.9 They had the disadvantage, however, ofproviding advice to only a narrow range of farmers, and, even then, usually ignored the traditionalfoodcrops that provided sustenance to many of their farmer clientele.

2.4 In the 1950s, governments became aware of a need to bring new knowledge and tech-niques to a broader spectrum of farmers, and new national extension services began to emerge.Deficiencies in overall sectoral services led to most of the public extension programs adopting notonly a technology transfer function but also services associated with input supply and often credit andproduct marketing. Further, realization of the need for capital investment in rural infrastructure andfor other social services to support the majority of smallholder farmers led to the evolution in the1960s and 1970s of multipurpose rural development projects and programs. This often resulted in aneven more diverse set of responsibilities for field staff (even to the extent of census-taking andselective human health services in some instances).

2.5 The broad responsibilities of field staff and lack of incentives to be effective in theirtechnology transfer function led to inadequate extension performance which became widelyappreciated. This encouraged the persistence of many commodity-specific services and the adoptionof technical assistance linked to state credit institutions in some countries. The main outcome,however, was the evolution of more specialized public sector extension which aimed at limiting advicelargely to technical agricultural matters. This occurred from the mid-1970s. Programs were usuallydeveloped in the Ministry of Agriculture, which may or may not have been separate from livestock,fisheries and forestry agencies. The technical extension services tended to be mostly focused onannual foodcrops, but some did cover other aspects of farming. In many countries, however, a largepublic extension service for foodcrops emerged, alongside smaller extension organizations for non-foodcrop activities which often operated in a different management system.10

THE BANK'S ROLE

2.6 The Bank's initial support for extension was through commodity-specific services (KenyaTea Development Authority in 1964). Subsequently, it also assisted many state credit organizationswith technical services that were linked to development investment. With the move to integrated ruraldevelopment projects, the Bank supported a variety of extension methods under these programs. Ofmost significance, however, in the evolution of the Bank's extension interventions, was the initiationand widespread adoption of a highly managed form of extension organization which became known asthe Training and Visit (T&V) system. T&V was initiated in Turkey in 1967 and was subsequentlyused in Bank projects in other Asian countries and spread to all the Bank's operating regions. Itbecame the main management system used in services specialized in technical agricultural extension.

9. A particular example is the Compagnie Francaise pour le Nvelopment des Textiles (CFDT) created by theFrench government in 1949 to develop cotton in the French colonies. After independence CFDT promoted theestablishment of national companies and provided technical assistance; it operates in 24 African and 5 Asiancountries.

10. A review of different extension approaches, their evolution and the pros and cons of each is provided by an FAO1988 publication 'Guide on Alternative Extension Approaches" (Axinn, G. 1988)

-4-

2.7 T&V was a hierarchically organized method of managing extension, designed to deliverselected and timely technical messages to farmers with strict regularity. The system encouraged linkswith agricultural research institutions, emphasized a professional approach to extension, and requiredan exclusive devotion to extension work. In principle, all extraneous activities such as those relatedto the physical handling of inputs and credit applications were to be avoided.11 These conceptswere developed to address the deficiencies which were perceived to be limiting the effectiveness andefficiency of public extension services: extension workers were charged with many non-extensiontasks which limited attention to technology transfer, often reported to more than one authority, wereusually poorly trained, had little incentive for extension performance, had little contact with theapplied research system, were subject to little supervision, often relied on development or inputsubsidies to entice adoption, usually had a narrow crop focus, tended to work with those farmershaving more resources rather than with poorer households, and frequently relied heavily on largedemonstrations of high and complete input packages which were beyond the resources of mostfarmers.

2.8 Although T&V has dominated Bank projects directed at enhancing the effectiveness ofpublic sector extension services over the last 15 to 20 years, assistance to other forms of extensionhas continued. Chapters 3 and 4 of this report which review completed Bank projects demonstrate therange of extension activities supported: various forms of T&V in free-standing extension as well asin rural development and irrigation projects; comprehensive services (advice, inputs, credit or credit-in-kind) directed at specific commodities; integrated services to farmers with packaged advice andinputs linked to investment and/or production credit; use of private sector consultants to providesubsidized technical services; and in rare cases, extension in which farmer participation in problemdefinition and resolution has been maximized. A 1989 Bank review12 of 361 agricultural projectswith extension components indicated that 38 percent had some form of T&V extension, 15 percentwere commodity-specific services, and 46 percent used "other methods" for extension investments.

BANK LENDING FOR EXTENSION

2.9 Bank lending to the agricultural sector in the period FY77 to FY92 amounted to $56,420million. This has involved directly productive agricultural investment and supporting infrastructure,services to facilitate access to production inputs and markets, human resource development, andimprovement in agricultural research and extension services. In the 1980s the Bank also embarked onstructural adjustment lending to assist in macroeconomic and sectoral policy changes appropriate forsustained sectoral growth. As a percentage of total agricultural lending in the four four-year periodscommencing in FY77, the allocation to extension has been about 6 percent, 4 percent, 5 percent and 6percent respectively.1 3 The total commitments to extension were approximately US$3,000 million

11. Benor, D. et al. (1977); Benor, D. et al. (1984).

12. Baxter, M., et al. (1989).

13. This compares with about 2 percent, 4 percent, 4 percent and 5 percent for research commitments in the samerespective periods.

-5-

in nominal terms (Figures 1, 2 and 3).14 This represents substantial support for extension but at thesame time its relatively low percentage of total sectoral funding indicates it is usually less capital-intensive than most other agricultural development projects, and that the Bank appreciates that exten-sion is only one of a number of factors that contribute to farm productivity. 15 (See Box 1.)

2.10 The commitments in free-standing extension projects as a percentage of total extensioncommitments in the period of analysis has averaged 55 percent and ranged from 41 percent to 62percent in the four four-year periods (Figure 4). The emphasis on free-standing projects, however,has evolved differently in the Bank's operating regions (Figure 5). Whereas East Asia and Pacific(EAP) and South Asia (SAS) were both prominent in this type of project in the late 1970s, SAS wasmost active in the early and mid 1980s, and Sub-Saharan Africa (AFR) region assumed dominance atthe .end of the 1980s. Relatively large commitments to Latin America and the Caribbean (LAC) inthe 1977-1980 and the 1985-88 periods were due to large loans to Brazil, rather than to a largenumber of projects in LAC countries. The significant intervention in the Europe and Central Asia(ECA) region in the 1981-1984 period was linked to a large project in Turkey.

BENCHMARK FOR EXTENSION PERFORMANCE IN THE 1980s

2.11 The Bank does not have an agricultural extension policy against which the implementa-tion and performance of agricultural extension investments can be measured. The 1983 OED report(para. 1.1) examined, inter alia, extension investments in the 1970s and does provide a frame ofreference to assess progress in the 1980s.

2.12 The OED report recognized the positive impact of extension programs which had beensupported by the Bank. The interventions had served to enhance the attention given by borrowers totechnology transfer to smallholder agriculture, and its projects had increased the coverage by exten-sion services of the smallholder subsector. Already by the end of the 1970s T&V was the majormethod used in free-standing extension projects, so progress in implementing this concept figuredprominently irn the OED analysis.

2.13 Despite the overall positive assessments, the report identified a number of seriousimpediments which detracted from project performance and which substantially reduced realization ofpotential benefits. These issues reflected not only on the immediate outcome of the extension invest-ments, but also on their sustainability in the circumstances of each country:

14. The database on extension commitments in free-standing projects is reliable. In the case of extension componentprojects, however, the databases (in AGR) is imperfect, and upward adjustments to the extension commitment hadto be made based on a sample of individual cases. Adjustments to this database have also been necessary in thecase of irrigation components in an ongoing OED study on irrigation investments.

15. Cernea, M.M. et al. (1983). This publication reported on the results of a symposium held in Thailand in late1982 which reviewed nearly 10 years of experience with T&V in Asia. It pointed out that improved technology,available resources and inputs to use technology, and attractive benefit-cost ratios for its use at acceptable riskwere essential ingredients for agricultural improvement before extension could be effective.

-6-

PIgure 1: TOTAL BANK COMMITMENTS TO AGRICULTURE

Iso

10000-

I-

0119770 1981-4 1985a-8 1989-92

Yaw of Approal

Figure 2: TOTAL RESEARCH AND EXTENSION COMMITMENTS

1200 ---- -

20

19770191-84 198 199-92Year of Approval

FgMt 3: RESEARCH AND EXTENSION COMMITMENTS AS APERCENTAGE OFAGRICULTURAL COMMITMENTS

10

2

0 - I1 -----1977-80 1981-84 1985-8 1989-92

Year of Approval

S- Extmsion ffM Researb

-7-

lgure 4: FREE-STANDING EXTENSION LENDINGRELATIVE TO TOTAL EXTENSION LENDING

100- -

60 - -- -- - - -- - - --

40 -

20 - --- -------

1977-80 198144 1985-88 1989-92Year of Approval

Free Standing E Component

Mgure 5: REGIONAL DISTRIBUTION OF COMMITMENTSTO FREE-STANDING EXTENSION PROJECTS OVER TIME

100 mmmmm

100 - -- - --------------- ------

80

j4 0 M3 ----............

20 -_ _ ......

17-01981-84 1985488 1989-92

FIsc-1 Year

gfflW1 AMR E SASM .AE L AC ECA

- 8-

Box 1: AGRICULTURAL EXTENSION HAS AN IMPORTANT BUT LIMITED FUNCTION

Farmers, regardless of their resource anl socioeconomic status, will adopt new technologies and modify resource usewhen they consider the change is relevant to their circumstances and can help them achieve their objective. An extensionservice can have an important function in increasing the rate of adoption of measures which can enhance productivityand producer welfare.

The rate of adoption of a technology (using technology adoption as a proxy for any desirable change in resource use)by a farming population will depend on:

a) the characteristics of the production circumstances of individuals (resources of land, labor and capital; climatic andother production uncertainties; and access to inputs and markets)

b) the characteristics of the technology itself;

- the extent to which the technology contributes to cost reduction, risk reduction, and/or production increase,

- its benefit/resource cost ratio (degree of profitability) at acceptable levels of risk,

- the skills needed for its adoption,

- the level of infrastructure and resources needed, and

- the degree of complexity in introducing it to the farming system;

c) the socio-cultural characteristics of the individual farmers (e.g. education, attitudes) and of the farming community(e.g. cohesiveness, values, attitudes to change), which can influence the perception of relevance of technology; and

d) how rapidly the population is made aware of the technology and its application to local production systems.

Consequently, extension has the potential to increase the rate of adoption by being directly involved in increasingawareness, in facilitating skill acquisition, and in assisting in understanding of a technology and its relevance to farmers'circumstances-the technology transfer process. It can have an important role in feeding back information on farmersconstraints and potentials to the technology generation system to encourage relevant research. Extension can alsointroduce the research system to appropriate innovations introduced by farmers to enrich the technology developmentand adaptation process. In addition, provided extension staff have adequate basic training and skills, they can work withfarmers and researchers in on-farm development of solutions based on Indigenous knowledge as well as externallygenerated technoloy.a

Extension cannot, however, create an attractive economic environment for production and technological improvement.It is unlikely to be able to sustain a useful function unless it is supported by an effective technology generation/adaptationinstitutional capacity, as there are limits to the potential for local development of indigenous knowledge. Nor can it hopeto achieve desirable changes if other ingredients for change are more limiting-access to resources, inputs, markets orcredit.

This emphasizes the importance of having a conducive macroeconomic, sectoral and market environment, and a techno-logy generation system in place if intended extension benefits are to be realized. To the extent that complementaryservices are needed to enable technology adoption, these must also be in place.

The array of factors influencing the rate of adoption implies that the extension process must be able to take into accountthe circumstances and constraints faced by farmers in a large number of different farming systems and environments.As scarcity of resources and production uncertainties increase, not only do the farmers' decision-making processesbecome more complex, but also there are increasing demands on an extension service to be responsive to the expandedinformation needs of farmers for sound decision-making.b Similarly, more commercially advanced farming systems arelikely to require comprehensive and up-to-date information to select the best opportunities for resource use. This doesnot imply a limitation of the potential impact of extension, but indicates the limitations of extension which relies ongeneral messages.

Rajasekaran, B. (1993). Ravnborg, H.M. (1993).

-9-

(a) Limited attention was paid to analysis of agricultural production constraints insectoral analysis, and to analysis of real needs of major farming systems inpreparing extension projects.

(b) Insufficient account was taken of the characteristics of the public sector culture,personnel situation and funding capacity in promoting and introducing extensionmethods, and this was often associated with little country input into the projectpreparation process.

(c) Although the concentration on technical messages in T&V services had advantagesin introducing a degree of professionalism to the advisory process, this concentra-tion, and the emphasis on organization and management aspects, often led to theprovision of technical advisory services in circumstances where other factors suchas lack of access to resources or markets were the limiting factors to production.All factors affecting the decision-making process of farmers needed to be consideredin design and implementation of extension programs.

(d) Relatively intensive and programmed services were often expanded into areas(especially in rainfed, resource-poor environments) for which very little technologywas available to improve productivity.

(e) A dearth of appropriate technology was exacerbated by generally poor linkagesbetween applied research and extension, and weak development of the intendedfeedback of the problems and needs of farmers through extension workers to theresearch complex.

(f) The capability of existing extension staff was often overestimated in project design;there was inadequate provision to upgrade the education level of staff; and in-service training was deficient in the practical aspects of technology.

(g) Extension programs did not take account of the important role of women in theproduction aspects of smallholder agricultural production.

(h) Monitoring activity usually concentrated on implementation of the extension methodand programmed activities, rather than measurement of technology adoption or agri-cultural output which might be linked to extension to assess its impact and justify itsinvestment expenditures.

Overall, the review concluded that extension investment was justified but that any extension programmust be tailored to the circumstances existing in each country/state/reiion if it was to make the bestuse of the resources allocated to extension by individual governments. 6

16. The review opined that the use of highly programmed extension systems (exemplified in T&V) at the national/state level appeared to be more suited to situations where there was a preponderance of smallholders of relativelyhomogenous circumstances operating at a significantly lower level of technology than was possible in theproduction environment, and where there was large number of under-utilized staff in the rural civil service.(These circumstances implied that the farming constraints and potentials were easily defined, that technology wasavailable to fill a significant technology gap, and that staff resources were available for the new system.)

- 10 -

3. IMPACT OF EXTENSION INVESTMENTS

THE EXTENSION PROJECT SAMPLE

3.1 The sample of projects subject to this review is restricted to completed and evaluatedprojects (paras. 1.1 and 1.2). This includes all free-standing, evaluated extension projects approvedafter FY77 (33)17 as well as 74 component projects (Annex 1, Tables 1 and 2). Due to the prepon-derance of free-standing extension projects in SAS and EAP in the late 1970s and early 1980s (Figure5), this has meant that the sample of completed free-standing projects is heavily weighted towardsthese regions. This is balanced by selecting many more components projects for review in AFR,LAC and Europe and Central Asia/Middle East and North Africa (ECA/MNA) regions. In relation tothe value of extension commitments, the number of component projects from AFR is increasedbecause of this region's emphasis on extension in the second half of the 1980s and its smaller loansize, while the LAC sample number is reduced due to the high proportion of large loans extended toBrazil in this region.18 The component projects were selected by initially identifying thosecompleted and evaluated projects which had extension components in excess of US$7.5 million, andprogressing to projects with lesser extension components until the required numbers were obtained ineach of the Bank regions.

3.2 The differences between free-standing and component extension projects are oftensubstantial (Box 2), creating some difficulties in the use of common assessment criteria for theextension investment (para. 3.6).

DIFFICULTIES IN MEASURING OUTCOMES

3.3 Regardless of the extension method used, quantification of the impact of extensioninvestments in economic terms is normally not undertaken in either ex-ante or ex-post analysis.Collection of reliable data to indicate outcome (as production or income benefits) is a constraint, butthe main difficulty lies in linking cause and effect in quantitative terms. Numerous factors contributeto a specific production response and to farmers' decisions in use of the resources available to them.Many of these factors are exogenous to the extension investment-changes in gross margins betweencrops, or in the profitability in using a particular input associated with policy interventions or marketforces; changes in rural infrastructure and accessibility through private or public investment; improve

17. The free-standing projects in Annex 1, Table 1 comprise extension projects in which extension absorbs close to100 percent of project costs, research and extension projects where extension commonly comprises 30 to 70percent of costs, and three supporting services projects in which about one third of costs are due to extension.

18. Sample of free-standing projects: 17 in SAS, 8 in EAP, 5 in AFR, 2 in LAC and 1 in MNA (Annex 1, Table 1).

Sample of component projects: 7 in SAS, 14 in EAP, 25 in AFR, 19 in LAC and 9 in ECA/MNA; the lastgroup represents the region which was known as Europe, Middle East and North Africa (EMENA) Region formost of the period under review (Annex 1, Table 2).

Overall sample number: 24 in SAS, 22 in EAP, 27 in AFR, 21 in LAC and 10 in ECA/MNA.

Sample numerical order: 1 AFR, 2 SAS, 3 EAP, 4 LAC and 5 ECA/MNA.

Bank extension expenditure order: 1 LAC, 2 AFR, 3 SAS, 4 EAP and 5 ECAIMNA.

- 11 -

Box 2: DIFFERENCES BETWEEN EXTENSION INVESTMENTSIN FREE-STANDING AND COMPONENT PROJECTS

COMMON CHARACTERISTICS OF FREE-STANDING EXTENSION PROJECTS:

(a) emphasis on developing an institutional capacity to provide effective public sector extensionservices,

(b) very large area coverage, often on a state or national scale,

(c) the majority of smallholder farmers are targeted,

(d) dedicated to technology transfer, although in many cases there is also participation in support ofcomplementary production services,

(e) execution by a state or national extension agency, and

(f) an intensity of service commensurate with its scope/coverage (usually in the range of 1 extensionworker : 500 to 1,500 smallholder households).

COMMON CHARACTERISTICS OF COMPONENT PROJECTS:

(a) most of these components had institutional development objectives for the implementing agency,but the emphasis was often on providing technical services to ensure that the benefits of the projectdevelopment investments were not constrained by lack of technology,

(b) more limited in area coverage,

(c) specific target groups ranging from the majority in a project area (e.g. many irrigation projects),to those individuals receiving credit or investment assistance, and to producers of a specific projectcommodity,

(d) an emphasis on technology transfer, but much greater use of extension staff for complementaryfunctions (e.g. input supply, credit),

(e) execution by a special extension unit in a project management entity, by a parastatal agency, byan extension wing of a credit agency or sometimes by the regular agricultural extension agencyof the Ministry of Agriculture with special project support, and

(f) an intensity of service which is designed to ensure satisfactory benefits from the other projectinvestments, so that it is often of higher intensity than the general state services.

- 12 -

ment or decline in the provision of complementary services for input supply, credit or marketing;improvement in the production and risk environment through expanded private or public irrigation; orprocessing investments which create new market opportunities.

3.4 Despite these constraints to analysis of impact, a number of empirical studies on exten-sion have been attempted with the assistance of econometric techniques. In nearly all cases theserepresent research studies by external entities rather than assessments by implementing agencies. Two1991 reviews of published studies 19 on the economic impact of extension investments showed thatthe majority, in at least some of the versions presented, had a significant and positive extension effect.The reviews, however, also pointed out that many of the studies could have econometric deficiencies,and stressed that empirical results from applied econometric analysis must be treated in a circumspectmanner to avoid potential overstatements of the economic impact of extension investments.Commonly, there have been serious data constraints in these studies and an inability to include all thecontributing variables in production equations (see Box 3). Two recent ex-post studies on Bank-supported T&V extension projects in Africa (Kenya and Burkina Faso) 20 illustrate the limitations ofeconometric techniques. In Kenya there were serious limitations associated with suitable baseline dataon pre-project farm performance and extension activities, and with the comprehensiveness of thefactors which could be reliably included in the econometric analysis.21 Nevertheless, acceptingthese limitations of the study, high rates of return were estimated on further investments, at themargin, in the current T&V extension system. The Burkina Faso study, while providing some veryinteresting observations, was subject to measurement and recall errors in farm level data, and anumber of biases in the analysis which could not be adequately addressed. It showed that adoptiontechnology was higher in T&V extension groups; it then inferred, however, that a doubling of thenumber of farmers entering T&V groups would have a similar marginal impact on benefits in thisexpanded population, without taking account of the differences in predisposition to adoption of thosewho had already participated and those who had not participated. One of the more robust studies wasundertaken by Bank staff in India.22 This compared the effects on productivity of extension in twosimilar, heavily irrigated districts (in the states of Haryana and Uttar Pradesh), one of which usedT&V and the other a traditional extension system. Of the two main crops, HYV wheat and rice,there was no significant difference in impact on HYV rice yields, but a productivity gain of 6-7percent in HYV wheat was attributed to T&V (with the aid of econometric analysis). Subsequentcost-benefit analysis showed that an internal rate of return of at least 15 percent was highly likely.

19. Birkhauser, et al. (1991); Evenson, R. (1991).

20. Bindlish, V. and Evenson, R. (1993), and Bindlish, V., Evenson, R. and Gbetibouo, M. (1993).

21. In addition to econometric interpretation of impact, this study provided interesting operational data, including the

importance of women in agricultural production and their responsiveness to working with extension in women's

groups.

22. Feder, G., et al. (1985).

23. It should be noted that it would not be difficult to obtain a satisfactory rate of return on the incremental extensioninvestment made in most Bank-supported projects. This is because the projects commonly rely heavily on re-organization and strengthening of an existing extension structure. The incremental cost is low relative to the totalcost, so that the additional benefits compared to the pre-project performance need not be very high. This, how-ever, is distinct from the economics of the total extension system in which all costs and benefits would have to beconsidered, which would normally be considerably lower if it could be reliably calculated.

- 13 -

Box 3: APPLIED ECONOMETRIC$ AND EVALUATION OF EXTENSION

There are many potential ways of trying to link investment and extension in a causal mannerto the outcomes in the target agricultural sector. One approach that has been widely used is to specify a"production function' in which a measure of production, perhaps at the level of an individual farminghousehold, is related to measured inputs of real factors of production, such as land, labor, capital,purchased inputs, and so on and, as well, to include additional explanatory variables related to variationin the natural environment, the agro-ecological circumstances, and other service variables, such asextension expressed on an extension input per farm basis. There are many difficulties inherent inmeasuring all such inputs into a model of production. One of the most significant difficulties relates to thelags in causal effects between a service input and the flow of supposed productivity effects. Extension notonly has potentially productive impacts in the immediate accounting period of, say, an agricultural year,but because of the persistence of the increments to knowledge in subsequent periods, an evaluation shouldtry to capture the effects of such transmitted knowledge over the life of that knowledge in the farmingenterprise. Here, there are not only the immediate measurement problems, but there are the conceptualdifficulties of dealing with the "depreciation' of knowledge as new information comes to hand. At best,such complex processes can only be represented rather crudely in the algebraic models that are used inapplied econometrics.

The difficulties do not stop here, however. Even if representative farm-level data can begathered and its collection funded, and samples can be taken that are sufficiently rich in the sense that thereis enough variation in the explanatory variables to "identify" a relationship, the analysts must then moveforward by introducing some particular functional forms to represent the processes that are being modeled.The functional forms that are used in most practical applications are greatly simplified representations ofreality that are used primarily because of analytical convenience and parametric parsimony. A reader ofsuch results must recognize that every simple functional form that is used implies some algebraicrestrictiveness and thus is an additional qualification to the interpretation of any results. One of the mostcommonly used fuictional forms is the so-called Cobb-Douglas form, which features constant partialelasticities of production of all included factors of production. At best this is an extremely stylized viewof the nature of agricultural production. More general specifications involve potentially much largernumbers of variables that must be included in estimating equations and, therefore, incur difficulties withhaving insufficient "degrees of freedom' for the estimation to proceed satisfactorily.

In choosing among the infinity of econometric specifications that are actually used inpractice, a high degree of subjectivity and judgment is required. The analyst must take particular care toemphasize the potential biasing and possibly uncertain implications of various econometric judgments.Some arbitrary assumptions may tend to bias the returns up, whilst others may work in the oppositedirection. If, for example, extension but not research is included in the list of services in a farm-levelproduction function, to the extent that research is a truly influential source of productivity gains and isstatistically highly associated with any investments in extension, that effect will tend to be picked up byextension as the included variable, and, thereby, contribute to potential overstatement of the positive effectsof investment in that particular service.

Consideration of all such potential sources of bias and subjective interpretation means thatempirical results from applied econometric analysis of phenomena, such as research and extension invest-ments, must be treated in an extremely circumspect manner to avoid (a) potential overstatements of pasteffects of the investments and thus (b) bias in anticipated returns and distortions in implied priorities forsubsequent investment activities that may be influenced by such findings.

- 14 -

However, even this study was unable to capture the possible effect of a reduced reliability of irriga-tion in the non-T&V district on farmer attitudes to expenditure on production inputs.24

3.5 The difficulties inherent in attempting to obtain robust estimates of economic benefitshave resulted in the use of alternative, less rigorous, but more practical methods to assess impact ofextension investments. It is important to note again that measuring the outcome of the extensioninvestment represents an assessment of the impact of the program on accelerating the rate of adoptionof technology (regardless of the origin of the technology) or the decision making by farmers to usetheir resources in a more productive way. The inherent difficulty lies in estimating what would havebeen the situation without the extension investment. Major improvements in production and incomecan be obtained without any significant contribution from public extension services. In the absence ofa quantifiable measure of attributable outcome, three elements which have major impacts on outcomecan be examined to estimate the worth of the investment - relevance of the program to the situation ofthe borrowei and the intended extension clientele, its apparent efficacy in meeting its objectives, andthe efficiency in use of resources available for the program. As virtually all free-standing extensionprojects and many component projects have a major institution builWing objective, these threeelements must be seen in relation to building a capacity to deliver appropriate services in asustainable system. In recent years OED has uniformly used these elements in evaluating extensionprojects in its PARs. In many PCRs, however, efficacy has received more attention than relevanceand efficiency criteria (paras. 3.8 and 3.9).5

PERFORMANCE RATINGS OF COMPLETED PROJECTS

3.6 In OED's Annual Reviews of Project Performance, agricultural research and extensionas a subset of the portfolio has performed, on average, better than the agricultural sector as a whole.Of the 33 free-standing extension projects included in this review sample, 70 percent have been ratedin their respective PCR/PAR as having a satisfactory outcome (Annex 1, Table 1). The 74 extensioncomponents in agricultural development projects were rated by the author based on observationsrecorded in ex-post evaluations of each project and any complementary data available. It was possibleto confidently allocate to an unsatisfactory category, but lack of specificity in the reports on the extentto which all extension objectives were met, necessitated two categories as "marginally satisfactory"and "fully satisfactory" for the extension components. This resulted in an unsatisfactory percentageof 30 percent, with the marginally and fully satisfactory judgments being 43 percent and 27 percentrespectively (Annex 1, Table 2). Amongst the component projects, area development dominated thesample and also had the highest proportion of unsatisfactory extension results (38 percent). 26 This

24. See OED Report No. 8808, Annex 2.

25. These paragraphs elaborate how a limited interpretation of efficacy prevailed in many judgments. This was partlyassociated with a frequent narrow focus on the extent to which the infrastructural and equipment support andprogram organization and methods included in the project design were implemented. The evaluations usually didnot attempt to judge the legitimacy of the project design and how more efficient use of resources may have beenprogrammed.

26. Of the component projects, area development (rural development and settlement) dominated, followed by credit(agricultural credit and agro-industry), commodity-specific and irrigation projects. Although sample numberswere low for the latter categories, the area development projects had the highest unsatisfactory proportion for theextension component: Project type ADP CR COM IR Total

Project number 48 12 8 6 74Unsatisfactory (%) 38 25 13 0 30

- 15 -

implies that, overall, 70 percent of free-standing and extension component investments were rated ashaving met at least some of their major objectives. This compares with an average satisfactory ratingof 63 percent for agricultural sector projects in OED Annual Reviews in the 1980s.

3.7 The difficulties in assessing the outcome or net benefits of extension investments havealready been elaborated (paras. 3.3 and 3.4). This must raise the question as to the robustness of therelatively favorable results indicated in the performance ratings.

3.8 Although the ratings are largely subjective assessments, they do represent judgments onthe extent to which project objectives are met. All the free-standing projects and most of thecomponent projects (especially those using the T&V method)27 had a strong institution-buildingobjective. This was usually expressed in attempts to improve the quantity and quality of extensionagency output. Increased quantity was usually to be effected by:

(a) adopting a well-defined extension method which obliged expanded and regularfarmer coverage and identified clear expectations and responsibilities at all levels inthe bureaucracy;

(b) providing supporting facilities and transport to enable better farmer coverage and anincreased proportion of time in contact with farmers;

(c) avoiding (or substantially reducing) those service functions of field staff which werenot associated with agricultural technology transfer; and

(d) ensuring staff numbers at various levels in the institution were adequate to enablethe intended farmer coverage and implementation of the method, by recruitment orreorganization of existing public sector staff or both.

Increased quality was to be effected by:

(a) improving the institutional human resource base by regular technical training offield staff, and by technical back-up for field staff;

(b) improved linkages between research and extension to ensure the continued avail-ability of relevant technology;

(c) obliging field staff to feed back farmer concerns into the extension and appliedresearch system; and

(d) increased technology demonstration in large numbers of small plots in farmers'fields.

27. Thirty of the 33 free-standing and 26 of the 74 component projects in the sample used some form of the T&Vmethod.

- 16 -

3.9 Most of these institution-building projects also had the ultimate objectives of increasingproduction and farmer income and welfare, with an explicit or implicit poverty alleviation implica-tion. However, the common absence of reliable data and the inability to link cause and effect inassessing net benefits has placed heavy reliance on institutional development as a final measure ofoutcome.28 The favorable 70 percent satisfactory rating in the free-standing projects (whichrepresent the most complete expression of projects designed to improve institutional performance) canbe linked to the following:

(a) the investment in projects of this nature implied that governments were willing togive increased attention to technology transfer, which was a positive outcome initself;

(b) at least reasonable implementation of the project design occurred and this meant thatthe quantity of output from the implementing agency was improved relative to thepre-project situation-enhanced farmer coverage, greater technology focus, strictermanagement and supervision to maintain the improved service; and

(c) regarding quality of output, projects invariably expanded staff training andintroduced programs of small demonstration plots to support technical recommenda-tions, both of which were mechanisms which had the potential to improve theeffectiveness of the extension staff.

3.10 Generally, then, the projects with major institution-building objectives did have theintended impact in increasing the quantity of output and in introducing some mechanisms to improvethe quality of output from the extension institutions. In a dichotomous rating system of satisfactory/unsatisfactory, these achievements were usually considered adequate to rate projects as satisfactory,despite less achievement of many of the quality-improvement objectives. It must be accepted,however, that the relatively favorable overall ratinss are " soft" and not based on empirical evidenceof net benefits caused by the extension investment. 9 The project ratings can be consideredlegitimate, but should not be allowed to disguise the qualitative deficiencies.

3.11 The reasons for unsatisfactory ratings of free-standing projects are analyzed in Chapter4. It is of interest, however, to note that five of the eight unsatisfactory T&V projects (Loan 2150-PER, Loan 1626-PHL, Credit 690-IND, Credit 761-IND and Credit 1669-RW) were sufficientlydelinquent in the quantity of institutional output alone to warrant an unsatisfactory rating. In the

28. In many PCRs, substantial production increases have been claimed based on area production statistics, surveys oreven extrapolation of demonstration plot responses. These assertions'have not been relied on, however, inmaking the rating assessment for the OED database due to the impossibility of linking cause and effect. To theextent that production, income or yield changes did occur, over time, this represented a response to a largenumber of factors, of which the extension service was only one (para. 3.3).

29. Of the extension investments in the sample, the Kenya Extension Project (Credit 1387-KEN) was the subject ofan empirical study; this had a number of data and analytical constraints which affect the robustness of itsconclusions. Another study was undertaken in the early 1980s in India in the state of Haryana which wassupported in the Composite Agricultural Extension Project (Credit 862-IND); this is considered one of the morerobust extension evaluation studies in the literature (para. 3.4).

- 17 -

others, very poor performance in qualitative aspects and/or lack of apparent real benefits obliged theunsatisfactory rating.

3.12 The "softness" of the overall judgments is evident when a closer examination is made ofthe major constraints to effective services that are nominated in both the satisfactory and unsatis-factory projects. In Table 1, the frequency of mention of serious problems in the PCR/PARs of thefree-standing projects is summarized. This illustrates that there are serious problems in funding ofoperations, in having a sound basis for technical recommendations, in improving the human resourcebase, in responsiveness to country and farming conditions in project design and implementation, andin program ownership. Although these problems are more frequently highlighted in unsatisfactoryprojects, they are also of serious concern in those projects rated as being satisfactory. It is also ofnote that two thirds of the satisfactory projects had "uncertain" or "unlikely" sustainability ratings.

3.13 The project ratings indicated that, on the average, extension investments in AFR faredworse than those in EAP and MNA/ECA, while those in LAC and SAS performed better than inother regions. 30 This observation, however, is not so useful in itself for the analysis, as the varietyof objectives and approaches used could influence the ratings, apart from any difficulties inherent inagricultural development in particular countries. The sample was more suited to analysis of factorswithin projects rather than comparisons of single, bottom-line ratings. Such an analysis of invest-ments in each region is made in Chapter 4 of this report.

4. PERFORMANCE IN EACH REGION

4.1 SAS Region. This region was most prominent in adopting free-standing T&V projectsin the late 1970s and early 1980s and has 17 projects in the sample, all based on T&V (Annex 1,Table 1). The countries included India, Pakistan, Bangladesh, Nepal and Sri Lanka. Thirteen of the17 projects had satisfactory ratings, while four were unsatisfactory-projects in the states of Bihar andW. Bengal in India (Credits 761 and 690-IND), in the hill area of Nepal (Credit 1101-NEP) and inthe state of Baluchistan in Pakistan (Credit 1533-PAK).

30. Performance Ratings for Extension InvestmentsAFR EAP MNAI LAC SAS

ECA

Free-standing Satisfactory 3 5 1 1 13Unsatisfactory 2 3 - 1 4

Component Satisfactory 2 4 3 9 2Marginally satisfactory 11 6 3 7 5Unsatisfactory 12 4 3 3 -

Unsatisfactory Total 14 7 3 4 4Projects Total 30 22 10 21 24Unsatisfactory Projects (%) 47 32 30 19 17

Source: Annex 1, Tables 1 and 2.

- 18 -

Table 1: FREQUENCY OF PROBLEKfS RAISED IN EX-POST EVALUATION REVIEWSOF FREE-STANDING PROJECIS

Of23 Of8 Of31Description of Serious Problems Satisfactory Unsatisfactory Total

(S) Projects (UNS) Projects Projects

Funding % % No.

* Recurrent cost fnding problems during and subsequent to the 87 10 28project which seriously inhibit field operations

Basis for Recommendations

* Inadequate research - extension linkage to ensure the techno-logical needs of some of the major farming systems are defined 74 10 26and addressed (especially for resource-poor and less predictableenvironments)

* Ins&fficient technology available to enable a major andprogressive progran to improve production in some important 39 63 12farming systems

* An entrenched *top-down' approach in developing recommenda- 48 75 17tions, despite objectives of continuous feedback from farmers

* Little or negligible consideration of production economics, risks 39 50 13and d(iferent degrees of access by farmers to resources

Hanan Resource Capacity

* Training program unable to ensure front-line extension staff hadsuff cient practical knowledge of production systems (and their 43 88 17constraints and potentials) and of relevant technology, to providethe desired level of interaction with farmers.

* Speafic mention of low education level offront-line staff limiting 22 38 8the potential for a more analytical and responsive service.

Selection of Methodologies

* The adoption of a methodological 'blueprint' approach over alarge area (region, state, nation) did not permit a desirableadaption of services to the circumstances of each area; greater 35 50 12resources should have been allocated in the preparation phase tomake the project more responsive to fiscalinstitutionallfarmingsystem conditions

* 'Contact farmer' system not very effective, or reference made to 43 75 16better results from working with farmer groups

Commitment-Ownership

* Government, implementing agency management or staff not fullycommitted to all the principles and procedures of the extension 48 50 15program

* Monitoring activities very weak, or, if developed, not effectively 35 38 11used for responsive management

SustainabUlty of the System*

* Likely 33 -* Uncertain 58 60* Unlikely 8 40

*of only 12 S and 5 UNS projects (evaluated from 1989) for which this assessment was made.

- 19 -

4.2 In the nine state extension projects in India all but Bihar and W. Bengal projects wereconsidered to have effectively introduced the T&V organization and management system to the stateservices. Coverage of farmers was improved, many ancillary functions were removed to permit agreater focus on extension of technological improvements, use of small technology demonstrationplots was widely adopted, and field staff of relatively low education level31 were subjected tofrequent technical training. Formal linkages were established with the state agricultural universities(which were the centers for agricultural research in the states) which were to generate relevant tech-nology assisted by the gradual adoption of an agro-ecological zone approach in researchprogramming. The positive results obtained in the empirical study undertaken in Haryana, in whichT&V was supported under Credit 862, have already been reported (para. 3.4).

4.3 Although the research-extension linkages were generally established, the PARs stated apersistent deficiency in the extent to which technology problems in the less favorable environmentswere addressed. This was further reflected in reference to inadequate relevant technology to meetfarmers' needs in five of the projects. Many of the projects were dissatisfied with the results obtainedin using the contact farmer principle, and states made various adjustments in this and in other aspectsof the method in attempts to accommodate local conditions without rejecting the basic framework ofthe system. 32 These adjustments included working with groups rather than contact farmers,adjusting the visit/training frequency, maintaining some service functions in addition to technologytransfer, and in some cases developing greater organizational and programming discretion to districtlevel. Some also made attempts to move away from the basic foodcrop emphasis to cover otheraspects of the farming system, such as soil conservation and the production aspects of livestock. Theprojects had generally been developed around a standard T&V model and involved the transfer oflarge numbers of staff from the existing community development programs to be village extensionagents (EA). Adequate political and administrative support to facilitate this process was not forth-coming in many states (necessitating new recruitment and additional budget requirements), and thiscontributed especially to poor implementation in Bihar and W.Bengal. The tendency to use a"blueprint" approach in design without adequate pre-project analysis was also evidenced in sixprojects which reported that construction of staff housing was overdone; it was financially moreattractive for EAs to not occupy government housing so that low occupancy rates occurred.

4.4 Although insufficient budgetary support hindered implementation in some projects, themore serious concern was for the willingness of the states to continue to finance the large recurrentcost extension budgets. The salaries of staff would be met, but staff would not be effective unlesssupported with sufficient operational expenditure; constraints in this area were evident in mostprojects. Other major problems were poor capability of the subject matter specialists to give apractical and economical orientation in EA staff training, and a lack of responsiveness by EAs toactual farmer problems (associated with low education levels, iimited experience and inadequatetraining).

31. Most field staff had secondary school education, and most eventually had some form of an agricultural certificate.

32. These were generally considered to be favorable developments, and represented an evolution of practices underthe T&V system.

- 20 -

4.5 Of the three projects in Pakistan, two earlier projects in five districts in each of thestates of Punjab (1978) and Sindh (1979) were rated as having satisfactorily established T&V despiteinitial reluctance by governments to accept this model. Recurrent cost funding concerns, however,were expressed, as was an insufficiency of appropriate technology, poorly developed research-exten-sion linkage, unwarranted repetitiveness of many technical messages, and the need for more flexibilityin the system to be more responsive to farmers' needs and to reduce costs. Excessive construction ofstaff housing was also criticized, as in many of the Indian projects. The third project (Credit 1533-PAK) was approved in 1984 for four districts in Baluchistan which contained over 60 percent of thestate's irrigated land (only six percent of crops were rainfed). Most of the physical implementationtargets were eventually met and a T&V extension system was installed and provided better farmercoverage. However, there was little evidence that the changed system was having any influence ofuptake of new or improved technologies, and a number of serious problems became evident to theextent that tls project was rated as unsatisfactory. Coordination of research and extension initiallyimproved but subsequently worsened, and the specific adaptive research program financed under theloan did little more than repeat trials from the research system on the project farms; there was ascarcity of technology for new extension messages; staff training was not very effective; the loweducation level of EAs was a constraint to developing more responsive services; and the contactfarmer concept did not function well. The PCR criticized the lack of attention to the existing institu-tional, infrastructural and farming conditions in design of the project and to the potential use of massmedia; it also queried the sustainability of the system because of its high recurrent costs and apparentinefficiencies.

4.6 In Bangladesh, an initial extension and research project (Credit 729-BAN) was followedby a second project (Credit 1215-BAN) - which expanded coverage from the productive northwest tomost of the country. Both projects were rated as satisfactory in PCRs, as they did bring better exten-sion organization and professionalism, increased farmer coverage, and improved somewhat thelinkages between research and extension. It was assumed that extension has probably had an impacton increased use of HYVs, fertilizer and plant protection, although the significant production increasein rice (the major crop) could be largely explained by the expansion in area under water control (withits associated incentives for use of higher technology). Many serious problems, however, wereidentified in the service: top-down recommendations, demonstrations with limited relevance, poorquality in subject matter specialists, lack of attention to economics and farming constraints, poor useof mass media and poor results with the contact farmer concept. Funding the recurrent costs was alsoa serious concern, although this has been alleviated temporarily by a follow-on project which also isattempting to address some of the problems of the service and to improve its responsiveness.

4.7 In Si Lanka, the statewide project was very similar to the earlier T&V extension andadaptive research projects developed in India, but also attempted to integrate livestock services withfoodcrop extension. The extension part of the project was rated as barely satisfactory, as it didestablish a T&V system of extension throughout the country. The inclusion of livestock services withfoodcrops was finally not approved by government. Funding support was deficient during the projectand even more serious subsequently; this, combined with major administrative changes whichdevolved extension responsibility to the provinces in late 1988, and a perception in government thatthe T&V system did not apparently have the ability to efficiently respond to the needs of diversifiedfarming, led to dissolution of the program and the creation of a new system in a new project. ThePAR noted that the "top-down" characteristic in regular message delivery using poorly educated EAs

- 21 -

appeared to have been associated with a positive effect when the main focus was improved riceculture, but was not effective.when more analytical and responsive services were required.

4.8 In Nepal, there was an extension project for selected provinces in the lowlands (Terai-Credit 1 100-NEP) and in the hill area (Credit 1 101-NEP). The Terai project obtained improvedcoverage of farmers, at least in foodcrops, and was rated as satisfactory; there was some evidence thata rice yield response might be linked to the project, but not so for wheat yield. Difficult conditionsof access made implementation of the method more difficult in the hill area and this, combined withlack of evidence of any significant impact, led the project to be rated as unsatisfactory. Theimportance of agriculture in the economy prompted government to give reasonable funding supportduring most of the project implementation periods, but fiscal constraints subsequently led to decliningextension budgets. The problems of frequent lack of relevance of recommendations to the circum-stances of farmers, "top-down" message formulation, and limited effectiveness of contact farmers andof contracted farmer-EAs were also serious constraints.

4.9 There are seven component projects from SAS in the sample. T&V was used in twoirrigation projects (Credit 736-IND and Credit 1316-NEP), in a sugar development project (Credit1399-BD) and in a second commodity-specific project towards the end of the implementation period(Credit 765-BD). The other three projects had intensive commodity-specific services. Allsatisfactorily met at least some of their extension objectives, although there was doubt regarding theimpact under Credit 1399-BD.

4.10 A review of the characteristics and outcome of SAS component projects is presented inAnnex 2. The observations from this sample are as follows:

(a) Intensive commodity-specific extension supported by credit is likely to be effectivein increasing production of the commodity (Credits 1012 and 1034-IND and Credit765-BD). The commodity must have a comparative advantage in using farmers'resources, however, to achieve widespread support (Credit 1399-BD). The programshould also at least be aware of how the commodity interacts with other componentsof the farming system, so that farmer concerns can be taken into account in thecommodity program (Credit 1012-IND).

(b) When commodity-specific services are provided to farmers in significantcommercial production of the commodity, this suggests that clients should pay fortechnical services though fees or market cess (Credit 1034-IND).

(c) A favorable production environment created by irrigation investment is conducive toa satisfactory outcome from T&V extension (Credit 1316-NEP). Even thoughappropriate technology is more likely to be available in these circumstances, this isnot always the case, however, and its absence can limit results (Credit 736-IND).

4.11 EAP Region. There were eight free-standing projects evaluated in EAP (Annex 1, Table1). T&V was used in two national. projects in Indonesia (Loan 1267 and Credit 996-INS), two inThailand (Loans 1393 and 1752-TH) and one each in Malaysia (Loan 1493-MAY) and the Philippines(Loan 1626-PH). All but the Philippines project were rated satisfactory.

- 22 -

4.12 In Indonesia, following earlier experience with T&V in some irrigation command areas,an initial national project (Loan 1267-INS) concentrated on foodcrop extension in 13 provinces whichcovered 80 percent of the farming population. This project successfully organized T&V foodcropextension services in the 13 provinces, but the focus was mainly on rice cultivation. Most extensionwas closely linked to a special program (BIMAS) which provided subsidized production inputs, creditand guaranteed markets for rice and, to a lesser degree, selected other crops. The PAR wasconcerned with the issue of continued budgetary support for the large organization and the inherentcost inefficiencies in having separate services for foodcrops, treecrops, livestock and fisheries. Asecond project (credit 996-INS) attempted to expand the T&V service to cover all subsectors through-out the country (through generalist "polyvalent" EAs). The project was rated as satisfactory, but thePAR considered that substantial improvements and modifications would be needed to attract continuedbudgetary support. The problems common to many of the SAS projects were also apparent here:lack of appropriate technology in many rainfed environments which severely limited impact in newtransmigration areas, "top-down" formulation of recommendations, poor contact farmer performance,inefficient use of demonstrations, and low education level in EA staff with poor support from subjectmatter specialists which limited the potential responsiveness of the service to farmers' needs. Thepolyvalent EA concept did not have the full commitment of the relevant agencies and was not wellimplemented (and was abandoned in 1992). The PAR concluded that the imposition of a uniformextension methods in a country with a wide range of production and socioeconomic circumstances asin Indonesia was unlikely to be the most cost-effective use of public extension investments.

4.13 In Thailand, an initial project covering 33 of the country's 72 provinces (Loan 1393-TH) evolved from earlier use of T&V in irrigation projects. Political pressure led to a second project(Loan 1752-TH) to cover the remaining provinces (which would benefit from significant fund transferunder the project) only two years after the start of the first project. The projects established a modelof T&V throughout the country and were rated as satisfactory. The overlapping of the projectscaused budget funding problems and delayed implementation of the key components of training andthe recruitment of subject matter specialists. Nevertheless, budgetary support subsequent to theprojects was reasonably good, although the decline in the operating support/salary expenditure ratiowas of concern. As for most other T&V projects, monitoring surveys indicated contact farmers werenot very effective, and there were problems in ensuring relevant technology was available for manycommon farming systems; farmers were aware of most recommendations, but non-adoption was dueto their not being appropriate to the production circumstances of the farmers. The governmentmodified the original format somewhat (use of farmer groups, training frequency changes, a day perfortnight in which EAs were "on-call") but retained the programming concept and regular visitschedule. More recently, the government has insisted on a more responsive and participatoryapproach, and considerable extension staff time is now allocated to working with farmers to help themdefine their needs (para. 5.19).

4.14 The national extension project in the Philippines (Loan 1626-PH) aimed at establishingT&V to promote integrated services for both foodcrop and livestock production. This was rated asunsatisfactory. The concept lacked the commitment of senior Ministry of Agriculture staff, and anadministrative reorganization also hindered progress. Fiscal constraints limited implementation.Overall, the intended integration of services did not occur, and only a loose form of T&V

- 23 -

programming and staff training concepts was introduced.33 Concentration on the government's riceand maize production programs meant that only a small proportion of farmers were contacted.

4.15 The Malaysian project (Loan 1493-MAL) followed on earlier research and extensionprojects. After some delays and reticence on the part of officials, a substantially modified form ofT&V was eventually adapted to Malaysian conditions. The government's emphasis on treecrops andhigh-value crops (as opposed to the foodcrops traditionally receiving at least the initial focus of T&Vprograms) demanded modifications which were developed with capable consultant support. Thefrequency and times of service varied depending on farming activities; farmers in relatively formalgroups were the recipients of service; attendance incentives were provided initially through treecropsubsidy programs, subsequently through marketing information, and finally in coordinated groupmarketing-these were such major departures from the typical T&V model that it could not beconsidered T&V.

4.16 T&V was not used in an agricultural research and extension project in Vanuatu (Credit1407-VAN) nor in an agricultural support services project in Papua New Guinea (Loan 2276-PNG).Both projects were unsatisfactory. The former had design faults in not catering adequately for thevery traditional nature of most of the farming clientele, and in overestimating the absorptive capacityof the implementing institution in a small island state. The latter had serious design problems inoverestimating the political commitment and institutional capacity to implement the project, andsuffered from poor management and only partial implementation.

4.17 There were 11 component projects in the EAP sample: four rural development, threeirrigation/drainage, four perennial crop and three settlement projects. In Indonesian, the three settle-ment projects used T&V implemented by the national extension service, but intensive commodity-specific services were employed in two treecrop projects. In Thailand, the national form of T&V wasused for a rural development project but not for a treecrop project. The projects in Myanmar(irrigation) and Philippines (rural development) used a T&V model, but not those in China (irrigation)and Papua New Guinea (rural development).

4.18 The observations from the overview of EAP component projects in Annex 2 are asfollows:

(a) Regardless of extension method, it is essential to have relevant technology for thetargeted farming systems (Loan 1227-PNG). A deficiency in technology is morelikely in the less predictable environments for rainfed annual crops (as in annualcropping in Loan 1499 and 1604-INS, and in the Indonesian transmigrationprojects) than in irrigated or other high potential environments.

(b) In more difficult production environment, the constraints faced by farmers are notonly more limiting, but also are not as obvious to outsiders and not as uniform as inbetter circumstances. Services must be aware of, and responsive to, these circum-

33. Mr. D. Benor, specialist consultant to the Bank on T&V, does not consider the systems in the Philippines andMalaysia as T&V.

- 24 -

stances in developing recommendations, and must be adequately supportedoperationally if they are to be effective (Credits 919, 2288, 1958 and 1898-INS).

(c) Irrigation investments usually represent a favorable production environment,enhancing the likelihood of an extension program being associated with increasedproduction. As benefits in these circumstances are often dependent on a relativelyhigh use of production inputs, however, it is necessary that farmers have access tothese inputs (Loan 815-PH, Credit 838-BA, Credit 1261-CHA).

(d) Intensive integrated services to smallholders in which all the requirements forimproved production are provided should be effective, but the problem lies with thecost per recipient and so raises issues of replicability and sustainability (Credit 946-INS). Intensive (and relatively costly) services, however, may be judged necessaryfor particular disadvantaged groups, who often warrant some form of governmentinvestment (as well as service) assistance. In these circumstances, it is necessary toensure this intensive support is as effective as possible, which implies emphasis oncommunity involvement and meeting the perceived needs of the community (Credit929-TH).

(e) A high degree of community involvement in the development process is conduciveto effective adoption of technology (Credit 1261-CHA).

(f) Commodity-specific services which provide the required crop improvement packageare usually effective in improving production of that crop (see also SASexperience-para. 4.10(b)); this service, however, is often provided without anyattention to the remainder of the production system (Loan 2078-TH). This lack ofknowledge of the total farming system can be especially problematical in promotionof a specific commodity after the investment phase (which is often the primaryfocus of a commodity package), when management of the commodity has to beintegrated with the demands of other components of the system (Loans 1499 and1604-INS).

4.19 AFR Region. There were five free-standing projects evaluated in AFR (Annex 1, Table1). A Somali project (Credit 905-SOM) was implemented in the early 1980s, while the remainingprojects (Credit 1387-KEN, Loan 235-ZIM, Loan 2765-IVC and Credit 1669-RW) were implementedfrom the mid-1980s. The three projects in Somalia, Kenya and Zimbabwe were rated as satisfactory.

4.20 In Somalia, the extension project envisaged introduction of T&V in three regions withfavorable natural or irrigated conditions. This was established effectively in the irrigated zones andless so in the rainfed areas. The system suffered from a poor research linkage and limited relevanttechnology, especially for rainfed crop production. The government initially tried to expand the T&Vcoverage into a wider area, but this had to be abandoned due to lack of technology, of trained staffand of a supervision capacity under poor infrastructural conditions.

4.21 In Kenya, the national extension project followed an earlier T&V pilot scheme in twodistricts. It successfully introduced T&V into the intended 30 districts in a phased program andachieved a limited integration of crop and livestock production extension. A major concern was the

- 25 -

difficulty in funding the recurrent costs of the program. This was exacerbated by an unanticipatedgovernment policy of employing all agricultural school graduates which increased the salary load inthe service and reduced operational support allocations. A serious constraint to effective operationswas poor linkages with research which affected the availability of relevant technology. The initialdesign was modified to focus on farmer groups, and a very positive development was that womenfarmers were progressively incorporated in the service, especially through women's groups. AFRcommissioned an impact study on the extension investment in Kenya.3 Although this study hadsignificant data and analytical constraints (para. 3.4), it maintained that T&V extension had made asubstantial positive impact. The survey undertaken as part of the study indicated a mostly favorablereaction of farmers to the service, but also indicated that non-adopting farmers in 1990 considered themain reason for non-adoption of the recommended technologies was poor access to production. inputs(especially through credit) rather than lack of awareness of technologies. This implies reducingmarginal returns for a staff-intensive service with high recurrent costs which cannot overcome thesemore limiting production constraints; new (useful) technology which is relevant to farmers circum-stances must be forthcoming to warrant continued, regular visits.

4.22 In Zimbabwe, a national extension and research project was designed to improve produc-tivity in communal farming areas (Loan 2335-ZIM), and was to complement the government'sprogram of improved credit, input supply and marketing assistance in this previously poorly-servicedarea. The T&V model was not used, but improvement were to be made in staff housing, transport,training, technical assistance, and more systematic programming of staff visits to farms with enhancedsupervision and technical support. The project was rated as satisfactory in transferring extensionresources from the commercial farming to the poorer smallholder communal areas. At the outset in1983, however, a "top-down" approach which promoted technology that was applicable to thecommercial farming areas met with little success, due to this technology usually being inappropriateto local circumstances. This subsequently (1987) evolved into an improved approach which workedmore closely with farmer groups in setting technology targets in accordance with different resourcelevels. In some areas, by 1989 a farming system diagnostic approach was being introduced as a basisfor defining technology requirements. This desirable development, however, had not been introducedinto staff training programs, and at project closure there was hardly any technology relevant to thepoorest 30 percent of farmers, especially those with limited labor or draft power. A major problemfaced by the extension system (with approximately 1 EA per 700 farm households) was serious diffi-culties in meeting its recurrent costs, despite government's strong commitment to extension incommunal areas.

4.23 The Rwanda project (Gitarama Agricultural Production - Credit 1669-RW) was todevelop a T&V system in the Gitarama prefecture as a national model. This was not successful.Major modifications were made to the appraised design during implementation in the use of existingteam leaders at community development centers to train other farmers under "development contracts";this could not be considered as an example of T&V. Smallholders lost interest in this scheme, andmanagement erroneously responded by concentrating its packaged technology on an even smallernumber of farmers in "cells". This "top-down" approach was complemented by weak linkage withresearch, relative ignorance of farming practices and constraints, and inappropriate technology.Successful propagation of improved cassava and sweet potato varieties by the project, and their

34. Bindlish, V. et al. (1993) op. cit.

- 26 -

dissemination with the help of the extension service in the face of a famine threat, demonstrated thateven relatively poor institutions can be effectively organized for specific campaigns during crisis situa-tions when everyone appreciates the problem and all parties are mobilized to avert the crisis (see alsoBrazil para. 4.31(d)). A follow-on project hopes to build on the lessons from the Gitarama project.

4.24 The Cote d'Ivoire project (Loan 2765-IVC) was designed to upgrade extension in theregional extension agencies by introducing T&V and improving research/extension linkages. T&Vwas adopted but the project was rated unsatisfactory. The training of EAs, who were generally of avery low education level, was substantial, but this was not enough to permit services which weresufficiently responsive to farmers' needs. Improved arrangements for adaptive research and linkageswith extension were adopted, but funding constraints hindered their implementation and insufficient,relevant technology was available to suit farmers' circumstances. Farmer coverage through contactgroups increased from about 25 percent pre-project to a peak of 64 percent, but subsequently declinedto pre-project levels. The degree to which various factors were responsible for this deterioration werenot quantified, but contributing factors were: significantly reduced funding affecting the regularity ofservices, reduced financial viability of some of the targeted crops (focus was on coffee, cocoa, cottonand foodcrops), an insufficiency of new relevant messages, and organizational changes whichintroduced coordination problems. 35

4.25 Rural development projects formed a major part of the agricultural portfolio in AFR inthe late 1970s and 1980s. Consequently, 22 of the sample of 25 component projects were in ruraldevelopment; two others related to perennial crops and one to irrigation/drainage. Half of the 22rural development projects used the T&V method, while the others used less programmed but moreintensive services. One of the rural development projects was in Kenya, which also had a free-standing T&V project; this project started with intensive traditional services but subsequently movedto a T&V method.

4.26 The observations from AFR component projects in Annex 2 are as follows:

(a) Regardless of the extension system used, there must be institutional commitmentand a reasonable level of management capability if worthwhile results are to beobtained (Credit 776-CAM, Credit 1193-ZAM, Loan 1667-NIG, Loan 1854-NIG).

(b) T&V extension used to complement other production services usually appeared tobe associated with some positive effects; the outcomes, however, were limited bybudgetary support constraints (the Nigerian ADPs, Credit 959-KEN, Credit 1343-MAL, and Credit 1128-SIL), weak research-extension linkage, poor appreciation offarming system characteristics with consequent problems in the relevance of recom-mendations (the Nigerian ADPs and Burundi projects), and often a "top-down"nature in the services offered.

35. A subsequent project attempts to address many of these issues including restructuring of responsible agencies.staff reduction, and the adoption of a farming system perspective in formulating technical recommendations.AFR has reported that there has been significant progress in the new project in agreement to staff retrenchmentand merging of some services, and attributes this to the experience under the earlier projects.

- 27 -

(c) The same sorts of constraints - budgetary deficiencies, ignorance of the farmingsystems, lack of attention to economics, unavailability of appropriate technology -often affected the non-T&V, intensive extension programs to an even greater extent,so that their impact was markedly reduced in reaching the majority of the small-holder population (the earlier Nigerian ADPs, Credit 1193-ZAM). A particularproblem in some cases was that reasonably good services were offered to a smallnumber of clients who were creditworthy but who had better resources and wereatypical of the majority, while the majority received negligible extension services(Credits 823, 857 and 1183-MAL).

(d) Experience with commodity-specific programs was consistent with that of SAS andEAP regions (paras. 4.10(b) and 4.18(f))-need for comparative advantage inproducing the commodity, and for an appreciation of the interaction of thecommodity with other components of the farming system (Credit 786-LIB). Wherecommodity staff are expected to provide an expanded technical advisory service,intensive training is likely to be required to give them theoretical and practicalknowledge of farming other than the specific commodity with which they have beendealing (Credit 1013-Burkina Faso, Credit 946-TOG, Credit 1127-BEN).

4.27 LAC Region. There were two free-standing projects in the LAC sample, one in Peru(Loan 2150-PER) and the other in Brazil (Loan 1568-BR).

4.28 The Peruvian project was well-designed and attempted to integrate research and T&Vextension into integrated services from a single institution. The project was unsatisfactory due to anbreakdown in the system associated with waning commitment by government, major fundingproblems in a deteriorating economic environment, and political change. Before this occurred, how-ever, the T&V system was faced with relevance problems in its technical messages for many areas, a"top-down" orientation and poor results with the use of contact farmers to spread technical messages.The change in institutional arrangements did not provide the opportunity to see if these constraintswould be overcome through system changes and progressive improvement in integration of researchand extension services.

4.29 The Brazil project was designed to support the national extension service (EMBRATER)through which federal and donor funds flowed to individual state extension services. The concept wasto develop extension methods which would expand services to more smallholders in the North, North-east and Central-west than was possible under the typical supervised credit projects (as in N.E.Brazil-para. 4.31). Twenty-five state services were involved. The project was rated as satisfactory:infrastructural investment to support extension was successfully completed; vehicle procurementreached 80 percent of target and mass media equipment purchase was substantial; pre-service and in-service training were extensive; and about 3,700 additional staff were recruited to expand extensioncoverage in the states. T&V was not utilized, but EMBRATER encouraged the development of agroup approach to extension, and heavy use of small demonstrations.36 Impact was constrained bylimited state budgets to support recurrent costs (despite support through EMBRATER), inadequate

36. A follow-on project (Extension I in 1986) also promoted the use of a programmed schedule of visits to groups;due to the dissolution of EMBRATER, however, and the transfer of implementation responsibilities toEMBRAPA and, subsequently, to the Ministry of Agriculture, this project has made limited progress so far.

- 28 -

research-extension linkages, lack of relevant technology in many of the common resource-poorfarming systems, and a lack of understanding by EAs of the socioeconomic factors affecting farmerdecisions. These deficiencies were obvious in the area development projects in N.E. Brazil approvedafter 1982 (para. 4.31(b)).

4.30 There were 20 component projects in the LAC sample. Brazil dominated with 11 ruraldevelopment projects, followed by Mexico with five credit and three rural development projects, andChile with one credit project which used subsidized private sector extension services to commercialsmallholders. None were on a formal T&V model, although the Brazilian rural development projectswhich commenced from the early 1980s had programmed activities and used some of the extensionprinciples that are also used in T&V, as promoted under the national free-standing project (para.4.29). All were personnel-intensive schemes.

4.31 The observations from the overview of LAC component projects in Annex 2 are asfollows:

(a) Integrated and intensive technical assistance complementing input supply andadequate investment and production credit in appropriate production programs, islikely to be effective in improving income for the recipient clientele; such services,however, are relatively costly per smallholder client. In addition, as was evidentfrom experience in other regions, smallholder technical assistance/credit programstended to be directed at the more progressive farmers with better resources. Thismeant that in public sector schemes, the normally scarce resources are concentratedon a limited number and often detract from attention to the households with lessresources (Mexican credit projects, Loans 1553 and 1945-MEX, and earlier N.E.Brazil rural development projects).

(b) In expanding intensive technical services from a limited clientele to a broader, lessintensive coverage, satisfactory results are unlikely without a knowledge of thetargeted farming systems to define technology requirements, effective linkages withresearch to ensure the development or adoption of appropriate improved technology,enough staff trained in not only the theoretical but also the practical aspects ofproduction, and adequate funding to support operations (later Brazilian ruraldevelopment projects, and Loan 2016-BR).

(c) Higher potential production environments (existing, or created through the invest-ment), are conducive to good results from integrated services which includeextension (Loans 1553 and 1945-MEX). At the other end of the productionpotential spectrum, however, the severity of constraints may warrant considerationof using public funds in poverty reduction programs other than agricultural exten-sion, unless public investment which is able to substantially improve the productionenvironment can be justified (later N.E. Brazil projects).

- 29 -

(d) Well organized campaigns to address widely recognized problems can avoid manyof the constraints faced by "permanent" services and be a very effective use ofpublic sector resources (OED study of N.E. Brazil area development projects).37

(e) Provision of extension services by the private sector can have the advantages ofbeing highly professional, attaining good results and reducing the public sectorservice cost. The clientele, however, must have the financial resources (i.e., be incommercial production with significant cash income) to pay for such services,although governments can provide subsidy incentives to promote the development ofsuch schemes (Loan 2481-CH).

4.32 ECAIMNA Regions. There was only one completed free-standing project in this region(Credit 1557-YEM). 3 1 It used a T&V model and was rated satisfactory in its PCR. It did faceconstraints in recurrent cost funding support and its potential impact was inhibited by problems in thesupply of fertilizer to farmers. It was unusual in that it had a high proportion of university graduatesas extension agents. There was no reference to any problem of poor linkage or communicationbetween research and extension staff.

4.33 Five credit and four rural development projects made up the component sample. Fourof the credit projects involved intensive public sector extension and credit to producers of high-valueexport crops in Greece, Morocco, Turkey and Cyprus; extension was unsatisfactory in the latter twoprojects and was judged to be only partially effective in Morocco. The extension under a fifth creditproject (Loan 1603-POR) was directed at large collectives as well as medium and small private farmsand was not very effective. T&V was used in the four rural development projects in Turkey,Morocco, Portugal and Yemen and in each of these extension was rated as having met at least someimportant objectives.

4.34 The observations from ECA/MNA component projects in Annex 2 are:

(a) In commercial farming systems based on high value export crops, farmers must takeadvantage of the most up-to-date technology and practices in commodity production,processing and marketing if they are to be competitive. This requires a highlyprofessional, economically-oriented service which is better performed by the privatesector as individual consultants or as specialists from producer marketing organiza-tions. Rarely would there be any justification for use of scarce public sector fundsfor this purpose (Loans 1988-CYP, 1588-GR, 1967-TK and 1757-MOR).


38. In April 1994, a draft PCR on the Agricultural Extension and Applied Research Project in Turkey (Lu. 2405-TU)was produced by ECA. Sixty-six percent of sampled farmers in the project area reported that they consideredextension had improved under the project (with a T&V model). There was some casual association betweenincreases in production and extension, especially in horticultural crops, which offered attractive economicopportunities. However, there was no significant difference between project and non-project provinces in the rateof change in wheat yields and the reduction in fallow area, which were probably the two most important changesin field crop agriculture in Turkey in the 1980s. Sustainability of the T&V system was questioned in view ofbudgetary limitations and the heterogeneity of production systems in Turkey.

- 30 -

(b) The potential impact of a T&V extension investment cannot be fully realized unlessstaff are competent in practical production technology, are supported by an effectivetechnology generation system, and are adequately funded to enable planned fieldoperations (Loan 2094-TK, Loan 2175-POR and Credit 1067-YEM).

5. FINDINGS

THE EX-PosT FINDINGS

The performance of the extension investments according to ex-post evaluations asdescribed in Chapters 3 and 4 permits a number of conclusions regarding not only the outcome ofthese projects but also the factors involved in performance. The main findings are as follows:

5.1 The Bank's support for extension in free-standing and component projects has substan-tially increased the attention given by governments to agricultural extension services, especially in thesmallholder farming population. Accepting the premise that effective extension services canaccelerate the adoption of improved technology, this is consistent with the Bank's and the commonly-expressed borrowers' strategy of improving agricultural productivity to contribute to economicdevelopment and the alleviation of poverty.

5.2 The focus of most investments on smallholders and improvements in the organizationand programming of services has enabled increased coverage of smallholders by public sectorextension services. Although there is scant empirical evidence documenting the extent to which thisincreased contact has contributed to any observed production increases over time, observationssupport a conclusion that the improved farmer coverage has been associated with an increased rate ofadoption of a number of significant technologies.

5.3 The effectiveness of an extension service was markedly reduced if its operational supportcosts were not adequately funded. Although extension budgets were not high relative to agriculturalGDP (0.2 to 0.4 percent of agricultural GDP in most countries), the majority of borrowers hadserious difficulty in meeting the recurrent cost expenditures of national extension services, to theextent that in many cases the sustainability of the instituted systems was in doubt. Fiscal realities andcompetition for limited funds meant that public extension relying on a relatively staff-intensive systemto provide what is usually not considered as a priority service suffered financially; 39 this, despite a

39. Uncertainties regarding sustainability usually involved reticence by central funding authorities to adequatelysupport the service, rather than its rejection by management of the implementing institutions or by centralauthorities. Nearly all of the national or state extension systems supported under the Bank projects havecontinued to-date, albeit with major modifications in some cases, and with continued external support in themajority of cases. There is usually no question about eliminating the extension service: extension existed beforethe Bank projects and is generally considered by governments as a necessary public service for smallholders; thecommonly used T&V management system suits most public sector institutions in which it has been introduced;and the T&V system is usually considered an improvement on earlier models which were regarded as very poorperformers. The pressure is for substantial reduction in recurrent costs and to make the extension service muchmore efficient and cost-effective.

- 31 -

recognition in the extension literature that sound extension investments usually give positive economicreturns.40

5.4 An insufficiency in technology which was relevant to farmer circumstances was a majorconstraint in many services, and was particularly prevalent for resource-poor environments.Regardless of the extension system used, a supply of appropriate technology is essential if extensioninvestments are to be worthwhile. This assumes even greater importance when the investment aims toestablish a "permanent" public sector extension service with substantial overhead in salary and othersupport costs. In these circumstances, without a continuous supply of technological improvements,the marginal returns on the cost of continued service quickly declines as the available technologiesreach a reasonable degree of acceptance in farming communities and new, relevant technologies arenot forthcoming. 41

5.5 The lack of available relevant technology was associated with an inability of appliedresearch institutions to develop or adapt technology which was appropriate to the constraints andpotentials of many of the common farming systems.4 2 This was associated in most cases with a

40. It is not suggested that an awareness amongst funding authorities of published accounts of extension benefitswould automatically generate more favorable funding. Even in a technocratic planning environment (which is notalways present), an inability to quantify all benefits and costs in economic terms for competing programs usuallymeans that other more tangible factors affect decision-making (degree to which a program is "essential', politicalor social consequences, or implications for continued recurrent cost financing).

41. Assuming that extension has been associated with increased knowledge and/or skills in the community, much ofthe accelerated adoption will have resulted from farmer-to-farmer contact This process occurs continuously at nopublic sector cost. When technologies become widely established, the incremental benefits to be obtained by anextension service promoting those technologies are markedly reduced; the impact of the farmer-to-farmerdiffusion process must be enhanced as the proportion of those with knowledge to those without knowledgeincreases. Continued frequent visits to promote standard technologies which are not relevant would not, ofcourse, be cost-effective.

AFR, in responding to this report, indicated that empirical evidence is not presented to validate the assertion ofdiminishing returns to services in the absence of new technologies. PCRs and PARs provide ample evidence ofproblems caused by lack of appropriate technology in repetitive and/or irrelevant messages. These are reflectedin the "Basis for Recommendations" section of Table 1. It is true that there is no empirical analysis to link thiscondition to reduced marginal returns. Nevertheless, OED considers the logic noted in the earlier paragraph ofthis footnote is especially strong.

In farming situations where pest control requires continued vigilance, however, frequent or strategic visits by awell trained extension agent can be appreciated and possibly be cost-effective if pest control is within the meansof the farmer. Unless they have been exposed to an effective educational programs, smallholders tend to needcontinued assistance in pest control, even if there is little to offer in the way of other relevant technologies. Thissituation, and the environmental and human health hazards of inappropriate use of pest control chemicals, suggestthe need for farmers or others to be specifically trained as pest observers, and to have well trained advisors onintegrated pest management (IPM) to educate extension staff and farmers on IPM. Indonesia has recently under-taken a large IPM training project with smallholders in its higher potential areas; this is an educational programwhich will progressively deal with communities, rather than a technical message program.

42. Although linkages were usually planned, and, to varying degrees, regular meetings between research andextension staff and joint "on-farm" trials were organized, rarely did this lead to a better understanding of theconstraints and potentials of many of the major farming systems by research and extension personnel.

- 32 -

poor understanding by researchers and extension staff of the circumstances of farm households (i.e. alack of farming system perspective) and poor linkages between researchers, extension staff andfarmers.

5.6 For smallholders in rainfed annual crop farming with significant unpredictability inproduction conditions, there were many instances of poorer apgarent performance of extensionservices in attaining adoption of "recommended technologies". This was linked to variousfactors: available technology being more suited to more favorable (less risky) production conditions;a strong sense of risk-aversion by smallholder farmers in these conditions in which production of mostof their subsistence food requirements was often a primary concern; and a lack of access to resourcesto adopt technology, partly associated with their marginal economic status in a difficult farmingenvironment. The reporting of poor adoption of "recommended technologies" under these circum-stances, of course, implies that technologies relevant to the farmers' circumstances (if there were any)were not being recommended, which represents a failing of the service.4

5.7 The concentration on face-to-face contact in most services necessarily was associatedwith increased participation of farmers in extension activities; on-farm technology demonstrationssimilarly promoted participation. Little attention, however, was given to involvement of farmers inproblem definition, problem solving and extension programming. In the rare cases whereparticipation was emphasized, results were very positive (Annex 2, paras. 6, 21 and 23).

5.8 In most developing countries there is a traditional "top-down" culture in interactionbetween government (through public servants) and the rural population. Extraordinary measures arenecessary to change this culture. Without its resolution, reliable farmer feedback cannot be obtained,and extension tends to be based on standard recommendations which cannot satisfy the informationneeds of many farmers in the diverse farming systems in the agricultural community. Thischaracteristic was prevalent in Bank projects, and had special significance in large programs cateringto diverse clientele. When a "top-down" attitude is combined with a poor appreciation of the inter-actions within farming systems and of production economics and risks, the relevance problem isexacerbated. Messages which happen to be relevant are adopted, but irrelevant messages arelegitimately rejected by farmers, representing a wasted extension effort.

5.9 Staff quality was a major constraint in nearly all of the free-standing projects and inothers which required very large numbers of staff. Most of the free-standing projects provided forcontinuous technical training for EAs, which was especially important due to their usual relativelylow level of education and basic agricultural training. However, serious difficulties occurred inhaving insufficient subject matter specialists who had adequate technical and practical experience toparticipate effectively in regular staff training in the state-wide programs, so that training was often

43. The converse of this observation is that extension performance appeared to be better in higher potentialenvironments. What has been observed, however, is that adoption of technology has been greater, but the extentto which this might be due to extension or other factors is not known.

44. There is little argument that if farmers become aware of technologies (or modifications in use of resources) whichare relevant to their circumstances and can improve their farm production/welfare, these changes will mostly beadopted (Box 1).

- 33 -

overly theoretical or too general. Training in communication theory and techniques was also usuallydeficient. In about one quarter of free-standing project reports, the evaluators specifically referred to

the relatively low level of basic education of the EA staff as an impediment to the adoption of a more

analytical and responsive service. The large number of personnel with limited capabilities was partlya consequence of adopting a staff-intensive program over a large area (state, nation). In five free-standing projects, evaluators recommended that an improved extension program (with its associatedcosts per operating unit) should only be expanded at the rate at which suitably trained personnelbecame available to implement it efficiently.

5.10 It was apparent that unless there was solid commitment by senior officials of central andimplementing agencies to implement the improved service, operational problems could be expected innot only funding support, but also in approvals for recruitment, procurement and establishingsatisfac-;ory terms of service. Problems of less than full commitment were noted in nearly half of thefree-standing projects.45

5.11 In general, monitoring was weakly developed in the free-standing projects. Most, how-ever, had some means to monitor the mechanics of implementation of the extension methods. Fewerhad survey systems to assess awareness of the adoption of recommended technologies. Even whenadoption surveys were done (e.g. many of the Indian state projects), very little use of the results wasmade byroject management, apart from the move from contact farmers to working with farmergroups. For component projects, with rare exceptions monitoring tended to be no more thanrequired for the mechanics of the system operation, whether this be in credit, investment or inputdelivery.

5.12 T&V system. Over 50 percent of all reviewed projects, including 90 percent of the free-standing projects, used the T&V management system. This model was, therefore, a major factor inthe positive elements of enhanced attention to extension and to coverage of smallholders. It cannot,however, be divorced from many of the above-mentioned serious concerns regarding program

45. In two unsatisfactory project (Loan 1626-PHL and Credit 1669-RW), the lack of commitment to the improvementprogram was given as a major cause of the poor performance.

46. Monitoring was reported to be very weak in countries like Bangladesh and Kenya, despite major T&V programs,and was weakly developed in most others (e.g. Nepal, Pakistan, Somalia). Awareness and adoption surveys,however, were noted as the main reason for changing from a contact farmer to a group approach in a number ofcases (e.g. Sri Lanka, Thailand, Somalia and some of the Nigerian states). In India, monitoring and evaluation(M&E) units were established in all projects, representing a very large investment. In India in general, wherestate departments of agriculture were well managed and appreciated the potential of M&E, staff recognition andfunding support were given to the M&E units, with positive results (e.g. Karnataka, Gujurat, Rajasthan andHaryana); elsewhere the units were not very useful.

- 34 -

efficiency and sustainability.47 This is evidenced in the reference to problems in free-standingprojects in para. 3.12 and Table 1. Specific factors concerning T&V are elaborated below:

(a) Although governments incurred significant recurrent costs for extension systemsbefore the introduction of T&V, the projects increased the operational support costper extension unit and maintained or increased the number of units. The additionalunit cost was usually justified to make the unit more effective. The large number ofunits, however, meant that the total system made substantial demands on therecurrent budget.4

(b) T&V exacerbated the deficiencies in lack of appropriate technology (on acontinuous basis), as its scale and intended coverage meant that a large array ofproduction systems and circumstances had to be catered for, including the moredifficult production environments.

(c) T&V increased participation in the limited sense of farmer contact and throughfarmers' test plots, but the scale of the program and traditional "message-centered"format was not prove conducive to intensive farmer involvement in problemdefinition and resolution. Similarly the hierarchical framework of T&V suited the"top-down" nature of most public services and there was insufficient incentive todepart from this culture in formulating recommendations. 49

(d) Again, the scale of the state or national T&V programs obliged the use of a largeproportion of staff at field, and, more importantly, support levels who were noteasily trained to provide responsive technical services.

(e) The degree to which T&V was accepted as an extension system in so manycountries was remarkable. It did fit in well with the public service culture in mostcountries (see Box 4 and Box 5). However, in projects which relied on massivestaff transfers from other programs, a lack of commitment to the T&V technology-focused, programmed and supervised mode of operation often occurred at the level

47. Since 1989, OED has rated projects as to the sustainability of their achievements. Due to the dominantinstitution-building objective of free-standing extension projects, this rating usually refers to the likelihood of thedeveloped services being sustained. For free-standing projects sustainability was rated as "likely" in only 3 of 17projects (18 percent), with 71 percent being "uncertain", and in 2 cases (12 percent) as definitely "unlikely"(Peru, because of a breakdown in the institutional arrangements and funding, and Rwanda, due to very poorimplementation with unsatisfactory outcome).

48. Although these demands and the funding deficits raised questions of sustainability, this does not imply that T&Vprojects were less sustainable than many of the more intensive, localized projects. The issue is that the nationalT&V projects were designed as institution-building projects so that institutional sustainability expectations werehigh.

49. The "top-down" phenomenon was specifically highlighted in over half of the free-standing T&V evaluations(Table 1).

- 35 -

Box 4: REASONS FOR WIDESPREAD USE OF T&V IN BANK PROJECTS

The relatively broad acceptance and continued use by the Bank of the T&V management system and itspromotion amongst borrowers in preparing extension projects can be related to a number of factors:

1. The earlier projects using T&V were associated with favorable production conditions (usually irriga-tion) and were able to capitalize on the spectacular benefits of HYV and higher input (especiallyfertilizer) use in these environments. Although the extent to which the extension system acceleratedthe adoption of these technologies (which, for example, was well underway before T&V in India) wasnot documented, it was assumed that extension contributed to the relatively high rate of adoption andproductivity gains obtained.

2. An unsatisfactory rate of adoption of improved technology was assumed to be substantially linked topoor performance in existing public extension services, implying that substantial improvements couldand should be made to permit better use of public extension funds.a T&V appeared to be addressingmany of the identified problems, and the early projects did show that it could result in enhancedgovernment attention to technology transfer.

3. Bank staff assumed that there was always appropriate technology available which was not beingadopted, and that a better organized delivery service would remove a key constraint to increased agri-cultural productivity.

4. The T&V system offered an apparently convenient vehicle for Bank intervention - supporting publicinfrastructure, vehicles, training, and technical assistance, and rather mechanical preparation basedon farming population service ratios, hierarchical staff ratios, and a set training/service method whichcould be monitored. It was assumed that all types of smallholder communities, farming systems, andproduction conditions would be more responsive to the T&V method than to the existing services, sothat there was no special need to look further into the different farming situations nor to the efficacyof the method in meeting production and welfare objectives in these situations.

5. In the early 1980s in AFR, major problems became evident in the widespread program of areadevelopment projects, and extension projects using T&V were seen as a potential, timely substituteto support agricultural sector growth, especially if aligned with economic and sectoral structural adjust-ment which should create an improved economic production environment.

6. A prominent and influential long-term consultant to the Bank (D. Benor) vigorously promoted theT&V concept within Bank management and with potential borrowers, convincing them of its benefits.The progressive spread of T&V from SAS and EAP to AFR coincides with the concentration ofpromotional effort by this remarkable individual.

Acceptance by borrowers can be partly related to T&V promotion by the Bank, and to the borrowers' ownperceptions concerning factors 1 and 2 above. Another important reason for borrower acceptance of T&V,however, is likely to have been its hierarchical, organized structure which was compatible with thedirected, 'top-down' mode of operation in most public services. This factor may not only be associatedwith its acceptance, but also with the difficulties in developing national T&V into a service which isresponsive to farmer needs. There is not the same incentive for the bureaucracy to make these adjustmentswhich represent a major departure from tradition.

a. As early as 1983, however, in a World Bank/UNDP symposium held to review the Asian experience withT&V (Cernea, M. et al. 1983), the conveners cautioned that 'e"ension is only one of a nwnber offactorsthat contribute to increased farm productivity, and not one of the essential ingredients-improved technoloV,available inputs, and attractive markets.

- 36 -

of middle management and field staff, at least in the initial phases of the project(especially in some of the Indian states). 50

(f) In the earlier, completed projects reviewed in this sample, the T&V project designswere frequently of a blueprint nature. They commonly had exclusive attention toface-to-face public sector extension services, and relatively standard staff:farmerratios, staff hierarchical systems, visit frequency, etc. The need for much greaterattention to individual area and country circumstances at the project design stagewas emphasized in nearly 40 percent of ex-post reports: fiscal, institutional,technology stock and farming system conditions should have been taken intoaccount in formulating extension components and approaches. 51

(g) Despite these criticisms of the T&V projects, many of the principles of the T&Vmanagement system are widely accepted as good practice in any face-to-faceextension services: programming of activities, responsibilities focussed on techno-logical improvement, continuous training of staff, supervision/accountability,effective research-extension linkage, and farmer feedback. Some of these principleshowever, were not adequately implemented in projects (poor training, weak linkageswith research, ineffective feedback). As noted above, this deficiency can be partlylinked to the large scale of the programs involving high numbers of staff withlimited abilities and experience. In the case of weak linkage with research,however, an additional factor was frequently an established research culture whichwas not accustomed to being responsive to farmers' needs in planning researchprograms.

(e) Other T&V practices, such as limiting recommendations to a number of predeter-mined messages and the use of contact farmers to spread messages have receivedmore criticism. The criticisms of the practice of a predetermined message programwere not related to the concept of a limited focus for each visit, but to an inabilityof the system to adjust recommendations to circumstances of farmers or farmergroups. This problem was connected to the "top-down" characteristic of most

50. In many cases where staff were transferred from a role in which they controlled inputs to a strictly technologytransfer function, they reacted negatively to the apparent loss of importance and prestige (and, possibly, "rent')associated with control of inputs to a limited clientele. Consequently, this reticence could be expected, as itwould have been virtually impossible to convince all staff of the worth of a new system before it wasimplemented.

51. Despite the relatively standard design format, it was common for significant modifications to be made duringproject implementation-change from the contact farmer principle to working with groups, adjusted visit andtraining session frequencies, and broadening of subject matter coverage-although all retained the principles ofprogrammed farmer visit schedules and regular training (see Chapter 4). These developments should not be seenas a criticism of T&V, but rather as expected adaptations of procedures in line with experience. More recently,designers of T&V projects have given greater recognition to local circumstances in formulating projects, but moststill adhere to the concept of having a nation-wide program with only minor deviations from a uniform set ofpractices (see paras. 5.23-5.27).

- 37 -

services and the poor development of a farming system perspective. 52 One halfof the reports were critical of the poor effectiveness of the contact farmer conceptand a change to working with farmer groups was widespread. 53 The poorperformance of contact farmers was likely associated with the selection of farmerswho were not representative of the resource circumstances of those they weresupposed to influence. This is not so much a criticism of the contact farmerconcept but of a lack of understanding by extension staff of the importance ofresource and other socioeconomic and cultural circumstances in farmer interactionsand decisions. Unfortunately, in most cases this same ignorance carried through tothe use of farmer groups as a means to overcome the poor results with contactfarmers. There was commonly little attempt to identify groups of farmers whomight be able to relate to each other in objectives, resources and potentials (oftentermed a "reference group"), or patron-client and other community relationshipswhich could influence decision-making.- This reduced the potential to usecommon characteristics and interests to enhance effectiveness by participation ofgroup members in problem definition and in derivation of solutions throughtechnological innovation or changed resource use, and by achieving group owner-ship of demonstrations. 55

5.13 Integrated rural development projects sometimes used the regular state extensionservices, but in most cases provided for extension services which were more intensive than state-wideprograms (using T&V or other methods). They also usually aimed to ensure the provision ofproduction inputs and credit to facilitate adoption of recommended technologies. Where relevanttechnology was available and complementary services were provided, extension results wereapparently favorable, as could be expected. Many of these projects, however, were in difficultproduction environments which incurred problems of relevant technology. They also were sometimesnegatively affected by deficient complementary services due to poor coordination of, and reliance on,other public agencies, thus removing any potential advantage over general extension services.

5.14 In most rural development projects, it was assumed that extension services would becontinued after project completion. This often incurred major funding difficulties when services were

52. Although not highlighted in reports, the combination of a 'simple message" and 'top-down" orientation has meantthat in most T&V services there has been limited farmer education in the more complex aspects of farmoperations, such as integrated pest management (IPM) or soil and water management. Recently in AFR, attemptshave been made to promote soil conservation through T&V, reportedly with some success. In Indonesia, thegovernment has adopted a more intensive educational approach for IPM in villages, departing from theirtraditional T&V system in these cases.

53. Most projects either adopted a group approach or experimented with the use of groups when the problems withcontact farmer influence became apparent.

54. See OED Report No. 8808, June 1990. (PAR on six Indian projects.) paras. 115-117 and Annex 4; and FAOInvestment Centre (1992). Technical Paper No. 9.

55. Quite often groups which did form could be classed as reference groups, but this was usually more by accidentthan design. Even then, there was no attempt to use structured interaction techniques to improve effectiveness(nor training to do so).

- 38 -

even more staff-intensive and more demanding in operational support funds than the regular publicextension services. Although this might lead to a conclusion that it would be better to use sustainableservice intensities in this type of project, particular circumstances should dictate requirements: theeffective "Plan Puebla" model in Mexico (Annex 2, para. 23) had an intensive service for targetedcommunities, but it was intended to progressively move from one community to another with well-trained staff; in the hill tribe area in Thailand (Annex 2, para. 6) the intensive (and effective) partici-patory service could be justified as a "permanent" service because of the government's specificdevelopment objectives with this marginalized target group.

5.15 Commodity-specific extension programs were staff-intensive and attempted to ensure thatthe package of inputs needed for improved production was provided. They usually were effective inattaining their objective with the targeted commodity, which was very often a perennial crop linked toestablished marketing or processing outlets. It was apparent, however, that there must be a compara-tive advantage to the farmer in producing the commodity if the desired response is to be obtained. Inaddition, the need for the specialist services to be aware of the interaction of the commodity withother components in the smallholder production system was demonstrated. Where a specialist servicewas expected to expand into general extension advice for the total farm operation, training to enhancecompetence of previously specialist staff and the overt support of service management are required ifthis expanded role is to be effective.

5.16 These finding would support the use of an integrated commodity-specific service whenthere are poor prospects for farmers to have access to all the inputs and technology required toproduce a priority commodity.5 As the growing of such commodities usually (but not always)implies commercial production conditions, cost recovery in some form should be anticipated if theseintensive specialized services are provided by a public or parastatal entity. Private entities would, ofcourse, obtain payment for such services. As the farm commodity becomes more specialized,moretechnologically advanced, and of higher value (e.g. vegetables, and export fruits and flowers), thetechnology transfer function should move to the private sector.

5.17 Apart from systems in which parastatals made deductions from commodity payments, orcredit agencies included an interest rate spread to obtain some cost recovery for services, there wasonly one (successful) example of private sector services by consultants being used to providesubsidized extension to commercial small farmers (in Chile - Annex 2, para. 24).57 There is littleargument that private technical assistance is likely to provide highly professional, economicallyoriented, relatively intensive and largely effective service; nor can there be any disagreement with thedesirability of passing the cost of technical service to the farmer clientele in order to reduce the(frequently unsustainable) fiscal burden of supplying public services. It can also be argued thatrelevant advice, by definition, means that it has an attractive benefit:cost ratio and therefore payment

56. The priority could relate to a government program to enhance exports, or the commodity may be 'required" by aprivate or public processing/marketing investment.

57. None of the reviewed projects elaborated on the use of services from private commercial input suppliers ormarketing agents. It is obvious, however, that every effort should be made to maximize the contribution of theseagencies to technology transfer, preferably in close interaction with public services to afford some protection torelatively uneducated smallholders against inappropriate commercial advice. This offers an opportunity to reducepublic sector costs while improving farmer coverage i.e., increasing extension efficiency.

- 39 -

for the information or skill acquisition is justified. The potential client, however, must have sufficientcash resources and be willing to pay for such services, implying commercial farm production withsignificant cash income. This is still not the case in the poorer smallholder subsector of manydeveloping countries, so that some public services for the private good of smallholder families islikely to be needed and justified in the foreseeable future on economic and poverty alleviationgrounds. 58 Change in the agricultural sector is likely to be rapid, however, in the coming decade,especially in the middle-income developing countries, so that any public extension service should beresponsive to the need to move to a system of payment for services as this capacity develops. Thiswill reduce fiscal costs and enhance the effectiveness of technical services through client ownership.

5.18 Extension services provided as an adjunct to projects with an agricultural credit focuswere usually effective for the clients who received the service.59 There was a strong tendency,however, for participation only by those smallholders with better resources, so that the majority offarmers were not attended (e.g. Malawi). This occurred whether the credit agency was private orpublic, as any serious financial institution must be interested in making loans of reasonable size (toallow cost coverage through interest-rate spread) to clients with an adequate resource base to reducerisks of repayment delinquency-a limited smallholder clientele.

5.19 The N.E. Brazil projects demonstrated that special extension campaigns to addresswidely appreciated problems (e.g., a disease or pest outbreak) in a limited timeframe can be veryeffective, assuming the required technology to address the problem is available. This is likelyassociated with an ability to avoid many of the factors which often negatively impact on extension-lack of relevant technology, funding constraints, lack of direction and a measurable output, lack ofsupport for the extension objectives by other public agencies and political representatives, andcomplex extension tasks requiring reasonably skilled personnel. Special, short-lived campaigns, how-ever, depend on the availability of suitable public sector personnel to respond to the crisis, and do notsubstitute for the development of a public sector institutional capacity to deliver effective extensionservices, should the latter be deemed necessary to enhance productivity in the smallholder subsector.

5.20 In the period of preparation of the reviewed projects, the importance of the role ofwomen in farming systems and associated implications for the extension program were not givenspecial attention. Consequently, this issue has generally not been highlighted in evaluation reports.However, the growing awareness of the need for special attention to women in farming householdsdid elicit mention of this issue in some of the more recent reports, and there were some cases ofextension beginning to work with women's groups (e.g. Kenya, Thailand). The need to specificallycater for women is receiving attention in most of the extension projects appraised over the last fewyears.6 Environmental issues also were not given special attention. Environmental concerns wereassumed to be covered within the objective of promoting sustainable agricultural development, and

58. Poor resource circumstances, however, should not prevent attempts to encourage farmer contribution in non-hterms so that their demands on, and expectations from, the service are increased.

59. This type of project usually ensured the delivery of the required inputs, at least to the clientele.

60. The extent to which now programs cater for women is often linked to the religious and other cultural values in acountry.

-40-

were not a significant subject in the reports. Similarly, the potential use of NGOs to complementpublic services was not topical in the design and implementation period of most of the projects andwas not addressed in the reports.6 1 As indicated in the Preface, many of these issues will bedeveloped further in the comprehensive review due in 1995.

5.21 The above findings do not support a contention that there is a single extension modelwhich has sufficient superior features to justify its uniform adoption for technology transfer in allsmallholder farming circumstances. Rather, they demonstrate that there are some key principleswhich determine the effectiveness of the extension services offered. Paramount amongst these is theprinciple that extension recommendations must be relevant to the conditions of the client, regardlessof the extension model used. This, in turn, means that the service must understand the characteristicsof the targeted farming systems and the factors which impinge on these systems,6 2 and must haveaccess to appropriate technology and to an effective technology generation/adaption capacity. Havingmet this basic tenet, methodological innovations such as programming and scheduling activities canenhance service efficiency and its efficacy in meeting objectives, as lon as a primary consideration intheir design and implementation is to be responsive to farmers' needs.6 1 The desired impact cannotbe obtained, however, unless a public sector extension investment is designed with full recognition ofother potentially limiting factors in fiscal, institutional and human resource constraints. Fiscalrealities in most developing countries dictate that priority must be given to containing recurrent costs,so that a service is not only economic in the true sense of representing the most cost-effective meansof achieving a particular result, but which is within an expenditure level which is able to be fiscallyand politically supported in the future. Just as the unavailability of a supply of new relevanttechnology will lead to diminishing returns on continuation of an extension service, so too will ascarcity of funds which limits the field activities in staff-intensive, face-to-face services.

5.22 In comparing the performance of completed projects in the 1980s with the findings andrecommendations of the 1983 OED report (para. 2.12), a number of observations can be made:

* The earlier contentions that the Bank's projects have increased the attention given totechnology transfer by governments and increased the coverage of the smallholderfarm sector were confirmed.

* The concern that sectoral, farming system, fiscal, institutional, technology stock,human resource and complementary service characteristics were not being sufficientlytaken into account, was not heeded significantly in project design and implementation

61. Realization of a need to more adequately address farmers' perceived problems, however, has led to some currentprojects (e.g. in Zaire and Bangladesh) attempting to make use of NGOs to enhance farmer participation indetermining the content of extension programs.

62. See also Antholt, C.H. (1991), Birkhauser D. et al. (1991) and Sims, H. et al. (1990) in which the authorsemphasize that awareness of new technologies is very often not the primary constraint to adoption of thetechnology by smallholders.

63. The apparent benefits of many T&V programs in successful irrigation schemes during the period of proliferationof HYV in Asia and associated greater use of inputs occurred when these principles were met: the farmingsystems were relatively well-defined; technology was available which had a high benefit:resource cost ratio; andresources and complementary services were available to make it very relevant to a large number of farmers in theirrigation scheme.

- 41 -

in the 1980s. The principle of ensuring that the extension system was relevant toexisting and anticipated circumstances did not receive adequate emphasis.

* The deteriorating fiscal situation of many developing countries in the 1980sexacerbated the problem of fiscal sustainability of the supported extension systems, sothat this issue is even more serious now than in the 1983 report.

* Concern that, despite many sound principles of the T&V method, it was being usedwithout adequate pre-project situation analysis, was generally not addressed in the1980s. This has led to a widespread adoption of T&V on a state or national scale,without sufficient consideration of the scale of services, of different methods, nor ofmajor modifications to T&V to suit particular circumstances. Some reasons for theinitial and continued widespread use of the T&V model in large national programs arepostulated in Box 4.

* The specific concerns in the 1983 report for developing cost-incurring extension net-works where there was insufficient technology, for having extension without sufficientresearch back-up and linkage, and for establishing large national programs based onstaff with very limited capabilities were not adequately addressed.

* The recommendation to give more attention to the important role of women in small-holder agriculture was gradually adopted in the late 1980s.

* The recommendation to concentrate monitoring activity on production results to linkthis to impact was generally not heeded. In view of the difficulty in linking cause andeffect for extension investments (paras. 3.3 and 3.4), it would have been inappropriateto use a large amount of scarce resources on measurements designed to provide datafor ex-post evaluation. The use of survey data for ongoing evaluation to enableresponsive management, however, could have been developed more extensively.

MORE RECENT DEVELOPMENTS

5.23 The nature of this product in the agricultural research and extension study (para. 1.2)does not permit a comprehensive analysis of the performance in ongoing extension projects. The finalreport will give more attention to current developments. However, in brief, the developments varyacross the Bank's six regions. In SAS and EAP, the attempts which were initiated in the 1980s tomake the dominant T&V model more suitable to local conditions have continued-changes in trainingand visit frequency,6 4 emphasis on farmer group contact, adjustments for tree crops in farmingsystems, broadening of the subject matter coverage, and experimentation with different degrees oflinkage with, or participation in, input supply services. Most countries have maintained some form of

64. The model of regular fortnightly visits over a number of years evolved in irrigated and higher potentialproduction environments with smallbolders in annual (seasonal) cropping systems, for which there was a stock ofimproved technology available or likely to be made available. In more marginal environments for which there isusually a smaller pool of appropriate technology available (or being produced), and in circumstances of pooraccess to complementary services (credit, inputs, markets), relevant measures to improve performance arelimited. Similarly, a dominance of perennial crops or livestock in the farming system dictate that adjustmentsmust be made in the programmed services.

- 42 -

programmed services in their state/national systems, and the majority appreciate the need for closeresearch-extension linkages (although this still has many imperfections in nearly all countries).Scarcity of recurrent cost funding is the most widespread, serious and persistent problem, followingclosely by poor quality of subject matter specialist support staff. Programs to upgrade the educationallevel of EAs are also underway in many countries in attempts to improve the capacity of staff toprovide more responsive services. Some countries are attempting to reduce recurrent costs, whileothers are trying to introduce greater farmer participation (e.g. Thailand and India)65 or adopt amore educational approach to technology transfer (e.g., Indonesia). The government of Sindh inPakistan is in the process of trying to make major changes to provide different types of services todifferent categories of farmers, with a reduced number of staff having better qualifications, andgreater use of mass media and collaboration with NGOs.6 China did not adopt T&V and hascontinued its concept of having extension staff accountable to groups of farmer clientele; in theongoing Agricultural Support Services Project (Cr. 2462-CHA), the use of farmer associations isbeing expanded to further link farmers to the technology development and transfer process.

5.24 In LAC, the T&V model is not widely used (although the programming of services iscommon in Brazil). Here, there has been more progress in involving the private sector in extensionfor commercial production by smallholders and in government contracting the private sector toprovide services to marginalized farmers (e.g., Chile), and in determining some degree of paymentfor public service at the individual or local community level (e.g., Colombia). Provision of servicesby farmer associations is also widely developed amongst smallholders in commercial production, oftenalong commodity lines with deductions from commodity prices paying for technical services. InECAIMNA there appears to be a realization that recurrent public sector costs must be contained, thatan effective farming system diagnostic capacity must be developed, and that project design must be

65. A review by the Bank's Country Department for India (Macklin, M. 1991) of the Indian extension investmentssince 1977 maintained that they had registered a worthwhile production response and had been appropriate forIndian conditions, but also highlighted most of the problems indicated in the OED audits of Indian projects.Recommendations were made for improvement in enhanced coverage of all socioeconomic groups and in thequality of services, and for adoption of methods to reduce recurrent expenditure demands and improve farmerparticipation. In 1993, a study was commenced by the government (with Bank project funding) of both theextension and research systems to develop pilot schemes in an attempt to address the key constraints of poorresponsiveness of both systems to farmers' needs, in preparation for a loan proposal. This has derived from agrowing realization in central government that (despite its apparent earlier impact-PAR author's addition) *majorinefficiencies exist in the Indian T&V extension system and the credibility and conviction of its social andeconomic impact has not been carried with policy and finance managers." In commenting on a draft of thisstudy, the head of the Indian review team (previously the Extension Commissioner in India) has agreed with thebasic analysis and recommendations, noting the characteristics which have tended to develop in India-'emphasison process rather than impact ... supply driven, with little pressure on, or incentive in, the system to perform ...poor understanding of farming systems with uniform and prescriptive dissemination resulting in irrelevance andloss of credibility with farmers ... concentration of technology in resource-rich rather than poorer farmers ...poor use of, and integration with, commercial services and useful NGOs ... need for understanding of groundrealities to design appropriate services ... system must be made to deliver in recognizable terms ... initialmarginal rates of return must be high when compared to the poor service system previously in place, but thisdoes not mean that continued use of the system after this initial impact is similarly attractive.

66. Eighty percent of farmers in the Sindh are sharecroppers on land for which many production decisions are madeby landowners; this was not taken into account in imposing a standard T&V model.

- 43 -

responsive to local conditions;67 the extent to which these principles are translated into actualpractice, however, remains to be seen.

5.25 AFR is currently the most active region in extension activities (para. 2. 10 and Figure 5).Here, the T&V model is very dominant. It has been common to pass through a two to three yearpilot phase before embarking on national programs, so that in 1993 there were 20 large-scale, free-standing T&V projects in Sub-Saharan Africa, in addition to extension components in other projects.On the positive side, the Bank's support has continued to (a) focus borrower and donor attention onthe need for enhanced technology development and transfer in this region which has more productionconstraints than other regions, (b) encourage the rationalization of many fragmented extensionservices within countries, and (c) increase coverage of smallholders by national extension systems. In1987 AFR realized the need to broaden its emphasis from an extension focus to cover the comple-mentary services and adopted an Agricultural Services Initiative. It also appointed extension servicespecialists in 19 countries to assist in project implementation (mostly from 1989). Recently, therehave been moves in the Sahel to link natural resource management operations with extension, whichwould introduce a more decentralized and participative approach in dealing with farmers, and rapidrural appraisals by joint research and extension teams have been commenced in Ghana to examineproduction constraints and potentials in production of specific commodities. All these developmentsreflect the massive effort that AFR is making to address poor agricultural sector performance throughinvestment in services, especially extension.

5.26 The AFR extension portfolio, however, is faced with many serious problems. Thesehave been highlighted by Bank staff and borrower representatives in a number of recent seminars andreports-regional workshops in Malawi (1991), Ghana and Cote d'Ivoire (1993), a review of theadequacy of staff appraisal reports on extension projects (1991), and a report on East African projects(1992). These reviews are summarized in Annex 4. The problems are not new nor peculiar to AFR.They are, however, exacerbated by generally more difficult fiscal, political, production and humanresource conditions than occur in other regions. Most staff appear to accept that the T&V programshave brought some rationalization into the use of existing extension resources in many countries in theregion (although concentration of efforts on T&V has meant that any alternative models using similarresources have not been tried), 68 but (b), despite ongoing attempts by dedicated field advisors toimprove efficiency, with some obvious successes, there are serious difficulties which have not yetbeen overcome in making nation-wide extension services efficient and sustainable in the circumstancesof individual AFR countries.

67. Zijp, W. (1991). Attempts to introduce a more analytical approach to project design have been made in Romanlaand Hungary.

68. This is undoubtedly one of the positive influences of the T&V national programs in AFR. This region, morethan any other, has been subject to the proliferation of numerous localized projects supported by a large array ofdonors. Althoughi many of these projects had a positive result in their area of influence during implementation,most were very intensive in nature (often with heavy expatriate participation), were not replicable andconsequently, had limited impact on a national scale while they did demand considerable counterpart support inscarce qualified human resources. The Bank's program in extension has enabled the development of a nationally-owned system into which various donors are expected to fit their support. The legitimacy of how these systemsare developed in various countries is another issue which is discussed below.

-44-

5.27 Overall, in all regions Bank staff and borrowers appear to be aware of the majorconstraints to increasing adoption of improved technology through sustainable public sector extensionservices, and most are trying to make adjustments in their extension systems to reflect at least someof these constraints. Plausible solutions, however,. in the large national T&V programs (which aredominant in Bank-supported project) are difficult to realize. This is evidenced by the experience inSAS where major efforts are still underway after 15 to 20 years to overcome the problem of poorresponsiveness of the services to the needs of many farming systems (para. 5.23 and its footnotes).Deficiencies and inefficiencies in the extension systems themselves, as well as factors exogenous tothe extension systems, limit the extent to which the ultimate extension objectives of increasingproductivity and alleviating poverty can be achieved. This is of special significance in the case ofAFR-scarcity of relevant technology, limitations on staff quality, severe fiscal limitations, weakdevelopment of the private sector in providing complementary services with a reliance on inefficientgovernment interventions, a poor credit delivery system to smallholders, and often political and/oreconomic instability-as the Bank is relying on extension investment as an important component of itssectoral support strategy. Such constraints severely limit the potential for large, national programsusing uniform extension practices to be efficient users of public funds in attaining technologicalchange in the targeted smallholder communities.

6. RECOMMENDATIONS AND ISSUES

RECOMMENDATIONS

6.1 This review did not identify an extension model which should be universally used by theBank to enhance technology adoption in developing countries. Rather, it supported a position thatconditions in each country should determine the scale and type of public sector extension services tobe developed. This implies that the Bank should place much more emphasis on the design of projectsto fit the particular circumstances of the borrower (fiscal, institutional and human resource) and of thefarming systems in the rural communities to be serviced, and must assess the availability ofappropriate, sustainable technology and the capacity for its continued generation.

6.2 Although project preparation is the formal responsibility of the borrower, the Bank, asthe dominant international development institution and the leading lender in the field of agriculturalextension, should give more attention to working with nationals in the upstream phase of projects.This would raise the costs of processing extension loans, but should have high development dividendsin creating sustainable extension systems, and would provide a valuable training exercise for nationalpersonnel.

6.3 Annex 4 elaborates a suggested process for developing a relevant extension strategy andthe preparation of an investment program representing a current phase in the strategy. This involvesa comprehensive analysis of:

- 45 -

* production conditions in agro-ecological zones, the characteristics of major farmingsystems69 in each zone, and the access by farmers in each system to resources andservices (public and private, including NGOs)

* the stock of technology which is appropriate for more profitable and sustainableproduction in the farming systems, and the technology generation capacity

* macroeconomic, sectoral and market factors which influence the use of farmers'resources and of technology to increase productivity

* government commitments to improved extension, existing and anticipated fiscalresources for extension, and institutional human resources.

6.4 The analysis would serve as the basis for developing an extension strategy and an initialinvestment program, provided a suitable macroeconomic and sectoral environment and borrower/agency commitment to improvement exist. The strategy would be for the total sector, so that agricul-tural development projects of any type should be accommodated within this strategy. This analysiswould involve:

* prioritizing target groups and areas for extension investment

* developing an overall strategy which takes into account private and public sectorservices, traditional mass media and the applicability of modern informationtechnologies, face-to-face communication systems, and fiscal capacity to maintainvarious levels of public extension services

* accepting that availability of trained human resources and of relevant technologyshould affect the rate at which a program is developed, and that although a loweducation level in front-line staff may permit the delivery of general technicalmessages, it would limit the capacity to provide comprehensive services which areresponsive to farmers' needs

* defining an initial program in the strategy which is consistent with resources and thetechnology base, captures any potential for cost recovery and encourages farmer"ownership" of the service

* ensuring that a satisfactory technology generation/adaptation capability is in place toservice the needs of all targeted farming systems, and that the long-run productivity ofthe agricultural resource base is taken into account in this process.

* defining the scale, type and intensity of face-to-face services for particular areas;application of many of the extension principles commonly used in the T&V model,

69. Farming system refers to not only on-farm activities but all household resources and use of these resources,including off-farm and non-agricultural pursuits, and factors influencing same.

-46-

with appropriate modification to increase their responsiveness to farmers' needs, 70

would usually be appropriate (Annex 4, Table 1), but mode of operation could varysignificantly between individual programs (Annex 4, para. 4(g) and footnote 3).

* preparing and initiating a needs-based staff training program in technical, economic,sociological and communication aspects71 of technology transfer72

" accepting that the extension program is not a permanent format but must beresponsive to changing circumstances, and ensuring that a properly trained extensionplanning unit is included and supported by a program monitoring capacity which willprovide the basis for responsive management

" insisting that borrowers and implementing agencies are intimately involved in strategyand program design to ensure their ownership of both, and that they are fully awareof its recurrent cost implications.

6.5 The above recommendation is based on the premise that an extension system must berelevan to the circumstances of the country and the intended extension clientele and make efficientuse of public sector resources if it is to be efficacious in meeting its objectives. Underlying thispremise is an assumption that public sector extension investment can give favorable economic returns

70. In many countries there has been an overt objective of providing extension services to attain governmentproduction targets, which may be as general as increasing the use of production inputs or as specific as promotingparticular commodities. It is common for governments to have specific production objectives within a sectoralWae gy. The findings of this review, however, would support an approach in which government's productionobjectives are promoted by creating an appropriate economic and sectoral policy environment to elicit the desiredresponse by farmers. Extension, on the other hand, would aim to educate and assist farmers to make best use ofthe resources available to them. This complementary approach of having government create an incentiveenvironment to promote any national production objectives, and an extension service asisting farmers to meettheir own objectives, is most appropriate if production, income and poverty alleviation goals are to be metefficiently. See OED Report No. 11980, 1993. Indonesia para. 5.20.

71. Expertise in sociological principles and the theory and practice of group dynamics is usually absent in extensionsystems and in their training programs. Technical assistance would normally be required to develop this capacityand experienced NOOs are a possible source. Similarly, rarely is there an appreciation of the benefits which canbe obtained by professional programming of radio and television; these activities will benefit not only fromtechnical assistance but also from contracting of particular services to the private sector to maximize efficiencyand impact. The common difficulty in retaining skilled media professionals in the public service also supports theuse of contracted services to work with the public extension program.

72. In commenting on this report, AGR indicated that the review has concentrated on extension in the narrowconfines of transfer of technology, rather than the broader view of extension in farmer education and organiza-tion, which would be beneficial even when relevant improved technology may be scarce. The report has, indeed,concentrated on technology transfer as this has been the character of Bank projects on which it was based. Inemphasizing the need for relevant technology and responsive services, however, the report strongly supportsgreater client participation in problem definition and resolution (para. 5.12(e), 6.7 (d), footnote 70 and Annex 4Table 1). The report contends that with reasonably educated extensionists, intensive training and appropriateleadership this capability can be developed, and the rate at which further investment is made in regular fac&-to-Wce services should be geared to a realistic expectation of trained staff availability. (China is currently using itsextension staff to expand the formation of farmer associations-para. 5.23.)

- 47 -

provided (a) the elements of relevance and efficiency are adhered to, (b) there is a significanttechnology gap between current practices and available, more profitable technology which is relevantto, and sustainable in, the targeted farming systems,,and (c) alternative, effective forms of technologytransfer are not able to be provided from private sources. The author believes that there is sufficientevidence in the literature to support this assumption.73 Adherence to these principles at the designstage of an extension strategy and any current program will not have the intended results, however,unless they are carried through into implementation. This assertion has implications for borrowersand the Bank.

6.6 For the borrower, this requires a recognition that the measure of the success of anextension service (and justification for public resource allocation) lies not merely in having extensionstaff in the field with adequate infrastructure and equipment implementing an established program, butin the extent to which it is efficiently providing relevant technology to meet the needs of the farmersin the targeted farming systems. Adherence during implementation to the design principles whichfocus on relevance (paras. 6.3 and 6.4 above) will facilitate this achievement. It is important,however, that the borrower provides the resources for a capacity to effectively monitor these factorsand provide the analysis required for responsive decision making by service management (Annex 4,paras. 7 and 8); client reaction should be a prominent part of the monitoring process.

6.7 In view of the importance of recommending relevant technology, and the contributionthat close interaction with farmers can make to this goal, the findings in Chapter 5 suggest fourpriority elements for the borrowers' extension operations:

(a) continuation of the process of identifying or verifying applicable technology andtechnology generation needs for major farming systems with farmer participation,and preferably including joint researcher-extension staff diagnostic surveys. 74

(b) maintenance of close linkages with research in joint field activities as well as intechnical meetings, and also linkages with commercial and other non-governmententities involved in technology development and transfer.

(c) installation of training programs for technical support staff which emphasize thepractical and economic aspects of production and the decision-making processes offarmers faced with constraints in their farming system i.e. developing a responsive,practical, farming system perspective in training of technical support staff.

(d) training of extension staff involved in face-to-face services in how to manage farmergroup interaction to ensure that farmers are involved in defining problems andconstraints, in analyzing the relevance of possible technological solutions, and in theownership of group test or demonstration plots i.e., ensuring greater participationwithin a programmed service which is farmer-centered, rather than message-

73. Birkhauser et al. (1991), Evenson, R. (1991) and Feder, G. et al. (1985).

74. Joint diagnostic surveys have commenced in some projects in a number of countries e.g. Indonesia, Thailand,Brazil, Mexico, Zimbabwe and, recently, Ghana. In most cases, the leadership and initiative in these operationshas been with the research system; this has possibly been associated with a more analytical approach to problemsolving by researchers with a higher level of education than most extension staff.

- 48 -

centered. This would not preclude the use of scheduled farm/group visits nor thedelivery of timely messages. Messages, however, would be in response to theconstraints and potentials periodically defined with each group, and the schedulewould be in accordance with group needs and service resources (Annex 4, Table 1).It could be expected that many widely applicable technologies would be part of mostof these responsive programs.

6.8 For the Bank, project supervision would continue to require monitoring of implementa-tion of investments, agreed extension organization and methods, and compliance with covenants, etc.Much more emphasis, however, would be given to assessing the effectiveness of the extension processin meeting the same primary objective identified for the borrower-accelerating the rate of adoptionof relevant technology in an efficient manner-and the responsiveness of the system to any changedcircumstanas; the latter could be due to an enhanced knowledge base or to other factors endogenousor exogenous to the extension system.7 5

6.9 While there has been a trend in more recent projects (Indonesia, Sri Lanka, Bangladesh)to increase the resources allocated to analysis at the preparation phase and to design a projectaccordingly, many ongoing projects did not have adequate pre-project analysis and have seriousproblems (paras. 5.19 to 5.22). The Bank's mid-term review process provides a convenientopportunity for project adjustments. To be most effective, however, these reviews and any subse-quent project revision would have to consider not only problems in implementing the project asdesigned, but also design issues which have not been adequately addressed previously (paras. 6.3 and6.4); the review would necessarily include "hands-on" random farmer visits and, preferably,organized rapid rural survey techniques to verify the qualitative aspects of the strategy and program inoperation.

ISSUES

6.10 Even if the foregoing recommendations are fully accepted as a legitimate approach forthe Bank to take in addressing its ongoing and future extension portfolio, a number of issues wouldhave to be confronted.

6.11 Fiscal, human resource, complementary service and technology stock constraints oftendictate that investment in state-wide, staff-intensive services are inappropriate at a particular point intime. The Bank (and the borrowers they advise) should accept that (a) while public sector extensionshould be beneficial, it is often not the most limiting factor to productivity increases (para. 2.9 andBox 1), and (b) staff-intensive services introduce a "permanent" overhead in recurrent costs whichmay be inconsistent with national fiscal restraint policies or capacity, and force extension to compete(usually unsuccessfully) in the recurrent budget with programs of higher priority in the view of fiscal

75. In most of the Bank's free-standing extension projects, a general objective of improving production and welfare inthe smalholder farming subsector has been commonly translated into an immediate project objective of"institution-building" (para. 3.9). This, in turn, has meant that much supervision has been dedicated tomonitoring the implementation of the institutional investments in facilities, staffing and agreed organizationalarrangements, without adequate attention to the efficiency of the program in meeting its primary purpose toaccelerate technology adoption. Particular emphasis on ensuring that the monitoring is being carried outefficiently, on using monitoring data to assist in analysis, and on interacting with a random selection of farmers inthe different type of farming systems would be essential for effective supervision.

- 49 -

managers. This would involve a departure from the current approach in some regions of installing(or continuing to support) relatively uniform national systems regardless of local circumstances.

6.12 A limited capacity to undertake the pre-project analysis and responsive designrecommended in paras. 6.3 and 6.4 must also be faced. It can be anticipated that the in-housenational capacity will be very limited without outside support. Experienced agricultural consultancyfirms and the FAO Investment Centre could do the required analysis and design, provided they areworking within appropriate terms of reference and guidance.7 6 The Bank, as an institution, shouldbe able to provide satisfactory guidance (provided appropriate regional policies are adopted), but thismay not be the case in individual operating divisions. Few Bank staff have had significant experiencein extension design and operations outside the Bank and most Bank staff dealing with extensionprojects have had little exposure to systems other than T&V due to the dominant role it has played inprojects over the last 15 years. This suggests a need for education within the Bank on many of theissues highlighted in this report with intervention by external experts with wide experience inextension operations and design.

6.13 The supervision capacity of the Bank is constrained by the same limitations on extensionexpertise, as well as by budgetary constraints. In AFR, considerable progress has been made inincreasing the supervision resources allocated to extension projects (para. 5.21) which potentiallyincreases the Bank's capacity to interact with borrowers in addressing problems and adjustingprograms. Unless the almost exclusive focus on nation-wide T&V systems in AFR is relaxed,however, this potential may not be realized.

6.14 The Bank and borrowers should understand the impact that changing economic, tradeand sectoral conditions will have on the public extension services that can be offered and will bedemanded; this, despite a large proportion of farmers still being near subsistence levels in regionssuch as AFR. Not only must the national strategy to encourage technological improvements beresponsive to changing circumstances, but any program to develop a cadre of extension staff must becognizant of the need for staff to be capable of meeting requirements for a more responsive andanalytical service, whether this be in public or private sector service. This has particular implicationsfor regions such as AFR where agricultural education is very often of a low standard, and even theability of the agricultural universities to produce well-trained graduates for technical extension supportstaff and adaptive research is seriously deficient in many countries. Conversely, the encouragementof a public sector establishment with large numbers of poorly qualified staff with limited potential tomeet emerging demands for services, but which incurs substantial recurrent cost financing in a tightfiscal environment, must be questioned. AFR does not completely accept some of the inferences inthis and the previous paragraph and these differences are elaborated in Box 5.

76. The FAO. Investment Centre has developed a method for integration of socioeconomic considerations intoagricultural investment project design. This has been applied in over 50 projects but only nine of these have beenfor Bank investment. IFAD has been the major client. See the references under FAO Investment Centre.

- 50 -

Box 5: DIFFERENCES BETWEEN OED AND AFR ON INTERPRETATION OF FINDINGS

A conclusion of the study is that the Bank has erred in the extent to which it has promoted the T&V extensionmanagement system in relatively uniform packages of investments and extension practices in large state andnational programs. Although some increased flexibility is now being introduced into the T&V projects, thestudy also concludes that a nation-wide application of the T&V system would not be the most appropriateapproach to improve extension in many African countries. OED considers there is a high probability that suchlarge, ambitious programs would not result in efficient and sustainable systems given the fiscal, humanresource, production environment and technology base characteristics commonly occurring in these countries.

AFR agrees that the location and rate of introduction of extension investmes should be in accordance withlocal circumstances. These are differences, however, between OED and AFR positions on how this principleshould be interpreted in practice. AFR maintains that there is always some improved technology which canbe extended; the existing extension programs are often diverse, inefficient or unreplicable, while the T&Vsystem is manageable and effective; T&V is flexible and can accommodate various extension techniques; andT&V is not an overly expensive system. Therefore, there is a strong case for nation-wide application of T&Vas a comprehensive system in most African countries, especially if it is introduced through a pilot phase.

The reasons for the major differences between AFR and OED on the interpretation of findings are elaboratedbelow:

a) AFR maintains the T&V management system should be introduced nation-wide to avoid the inefficientuse of scarce resources on a large array of individual projects which introduce staff-intensive (and oftenexpatriate-intensive) technology transfer systems which are unreplicable on a large scale, and to developa clearly defined, single extension framework to focus available resources. In addition, the T&Vmanagement system is very flexible, can incorporate such practices as complementary mass mediamethods and commodity-specific technical services, and is continually evolving.

OED concurs that it is desirable to have a national extension strategy. Annex 4 describes how such astrategy could be developed. It would determine where and how public sector extension investment shouldbe involved, and lead to a program of specific extension practices in accordance with assessed needs andavailable resources. The common T&V format of regular face-to-face services, numerous smalltechnology demonstrations and regular training is likely to be an important component of the extensionprogram in specific instances. Equal consideration would be given, however, to the approaches such as(i) short-term campaigns taking advantage of all available sectoral staff to address particular problems, andpossibly (ii) commodity-specific extension programs aligned with delivery of an input package, preferablywith cost recovery, or (iii) highly participatory schemes involving a lesser number of better-educatedextensionists and farming communities in an infrequent but scheduled program of interaction for problemdefinition and resolution, or (iv) reliance on well-organized mass media programs to complement existingcommercial services in appropriate circumstances, all in accordance with the situation analysis as outlinedin Annex 4. Although there are instances where commodity-specific services have been allowed to persistalongside the T&V general service in Bank projects (tea in Kenya, cotton in Cote d'Ivoire) and radio andwritten mass media have received some attention, in practice the concentration of effort in T&V projectscontinues to be on the implementation and improvement of the regular visit-demonstration format by whichthe system is commonly recognized. Modifications are usually only variations on the main format andhave not given rise to substantial, widespread innovations. Indeed, if the "T&V management system' wereso flexible as to accommodate the open analysis and selection of extension techniques as suggested inAnnex 4, this would be a substantial departure from the training and farm visit concepts on which it wasoriginally based; it would simply be a responsive extension management system.

b) AFR maintains there is always some improved technology which can be promoted so that there isusually no need to delay the introduction of the T&V format in an area for want of technology.

continued ...

- 51 -

Continued

OED agrees that in most (but not all) instances an analysis of existing farming systems will reveal sometechnology which can be profitably introduced using the resources available to the farmer. However, thisneed not justify the establishment of a "permanent" extension force to promote its adoption unless thereare good prospects for further improved, relevant technologies being made available. Temporarycampaigns may be appropriate in cases of limited technology. Most of the research systems in AFR areweaker than in other regions, so that the normal development delays in producing relevant research forparticular farming systems are exacerbated on this continent. Although borrowers (with Bank and donorassistance) are investing in research system improvement, enhancement of this capability will be a veryslow process.

c) AFR does not consider sustainability a serious issue, as T&V systems established in Asia have persistedfor nearly 20 years, and empirical studies indicate that extension is generally a sound economicinvestment.

OED agrees that most national T&V extension systems have persisted, albeit in drastically modified formsin some cases. However, this persistence should more correctly be viewed as a commitment to some formof public sector extension; governments appreciate that extension is warranted, and the national T&Vsystem, despite its problems, is in place to give advice to all farmers.

It is likely that any public sector program which extends profitable, relevant, improved production tech-nology will give favorable economic returns. Reorganization of a large unproductive public sector to makethe same people more effective would invariably give high marginal returns as the improved efficiencytakes effect. Unless new relevant technology continues to be generated, however, the marginal return mustdecrease (para. 5.4). Nor should it be assumed that a favorable economic rate of return will generate therequired annual recurrent budget support. Fiscal constraints and political realities will often reduce thepriority given to a service such as extension, regardless of its economic worth; a system which is perceivedas lean and efficient by finance managers, and which is appreciated as necessary and effective by publicbeneficiaries, has the best chance of attracting the necessary funding to keep it efficient and responsive toa recognized demand.

d) APR maintains extension investments are not excessively costly in relation to agricultural GDP, anddo not represent a high cost per beneficiary household, so that service costs should not be considereda deterrent to large nation-wide programs. In any case, in many instances existing agricultural sectorstaff are used; staff reductions have even occurred in Burkina Faso and are planned in Cote d'Ivoire.

OED agrees that most extension budgets are not high in relation to agricultural GDP nor, implicitly, topotential impact (although reliable data on full costs are extremely difficult to find). Annual costs havebeen calculated at US$4 to US$8 per beneficiary family for some services. However, (i) what appearsto be a low cost per beneficiary does not mean much if the same result can be achieved more cheaply, ora better result at the same cost, and (ii) the reality is that fiscal managers normally are under pressure toallocate recurrent costs only to the extent needed to satisfy minimum 'requirements". The latter aredetermined by a complex array of factors, many being socio-political rather than economic, and theextension lobby is normally not strong. The size of the recurrent budget demand for staff-intensiveservices rather than particular economic ratios influences decision-making under these circumstances.Where inefficiencies are obvious in the extension system, the pressure is greater to reduce recurrent costs,hence the common occurrence of service funding being squeezed to little more than the salary budget. Theobjective should be to develop a service which is economic in a true sense, meaning most cost-effectiveto obtain a particular result, and which is within an expenditure level which is able to be politically andfiscally supported in the future.

e) AFR maintains the T&V management system is not "top-down", and farmer feedback and the emphasison a field orientation are integral parts of the system.

continued ...

-52-

Continued

The objective of the T&V system is undoubtedly to take account of farmer circumstances in providingadvice, and there are plenty of cases where individual field and support staff have developed a responsiveservice. Experience has shown, however, that in large, state or national programs this is the exceptionrather than the rule, even after 10 to 15 years of service development. A 'top-down" system suits thepublic service bureaucracy and there is normally no major incentive to institute more responsive services(Box 4). It is also difficult to develop the required analytical capacity and confidence in very largenumbers of field staff with a minimum of agricultural education (and, very often, an urban background)to confidently analyze farmers problems and potentials with them and interact accordingly with higherlevels in the technical support structure. In any public sector extension format, the "top-down' culturemust be actively resisted; in T&V, the predetermined message concept tends to be conducive to thisattitude, and extraordinary measures must be adopted to develop an effective "farmer-centered" culture(para. 5.8).

f) AFR considers the need for increased production is so great, that it makes sense to introduce a nation-wide program in the T&V system, admitting that the capability of the large number of existing fieldstaff and the availability of skilled specialists to support them are very limited, but relying on a gradualimprovement in human resources as the system develops.

OED agrees the need for improved productivity is unquestionable. The study argues for an approachwhich does not seek perfection, but which does not overstretch the human resource and fiscal capacity sothat serious inefficiencies become institutionalized (even if these are less than before the organizationalchanges takes place)-para. 5.12. Once a hierarchical bureaucracy is established, experience indicates itis very slow to be dismantled, regardless of efficiency level; there will be an obligation to pay salaries,but apparent inefficiency will tend to reduce the necessary operational support funding by finance managers(paras. 5.3 and 5.23).

One of the key elements for improvement in any service, regardless of the extension approach used, is theavailability of skilled technical specialists with an appreciation of how the socioeconomic characteristicsof farming systems influence the relevance of a technology for each farming system. Top priority shouldbe given (up front) to training such a cadre, creating a conducive career structure for it, and developingthe necessary linkages with the research system (which must also adopt a farming system, demand-drivenperspective). The study suggests that the restricted number of such staff in the early years of animprovement strategy should be used where they will be most effective (see (a) above). This wouldinclude support for face-to-face services using the typical T&V format in areas where this is consideredthe most appropriate approach, but numbers of skilled specialists would not be available in at least themedium term to service T&V field staff in a nation-wide program so that the intended responsive servicecould be offered. The latter is currently evidenced in the long-established systems in Asia. One of thevaluable areas in which skilled technicians can be used is in interacting with farmers in farming systemdiagnostic work (preferably with researchers) to develop appropriate key improvement areas which canform the basis for focused campaigns using all available sector staff, and for mass media programs andinformation systems designed to satisfy established farmer needs.

The use of the usually large numbers of existing sectoral service staff is a difficult issue. Depending ontraining and aptitude, these could be absorbed as village extension workers under T&V, be used in anycommodity-specific program which may be part of the national strategy, be used temporarily in generalextension campaigns but be destined for use in other agricultural or related services, or even retrenchedif their aptitudes are low and this is politically feasible. Use of existing government sectoral humanresources (and other non-government resources) would be part of the preparation for an extension strategyand program.

AFR has no disagreement with the basic extension principles highlighted in this report, including the needto give priority to the development of extension services which are fully responsive to the needs of farmersin the targeted farming systems, and the need for a close linkage with a research service which is itselforganized to be similarly responsive.

-53- ANNEX 1Table 1

FREE STANDING EXTENSION PROJECTS APPROVED FROM FISCAL YEAR 1977 WHICH RAVE EX-POST EVALUATIONS

PCR PAR (a) L/C L/C Project Extension Orig Actual Project ExtensionReport Nos. No. Country Description US$m cost cost Approv. Close Close Rating (c) Methodology (d)Africa

(b) L2765 Cote D'Ivoire Ag. Extension 34.0 58.5 40.8 Dec-86 Jun-92 Jun-92 UNS T11467 C1387 Kenya National Extension 15.0 28.6 26.9 Jun-83 Dec-87 Jun-91 S T11515 C1669 Rwanda Gitarama Ag. Svcs. 12.7 22.6 7.6 Mar-86 Sep-91 Sep-91 UNS T8579 C0905 Somalia Extension & Training 10.5 32.4 29.9 May-79 Jun-85 Jun-87 S T

11835 L2335 Zimbabwe Ext & Res - MAD 13.1 180.9 115.8 Jul-83 Mar-88 Sep-91 S 0 (c)

East Asia Pacific5158 L1267 Indonesia Nat Food Crps Ext 22.0 43.5 43.5 Jun-76 Jun-81 Dec-82 S T

7770 11980 CO996 Indonesia Ag. Ext II 42.0 81.9 81.9 Mar-80 Mar-86 Mar-87 S T7287 L1493 Malaysia National Ext 19.0 46.5 42.3 Nov-77 Dec-83 Dec-85 S T

8552 L2276 PNG Ag. Support Scrv. 14.1 24.4 5.4 May-83 Oct-83 Oct-83 UNS 07286 L1626 Philippines Nat'l Extension 35.0 70.1 70.1 Nov-78 Jun-83 Jun-85 UNS T6263 L1393 Thailand Nat'l Ag. Ext 28.0 56.5 56.5 Mar-77 Dec-82 Sep-83 S T

10101 8618 L1752 Thailand Ag. Ext II 40.0 70.0 70.0 Aug-79 Dec-85 Dec-87 S T9912 9615 C1407 Vanuatu Ag. Res & Ext 2.0 3.8 3.6 Aug-83 Jun-89 Jun-89 UNS 0

Middle East & North Africa11731 C1557 Yemen Ag. Res & Ext 5.0 7.0 3.8 Mar-85 Dec-90 Dec-91 S T

Latin America and the Caribbean7331 L1568 Brazil Ag. Extension 100.0 284.9 284.9 May-78 Dec-82 Dec-85 5 0 (f)

10630 L2150 Peru Ag. Res & Ext 40.6 83.3 48.1 May-82 Mar-88 Mar-88 UNS T

South Asia5071 C0729 Bangladesh Ag. Ext & Res I 10.0 16.3 13.0 Jun-77 Apr-81 Jun-82 S T

11363 C1215 Bangladesh Ag. Ext & Res II 27.0 40.9 32.2 Mar-82 Jun-88 Jun-91 S T6654 8808 C0690 India W. Bengal Ext & Res 12.0 28.1 17.1 Mar-77 Sep-82 Mar-85 UNS T6800 8808 C0761 India Ext & Res - Bih,ar 8.0 16.0 13.8 Dec-77 Oct-83 Mar-85 UNS T

5618 C0737 India Ext & Res - Rajas 13.0 26.6 21.5 Jul-77 Jun-83 Jun-83 S T6281 C0862 India Composite Ag. Ext 25.0 53.4 53.4 Dec-78 Dec-84 Jun-85 S T

8258 8808 C1028 India Kerala Ag. Ext 10.0 15.3 15.3 May-80 Jun-86 Jun-86 S T8258 8808 C1137 India Tamil Nadu Ag. Ext 28.0 45.6 45.6 Apr-81 Jun-87 Jun-87 S T8258 8808 C1135 India Maharashtra Ag. Ext 23.0 38.2 38.2 Apr-81 Jun-87 Jun-87 S T9189 C1138 India M.P. Ag. Extension 37.0 60.0 60.0 Apr-81 Jun-87 Jun-89 S T9033 C1219 India A.P. Extension 6.0 9.7 9.7 Mar-82 Mar-88 Mar-89 S T9994 10661 C1100 Nepal . Ext & Rea 17.5 20.8 6.3 Feb-81 Jul-87 Jul-90 S T9994 10661 C1101 Nepal Hill Food Prod I 8.0 9.7 3.2 Feb-81 Jul-87 Jul-90 UNS T

7288 C0813 Pakistan Punjab Ext & Ag. Dev 12.5 20.8 15.5 May-78 Jun-84 Jun-85 S T7288 C0922 Pakistan Ext & Adpt. Res (Sind) 9.0 14.3 9.1 May-79 Jun-85 Jun-85 S T

(b) C1533 Pakistan Ag. Ext & Res (Bal.) 8.3 11.5 9.8 Dec-84 Jun-90 Jun-92 UNS T7684 10632 C0931 Sri Lanka Nat'1 Ext & Res 15.5 22.4 18.1 Jun-79 Jun-85 Jun-86 S T

Note: (a) The majority of ex-post evaluations done before 1987 are registered as PARs, as all PCRs were subject to a Project PerformanceAudit Memorandum which was attached to the PCR to become the PAR. Subsequently, the PCR and PAR have been presented asseparate products.(b) PCR is in process.(c) S - satisfactory; UNS - unsatisfactory(d) T - some form of T&V; 0 - other methodology(c) This project included T&V as one pilot scheme only, but did include the features of improved programming of services and training.(f) Although not labeled as T&V, this project promoted programming of regular visits to farmer groups and heavy emphasison small technology demonstrations.

-54- ANNEX 1 - Table 2

Page 1 of 2

SAMPLE OF AGRICULTURAL PROJECTS WITH EXTENSION COMPONENTS, WHICH HAVE EX-POST EVALUATIONS

PCK PAR UC UC Projec Orig Acual Sub- EPaahn Esandam

Report Noa. No. County De0aiption USSm cost Approv. Cose Clo~ s~r (a) Ratig (b) Møb ~cagy (c)Afnum

10478 C1127 Benin Rura Dev/ Co~a 21.9 41.0 Apr-S1 Dec-S8 Dec-88 ADP MS 010511 C116.5 Burundi Kinmro Rural Dev 19.3 21.4 Jun-91 Jun-87 Jun-88 ADP MS T

10510 C1192 Burundi Int. R/D Ngozi II 16.0 28.9 Dec-gl Mar-87 Dec-89 ADP MS T

7904 C1013 Burkina Fam Niena Dionkele Rice 6.5 7.1 Apr-80 Der-84 Dec-86 IR S 0

6808 8690 0776 Cameroon Zapi Eas Rural Dev 8.5 12.2 Mar-78 Jun-84 Dec-84 ADP UNS T

6962 8690 00784 Caneroon Rural Dev Weat High 13.0 25.0 Mar-78 Jun-83 Jun-84 ADP S 0

9630 C1376 CAR Rurai Developmeni 10.4 51.2 May-93 Mar-88 Sep-89 ADP MS 0

8631 C0959 Kenya lntegrated Ag. Dev 46.0 91.7 Dec-79 Apr-86 Apr-86 ADP MS T

C006 0786 Liberia Rubber Development 13.0 29.6 Mar-78 Jun-84 Dec-85 PER UNS 0

6171 CO23 Malawi Shire Consolidation 10.7 12.6 Jun-78 Dec-82 Dec-92 ADP UNS 06797 9750 C0857 Malawi Nat'l Run Dev I 22.0 66.0 Nov-78 Sqp-83 Jun-86 ADP UNS 08638 9750 C1183 Malawi NRDP IE 7.3 8.6 Oct-91 Sep-86 Sep-88 ADP UNS 0

10211 9750 C1343 Malawi NRDP IV 10.6 12.5 Apr-83 Sep-88 Apr-91 ADP UNS T

7315 8557 L1667 Nigena Bida Ag. Dev 23.0 64.4 Mar-79 Jun-85 Jun-6 ADP UNS 010287 12016 L1668 Nigeria Ag. Dev - Ilorin 27.0 64.4 Mar-79 Jun-45 Jun-88 ADP UNS 0

10328 12016 L1838 Nigeria Ag. Dev & Tech. Asst. 69.4 28.0 Apr-80 Mar-85 Sep-88 ADP UNS 08520 8557 L1854 Nigeria Eiu Akodo Ag. Dev 32.5 80.3 May-80 Sep-85 Sep-85 ADP UNS 0

10986 12016 L1981 Nigeria Bauchi Ag. Dev 132.0 350.6 Apr-81 Dec-86 Dec-89 ADP MS T10980 12016 L1982 Nigeria Kano Ag. Dev 142.0 482.2 Apr-81 Dec-86 Dec-89 ADP MS T10981 12016 L2185 Nigeria Sokoto State Ag. Dev 147.0 498.7 Jun-82 Jun-87 Dec-90 ADP MS T9245 C1128 S. Leone Northermn ADP [1 8.5 27.7 Apr-81 Jun-87 Jun-87 ADP MS T

7262 0804 Sudan Mechanized Farming I 16.0 26.4 May-78 Mar-84 Mar-95 ADP UNS 08789 0945 Togo Cocod/Coffee i 14.0 34.2 Jun-79 Dec-85 Jun-87 PER MS 09152 C1193 Zambia Ag. MD (So. Prov.) 18.0 24.6 Dec-Sl Dec-87 Jun-91 ADP UNS T

8582 L201 Zambia Ag. Dev (E. Province) 11.0 29.1 May-Sl Jun-87 Jun-88 ADP MS T

Edst Asia Padfic

9571 C0835 Myanmar Lowcr Burma Paddy 34.5 56.0 Jul-79 Apr-86 Apr-90 IR MS T10698 C1031 Myamar lrg. t1 Nyaunggyat 90.0 235.0 M ay-90 Jun-87 Mar-91 IR MS T

7736 C1261 China N. China Plain 1 60.0 177.5 Jun-82 Dec-87 Dec-87 IR s 07771 11431 0919 Indonesia Transmig. U 157.0 242.0 May-79 Dec-85 Dec-86 SET UNS T9246 11981 C0946 Indonesia Yogyakaru ural Dev 12.0 19.0 Jul-79 Mar-87 Mar-89 ADP s 07794 7995 L1499 lndoneaia . NES I 65.0 134.0 Nov-77 Jun-82 Jun-84 PER MS 07794 7995 L1604 Ladonesia NES Il 65.0 100.5 Jun-78 Dec-83 Dec-86 PER MS 08135 9660 L1898 Indoneaia Smllhidr Coconut Dev 46.0 95.2 Aug-80 Dec-S5 Jul-87 PER MS T7930 11431 L1958 Indonesia Swamp Reclamation 1 22.0 44.0 Mar-81 Dec-86 Dec-87 SET UNS T

10768 11431 L2288 Indoneaia Transmig. IV (Kalm.) 63.5 121.0 May-83 Dec-87 Mar-91 SET UNS T8104 C1227 PNG Enga Prov. Dev 8.0 16.0 Apr-82 Sep-86 Sep-S8 ADP UNS 0

7949 L1815 Phihppinea Rainfed Ag. Dev 12.0 21.4 Mar-80 Jun-86 Jun-86 ADP s T10430 C0929 Thailand Northern Ag. Dev 25.0 47.5 Jun-79 Dec-85 Dec-88 ADP MS T8173 10051 L2078 Thailand Tree Crops Il 142.0 358.8 Jan-82 Mar-87 Mar-87 PER s 0

-55 - ANNEX 1 - Table 2Page 2 of 2

SAMPLE OF AGRICULTURAL PROJETS WITH EXTENSION COMPONENTS, WHICH HAVE EX-POST EVALUATIONS

PCR PAR IJC UC Project Orig A l Sub- Estanso Extendon

Report No. No. Country Description US$m cost Approv. Close Close sector (a) Rating (b) Methodology (c)

Europe, Central Amia, Middle East & North Africm

9859 8885 L1988 Cyprus Fruit & Veg. 14.0 32.9 May-81 Jun-86 Dec-88 CR UNS 0

6013 88&5 L1588 Greece Veg. Production MkL. 30.0 84.0 Jun-75 Jun-43 Jun-84 CR S 0

6155 L1603 Portugal Ag. Credit 1 70.0 256.5 Jun-73 Jun-84 Jun-84 CR UNS 0

11078 1.2175 Portugal Rural Dev 51.0 154.8 Jun-12 Dec-88 Dec-89 ADP MS T

8157 8885 L1967 Tukey Fruits & Veg. II 40.0 107.0 Mar-81 Jun-86 Jun-86 CR UNS 0

11410 L2094 Turkey Erzurum Rural Dev 40.0 137.0 Feb-82 Jua-87 Dec-89 ADP MS T

10808 L1602 Morocco Karia-Tasa Rafd Ag. 65.0 161.5 Jun-7 8 Jua-86 Jun-90 ADP S T

7821 8885 L1757 Morocco Veg. Prod. & Mks. 58.0 129.6 Sep-79 Juo-34 Dec-85 CR MS 0

364 C1067 Yemen S. Uplands Rural Dev 17.0 11.6 Sep-80 Dec-85 Jun-87 ADP S T

lAtia Amrwc and the Caribbeam

7331 L1362 Brazil Minais Gerais I R/D 42.0 139.0 Jan-77 Dec-81 Dec-84 ADP S 0

7910 L1537 Brazil Rural Dev - Paraiba 24.0 67.3 Aug-73 Sep-83 Sep-86 ADP MS 0

7910 L1488 Brazil Rural Dev - Cears 17.0 55.8 Feb-73 Dec-82 Dec-85 ADP MS 0

7331 L1589 Brazil Integrated RID IV - Ba. 37.0 106.6 Oct-73 Dec-33 Dec-45 ADP S 0

7910 L1728 Brazil Rural Dev V - Pernam. 40.0 116.7 Oct-79 Dec-34 Dec-86 ADP MS 0

3380 10636 L1877 Brazil Minais Gerais II R/D 63.0 184.6 Dec-80 Dec-85 Dec-87 ADP MS 0 (d)

3380 10636 L1924 Brazil Ceara II Intgr. R/D 56.0 163.2 Apr-81 Dec-85 Dec-87 ADP UNS 0

3488 9700 L2015 Brazil Piaui Rural Dev 29.0 34.4 Nov-81 Dec-86 Dec-86 ADP UNS 010506 12177 Brazil Maranhao R/D 42.7 122.2 Sep-82 Dec-87 Dec-88 ADP UNS 0

9382 12116 Brazil NW Day II - M. Grosso 26.4 76.4 Aug-82 Dec-86 Dec-88 ADP MS 0

9597 L2269 Brazil Bahia R/D II 67.8 173.6 Aug-83 Dec-38 Dec-89 ADP MS 0

L2481 Chile Ag. Services/Credit 56.0 157.2 Mar-85 Mar-90 Oct-91 CR S 0

4240 8860 L1569 Mexico Livestock Credit VI 200.0 627.2 Dec-78 Jun-82 Jun-82 CR S 05997 L1553 Mexico Tropical Ag. Dev 56.0 149.0 Dec-78 Dec-83 Jun-84 ADP S 07931 8860 L1891 Mexico Ag. Credit/FIRA VII 325.0 1179.0 Nov-80 Mar-84 Dec-84 CR S 08335 8880 L1945 Mexico Rainfed Ag. Dev 280.0 797.0 Jun-81 Jun-86 Jun-37 ADP S 0

9175 12043 Mexico Pider III R/D 175.0 505.7 Feb-82 Sep-85 Dec-U ADP MS 0

8926 860 12454 Mexico Ag. Credit/ FIRA VIII 300.0 763.0 Nov-34 Sep-U Sep-U CR S 0

8926 8860 12610 Mexico Ag. Credit/FIRA VIII-A 180.0 473.0 Jul-85 Dec-86 Sep-87 CR S 0

South Asia

5746 C)765 Bangladesh lute Development I 21.0 33.3 Jan-78 Jun-33 Jun-83 ADP MS T

9020 C1399 Bangladesh Sugar Dev 20.0 25.7 Jul-33 Dec-83 Jun-39 A.IN. MS T

5940 C0736 India CAD Maharashtra 70.0 140.0 Jul-77 Mar-83 Mar-84 IR MS T

9375 C1034 India Karnataka Sericulturt 54.0 95.1 Jun-80 Dec-85 Sep-88 SER MS 0

9415 CD963 India Fisheries IV - Inland 20.0 40.8 Dec-79 Sep-85 Scp-88 FISH S 0

9021 C1012 India Cashewut Proj. 22.0 45.7 Apr-80 Sep-85 Sep-87 A.IN. MS 010477 C1316 Nepal Irrig. VII (Bhairawa) 16.0 19.4 Jan-83 Oct-38 Oct-90 IR S T

Note: (a) ADP - area development; A.IN - agroindustry; CR - credit; FISH - fishenes; IR - irrigation/drainage; PER - perennial crops;

SER - sriculture; SET - settiement

(b) S - satisfactory; UNS - unsatisfactory; MS - marginally satisfactory

(c) T - some for of T&V; 0 - other methodology

(d) The Brazilian arta development projects commenced from the early 1980s incorporated the concepts promoted by the federal

extension agency (EMBRATER) under Lnl56&-BR. These included programming of regular visits to farmer groups and heavy

emphasis on small technology demonstrations.

-56- ANNEX 2Page 1 of 8

REVIEW OF COMPONENT PROJECTS BY REGION

1. The basis for selecting this subset of projects is described in para. 3.1 of the text, andthe list of projects detailing some of their characteristics appears as Annex 1 Table 2.

South Asia Region (SAS)

2. In a sample of seven component projects, two were irrigation/drainage projects whichemployed a T&V model, as did a sugar development project. The remaining projects werecommodity-specific and used intensive extension services aligned with credit (and often input supplyand adaptive research) to encourage development.

3. Of the three projects designed to use T&V, the project with a favorable irrigationproduction environment and appropriate technology (Credit 1316-NEP) had very favorable results, towhich the extension program very likely made significant contribution. The T&V extension in asecond irrigation project (Credit 736-IND) had limited results, largely due to uncertainty regardingthe most appropriate cropping patterns for the local environment. Extension in the third project withT&V (Credit 1399-BD) appeared to have little impact due to the constraints imposed by aninappropriate pricing/profitability environment for sugar as the targeted commodity, although the PCRsuggested there may have been some yield increases which could have been linked to the specializedT&V services.

4. The intensive commodity-specific services in the other four projects generally hadpositive impacts. However, in one (Credit 1012-IND) the impact was constrained by a lack ofattention to the socioeconomic characteristics of the common farming systems in which thecommodity (cashew nuts) interacted. In another (Credit 1034-IND) the service was for commercialinland fisheries and while effective, the justification for having a public sector service for this type ofenterprise was queried as the beneficiaries could well afford to pay for technical services. Theintensive commodity extension in a third project (Credit 765-BD) was absorbed into the regular T&Vstate system towards the end of the project, and a reduced effectiveness of extension on juteproduction was reported (without any hard evidence to this effect).

East Asia and Pacific Region (EAP)

5. The EAP sample included four area development, three irrigation/drainage, fourperennial crop, and three settlement projects.

6. Of the four area development projects, three were personnel-intensive. Of these, aPapua New Guinea project (Credit 1227-PNG) had separate extension services for food crops andcash crops and suffered from a deficiency of appropriate technology and budgetary support problems;


the service was unsatisfactory. The extension in an Indonesia project (Credit 946-INS) wasaccompanied by substantial investment grants, in addition to being very staff-intensive, and achievedreasonable results. The system, however, was very costly and did not represent a model forreplicability on a large scale. A third project (Credit 929-TH) had two separate intensive extensioncomponents. One was associated with land development with ethnic Thai farmers in their uplandparts of their holdings using a modified T&V approach (450:1); the other was largely with hill tribesinvolving a reduction in shifting cultivation, increase in cash cropping and vegetative soil conservationmeasures. The Thai component achieved some immediate impact with assistance from investmentsubsidy, but this was substantially diluted on reversion to regular Thai Department of AgricultureT&V services at project closure. The hill tribe component, however, developed (with effectivetechnical assistance) a system of working closely with communities in defining investment andproduction priorities, including use of group problem census techniques, and facilitated the creationand successful operation involving funds which continue to be supported by farmer/communitypayments for project investments. The extension service (by the Hill Tribe Welfare Division of theMinistry of Interior) continued its activities after the Bank project closure with a further technicalassistance project from AIDAB, and has persisted in the same effective program subsequent to theAIDAB project. The fourth project was modelled on T&V (Loan 815-PH). It was located in a highpotential zone and obtained reasonable results; extension did suffer, however, from budgetary supportproblems and scarcity of the credit needed for adoption of the technologies appropriate for the zone.

7. Of the irrigation/drainage projects, one was in China (Credit 1261-CHA). This projectand its extension component were very successful. A project unit of the Ministry of Agricultureworked with and mobilized the recipient communities for construction work; credit for inputs andcapital item was provided to complement the technical extension services which had appropriatetechnology for the high potential production environment. The other two irrigation projects were inMyanmar and used the T&V model to complement the irrigation investments (Credit 835-BA and1351-BA). These extension components were rated as partially successful without much comment,although one reported problems with a scarcity of production inputs and poor provision of transport.

8. One of the perennial crop projects was in Thailand (Loan 2078-TH). The projectincluded a total package of investment and production inputs for rubber replanting (financed by amarket cess) and relatively intensive technical extension by the specialized rubber agency until treesreached production. The extension was successful; the favorable technical orientation of the service,however, was not matched by an awareness of socioeconomic problems being faced by the majorityof sharecroppers which formed the main labor force in the industry. The other three perennial cropprojects were in Indonesia. One (Loan 1898-INS) had a packaged delivery of inputs and services bya specialized unit of the Ministry of Agriculture, with financing as credit-in-kind. A T&V model wasused for extension after the planting of the tree crop. This was partially successful but suffered fromstandard "messages" which did not adopt to the different client circumstances nor consider economicfactors; the use of production inputs (especially fertilization) after the credit-in-kind period was notgood. The two other projects (Loans 1499 and 1604-INS) were implemented through parastatalcompanies providing packages of investment inputs and advice on tree crops, also financed by credit-in-kind. The intensive extension services by the parastatals, however, although relatively successfulin the establishment period of tree crops, largely failed with annual food crops; the latter sufferedfrom a lack of appropriate technology to fit the socioeconomic circumstances of the smallholders in a


difficult soil environment, and less than full commitment by the management and staff of thespecialized tree crop parastatals to this element of the program.

9. The three voluntary settlement projects in Indonesia (Credits 919, 2288, 1958-INS)provided unsatisfactory extension services. These were in a T&V model as part of the nationalextension system. The main problems were a lack of appropriate technology for the circumstances ofthe settlers (in a difficult upland soil environment), a "top-down" orientation in recommendationswhich did not appreciate the real constraints in the farming systems, lack of practical and economicorientation by E.A.s, and transport difficulties.

Africa Region (AFR)

10. Area development projects formed a major part of the agricultural portfolio in thisregion in the late 1970s and 1980s. Consequently, of a sample of 25 component projects, 22 are areadevelopment projects, two relate to perennial crops and one to irrigation/drainage. In half of the 22area development projects the extension component was based on T&V, while in the other half a less-organized but intensive service was offered.

11. Of those 11 area development projects with T&V, all included rural infrastructure, allbut one had production input supply, eight had a research component, seven had credit and fourpromoted cooperative development - a typical mix for this type of project. Extension was rated asunsatisfactory in two projects (Credit 776-CAM and Credit 1193-ZAM) and could be linked to verypoor implementation by the institutions involved. The extension components in the others (9) wererated as partially effective, but there were a number of frequent problems which detracted fromperformance. Seven faced budgetary support constraints, three had staff transport difficulties, andfive suffered from poor linkage with research. Four of the project reports noted inappropriatetechnology, four indicated inadequate knowledge of the farming systems involved, in three a poorappreciation of production economics was highlighted, and three specified a "top-down" characteristicin formulating recommendations with consequent relevance problems. Insufficient access to keyinputs caused problems in three projects.

12. In the other 11 area development projects (which had typical component mixes), theextension services in two were judged unsatisfactory primarily due to very poor management andimplementation, and compounded by poor technology and limited funding and research support (Loan1667-NIG and Loan 1854-NIG). One project was implemented by a well established coffeecooperative (Credit 784-CAM) which provided relatively effective services for cash and food crops ina well organized program with regular service delivery; even then the services could have beenimproved by better adaptive research support and a greater knowledge of the resource economic(including labor) of the common farming systems (more attention to the food crop systems waswarranted during project preparation). Two other partially effective, very intensive services wereprovided in West African projects, Credit 1127-BEN and Credit 1376-CAR. Both projects had betterresults with the dominant cash crop (cotton) than with food crops. In the Benin project, the food cropextension suffered from poor research back-up, insufficient technology and lack of attention toproduction economics. In the CAR project, extension was provided by a cotton parastatal which waslinked to excessive attention to cotton; this continued despite the apparent need for diversification as


cotton prices collapsed, and, as for the Benin projects, insufficient technology support and awarenessof production constraints limited effectiveness.

13. The remaining six non-T&V, intensive extension components in rural developmentprojects were judged unsatisfactory as providers of services to the smallholder subsector. All sufferedfrom inappropriate technology, which was oftea associated with poor linkages with research (4),inadequate appreciation of farming system circumstances (3), or disregard for production economics(2). Budgetary constraints also affected the extension operations in four cases. Three of theseprojects, however, were in Malawi (Credits 823, 857 and 1183-MAL) and did provide relativelysuccessful, intensive services along with production credit and inputs to larger, creditworthy farmers;the problem was that the objective of providing services to the majority of smallholders was not met,and even hindered by the exclusiveness of the intensive service (300:1) offered to the less-than-thirtypercent of farmers with better resources in credit groups. The last (fourth) Malawi national ruraldevelopment project (Credit 1343-MAL) did attempt a T&V model working with farmer groups notexclusively linked to credit. The extension performance in this project improved a little and isincluded as partially effective in those projects described in para. 23, but the tendency to work withthose farmers with better resources continued.

14. The two perennial crop development projects had extensive extension services directedat specific commodities (rubber for Credit 786-LIB, and cocoa and coffee for Credit 945-TOG) withcredit for development and the provision of production inputs. Extension in the Liberian project wasineffective as price/profitability factors severely limited farmer participation. In the Togo project, theextension in cocoa was hampered by a lack of appreciation of the sharecropping characteristics of thefarming system, but the program with coffee was successful as technology and price incentives werefavorable.

15. The irrigation project (Credit 1013-BUR) was implemented by a relatively autonomousproject management unit and had intensive extension, credit and inputs supplied through servicecenters. Extension worked relatively well on rice culture in which the staff were well-trained, butpoor results were obtained in efforts to diversify out of rice, despite water supply problems. The ex-post reports did not specify the reasons for this failure, but it could be inferred that the staff were notadequately informed on the resource and management implications of a change from rice culture intothe various diversifications.

Latin America and the Caribbean Region (LAC)

16. Twenty projects were selected in the LAC region - 11 in Brazil and 8 in Mexico as thedominant regional borrowers - and one in Chile.

17. Of the 11 projects in Brazil, all were area development projects. This meant that, inaddition to extension, components in most cases included rural infrastructure and social services,production input supply and agricultural credit. In many cases there was an investment grant element(in the areas with very poor resources) to complement credit, and components for soil conservation,development of cooperatives and small-scale irrigation were included in specific projects.


18. The extension components of the Brazilian projects were very personnel-intensive.Many of the earlier projects operated with only 100 to 200 clients per EA. The technical extensionservices were usually associated with on-farm development investments supported by grant or creditfinance, and the EA worked with these individual clients who were normally from those smallholderswith a relatively high level of resources. These earlier projects had some success in working with thevery limited clientele, but they often suffered from long periods of restricted grant/credit availabilitywhich further restricted the scope of activities and service morale. There were also criticisms that thetechnology packages promoted/financed were too rigid and often not the most appropriate for thecircumstances of the client. The intensive personnel/client ratio was partly linked to the multitude offunctions that the EA had in the area development programs.

19. Later Brazilian projects attempted to obtain a wider client coverage (but still withintensive services - 350 to 500 households:EA) by working with groups rather than individualsmallholders (although usually with contact farmer "representatives' of the groups), by givingemphasis to technology demonstrations, and by introducing the concept of programming the EA visitsto the intended clientele, i.e., some of the elements of T&V.1 Overall, these attempts to develop abroader agricultural extension service were not as effective as expected. Common problems were ascarcity of appropriate technology for the targeted farming systems, poor research-extension linkages,a strong "top-down" tendency a lack of practical orientation in staff, budgetary constraints limitingoperations and affecting morale, and poor results with the concept of using a contact farmerrepresentative.

20. In an overall review by OED of the series of area development projects in NortheastBrazil, 2 it was pointed out that the more successful episodes of technology dissemination wereusually associated with specific campaigns against crop diseases or pests (boll weevil in cotton, anorange disease in Sergipe, and banana-root fungi in Paraiba). These campaigns avoided theconstraints which affect many general extension services: (a) there was a clearly defined, singleproblem to address, in a relatively simple task rather than the more complex task of understandingand dealing with a whole farm situation; (b) the problem was very evident and demanded the attentionnot only of farmers, but also of political entities and officials of all agencies dealing with the ruralsector, so that public sector entities (including the extension agencies) mobilized around the programto ensure all necessary inputs were guaranteed; (c) performance could be (and was) judged on thebasis of outputs/results rather than inputs delivered; (d) it covered a limited timeframe (usually one ortwo crop seasons) which was conducive to sustained, priority effort; and (e) the technology to addressthe problem was available (avoiding the need for close research-extension collaboration). Althoughthese are legitimate explanatory variables for the success of the campaigns, they do not offer generalsolutions to the constraints faced in developing an effective smallholder extension service, but rather

1. These principles reflected the Brazilian intention to improve national extension services, and were consistent withthe Agricultural Extension Project (Loan 1568-BR, from 1978 to 1985) which supported the federal extensionagency (EMBRATER) in assisting the state extension agencies (EMATERs) to improve services; the EMATERsprovided extension in the area development projects.


-61- ANNEX 2

Page 6 of 8

underline the difficulties incurred.3 The very limited improvement possibilities posed by theharshness of the environment and the land tenure and smallholding size problems in the semi-aridareas of N.E. Brazil, however, do raise the issue of the extent to which public funds should be usedon agricultural extension relatively to other programs which might more appropriately address thepersistent poverty issue.

21. It is also noteworthy that under the Second Agricultural Research Project in Brazil (Loan2016-BR) a number of extension subprojects with French bilateral assistance were developed with theparticipation of both the state EMATERs and the staff from research centers of EMBRAPA, thenational research agency. An initial farm system research program evolved into a pilot program ofrapid field assessment to define communities and their farming system characteristics, constraints andpotentias followed by working with groups in problem definition, technology review and verification,and an action/adoption program to resolve problems and realize potentials. These programs wererated as very successful.

22. Five of the eight Mexican component projects were classified as agricultural credit andthree as area development projects. Although most of the credit projects were rated as unsatisfactory,the intensive technical assistance provided with the credit performed reasonably well; the problem wasthe very limited coverage of the smallholder population in this type of credit/extension systems,especially those smallholders with less resources.

23. Each of the area development projects had intensive services which were reasonablyeffective. Two attempted to use the principles of the successful "Plan Puebla" model. This modelinvolved initial technical and socioeconomic diagnostic work in an area by an interdisciplinary projectteam, followed by testing and promotion of relevant technology in selected communities with theirheavy participation. Coordination.with credit agencies to facilitate procurement of inputs bysmallholders was organized, and the concept was to progressively work through communities in thearea using well-trained university graduates to "educate" the farmers to help themselves. In the areadevelopment projects (Loan 1553 and 1945-MX) these principles were considerably diluted, especiallyin the latter, very large project, in which production and social infrastructure investments wereincluded and subsidized credit dominated the program. Diagnostic surveys were undertaken,however, and adaptive research and extension work were well-coordinated, but the degree ofexperience in staff declined and did reduce effectiveness. 4 Both projects were in relatively highpotential areas. A real problem was the relatively high public sector cost of this type of areadevelopment (which involved not only extension but also major infrastructural and credit investments)for a limited number of districts, which limited its potential as a model to address the majority of therural poor. Extension in the third area development project was similarly linked to productioninvestments but did not have the same amount of emphasis on diagnostic work or research-extension

3. The same report (Tendler, 1993) concludes that the most effective development projects (as opposed to extensionservices) should involve single-crop, highly-focus interventions to be implemented in a specific, limited timeframewhich are measured in terms of output.

4. The T&V method was to be implemented in Loan 1945-MX, but project management preferred to stay with thePlan Puebla model.


linkage (the PIDER program). and was not as effective as the models previously described; thetechnical follow-up to investment sub-projects was also less adequate.

24. The project in Chile (Loan 2481-CH) included credit with private sector consultantsproviding the technical extension service to small commercial farmers. 5 The technical assistance wasobligatory for credit recipients for three years; it was subsidized by a government agency (INDAP)under a graduated cost sharing formula; consultants were approved by INDAP and their performancewas monitored. The service was rated very effective in achieving adoption of technology andresulting commodity production levels. Towards the end of the project, the program was redirectedto smaller farmers with less resources, and support for this clientele is being continued under anongoing project (Loan 3473-CH). It is yet too early to assess results of the model with this group butpreliminary results are encouraging. It is already apparent, though, that there is a need for moreadaptive research to provide technological innovations more appropriate to farmers with lessresources.

Europe and Central Asia/Middle East and North Afica Regions (ECAIMNA)

25. The sample of nine projects in these two Bank regions included five credit and four areadevelopment projects.

26. Four of the credit projects involved intensive public sector credit and extension servicesto producers of high-value export crops (Loans 1988-CYP, 1588-GR, 1967-TK AND 1757-MOR).The extension component of the Greek project was rated as satisfactory, but extension in the Turkishand Cypriot projects were unsatisfactory; the Moroccan project extension component was partiallyeffective. Common problems, which were exhibited to different degrees in specific projects, includeda) an inadequate appreciation of the most up-to-date technology suitable to support the production,processing and marketing of the high-value fruit and vegetable crops, b) staff who were notadequately trained or given incentives to meet the demands of a technical and economic service of thistype, and c) a lack of cost recovery for this intensive service. The conclusion was that beneficiariesshould pay for services of this type, and that they would be best provided through private sectorentities. The fifth credit project (Loan 1603-POR) was directed at agricultural development of largecollective farms and medium and small sized private farms, as well as investments in agro-processingand fishing vessels. The extension services were not very effective due to a) the collectives not beingviable, b) excessive rigidity in development models, and staff not appreciating the economics ofproduction on private farms, and c) diverse duties of staff (including land reform activities) whichdetracted from the technical orientation of the service.

27. The four area development projects used the T&V model to provide extension services;one of those projects (Loan 1602-MOR) had more intensive T&V services. All projects had thecommon array of components for area development projects, although a Turkish project (Loan 2094-TK) did not specifically include production inputs. Extension in the Turkish project and in a project

5. The project was directed at small farmers with holdings of 8 to 12 potential irrigated ha equivalents (a localmeasure); this is a larger farm size than most smallholders in Asia and Africa.


in Portugal (Loan 2175-POR) was rated as partially effective. Both suffered from inadequatepractical, technical training of staff and had staff mobility problems, and extension in the Turkishproject suffered from budgetary support constraints towards, and subsequent to, the end of the projectperiod. The Moroccan project and a project in Yemen (Credit 1067-YEM) were rated as havingsatisfactory extension results. The Yemen project, which was in YAR-North Yemen, had problemsin linkages with the national research system (whereas this was not reported as a problem in thenational extension project in PDRY-South Yemen, which had a high proportion of universitygraduates in its extension service), and also suffered from insufficient transport and limited coveragethrough the contact farmer approach. The limited coverage, however, was partly overcome by thehigh exposure of the project in villages (drinking water provision and bilharzia control) and byextension centers supplying inputs at a ratio of one center to 20 villages.


RECENT REVIEW OF EXTENSION EXPERIENCE IN AFR

1. Not only is AFR the most active Bank region in promoting free-standing nationalextension projects, but it also has almost exclusively used the T&V model of extension. A number ofpositive effects are obvious:

* The Bank's support for extension in Africa continues to focus attention on the needfor a faster rate of adoption of appropriate technology if agricultural productivity is tobe significantly improved.

* The programs are encouraging the adoption of national extension strategies (in lieu offragmented projects with "temporary" extension components supported by a multitudeof donors, each with its own ideas on technology transfer) which encourages betteruse of institutional and fiscal resources. Realization of this goal would have theadvantage of creating the potential to develop a cadre of career extension personnelwith appropriate training and motivation.I

* Better coverage of smallholders is being obtained and indications are that the serviceshave been associated with an increase in the rate of adoption of some importanttechnologies.

2. Recognizing the enormity of the capacity-building task in AFR, regional managementresponded with a number of rational measures:

* In 1987 it adopted an Agricultural Services Initiative which had the objective ofbroadening the coverage of the national projects to include, besides research andextebsion, other agricultural services, such as supply of improved seeds, credit,marketing etc.; if properly used, this Initiative has the potential to identify and addressthe most limiting constraints to technology adoption.

* As part of its efforts to strengthen the capacity of the countries to implement extensionprojects, it appointed agricultural extension services specialists in some 19 countries(most of these field staff have been in position since 1989).

* A regional extension network was established and a series of extension workshops forBank staff and national management personnel was organized to promote exchange of

1. In some cases, the rationalization of extension components in numerous projects and programs into a nationalsystem has been able to reduce (in theory, at least) public sector extension personnel, and so, potentially,extension costs. The realization of this potential is awaited. If it were to be substantially realized, however, it isunlikely to alleviate the problem of recurrent cost funding for a comprehensive national service, as the donor-financed projects traditionally carried a large share of costs, or made sufficient contribution to investments toprovide adequate incentive for government to temporarily allocate extension support costs under the particularproject.


experiences; an extension newsletter was also recently initiated and a concerted efforthas been made to develop a "field/implementation" culture.

3. Despite these positive aspects, there are a number of serious problems with currentextension development in AFR. In 1991 AFR held a workshop at Lilongwe2 in Malawi for its staffinvolved in extension to discuss country experience in implementing the Agricultural ServiceInitiative. In 1993 it held two similar workshops in Ghana and Cote d'Ivoire with national extensionand research staff. One of the keynote papers at the Lilongwe workshop highlighted some of themain issues affecting the AFR extension program:

" The frequent use of standard packages of recommendations for major agro-climaticzones meant that these were often irrelevant to the circumstances of farmers; thiscould be avoided if greater emphasis was given to understanding farming systems inproject preparation and in the formulation of recommendations during implementation.

* Continuous collaboration was essential between research and extension to ensure thattechnology was developed which was appropriate to farmer circumstances.

* Greater emphasis was needed in training staff to have a capacity to diagnose fieldconstraints and to involve farmers directly in resolving defined problems.

* A standard "template" approach to setting up national extension institutions should beavoided.

The Bank's agricultural extension field staff participating in the Lilongwe workshop reported on thestatus, problems, impact and recommendations for improvement in extension projects in 16 countries.Staff were encouraged to be self-critical, and a range of operational problems were mentioned, but thetwo most frequently mentioned issues were (a) inadequate technical capacity of staff and poordiagnostic skills, and (b) a scarcity of relevant technology on which to base an effective service.

4. National project staff in the 1993 Workshops4 in West Africa were also used to beconstructively critical and expressed a number of concerns about the progress of implementing theT&V system in their countries: inadequate provision of funds to operate the national services, needfor modification to suit the socio-cultural-administrative context of the particular country, need for

2. Venkatesan, V. et al. (1991).

3. The issue of scarcity of relevant technology for the majority of smallholders is an overriding constraint tosustained growth of agricultural production in AFR, and, consequently to the potential impact of extensioninvestments. Much of the research effort in the region has been devoted to crop breeding, following thesuccesses of the Green Revolution in Asia. The Asian success, however, relied largely on younger, more fertilesoils in favorable production environments. Africa, with its ancient soils and erratic rainfall over much of thecropping zones, has a different set of problems. Although germplasm is important, the more intractable problemsof crop husbandry and management (for which research is likely to generate less economic return than didgermplasm research for Asia), and these have not yet had the required research attention. Lynam, J.K. andBlackie, M.J. (1991).

4. Aruna Bagchee (1993).


congruence with farmers' needs in adopting a unified (polyvalent) extension system through T&V,poor research-extension linkage which was limiting the effectiveness of extension through slowgeneration of relevant technology, need for staff training sessions to adopt a problem-solving methodand to include communication skills, and the need for specific diagnostic surveys for specialcategories of farmers with particular problems (e.g., women farmers, pastoralists). The projectimplementers did, however, support the need for a single unified national extension system (asopposed to crop-specific services which had been very common in West Africa but also were used, tosome extent, in East Africa), and for the acceptance of a strategy to prevent fragmented approaches toextension through enclave projects. It was agreed that political and economic stability had anoverriding influence on the outcome of extension investments (thereby intimating that this should be aprerequisite for major extension investment).

5. A 1991 review of staff appraisal reports (SAR) 5commissioned by the TechnicalDepartment of AFR6 indicated a number of favorable factors. The most encouraging were a)support in 80 percent of SARs for adaptive on-farm research, including diagnostic surveys of typicalfarming systems, and b) in one case (Zaire pilot extension project), the involvement of NGOs andcotton companies in adaptive research and technology transfer. However, the report also highlighted:(a) some pilot projects (3 of 7) were not allowed to mature and indicate lessons before embarking ona national project (suggesting to OED that there was pressure to introduce a T&V national systemregardless of the pilot experience); (b) half of the SARs did not address the availability of technologyfor transfer; (c) training needs of staff were not addressed in half of the projects; (d) obviousdifficulties in meeting recurrent cost funding in earlier projects were not addressed, while in others(half) no information was provided on how the key research-extension linkage would be funded; (e)although some SARs provided support for mass media facilities and programs, none emphasized theuse of professional programming to maximize relevance and cost-effectiveness; (f) in most free-standing projects the status of complementary services (credit, input supply) was not analyzed; and (g)most provided some indication of the general smallholder target beneficiaries, but little description ofthe characteristic of the different major farming systems. Although it may be argued that theseapparent deficiencies were associated with an attempt to prepare a short SAR, the importance of theissues for effective and sustainable extension systems should warrant mention or specific reference tosupporting documents.

6. A further review of the current problems confronting T&V extension programs in sixEast African countries was provided by one AFR country department in 1992.7 This highlightedsimilar types of constraints to those noted in other internal AFR reviews, made constructive anddemanding recommendations, and concluded that "the T&V system has yet to show that it hassustainable answers for providing efficient extension services to small-scale farmers, particularly inrainfed agricultural (and livestock) systems, and the search for more effective mechanisms for thegeneration and dissemination of new technology continues."

5. Twenty-one SARs comprising fre-sUmdmg and extension component projects.

6. World Bank AFRTD Internal Report (1991).

7. Swartz, Lisa et al. (1992).


A PROCESS TO PREPARE AN EXTENSION STRATEGY AND INVESTMENT PROGRAM

1. The findings suggest that the Bank as a development institution should allocate moreattention and resources to the preparation phase of extension projects. This would raise the costs ofprocessing loans, but should have high development dividends in creating sustainable extensionsystems, and would provide a valuable training exercise for national personnel who could continue tocontribute to the management of responsive services.

2. The identification/preparation process would initially involve a comprehensive analysisof:

(a) production conditions: agro-ecological zones, characteristics of the major farmingsystems in each (including resource levels, constraints, labor usage, off-farmactivities, influence of gender or caste, education levels, land tenure andenvironmental issues), supporting infrastructure and the level and accessibility ofprivate and public sector services in each zone (inputs, markets, credit), andrelevant demographics;

(b) technology being used in, and the stock of technology relevant to, the majorfarming systems, and the national technology generation and/or adaptation capacity(i.e. private and public research system and its external and internal linkages);

(c) existing institutional rural services and communication:

* public sector organizations (resources, operational systems, coverage, linkagesand performance)

* existing human resources for agricultural services and the capacity of theeducational system to produce the required skills

* private sector commercial technical services (coverage of commodities orsystems, scale, competence, potential for expansion)

* NGOs working with farming communities (location, scale, specializations,resources, relationships with officialdom)

* farmer and community organizations (official and unofficial)

* literacy and education levels of the rural population and access to various formsof communication (radio, TV);

(d) macro and sectoral environment, funding and commitment:

* extent to which macroeconomic and sectoral policies and market conditions areconducive to use of improved technology to increase productivity


* expressed commitment of government to adopt and fund improved serviceswithin an agreed strategy and with acceptable terms of service for personnel

* historical analysis of funding of agricultural extension (and research) and ratiosagainst sectoral spending and agricultural GDP to assess historical commitmentand consistency in support for these services

* past, current and anticipated involvement of other donors in the subsector.

3. Various means can be used to undertake the pre-project analysis including collection andcollation of primary and secondary data. It is likely that rapid rural survey techniques to obtain aninitial understanding of farming systems would be required. Strong local participation by theborrower and implementing agencies to ensure commitment and ownership of the program would beessential.

4. Assuming that macroeconomic and sectoral policies are conducive to additionalinvestment in technology transfer, and the required commitment to improvement exists, then astrategy/investment program can be developed. This would include:

(a) agreement with government on which kinds of farmers and production systemsshould have priority for public extension. Considering the normal fundingconstraint, it is usually impractical to consider major extension investments toimprove services (particularly face-to-face services) to all farmers at once. Foodor export commodity production objectives, as well as equity and regionaldevelopment considerations may have to be taken into account.

(b) determining the justification for additional extension investment in particular areasin view of priority objectives, potential for improvement through improvedtechnology, and existing status of infrastructure and complementary services.

(c) developing an extension strategy which accounts for both private and public sectorservice interventions, and mass media and face-to-face communication systems.

(d) assessing the availability of trained human resources to implement the envisagedstrategy as an important determinant of the rate at which a program could bedeveloped, and accepting that a low basic education level would limit the capacityof a system to provide comprehensive services and to be responsive to farmers'needs.

(e) ensuring that a capacity to continually generate and adapt new, appropriatetechnology exists (but it must be realized that there would be substantial delays

1. The analysis of complementary services may lead to a decision to include the development or encouragement ofspecific complementary services in a primarily extension investment project or in a parallel project, or to a delayin increasing extension investment in the affected zone.


before new technology could be produced for environments in which relevantimproved technology is scarce or unavailable).

(f) with restraint on recurrent costs of the public extension service as an importantconsideration, defining and agreeing on the role of the private sector and NGOentities and their linkages with public extension in an initial program. Similarly, amass media and communication system would be developed (to include specializedassistance) to complement and, where appropriate, reduce the intensity of face-to-face services for particular types of clientele and subject matter. 2 The capacity offarmers of different socioeconomic status (or organizations representing them) tocontribute to the cost of public services (in cash or kind) would determine theextent to which this concept would be introduced to reduce budgetary costs andincrease accountability of staff to their clients; even small contributions bybeneficiaries can have significant benefits through participation/ownership.

(g) defining, for particular areas, the initial intensity and frequency of programmedface-to-face services (and, therefore, staffing and support) in accordance with theforegoing factors. This implies that there could be major variations in face-to-faceservices (reflected in farmer-extensionist ratios and mode of operation) throughoutthe areas serviced by public extension. Services to meet the established needs ofparticular groups according to gender or social class would receive specialattention. Application of many of the extension principles commonly used in T&Vextension models could be used, with appropriate modification, to develop theinitial phase of an improved face-to-face extension program; the attached Table 1illustrates this.

(h) realistically projecting the development and the recurrent cost implications of theextension program and each of its components.

5. An essential requirement in formulating the extension program would be to assessthe availability of human resources (as in 4(d) above) and construct a needs-based training program toattain the required skills and understanding in all facets of the program (this is distinct from regulartechnical training sessions for front-line extension staff in T&V models). Training should commenceat the start of the investment project (or pre-project). Key areas would be (i) practical training of

2. As for face-to-face services, the relevance of mass media material to the targeted audience would be critical.This implies that mass media programming should be significantly dependent on feedback from face-to-facesurveys or services. See also Academy for Educational Development reports on USAID-funded project"Communication for Technology Transfer" for examples of close integration of mass media and farmerinteraction in pilot programs in Peru, Honduras, Indonesia and Jordan. A guide for use of these techniques hasbeen prepared by the Academy. See Mata, 1.1. (1992).

3. These might include models close to T&V for particular locations and time periods, extension campaigns onparticular issues, commodity-specific extension programs aligned with delivery of an input package, or highlyparticipatory schemes involving extensionists and farming communities (and often researchers) in an infrequentbut scheduled program of interaction for problem definition and resolution.


technical specialists; (ii) training of trainers in small group dynamics and problem census4 techniquesto show how to increase client participation5 and utilize interaction within farmer groups to assist inthe technology transfer process in face-to-face services; and (iii) education of managers and policymakers in the principles being used in formulating an effective and efficient service.

6. The program developed would not be seen as a permanent format. Emphasis on localinvolvement in project preparation provides the opportunity to train personnel in the management andextension principles used in the strategy, so that a capable extension planning unit can be establishedto monitor and analyze implementation. This unit should make recommendations for programmodifications in accordance with experience and changing circumstances to provide a responsiveservice.

7. The expenditure of a large amount of scarce resources on extensive surveys to quantifyextension impact and economic benefits appears questionable for many countries, due to the extremedifficulty in linking cause and effect even if reliable statistical data can be obtained. It is probablymore appropriate to concentrate monitoring and evaluation (M&E) survey resources on (a)measuring the change in adoption of technologies promoted by the extension agents in their specificareas of influence; (b) understanding the reasons for significant non-adoption (implying the proceduresto ensure relevant recommendations may have fallen down); and (c) questioning the intended farmerclientele on how they perceive the extension services offered by the public extension organizations, bytheir particular extension agents, and by other private or public service institutions. In other words,M&E resources would be concentrated on improving the effectiveness and efficiency of the extensionsystem and directly asking the intended beneficiaries of the impact it is having. It would be geared toongoing evaluation rather than quantitative ex-post evaluation, but would contribute to ex-postassessment.

8. Surveying of farmers would be complemented by collation of other primary andsecondary data which provide (a) indications of how the program is being implemented, (b) input andoutput indicators of changes in use of technology and in production to which the extension programhas likely contributed, (c) analysis of factors involved in effectiveness (funding, linkages, training),and (d) sustainability indicators (progressive annual budgets and expenditures, staffing, involvement ofthe private sector). The use of performance indicators and their relationship with M&E measures willbe elaborated in the final study report.

4. 'Problem census* is a process of having farmers define their own objectives and problems through structuredgroup discussion (see OED Report No. 10661. 1992. Nepal, para. 5.45; OED Report No. 11980. 1993.Indonesia, para. 5.1; and Crouch, B. 1984).

5. There is little argument concerning the advantage of high participation by the extension beneficiaries in theextension process.

-71- ANNEX 4 - Table 1Page 5 of 5

EXTENSION PRINCIPLE APPLICATION IN AN IMPROVED T&V SYSTEM

Programming of Activities maintain, but program to be determined according to the needs ofthe farmer group (as agreed with them) as well as the availabilityand appropriate timing for treatment of specific technologies.

Fixed visit schedule maintain, but in accordance with a schedule agreed as appropriatebetween farmers, field staff and supervisor, rather than as ageneral fortnightly standard.

Supervision maintain, but with more emphasis on assessing reactions of farmerclientele than on the mechanics of the visit schedule.

Single line of command maintain (but not as an essential requirement), but develop anunderstanding that the extension staff must be responsive to theneeds of the farmer clientele whose reactions will be monitored bysupervisory staff as a measure of staff performance.

Concentration on maintain, but not to the exclusion of facilitating the provision ofextension of technology services by other agencies (private and public) to meet the needs

of the farming communities being attended, nor of assistingfarmers to develop a self-help capacity (e.g. seedling production,group marketing).

Technical Content and maintain timely training sessions for field staff geared to subjectsTraining of proximate visits, but not as district-wide general

recommendations associated with dominant commodities; thewidely applicable recommendations would be taken into accountby training officers, but training would be specific to the require-ments of the program previously agreed between staff and farmergroups; general recommendations would be adapted andincorporated where possible into the solutions to the problemsidentified with farmer groups.

Use of Contact Farmers or adopt a concept of identifying existing groups of farmers withGroups for Visits and similar resource circumstances, problems and attitudes as theDemonstrations medium for interaction with the extension staff (as groups or

subgroups in the recipient community), and use these groups todefine major constraints, analyze the relevance of technologicalsolutions, and assume ownership of demonstrations.

Research-extension linkage maintain and expand, with emphases on ensuring relevance inapplied/adaptive research; this implies a farming systemperspective in both research and extension, which would benefitby joint farming system diagnostic surveys.

-72- ANNEX 5

Page 1 of 3

BIBLIOGRAPHY

Antholt, C.H. (1991). Agricultural Extension in the 21st Century: Lessons from South Asia.Agricultural Extension: Worldwide International Evolution and Forces for Change, eds. W.Rivera and D. Gustafsen, New York. Elsevier.

Aruna Bagchee (1993). Agricultural Extension in Aftica. Report of African Extension workshops Accra,Ghana and Abidjan, Cote d'Ivoire. AFR Technical Department and Economic DevelopmentInstitute.

Axinn, G. (1988). Guide on Alternative Extension Approaches. FAO.

Baxter, M., R. Slade, and J. Howell (1989). Aid for Extension. Washington, D.C. World Bank.

Benor, D. and J.Q. Harrison (1977). Agricultural Extension: The Training and Visit System.Washington, D.C. World Bank.

Benor, D. and M. Baxter (1984). Training and Visit Extension. Washington, D.C. World Bank.

Bindlish, V. and R. Evenson (1993). Evaluation of the Performance of T&V Extension in Kenya. WorldBank Technical Paper No. 108, Africa Technical Department Series.

Bindlish, V., R. Evenson and M. Gbetibouo (1993). Evaluation of T&V-Based Extension in BurkinaFaso. World Bank Technical Paper No. 226, Africa Technical Department Series.

Birkhaeuser, D., R.E. Evenson and G. Feder (1991). The Economic Impact of Agricultural Extension: AReview. Economic Development and Cultural Change. Vol. 39, No. 3.

Cernea, M.M., J.K. Coulter and J.F. Russell (1983). Strengthening Extension for Development: CurrentIssues and Prospects. World Bank/UNDP Symposium.

Crouch, B. (1984). Problem Census: Farmer-centered Problem Identification. Training for Agricultureand Rural Development. Rome, FAO.

Evenson, R.E. (1991). Research and Extension in Agricultural Development. International Center forEconomic Growth. Occasional Paper No. 25. Panama.

FAO Investment Centre (1992). Sociological Analysis in Agricultural Investment Project Design.Technical Paper No. 9.

FAO Investment Centre (1991). China: Development of Low-lying Lands in Baicheng Prefecture (JilinProvince). Socio-economic and Farming Systems Study. FAO/IFAD Cooperative Programme.Report No. 1/91 IF-CPR 42.


FAO Investment Centre (1992). China: Yunnan (Simao) Minority Areas Agricultural DevelopmentProject. Socio-economic and Production Systems Diagnostic Study. FAO/IFAD CooperativeProgramme. Report No. 64/92 IFAD-CPR 47.

FAO Investment Centre (1992). Myanmar: Border Hills Area Development Project Socio-economic andProduction System Study. Identification Mission. FAO/IFAD Cooperative Programme. ReportNo. 41/92 IFAD-MYA 35.

FAO Investment Centre (1993). Zambia: Southern Province Food Security Project. Socio-economic andProduction Systems Study. FAO/IFAD Cooperative Programme. Report No. 137/93 IFAD-ZAM SEPSS.

Feder, G., L.J. Lau and R.H. Slade (1985). The Impact of Agricultural Extension. A Case Study of theT&V System in Haryana, India. World Bank Staff Working Paper No. 756.

Lynam, J.K. and M.J. Blackie (1991). Building Effective Research Capacity: The African Challenge.Agricultural Technology for Sustainable Economic Development in the New Century: PolicyIssues for the International Community. World Bank seminar at Airlea House, Virginia.

Macklin, M. (1992). Agricultural Extension in India. World Bank Technical Paper No. 190.

Mata, Jose Ignacio (1992). CTTA: A method for the Transfer of Technology to Farmers: A Guide forPlanning and Implementtion. An Academy for Educational Development Publication (Spanish andEnglish). USAID contract D.P.E. 5826-c-00-5054-00.

Rajesekaran, B., R.A. Martin and D.M. Warren (1993). A Framework for Incorporating IndigenousKnowledge Systems into Agricultural Extension. Indigenous Knowledge and DevelopmentMonitor. Vol. 1, No. 3.

Ravnborg, H.M. (1993). Agricultural Research and the Peasants: The Tanzanian Agricultural Knowledgeand Information System. Centre for Development Research. Denmark.

Sims, H. and D. Leonard (1990). The Political Economy of the Development and Transfer ofAgricultural Technologies. Making the Link.* Agricultural Research and Technology Transfer inDeveloping Countries, ed. David Kamowitz. Boulder: Westview Press.

Swartz, Lisa and Jacob Kampen (1992). Agricultural Extension in East Afica. World Bank TechnicalPaper No. 164.

Venkatesan, V. and L. Swartz (1991). Agricultural Services Initiative in Africa. Internal AFRTDdocument. World Bank.

World Bank. 107 PARs and PCRs as listed in Annex 1, Tables 1 and 2. These reports are available toBank staff and the respective borrowers, but others can access individual documents through theirrespective Bank Executive Directors.


World Bank (1983). Strengthening Agricultural Research and Extension: The World Bank Experience.OED Report No. 4634.

World Bank (1990). Agricultural Extension: The Next Step. Agriculture and Rural DevelopmentDepartment. Policy and Research Series 13.

World Bank, AFRTD Internal Report (1991). Staff Appraisal Reports for Extension Projects in theAfrica Region: Characteristics, Issues and Proposal Guidelines.

World Bank (1993). Dynamics of Rural Development in Northeast Brazil: New Lessons from OldProjects. OED Report No. 10183.

Zijp, W. (1991). From Extension to Agricultural Infomation: Issues and Recommendations from WorldBank Experience in the Middle East and North Africa. World Bank. Europe, Middle East andNorth Africa Region, Technical Department Paper 1.2.

Report No. Agricultural Extension: Lessons from Completed...

Documents

Transcript of Report No. Agricultural Extension: Lessons from Completed...