VORLD BANK OPERATIONS EVALUATI ON DEPARTM E NT 20744 June 2000 Agricultural Extension The Kenya Experience An Impact Evaluation lAVM OPERATIONS EVALUATION DEPARTMENT ENHANCING DEVELOPMENT EFFECTIVENESS THROUGH EXCELLENCE AND INDEPENDENCE IN EVALUATION The Operations Evaluation Department (OED) is an independent unit within the World Bank; it reports directly to the Bank's Board of Executive Directors. OED assesses what works, and what does not; how a borrower plans to run and maintain a project; and the lasting contribution of the Bank to a country's overall development. The goals of evaluation are to learn from experience, to provide an objective basis for assessing the results of the Bank's work, and to provide accountability in the achievement of its objectives. It also improves Bank work by identifying and disseminating the lessons learned from experience and by framing recommendations drawn from evaluation findings. WORLD BANK OPERATIONS EVALUATION DEPARTMENT m Agricultural Extension The Kenya Experience An Impact Evaluation Madhur Gautam 2000 The World Bank www.worldbank.org/html/oed Washington, D.C. Copyright @ 2000 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First edition June 2000 The opinions expressed in this report do not necessarily represent the views of the World Bank or its member governments. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and will normally grant permission promptly. Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use is granted by the World Bank, provided that the appropriate fee is paid directly to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, U.S.A., telephone 978-750-8400, fax 978-750- 4470. Please contact the Copyright Clearance Center before photocopying items. For permission to reprint individual articles or chapters, please fax your request with complete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the Office of the Publisher, World Bank, at the address above, or faxed to 202-522-2422. Photo credits: cover, page 10, and page 28, Curt Carnemark. ISBN 0-8213-4758-6 Library of Congress Cataloging-in-Publication Data Gautam, Madhur, 1960- Agricultural extension : the Kenya experience / Madhur Gautam. p. cm. - (Operations evaluation study) Includes bibliographical references. ISBN 0-8213-4758-6 1. Agricultural extension work-Kenya. I. Title. II. World Bank operations evaluation study. S544.5.K4 G38 2000 630'.715'096862-dc21 00-032087 4 Printed on recycled paper. Contents v Acknowledgments vii Foreword, Prefacio, Pr6face xi Executive Summary, Resumen, R6sum6 Analytique xxi Abbreviations and Acronyms 1 1. Introduction 2 Study Objective 2 The National Extension Projects 3 Study Strategy and Outline 7 2. Institutional Development 7 Before 8 And After 10 Conclusions 13 3. Beneficiary Assessment 13 Welfare and Productivity 13 Access to Services 14 Interactions Between Farmers and Agents 14 The Relevance of Agents' Advice 15 Farmers' Priorities 15 Conclusions 17 4. Efficacy: The Quantity and Quality of Contact 17 Program Design 17 Access to Information 18 The Contact Farmer Approach 19 System Performance 19 Biases 20 Farmer Assessment of Extension Advice 20 The Suspension of NEP II's Disbursements 20 Conclusions 23 5. Outcomes: Farmer Awareness and Adoption 23 Awareness 24 Adoption 25 Statistical Tests of Extension's Impact 25 Conclusions 27 6. Results 1: Farmer Efficiency and Productivity Change 28 Relative Efficiency 29 Productivity Change 30 Conclusions 31 7. Results II: Production Effects of Extension 31 A Fixed-Effects Approach 32 Findings 32 Conclusions 111 Agricultural Extension: The Kenya Experience 33 8. Client Focus: Farmer Valuation of Extension Benefits 33 Desired Frequency of Visits 34 Willingness to Pay 34 Factors That Influence Willingness to Pay 34 Conclusions 37 9. Conclusions and Lessons 37 Lessons and Recommendations 39 Annexes 39 Annex A. Background 41 Annex B. The Design of NEP I and NEP II 43 Annex C. Institutional Features 47 Annex D. Focus Group Briefing on NEP's Impact 49 Annex E. Awareness and Adoption of Extension Messages 51 Annex E The Contingent Valuation Method and Its Application 53 Annex G. Comments from the Government of Kenya 57 Annex H. OED's Response to Borrower Comments 59 Annex I. Household Survey Data 61 Endnotes 65 Bibliography Tables 8 2.1 NEP I and II: Borrower Perspectives 14 3.1 Changes in Farmers' Access to Services and Service Quality Figures 3 Figure 1.1 Stylized Impact Model 18 Figure 4.1 The Availability of Information: Now, Relative to 10-15 Years Ago 18 Figure 4.2 Normal Place and Frequency of Extension Agent-Contact Farmer Meetings 19 Figure 4.3 Extension Staff Allocation and Poverty by District 20 Figure 4.4 Impact of the Suspension of NEP II Disbursements 24 Figure 5.1 Farmers' Awareness of Simple-to-Complex Messages, 1982-97 25 Figure 5.2 Farmers' Adoption of Extension Recommendations, 1982-97 29 Figure 6.1 Productivity Change, 1982-97: Trends in Cluster Averages Boxes 39 Box A.1 The Difficulty of Drawing Inferences from Field Visits 1V YW. Acknowledgments This study presents the main findings of an evaluation Evenson, the Rural Household Budget Survey (RHBS) of the impact of the World Bank's support of projects data for 1982 by Steven Block. for agricultural extension services in Kenya during So many people in Kenya contributed to this 1982-98. Madhur Gautam, Evaluation Officer, pre- evaluation that it is not possible to acknowledge them pared this report; it was peer reviewed by Gershon all individually. The many Kenyan farmers who Feder, John Heath, and David Nielson. William patiently participated in the household survey and its Hurlbut provided editorial assistance. Helen Claverie, pretests, and those subjected to the many field visits Constance Frye, Maisha Hyman, and Janet Wyse over the course of this evaluation made invaluable provided administrative assistance at various stages of contributions. The support and assistance of the Minis- the evaluation. Detailed analyses of the Bank's agricul- try of Agriculture, particularly of the extension staff at tural extension work in Kenya are contained in the six Kilimo House, in the study districts and in other working papers prepared in support of this study (see districts visited were instrumental as the evaluation the Bibliography). was being conducted. Generous contributions were Madhur Gautam performed the overall analysis, made by many staff of the Kenya Agricultural Research with contributions from Jock Anderson in Working Institute and the International Research Centers in Paper 3, Jonathan Alevy in Working Paper 4, and Amy Nairobi, as well as by members of the donor commu- Gautam in Working Paper 6. The Tegemeo Institute of nity and civil society in Kenya. Finally, the author is Egerton University undertook the household survey and grateful to his many colleagues in the resident mission data processing under the overall supervision of Gem and at the Bank's headquarters for the support they Argwings-Kodhek. The team included supervisors provided. Thomas Awuor and Francis Karin and assistant super- This study was published in the Partnerships and visors Betty Kanunga and Bridget Ochieng, while Knowledge Group by the Dissemination and Outreach Joshua Ariga performed the editing. Enumerators were unit, directed by Elizabeth Campbell-Pag6 (task team Stephen Irungu, Joshua Kerima, Isaack Maritim, Salim leader), including Caroline McEuen (editor), Kathy Mgallah, Geoffrey Oganga, Daniel Ouma, Tony Strauss and Aichin Jones (graphics and layout), Diana Sisule, Milka Wangare, Alex Wanjohi, and Lucy Qualls (editorial assistant), and Juicy Qureishi-Huq Wanyanga. Willis Oluoch-Kosura of the University of (administrative assistant). Nairobi conducted the staff survey and data analysis, and Lutta Muhammad of the Kenya Agricultural Research Institute at the Katumani Research Station Director-General, Operations Evaluation: Robert Picciotto compiled the inventory of technology and extension messages. Allan Musamali and Martin Mutuku of the Ministry of Agriculture compiled the time series data Manager, Sector and Thematic Evaluations: Gregory Ingram on extension services in the study districts. The Africa Task Manager: Madhur Gautam Technical Department (ATD) household survey data for 1990 were graciously made available by Robert v  Foreword FOREWORD .PREFACIO PREFACE The Training and Visit (T&V) El sistema de gerencia que se En 1982, la Banque mondiale a system of management was intro- basa en la formación y las visitas (el mis en place au Kenya le système de duced in Kenya by the World Bank sistema T&V) fue introducido en gestion basé sur la formation et les in 1982 as a pilot, financed by the Kenia por el Banco Mundial en 1982 visites. À ses débuts, ce projet pilote Second Integrated Agricultural como práctico financiado por el était financé par le deuxième projet de Development Project. It has since Segundo Proyecto Integrado para el développement agricole intégré. been supported by the National Desarrollo de la Agricultura. Desde Depuis, il a bénéficié de l'aide du projet Extension Project (NEP 1), approved entonces, ha sido sostenido por el NEP I (approuvé en 1983) et du Projet in 1983, and the Second National Proyecto Nacional de Extensión (el NEP II (mis en œuvre en 1991). Extension Project (NEP II), which NEP 1), aprobado en 1983, y el Les projets de la Banque avaient became effective in 1991. Segundo Proyecto Nacional de deux objectifs : la consolidation The Bank projects had two Extensión (el NEP II), que entró en institutionnelle des services de objectives: institutional develop- efecto en 1991. développement et l'augmentation ment of the extension service and Los proyectos del Banco tuvieron soutenue de la productivité agricole. sustained increases in agricul- dos objetivos: el desarrollo L'efficacité des services de tural productivity. The effective- institucional del servicio de extensión développement est cependant ness of the extension approach y unos incrementos sostenidos de la controversée. Certains trouvent qu'ils adopted by the projects has been productividad agricola. Se ha coûtent fort cher pour le peu d'impact a subject of debate because of the discutido la eficacia del enfoque sobre qu'ils ont sur la production agricole. perceived high cost and an appar- la extensión adoptado por estos Cette controverse fait partie d'un ent lack of impact on agricultural proyectos a causa de sus altos costos débat plus large sur l'efficacité au production. percibidos y su falta de tener un efecto Kenya du système formation-visites. This debate has been part of a apreciable en la producción agricola. Les discussions portent largement sur broader disagreement on the effec- Este debate forma parte de un leur efficacité comparée à celle d'autres tiveness of the T&V approach to desacuerdo más amplio cuyo sujeto es systèmes d'encadrement. Alors qu'on extension. The arguments have la eficacia del uso del sistema T&V en s'accorde généralement pour dire que focused on its efficacy relative to la extensión. Los argumentos se han le système formation-visites coûte cher, alternative mechanisms for deliv- enfocado en la eficacia de este sistema la controverse porte essentiellement ering extension advice. While it is relativa a los mecanismos alternativos sur leurs répercussions sur la generally agreed that the T&V para entregar avisos de extensión. production agricole. Malgré l'intensité system is costly, the controversy Mientras que todo el mundo està de de la discussion, on ne s'est guère centers on its impact on agricul- acuerdo en que el sistema T&V es efforcé d'évaluer avec rigueur l'impact tural production. Despite the caro, la controversia se concentra en des projets de type formation-visites. intensity of the debate, however, su efecto en la producción agricola. La présente évaluation a adopté there have been very few attempts Sin embargo, a pesar de la vehemencia une approche théorique afin de réunir to rigorously establish the impact de este debate, pocos ensayos se han des preuves scientifiques qui of T&V projects. hecho para rigurosamente determinar démontreront avec rigueur les This evaluation adopted a el efecto de los proyectos T&V. répercussions des projets. En utilisant theory-based approach to gather a Esta evaluación adoptó un un cadre de gestion axé sur les credible body of empirical evi- enfoque basado en la teoria para résultats, on y évalue les principaux dence to rigorously establish the recoger una colleción convincente de indicateurs des résultats à diverses impact of the projects. Following a evidencia empirica para rigurosamente étapes du projet afin de déterminer le results-based management frame- establecer el efecto de estos proyectos. rendement du système de vii Agricultural Extension: The Kenya Experience work, key indicators are Siguiendo un estilo de gerencia développement kenyan. evaluated at the varlous que se basa en los resultados, se Les résultats de l'évaluation stages of the results chain to evalúan los indicadores claves a indiquent que les projets ont eu assess the performance of the varias etapas de la cadena de des répercussions limitées sur la Kenyan extension system. resultados para valorar el consolidation institutionnelle. The evaluation finds funcionamiento del sistema de Par contre, ils ont eu des effets that the projects have had extensiôn en Kenia. positifs : ils ont étendu leur limited institutional devel- La evaluación halla que los couverture géographique, opment impact. They have had proyectos tuvieron un efecto limitado resserré (bien que tardivement) les some beneficial impacts in in- en el desarrollo de las instituciones. liens entre le développement et la creased geographical coverage, Tuvieron algunos efectos beneficiales recherche, et amélioré la qualité du improved research-extension link- en cuanto al alcance geográfico más personnel par la formation. Mis à part ages (albeit belated), and amplio, los enlaces más fuertes entre la ces quelques points positifs, improved staff quality through extensión y la investigación (aunque l'évaluation a conclu que les systèmes training. Overall, however, this con retraso), y la mejor calidad del de développement actuels ne sont ni evaluation found the current personal gracias a la formaciôn. Sin efficaces ni rentables à offrir aux extension system to be ineffective embargo, los resultados de esta paysans les services dont ils ont and inefficient in delivering the evaluaciôn indicaron que, en besoin. Cependant, la conception needed services ta farmers. The conjunto, el sistema de extensión institutionnelle, qui reflète les objectifs institutional design has lacked a corriente es ineficaz e incapable de des projets, ne tenait pas suffisamment focus on farmer empowerment. As entregar a los campesinos los servicios compte de la nécessité d'augmenter such, inappropriate incentives que necesitan. Le ha faltado al disefño l'autonomie des paysans. De plus, des have resulted in a lack of account- institucional un efoque en apoderar a incitatifs inadaptés aux besoins ont ability or responsiveness ta the los agricultores. Como tanto, engendré un manque de clients' needs. Most important, incentivos inapropiados resultaron en responsabilisation et de sensibilisation the system is not financially una carencia de contabilidad o aux besoins des clients. Enfin, ce sustainable. sensibilidad a las necesidades de los système n'est ni viable ni rentable. A distinction needs to be main- clientes. Más importante, el sistema no Il ne faut pas confondre la tained between the relevance of se sostiene financieramente. pertinence des services de extension services per se and the Hay que hacer una clara développement avec celle de la relevance of the projects' design. diferenciación entre la relevancia de los conception des projets. Par exemple, The rationale for providing exten- sistemas de extensión de por si y la l'offre de services de développement sion services is still relevant; how- pertinencia del disefño de los aux petits exploitants du Kenya se ever, several features of the proyectos. Aunque todavia vale la justifie toujours. Toutefois, les projects' design proved to be inap- pena ofrecer servicios de extensión, données recueillies donnent à penser propriate. The available evidence algunas características del diseio de que l'approche adoptée manque suggests that the extension los proyectos resultaron ser d'efficacité. Ces mêmes données ne approach applied in Kenya was not inapropiadas. La evidencia disponible permettent pas de déterminer les efficacious. The physical impact of sugiere que el enfoque en la extensión répercussions physiques des services the extension services also cannot que se llevaba a cabo en Kenia no fue de développement. Même s'il est be established with the current- efectivo. Tampoco se puede establecer probable que le NEP I ait eu, à ses data. While it is likely that there el efecto fisico de los servicios de débuts, des répercussions positives sur was a positive impact on farmer extensión utilizando los datos la productivité et l'efficacité des productivity and efficiency in the corrientes. Mientras es probable que el fermiers, cela n'a pas duré. En effet, les initial years of NEP 1, the benefits NEP I rindió algunos beneficios données recueillies en 1990 appear to have been short-lived, as tempranos, no parecen haber durado n'indiquent pas de progrès notables. even the 1990 data do not indicate mucho, y atn los datos de 1990 no La présente évaluation n'indique any significant impact. This evalu- indican ningún efecto significativo. aucune répercussion appréciable sur ation thus could not establish a Como consecuencia, esta evaluacién les résultats ni sur les indicateurs de viii Foreword significant impact for either no podia establecer ningùn résultats. Elle n'a pas non plus the key outcome or the re- efecto significativo para los engendré un taux de rendement sults indicators; nor could it indicadores claves del desenlace positif dans les dépenses de establish a positive rate of o de los resultados, ni tampoco développement agricole. return to the expenditures on era capaz de establecer una tasa Cinq grandes leçons agricultural extension. positiva de rendimiento con ressortent de cette évaluation. Five main lessons emerged respeto a los gastos para la Premièrement, il est from this evaluation. First, there is a extensión agricola. nécessaire de mieux cibler les zones et need for more efficient targeting of Cinco lecciones principales se les groupes où l'impact marginal est extension services to focus on groups aprendieron de esta evaluación. La susceptible d'être le plus important. and areas where the marginal impact primera lecciôn es la necesidad de D'où le besoin d'un système plus is likely to be the greatest. This, in seleccionar màs eficientemente las souple et plus « ingénieux », capable turn, calls for a more flexible and a áreas y los grupos para los cuales el d'identifier le fossé qui sépare les "smart" system that can identify gaps impacto marginal de los servicios de meilleures pratiques de celles between average and best practices, extensión va a ser el màs importante. courantes et d'allouer de façon and allocate scarce resources more Para esto, se necesita un sistema más rationnelle les ressources limitées. rationally. flexible e "inteligente" que pueda Deuxièmement, pour bien Second, to guide the "smart" identificar las diferencias entre las orienter le système « ingénieux » et system and to target better, it is mejores prácticas y las prácticas mieux cibler l'action, les services de necessary to have timely flows of medias y asignar los recursos escasos développement doivent disposer relevant management information de una manera más racional. régulièrement d'information de and continuous evaluation to pro- Segundo, para guiar el sistema gestion pertinente et procéder à une vide in-time feedback. Hence, there "inteligente" hacia una selecciôn más évaluation continue pour fournir de la is a need for a reliable monitoring eficiente, se necesita un flujo oportuno rétroaction au moment opportun. Il and evaluation system. de datos relevantes para la gerencia y est donc indispensable de mettre en Third, commensurate with the una evaluaciôn continua que pueda place un système d'évaluation et de pace of technology generation and proveer información de vuelta en contrôle fiable. the demand for advice, the inten- buena hora. Para esto es necesario un Troisièmement, le service de sity of the extension service needs sistema fiable de monitoreo y développement doit s'adapter aux to be suited for particular circum- evaluacion. circonstances particulières en tenant stances. In some areas, a leaner Tercero, proporcionado con el compte du rythme du développement and less-intensive presence (in paso de la generación de la tecnologia technologique et de la demande de extension staff per farm house- y la demanda de avisos, la intensidad conseils. Dans certaines régions, une hold) would allow a wider del servicio de extensién tiene que présence réduite et modérée (nombre geographical coverage and may adaptarse a las circunstancias d'agents de développement par ferme) be more cost-effective. This calls particulares. En algunas áreas, una étendrait la couverture géographique for a more demand-driven and presencia menos fuerte e intensa (en et peut-être la rentabilité. Cela responsive delivery system. términos del número de personal de nécessite un système mieux adapté et Fourth, a blanket approach extensiôn por cada granja) permitiría axé sur la demande. using a uniform methodology in un alcance geográfico más amplio y Quatrièmement, une approche al circumstances is also unlikely puede ser más rentable. Esto exigiria générale qui utilisera la même to be effective. A more pluralistic un sistema de entrega más sensitivo y méthodologie quelles que soient les approach that exploits the syner- empujado por la demanda. circonstances a peu de chance d'être gies among the extension service, Cuarto, un enfoque general que efficace. Une approche imprégnée de low-cost modern communications, se fija de una metodologia uniforme pluralisme pour exploiter les synergies demonstrations, printed media, para todas las situaciones tampoco va qui existent entre les services de and partnerships with civil society a ser efectivo. Hay que desarrollar un développement, les moyens de and the private sector needs to be enfoque de naturaleza más pluralista communication modernes peu developed. que se aprovecha de las sinergias entre coûteux, l'organisation de ix Agricultural Extension: The Kenya Experience And finally, the service el servicio de extensiôn, las démonstrations, le recours aux delivery mechanism needs to comunicaciones modernas de journaux et la création de fully incorporate client focus. bajo costo, las demostraciones, partenariats avec la société civile This requires considering la prensa, y las asociaciones con et le secteur privé doit être mise alternative options such as 1 la sociedad civil y el sector au point. cost-sharing, farmer organi- privado. Cinquièmement, le zations, decentralization, and Finalmente, el mecanismo mécanisme d'offre de services the like as an integral part of the para entregar los servicios tiene doit intégrer des moyens pour les delivery mechanism. que incorporar un enfoque en el paysans de se faire entendre. Cela cliente. Esto requiere la consideración exige d'envisager d'autres options de alternativos tales como la comme le partage des coûts, repartición de costos, el apoyo a los l'incitation à l'établissement organismos que representan a los d'organismes de paysans et la campesinos, y la decentralizaciôn, décentralisation entre otros, como parte integral del mecanismo de entrega. Robert Picciotto Director-General, Operations Evaluation Department X Executive Summary EXECUTIVE RESUMEN RESUME SUMMARY ANALYTIQUE Kenya's extension services En 1982, los servicios de Les services d'extension du adopted the training and visit (T&V) extensiôn de Kenia adoptaron un Kenya ont adopté, en 1982, un management system in 1982, and the sistema de gerencia basado en la système de gestion basé sur la World Bank supported the system formacién y las visitas ("training and formation et les visites. La Banque through the first and second National visit" o "T&V"), y el Banco apoyô a mondiale a appuyé ce nouveau Extension Projects, NEP I and NEP H. este sistema en el transcurso de los système dans le cadre des deux The projects had two objectives: the proyectos nacionales de extensión, el premiers projets, NEP I et NEP Il, qui institutional development of the NEP I y el NEP Il. Estos proyectos visent la consolidation institutionnelle extension services and sustained tuvieron dos objetivos: el desarrollo des services d'extension et increases in agricultural productivity. institucional de los servicios de l'augmentation soutenue de la The effectiveness of the exten- extensión y unos incrementos productivité agricole. sion services has been subject to sostenidos de la productividad agricola. L'efficacité des services d'extension est cependant controversée. Certains trouvent qu'ils coûtent fort cher et qu'ils ont peu d'impact sur la production agricole. Ce débat fait partie d'un différend plus large sur l'efficacité au Kenya des systèmes formation-visites. Les discussions portent largement sur leur efficacité comparée à celle d'autres systèmes d'encadrement. Alors qu'on convient généralement que le système formation-visites coûte cher, les P retombées des gros investissements des pays emprunteurs et, par conséquent, leurs répercussions sur la production agricole animent la controverse. Malgré l'intensité de la discussion et les prêts importants de la Banque, on ne s'est guère efforcé debate, however, because of its pet- Sin embargo, se ha discutido la d'évaluer avec rigueur l'impact des ceived high costs and apparent lack eficacia de los servicios de extensión projets basés sur la formation et les of impact on agricultural produc- porque existe la percepciôn de que visites. tion. This debate is part of a broader cuestan mucho y tienen poco efecto en la En adoptant une approche disagreement about the effectiveness producción agricola. Este debate forma théorique, cette étude a réuni de façon of the T&V approach. Arguments parte de un desacuerdo más amplio méthodique des renseignements qui have focused largely on its efficacy cuyo sujeto es la eficacia del uso del ont permis de déterminer l'impact relative to alternative mechanisms sistema T&V en Kenia. Los argumentos probable des projets. En utilisant un for delivering extension advice. se han enfocado sobre todo en la cadre de gestion axé sur les résultats, While it is generally agreed that the eficacia de este sistema relativa a los l'étude a essayé de relier les résultats T&V system is costly, the contro- mecanismos alternativos para entregar observés dans les champs des fermiers xi Agricultural Extension: The Kenya Experience versy centers on returns to avisos de extensiôn. Mientras que iaux intrants des projets. De borrowing countries' high lev- todo el mundo está de acuerdo en plus, des indicateurs els of investment, and hence que el sistema T&V es caro, la intermédiaires de la production on its impact on agricultural controversia se concentra en el et des résultats ont été mesurés production. Despite the inten- rendimiento de las altas tasas de afin d'évaluer le rendement du sity of the debate and the high inversión de los países que piden système d'extension par volume of the Bank's lending, préstamos y asi en su efecto en la rapport aux résultats. very few attempts have been made to producción agricola. A pesar de rigorously determine the impact of la vehemencia del debate y el gran Constatations principales T&V projects. volumen de los préstamos concedidos Consolidation institutionnelle Adopting a theory-based por el Banco, pocos ensayos se han Les projets NEP I et NEP Il ont abouti approach, this study systemati- hecho para rigurosamente determinar el à une consolidation institutionnelle cally gathered a credible body of efecto de los proyectos T&V. limitée. Le projet NEP I a introduit le evidence to establish the likely Adoptando un enfoque basado concept formation-visites comme impact of the projects. Following en la teoría, este estudio étant « d'abord et avant tout un a results-based management sistemáticamente recogió una colleción système de gestion ». Dix-sept ans framework, the study sought to de evidencia empirica para establecer el plus tard, l'efficacité de ces services relate the results observed in the efecto probable de estos proyectos. d'extension ne semblent pas avoir fait farmers' fields to the projects' Siguiendo un estilo de gerencia que se de progrès notables. Il leur manque inputs. In addition, intermediate basa en los resultados, el estudio une vision stratégique pour l'avenir, et output and outcome indicators intentó relacionar los resultados leur gestion demeure insuffisante et were measured in order to assess observados en los campos de los presque dépourvue de systèmes the performance of the extension agricultores a los ingresos de los d'information. Une mauvaise gestion system along the results chain to proyectos. Además, se midieron los financière empêche les capitaux confirm the potential for impact. indicadores intermedios de productos d'exploitation d'être disponibles à y resultados para evaluar el temps. L'inefficacité persistante des Principal Findings funcionamiento del sistema de services d'extension pousse le Kenya, Institutional Development extensión a lo largo de la cadena de appuyé par la Banque et d'autres Institutional development under NEP resultados para confirmar el potencial bailleurs de fonds, à tenter de les I and NEP Il was limited. NEP I del sistema de tener un impacto. rationaliser en adoptant d'autres introduced T&V as "first and fore- méthodes. most a management system." After Conclusiones principales Une fois lancés à l'échelle 17 years, however, the effectiveness Desarrollo institucional nationale grâce au système de gestion of extension services does not appear El desarrollo institucional bajo los formation-visites, les projets ont to have improved appreciably. A NEP I y Il fue limitado. El NEP I étendu leur couverture géographique, strategic vision for the future develop- introdujo el sistema T&V como "un resserré (bien que tardivement) les ment of the extension system is lack- sistema de gerencia ante todo." Sin liens entre l'extension et la recherche, ing, and management continues to be embargo, después de 17 afños, la et amélioré la qualité du personnel par weak, with virtually nonexistent in- eficacia de los servicios de extensión la formation. Cependant, la formation systems. Inadequate finan- no parece haberse mejorado conception institutionnelle, qui reflète cial management disrupts the timely apreciablemente. Se falta una visión les objectifs des projets, ne tenait pas flow of operational funds. The pro- estratégica para el desarrollo futuro suffisamment compte de la nécessité longed ineffectiveness of the exten- del sistema de extensión, y la gestión d'augmenter l'autonomie des paysans. sion services recently led Kenya, with continúa a ser débil, con una carencia De plus, des incitatifs inadaptés aux the help of the Bank and other donors, casi total de sistemas de información. besoins ont engendré un manque de to try to rationalize them with alter- La ineficiencia prolongada de los responsabilisation et de sensibilisation native approaches. servicios de extensión recientemente aux besoins des clients. Enfin, la The projects, having established motivó a Kenia, con la ayuda del structure hiérarchique de ces projets a national system based on the T&V Banco y de otros donantes, de tratar décourageait l'innovation, le Xii Executive Summary system of management, did de racionalizarlos utilizando partenariat et l'efficacité, et le increase geographical cover- enfoques alternativos. modèle d'extension à partir du age, improve the links Después de establecer un sommet était axé sur l'offre et between extension and re- sistema nacional basado en el non sur la participation. search (although belatedly), sistema de gerencia T&V, los and raise staff quality through proyectos si aumentaron el Viabilité training. But the institutional i alcance geográfico, fortalecieron Ce système n'est ni viable ni design, reflecting the projects' objec- los enlaces entre la extensión y la rentable. Il est beaucoup plus coûteux tives, lacked a focus on the critical investigación (aunque con retraso), y et pas plus efficace que le système qu'il issue of empowering farmers, with perfeccionaron al personal por medio de a remplacé. L'État continue de réduire inappropriate incentives leading to a la formación. Pero al diseñio les fonds qu'il alloue à l'extension, de lack of accountability or responsive- institucional, que reflejaba los objetivos même que ses autres dépenses, rendant ness to the clients' needs. Further- de los proyectos, le faltaba el enfoque en ainsi le système très dépendant des more, the projects' hierarchical struc- el asunto critico de apoderar a los bailleurs de fonds. ture yielded disincentives for campesinos; ademàs, los proyectos Les salaires du personnel innovation, partnerships, and effi- proporcionaban incentivos absorbent la majeure partie (80 %) du ciency, and the top-down extension inapropiados que condujeron a una budget de fonctionnement. Par system bas been supply-driven and falta de contabilidad o sensibilidad a las conséquent, nombre de problèmes qui nonparticipatory. necesidades de los clientes. Además, la limitaient l'efficacité du système estructura jeràrquica de los proyectos précédent subsistent, et le personnel en Sustainability no favoreció a las asociaciones, la est revenu aux anciennes méthodes de The system is neither financially inovaciôn, o la eficiencia, y el sistema de diffusion de l'information. sustainable nor cost-effective. It is extensiôn, que funciona de arriba en Les contacts fréquents avec un significantly more costly, and no abajo, ha sido impulsado por el nombre limité de paysans coûtent cher more efficient, than the system it aprovisionamiento y no ha sido et ne sont pas justifiés, car les replaced. The government's alloca- participatorio. messages à transmettre sont peu tions for extension, as for other nombreux, et le rythme d'introduction public expenditures, continue to Sostenibilidad des nouvelles technologies, lent. decline, leaving the system heavily El sistema no es ni financieramente dependent on donor funding. sostenible ni rentable. Es Pertinence An overwhelming proportion significativamente màs caro, pero sin Il ne faut pas confondre la pertinence (80 percent) of the operational bud- ser màs eficiente, que el sistema que des services d'extension avec celle de la get is consumed by staff salaries. As reemplazó. Las afijaciones del conception des projets. Par exemple, a result, many of the problems that gobierno para la extensión, como las l'offre de services d'extension aux limited the effectiveness of the pre- para otros gastos públicos, continúan petits exploitants du Kenya se justifie vious system have persisted, and a ir en declive, dejando el sistema toujours. Par contre, plusieurs aspects staff have reverted to the methods of excesivamente dependiente del de la conception des projets se sont dissemination used earlier. financiamiento por donantes. révélés inadéquats. Les paysans The projects' approach of high- Un porcentaje asombrante (unos apprécient l'accès à ces services au intensity contact with a limited 80%) del presupuesto operacional es point d'être disposés à payer pour, number of farmers bas been costly consumido por los sueldos del personal. mais la demande n'est pas satisfaite. and unwarranted, given the inad- Como resultado, muchos de los Les solutions de rechange aux services equate stock of messages for dis- problemas que limitaron la eficacia del publics d'extension sont rares, bien semination and the slow pace in sistema anterior han persistido, y el qu'une offre se dessine de ce côté. generating new technology. personal ha revertido a los métodos de La plupart des paysans, y compris diseminación que se empleaban antes. ceux qui assurent la liaison avec les Relevance La estrategia de contactos services d'extension, ne veulent pas A distinction is needed between the intensos y frecuentes con un número rencontrer les agents trop souvent, ce relevance of extension services and limitado de campesinos que qui porte à s'interroger sur la xiii Agricultural Extension: The Kenya Experience the relevance of project de- practicaban los proyectos ha pertinence de rencontres sign. While providing exten- sido cara e injustificada, dado el bimensuelles, voire mensuelles. sion services to smallholders inventario inadecuado de Même les agents des services still has a rationale in Kenya, mensajes para diseminar y el remettent en question le several features of the paso lento de la creación de caractère répétitif et l'inefficacité projects' design have proved nueva tecnologia. de leurs visites. inappropriate. The farmers Uutilisation pour presque value access to extension services Pertinencia toutes les zones de production d'une enough to be willing to pay for Hay que distinguir entre la relevancia approche générale et de messages them, but their demand is unmet. de los servicos de extensiôn y la passe-partout se révèle inefficace et The alternatives to government- pertinencia del diseñio de los improductive. Les quelques provided extension are few at proyectos. Aunque es todavia lógico expériences tirées d'initiatives pilotes present, although some providers ofrecer servicios de extensión a los du projet NEP Il confirment la are emerging. minifundistas en Kenia, se ha probado nécessité d'adapter les actions aux Most farmers, including the con- la impropriedad de algunas besoins. tact farmers, do not want to meet caracteristicas del diseio de los extension agents often, which ques- proyectos. Tanto valen los graneros el Efficacité tions the relevance of a biweekly-or acceso a los servicios de extensión que En 1982, les paysans ne pouvaient pas even monthly-visit schedule. Even están dispuestos a subvenirlos, pero accéder aux services d'extension the staff are wary of the repetitiveness no se satisface la demanda. comme ils l'auraient dû, et c'est encore and ineffectiveness of the visits. Actualmente, hay pocos alternativos a le cas aujourd'hui. Toutes les données Blanket coverage of the major- la extensión proporcionada por el accessibles démontrent l'inefficacité ity of the production areas with a gobierno, aunque algunos des dispositions institutionnelles single approach and standard mes- proveedores alternativos están actuelles pour fournir des services aux sages has proven inefficient and comenzando a aparecer. paysans. Cela reflète en partie le unproductive. Limited experience Visto que la mayoría de los contexte peu favorable dans lequel les from pilot initiatives under NEP Il agricultores, incluso los campesinos que services d'extension ont dû confirms the potential usefulness of se contactaron, no quieren reunirse con fonctionner ces dernières années. alternative and more responsive los agentes de extensién muy a menudo, Toutefois, leurs piètres résultats ne approaches. la relevancia de un programa de visitas découlent pas uniquement de facteurs que tienen lugar cada dos semanas, o externes, puisque tout indique que la Efficacy aun cada mes, es dudosa. Aún el situation a très peu changé depuis The farmers did not have adequate personal se da cuenta de la repetitividad 1990. access to extension advice in 1982, y la ineficiencia de estas visitas. Les évaluations qualitatives et and it appears that they do not have La tentativa de proveer un alcance quantitatives indiquent que le système it now. All methods of applying the general para la mayoria de las areas de d'extension ne fournit pas aux available data indicate that the producciôn utilizando un sôlo paysans le type d'information qu'ils current institutional arrangements enfoque y mensajes convencionales ha veulent. Elles montrent aussi que les for delivering services are ineffec- resultado ser ineficaz e improductiva. conseils prodigués à nombre d'entre tive. This is partially a reflection of La experiencia limitada de las eux sont discutables. Les services the poor enabling environment in iniciativas pilotas del NEP Il confirma d'extension continuent de privilégier la which the extension services have la utilidad potencial de enfoques diffusion de messages agronomiques operated in recent years. The poor alternativos más sensibles. simples portant notamment sur le performance, however, is not maïs. Leurs activités dans le cadre des entirely a function of these external Eficacia projets NEP I et NEP II n'encouragent factors, as evidence suggests that Los campesinos no tuvieron acceso guère la prise de conscience des the current situation differs little adecuado a los avisos de la extensiôn en paysans et l'adoption par ceux-ci des from that of 1990. 1982, y parece que todavia no bo tienen. recommandations. L'incapacité à leur Both qualitative and quantita- Todos los métodos de aplicar los datos offrir des conseils plus élaborés et xiv Executive Summary tive assessments indicate disponibles indican que los adaptés à leur environnement that the extension system has arreglos institucionales corrientes réduit la rentabilité des projets, not responded with the kinds para suministrar estos servicios notamment des rencontres of information the farmers son inefectivos. Esto es en parte individuelles avec les paysans. want, and the relevance of un reflejo de las pocas C'est particulièrement vrai pour the advice it offers to a posibilidades de apoderarse que le projet NEP IL Quand ce broad range of farmers is existen en el ambiente en que los programme a été lancé, on questionable. The focus of the ex- servicios de extensiôn han savait que la plupart des paysans tension services has remained on operado en afios recientes. Sin embargo, avaient déjà assimilé les messages les disseminating simple agronomic este pobre funcionamiento no se debe plus simples concernant le maïs. and maize-related messages. Ex- totalmente a estos factores externos, Constatation significative, la très tension activities under NEP I and porque la evidencia sugiere que la grande majorité des paysans qui ont NEP Il had little influence on the situación presente no es muy diferente reçu des messages, même sur les evolution of patterns of awareness de la de 1990. méthodes plus complexes, les ont and the farmers' adoption of rec- Ambas las evaluaciones cualitativas adoptées. Bien que des facteurs ommendations. The failure to de- y cuantitativas indican que el sistema de étrangers aux services d'extension, liver more advanced and context- extensiôn no ha respondido con el tipo comme le manque de ressources specific advice reduced the de información que quieren los financières ou l'incapacité d'obtenir cost-effectiveness of the main fea- campesinos, y la pertinencia de los des crédits, fréquemment mentionnés ture of the projects' design-that is, avisos que se les ofrecen a un rango par les paysans, puissent les empêcher face-to-face extension. This was amplio de agricultores es dudosa. Los d'adopter certaines méthodes particularly true for NEP Il. When servicios de extensión continùan a complexes (p. ex. l'emploi d'engrais et this project began, most farmers enfocarse en diseminar mensajes simples de pesticides), il va de soi que le were known to have already y agronómicos que muchas veces tratan manque d'information reste un adopted the simpler messages del maiz. Las actividades de extensión obstacle important. about maize. bajo los NEP I y Il tuvieron poca Malgré les progrès réalisés en ce A significant finding is that a influencia en la evoluciôn de los modelos qui concerne l'accès aux services, ce very large proportion of those who de conocimiento y la adoptaciôn por los dernier demeure limité, surtout pour are aware of the messages on even the campesinos de sus recomendaciones. Su les paysans pauvres et peu éduqués. more complex practices have adopted inhabilidad de ofrecer avisos màs Les paysans affirment que them. Thus, although factors not avanzados y especificos dentro del l'information est plus rare et que les related to extension, including a fre- contexto redujo la rentabilidad de la services sont de moindre qualité quently cited lack of financial caracteristica principal del disefño del depuis le début des années 80. resources or access to credit, may be proyecto¾es decir, la extensión cara a Les progrès pour l'égalité des important in preventing farmers from cara¾sobre todo en el caso del NEP Il. sexes sont mitigés. Les préjugés dont adopting certain complex practices Cuando se inició este proyecto, se sabia étaient victimes les paysannes autrefois (for example, using fertilizers and que la mayoria de los campesinos ya ont disparu, mais certains préjugés pesticides), it is evident that the lack habia adoptado los mensajes más demeurent quand vient le moment de of information continues to be an sencillos sobre el maíz. choisir des contacts dans la important constraint. Un hallazgo importante es que una population paysanne. La proportion Access to existing services is proporción muy grande de los de femmes parmi les agents limited, especially for the poor and campesinos que se han dado cuenta de d'extension sur le terrain est the less educated, despite the overall los mensajes, o aun de las pràcticas màs sensiblement la même depuis 1982. increase in coverage under NEP I complejas, los ha adoptado. Asi, aunque La pénétration des projets reste and Il. Most farmers also report that factores que no tienen nada que ver con bien inférieure aux niveaux prévus au the availability of information and la extensión, tales como la moment de leur conception, car the quality of services have declined frecuentemente citada falta de recursos seulement 7 % des paysans qui y sont since the early 1980s. financieros o acceso al crédito, pueden associés et 2 % de l'ensemble des Progress toward gender equity ser importantes en impedir a los paysans rencontrent régulièrement les xv Agricultural Extension: The Kenya Experience has been mixed. The earlier campesinos de adoptar ciertas agents d'extension au cours des bias against women farmers prácticas complejas (por ejemplo, sessions prévues dans le cadre bas been rectified, but some el uso de los fertilizantes y las des projets. bias persists in the selection of pesticidas), es evidente que la falta contact farmers. The propor- de información continúa a ser Rentabilité tion of women field-extension una limitaciôn importante. Dans l'ensemble, on ne peut agents has remained largely El acceso a los servicios établir un taux de rendement unchanged since 1982. existentes es limitado, sobre positif pour les dépenses courantes Outreach is well below the levels todo para los pobres y los des projets d'extension. Il se peut que anticipated in the project design- ineducados, aunque su alcance ha le projet NEP I ait récolté quelques only about 7 percent of the contact crecido. Los agricultores informan que profits au début, mais cela n'a pas farmers and 2 percent of all farmers la disponibilidad de informaciôn y la duré. Les données ne révèlent aucun meet regularly with extension agents calidad de los servicios son peores que impact significatif, même depuis 1990. in settings envisioned in the design. a principios de los afios ochenta. L'efficacité technique des paysans El progreso hacia la igualdad de s'est quelque peu améliorée depuis Efficiency los géneros no ha sido totalmente 1982, mais le rendement d'ensemble Overall, a positive rate of return on positivo. El perjuicio anterior contra reste faible. Selon les données, les expenditures for extension cannot be las campesinas se ha remendado, pero services d'extension ont eu un très established. It is likely that NEP I persiste este perjuicio en la selecciôn de léger effet positif sur l'efficacité yielded some early benefits, but they los agricultores que se contactan. La technique, mais le degré de confiance appear to have been short-lived, as proporción de agentes de extensiôn en est faible. D'après ces mêmes données, the data do not indicate any signifi- el campo que son mujeres ha les services d'extension n'ont eu cant impact, even by 1990. cambiado poco desde 1982. aucune influence visible sur l'efficacité Farmers' technical efficiency El alcance es bastante debajo de économique. Le faible rendement has improved somewhat since los niveles anticipados en el disefño del économique des paysans laisse 1982, but overall efficiency con- proyecto: sôlo unos 7% de los supposer qu'il suffirait d'adopter une tinues to be low. According to the campesinos contactados y unos 2% de meilleure combinaison d'intrants pour data, extension services have had todos los agricultores se reunen réaliser des économies considérables a very small positive impact on regularmente con los agentes de dans les conditions actuelles du the level of technical efficiency, extensión bajo las condiciones que se marché. but the level of confidence in the conceptualizaron en el disefño. Les données ne permettent pas de result is low. The data also indi- dire si l'offre des services d'extension a cate that extension services have Eficiencia eu une influence significative sur la had no discernible impact on the En conjunto, no se puede establecer productivité des fermes. Il est level of economic efficiency. una tasa positiva de rendimiento en cependant possible de souligner Farmers' low economic efficiency cuanto a los gastos para la extensión. l'inefficacité de l'allocation des suggests the potential for realizing Es probable que el NEP I rindiô ressources de ces services et le mauvais significant savings by simply algunos beneficios tempranos, pero choix de leurs objectifs. La production moving to a more economical mix parece que no duraron mucho, visto agricole affiche une croissance plus of inputs under current market que los datos no indican ningùn efecto forte dans les zones jusque là moins conditions. significativo aún en 1990. productives, tandis que la mise en The current data do not indi- La eficiencia técnica de los place d'agents d'extension favorise les cate a significant impact of the campesinos se ha mejorado algo desde zones plus productives. Les services supply of extension services on pro- 1982, pero su eficiencia en general d'extension ont peut-être aidé à ductivity at the farm level. What the continúa a ser baja. Según los datos, los transmettre des messages data do indicate is that extension servicios de extensiôn han tenido un technologiques simples dans les zones resources have been allocated ineffi- efecto positivo muy pequefño en el nivel auparavant mal desservies et moins ciently and poorly targeted. Growth de eficiencia técnica, pero hay un bajo productives, mais les données actuelles in agricultural production bas been nivel de confianza en los resultados. ne permettent pas de l'affirmer. xvi Executive Summary higher in the previously less- Estos datos también indican que Une grande proportion de productive areas, while the los servicios de extensión no han paysans est prête à payer pour placement of extension staff tenido ningún efecto perceptible les services d'extension, has favored the more produc- en el nivel de la eficiencia montrant ainsi leur intérêt pour tive areas. Extension might económica. La baja eficiencia les conseils qui leur sont have helped spread simple econàmica de los campesinos prodigués. Mais la valeur qu'ils technological messages to the sugiere la posibilidad de realizar y attachent, comme l'indique les formerly underserved and less-pro- ahorros significantes sólo por montants qu'ils sont prêts à payer, ductive areas, but this cannot be medio de introducir un conjunto de reste cependant bien inférieure à que conclusively determined with the ingresos que sería màs económico bajo l'État investit par ferme dans les current data. las condiciones del mercado presente. services d'extension. A significant proportion of Los datos corrientes no indican farmers is willing to pay for exten- ningún efecto importante de la oferta de Recommandations principales sion services, which indicates that la extensión en la productividad al nivel Les leçons que l'on peut tirer de la they value the advice they receive. de la granja. Lo que si indican es que los présente étude, et dont on devrait tenir But the perceived value, as revealed recursos de la extensión han sido compte dans la conception des futurs by what the farmers are willing to asignados de manera ineficiente a projets d'extension agricole, pay, is well below what the govern- objetivos malamente escogidos. El débouchent sur les recommandations ment currently spends on extension incremento en la producción agricola ha suivantes. services per farm. sido más grande en las areas que antes eran menos productivas, mientras que la Choix des objectifs Principal Recommendations colocación del personal de la extensión La première leçon porte sur la The lessons emerging from this ha favorecido las areas más productivas. nécessité de mieux cibler les zones et study that can be applied to the Es posible que la extensión haya les groupes où l'impact des services design of future agricultural exten- ayudado a diseminar simples mensajes d'extension devrait être plus fort. sion projects suggest the following técnicos a las areas menos productivas D'où le besoin d'un système plus recommendations. que antes no se servian bien, pero esto souple et plus « ingénieux », capable no se puede determinar conclusivamente de déterminer le fossé qui sépare les Targeting a base de los datos disponibles. meilleures pratiques des pratiques The first lesson is the need for more Una proporción significativa de los courantes et d'allouer de façon efficient targeting of extension campesinos está dispuesta a pagar pot rationnelle les ressources limitées. En services to areas and groups where los servicios de extensión, bo cual indica outre, les services d'extension the marginal impact is likely to be que valen los avisos que reciben. Pero el devraient sélectionner des paysans the greatest. This calls for a more valor percibido, cuando se representa plus représentatifs du milieu flexible, "smart" system that can por la suma que los campesinos están socioéconomique local, de manière à identify the gaps between best prac- dispuestos a pagar, es bastante menos pouvoir fournir des conseils mieux tices and average practices and que bo que el gobierno desembolsa por adaptés aux catégories de paysans. allocate scarce resources more ra- granja para los servicios de extensión. tionally. Further, the farmers who Systèmes d'information are selected for interaction should Recomendaciones principales Afin de mieux cibler leur action, les be more representative of the local Las lecciones que se han aprendido de services d'extension doivent fournir socioeconomic environment, which este estudio y que se pueden aplicar al au bon moment une information will ensure the delivery of more disefño de proyectos futuros de fiable, ce qui souligne l'importance du relevant advice to the various cat- extensiôn agricola sugieren las monitorage et de l'évaluation. Une des egories of farmers. recomendaciones siguientes. grandes leçons à tirer de l'expérience kenyane est qu'il faut cerner les Information Systems Selección demandes des paysans et leur offrir Targeting calls for appropriate flows La primera lección es la necesidad de des services adaptés au contexte of timely and reliable information, seleccionar màs eficientemente las areas technologique et économique de leur xvii Agricultural Extension: The Kenya Experience which brings the focus to moni- y los grupos para los cuales el région. Le monitorage et toring and evaluation (M&E). impacto marginal de los servicios l'évaluation sont aussi An important lesson of the de extensiôn va a ser el más indispensables pour déterminer Kenyan experience is that farm- importante. Para esto, se necesita les lacunes et orienter le système ers' demands should be identi- un sistema mis flexible e « ingénieux » vers un meilleur fied and that services should be "inteligente" que pueda choix de ses cibles, en fonction tailored to suit local technologi- identificar las diferencias entre las des besoins. cal and economic conditions and cir- mejores prácticas y las prâcticas cumstances. M&E is also critical for medias y asignar los recursos escasos de Intensité identifying the gaps and guiding the una manera más racional. Además, los En se basant sur leur expérience, les "smart" system, as needed, toward campesinos que se escogen para la paysans ont clairement exprimé qu'ils more efficient targeting. interacción deben ser más ne veulent pas rencontrer trop representativos del ambiente souvent les agents des services Intensity socioeconômico local, bo cual asegurará d'extension, notamment parce qu'il Based on their experience, farmers el suministro de avisos mis relevantes a n'y a pas assez de nouvelles have clearly indicated that they do las varias categorías de agricultores. technologies pour justifier des visites not want to see extension agents too rapprochées. Il serait plus rentable often, and there are not enough new Sistemas de información d'instaurer un système qui allégerait la technological recommendations to Para la selección se necesita un flujo de présence et qui couvrirait des zones sustain a highly intense visit sched- datos oportunos y fiables, bo que se plus vastes. On peut supposer que ule. It would be more cost-effective hace pensar en el monitoreo y la l'amélioration de la qualité des services to establish a leaner and less-inten- evaluaciôn (M&E). Una lección augmenterait la demande. Cela sive presence with wider coverage. importante de la experiencia de Kenia permettrait d'accroître l'importance It is conceivable that improving es que las demandas de los campesinos des services et de mettre en place un service quality could increase tienen que identificarse y que los système dynamique et axé sur les demand. This would amplify the servicios se deben hacer a la medida de besoins de la population. importance of a responsive and las condiciones y circunstancias locales dynamic delivery system. tecnolôgicas y econômicas. El M&E es Pluralisme critico también para identificar las Des services d'extension qui Pluralism faltas y guiar el sistema "inteligente," transmettent des messages standards Delivering standard messages by cuando sea necesario, hacia una inspirés d'une méthodologie unique means of a single or uniform method- selección mis eficiente. ou uniforme risquent d'être moins ology is likely to limit the effective- efficaces. Des paysans plus jeunes et ness and efficiency of extension ser- Intensidad plus instruits remplacent vices. Younger, more educated A base de su experiencia, los progressivement leurs parents. Les farmers are taking over from their campesinos han claramente indicado émissions radio qui diffusent de parents. Radio programs that dis- que no quieren ver a los agentes de l'information nouvelle sont seminate new information are popu- extensiôn con demasiada frecuencia, y populaires, et d'autres fournisseurs de lar, and alternative providers are no hay bastantes nuevas services font peu à peu leur apparition beginning to emerge in rural Kenya. recomendaciones técnicas para dans le Kenya rural. Il pourrait Exploiting low-cost modern commu- sostener un programa muy intensivo s'avérer rentable d'exploiter les nications, demonstrations, print me- de visitas. Seria mis rentable establecer communications modernes à bas prix, dia, and partnerships with civil soci- una presencia menos fuerte e intensa d'organiser des démonstrations, de ety and the private sector could con un alcance mis amplio. Es posible recourir aux journaux et de créer des be cost-effective. This would leverage que un adelanto en la calidad de los partenariats avec la société civile et le resources to increase outreach, and servicios pueda incrementar la secteur privé. Ainsi, on multiplierait les would be likely to have a greater demanda. Esto agrandaria la ressources et étendrait la portée des impact with the same or reduced importancia de un sistema de entrega activités, ce qui accroîtrait demand on government resources. mas sensitivo y dinámico. probablement l'impact des ressources xviii Executive Summary Client Focus Pluralismo de l'État, qui pourraient peut- The central focus of the insti- La entrega de mensajes être méme être réduites. tutional design should be to convencionales por medio de empower farmers. An effec- un método único o uniforme Priorité aux clients tive way to incorporate client probablemente va a limitar la Les services d'extension focus is to integrate alterna- eficiencia y la eficacia de los devraient viser l'autonomie des tive means of giving farmers servicios de extensión. Los paysans. Une bonne façon d'y a voice-such as cost-sharing, fos- campesinos más jóvenes y parvenir consiste á intégrer aux tering farmers' organizations, and educados se están encargando de las mécanismes d'offre des services des decentralization-into the delivery tareas de sus padres. Programas de la moyens permettant aux paysans de se mechanisms. Cost recovery (even if radio que diseminan nueva faire entendre - partage des coûts, only partial), in particular, offers información son populares, y los incitation á l'établissement several advantages. It provides ap- proveedores alternativos empiezan a d'organismes de paysans et propriate incentives, addressing the aparecer en la Kenia rural. La décentralisation. Le recouvrement des issues of accountability and quality explotación de comunicaciones coúts (même partiel) présente control; it makes service more modernas de bajo costo, plusieurs avantages. Il sert d'incitatif et demand-driven and responsive; it demostraciones, la prensa, y introduit la responsabilisation et le provides some budgetary respite; asociaciones con la sociedad civil y el contróle de la qualité. Il axe davantage and it encourages alternative pro- sector privado puede ser rentable. Así les services sur la demande, procure viders. Such institutional arrange- se aprovecharía de los recursos un certain répit budgétaire et ments remain unexplored in Kenya. disponibles para el alcance para tener encourage d'autres fournisseurs de un máximo efecto con la misma o aún services. De tels arrangements menos demanda para los recursos institutionnels demeurent inexplorés gubernamentales. au Kenya. Enfoque en el cliente El enfoque central del diseño institucional debe ser de apoderar a los campesinos. Una manera efectiva de incorporar un enfoque en el cliente es la integración dentro de los mecanismos de entrega de métodos alternativos para dar una voz a los agricultores, métodos tales como la repartición de costos, el apoyo a los organismos que representan a los campesinos, y la decentralización. En particular, la recuperación de los costos (aunque sea solamente en parte) ofrece varias ventajas. Provee incentivos apropiados, dirigiéndose a los temas de contabilidad y control de la calidad; resulta en un servicio más sensitivo y empujado por la demanda; ofrece una tregua presupuestaria; y estimula a los proveedores alternativos. Arreglos institucionales de este tipo se quedan a explorar en Kenia. xix  ABBREVIATIONS AND ACRONYMS AIC - Agriculture Information Center ASIP - Agricultural Sector Investment Program ATD - Africa Technical Department CVM - Contingent valuation method DEA - Data envelopment analysis FEW - Frontline extension worker FTC - Farmer Training Center GDP - Gross domestic product IDA - International Development Association IFAD - International Fund for Agricultural Development KARI - Kenya Agricultural Research Institute KP - Kenya pounds (currency) Ksh - Kenya shillings (currency) MALDM - Ministry of Agriculture, Livestock Development, and Marketing M&E - Monitoring and evaluation NASSEP - National Sample Survey and Evaluation Program NEP - National Extension Project NGO - Nongovernmental organization OED - Operations Evaluation Department PPA - Participatory poverty assessment RHBS - Rural Household Budget Survey SAR - Staff appraisal report SMS - Subject matter specialist T&V - Training and visit WTP - Willingness to pay xx  Introduction he focus on agricultural extension in Kenya is rooted in agriculture's strategic importance in development. With 70 percent of the population living in rural areas, almost half in absolute poverty, the centrality of rural development to any strategy for poverty alleviation is clear. Critical to rural development is the performance of agriculture, both because of its direct contribution to the national economy (28 percent of gross domestic product; 60 percent of export earnings; and 80 percent of national employment) and because of its indirect multiplier effects. Within agriculture, 81 percent egy for Africa, and the large investment made, there of the farmers are smallholders (owning less than 2 have been few attempts to rigorously measure the hectares), which makes it imperative to improve small- impact of T&V extension. holder productivity. In the Kenyan context, the debate has been elevated Accordingly, the government of Kenya has long by the estimate of very had agricultural extension on its development agenda. high returns to T&V The performance of With World Bank support, Kenya adopted a T&V extension by a Bank Af- system of management for its extension services in rica Technical Depart- 1982 as part of its growth strategy. The Bank then ment (ATD) study system is controversial, financed the extension system with NEP I (1983-91) (Bindlish and Evenson mirroring the broader and NEP 11 (1991-98). 1993, 1997), on the one disagreement about the The performance of Kenya's extension system is hand, and the lack of cost-effectiveness of controversial, and this debate is part of a broader visible results on the disagreement about the cost-effectiveness of the T&V ground, on the other.' approach. Debate has focused largely on the efficacy of Agriculture's poor per- the mechanisms for delivering extension advice formance in Kenya in recent years, the country's (Picciotto and Anderson 1997). It is generally agreed that declining budgetary resources, and efforts to rationalize the T&V system is costly. The controversy centers on the the structure of the Ministry of Agriculture, Livestock returns to the high levels of investment by borrower Development, and Marketing (MALDM) have increas- countries in the T&V system, and hence on its impact on ingly called into question the effectiveness of extension agricultural production (Purcell and Anderson 1997). services.2 (The general nature of the debate is discussed in Annex Notwithstanding the ATD's estimates of high A.) Despite the debate's intensity, the importance of marginal returns to extension, an Operations Evalua- agricultural extension in the Bank's development strat- tion Department (OED) review of NEP I raised I Agricultural Extension: The Kenya Experience questions about the effectiveness of the T&V approach objective, empirical approach.5 Most of the conclusions in Kenya. The review concluded that NEP I had some are based on the results of a 1997 OED household survey beneficial aspects but several operational deficiencies, and a survey of the extension staff, supplemented by and, most important, that it was not financially secondary data and information from several recent sustainable. With little evidence to suggest any sig- studies by MALDM. OED's household survey (see Annex nificant impact on agricultural growth to justify I) covered the same population surveyed for the 1990 ATD T&V's high fiscal costs, OED questioned both the study, which used a subsample of the 1982 Rural appropriateness of the extension format developed as Household Budget Survey (RHBS) data. In 1997, inter- a uniform approach throughout Kenya's cropping viewers revisited as many of the respondents as could be areas and the overall outcome of the project. The contacted in the clusters that had been sampled by the ensuing discussions with the Africa Region led to ATD study. This evaluation thus has the advantage of OED's final rating of the project outcome as margin- baseline data, even if they are somewhat limited. To ally satisfactory. The Africa Region contested the dissociate the OED survey from the government extension rating, and the disagreement has persisted. This study service and the Bank, it was implemented by the Tegemeo was undertaken to inform the discussion. Institute of Egerton University in Kenya. The discontent with the current system of extension is not limited to the Bank. Criticism is sharpest among Study Objective the donor and nongovernmental organization (NGO) The goal of this study was to make an empirical communities, which seek assessment of the impact of the NEP I and NEP 11 T reform toward a demand- projects in Kenya. Following a theory-based evaluation driven system that better approach, it combined qualitative and quantitative the current system of accommodates alternative methods to arrive at a credible body of evidence on the extension is not ways of reaching farmers. projects' likely impact. In so doing, it provides an limited to the Bank. The Kenyan government independent review of earlier findings. Criticism is sharpest is also wary of continuing To allow for appropriate policy conclusions, the with the large allocation study distinguishes clearly between the impact of exten- among the donor and of resources to extension, sion and the impact of a particular system of extension. nongovernmental given agriculture's poor This has implications for interpreting the results to organization performance since the illuminate the three key aspects of the study: the rel- communities. start of the 1980s. evance, efficacy, and efficiency of the Kenyan extension It is not clear, however, system. In Kenya, impact evaluation is complicated by whether the poor record of the introduction of the T&V system on a national scale, agriculture necessarily reflects the performance of exten- which precluded a with-and-without comparison. The sion; it is possible that agriculture could have done worse system was also introduced rapidly, over the course of were it not for extension. Establishing the impact of only 3 years, so that now, after some 15 years, the data extension from the sector's aggregate performance is allow only limited before-and-after comparisons. difficult because, much like casual observations from field visits, it lacks an appropriate counterfactual. The prob- The National Extension Projects lem can be demonstrated with maize yields.' Between Before NEP I and NEP 11, the traditional system of 1970 and 1989, the annual growth rate of maize yields in extension in Kenya suffered from several weaknesses.' Kenya was 4.7 percent. This rate rises to 5.4 percent when To overcome them, the T&V system of extension was rainfall is factored in.4 Comparing the periods before and introduced with the intent of providing "competent, after 1982-that is, before and after T&V-growth well-informed village-level extension workers who will slowed from 6.3 percent to 5 percent, a statistically visit farmers frequently and regularly with relevant significant difference. The difference, however, ceases to technical messages and bring farmers' problems to be significant when rainfall is factored in. research" (Benor and Baxter 1984). (The design fea- Given the strong and divergent opinions on the tures of the T&V system implemented by NEP I and perceived performance of NEP I and NEP and the lack NEP 11 are described in Annex B.) T&V was introduced of evidence to determine their impact, this study takes an as a brief pilot project in two districts in 1982. Starting 2 Introduction in 1983 with NEP I, it was rapidly expanded to cover to increase their productivity and incomes. In addition about 90 percent of Kenya's arable land. to continuing to support the work of NEP 1, NEP I1 The objective of NEP I was to achieve sustained introduced T&V to 6 new areas, and when 4 of the increases in agricultural production in 30 of Kenya's original districts were split, T&V extension was 41 districts, covering all medium- and high-potential applied in 40 of Kenya's 45 districts. The project was arable areas. The approach entailed reorganizing designed to consolidate and fortify the gains made and strengthening extension services through the under NEP I, to increase direct contact with farmers, to adoption of T&V over a period of three years and improve the relevance of extension information and improving the link between research and extension. technologies, to upgrade the skills of staff and farmers, NEP I was considered an institution-building project, and to introduce pilot innovations into the extension and the Bank expected to provide external assistance system. for 10 to 15 years to ensure that the necessary institutional reforms and improvements in staff skills Study Strategy and Outline were made. The measurement of the impact of the two projects In 1991, NEP II succeeded NEP I. The objective of focused primarily on their key objectives-institutional NEP II was to stimulate the development and adoption development and sustained increases in agricultural of technical packages that would enable smallholders productivity. While important economic, social, and FIGURE 1.1. STYLIZED IMPACT MODEL Food security Land quality Friends, neighbors, w undeinnovative farmers Credit Risk Adaptive thPublic extension service Spillovers research introduceilotinnovtionintoWeather and pests Farmer organizations, NGOs Private sector: input suppliers, __Household Output HH processors, consultants obectives welfare Media: audio, video, print Basic Indigenous Field days research systems Friends, eighbors iemnovativnfel fmrasCrdtRs Adaptiverucur Pubicextnsonieric Technology Generation Knowledge Delivery Farm Decisionmaking Impact Inputs Acaities Output Outcomes Results Research- Institutional extension links development Accesssupphers,sHouseh Recommendations Sustainabilisy Contact Aarene s Prdciey Training Efficacy Distribution Feedback Plurality Note: Broken line: feedback from farmers to decisionmakers. Solid line: delivery of time-bound "messages" by extension workers from researchers to farmers. RH: household. 3 Agricultural Extension: The Kenya Experience environmental impacts are implicit in the projects' Institutional Development rationale, they were not explicitly stated as major The focus of institutional impact is on the left half of objectives. But the key social issues of gender and the the model in figure 1.1, specifically on the institutional distribution of benefits are naturally addressed because arrangements, or input, used by NEP I and II to deliver of the large proportion of Kenyan farmers who are their output. The analysis is presented in Chapter 2, women or smallholders. and combines secondary data with the findings of The study's theory-based approach used the styl- several recent reviews of extension in Kenya and results ized causal flow model depicted in figure 1.1. The from the OED staff survey. model reflects the reliance of NEP I and NEP II on extension agents successfully delivering time-bound Beneficiary Assessment "messages" (indicated by solid arrows) from research- An important element of the strategy was to obtain ers to farmers.7 The task participatory beneficiary assessments of extension ser- As implemented, the of delivering feedback vice. Although the beneficiary assessments for this program was from the farmers to the evaluation were not conducted, findings from two decidedly researchers (indicated by recent participatory assessments in Kenya are perti- decdely broken arrows) was also nent, and these, along with the farmers assessments nonparticipatory- to be carried out by the from the OED household survey, are discussed in there was no pretense extension workers. As Chapter 3. of involving farmers in implemented, the pro- the development of gram was decidedly Efficacy: The Quantity and Quality of Contact nonparticipatory-there The debate on the effectiveness of T&V in Kenya has technology or the was no pretense of centered primarily on the "real" side of the equation- messages, nor was involving farmers in the that is, T&Vs impact on agricultural productivity at there an opportunity development of technol- the farm level. This impact is evaluated following the for the farmers to ogy or the messages, nor stylized flow expressed in the model in figure 1.1. The selectwas there an opportunity first part is an analysis of the outreach and the quality selet tpic of for the farmers to select of interaction between extension agents and farmers, interest to them. topics of interest tothem which is discussed in Chapter 4. (MALDM 1997b). The impact of the projects can be Outcomes assessed at several points along the continuum, from Next on the continuum leading to ultimate results is the the knowledge complex to the final change in welfare. measurement of the proximate outcomes of extension Following the results-based management framework, efforts, which indicates potential for impact. This this evaluation sought to relate the results observed in entailed measures of farmer awareness and adoption of the farmers' fields back to project inputs. In addition, extension outputs (that is, its technological recommen- intermediate output and outcome indicators were dations, or "messages"). These measures, and how they measured to assess the performance of the extension relate to the supply of extension services, are discussed system along the results chain to confirm the potential in Chapter 5. for impact. Accordingly, the five boxes along the bottom of the model present the key indicative Results measures corresponding to inputs, activities, outputs, Determining the physical impact, or results on the outcomes, and results. The evaluation strategy was to ground, required relating the supply of extension use a mix of qualitative and quantitative methods to services to changes in productivity and efficiency at measure and analyze these key indicators. No attempt the farm level. To do this, the evaluation followed was made to study the projects' impact on household two lines of inquiry. One was a nonparametric welfare, because it was likely that it had been affected measurement of farmers' efficiency in production and by a number of factors beyond the scope of how farm productivity changed over time. (This is extension's activities. The outline of this volume discussed in Chapter 6.) The other was an econometric reflects the study approach. measurement of the impact of extension on farm 4 Introdu c tion production (see Chapter 7). These estimates reveal the returns to investment in extension and are needed to establish its efficiency. Willingness to Pay This analysis revisits the farmers with a new method- ological tool-the contingent valuation method, which is used to directly elicit their willingness to pay for extension services (see Chapter 8). The findings from this exercise have implications for the relevance of the design of past projects and provide insights for future project design. 5 s 1 Institutional Development Akey objective of both NEP I and NEP II was the development of organizational and institutional arrangements for the efficient and effective delivery of agricultural services to Kenyan smallholders. In 1982, it was expected that the extension system would need 15 years to mature. The period has now passed, and it is natural to ask how the extension system is performing in comparison with the system it replaced. The discussion in this chapter is based on reviews undertaken or commissioned by MALDM over the past two to three years in an attempt to develop a national policy and a frame- By 1982, the year before NEP I was introduced, work to rationalize the current extension service.' several disparate and uncoordinated donor-supported These reviews are supplemented with conclusions from efforts were in the field. The various extension arrange- discussions with extension and ministry staff, focus ments lacked a consistent national strategy and were groups, individual district-level staff, and former exten- essentially ad hoc project components. This prolifera- sion staff familiar with the early days of NEP I and tion of programs was viewed as expensive, inefficient, with findings from the 1997 staff survey conducted by and largely ineffective. OED for this study. The extension services did, however, have a It is instructive to review the before-and-after well-defined line of command, from the director of pictures of Kenya's extension system. agriculture down to the field-level, frontline exten- sion worker (FEW). The staff numbers were seen as Before adequate, but the services were thought to be per- Agricultural extension in Kenya dates back to the early forming well below their potential (World Bank staff 1900s. Several approaches were tried, including indi- appraisal reports).2 vidual visits, group methods, unified extension, farm There were several management, integrated development, and specialized reasons for this: the commodity extension programs. Except for the last, junior staff were not system was highly none endured. Nevertheless, the traditional extension well trained; there successful in the system was highly successful in the dissemination of was a lack of field dissemination of hybrid hybrid maize technology. Starting in 1965 with a emphasis; and FEWs nationwide program of demonstrations and field days, visited few farms, hybrid maize was promoted so successfully that by and most of their vis- 1977, 50 percent of even smallholders were growing its were to progressive farmers, who represented hybrid maize (Johnson and others 1980). about 10 percent of all smallholders. Other probliems 7 Agricultural Extension: The Kenya Experience included insufficient operational funds, which lim- TABLE 2.1. NEP I AND II: BORROWER PERSPECTIVES ited mobility (except for projects and programs Weak points Strong points funded by donors), resulting in lax supervision and Very broad objectives %Ide coverage support of the field staff, and generally inadequate trong top-down Co%rage of all types of allowances for the FEWs, which left them to travel planning f.$rmers on foot and unable to cover a large area. 1Trget not specific Strmg staff training. The extension system in place before the pa rticipation ~ agroecological endowments, district-specific measures ewere also calculated for 1997. As expected, the average t n level of technical efficiency increased (with an average trefficiency level of 69 percent), but the economic 1efficiency was still very low (30 percent). The low level of cost-efficiency implies that the allocative efficiency oeahouseholds is quite low, which further indicates that rjcthe farmers are not using economically optimal levels Th oof inputs. A statistical analysis using the 1997 district-specific efficiency measures fails to reveal a statistically signifi cant relationship between any of the efficiency measures (cost, technical, or allocative) and the supply of exten- sion services. In these tests, the cluster average technical ~ ~ 4 efficiency for 1982 is used to control for the regional effects for each location, and extension supplyis Field consultations measured as a weighted average of lagged extension staff-farm ratios. Qualitatively, extension has a small All observations are ranked against the most efficient positive coefficient in the cost and technical efficiency farmer in the sample. relationships, but a negative coefficient for allocative efficiency. The 1982 efficiency level is positive and Relative Efficiency significant for cost and technical measures, and positive The average farm was operating at a very low level of but weakly significant for allocative efficiency. technical efficiency in 1982 (31 percent), and although Although there is no clear-cut rationale for includ- there was some improvement, relative efficiency was ing district-specific effects, since efficiency measures still quite low in 1997 (45 percent). The scale of are calculated by district, the consequence is that the farmers' operations has improved, but a majority still effect of the supply of extension on technical efficiency operate at a suboptimal scale; statistical tests fail to is still low and positive (0.056), but is now significant reject the hypothesis of nondecreasing returns to scale, at the 10 percent level. Overall, the results do not A comparison of cost change much, but they do reinforce hints of mild The low level of cost- or allocative efficien- extension effects on technical efficiency. Cost and efficiency implies that cies between 1982 and allocative efficiency results do not change. 1997 is precluded by a To test for the effectiveness of alternative extension the allocative efficiency lack of price data for methodologies, indicators for the normal place and of households is quite 1982. For 1997, the frequency of meetings were tested in a separate set of low, which indicates that results show that the regressions. The frequency of meetings has no influence the farmers are not using average level of eco- on efficiency. Those who interact with extension agents fnomic (cost) efficiency at cooperative society meetings have a large but lwas very low (15 per- weakly significant effect on both cost and technical levels of inputs, cent), much lower efficiencieS.2 it might also be expected that contact than technical effi- farmers would be more efficient, but the current data ciency. Thus, even with the current level of technology, do not show this for any measure of efficiency. it appears that a simple change in the input mix, to one Among other variables, farm size has a strong that is more economical given the current market negative effect on cost and technical efficiency; that is, conditions, would offer farmers the potential for signifi- smaller farmers are more efficient. Distance to markets cant savings, has a significant negarive effect on allocaive effi- 28 Results 1: Farmer Efficiency and Productivity Change FIGURE 6.1. PRODUCTIVITY CHANGE, 1982-97: TRENDS IN CLUSTER AVERAGES Index Staff/farm ratio (x 1,000) 6 3 5 2.5 4 2 3 1.5 1 0.5 0 m 1982 Staff/farm ratio 1997 Staff/farm ratio noon noun Efficiency change Technical change - a no- a Malmquist index Note: Clusters sorted by average 1982 efficiency scores. Source: Results of DEA analysis using OED survey data. ciency, no effect on overall cost efficiency, and a tivity increased by 28 percent over the period 1982- positive effect on technical efficiency. This suggests 97. This increase stemmed largely from positive techni- that farmers farther from markets may be specializing cal change, which raised productivity by about 56 in specific crops, and hence may be more efficient; percent. The technical improvement, however was those close to markets may be diversifying, and while moderated by a decline in efficiency of 31 percent, on they gain in allocative efficiency, they compromise average, relative to 1982 technical efficiency. As may also be expected, levels. By district, the agroecological variables variously affect all measures more productive districts of efficiency. Larger families and land fragmentation (Kericho, Muranga, and districts have have a negative influence on cost and technical Trans Nzoia) experienced experienced efficiency, and farmer age has a weak negative effect net productivity declines, substantially greater on technical and cost efficiency. Social capital has a while the others experi- weakly positive effect on allocative efficiency, while enced substantial gains distance to dirt roads has weak negative significance (with the largest gains in Most districts also for both cost and allocative efficiency. Households Kisumu and Machakos, show technical whose heads have attained higher levels of education followed by Bungoma progress, hut also have a lower cost efficiency, but the result is only and Taita Taveta). Most declines in efficiency. weakly significant. districts show technical progress, but declines in Productivity Change efficiency. The exceptions are Kericho, with no techni- Relative measures for individual years do not indicate cal change but a decline in efficiency, and Trans Nzoia, how efficiency or productivity has changed. These with significant technical regression but a modest gain changes are measured with the Malmquist index for in efficiency.4 productivity change, which is also decomposed into The measures reveal some unexpected trends. Figure indexes that measure technical and efficiency changes. 6.1 plots the linear trends in the cluster-level averages of The indexes are calculated by district to control for the three Malmquist indexes.' The data are sorted in regional effects in production and the economic envi- ascending order by the 1982 cluster-level average relative ronment. The results show that, on average, produc- efficiency. The trends show that clusters that had high 29 Agricultural Extension: The Kenya Experience average efficiency levels in 1982, generally in the more bound, and with little new technology forthcoming to productive areas, gained less in total productivity, and substantially raise production, their productivity has some may have regressed.' The trend in efficiency is stagnated. However, the efforts of the extension ser- similar, but shows relatively smaller changes. The techni- vices have been consistently focused on these areas. At cal change trend is much flatter but, again, negatively the same time, a general lack of improvement in correlated with the 1982 level of efficiency. Without any efficiency has meant that even in districts that have major technological advances, these results show a seen technological progress, the overall potential for convergence across regions toward homogeneity in the productivity gains has been compromised. level of productivity. Combined with the still very low overall efficiency How does this relate to extension? Figure 6.1 also (as measured by cost-efficiency) and the high degree of shows the trends of the variation in efficiency among farmers, these results If the potential for staff-farm ratios in 1982 suggest that, despite room for improvement, the exten- technical gains had and 1997. Other ratios sion resources have not been used as efficiently as they are not shown since these might have. While extension may have contributed to been properly trends are similar-that growth in the less-productive areas, its overall effective- assessed, a greater is, positively correlated ness appears to have been limited. The minor differences deployment of with the 1982 level of in the cost, technical, and allocative efficiencies, even extension staff in less- efficiency. The 1990 though the estimates are statistically insignificant, sug- productivetrend is steeper than that gest that extension has generally concentrated on dis- prodctie aeas of 1982, indicating that seminating technical messages rather than helping farm- might have been more the allocation of front- ers optimize their resource use or tailoring its messages cost-effective. line staff during NEP I to the prevailing economic environment. generally favored the The lesson that emerges is that extension services more productive regions. During NEP II, recruitment could have been allocated more efficiently. If the was frozen. The decline in the slope of the 1997 trend potential for technical gains had been properly assessed may thus reflect natural attrition in the frontline work (especially in the main areas of maize production that force. Productivity change is clearly inversely corre- had already benefited substantially from past research lated with the allocation of extension staff, and extension efforts), a greater deployment of exten- sion staff in less-productive areas might have been Conclusions more cost-effective. Considering the significant re- The results of this analysis, especially as summarized sources that are needed to sustain the current system in figure 6.1, are striking. The analysis shows that and MALDM's extremely tight budget constraints, it is there has been little change in the areas that were apparent that fewer resources could have been used to relatively more productive in 1982, while the other achieve the same results. And the returns to the regions have been catching up. This suggests that the investment in extension could have been much higher. more productive areas may have reached an upper 30 Results II Production Effects of Extension etermining returns to an investment in extension requires an estimate of its impact on agricultural production. Chapter 6 showed that the allocation of extension has been biased in favor of the relatively more efficient-and likely more productive-areas. This bias makes measuring the impact of extension difficult in a cross-sectional framework-that is, with the use of data from one point in time-and necessitates the use of more reliable methods using panel data. This problem was demonstrated in the context of ATD's evaluation of the impact of extension in Kenya (Bindlish and Evenson A Fixed-Effects Approach 1993). Working Paper 3 in support of this study The 1997 data confirm the problem with using data reconsiders the results of the ATD study, and discusses from a single cross-section. Statistical tests show that the technical details of the difficulty of interpreting its while the marginal effects of variable production inputs results.' The main finding is that the high estimated on farm production are stable with respect to alternative returns in the ATD study are very sensitive to regional regional and agroecological indicators, the coefficient effects. At the same time, correcting for inadvertent on the extension variable is sensitive. To overcome this data errors makes the results less robust. The sensitivity problem, the panel nature of the data is exploited.2 of the results precludes any judgment that the returns A difference model are positive. can control for the unob- To overcome the methodological limitations of a servable regional and cross-sectional framework, the current study used a agroecological factors, extension has been more robust method. It combined the 1982 RHBS Using this specification, biased in favor of data and the 1990 ATD data with a fresh survey of a separate model is esti- more productive the same households to develop a panel data set. mated for each of the (The technical details of the analysis and results are three two-year panels presented in Working Paper 5.) The objective of the (1982 and 1990, 1982 analysis was to identify the impact of extension on and 1997, and 1990 and 1997). In addition to the crop production by appropriately controlling for as variable production inputs, household characteristics many unobserved factors as possible. Of these, the are included in differenced form, since these also primary concern was with the unobserved natural changed over time. Varying weather conditions are productivity effects and other inherent regional controlled for by including farmer-reported crop perfor- socioeconomic or agroecological effects. mance indicators (normal or poor, relative to good); 31 Agricultural Extension: The Kenya Experience these indicators, however, are available for 1990 and Working Paper 5). Average cluster-level yields for the 1997, but not 1982. base year are used as a proxy for the initial condi- The remaining complication is the extension sup- tions.' Applying the model to the three panels confirms ply variable. The data on staff-farm ratios, used to that the resources have been allocated inefficiently. The measure the supply of extension services in each base year yields have a significantly negative coeffi- location, go back only to 1982. Thus, for 1997 and cient. The extension effect, however, is not significant 1990, it is possible to in either the production function or the reduced-form The less productive model lagged impact of supply function specifications for any of the panels.6 farmersextension, using weighted To confirm the hypothesis of the confounding effect farersan aras lags going back seven of initial conditions on the impact of the supply of have been years each.' For 1982, extension services, the simple difference model for all catching up as only the single-year mea- specifications yields a result similar to that described in new technology sure can be used. While it Chapter 6. That is, in the pure difference specification, reaces tem. would be desirable to productivity change is negatively and significantly have data on the previous correlated with the extension variable. A naive inter- years' supply of exten- pretation of this result would be misleading, since it sion, such data are not available. One way around the suggests that extension has a negative impact on problem is to assume that past extension efforts are production. embodied in the 1982 level of production. The post- 1982 changes in the supply of extension thus help Conclusions identify the impact of the new system. In addition, the The main finding of both this analysis and the material 1982 staff-farm ratio provides additional control for presented in Chapter 6 is that extension resources in the base level of extension supply. What the difference Kenya have been inefficiently allocated and poorly model measures, then, is the change in productivity targeted. At the same time, once the initial conditions that can be attributed to changes in extension after are controlled for, a statistically significant impact 1982.i Extension supply can thus be modeled either as cannot be established for extension. the difference between the cumulative extension supply It appears that the less productive farmers and for one of the later years (1990 or 1997) and the 1982 areas have been catching up as new technology reaches supply, or by allowing the coefficient to vary over time them. And while it is likely that extension has played a by including both variables independently. role in extending these technologies, this cannot be firmly established with the data in hand. Findings It is likely that more rational allocation of The results detailed in Chapter 6 point to the impor- resources would have achieved the same results more tance of distinguishing between the program effect-or cost-effectively. The lesson that emerges is that Kenya more precisely, efficiency in the allocation of extension needs to build a flexible and responsive system. With resources-and the direct extension effect. Given that little new technology forthcoming, as in the case of the allocation of extension staff has been, and continues maize, it is not economical for extension to maintain a to be, biased in favor of the more productive areas, and high-level presence. Instead, reaching new areas or that growth in agricultural productivity has been farmers previously not covered by extension services uneven from area to area, it is necessary to control for would have a greater marginal impact on both produc- the initial conditions in order to properly identify tion and poverty. What is needed, perhaps, is a "smart," extension's impact. flexible system that responds to imbalances in the To control for the effect of initial conditions, a delivery of information, targeting existing or emerging more flexible approach is used than a simple difference gaps between average and best practices. At other times, model. The analytical model is extended to yield an it would be more efficient to keep a leaner presence to empirical model that allows a distinction between maintain a local equilibrium, rather than blanketing all program efficiency, or the "program effect," and the regions with intensive coverage at all times. impact of extension, or the "extension effect" (see 32 Client Focus Farmer Valuation of Extension Benefits he benefit of public service to the ultimate beneficiaries is a critical issue for policy. If the beneficiaries were paying clients, the value of the services provided to them would simply be their market price. For most public services, however, there is no market. Traditionally, extension advice has been free because of its nature as a public good (low excludability and rivalry), with substantial positive externalities. But with limited resources, the issue of the efficiency of allocation across a number of possible public goods remains, and hence the need to measure the benefits of the services provided. The approach this evaluation uses is briefly dis- A measure of benefits would also allow consideration of cussed in Annex F Working Paper 6, on farmers' cost recovery measures. Even if it is only partial, cost willingness to pay, presents a more detailed discussion recovery has several benefits: it provides appropriate of the survey design, incentives, and therefore accountability and client respon- an important part of Cost recovery, even if siveness; it brings budgetary respite; and it promotes CVM; the tests and pluralism by allowing alternative providers, particularly controls to check the private suppliers, to enter the market. consistency and reli- appropriate incentives, With respect to the relationship between extension and ability of the farmers' and hence accountability poor farmers, some pertinent issues are their demand for responses; and the de- and client responsive- advice, their willingness to pay for it, and their ability to tailed descriptive and afford the payments. Theoretically, the upper limit of what analytical results. n an individual would be willing to pay for a service would The key findings are respite; and it promotes be the maximum private net benefit derived from it. This summarized here, pluralism by allowing benefit can be estimated either directly or indirectly. One alternative providers, indirect method is to estimate benefit from the impact of the Desired Frequency particularly private service on a farmer's productivity, as discussed in Chapter of Visits 6. This method, however, assumes that the service is Some farmers (4 per- suppliers, to enter the delivered, and delivered in a manner that is efficient and cent) indicated that market. effective. More important, it does not reveal whether the they do not want any farmer is willing to pay for it. A direct method is the extension advice, and some (another 4 percent) do not contingent valuation method (CVM), which elicits from want the current service to continue. For the remainder, farmers their willingness to pay for the service, giving some the median number of desired visits is three each year, idea of what they perceive to be its benefits. with a modal value of two. More than two-thirds of 33 Agricultural Extension: The Kenya Experience the farmers want fewer than one visit every three tested. Households that meet extension agents either in months. Even among contact farmers, almost half want their own fields or in a baraza do not have a to meet the extension agent no more than once every significantly higher WTP than those who do not three months. These responses suggest that the norm of normally meet extension agents or who meet them very biweekly or even monthly visits under NEP I and NEP infrequently. Farmers who normally meet at a coopera- II exceeds what farmers want. tive society have a negative and significant effect. Finally, a variable indicating whether a farmer is a Willingness to Pay contact farmer was tested, but it failed to attain Overall, a small proportion (9 percent) of the farmers significance. who would like to receive advice (including 12 percent of Listening to radio programs has a strong positive the contact farmers) are not willing to pay for it. More and highly significant influence on WTP, This could be than half of the farmers are willing to pay individually, because hearing information on the radio encourages while the rest prefer farmers to seek additional or more detailed informa- A vast majority of farmers to pay in a group.' tion-enough to convince them that it would be are willing to pay for The mean willing- worthwhile to pay for it. It may also reflect their ness to pay (WTP) current lack of access to such additional information, extension services. Of for individual con- and their consequent inability to follow up on what these, over half are tributors is Ksh 67 they hear on the radio. willing to pay for each visit, and The results also show that VTP is significantly individually, the rest to for group contribu- higher among those who are willing to pay in a group. p tors, it is Ksh 51 for Both gender and education make a difference: house- pay in a group. each visit; the mean holds headed by women and households whose heads WTP is Ksh 60. At have a primary or, especially, higher-level education the time of the survey, the daily wage rate for agricultural have a significantly higher WTP. Other variables do labor was Ksh 60. The farmers who are unwilling to pay not appear to influence WTP. Social capital variables (at all, or as individuals) most frequently cited lack of show weak but mixed effects. The membership of funds as the reason. households in groups (either of farmers or their The average total annual WTP is Ksh 346, with a spouses) increases WTP, but a higher incidence of median of Ksh 160. The range is quite wide, from 0 to groups within a location has a generally negative 8,640, but the mean WTP is significantly different from effect. Perhaps the latter reflects the benefit of having 0 (standard error of 32). The interquartile range, greater access to information, while the former re- however, is tighter, between 60 and 360. flects the benefit of collective action. Among infrastructural variables, only access to dirt roads Factors That Influence Willingness to Pay affects WTP (households that live farther from a dirt Systematic variation in WTP by socioeconomic or road are willing to pay more), although the effect is agroecological characteristics, or with the alternative weak. Households living in lower-potential zones extension methodologies currently in use, is important have a lower WTP, as do those living on hills and for policymakers to more effectively and efficiently undulating terrain. target future services. It is also important in determin- ing the perceived benefit from the services that are Conclusions available to different farmers. A significant proportion of farmers would like to The influence of existing extension services on receive extension services and are willing to pay for WTP is viewed from three perspectives. One is the them. The perceived benefit, however as reflected in effect of its current supply of extension services. This the total amount that farmers are willing to pay, is well has a small positive effect, but does not attain statisti- below what the government currently spends per farm cal significance in any specification tried in the on extension services. Also, the frequency of visits the analysis. Next, to test the influence of alternative farmers desire is much lower than was presumed in methodologies, variables indicating the households' NEP I and NEP 11 designs. An econometric analysis normal meeting place with extension agents were suggests that TP is not related systematically to the 34 Client Focus: Farmer Valuation of Extension Benefits level or methods of extension currently in use. One that farmers value agricultural advice and are willing interpretation of these findings is that the farmers' WTP to share its cost. It also reflects the current lack of an reflects an unmet demand for services, which is also alternative source of information. indicated by the close, statistically indistinguishable At the same time, willingness to pay of contact farmers and those who the uniformity of the Farmers value have never before received advice. Another indicator is level of demand (that agricultural advice the desired frequency of visits, which is approximately is, the frequency of vis- and are willing to share the same across all categories of farmers. Finally, the its) and W" suggests strong influence on WTP of farmers' listening to the that it would be more its cost. radio probably reflects their inability to follow up on efficient to cover a the information they get through that medium. larger number of farmers, but with lower intensity and a The study results have important implications for higher quality of contact. The large positive influence of the design of future extension services. The most radio programs on WTP suggests that complementary important is the implication for cost recovery and the extension approaches should be exploited for potentially possibility of incorporating an endogenous quality significant synergy. It is also possible that radio programs control mechanism in the delivery system. Considering could be used to whet the farmers' search for information, that even those who do not regularly receive extension which could then help promote the provision of more services, or who meet agents only infrequently and in specialized private extension, or extension for a fee. public gatherings, are willing to pay clearly indicates 35  Conclusions and Lessons he rationale for providing extension services in Kenya is still relevant, but the evidence suggests that the extension approach used by NEP I and II was not efficacious. The overall record of the T&V extension system implemented in Kenya has been disap- pointing. The extension approach adopted by NEP I and NEP II has not proven to be effective, and the current system is not sustainable. Although the system's geographical coverage, research- extension linkages, and staff skills have been improved by the projects, the outreach of the system is low, and the interaction between the extension farmers are willing to pay for them, is well below what agents and the farmers is qualitatively well below the government is currently spending per farm to what was anticipated. deliver the services. The findings suggest that a more The evaluation reveals that there is an unmet rational allocation of extension resources would have demand for extension services, and the farmers value been more cost-effective. access to advice enough to be willing to pay for it. Despite the substantial scope for improvement, how- Lessons and Recommendations ever, the data do not provide evidence of any signifi- The main lessons and recommendations to emerge cant impact of the current extension system on farmer from this evaluation include the following. efficiency or crop productivity. On the contrary, all Targeting. The first lesson is the need for more approaches indicate that the current institutional efficient targeting of arrangements have been ineffective in delivering the extension services to much-needed services to the vast majority of Kenyan focus on areas and Teran unet farmers. It is likely that NEP I had some beneficial groups where the mar- impact early in its implementation period. The ben- ginal impact is likely to services, and the efits, however, appear to have been short-lived. The be the greatest. This farmers value access available evidence does not indicate any significant calls for a more flex- to advice enough to be impact, even by 1990. The results do show that ible, "smart" system extension resources have been allocated inefficiently. that can identify the The various estimates obtained in the evaluation's gaps between existing analysis show that a positive rate of return to the best practice and average practice and allocate expenditures on extension cannot be established. Fur- scarce resources more rationally. Further, the farm- ther, the worth of the perceived benefits from the ers selected for interaction should be more represen- current services, as indicated by the amount that tative of the local socioeconomic environment so 37 Agricultural Extension: The Kenya Experience that more relevant advice can be delivered to taking over from their parents; radio programs are different categories of farmers. popular among farmers; and alternative providers are Information systems. Targeting calls for appropri- beginning to emerge in rural Kenya. It would be ate flows of timely and reliable information, and hence advisable to adopt a more cost-effective strategy that for monitoring and evaluation (M&E). An important exploits the synergistic effects of low-cost modern lesson emerging from the Kenyan experience is the need communications, demonstrations, printed media, and to identify farmer demands and tailor the service to suit partnerships with civil society and the private sector. local technological and economic conditions and cir- This would leverage the resources to increase outreach, cumstances. M&E is also critical in identifying the and is likely to have a greater impact with the same or gaps and guiding the "smart" system for more efficient lessened demand on government resources. targeting of services. Client focus. The central focus of the institutional Intensity. Reflecting their current experience, farm- design should be on empowering the farmer. An ers do not want to see the extension agent too often; and effective way to incorporate client focus is to consider there are not enough new technological recommenda- alternative options that give a voice to the farmer, tions to sustain a high intensity of visits. It would be such as cost-sharing, farmer organizations, decen- more cost-effective to establish a leaner and less- tralization, and the like, as an integral part of the intensive presence, but with wider coverage. It may be delivery mechanism. Cost recovery (even if only that with improved quality of service, the demand will partial), in particular, would be advantageous: it increase. To be ready for such a change, it is impera- provides appropriate incentives, addressing the issues tive that a responsive and dynamic delivery system be of accountability and quality control; it renders the in place (as in targeting, above). service more demand-driven and responsive; it pro- Pluralism. A blanket approach, using a single or vides some budgetary respite; and it encourages uniform methodology to deliver standard messages, is alternative providers. Such institutional arrangements likely to limit the effectiveness and efficiency of remain unexplored in Kenya. extension services. Youngean more educated farmers are 38 ANNEXES ANNEX A. BACKGROUND Among the several goals often cited for agricultural In Africa, agricultural extension has been central to extension services, the most common is agricultural the Bank's development strategy (Cleaver 1993). The development (Feder, Willett, and Zijp 1999). The strategy for the new millennium, designed to "focus on objective of extension services is to sustainably increase a few selected national and thereby systemic programs agricultural productivity by, among other things, of high impact," also lists extension as a key area for expanding the knowledge farmers have about new Bank support in Africa (World Bank 1997). In the past, crops, crop varieties, inputs, and better husbandry and this strategy largely relied on the T&V system, with management practices. The importance of science-based national programs in more than 22 countries designed technological advances in raising farm productivity to follow its guidelines (Venkatesan and Kampen 1998). makes agricultural extension key to development, and The effectiveness of the T&V system of extension, has brought about consistent Bank support for such particularly its cost-effectiveness, has been subject to activities in many of its borrowing countries. Over the much debate. The central issue has been the institu- past two decades, the Bank has invested about US$4 tional design and efficacy of T&V relative to alterna- billion worldwide in extension projects. A large number tive mechanisms for delivering extension advice of these projects have used the T&V system of manage- (Picciotto and Anderson 1997). Within the Bank, the ment (along the principles laid out by Benor, Harrison, debate has been passionate, and often emotional. The and Baxter 1984). focus has largely been on conceptual issues, but little BOX A.I. THE DIFFICULTY OF DRAWING INFERENCES FROM FIELD VISITS ost of the more successful farmers, extension services were now with the bilateral field visits or those who had ben- positive, pointing to the agency) about their for this efited from extension, benefits of the system. dilemma. The diffi- evaluation were were visited. The farmers The visit organized by culty was in deciding arranged through the visited apparently the bilateral agency, which program to extension service. Each received many missions, however, was to a district discuss and, more visit typically entailed since most of them kept that received both NEP important, which pro- an entourage of mission an impressive diary that and bilateral aid funds gram to praise, since members; resident mis- the visiting "dignitaries" At the start of the visit to their approaches were sion staff; ministry were obliged to sign. the district extension very different. The representatives; provin- Such experiences are office, the extension staff project staff resolved cial or district staff, or unlikely to be insightful, seemed uncomfortable, the issue by noting that Iboth, often including however. Their problem was that the Bank staff were the officer in charge; The moral hazard the "mission" included actually from OED, several subject matter confronting the field representatives of both and that the extension specialists; divisional extension staff was funding sources, the local staff should feel free staff; and the local revealed during a visit program head of the express their feelings frontline staff. The organized by a bilateral bilateral agency, and a about NEP. Of course, group usually arrived donor to view an alterna- Bank staff member. The the rest of the discus- in a motorcade of three tive extension approach. district officials confided sion revealed that the to five vehicles to visit Most previous visits had in the accompanying program funded by the with farmers or groups been to districts funded local staff (some of bilateral agency was that normally worked only by NEP, and most whom had previously the program of choice. wihextension agents. discussions of the issues worked for the govern- Almost always the and problems with the ment service but were 39 Agricultural Extension: The Kenya Experience supporting evidence has been offered. The limited The Kenya study (Bindlish and Evenson 1993), of evidence that is put forth usually comes from assess- particular interest here, was part of an effort by the ments made during field visits. As expected, support- Africa Region to assess the impact of the large amount ers of T&V generally make positive assessments; its of development resources going to extension in Africa. critics generally draw negative conclusions. While The Africa Technical Department (ATD) undertook the there is probably some truth in all the assessments, the study to evaluate the impact of the agricultural exten- reality is largely obscured-most field visits are sion projects it had supported in Kenya and Burkina unlikely to be representative or unbiased, either in the Faso in 1990.2 As noted, the studies estimated very high manner in which the data are obtained or in their returns to extension, especially in Kenya. The findings interpretation. The facts are also often colored by have been controversial, however because of their moral hazard on the part of the local extension staff various limitations, some of which the authors them- when dealing with so-called random donor visits (see selves noted (Bindlish and Evenson 1993, p. 29). box A.1). At a time when many borrower countries were Despite the intensity of the debate, very few becoming concerned with the high costs of the T&V attempts have been made to rigorously measure the approach, and concern was increasing within the Bank impact of T&V, or the lack of impact. Three notable about the development effectiveness of its extension exceptions attempted to estimate the returns to T&V portfolio, the high estimated returns were greeted with investments.' All were conducted by the Bank, and mixed feelings, and even skepticism, in some quarters included work in India (Feder, Slade, and Lau 1985), (World Bank 1994; Purcell and Anderson 1997). Burkina Faso (Bindlish, Evenson, and Gbetibouo 1993), Nevertheless, since the evidence was based on house- and Kenya (Bindlish and Evenson 1993). Positive but hold survey data and formal statistical methods were varying degrees of impact were found. All three studies used, the estimates of high returns lent credibility to the used survey data, but were subject to limitations claims of T&V supporters. 3 The findings vindicated imposed by the available data. Other studies have the Bank's stated policy of using extension as a major considered the effectiveness of the T&V approach in plank in the overall rural development strategy for other settings-for example, Hussain, Byerlee, and Africa (Cleaver 1993) and justified speeding up the Heisey (1994) in Pakistan-and the findings have been already rapid introduction of the T&V system in generally mixed. A number of studies of T&V's Africa. At the end of 1997, 22 countries had a national operational aspects, most of them critical of the extension program with a T&V system of manage- approach, have failed to assess the full impact of the ment, with active Bank projects supporting a total extension system. investment of more than US$700 million. 40 Annexes ANNEX B. THE DESIGN OF NEP I AND NEP II The Bank introduced the T&V system of management important aspects of crop production activities, as well as a pilot in two districts in 1982.1 Following the brief as low-cost improvements that the majority of farmers pilot, the system was expanded to 30 of Kenya's 41 could afford. The implicit assumption was that once districts over a 3-year period, covering all high- and productivity and revenues had increased, the farmers medium-potential areas. NEP I was designed as the could graduate to more costly technological compo- first phase of a longer-term institutional development nents with the additional income generated. plan: the T&V system was to be introduced, and then Technical officers and subject matter specialists improved over time. It was primarily an institution- were to supervise and back up each FEW Every two building project, and it was anticipated that external weeks, the FEWs were to receive a full day of intensive assistance would be required for 10 to 15 years. technical training on the messages they were to deliver The project design followed standard T&V prin- during the following fortnight. The subject matter ciples. Project activities would provide farmers regular, specialists were to upgrade their knowledge and skills systematic, up-to-date advice on the farming practices through monthly training workshops attended by best suited to their specific conditions. The program research scientists. By design, at the district level alone, was initially restricted to the crop extension service. the ratio of non-FEW to FEW staff was 1:3.2 The Frontline extension workers (FEWs) would visit farmers project also provided funding for transport to increase regularly and receive systematic training and technical staff mobility; allowances for field staff; audiovisual support from research staff. Each FEW was to divide equipment; civil works to build office space where none the farm families in their jurisdiction (then anticipated existed; incremental operating costs; and the produc- to be between 400 and 800 farmers) into 8 groups. Each tion, publication, and updating of extension manuals group was to be visited every fortnight (four one week, for all staff by the Agricultural Information Centers and four the next). Since it was impossible to visit every (AICs). group member on any one day, five to ten contact NEP 11 sought to further strengthen extension farmers were to be selected from each group. About 10 services and support their expansion to uncovered percent of the farmers would thus be designated contact areas, including the dryer zones; provide funds to farmers, and the extension workers were to work improve staff transportation; foster the use of mass mainly with them, demonstrating practices that would media and communications; rehabilitate and refurnish be followed in the next two weeks, and were to involve FTCs; and promote links between research and exten- as many other farmers as possible in the demonstra- sion by funding transportation and equipment, allow- tions and discussions. The selection process was to ing greater participation of research staff in extension entail an inventory of all farmers, identifying the training sessions, demonstrations, and farm trials. The particularly poor farmers, whose progress was to be project's goal was to effectively deliver technical monitored and evaluated. messages tailored to the needs of smallholder farmers, The initial focus was to be on simple messages, especially women, and increase yields of both staple concentrating on a few important crops and the most and export crops. 41  Annexes ANNEX C. INSTITUTIONAL FEATURES Management multifaceted pressures, including involvement of the An immediately notable feature of Kenya's extension Bank and other donors in the context of the preparation system is that not only is monitoring and evaluation of the Kenya Agricultural Sector Investment Program, (M&E) nonfunctional, but even basic management as well as support from other Bank projects. The information is missing. Data are not readily available resulting studies, as MALDM documents, reflect re- on the number of extension staff, their operational newed thinking about extension and deeper and candid capacities, or even on extension's annual expenditures. assessments of the current state of Kenya's extension While some of the underlying factors go beyond the efforts. Progress toward developing a national policy extension department, and indeed beyond MALDM, it has nevertheless been slow, and a final policy paper has is apparent that NEP I and NEP II had no impact on not yet been produced. this important aspect of management. Overall manage- ment of the projects was also weak. Poor financial Incentives arrangements compounded the problem of inadequate Another consequence of the mechanistic implementa- resources, and poor implementation arrangements tion of the projects' design has been inappropriate have impeded the functioning of extension services. incentives, both institutional and individual. The This proved particularly significant during NEP II, "rules of the game" are key elements of institutional during which the management of extension services development, especially for service delivery, as they rested with a working group with insufficient authority; determine the incentive structure. While the bureaucra- it was unable to coordinate the activities of the tization of the extension services cannot be attributed to agriculture, livestock development, and veterinary the projects, the hierarchical structure of the T&V departments. As a result, the management during most design has not improved the situation over that before of NEP II was ineffective. This is particularly signifi- NEP. The result is that the extension service is both top- cant since the staff appraisal report for NEP I promi- heavy and headquarters-centric (supervisory staff is nently stated that T&V was first and foremost a excessive, with a large concentration in Nairobi). management system. Extension's relative success and free flow of funds Another shortcoming of the projects' work in in the beginning led to an intolerance of dissent and institutional strengthening is the continuing lack of a neglect of emerging problems. The perceived manage- strategic vision or national policy for agricultural ment benefits of the projects, in the monitorability of extension. Several observers have noted that this project outputs and accountability of staff (for example, reflects a preoccupation of extension's management in terms of number of visits, number of training with the modalities of the delivery mechanisms, with- sessions, and whether or not FEWs were strictly out regard to policy, planning, or management of the following their assigned routes) and the focus on extension services.' The new institutional paradigm delivery of specific and well-defined messages, put in introduced by NEP I helped increase the level of energy place adverse incentives. As in any bureaucratic orga- throughout the program in the early years because of nization, staff accounted to their supervisors, not to the the large influx of operational and development funds, clients, and what was monitored was the number of the availability of new vehicles, payment of allow- visits, not the quality of the meetings. ances, and significant training. These changes raised This lack of accountability to the farmers is morale and motivated field staff, and the detailed observable in both the household and the staff surveys. implementation program, with its clear chain of com- The household survey shows that the proportion of mand and well-defined bureaucratic staff assignments, farmers who need advice-that is, the demand for dispensed with the need for policy or planning. information-on more complex messages is twice that But with a deteriorating financial situation and of the farmers who need information on simple agro- ineffectiveness of the services, efforts were begun nomic practices. In contrast, the proportion that finds recently toward developing a national policy and a simple agronomic messages most applicable, which framework for the future development of extension in reflects the effective supply of information, is twice that Kenya. Progress so fare however, has been driven by finding the more complex messages applicable. Simi- 43 Agricultural Extension: The Kenya Experience larly, the mismatch between what the farmers want and at the end of 1996, more than 48 percent of all ministry what extension services supply is reflected in staff were engaged in extension. At the field level, the extension's methodology. According to the FEWs them- agricultural FEWs numbered approximately 6,841, selves, field/home visits constitute the least popular more than double the 3,328 of 1982. Veterinary and method among the farmers (only 5 percent prefer livestock production FEWs numbered 357 and 547, them), and demonstrations and field days are the most respectively, in 1996 (comparable estimates for 1982 popular. Yet the majority of the FEWs prefer to use field are not available). This yields an average of about 500 visits. farms per agricultural FEW The number of technical (non-FEW) agricultural staff is estimated at about Sustainability 1,577-that is, there are about 4 FEWs for each The available estimates of total expenditure on exten- supporting technical staff member. sion do not present a complete picture.2 How expendi- The unsustainable growth in staff through the tures have evolved over time is not known.' The most 1980s, which currently takes up about 80 percent of reliable are the "printed" estimates, which suggest that operating costs, led to a recruitment freeze in 1990. for the year 1996-97, expenditures were approximately While the increase in staff numbers was the result of KP 156-177 million (or US$54-61 million)4 out of a government policy and presumably unrelated to NEP total ministry budget of KP 340 billion.s That is, about I (during which most of the increase occurred), it is 46 percent of the ministry's budget goes to extension noteworthy that the ratios of farm families to FEWs activities. A review of public expenditures for agricul- and FEWs to technical staff are very close to the ture also estimated that about 60 percent of the original NEP I prescription of 500:1 and 3:1, respec- agricultural budget is devoted to extension activities, of tively. It appears that the number of farm families was which 70 percent is donor-funded. underestimated at the start of NEP 1, when the farm- With Kenya's total of about 3.44 million farm to-FEW ratio was much higher than the reported families, according to the Welfare Monitoring Survey 500:1. and the Staffing Norms Study, these estimates suggest These estimates of staff strength, however, are that Kenya spent an average of about US$15.11 per farm inconsistent with the data provided by the districts. The family in 1996-97 for extension services, or Ksh ratio of farms to FEWs is, on average, about 1,100:1. 876.38.6 Comparable estimates for 1982 are not avail- The discrepancy is significant and not easy to explain. able, but estimates for the districts in the 1990 ATD It may stem from the poor identification and classifica- study were US$3.92 for 1982 and US$4.67 for 1990 (in tion of staff by their current assigned duties (which do constant 1991 dollars). These figures compare with the not necessarily correspond to the assigned job codes), current estimate of US$13.29 (in 1991 constant shillings and probably also reflects a concentration of staff at at the 1991 average exchange rate of Ksh 27.5 per US$1) headquarters, provincial, and district offices, where or US$15.11 (in 1997 dollars, Ksh 58 per US$1). staff are involved in essentially nonfield activities. This While the optimal extension expenditure level is is reflected in the staff survey: a majority of the SMSs debatable, the problem facing MALDM is that the indicated that they had responsibilities in subject areas current system is too expensive and not financially other than those of their expertise. But regardless of the sustainable. Even toward the end of NEP II, project particulars, it is apparent that the current system is funds were financing 90 percent of the system's overstaffed and expensive. nonsalary operating costs. The government budget is insufficient to keep the staff mobile and effective. A Pluralism vast majority of both FEWs and SMSs confirm that Pluralism was clearly not a characteristic of T&V as funds for transportation and allowances are a "serious implemented in Kenya. NEP I introduced the contact or very serious" constraint on the effective delivery of farmer, and later the contact group, approach, with all extension services. Similarly, training sessions and the extension staff time devoted to delivering advice to monthly workshops have been reduced significantly the contact farmers. This was achieved by eliminating because of a shortage of funds. all other extension activities. which may not have been A major reason for the inadequacy of funds is the effective in all cases, but were nevertheless efficient large number of extension staff. Estimates indicate that alternatives for delivering certain types of general 44 Annexes information. The NEP I and NEP II approaches, requires a significant amount of training to convert however, were not very successful in reaching specialists into generalists and vice versa, which the noncontact farmers (the briefing in Annex D, prepared already trained and experienced livestock staff resent. by extension staff in one of the districts for a focus group meeting, gives some insight into the reasons, and Training the effectiveness of the NEP I and NEP II approaches), The benefit of NEP I and NEP 11 that is most widely but institutional energy has remained focused on the agreed on is the upgrading of staff skills through T&V approach. training.' NEP I provided substantial training to older A vast number of FEWs and supervisors acknowl- staff and newer untrained staff. The regular training edge working on projects of other donors and working schedules were effective and had a positive impact on with NGOs and the private sector, mostly for the staff quality. After the initial years, however, funding additional incentives such as allowances, mobility, and constraints, strained research-extension links, and the training. Most such activities are also differently lack of new technology reduced the effectiveness of organized, and staff generally consider them to be training sessions. The quality of extension staff is also more effective. Whether or not they truly are, or if this attested to by farmers' positive assessments of their belief is a reflection of better working conditions, is competence (as noted in Chapters 3 and 4) and by unknown. However, there has been no effort to institu- assessments from NGOs and other donor projects that tionalize the significant level of ad hoc activity to make often use extension staff (although with additional the system more effective or rational. training). A negative impact of NEP I on an alternative Despite these significant efforts, however, a major- institutional arrangement was the discontinuation of ity of the SMSs feel that FEWs are not qualified to the soil conservation program (Tiffen and others carry out their responsibilities. A majority also feel that 1996). The program had been well established and there are too many subjects to handle effectively; a functioning reasonably well since 1974, but the large number of FEWs also voiced this sentiment. collective action required for soil conservation exten- Finally, limited funds have restricted training sessions, sion was not amenable to the contact farmer or even a although the primary reason the vast majority of SMSs small group, approach. It cannot be demonstrated on give for reduced frequency was that there was "nothing small plots, nor can it be reduced to simple messages. new to say." This reduced frequency of training, The program was therefore discontinued. It was however, is reflected in the FEWs' demand for more reintroduced in 1988 as a separate branch office, training sessions, since they see their effectiveness supervised and supported by the Swedish International declining. Development Authority. The new program introduced the catchment area approach. Research-Extension Links A more recent unintended impact has been on A critical element in the high-intensity T&V approach livestock extension. NEP I had concentrated on agricul- is a regular flow of messages from research. In the ture. Even though the Ministry of Livestock was early years, the link between research and extension merged with the Ministry of Agriculture briefly in the was weak, but sufficient to ensure a supply of simple early 1980s, the livestock extension department oper- messages. Over time, as funding became tight, prob- ated separately until 1991, using its own approach, lems started to emerge. Eventually, when KARI was largely funded by other donors. Early attempts at separated from MALDM, the link was totally severed. including livestock extension in NEP did not succeed, In 1993, renewed efforts under NEP 11 led to a because the livestock department resisted the T&V memorandum of understanding between KARI and approach. The reamalgamation of the ministries and MALDM to reestablish the link. The staff survey, the more recent unified approach to extension pro- however, reveals that the link is still very weak. The moted by NEP II are creating significant tensions. In majority of both FEWs and SMSs have noted that general, livestock extension is not amenable to deliver- meetings with researchers are inadequate and infre- ing messages at predetermined times, and it does not quent and that participation in field trials is limited. require frequent visits. Livestock advice is based on One constraint is the lack of adequate adaptive solving specific problems and is not seasonal. It research to generate new messages. Limited feedback 45 Agricultural Extension: The Kenya Experience from farmers through extension has led several regional research centers to conduct their own partici- patory rural appraisals to identify the farmers' prob- lems and to target adaptive research. More recently, under the Farming Systems Approach to Research, Extension, and Training Initiative, with KARI's leader- ship, the link appears to be improving, and some new technologies are emerging. 46 Annexes ANNEX D. FOCUS GROUP BRIEFING ON NEP'S IMPACT A field extension staff member prepared the following Frontline extension workers meet with four groups a briefing points for a focus group meeting that discussed day, primarily women's groups. They meet once a the impact of NEP I and NEP II in one of the districts. month, and therefore the FEW would either meet the owner of the home (a chairlady) alone or very few NEP I members of the group. * Farmer selection was not handled well. In most cases, The lack of serious technical packages resulted, the farmers were handpicked by extension staff. again, in repetitive and boring messages. * The farmers were unwilling to go to the same home Most of the demonstration sites were based at the every time, so that a program of fortnightly visits chairladies' homes, with little benefit to the members. was not well received. No cleat packages came from research except for the * In most cases, poor follow-up led to the failure of normal agronomic messages that the farmers had follower farmers to take up agents' messages and already practiced for a long time. replicate them in their homes. The farmers' problems have still not been solved by * Repetitive messages and the lack of clear technology the existing extension approach. For example, (a) packages led to monotony in the information being correct seed varieties for beans, maize, sunflowers, passed to the farmers. and the like; (b) crop pests and diseases are still being * Feedback mechanisms-from the farmers, through researched, and there is little assurance that success extension, to research, and vice versa-were not will come soon; and (c) the lack of credit supply has effective. led to poor or low adoption rates. * Research programs rarely addressed the farmers' Mobility has continued to be the greatest problem at needs in the field, which led to poor linkages. the frontline, leading to very poor coverage at the * Individual farmers who were left out felt that the field level. The districts have good vehicles but extension service was aimed only at a few well-to-do insufficient funds to maintain them. Field staff are farmers. This is the reason for the administration's poorly remunerated. permanent attack on extension-that its agents were The administration of funds, particularly problems never seen. with district treasuries' liquidity, always led to the * Division and district staff supervision was difficult failure of demonstration plots compared with those of because transportation was lacking and route maps the farmers. were unrealistic. Funding for research programs depends on the * Integration with other programs such as soil conser- interest of the donor agencies, which generally do not vation and home economics was minimal. address farmers' needs. For example, only one cluster * While funding was adequate for program activities, was selected for an entire district of 14 divisions that more than 90 percent of support went to staff had very varied farmer needs. activities, and less than 10 percent went directly to The top-down approach resulted in farmers' expect- the farmers. This was the reason for low adoption ing free things, and the projects' approach as it was rates. Extension packages were judged impractical implemented seemed to be imposed on the farmers. for simple farmers. Farmers' views on how extension should be con- ducted should have been taken into account. NEP II * Coverage was fairer than the contact farmer approach, but most areas did not have worthwhile groups. Groups made up mostly of women had a "merry-go-round" agenda with very little agricultural activity. 47  Annexes ANNEX E. AWARENESS AND ADOPTION OF EXTENSION MESSAGES This annex briefly summarizes the data used for the Survey Design awareness and adoption analysis; results are discussed OED's survey covered four crop categories: the main in Chapter 5. Complete details on OED's survey, the crop, which was maize, cash crops, minor food crops, methodology of the analysis, and the results are given and new or promotional crops.' The questionnaire, in Working Paper 2. which covered messages on 13 cropping activities, was designed to gain an appreciation of the sophistication Stock of Messages of the farmers' knowledge.2 The coverage of the The first step of this strategy was to establish an different crop categories was motivated by the status of inventory of extension messages and technologies that the main crop, which was maize for all study districts. are available from the research system. This effort Maize has been the target of extension activities since yielded limited results. Few new technologies were the mid-1960s, and therefore may not be very useful in recommended during NEP I and NEP II, and those few determining the impact of extension services at the time generally took the form of updated varieties, without of the survey. Cash and minor (or non-maize) food major changes in practices. The recommendations for crops were covered to assemble a more complete most practices have remained essentially the same for picture, and the new and promotional crop category the past 15 years. Obtaining specific extension mes- was included as a test of the effectiveness of the sages proved to be difficult, because district farm information dissemination system. management guidelines have not been updated. That Some limited comparative results for 1990 are the technology stock and associated messages have available from the ATD survey. It should be noted, remained fairly constant is noteworthy. The evalua- however that the ATD data have information on only tion, however, was able to establish that recommenda- one crop (mainly maize) for the vast majority of the tions for the 13 main activities for the crops grown sample of 420.1 Results from the ATD and OED most commonly in the study districts do exist, surveys must thus be compared with caution. 49  Annexes ANNEX F THE CONTINGENT VALUATION METHOD AND ITS APPLICATION The contingent valuation method (CVM) is a tool for This was followed by a question on whether the eliciting individuals' use and nonuse values for a farmers wanted the extension program to continue. The variety of public and private goods and services. The farmers who did want it to continue were reminded of method relies on describing a hypothetical situation to the number of annual visits they wished to have and a sample of individuals and asking them to state their asked whether they would be willing to pay individu- willingness to pay to avoid a particular change in that ally for extension services. Those who said no were situation, or their willingness to accept compensation then asked whether they would be willing pay as a for a proposed change. The name of the technique group member, if such a group were to be organized. derives from the fact that what the individuals report is The farmers were then asked how much they would be contingent on the scenario that is described to them. willing to pay per visit. In its simplest form, CVM asks individuals directly The format of the CV questions was altered about their willingness to pay to maintain the status between the first and second rounds of the survey in quo; that is, to avoid a change in the provision or order to test for various biases. In the first round, about quality of goods or services. The method has been used half of the farmers were asked double-bounded referen- widely in industrial countries to estimate nonuse dum, or closed-ended, questions, and the other half values, typically for environmental and public goods. were asked open-ended questions. In the second round, Its applications in developing countries are growing. It farmers asked closed-ended questions in the first round has been used to value, among other things, improved were asked the open-ended questions, while the rest sanitation services, household water services, surface were asked a costless choice question-that is, whether water quality improvements, tsetse control, forest they would prefer to receive extension services or a protection, and wildlife viewing. This is the first known market good (such as sugar or kerosene) worth approxi- application of CVM to elicit the willingness to pay for mately 100 Ksh. The results reported in this evaluation agricultural extension services. This is also the first are from the pooled responses to the open-ended known application for an impact evaluation of a questions. The detailed description of the survey design project. and results are given in Working Paper 6. In the OED survey, the farmers were first asked whether they wanted to continue receiving extension advice, or would like to start receiving advice. Those who answered in the affirmative were then asked how many extension visits they would like to receive each year. The following statement was then read to the farmers: The cost of providing extension advice (including transport costs, salaries, etc.) has been mostly financed by the government. The lack of funds is a major obstacle in providing extension services. This could lead to irregular visits by the extension workers, and a deterioration in the quality of the service. There is also the possibility that the extension program could be eliminated altogether. 51  Annexes ANNEX G. COMMENTS FROM THE GOVERNMENT OF KENYA REPUBLIC OF KENYA MINISTRY OF AGRICULTURE Telephone 718870, Fax 720568 THE PERMANENT SECRETARY When replying please quote KILMO HOUSE Ref. No. OFTA/9 CATHEDRAL ROAD and date P.O. BOX 30028 NAIROBI 28 June, 1999 Gregory K. Ingram Manager Sector and Thematic Evaluations Group Operations Evaluation Department The World Bank Washington, DC 20433 USA Fax 202-522 3123 Dear / RE: COMMENTS ON IMPACT EVALUATION OF AGRICULTURAL EXTENSION PROJECTS Refer to your letter dated 10th June 1999 on the above subject. Please enclosed herewith final our comments on the document for your consideration and incorporation to the final document. Yoursus AMB. JOSHUA K. TERER PERMANENT SECRETARY Encl. 53 Agricultural Extension: The Kenya Experience Ministry of Agriculture's comments on the Impact Evaluation of Agricultural Extension Projects in Kenya [Numbers refer to original paragraph numbers in draft questionnaire is appended hence it is difficult to submitted.] establish the quality of data. There is also no indication of use of informal survey methods to 1.0 Introduction verify the results from the formal survey. 1.1 Generally the document has captured most of the 2.5 Comparison has been made using data and areas in evaluating the Agricultural Extension information of 10-15 years ago. There is no Projects in Kenya. However, we have the follow- indication of any survey done before the start of ing comments to make. project implementation, hence the rationale 2.0 Specific Comments mentioned here does not hold. It is also stated in 2.1 In 1994, the animal health extension was incor- the document that there was no baseline study porated in NEP II activities after the Mid-Term done to ascertain pre-project situation. Review of August, 1994, however, the contribu- 2.6 The document has concentrated basically on tion of T&V towards livestock activities has not information access/dissemination within the been captured in the document. household and no other stakeholders are in- 2.2 After four joint IDA and GOK monitoring and volved. In addition, the indication that only 47% evaluation Missions. It was concluded that man- of respondent say that information is less avail- agement of NEP II was the main constraint able now than it was before relates to asking towards unification and co-ordination of exten- direct questions or through a questionnaire sion services. Firstly, Management of NEP II which may not be conclusive. Other forms of was through a Project-Working Group and this verification should have been explored. may have had a far reaching effect on the project 2.7 The document also indicates that key findings of implementation. This was a very important the survey are that there has not been apparent conclusion that the evaluators of efficacy of improvement in the quality of delivering of exten- T&V should have considered. Secondly, the sion services. There are no conclusive indication of implementation process followed the procedures the same. The kind of data the evaluators have laid down by the Ministry of Agriculture. Pro- obtained from the questionnaire can not be conclu- curement was an important management tool in sive. Other verifiable indicators are required. NEP I and NEP II. Motor vehicles, motor cycles, 2.8 Extension Services covers a wide range of agro- bicycles and office equipment and furniture were ecological zones. In all places it is not possible to procured. The extent to, which it was realized apply contact farmer and follow-up farmers should evaluated. principles of T&V. So the extension agent has to 2.3 Extension Services provision in NEP II were mainly apply other methods to deliver information to the confined to monthly workshops, monthly training farmers. In ASAL areas they can use extension sessions, staff seminars and courses, staff/farmer barazas but it should be noted that these are not educational tours, on-farm demonstrations farmers administrative barazas. Also in administrative field days, agricultural shows, farmers courses and barazas, extension agents are invited to address seminars, group visits and meetings. These were the public but this should not be construed to extension methods used by extension agents to mean it is an extension barazas. create awareness and encourage adoption of tech- 2.9 One year of suspension is too short to notice any nologies by farmers. However, one page 6 of the significant change in delivery of extension ser- document in table 1, it has been observed that T&V vices as reflected in the document. There were had its strengths and weaknesses that the evalua- other parallel extension related projects, which tors should have considered to be able to arrive at a may have supported extension services in 1996, balanced evaluation report about the efficacy of when NEP 1I funds were suspended. The Govern- T&V Management of Extension Services. ment may have redirected the recurrent funds 2.4 In the document enumerators have used a struc- from other sources towards extension services tured questionnaire. Unfortunately no sample of during that period of suspension. 54 Annexes 2.10 Regarding the outreach, first contact with farm- confounding variable like attitudes and past ers has been used to analyze the effectiveness of experiences of farmers that will very much an Extension Services. This measure does not say influence farmers probability of becoming aware much about the efficacy of T&V. It leaves a lot of technology or information and adopting it. of room for criticism because the baseline data This section is an academic exercise and does on pre- T&V situation were not available. This not explain actual reality. is may therefore be treated as a proxy output 2.14 The Ministry of Agriculture through the Division indicator. Extension Service also recognized the constraints 2.11 The document has largely covered crops at the of non-adoption due to irrelevant technology. A expense of livestock which is often a major linkage was established between extension and enterprise in some farms, therefore it may not KARI with the mandate to address the problem of represent a holistic picture of the Extension inappropriate technology. Indeed some NEP I Services. funds were allocated to KARI Regional Research 2.12 In T&V extension approach, it is assumed that Centers (RRC) for linkage activities. The results farmers learn about innovation/technology by may not have been captured during the survey. becoming aware of it, become interested in it and 4.15 The evaluators conducted T&V systems as stated adopt it. Learning occurs by adopting ideas that in theoretical books and did not consider the come from outside. The evaluation has not modification that this approach had undergone, considered the possibility that farmers are active the complex farming system and socioeconomic problem solvers on their own and that they are circumstances of smallholders. Therefore the not passive consumers of technology, but part of evaluation is academic. its development born out of long experiences in 4.16 Despite the increased farmer/extension contact farming. The term "message" here could acquire and significant increase in level of technology a more pluralistic meaning if the term "informa- awareness as observed in the documents, the tion" was used instead of message. evaluators have not considered that adoption 2.13 The use of statistical analysis to be able to rate for these new technologies was also lim- explain process such as awareness and adoption ited to a greater extent by low returns from is not plausible. It is because there are other farm produce. 55  Annexes ANNEX H. OED'S RESPONSE TO BORROWER COMMENTS This report summarizes the main findings of the OED Para. 2.3. OED assesses efficacy by comparing a evaluation, detailed descriptions of the analysis are project's outcomes with its goals. The findings in included in the six supporting working papers. The Chapters 3 and 4 show that project design was not evaluation is based on evidence obtained through a efficacious, for reasons outlined in table 2.1 and noted broad-ranging household survey and on the information in the conclusions section of,Chapter 2. and data available from two earlier surveys, in 1990 and Para. 2.4. The questionnaire was deliberately 1992. The evaluation also surveyed frontline extension omitted from the final report for brevity, but is staff, subject matter specialists, and supervisors in the available on request. A copy was sent to the Ministry of study districts. As noted in the report, these data were Agriculture at the time the survey was conducted. combined with secondary sources of information ob- Para. 2.5. The comparison with the situation 10 to tained from various reports and documents, including 15 years earlier was a subjective assessment by the users several reports of the Ministry of Agriculture (MOA). of the extension services, the farmers, of the change in The historical evidence was collated from studies reach- their access to extension services and their quality. OED ing back to the early 1970s (see references in the report took these beneficiaries' perceptions at face value, but and the Working Papers). In addition, the evaluation their views were not the only basis for the analysis. In the used the government of Kenya's participatory poverty absence of appropriate baseline data on various aspects assessments (as noted in Chapter 3) and an independent, of the extension services, OED used recall data where beneficiary assessment conducted by an NGO in full feasible. Formal statistical tests in Working Paper 1 cooperation with the MOA. Responses to specific com- compare recall estimates with estimates from three ments from the government of Kenya follow. (All independent surveys for different time periods between paragraph references refer to the original draft.) 1982 and 1993 on contacts made with the extension Para. 2.1. The household and staff surveys and the system. The tests demonstrate that recall bias is unlikely beneficiary assessments covered all extension activities, to be significant. Further comparisons in Working Paper including livestock production, animal health, soil 2 on the awareness and adoption of maize practices also conservation, and the like. Thus, the analysis in Chap- suggest the same. More important, comparable findings ters 2-4 is not restricted to crop production. In other from a number of studies noted in Working Paper 1, chapters, dealing more specifically with physical or some from as early as 1972, provide a significant degree technical impact, the analysis is restricted to crop of confidence in the broad results emerging from the activities. Accordingly, the text in Chapters 5-7 has been OED survey. As for the technical impact on farmers' modified and a footnote added to clarify this. The efficiency and productivity, the 1982 RHBS provided an primary reason for focusing on crop-related activities appropriate baseline for NEP I, the 1990 ATD survey for was that livestock production was included in NEP II NEP 11. only after 1994. For most of the period under study, Para. 2.6. As noted above, the evaluation included livestock extension activities were financed and man- a survey of extension staff. Most staff reported a aged under a different system. Also, a preliminary decline in the system's effectiveness after 1982 and analysis of the awareness and adoption of livestock 1990, as noted in Chapter 2. Table 3.1 summarizes an production practices showed that almost all farmers had alternative assessment by the beneficiaries that queried heard of and adopted those practices by 1994. A deeper 19 services in a contextually independent manner. analysis of livestock production activities will be under- There is no compelling reason to believe that the taken and communicated separately, although the asso- responses for extension services are biased. In this ciation with NEP II is likely to be tenuous. Further, as context, it should be noted that the evaluation strategy discussed in Annex C, the limited experience of the included beneficiary assessments. It is unfortunate that livestock extension staff with T&V has not been positive. the beneficiary assessments were not carried out. It was Para. 2.2. OED considers procurement to be a agreed that beneficiary assessments would be con- function of project management, not a management ducted by the MOA with the full support and coopera- tool. The text in Chapter 2 and Annex C has been tion of OED (including financial contribution) and the suitably modified. Bank's Kenya Country Department. Howevei after 57 Agricultural Extension: The Kenya Experience initial preparations, the ministry unilaterally decided Para. 2.14. The report takes note in Chapter 2 that not to implement the beneficiary assessment. staff training and improved links between research and Para. 2.7. The qualitative aspect of the delivery of extension are benefits of the projects. As inputs into the services refers to the projects' design, and the text has extension system, the results of these activities would been modified to clarify this point. The projects' design naturally affect the flow and content of the extension anticipated high-intensity and qualitatively different activities in the field. These would thus be fully meetings between extension agents and the contact reflected in the impact at the farm level in greater groups. As discussed in Chapters 3 and 4, the outcome awareness and adoption of practices and improved for this aspect is unsatisfactory. productivity. Para. 2.8. OED agrees with the assessment, and Para. 2.15. Farmer characteristics are included in the report is consistent in noting that a uniform method all aspects of the evaluation. The analysis is, for the should not be used in all locations, as advocated in the most part, general and encompasses alternative meth- T&V design. The report has been modified to reflect ods, as long as they were carried out by government the correct interpretation of baraza. extension agents. However, as an evaluation of NEP I Para. 2.9. The impact of the suspension of funds is and NEP 11, the focus had to be on the projects' discussed in further detail in Working Paper 1. Consider- objectives and design as contained in their Staff ing the prescribed schedule of meetings between exten- Appraisal Reports and as presented in the various sion agents and contact farmers and contact groups MOA documents noted earlier. Therefore, the evalua- envisaged by T&V, the suspension would be noticeable if tion paid particular attention to the specific aspects of the system were working as anticipated. While it is the extension methodology, but the farmers' quantita- possible that other funds could have been used, this did tive and qualitative assessments were generally inde- not happen. None of the officials met during the pendent of any particular method. evaluation indicated that alternative funds were made Para. 2.16. The results show increases in aware- available. In any event, if other funds had been used, it ness and adoption for only a few simple agronomic would have indicated that the Bank's funds were not messages. As noted in the relevant sections of Chapter needed, raising the issue of the additionality of NEP II 5, these are also known to have been relatively high funds. even at the start of NEP 1, and especially during NEP 11. Para. 2.10. OED agrees with the comment, Work- The levels for the more complex messages are still very ing Paper 1 makes the same point, and the report has low. It is true that low returns, along with other been modified to further clarify this point. The analy- constraints such as the availability of credit or inputs, sis, however, does go beyond first contact and covers are important factors. However a negligible propor- more meaningful aspects of meetings between exten- tion of farmers indicated low returns as the reason for sion agents and farmers. not adopting agents' recommendations (see Working Para. 2.11. The response is the same as that for Paper 2). A number of farmers did note financial paragraph 2.1. constraints, and some noted input market problems. Para. 2.12 OED fully agrees with the premise of Also, almost 80 percent of those who were aware of the the comment, which is the basis for judging the recommendations (including the more complex ones) incremental impact attributable to NEP I and NEP II. It adopted them. This clearly shows that while, other should be reiterated that this is an evaluation of the concerns may be valid, lack of information is a serious impact of NEP I and NEP II; both relied on the time- constraint. Finally, according to the 1998 Economic bound delivery of messages. Survey published by the Government of Kenya, the Para. 2.13. The analysis of awareness and adoption in agricultural output price index for 1997 was 598.9 Working Paper 2 includes a number of social and (1982=100), the total input price index was 520, and agroecological factors in addition to extension. The statisti- the fertilizer price index was 314.3. Thus, it is unclear cal results are consistent with the descriptive findings and whether adverse market conditions were behind the the focus of NEP I and NEP 11 activities, as noted by several failure of farmers to adopt the recommended practices. senior extension staff and other observers. 58 Annexes ANNEX 1. HOUSEHOLD SURVEY DATA OED conducted a comprehensive household survey Another 62 of the households belonged to descendants in 1997 to collect data on a wide range of topics.' The of the original respondents, and these were retained in survey was designed to generate panel data by revisit- the sample. In addition, 241 new households were ing as many households as could be relocated from a randomly picked from the same clusters as the missing 1990 household survey conducted by the Africa Techni- households, bringing the total sample size for 1997 to cal Department (ATD), which drew from a subsample 596 households.4 The purpose of selecting replace- of the 1982 Rural Household Budget Survey (RHBS).2 ment households was to develop a sample of reason- As may be expected, the number of households com- able size, but at the same time to maintain the mon to all three surveys is considerably smaller than locational and socioeconomic characteristics of the the total observations in each year. The intervening sample. In all, the OED survey collected data from 73 periods are relatively long and it is not surprising that a clusters spread over 12 (current) districts. number of households could not be traced (because of The OED survey was structured to begin with deaths, migration, and the like). contextually general questions-that is, questions that The 1982 survey was a stratified random sample were not specific to any particular agency or organiza- representative of about 95 percent of the national tion. The enumerators were specifically instructed to population, and drawn from the existing national record unprompted responses whenever feasible.' sampling frame (NASSEP I).1 The sample represented Farmers were prompted only when it was necessary to 2.8 percent of the rural population, drawn from 640 test for specific issues or to pursue certain preconceived clusters from all but the North-Eastern province notions based on past experiences within and outside (covering about 54 percent of the land area). The 1990 Kenya. Special care was taken to phrase and sequence survey randomly picked clusters in seven of the the questions to ensure that there were no leading districts, representing a broad cross-section of Kenyan questions. Accordingly, it was important to maintain agriculture, but excluded districts in which NEP I had the identity of the surveyors as far removed from the not been implemented. From the list of households government or the extension service. surveyed in 1982 in these districts, about 700 house- To dissociate the survey from the extension ser- holds were randomly picked for the ATD survey (all vice, the government, or the World Bank, the OED households with household identifiers ending in 7). survey was conducted by the Tegemeo Institute of The 1997 survey targeted the same subsample as the Egerton University. The enumerators were college 1990 survey, using the original list of respondents graduates, selected for their familiarity with agricul- from the 1982 survey. tural issues and fluency in the local dialects of the For this analysis, only the clusters included in the survey sites. The enumerators were trained in the subsequent surveys are retained from the RHBS data. survey methodology, particularly for questions relat- This yields about 611 observations for which agricul- ing to the extension and contingent valuation mod- tural input and output information is available. The ules. The survey instrument was pre-tested to fine-tune 1990 survey targeted about 700 households, but the nuances of the questions and language. The survey complete data for production analysis are available was implemented in two rounds, one at the start of the for about 450 households. The 1997 survey targeted 1997 long rains season (August/September 1997) and households belonging to the same clusters retained for the second after harvest (November/December 1997). the 1990 ATD survey, but clusters with fewer than four In the second round, 34 of the respondents could not target households were dropped. Another 11 clusters be contacted again. could not be traced, following the many changes in The type of data available from the three surveys administrative boundaries that have occurred since varies. But although the surveys are not fully compa- the original sample selection. Of these, two were rable, they are sufficiently so to allow some meaningful discarded altogether; for the other nine, replacement analysis. The 1997 data is the most detailed, by design. households were randomly selected in the same In the planning stages, the objective of the OLD survey location as the original clusters. Overall, 293 of the was to be able to allow a comparison with the 1990 original respondent households could be contacted. data. Howeves the questionnaire was expanded to 59 Agricultural Extension: The Kenya Experience collect additional information on various aspects of the production process. The details of the differences among the surveys are discussed in the Working Papers as needed for analysis or comparisons. The OED survey questionnaire is available from the author on request. 60 ENDNOTES Chapter 1 goods or services from the suppliers' perspective. Thus, the 1. The ATD Kenyan study is one of the three notable subtractability and rivalry of benefits from goods and services attempts to rigorously estimate the returns to T&V investments, provide guidelines for the optimal delivery mechanism. But even as described in Annex A. But as discussed in Chapter 7, the for a public good, viewing benefits from the demand side helps estimates of high returns are not robust. conceptualize efficiency gains from cost recovery, even if cost 2. A public expenditure review for agriculture revealed that recovery is only partial. in 1996-97, extension claimed 61 percent of the development resources allocated for MALDM's core services. Extension also Chapter 3 accounts for about 45 percent of MALDM's total expenditure 1. In addition to complementing the quantitative, survey- and more than half of its staff. based findings of this study, the beneficiary assessment was 3. Aggregate crop statistics are notoriously poor in Kenya. intended was to allow a comparison of the quantitative results Maize data are likely to be the most accurate because of the with those from a more qualitative and participatory approach. importance of maize in Kenyan agriculture and the attention it 2. Even though these were "poverty" assessments, a substan- receives from all quarters. Nevertheless, the accuracy of even tial number of people belonged to land classes that are compa- these data cannot be affirmed. rable to the OED survey households. 4. Average growth between 1970 and 1996 is estimated at 3. On average, once a year in low-potential zones, twice a 2.7 percent, reflecting a deterioration of yields in the 1990s. The year in medium-potential zones, and often in higher-potential rainfall-controlled growth rate cannot be calculated from 1990 zones. onward because rainfall data are lacking. The rainfall data used 4. The ranking used a simple tabulation of the reported first here are average annual millimeters of rain from 14 stations from choices. Considering that basic preferences are likely to be given the south, southeast, central, and western parts of Kenya. higher priority, alternative rankings were tried that used the top 5. Most opinions are based on anecdotal evidence from field three and five choices and the assigned ranks as weights for visits. See box A.1 in Annex A on the difficulty of using such aggregating across observations. These results were consistent evidence to draw inferences. with extension retaining the fifth or sixth ranking. 6. These weaknesses, of course, were not specific to Kenya 5. The only exception was electricity. In declining order, the (Feder, Slade, and Lau 1985). rankings were piped water, public health dispensaries, dry season 7. As the perception of the roles and functions of extension road access, electricity, and tarmac roads. has evolved, so have the models used to capture the interactions of research, alternative extension providers and methodologies, Chapter 4 and farmers. The simplistic model here is meant to reflect the 1. It is possible that some noncontact farmers do not know interactions assumed for NEP I and NEP II. that a neighbor is a designated contact farmer, and consequently report their source of information as "friends and neighbors." Chapter 2 However, it is unlikely that in a small community, especially 1. The key reference reports are MALDM 1997b, c, and where group activities are reportedly common, that farmers Kandie 1997. would not observe the regular and frequent visits of an extension 2. It was found later that a miscalculation of the number of agent or other farmers to one particular farm. In either case, the farm families had led to a farm-to-staff ratio much higher than lack of publicity about extension activities in given locations is the 500:1 that the staff appraisal report judged to be adequate. likely to have reduced their potential for impact. 3. Outreach is defined here as any type of contact between 2. These results are consistent with recent findings from farmers and the extension system. As discussed in Chapter 4, beneficiary assessments in several African countries, which show however, this can be a poor measure of effectiveness. It is also that contact farmers are likely to be less well connected with the inconsistent with the role of "contact farmer" or "contact group" rest of the community than hoped (Salmen 1999). that T&V advocates. Nevertheless, for reasons discussed below, 3. This analysis is based on farmers' recall of their first field staff are using alternative methods, particularly barazas, to meeting with extension agents. Comparisons with three indepen- increase their outreach. dent data sets for three different time periods since 1992 show 4. About 26 percent of those who were in service in 1990 that the recall bias is unlikely to be significant. thought that the system was more effective at the time of the 4. As noted earlier, however, the bias against women in the survey; 49 percent thought that it was less effective, and 25 selection of contact farmers appears to have persisted. percent thought that it had remained the same. 5. That is, among the farmers who reported receiving advice 5. The general principles are professionalism, a single line of at least once a year, or about 41 percent of the sample. command, concentration of effort, time-bound work, client orien- 6. In part, the problem is that "useful," a term used often to tation, and regular training. determine the effectiveness of extension services, is ill-defined and 6. Most institutional analyses focus on the characteristics of vague. Farmers are also reluctant to volunteer criticism. 61 Agricultural Extension: The Kenya Experience 7. Supply is defined as the effective supply of information as cumulative effect of extension advice over years is captured by revealed by the farmers' responses about the recommendations using a weighted, lagged structure of extension supply over a that they find most applicable. It is assumed that these recommen- number of years. dations are either those they receive advice on or those that are 12. For spacing, seed rates, and chemical plant protection relevant to their circumstances. measures, the predicted contact farmer variable is significant at 8. Of the contact farmers who reported noticing a change, the 5 percent level; for crop variety and cultural pest control, it is about 23 percent actually reported an increase in extension visits, significant at 10 percent. 37 percent reported fewer visits, and 32 percent reported no visits 13. This new specification, a multiplicative parametric speci- at all. The remainder gave unspecified or other responses. Surpris- fication of the supply of extension over discrete time intervals ingly, the pattern of the contact farmers' responses is almost the between 1982 and 1997, was also tried for the awareness same as that of the noncontact farmers. estimation. The results for awareness with respect to extension were similar to the adoption results. Chapter 5 1. Assessing the actual impact of extension requires measur- Chapter 6 ing the associated increase in farmers' agricultural productivity. 1. Technical efficiency measures physical productivity-that This is dealt with in Chapter 6. is, it relates physical inputs to output. Scale efficiency measures 2. Increased productivity can take the form of diversification the deviation of each farm from the optimal size of operation. into higher-valued crops, an increase in the efficiency of input use, Allocative efficiency measures the deviation from the optimal of a change in the use of productive inputs, or a combination of the the input mix given the current market conditions as reflected by three. the current prices. For 1982, input price data are not available, 3. A farmer's decision to adopt a particular technology is and hence only the technical and scale measures are calculated. influenced by a number of economic and technological factors, 2. Weak significance refers to significance at the 10 percent level. including extension advice (Feder, Just, and Zilberman 1985). 3. Since only 285 observations were used, some tests were done 4. This process is akin to the spread of an epidemic or an to check for potential selection bias for the retained observations (a infectious disease (Feder and Umali 1993). result of inability to contact all the households in the 1997 survey 5. Livestock extension services were included in NEP II only locations). A simple t-test of the 1982 level of relative efficiency after 1994. A separate, preliminary analysis shows that the vast shows that the average efficiency of retained observations is slightly majority of the households that were aware of livestock-related lower than that of the rest of the 1982 sample, and significant at the recommendations were already aware of them by 1994. Attributing 5 percent level. But when agroecological factors are controlled for, in the awareness or adoption of these recommendations to NEP II is a Probit regression, the 1982 difference in efficiency level is no longer thus likely to be tenuous. The analysis of the awareness and adoption significant at the 5 percent level. of livestock recommendations will be conducted separately. 4. A drawback of DEA is its sensitivity to measurement 6. Less than a third of the sample was aware of the errors. To minimize these, the analysis used only observations recommendations on fertilizer and other chemical inputs. with positive outputs, and for all variables, observations in the 7. Several early surveys showed that activities in many of the top and bottom 1 percent of distribution of intensities (output or simpler and maize-related recommendations were already being input per unit of area) were eliminated; the exception to this was performed widely at the start of NEP I (see Gerhart 1975 and observations with zero nonlabor cash inputs. Ongaro 1990). 5. Malmquist indexes are calculated so that scores below 1 8. Since maize has been the focus of government extension represent a positive change or gain, while scores above 1 represent since the mid-1960s, the cumulative impact of its efforts on regress. To make the presentation mote transpatent, the graphs relative levels of awareness for maize practices is to be expected. depict the inverse-that is, scores greater than 1 represent gains in 9. The results could also be interpreted as reflecting the productivity or efficiency. efficient working of the spread effect of the contact farmer 6. These trends are consistent with farmer complaints in the approach. But if this were the case, there should be a corre- high-potential districts, such as Trans Nzoia, that they are sponding increase in the share of friends, neighbors, and family obtaining lower yields with the same or more inputs. as a source of information. The results show a contrary trend: the share of friends, neighbors, and family has steadily declined Chapter 7 for all activities. 1. The working paper has been recently published as World 10. It is reasonable to assume that, with more than 80 Bank Policy Research Working Paper 2098. percent of those who are aware adopting recommendations, if the 2. Because of attrition and incomplete data for some van- remaining two-thirds who are currently unaware of the recom- ables, the number of observations common to the 1982 and 1990 mended practices were provided with appropriate advice, almost data are 306; to the 1990 and 1997 data, 216; to the 1982 and 50 percent of the sample could be potential adopters. 1990 data, 258. 11. These include the supply of extension, measured as the 3. Alternative lag structures and lengths were tried, but the ratio of extension worker to farm families in each location. The qualitative results did not change significantly. 62 Endnotes 4. To the extent that there are any carryover effects from the other projects (a number of which are donor-funded) are not yet previous system, the impact of NEP I's early years is likely to be accounted for. overestimated, but this is not considered to be a major limitation. 3. The data MALDM provided to OED for NEP I are 5. For each observation, the average was taken over all other inaccurate. A cross-check with the appropriation accounts for observations in the cluster to avoid spurious statistical associa- 1994-95 revealed that those data were for expenditures (develop- tion, since household production level enters the dependent ment and recurrent) for the whole ministry, and not just for variable calculation. extension. 6. Note that district dummy variables are included, in 4. The conversion rate is US$1 to KP 2.9 (Ksh 58) for 1997. conjunction with distances to market and roads, as proxy for local 5. The exact total depends on how much of the KP 21.7 prices, which are not available for the 1982 and 1990 data. million that was spent on information management was attribut- able to extension activities. Chapter 8 6. An alternative estimate can be derived from the Staffing 1. The survey followed the professionally accepted guide- Norms Study. Assuming that the short-term staffing norms pro- lines for CVM questioning: interviews were conducted in person; posed for 1998-99 reflected the current staffing levels, personnel the more conservative willingness to pay (rather than willingness costs for the study districts for district-level staff alone were to accept) was elicited; sample and item nonresponse were quite about Ksh 374 per farm family. Adjusting for 15 percent inflation, low; open-ended as well as double-bounded referendum questions assuming that these costs represented 80 percent of the recurrent were asked; the program was well understood by the farmers; and budget, and doubling it to account for the development budget, the farmers were given opportunities to explain their responses yields an estimate of Ksh 794 per farm family for 1998-99. This is (see Arrow and others 1993). close to the estimate from the 1996-97 budget, considering that it 2. Eleven respondents (less than 2 percent of the sample) is an underestimate because it omits staff expenses for all staff mentioned "on-demand" in response to the desired frequency of above the district level. visits. These observations have been dropped for the rest of the 7. Part of the improvement in staff quality can also be analysis. attributed to the rapid increase in staff numbers through the 3. Farmers were also asked the mode of payment they 1980s, which brought in better trained graduates from the preferred, and the majority chose cash. agricultural colleges. Annex A Annex E 1. A larger number of studies have attempted to measure the 1. The questions on awareness and adoption of practices effectiveness of extension in general, including several studies of were administered in the second round of the OED survey, when T&V systems. However, most of the studies suffer from concep- 34 households from the original sample of 596 could not be tual or methodological limitations, as Birkhaeuser, Feder, and contacted again. The following analysis is based on responses Evenson (1991) and Feder and Umali (1993) have noted. from the remaining 562 households. 2. Kenya was the first country in Africa to reform i 2. Targeted recommendations ranged from simple messages national extension service along T&V lines. (on crop varieties, planting time, spacing, seed rate, weeding 3. The Central Bureau of Statistics, an independent agency, time, number of weedings, and the like), to practices of intermedi- collected the data, not MALDM. ate complexity (such as applying types of basal and top-dress fertilizers), to more complex messages (such as knowing the quantity of basal and top-dress fertilizers, time of top-dressing, Annex B and chemical and cultural pest and disease control measures). 1. The pilot focused only on maize and was conducted with 3. The results reported in Bindlish and Evenson (1993) are weekly visits rather than the fortnightly visits prescribed by T&V mostly for maize, with beans as the only second crop for about 25 standards. percent of the 1990 sample. Coverage of crops other than maize 2. Non-FEW staff included the DAO, SMS/AO, DEO, TO, intercrops was negligible. senior account clerk, clerks and enumerators, and drivers. Annex I Annex C 1. The survey included modules on the nature and extent of 1. Such observations were made during individual and group interaction with agricultural extension services and other extension meetings of current and past extension and nonextension staff of activities; detailed input and output data for crop production; the ministry, academics, researchers, and donors. animal health and livestock production data; household demograph- 2. Expenditure estimates for wholly extension projects can ics; farm equipment and agroecological characteristics; awareness be identified, but expenditures on extension as a component of and adoption of extension messages; a contingent valuation module 63 Agricultural Extension: The Kenya Experience to elicit willingness to pay for extension services; infrastructure and other services; credit; and group activities. 2. These three surveys generate a panel data set for approxi- mately 300 households. The surveys cover household demograph- ics, farm characteristics, and input-output data on agricultural production. The 1990 and 1997 surveys also collect information on contact with extension services, including awareness and adoption of extension messages. 3. NASSEP stands for the National Sample Survey and Evaluation Program. 4. The selection criteria for the replacement households was to select the fourth household to the left (facing out at front gate) of the missing household's farm. 5. With subsequent post-coding of the responses. 64 BIBLIOGRAPHY Working Papers of Kenya and Burkina Faso." The World Bank All Working Papers are available on request. Research Observer 12(2): 183-202. __ 1993. Evaluation of the Performance of T&V 1. Gautam, Madhur. 1999. "The Efficacy of the T&V Extension in Kenya. Technical Paper 208, World System of Agricultural Extension in Kenya: Results Bank Africa Technical Department. Washington, from a Household Survey." OED, World Bank, D.C. Washington, D.C. Photocopy. Bindlish, Vishva, R. E. Evenson, and Mathurin 2. Gautam, Madhur. 1999. "Awareness and Adoption Gbetibouo. 1993. Evaluation of T&V Extension in of Extension Messages." OED, World Bank, Wash- Burkina Faso. Technical Paper 226, World Bank ington, D.C. Photocopy. Africa Technical Department. Washington, D.C. 3. Gautam, Madhur, and Jock Anderson. 1999. "Re- Cleaver, Kevin. 1993. A Strategy to Develop Agricul- considering the Evidence on Returns to T&V Exten- ture in Sub-Saharan Africa and a Focus for the World sion in Kenya." OED, World Bank, Washington, Bank. Technical Paper 203, World Bank Africa D.C. Photocopy. (Reprinted as World Bank Policy Technical Department. Washington, D.C. Research Working Paper 2098.) Feder, Gershon, and R. H. Slade. 1986. "The Impact of 4. Gautam, Madhur, and Jonathan Alevy. 1999. Agricultural Extension: The Training and Visit Sys- "Farmers' Efficiency and Productivity Change in ter in India." World Bank Research Observer 1(2): Kenya: An Application of the Data Envelopment 139-61. Analysis." OED, World Bank, Washington, D.C. Feder, Gershon, and Dia Umali. 1993. "The Adoption Photocopy. of Agricultural Innovations: A Review." Technologi- 5. Gautam, Madhur. 1999. "The Impact of Extension cal Forecasting and Social Change 43: 215-39. on Production in Kenya." OED, World Bank, Wash- Feder, Gershon, R. just, and D. Zilberman. 1985. ington, D.C. Photocopy. "Adoption of Agricultural Innovations in Developing 6. Gautam, Madhur, and Amy Gautam. 1999. "The Countries: A Survey." Economic Development and Willingness to Pay for Agricultural Extension Cultural Change 33: 255-98. Services in Kenya: An Application of the Contin- Feder, Gershon, L. J. Lau, and R. H. Slade. 1987. gent Valuation Method." OED, Washington, D.C. "Does Agricultural Extension Pay? The Training and Photocopy. Visit System in Northwest India." American journal of Agricultural Economics 69(3): 677-86. References Feder, Gershon, R. H. Slade, and L. J. Lau. 1985. The Actionaid Kenya. 1997. Kwa Maoni Yetu (From Our Impact of Agricultural Extension: The Training and Perspective). Washington, D.C.: World Bank, Opera- Visit System in Haryana. World Bank Staff Working tions Evaluation Department. Paper 756. Washington, D.C. Arrow, K., R. Solow, P. R. Portney, E. E. Leamer, R. Feder, Gershon, A. Willett, and W Zijp. 1999. "Agri- Radner, and H. Shuman. 1993. "Report of the cultural Extension: Generic Challenges and the Ingre- NOAA Panel on Contingent Valuation." Federal dients for Solutions." World Bank Policy Research Register 58: 4601-14. Working Paper 2129, Rural Development Depart- Benor, D., and M. Baxter. 1984. Training and Visit ment. Washington, D.C. Extension. Washington, D.C.: World Bank. Gautam, Madhur, and Jack R. Anderson. 1999. "Re- Benor, D., J. Q. Harrison, and M. Baxter. 1984. considering the Evidence on Returns to T&V Exten- Agricultural Extension: The Training and Visit Sys- sion in Kenya." World Bank Policy Research Work- tem. Washington, D.C.: World Bank. ing Paper 2098. Washington, D.C. Birkhaeuser, D., R. E. Evenson, and G. Feder. 1991. Gerhart, J. 1975. The Diffusion of Hybrid Maize in "The Economic Impact of Agricultural Extension: A Western Kenya. Abridged by CIMMYT Mexico Review." Economic Development and Cultural City: Centro Internacional de Mejoramiento de Maiz Change: 607-50. y Trigo. Bindlish, Vishva, and R. E. Evenson. 1997. "The Hussain, Syed Sajidin, Derek Byerlee, and Paul W Impact of T&V Extension in Africa: The Experience Heisey. 1994. "Impacts of the Training and Visit 65 Agricultural Extension: The Kenya Experience Extension System on Farmers' Knowledge and Adop- Kenya. World Bank Environment Department Par- tion of Technology: Evidence from Pakistan." Agri- ticipation Series Paper 0341996. Washington, D.C.) cultural Economics 10: 39-47. Ongaro, W 1990. "Modern Maize Technology, Yield Johnson, Charles W, Keith M. Byergo, Patrick Fleuret, Variations and Efficiency Differentials: A Case of Emmy Simmons, and Gary Wasserman. 1980. Kitale Small Farms in Western Kenya." Eastern Africa Maize: The Limits of Success. Project Impact Evalua- Economic Review 6(1): 11-29. tion No. 2, U.S. Agency for International Develop- Picciotto, Robert, and J. R. Anderson. 1997. "Recon- ment. Washington, D.C. sidering Agricultural Extension." World Bank Re- Kandie, E. K. 1997. "Proposals to Improve Effective- search Observer 12(2): 249-S9. ness of the Agricultural Extension Services in Purcell, D. L., and J. R. Anderson. 1997. Agricultural Kenya." Agricultural Extension Policy Project, Min- Extension and Research: Achievements and Prob- istry of Agriculture, Livestock Development, and lems in National Systems. Operations Evaluation Marketing, Nairobi, Kenya, and German Agency for Study. Washington, D.C.: World Bank. Technical Cooperation. Nairobi. Salmen, L. 1999. "The Voice of the Farmer in Agricul- MALDM (Ministry of Agriculture, Livestock Develop- tural Extension: A Review of Beneficiary Assessments ment, and Marketing). 1997a. Alternative Ap- of Agricultural Extension and an Inquiry into Their proaches to Agricultural Extension: ASIP Prepara- Potential as a Management Tool." World Bank AKIS tion Study. Nairobi. Thematic Team. Washington, D.C. Photocopy. _. 1997b. Proposal on the National Agricultural Tiffen, M., and others. 1996. National Soil and Water and Livestock Extension Programme. Nairobi. Conservation Programme, Kenya. SIDA Evaluation 1997c. "Staffing Norms Analysis," Vols. 1-3. 96/2S, Department for Natural Resources and the Agriculture Sector Management Project II. Nairobi. Environment. Stockholm. Mukui, J. T. 1994. "Kenya: Poverty Profiles." Office of Venkatesan, V., and J. Kampen. 1998. Evolution of the Vice-President and Ministry of Planning and Agricultural Services in Sub-Saharan Africa. Discus- National Development. Nairobi. sion Paper 390, World Bank Africa Region Series. Narayan, D., and D. Nyamwaya. 1995. A Participa- Washington, D.C. tory Poverty Assessment Study-Kenya. Report pre- World Bank. 1997. Rural Development: From Vision to pared for the World Bank, sponsored by British Action. Washington, D.C. Overseas Development Association, United King- . 1994. Agricultural Extension: Lessons from dom, and UNICEE (Revised, 1996. Learning from Completed Projects. Operations Evaluation Depart- the Poor: A Participatory Poverty Assessment in ment. Washington, D.C. 66 OPERATIONS EVALUATION DEPARTMENT PUBLICATIONS The Operations Evaluation Department (OED), an Documents listed with a stock number and price independent evaluation unit reporting to the World code may be obtained through the World Bank's mail Bank's Executive Directors, rates the development order service or from its InfoShop in downtown impact and performance of all the Bank's completed Washington, D.C. For information on all other docu- lending operations. Results and recommendations are ments, contact the World Bank InfoShop. reported to the Executive Directors and fed back into For more information about this study or OED's the design and implementation of new policies and other evaluation work, please contact Elizabeth projects. In addition to the individual operations and Campbell-Pag6 or the OED Help Desk. country assistance programs, OED evaluates the Bank's policies and processes. Operations Evaluation Department Summaries of studies and the full text of the Pr9cis Partnerships & Knowledge Programs (OEDPK) and Lessons & Practices can be read on the Internet at E-mail: ecampbellpage@worldbank.org http://www.worldbank.org/html/oed/index.htm E-mail: OED Help Desk@worldbank.org Telephone: (202) 473-4497 How To Order OED Publications Facsimile: (202) 522-3200 Operations evaluation studies, World Bank discussion papers, and all other documents are available from the World Bank Infoshop. Ordering World Bank Publications The World Bank InfoShop serves walk-in customers Customers in the United States and in territories not only. The InfoShop is located at: served by any of the Bank's publication distributors may send publication orders to: 701 18th Street, NW Washington, DC 20433, USA The World Bank P.O. Box 960 All other customers must place their orders through Herndon, VA 20172-0960 their local distributors. Fax: (703) 661-1501 Telephone: (703) 661-1580 Ordering by e-mail The address for the World Bank publication database If you have an established account with the World on the Internet is: http://www.worldbank.org (select Bank, you may transmit your order by electronic mail publications/project info), on the Internet to: books@worldbank.org. Please in- E-mail: pic@worldbank.org dude your account number, billing and shipping Fax number: (202) 522-1500 addresses, the title and order number, quantity, and Telephone number: (202) 458-5454 unit price for each item. 67  OED STUDY SERIES 1999 Annual Review of Development Effectiveness Agricultural Extension and Research: Achievements and Problems in National Systems Bangladesh: Progress Through Partnership Developing Towns and Cities: Lessons from Brazil and the Philippines Financial Sector Reform: A Review of World Bank Assistance Fiscal Management in Adjustment Lending India: The Dairy Revolution Investing in Health: Development Effectiveness in the Health, Nutrition, and Population Sector Mainstreaming Gender in World Bank Lending: An Update Nongovernmental Organizations in World Bank-Supported Projects: A Review Paddy Irrigation and Water Management in Southeast Asia Poland Country Assistance Review: Partnership in a Transition Economy Reforming Agriculture: The World Bank Goes to Market The World Bank's Experience with Post-Conflict Reconstruction Zambia Country Assistance Review: Turning an Economy Around Proceedings Lessons of Fiscal Adjustment Lesson from Urban Transport Evaluation and Development: The Institutional Dimension (Transaction Publishers) Monitoring & Evaluation Capacity Development in Africa Public Sector Performance-The Critical Role of Evaluation Multilingual Editions Assessing Development Effectiveness: Evaluation in the World Bank and the International Finance Corporation Appr9ciation de l'efficacit6 du d9veloppement: L'valuation a la Banque mondiale et a la Socit9 financibre internationale Determinar la eficacia de las actividades de desarrollo: La evaluaci6n en el Banco Mundial y la Corporaci6n Financiera Internacional C6te d'Ivoire : Revue de l'aide de la Banque mondiale au pays Philippines: From Crisis to Opportunity Filipinas: Crisis y oportunidades Rebuilding the Mozambique Economy: Assessment of a Development Partnership Reconstruir a Economia de Mogambique www.worldbank.org/html/oed @ THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. Telephone: (202) 477-1234 Facsimile: (202) 477-6391 Telex: MCI 64145 WORLDBANK MCI 248423 WORLDBANK Internet: www.worldbank.org E-mail: books@worldbank.org 4 1 47 5 8 9 7 8 0 8 2 3 4 7 5 8 4 S N 8~ 3-4 58-6