Seeing beyond fertiliser trees : a case study of a community based participatory approach to agroforestry research and development in western Kenya

Key words: village committee approach, agroforestry, improved tree fallows, biomass transfer, realist evaluation, soil fertility, adoption, dissemination. The thesis explores and describes various processes that take place in the implementation of a community based participatory initiative known as the village committee approach by a collaborative agroforestry programme between the Kenya Forestry Research Institute (KEFRI), Kenya Agricultural Research Institute (KARI) and the World Agroforestry Centre (ICRAF) which has since 1988 been undertaking soil fertility research to address the problem of nutrient deficiency in smallholder farms within the western Kenya highlands. Over the years, various agroforestry technologies have been developed to address the problem of soil fertility. Issues that this thesis explores in detail are the processes of participatory learning, adoption/adaptation/non-adoption, dissemination and diffusion of the technologies. Overall, the thesis is guided by the technographic approach which makes use of diverse observational and analytical methods and frameworks to arrive at hypotheses about likely mechanisms affecting the operation, transformation or adoption of technological processes. One such framework adapted to the needs of this thesis is the context-mechanism-outcomes configuration (CMOC). This framework rests upon realist assumptions. This study drew upon the qualitative methods used by ethnographers. But some issues to do with learning and adoption were assessed from the perspective of a sampling approach. Attention was paid to the integration of quantitative and qualitative approaches. Multiple sources of data were used, including formal and informal surveys involving structured/semi-structured/unstructured interviews with farmers, in-depth interviews with key informants, case studies, participant observation and secondary data. Findings presented show that the use of the village committee approach was misplaced as the approach assumed that groups are fully appropriate vehicles for technology development and dissemination. The groups did not play a major role in agroforestry dissemination, as was hoped by the programme. This may partly be attributed to the fact that agroforesty as a technology was not high on the agenda of most groups and therefore farmers did not give it much thought. In addition, group formation and success depended on being able to exclude some of the most needy persons through imposing membership requirements, such as fees. The thesis shows that groups work, but only for those who have assets to begin with. This suggests the possibility that wealthier farmers benefit from cooperation only when they can exclude poor resource farmers. Second, agroforestry is apparently treated as a kind of ritual requirement helping groups access assets that really make sense – namely livestock distribution through the pass-on system. The possibility must be faced that agroforestry in western Kenya is valued more as a networking opportunity than as a mechanisms for transforming land management. In short, the context was not thoroughly understood, and unanticipated mechanisms (associated with village power politics) kicked into play, resulting in outcomes that diverged from those intended by the agroforesters. As regards to participatory learning this thesis shows that achieving genuine participation has remained elusive. Some people were virtually excluded from the learning process. Exclusion was either by choice (self exclusion) or a product of village power politics. Part of the reason lies in the fact that despite the shift from top down to bottom up in development circles, community structures have remained paternalistic, with a few (better educated, better connected) elites (often older farmers retired from urban employment) controlling development initiatives. This is a major obstacle to participation, and unless it is tackled, efforts being made to involve marginalized members of society through up-scaling of development initiatives will have disappointing results. As regards to farmer to farmer dissemination, findings show that informal social networks were more effective for seed dissemination than for knowledge sharing. This calls for simplification of technical information by development professionals, in order to help support farmers’ understanding and communication of complex principles. In relation to agroforestry adoption, the results show that the process of adoption is highly variable and dynamic, with farmers taking up or discontinuing the use of soil fertility management technologies due to a whole range of factors of which soil fertility management is just one. Mechanisms of adoption are complex, and switched on and off by contextual factors. For this reason adoption research needs to probe beyond categorization and correlation, and frame its analytical questions in terms of the context-mechanism-outcome configurations advocated as the basis of a realistic evaluation methodology. Adoption and diffusion of these technologies has been disappointly very low. This sends a strong message to researchers; for agroforestry-based soil fertility replenishing technologies to be attractive to farmers, they must provide other tangible economic benefits besides soil fertility improvement. Secondly, the marginal superiority of a complex technology is not good enough, it must either be so superior as to sell itself, or it must be sufficiently clear in terms of how the basic mechanism works in local context that pioneer farmers could indeed disseminate it to other farmers if they wished. But this raises the question of whether the social mechanisms of community participation do indeed work as agroforesters had hoped. Spontaneous spread of agroforestry soil-improving agroforestry innovations remains a goal in western Kenya, but basic effort is still needed to specify a plausible scenario linking context (exteme poverty), mechanisms (technical knowledge) and favourable outcome (adoption of agroforestry innovation). What this thesis shows is that if the context is not conducive for the technology, diffusion and adoption simply can not take place. All in all the thesis concludes in the final chapter, that soil fertility management is a function of socio-economic processes within a community, and it is therefore imperative that researchers develop a realist awareness of the contexts, mechanisms and outcomes governing participatory technology development so that there is rapid correction when evaluation reveals evidence of negative or perverse outcomes. By establishing good feedback to R&D, there will be a better chance of avoiding a situation in which a lot of time and resource is wasted on promoting technologies that are ‘not good enough’ in the eyes of farmers.