Determinants of yield impact and adoption of conservation agriculture among smallholder farmers in Zimbabwe

Abstract

The thesis assesses the yield advantage and adoption dynamics of conservation agriculture (CA) as a sustainable farming method that was introduced in Zimbabwe to address the problems of low productivity and declining soil quality. This study is based on five-year panel survey that was intended to monitor the impacts of CA on adopters. The study focused particularly on basin CA, which involves digging small pits with hand hoes during the off-season. This technology allows for early planting and the concentration of soil nutrients within the planting basin in order to reduce the risk of crop failure. Specifically, this study attempts to: a) Provide evidence that shows that CA adoption has a positive impact on maize yield; b) Determine factors that condition farmers to apply more components of the CA package ; and c) Answer the question why some farmers are abandoning CA, which they had adopted earlier. The first part of the thesis used plot level data to model a single equation yield function where CA was assumed to have an intercept effect. Through a household fixed effect model, the impact on yield was measured and verified through ordinary least squares. The evaluation showed that the input with the greatest impact on yield was nitrogen fertiliser. The unambiguous finding of this analysis is the positive significant impact of CA technology on maize yield. The second part of the thesis examined the determinants of adoption intensity using count regression models, specifically Poisson and negative binomial regression. The evaluation showed that more intense users of CA had higher productivity, lived in areas with higher production potential and received some form of input support from non-governmental organisations. There is a general tendency towards dis-adoption as farmers reduce the number of CA practices applied with time. However, the number of techniques applied in the current season increases albeit at a diminishing rate. This implies that CA is becoming more intensively practised in a relatively endogenous manner. However, unless conditions that make the practice easier to apply, CA cannot be expected to be maintained in Zimbabwe. Finally the thesis applied a random effects logit model to measure abandonment of CA. Study findings suggest that poor vulnerable households are more likely to persist with CA confirming that CA is accessible to the poor who are the target group for this technology. Loss of input support through programmes has contributed to dis-adoption but it is not clear whether commercial fertiliser has been available in the absence of NGO programmes. In addition, there is a strong tendency toward dis-adoption in semi-arid and arid regions, raising the question about the suitability of CA in those regions. The study finds results that appear to be at odds with each other: that the practice of CA leads to significantly higher yields of the most important crop, yet there is evidence of farmers discontinuing the practice. There is therefore need to explore the factors that constrain adoption and encourage abandonment in order to understand whether the future of sustainable agriculture in Africa lies in CA.