Abstract/Summary

Conservation Agriculture (CA) is an agro-ecological approach to sustainable production intensification. Low rates of adoption have plagued Sub-Saharan-Africa despite years of promotion. A polarised debate has emerged centred on the farm-level costs/benefits (particularly for the poorest farmers), including when benefits occur, labour requirements (including weeding) and in particular whether CA requires high inputs. The thesis draws on a household survey of 197 farmers in Metuge district (Cabo Delgado, Mozambique) in tandem with participatory stakeholder interviews administered in 2014. Probabilistic cash flow analysis compares CA and conventional cropping for different crop mixes and planning horizons. Secondly, a socio-psychological model explores intention to use CA. A novel Monte-Carlo Markov chain algorithm using socio-psychological factors and conventional determinants of adoption is also incorporated in order to explore adoption dynamics. The thesis finds evidence of benefits for the poorest farmers and in the short-term under CA (without high inputs) but which are dependent on crop mix and opportunity cost of labour assumed. Socio-psychological factors play a strong role in the adoption process; farmers’ attitude is found to be the strongest predictor of intention to use CA mediated through key cognitive drivers such as increased yields, reduction in labour, improvement in soil quality and reduction in weeds (which are precisely the areas of current contention). Interestingly, Farmer Field School participants have a significantly stronger positive attitude towards CA. The employment of the novel Monte-Carlo estimation (as do the stakeholder interviews) also identify Farmer Field School membership, the role of village facilitators in engaging with farmers on CA and willingness to be part of a group play an important role in adoption. Importance of labour reduction, soil quality improvement and perceptions of pests also significantly influence adoption suggesting social learning interactions (taking account of these issues) vis-à-vis an appropriate innovation system are critical to CA usage.