The role of social networks in the adoption of competing new technologies in Ghana

We use a detailed dataset to examine the impact of social networks, conditional on contextual and individual confounders, on farmers' adoption of competing improved soybean varieties in Ghana. Based on the contagion conceptual framework, we employ a spatial autoregressive multinomial probit model to examine how neighbours' varietal and cross-varietal adoption of improved varieties affect a farmer's adoption decision in the social network. Our results show that adoption decisions in a network tend to converge on one variety, such that beyond a threshold of adopting neighbours of that improved variety, the cross-varietal effects tend to lose significance in the network. If the shares of adopting neighbours of the improved varieties are equal, we find evidence that farmers are not more likely to adopt either improved variety compared to farmers with no neighbours who have adopted the improved varieties. The findings demonstrate the significance of neighbourhood effects in the adoption of competing technologies.     

Utilising farm-level panel data to estimate climate change impacts and adaptation potentials

We combine farm accounting data with high-resolution meteorological data, and climate scenarios to estimate climate change impacts and adaptation potentials at the farm level. To do so, we adapt the seminal model of Moore and Lobell (2014) who applied panel data econometrics to data aggregated from the farm to the regional (subnational) level. We discuss and empirically investigate the advantages and challenges of applying such models to farm-level data, including issues of endogeneity of explanatory variables, heterogeneity of farm responses to weather shocks, measurement errors in meteorological variables, and aggregation bias. Empirical investigations into these issues reveal that endogeneity due to measurement errors in temperature and precipitation variables, as well as heterogeneous responses of farms toward climate change may be problematic. Moreover, depending on how data are aggregated, results differ substantially compared to farm-level analysis. Based on data from Austria and two climate scenarios (Effective Measures and High Emission) for 2040, we estimate that the profits of farms will decline, on average, by 4.4% (Effective Measures) and 10% (High Emission). Adaptation options help to considerably ameliorate the adverse situation under both scenarios. Our results reinforce the need for mitigation and adaptation to climate change.