The slow adoption rates for scientific practices in rural settings, such as precision agriculture in developed regions, are a source of frustration for scientists and technology advocates. When a new technology or farming... [ view full abstract ]
The slow adoption rates for scientific practices in rural settings, such as precision agriculture in developed regions, are a source of frustration for scientists and technology advocates. When a new technology or farming practice offers genuine benefit to farmers, slow rates of adoption cause an acute loss of benefit to individual producers and the wider industry. A great deal of attention across a range of disciplines is now focused on understanding what prevents the adoption of new technology among farmers in rural areas. We offer a new perspective to this field of research.
Individual choice and behavioural expectations are fundamental in understanding broader rural societal motivations. Limitations to fully rational decisions and the role of behavioural factors in individual decision making have motivated a shift in assessing societal behavior from a representative agent with rational expectations to boundedly rational agents with heterogeneous expectations. This shift reflects the growing evidence on the theoretical limitations and empirical challenges in the traditional view of homogeneity and perfect rationality of the ‘economic agent.’
In this analysis, we use a spatial heterogeneous agent programming model (HAM) to examine the range of policy options available for assessing the diffusion of innovations in rural settings governed by typical resource constraints. Unlike conventional simulation tools used in rural sociology, HAM applies a multi-agent cellular automata (CA) approach by using heterogeneous farm-household models and capturing their social and spatial interactions explicitly. The individual choice of the farm-household among available production, consumption and investment alternatives is represented using novel HAM methods and tested against recursive linear programming models for comparison. We model the adoption decision-making process to better understand why some innovations, even if apparently highly profitable, may not be adopted.
Adoption constraints are introduced in form of network-threshold values that reflect the cumulative effects of experience and observation of peers' experiences. The model's economic and resource components are tightly connected within a spatial framework, reflecting the experience of highly productive rural areas. The integration of economic and resource-use processes facilitates the consideration of feedback effects in the use of resources, while demonstrating how resource limits can be overcome through precision agriculture techniques. The model is simulated using empirical data from various data sources in agricultural regions in Australia, Canada and the US. We find that the adoption decision consumes two highly-valued but limited resources: time and the capacity to integrate new information. Readily available quality information with high reliability and relevance to decision-maker reduces these information seeking and learning costs. We show that information quality and effectiveness need to be greatly increased to achieve more responsive adoption decisions. We then consider the rise of formal ‘farmer groups’ as part of regional agronomic research and precision agriculture extension networks in this context. While engagement between agents is confined to relatively traditional modes of communication, their value is grossly underestimated. There is a very high value placed on localised information. Opportunities to achieve more responsive adoption by reducing information and learning-related costs need to be exploited to improve rural technology diffusion.
