In the short term
  • Heterogeneity: Within AGRICORE each farm will be modelled as an autonomous entity with capacity to interact with their environment and other agents, and to evolve according to its decisions and their context dynamics. This will provide a means to better account for heterogeneity among farms, to enable modelling at a geographic scale ranging from individual farm to regional and to global, and to reduce bias when averaging multiple farms.
  • Structural change: The configuration of the AGRICORE agent-based model will allow the explicit consideration of structural change adjustments as a consequence of agricultural policies. Specifically, the AGRICORE model will explicitly consider the transfer of production factors such as land and the impact of policies on variables like number of active farms, territorial distribution of farms and their evolution over time; farm size distribution and its evolution over time; land rental process; hired and family labour allocation; on-farm and off-farm labour allocation; and farm income disaggregated by farm type/size/regions.
  • Financial viability: The outputs of the AGRICORE agent-based model will include all major financial indicators of the farm business derived from the FADN database which accounts for growth, profitability, liquidity, leverage and activity. Moreover, the dynamic nature of the AGRICORE model will allow considering the impacts of policies on solvency, liquidity and other financial indicators which ultimately affect the structure of production of modelled farms and their survival rates.
  • Dynamic investment: The farmers’ simulation process implemented in AGRICORE will rely on a multiperiod optimisation problem in which the investment demand for land and capital goods (such as machinery, buildings, facilities and equipment) are to be explicitly modelled at various time scales – from the short to long term. This will allow analysing the impacts that policy measures have in the investment decisions of farmers in the long run.
  • Risk and uncertainty: The AGRICORE’s AI-based behavioural rationale module will explicitly integrate uncertainty by considering the mean, variance and skewness of likely future framework conditions (prices, yields and input demands). This will allow not only to model policy tools explicitly targeted to farmers’ risk management (such as income stabilisation schemes as a response to high losses in crop incomes) but also to effectively assess the impact on producers’ attitude toward risk.
  • Interaction between farms: AGRICORE will explicitly consider farm agents not only to interact indirectly by competing on factor and product markets but also directly by sharing information related to expected production outcomes and expected technology adoption. This will allow the model to simulate the interactions that take place when farmers are brought into a dynamic relationship through a set of reciprocal actions and which has effects on their future evolution.
  • Transfer of production factors: The structure of the AGRICORE agent-based model, along with the explicit handling of interconnecting variables, will enable to effectively consider the transfer of production factors between agents. Specifically, the transfer of land will be implemented through the dedicated Land module, and the transfer of other production factors will be implemented through the dedicated Markets module. Moreover, and given that land is essential to most agricultural and livestock production, AGRICORE will exploit today’s availability of computing power and GIS to implement realistic representations of land properties. This way, AGRICORE will generate a 2D mesh of the area of interest using satellite information with dynamically-sized cells attending to the extension of the area of interest and the available computing power.
  • Market inter-linkages: The AGRICORE approach to consider market inter-linkages is twofold. On the one hand, the AGRICORE Markets module will account not only for final products but it will also consider other market feedbacks such as those for labouring, water, manure, fodder and young animals. On the other hand, it will implement market equilibrium model dynamics linked to the aggregated feedback from the demand side, so prices that seem exogenous at farm scale become endogenous when simulating policy impacts at larger geographical scales such as at EU extent.
In the medium to long term
  • Improvement of policy design:
    • Positive purpose: The AGRICORE tool is tailored to be used for a positive purpose, that is, for representing or imitating the real system in response to a given policy as closely as possible.
    • Normative purpose: As an alternative, the AGRICORE tool can also be configured to account for a normative purpose, that is, to automatically choose the best parameters defining a policy scheme in terms of a certain objective function (for instance, to automatically choose the optimal value of a price support policy to maximise the survival of SMEs under a given market context). To do so, the “positive” configuration of the agent-based simulation module will be completed by an additional optimisation layer accounting for the policy instrument optimisation.
    • Ex-ante analysis: AGRICORE will allow to run ex-ante evaluations of different policy schemes with heterogeneous populations of farmers. To do so, the user (for instance, a policy maker) would define the population of interest along with the specific policy to be evaluated and a corresponding context. Then, the AGRICORE tool would simulate the evolution of the farmers composing the population as the response to the given policy. This, along with the corresponding impact assessments, will allow the user to have an informed decision on which policy schemes are best to be promoted.
  • Monitoring:
    • Ex-post analysis: Additionally, the AGRICORE tool can be used for the monitoring of a given policy during the time period in which the policy is implemented. Indeed, the receding horizon strategy executed in AGRICORE allows to, every time step, integrate real feedback of the policy performance. Such feedback would be incorporated in the first stage of the optimisation process implemented by AGRICORE, so the state of the population is updated accordingly. On the one hand, this will allow comparing the expected status of the population with its actual status (along with the expected and actual impacts). On the other hand, this will allow refining the assessment of the impacts of the policy in subsequent time steps considering the real feedback collected in the current time step (for instance, in a mid-term policy evaluation).
  • Impact assessment:
    • Environmental and climate: AGRICORE addresses the environmental and climatic impact assessment of policies by means of a dedicated module aimed to establish links between targeted policies and the corresponding impact KPIs on farmers’ practice. Additionally, this dedicated module will provide regional climatic patterns to the agent-based models as a means to take account of the bi-directional relationship between farming and environment.
    • Socioeconomic: AGRICORE will include a dedicated module for assessing policy outcomes on the delivery of ecosystem services. To do so, it will establish a connection between the response of agents to a given policy (for instance, a scheme aimed to improve the resilience and environmental value of a certain environment) and the related ecosystem service KPIs. In this aspect, AGRICORE will consider four main areas of ecosystem services: supporting, provisioning, regulating and cultural services.
    • Delivery of ecosystem services: The AGRICORE agent-based modelling approach, in connection with a dedicated module, will provide a powerful means to assessing the relationships between policy initiatives and the integration of agriculture in rural systems. Such relationships are to take account for regional heterogeneity on specific KPIs related to the integration of agriculture in rural systems.