From September 2018 to January 2024 MIND STEP has developed tools and models at different scales to do exactly this: monitor and assess policies related to agriculture and take aspect of individual decision-making into account. In describing and modelling the complexities and ever-evolving landscape of agriculture the MIND STEP project has contributed to innovations and progress in:
- Better representation of the diversity of farms heterogeneity in modeling
- Modelling interactions between farms
- Improved interfaces between data and models at different scales (farm, regional, national, EU)
- Transparency of methods, sustainable software development and model validation
MIND STEP has developed detailed bio-economic, farm level, mathematical programming and econometric optimisation and simulation models. These models are firstly based on individual farm data from the EU Farm Accountancy Data Network (FADN). Econometric activity-based cost accounting tools have been developed to assign different cost components in the EU FADN from the farm to the agricultural activity level. In a second stage, MIND STEP has developed tools to combine farm level data (FADN) with biophysical data. Using biophysical data, MIND STEP among other developed probabilities regarding spatial allocation of representative farms in the EU FADN and improved grassland yield response curves using remote sensing data. Surveys have been conducted to combine statistical data with socio-psychological data helping to understand individual farmers preferences, behavior and adoption of risks. In particular the concept of modularity for the design of a core bio-economic farm model has received substantial attention in MIND STEP as it is crucial for the combined use of rather complex analytical tools at farm level in different contexts. The addressed challenges in this area were, among others, to define basic functionalities of a core bio-economic farm model (BEFM), to facilitate and regulate the communication of the involved researchers, and establishing standards for computer codes and interfaces between models in the MIND STEP model toolbox. Besides individual farm models, MIND STEP also developed innovative Agent Based Models (ABMs) including interaction between individual farms e.g. used to explain farm exit and adoption of Agri-Environmental Schemes. The development of machine-learning-based surrogate models, allowed efficient and consistent integration of detailed farm models in an ABM, capturing structural change implications of policies for the farm population. This is an important step towards upscaling of agent-based models to larger regions that is normally restricted by computation time. MIND STEP made important contributions to improved micro-economic underpinnings of models at various scales, frequently used by the European Commission for assessments of policies related to agriculture. As a roadmap, MIND STEP gained important experiences in the field of a) validation and proof of concept, b) importance of stakeholder workshops and c) policy evaluation.
