Are agricultural land use patterns influenced by farmer imitation?

This paper proposes a methodology based solely on spatial data to analyse whether and, to what extent, farmer imitation leaves an observable footprint on an agricultural landscape. Geographical Information System (GIS) analysis of parcel and farm location data of a study region in central Belgium was developed as an alternative methodology to farmer interviews. Results suggest that imitation is not an important determinant of agricultural land use patterns in the study area. The effect of imitation on landscapes is limited to the extent of being hardly significant. Neighbouring parcels cultivated by farmers who live in close proximity are only slightly more similar than neighbouring parcels cultivated by farmers who live further away from one another. The results question the validity of the assumptions underlying agent-based models that try to explain agricultural land use through imitation behaviour. The results should, however, be considered with caution as the proposed methodology has two limitations. First, comparison between neighbouring parcels could not identify the imitation effect from all factors that influence agricultural land use. Relative space was not accounted for, which led to two possible explanations for the similarity of neighbouring parcels: imitation or the location of a parcel relative to the farm. Secondly, the method was applied to aggregated land use classes for a single year, which did not allow for the effect of crop rotations in understanding imitation behaviour.