Communicative farm-specific diagnosis of potential simultaneous savings in costs and natural resource demand of feed on dairy farms

Feed plays a key role for dairy farmers to produce in an environmentally sustainable and competitive way. It determines both costs and natural resource demand. In this paper, we investigated whether and how dairy farms could simultaneously reduce feed costs and overall natural resource use in the feed supply chain without reducing farm revenues. We applied the frontier method Data Envelopment Analysis (DEA) on a data sample of specialized dairy farms in the region of Flanders (Belgium). Results showed potential simultaneous savings in costs and natural resources (up to 48%). This could mainly be achieved by increasing technical efficiency (proportionally minimizing the feed inputs, i.e. (i) on-farm produced roughage feed and (ii) purchased feed, consisting of concentrates and by-products) and to a lesser extent by increasing allocative efficiency (substituting these feed inputs up to a cost and/or natural resource use minimizing allocation). We offered farm advisors starting points to identify concrete improvement actions for individual farms, by graphically presenting improvement paths and by relating DEA's outcomes to Key Performance Indicators they are familiar with. High cost and natural resource efficiencies were related with (i) high milk production per cow obtained with as little as possible purchased feed, and (ii) low on-farm roughage production costs per ha associated with lower contract work costs and a lower proportion of grasslands in the available on-farm area. Finding a good equilibrium of purchased feed amounts and stocking density seemed to play a substantial role in optimizing allocative efficiency. Analysis with different frontier methods showed that the shape of the frontier influences the quantified improvement margins and the diagnosis of win-win and trade-off situations. Further research should focus on (i) the accuracy of the constructed frontier, (ii) relations with emission-related impacts and (iii) possible trade-offs between different resource types.