Technical efficiency and fertilizers use in Italian farms using a machine learning approach

Abstract

The Common Agricultural Policy has supported a less intense use of fertilizers and chemicals in agriculture in the next five years (2023 – 2027) because of the European Green Deal proposals that are environmental protection-oriented. The most common consequence of input reduction in farms was a direct effect to the technical efficiency in farm. The main purpose of this research was to assess the technical efficiency in a sample of Italian farm part of the Farm Accountancy Data Network (FADN) dataset using the Data Envelopment Analysis (DEA) input-oriented approach and by the machine learning approach such as the iterative decision tree evaluating which quantity of chemical fertilizer in terms of nitrogen (N), phosphorus (P) and potassium (K) has to be reduced to improve the technical efficiency. Results have pointed out that between a reduction of chemical fertilizers and technical efficiency there is a fundamental link and the drop in chemical fertilizers has impacted the technical efficiency in Italian farms part of FADN dataset. Based on these results emerged the need of putting into practice some actions towards farmers to compensate the reduction in technical efficiency and the produced output in the productive process as well.