The productivity-environment nexus in space. Granularity bias, aggregation issues and spatial dependence within Italian farm-level data

This paper looks for empirical support to the existence of a positive nexus between economic and environmental performance in farming as implied by the Sustainable Intensification hypothesis. As the ecological scale at which this nexus actually occurs is unobservable, the paper juxtaposes its estimation at three spatial scales, the farm level and two regional levels. Starting with a common theoretical background, the paper estimates a dynamic spatial panel model on these alternative scales. Identification issues (granularity, aggregation bias and spatial dependence) may generate significantly discrepant estimates eventually questioning the reliability of these results. The empirical study investigates the relationship between total factor productivity, greenhouse gas emissions and crop diversity using a 2008-2018 panel of Italian farms. Results show that the productivity environment nexus changes and may even revert its sign when passing from farm-level to aggregate data. The implications of these results for evidence-based policy making are discussed.