This paper proposes a stylized model that combines a Ricardian framework of land heterogeneity with the ecological concept of a density–yield curve introduced by Green et al. (2005). Biodiversity responds to agricultural choices through two channels: an intensity channel, governed by the curvature of ecological damages, and a reallocation channel, as effort is redistributed across land qualities. I characterize how biodiversity concerns reshape the efficient allocation of land exploitation as a function of damage curvature and food-demand elasticity. In a Ricardian landscape, heterogeneity creates a baseline incentive to concentrate production on high-quality land, so that land-sparing adjustments arise for concave and for moderately curved damages. Land sharing emerges only when damages are sufficiently convex, in which case biodiversity concerns flatten the intensity profile and can expand cultivation at the extensive margin when demand is inelastic. Land heterogeneity further implies non-uniform local adjustments, especially near the cultivation threshold. I then derive policy implications for spatially uniform taxes on land use, effort, and output: the optimal mix trades off average internalization against better targeting of high-intensity land, and under quadratic costs and damages it implements the first-best allocation even with heterogeneous land. Overall, the analysis highlights the role of indirect production reallocation and market-mediated feedbacks in biodiversity-oriented agricultural policy.
JEL Classification: H23; Q15; Q57
Keywords: Biodiversity; Agriculture; Second-best policy.
