Does land fragmentation affect farm performance? A case study from Brittany

Agricultural land fragmentation is widespread and may affect farmers’ decisions and impact farm performance, either negatively or positively. We investigated this impact for the western region of Brittany, France, in 2007. To do so, we regressed a set of performance indicators on a set of fragmentation descriptors. The performance indicators (production costs, yields, revenue, profitability, technical and scale efficiency) were calculated at the farm level using Farm Accountancy Data Network (FADN) data, while the fragmentation descriptors were calculated at the municipality level using data from the cartographic field pattern registry (RPG). The various fragmentation descriptors enabled us to account for not only the traditional number and average size of plots, but also their geographical scattering. We found that farms experienced higher costs of production, lower crop yields and lower profitability where land fragmentation (LF) was more pronounced. Total technical efficiency was not found to be significantly related to any of the municipality LF descriptors used, while scale efficiency was lower where the average distance to the nearest neighbouring plot was greater. Pure technical efficiency was found to be negatively related to the average number of plots in the municipality, with the unexpected result that it was also positively related to the average distance to the nearest neighbouring plot. By simulating the impact of hypothetical consolidation programmes on average pre-tax profits and wheat yield, we also showed that the marginal benefits of reducing fragmentation may differ with respect to the improved LF dimension and the performance indicator considered. Our analysis therefore shows that the measures of land fragmentation usually used in the literature do not reveal the full set of significant relationships with farm performance and that, in particular, measures accounting for distance should be considered more systematically.