Combining farm simulation with frontier efficiency analysis

The model used, GAMEDE, is a “whole-farm” dynamic model composed of 6 biophysical modules and a management system (Vayssieres et al, 2009). This simulation model gives accurate predictions for various sustainability indicators (labor, energy consumption, production, nitrogen leaks to the environment…) to characterize observed or hypothetical farms. As GAMEDE is based on a stock-flow approach, we can monitor the farm stocks (slurry, fodders…) over time. GAMEDE also gives a full description of management operations of the production system. The GAMEDE model is randomly parameterized with the objective to cover the realm of the possible production systems by simulation. Key issues of the methodology are selecting input parameters and defining lower and upper bounds to these parameters. Expert knowledge is very useful to define these bounds. Even though the simulation approach constitutes a relevant tool for describing the production system, it can not provide a full efficiency analysis taking into account multiple input parameters. We suggest combining GAMEDE with “Data Envelopment Analysis” (DEA) to assess the efficiency of a large variety of simulated farms (frontier efficiency analysis). Each farm is characterized by different structural and management inputs parameters, inflows and sustainability indicators (including outflows). The last two types of variables are respectively inputs and outputs used in the DEA model. For inefficient farms, potential efficiency progress is calculated as the distance between these farms and the frontier. The main advantage of our methodology is to benefit of the synergy between simulation and efficiency frontier modelling, as drawback of each method are balanced by the asset of the other method.