Measuring absolute efficiency is difficult because it is often not possible to include all factors which influence the performance. This is especially true for the agricultural sector, where the performance of farms is strongly influenced by bio-geophysical site conditions. To overcome this problem, efficiency is often measured in relative terms in the form of benchmarking. 

Benchmarking is applicable to several different sectors, and plays also an important role in agriculture, in terms of assessing environmental impacts that are connected to the efficiency of their performance (Kuo et al. 2014). Not only policymakers benefit from the assessment of efficiency and productivity but also the farmers can learn from the best operators how to handle resources efficiently. An important goal is to use as few inputs (resources) as possible to produce the most outputs.

Benchmarking methods can be applied to evaluate the impact of multiple policies on regional agricultural performance, technical efficiency and environmental sustainability (Quiroga et al. 2017). To achieve improvements and for moving towards a sustainable agriculture, benchmarking consider effects on environmental issues such as water conservation or soil protection (Kuo et al. 2004).

This is solved by comparing similar farms with each other and by identifying very efficient farms. They are used as examples of good management practices to achieve a better performance.

Modern benchmarking analyses use either best practice, where best performers are identified and used as an example, or frontier analysis methods. For the latter, Stochastic Frontier Analysis (SFA) and Data Envelopment Analysis (DEA) are the most common parametric and nonparametric methods. They have been developed as economic tools to evaluate the performance of individual companies or enterprises, referred as “decision-making units (DMU)”  (Malana & Malano 2006). DMUs must be defined and can stand for instance for organizations, which use several inputs to transform them into certain outputs (Bogetoft & Otto 2010). Both methods are characterized by frontier lines which stand for the optimal combination of inputs and outputs (Oh & Hildreth 2016). The efficiency of each farm is calculated as distance to the efficiency frontier. They can be used to estimate overall resource use efficiency and rank farms based on their performance, or to identify areas of inefficiency in order to support improvement.

SFA
Parametric models, such as SFA, are defined as a-priori except for a finite set of unknown parameters that need to be estimated from data. These parameters could refer to the parameters in the possibly random noise and efficiency distributions (Bogetoft & Otto 2010). With SFA approaches hypotheses about production structure and degree of inefficiency can be performed using statistical tests. However, the method requires explicit parametric functions and distributional assumptions.

The best practice frontier is estimated through a mathematical function. Actual input/output relationships are used to estimate optimal input/output conditions for a certain time period. For considering measurement errors and effects which cannot be controlled by the individual farms, such as weather, a random error is incorporated. The validity of SFA depends on the accuracy of the assumptions made (Odeck 2007).

DEA
As a difference to parametric models, nonparametric approaches such as DEA do not require any a-priori assumptions regarding the quality or distribution of the data, avoiding potential errors in this estimation (Oh & Hildreth 2016). Instead, all parameters are calculated from the data in the sample. However, it is also sensitive to measurement errors or other noise affecting the dataset. Environmental factors are not explicitly considered. Therefore, it is important to carefully select the sample to avoid comparing farms that work under different environmental conditions (Odeck 2007).

DEA calculates the best practice-frontier line (highest outputs from fewest inputs) from linear combination of farms in the sample. Two types of DEA models can be distinguished: The input-oriented model aims towards minimizing inputs while maintaining the amount of outputs. Output-oriented models on the other hand aim to increase outputs with the same amount of inputs (Malana & Malano 2006). 

Both of these methods show advantages and disadvantages that may influence the choice of method. The possibility of a combination of both methods might help to overcome disadvantages and has been applied to several studies dealing with benchmarking processes (Oh & Hildreth 2016). Both methods allow us to work with multiple inputs and outputs to accomplish an overall evaluation of different production entities (Bogetoft & Otto 2010).

Bogetoft, P. & Otto, L. 2010. Benchmarking with DEA, SFA, and R. Springer Science & Business Media. 352 pp

Kuo, H., Chen, H., Tsou, K. 2014. Analysis of Farming Environmental Efficiency Using a DEA Model with Undesirable Outputs. APCBEE Procedia 10: 154-158

Malana, N.M., Malano, H.M. 2006. Benchmarking productive efficiency of selected wheat areas in Pakistan and India using Data Envelopment Analysis. Irrigation and Drainage 55: 383-394. DOI: 10.1002/ird.264

Malano, H., Burton, M., Makin, I. 2004. Benchmarking Performance in the Irrigation and Drainage sector: a tool for change. Irrigation and Drainage 53: 119-133. DOI: 10.1002/ird.126 

Odeck, J. 2007. Measuring technical efficiency and productivity growth: a comparison of SFA and DEA on Norwegian grain production data. Applied Economics 39: 2617-2630. DOI: 10.1080/00036840600722224

Oh, S.-C., Hildreth, A.J. 2016: Analytics for Smart Energy Management. Springer Series in Advanced Manufacturing. DOI: 10.1007/978-3-319-32729-9_2

Quiroga, S., Suárez, C., Fernández-Haddad, Z., Philippidis, G. 2017. Levelling the playing field for European Union Agriculture: Does the Common Agricultural Policy impact homogeneously on farm productivity and efficiency? Land Use Policy 68: 179-188. DOI: 10.1016/j.landusepol.2017.07.057