Enhanced data envelopment analysis with undesirable outputs and interval data for efficiency measurement of rice farmers

Agricultural activities, including rice farming, produce two types of outputs: desirable and undesirable. These outputs are to be considered to correctly evaluate agricultural efficiency score. However, the undesirable outputs have been ignored by most previous studies on efficiency measurement of rice farmers in spite of the fact that they can significantly produce adverse impacts on the environment. In order to simultaneously increase desirable outputs and reduce undesirable outputs in a chosen direction, a
Directional Distance Function (DDF) approach has been introduced. The approach, however, arbitrarily chooses the direction vector to the production boundary, does not consider input and output slacks and assumes all data values are deterministic.
Therefore, this research developed two enhanced slack-based DDF Data Envelopment Analysis (DEA) models to measure the efficiency of rice farmers in addressing the drawbacks. The first model incorporated undesirable outputs and non-discretionary
inputs to measure efficiency. The second model was formulated using the interval data approach to represent data uncertainty and measure interval efficiency. Both models were tested to evaluate the efficiency of 160 rice farmers and rank efficient farmers in
northern Kedah, Malaysia. The models were assessed and validated to identify the ideal DEA model via sensitivity analyses. The overall performance of the models was compared with existing DEA models. Empirical findings showed that both models
successfully identified efficient rice farmers, and computed input savings, undesirable output reduction and desirable output augmentation for inefficient rice farmers. These models have been formulated in an effective manner and produced pertinent results.
The models help the decision makers evaluate rice farmers’ efficiency and suggest necessary actions to improve their performance, especially in uncertain operating conditions. To further improve the performance, the best practices employed by
efficient farmers can be identified and implemented by inefficient farmers.