Computing confidence intervals for output-oriented DEA models: an application to agricultural research in Brazil

Abstract

We define and model the research production at Embrapa, the major Brazilian institution responsible for applied agricultural research. The main theoretical framework is data envelopment analysis. We explore the economic interpretation and the statistical properties of these models to compute confidence intervals for output-oriented efficiency measurements. It was based on a parametric flexible model, defined by the truncated normal distribution. Intervals are calculated exploring a bootstrap approach. These results provide a better insight on the efficiency classification and allow comparisons among the decision making units involved in the evaluation process, taking into account random errors and inefficiency random variation.

Keywords:

• data envelopment analysis
• confidence intervals
• bootstrap
• agricultural research

Acknowledgements

To the National Council for Scientific and Technological Development (CNPq), for the financial support.