Technical, Allocative and Overall Efficiency: Inference and Hypothesis Testing
(2018) , 43 pages
Files
ISBADP2018_18_simar_TechnicalAllocativeandOverallEfficiencyInferenceandHypothesisTesting_dial.pdf
Open Access - Adobe PDF
- 544.81 KB
Details
- Authors
- Abstract
- Nonparametric estimates (based on data envelopment analysis or DEA estimators) of cost, revenue or profit efficiency are typically obtained in the applied literature while working in the full (p+q)-dimensional space of p inputs and q outputs. In the case of cost efficiency, minimum cost is first estimated and then divided by observed cost, while in the case of revenue efficiency, maximum revenue is estimated and then divided by observed revenue. The statistical properties of these estimators are unknown, and consequently no methods for inference exist. This paper shows that cost efficiency can be estimated in a (1 + q)-dimensional space by a simple distance function estimate. Similarly, revenue efficiency can be estimated in a (p + 1)-dimensional space by a distance function estimate. This improves the rate of convergence when p > 1 or q > 1. In addition, properties of standard DEA input and output oriented estimators established by Kneip et al. (2008, 2015) as well as properties of free-disposal hull (FDH) estimators established by Park et al. (2000), Daouia et al. (2017) and Kneip et al. (2015) are shown to hold for the new distance function estimators, permitting statistical inference.
- Affiliations
Citations
Simar, L., & Wilson, P. (2018). Technical, Allocative and Overall Efficiency: Inference and Hypothesis Testing (ISBA Discussion Paper 2018/18). https://hdl.handle.net/2078.5/174094