Accounting for Heterogeneity in Environmental Performance Using Data Envelopment Analysis
This paper proposes a novel way of modeling heterogeneity in the context of environmental performance estimation when using Data Envelopment Analysis. In the recent literature estimation of productive efficiency is common and relies on inputs and outputs identifying an environmental production technology in cases of joint production of good and bad outputs. However, heterogeneity is an important issue in this context. Our proposed novel approach relies on identification of different groups using a multivariate mixture-of-normals-distribution. The new techniques are applied to a data set of 44 countries during 1996–2014 concerning the finance of environmental efforts where significant problems of heterogeneity both in cross-sectional as well as in the time dimension are anticipated. For this purpose, apart from the usual variables of the production function, proxies of environmental investments like renewable electricity output and research and development expenditures are used. The sampling properties of the new approach are investigated using a Monte Carlo experiment. The problem of structural breaks over time is also considered with a penalty term in local likelihood estimation.
KeywordsEnvironmental performance Heterogeneity Production Data Envelopment Analysis Multivariate-mixture-of-normals-distributions
JEL ClassificationQ56 Q42 C13 C14
We would like to thank the Editor Professor Hans Amman and three anonymous reviewers for the comments provided in relation to an earlier version of our paper. Any remaining errors are solely the authors’ responsibility.
- Barr, R., Durcholz, M., & Seiford, L. (1994). Peeling the DEA onion: Layering and rank-ordering DMUs using tiered DEA. Southern Methodist University Technical Report, 1994/2000.Google Scholar
- Sakris, J. (2007). Preparing you data for DEA. In: J. Zhu & W. D. Cook (Eds.), Modelling data irregularities and structural complexities in DEA (Chapter 17 pp, pp. 305–320). Springer.Google Scholar
- Sarrico, C. S., & Dyson, R. G. (2000). Performance measurement in UK universities—The institutional perspective. Journal of Operational Research Society, 51(7), 789–800.Google Scholar