Recognized as the central challenge of the 21^st century, the sustainable use of resources and the limitation of toxic emissions have become an indisputable goal for the conservation of biodiversity and Homo sapiens’ permanence.  The notion of eco-efficiency has been seen as a pragmatic approach to better represent the principles of sustainable development (Schaltegger and Synnestvedt, 2002). It is also considered as a measure of the performance of multi-level sustainable development policies, ranging from a micro analysis to a macro analysis, whose aim is to bring a certain harmony between natural ecology and human ecology (Li et al., 2017).

This paper addresses the problem of eco-efficiency assessment, using the DEA method, when desirable and undesirable outputs are jointly generated by the firm’s activity. It uses the data of 19 European listed companies in the chemical sector, for the period 2012-2016. This business sector is considered as one of the most polluting industrial sectors on the planet. 

The approach adopted in this research is at once descriptive and analytical, prospective and propositional because its interests are:

(1) modeling and evaluation of eco-efficiency in the presence of undesirable outputs. Different DEA models are applied for estimating eco-efficiency scores. The CCR and BCC models will be used to analyze the returns to scale and the efficiency of scale; the SBM model will be used to analyze eco-efficiency scores, in addition to output and input slacks.

(2) recognizing the existence of good practices between homogeneous enterprises and the possibility of dissemination of these practices. The DEA method makes it possible to identify benchmarking partners from which managers put in place corrective and adaptive strategies for better eco-efficiency.

(3) measuring the change in eco-efficiency between 2012 and 2016 (following the numerous European regulatory constraints) by using two methods: the Window analysis and the analysis by the Malmquisit productivity index. This index allows us to understand if the evolution of productivity (eco-efficiency score) is due to technical efficiency or technological progress.