Monash University

The robustness of data envelopment analysis to omitted and/or irrelevant production inputs

Email address: Tissa.Galagedera@buseco.monash.edu.au

Keywords: Relative efficiency, non-parametric, data envelopment analysis

JEL Classifications: C44, C61, C67

Abstract:
In the recent literature on efficiency measures, non-parametric data envelopment analysis (DEA) has been quite extensively used for measuring relative performance of, for example, universities and schools in the education sector, and banks, mutual funds and superannuation funds in the financial sector. However, the robust properties of DEA to omission of relevant production inputs and inclusion of irrelevant production inputs are largely unknown. In this paper, a limited simulation study is carried out to address this issue. Data is simulated from the Cobb-Douglas production process under different returns-to scale (RS) assumptions. Omission of relevant inputs and inclusion of irrelevant inputs appear to adversely affect the results of DEA, depending on the assumptions on RS. The DEA with irrelevant inputs perform better under the correct RS assumption. Further, in the production process with constant RS specification, the DEA with irrelevant inputs tend to over estimate the efficiencies in almost all production units studied in this paper. A similar observation is made in the presence of irrelevant inputs when the underlying process is decreasing RS but unwittingly assumed variable RS. The DEA with omitted inputs perform better under the variable RS assumption irrespective of the nature of the true specification.

PDF file of paper: galagedera.pdf

Session: Econometric Methods II

Time: Saturday, 7 July, 2:15pm - 3:45pm

Room: B