A Conceptual Foundation for Financial DEA

Abstract

This study aims to provide a conceptual foundation for Financial Data Envelopment Analysis (DEA), which refers to DEA models that exclusively use accounting data from external financial reports to measure firm performance. Researchers have argued that using accounting measures in DEA models without appropriate methodological rationalisation might distort performance evaluations (Färe et al., 1985; Färe et al., 2017). Researchers have also questioned the kind of efficiency measures being calculated by DEA when the inputs and outputs are accounting measures since they cover both quantities and prices (Banker et al., 2007; Cross & Färe, 2008; Portela, 2014; Zelenyuk, 2020). This study is motivated by the growth in Financial DEA research, where DEA model are potentially being applied without underpinning by an articulated conceptual foundation A two-phase analytical approach is used to examine the uses of Financial DEA and develop a conceptual foundation to assist the design and interpretation of Financial DEA. Phase I provides a typology of Financial DEA literature and reviews the methodological issues in the application of Financial DEA. Phase II quantifies selective methodological issues with empirical illustrations, using Monte Carlo simulations and analysis of archival data. The key findings and contributions are three-fold. First, the typology identifies 12 dimensional constructs of firm performance and nine indicators. This contributes a conceptual framework at the construct level, which provides an overview of the scope of Financial DEA and can be used to position the application of Financial DEA and apprise Financial DEA practice. Second, the methodological issues are formed into a four quadrant framework of measurement models with various measurement errors in Financial DEA. This contributes a conceptual framework at the modelling level, which can be used to highlight potential methodological pitfalls in Financial DEA. Third, the empirical tests demonstrate the quantitative magnitudes of selective methodological issues on Financial DEA results, exploring how research contexts affect results to varying degrees. This contributes a conceptual framework at the factor level, which can guide accounting variable selection in Financial DEA practice so that the measurement errors can be reduced. These three frameworks form a conceptual foundation for future studies of Financial DEA.