Performance comparison of retailing stores using a Malmquist-type index

Abstract

This study develops a framework that combines different management science methods to provide insights concerning the performance of retailing stores. First, the framework enables to specify appropriate targets for stores of a retail network using data envelopment analysis. This involves comparing stores within homogenous groups, that is, supermarkets and hypermarkets. Second, the framework compares the overall performance of these two groups. This requires the combined use of a Malmquist-type index and statistical tests. This index is decomposed into sub-indices for comparing the differences between groups in terms of the efficiency spread in each group of stores and the productivity differences between the best-practice frontiers spanned by the benchmark stores from each group. The hypothesis tests are used to verify if the differences between groups captured by the sub-indices are statistically significant.

Performance comparison of wind energy conversion system technologies

“…The original DEA model [17] is used to assess the farms efficiency in producing electric energy from the physical resources and the wind available in each farm. The efficiency scores can be used to calculate the Malmquist-type index for groups comparison [6,18], which enables to compare the overall performance of wind farms using the two conversion system technologies under analysis. Malmquist index can be decomposed in two components: the IE for comparing the differences between groups in terms of the efficiency spread in each group of farms and the IF for comparing the productivity differences between the "best-practices" frontiers defined by the benchmark farms from each group.…”

“…The Malmquist index (MI) was introduced by Caves et al [19], developed further in the context for measurement of productivity change over time by Färe et al [20] and extended by Camanho and Dyson [18] for comparing different groups. This study uses the output oriented version of the MI, presented in [6], which is consistent with the objectives of the problem under analysis. MI computation is based on radial measures, defined by output distance functions, which are equal to the efficiency score estimated by the DEA model previously introduced, given by o θ , for each DMU o .…”

“…The efficiency compares the ratio output/input in each farm with the best ratio observed among all farms (relative concept). The DEA models enable to compute the Malmquist index to compare groups [6] by assessing the efficiency and productivity of wind farms. The robustness of the efficiencies scores derived from DEA models can be tested by using bootstrapping framework [7].…”

A note on measuring group performance over time with pseudo-panels

“…The main advantage of this approach compared with others (Charnes et al, 1981), is that it does not assume convex combinations of group specific frontiers to be feasible. This appealing property has allowed to the CDMI to be used successfully in empirical applications, to provide performance comparisons of groups of units in crosssectional studies (see, for example, Vaz and Camanho, 2012, Ferreira and Marques, 2014and Thanassoulis et al 2015.…”

Factor-specific Malmquist productivity index based on common weights DEA

“…DEA-based MPI has been widely used in various application areas, such as commercial banks (Bi et al 2011), traffic safety (Odeck 2006), construction industry (Xue et al 2008), retailing stores (Vaz and Camanho 2012), artisanal fishing (Oliveira et al 2009), and so on. It is especially useful to examine the effect of a reform or act.…”