Mohsen Afsharian

Generalized DEA: an approach for supporting input/output factor determination in DEA

Purpose The determination of input and output factors is a well-known source of pitfalls when applying data envelopment analysis (DEA). The purpose of this paper is to contribute to overcome the respective problems of input/output factor determination related to factor selection, dual-role factors and undesirable factors. Design/methodology/approach The problems of input/output factor determination are discussed from a goal-oriented perspective, shedding a new light on the role of input/output factors in DEA. This is exemplified by the case of measuring pharmacy stores’ efficiency concerning their goal of customer retention. Findings The findings suggest to applying a generalized DEA (GDEA). The three steps of this approach include the development of a system of objectives, the derivation of corresponding performance criteria as well as the construction of cost and benefit functions. These functions build the basis for GDEA models, of which one is exemplarily described and applied to the customer retention case. Research limitations/implications While traditional DEA implicitly assumes linear cost and benefit functions, GDEA requires to explicitly specifying these functions. In doing so, the approach contributes to solve the problem of factor selection, the problem of dual-role factors and the problem of undesirable factors. Practical implications For determining input/output factors in a consistent and transparent manner, it is recommended to apply GDEA in practical benchmarking studies. Originality/value GDEA integrates well-known concepts of multi-criteria decision making into traditional DEA. The new approach helps to cope with the challenges of input/output factor determination in DEA.

A DEA-based incentives system for centrally managed multi-unit organisations

In multi-unit organisations such as a bank and its branches or a national body delivering publicly funded health or education services through local operating units, the need arises to incentivize the units to operate efficiently. In such instances, it is generally accepted that units found to be inefficient can be encouraged to make efficiency savings. However, units which are found to be efficient need to be incentivized in a different manner. It has been suggested that efficient units could be incentivized by some reward compatible with the level to which their attainment exceeds that of the best of the rest, normally referred to as “super-efficiency”. A recent approach to this issue (Varmaz et. al. 2013) has used Data Envelopment Analysis (DEA) models to measure the super-efficiency of the whole system of operating units with and without the involvement of each unit in turn in order to provide incentives. We identify shortcomings in this approach and use it as a starting point to develop a new DEA-based system for incentivizing operating units to operate efficiently for the benefit of the aggregate system of units. Data from a small German retail bank is used to illustrate our method.

Developing selective proportionality on the FDH models: new insight on the proportionality axiom

In many cases of DEA-based efficiency measurement systems, only a set of outputs has to be assumed to be proportional to a set of inputs. The assumption of constant returns to scale (CRS) is required with respect to the selected sets of inputs and outputs, preserving the variable returns to scale (VRS) assumption for the remaining factors. In such situations, neither CRS nor VRS-based models can provide valid results. In contrast to that, the selective proportionality axiom allows applying any desired combination of CRS and VRS. This paper proposes free disposal hull (FDH) technologies which incorporate the selective proportionality axiom. The considered technologies do not restrict themselves to convex technologies and are built solely on minimal axioms of non-emptiness and free disposability. The resulting FDH models are formulated as linear programming problems which are not only simple to solve but also provide an intuitive interpretation corresponding to the results. An illustrative numerical example is presented to explain the properties and features of the suggested FDH models.

A heuristic, dynamic programming-based approach for a two-dimensional cutting problem with defects

This paper deals with a two-dimensional cutting problem in which small rectangular items of given types are to be cut from a rectangular large object which contains several defects. It is assumed that the number of pieces of each small item type which can be cut from the large object is not limited. In addition, all cuts are restricted to be of the guillotine-type and the number of stages necessary to perform all cuts is not limited. Furthermore, no small item must overlap with a defective region. The objective is to maximize the value of the cut small items. For the solution of the above-described problem, a heuristic, dynamic programming-based approach is presented which overcomes the structural and computational limitations of previously-proposed methods. In the presence of defects, the representation of the defective regions and the definition of discretization sets are revisited. This allows for an improvement of the computational efficiency as well as of the storage space requirements for solving the given problem with any number of defects in this approach. We further analyze the computational complexity of the algorithm and identify the factors which affect its running time. The proposed method is evaluated by means of a series of detailed numerical experiments which are performed on problem instances extracted from the literature, as well as on randomly generated instances. The experiments do not only illustrate how the suggested method is able to identify optimal solutions of the test problem instances, but they also explain why already existing methods fail to do so. Furthermore, the computational results indicate that the proposed method, equipped with the newly-proposed discretization sets, is capable of efficiently generating a high percentage of optimal solutions to the corresponding problem with defects.

The Luenberger indicator and directions of measurement: a bottoms-up approach with an empirical illustration to German savings banks

The Luenberger productivity indicator applies directional distance functions which allow to specifying in what direction (i.e. direction of measurement) the operating units will be evaluated. In the presence of a change in the direction of measurement, the standard components of the existing Luenberger productivity indicator may provide values which are not compatible with reality. In order to eliminate this pitfall, the so-called bottoms-up approach is used to revisit the definition of the indicator and its components. We start with a list of selected sources of productivity change, namely efficiency change, technical change and direction change, then examine the best possible way of measuring each of the sources and combine them to derive a new measure of productivity change. The proposed indicator will be illustrated by means of an empirical application to a panel of 417 German savings banks over the time period 2006–2012. The example explains how the proposed approach is able to properly measure efficiency change, technical change and direction change. The results also provide conclusive evidence about the effect of the change in direction of measurement on the results of the productivity over time in a centralised management scenario.

A linear programming approach to efficiency evaluation in nonconvex metatechnologies

The notions of metatechnology and metafrontier arise in applications of data envelopment analysis (DEA) in which decision making units (DMUs) are not sufficiently homogeneous to be considered as operating in the same technology. In this case, DMUs are partitioned into different groups, each operating in the same technology. In contrast, the metatechnology includes all DMUs and represents all production possibilities that can in principle be achieved in different production environments. Often, the metatechnology cannot be assumed to be a convex set. In such cases benchmarking a DMU against the common metafrontier requires implementing either an enumeration algorithm and solving a linear program at each of its steps, or solving an equivalent mixed integer linear program. In this paper we show that the same task can be accomplished by solving a single linear program. We also show that its dual can be used for the returns-to-scale characterization of efficient DMUs on the metafrontier.

Multi-period productivity measurement under centralized management with an empirical illustration to German saving banks

In this paper, we revisit the structure of the centralized Malmquist indices which apply inter-temporal benchmark technologies coupled with a relaxed assumption that the technology remains unchanged between the start and the end of the analysis. From a theoretical point of view as well as with an empirical application to a panel of German savings banks over the time period 2006–2012, we discuss this premise as the technology—which is naturally under the influence of different external and internal conditions—can change over time. This may hence result in an inappropriate estimate of the benchmark technology, generate questionable sets of common-weights and lead accordingly to misleading results and managerial conclusions. To eliminate this pitfall, we propose a new centralized framework in which individual characteristics of the technology, represented by different contemporaneous technology sets over time, can be preserved and later traced in measuring productivity change. Details of our empirical results, determined by the proposed Malmquist index, reveal that the productivity of the group of German savings banks has always been increasing during the whole period analyzed. The positive rates of growth highlight the fact that this group had a stable financial system even when the financial crisis hit the international monetary and financial market. The best practice change component of the suggested Malmquist index also verifies the significant effect of change in the technology on the performance of these banks over time. Although the group of German savings banks reduced its fixed assets over time, our analysis of productivity change shows how successfully these banks could improve even in a highly customized and growing digital business environment. However, looking at the slowdown in the growth of productivity between 2011 and 2012, captured by our results, it seems advisable that they accelerate the adaptation of their business strategy, e.g., by investing more in high-quality and diverse internet-based products and services to catch up with the rapid developments in information technologies.

Consistent proportional trade-offs in data envelopment analysis

Abstract Proportional trade-offs – as an enhanced form of the conventional absolute trade-offs – have recently been proposed as a method which can be used to incorporate prior views or information regarding the assessment of decision making units (DMUs) into relative efficiency measurement systems by Data Envelopment Analysis (DEA). A proportional trade-off is defined as a percentage change of the level of inputs/outputs so that the corresponding restriction is adapted with respect to the volume of the inputs and outputs of the DMUs in the analysis. It is well-known that the incorporation of trade-offs either in an absolute form or proportional form may lead in certain cases to serious problems such as infinity or even negative efficiency scores in the results. This phenomenon is often interpreted as a result of defining the set of trade-offs carelessly by the analyst. In this paper we show that this may not always be the case. The existing framework by which the trade-offs are combined mathematically to build a corresponding production technology may cause a problem rather than the definition of the trade-offs. We therefore develop analytical criteria and formulate computational methods that allow us to identify the above-mentioned problematic situations and test if all proportional trade-offs are consistent so that they can be applied simultaneously. We then propose a novel framework for aggregating local sets of trade-offs, which can be combined mathematically. The respective computational procedure is shown to be effectively done by a suggested algorithm. We also illustrate how the efficiency can be measured against an overall technology, which is formed by the union of these local sets. An empirical illustration in the context of engineering schools will be presented to explain the properties and features of the suggested approach.