This thesis gives an overall view of the two most commonly used approaches for measuring the relative efficiencies of organisational units. The two approaches, data envelopment analysis (DEA) and stochastic frontiers (SF), are supposedly estimating the same underlying efficiency values but the natures of the two methods are very different. This can lead to different estimates for some, or all, of the units in an analysis. By identifying the nature of these differences this work shows that it is possible to gain some insight into the nature of the underlying data and to say more confidently which of the two estimates is closer to the true efficiency for individual units. In order to investigate the differences between the methods across different facets of the technology two important dimensions are chosen. Firstly differences across scale size are investigated. It is shown how it is possible to define a measure of scale size in both the single output and multiple input and output cases. This measure of scale size can then be used to split the technology into regions of differing scale size enabling, for example, tests for the true nature of returns to scale in DEA. The measure of scale size developed in multiple dimensions necessitates a method for estimating an homothetic, constant returns to scale function. Differences between the approaches across input mix are also investigated. These differences may highlight the abilities of the methods to correctly identify the elasticity of substitution between the inputs. The results of the comparisons between the methods are summarised. This summary gives possible reasons for differences which may be found between the results of the two approaches, and an indication of what the nature of the estimates may be to the true efficiency values. An algorithm is then developed for using a comparison of the results from the two methods to help to identify the better estimates.

Data Envelopment Analysis (DEA) is a linear programming technique for measuring the relative efficiencies of a set of Decision Making Units (DMUs). Each DMU uses the same set of inputs in differing amounts to produce the same set of outputs in differing quantities. Weights are freely allocated in order to allow these multiple incommensurate inputs and outputs to be reduced to a single measure of input and a single measure of output. A relative efficiency score of a DMU under Constant Returns to Scale is given by maximising the sum of its weighted outputs to the sum of its weighted inputs, such that this ratio can not exceed I for any DMU; with the weights derived from the model being taken to represent the value attributed to the inputs and outputs of the assessment. It is well known in DEA that this free allocation of weights can lead to several problems in the analysis. Firstly inputs and outputs can be virtually ignored in the assessment; secondly any relative relationships between the inputs or outputs can be ignored, and thirdly any relationships between the inputs and outputs can be violated. To avoid/overcome these problems, the Decision Maker's (DM) value judgments are incorporated into the assessment. At present there is one main avenue for the inclusion of values, that of weights restrictions, whereby the size of the weights are explicitly restricted. Thus to include the relative value of the inputs or outputs, the relative value of the weights for these related inputs or outputs are restricted. The popularity of this approach is mainly due to its simplicity and ease of use. The aim of this thesis is, therefore, firstly, to demonstrate that, although the weights restrictions approach is appropriate for many DMs, for a variety of reasons some DMs, may prefer an alternative form for the expression of their values, e.g. so that they can include local values in the assessment. With this in mind, the second aim of this thesis is to present a possible alternative approach for the DMs to incorporate their values in a DEA assessment and, thirdly, it aims to utilise this alternative approach to improve envelopment. This alternative approach was derived by considering the basic concept of DEA, which is that it relies solely on observed data to form the Production Possibility Set (PPS), and then uses the frontier of this PPS to derive a relative efficiency score for each DMU. It could be perceived, therefore, that the reason for DMUs receiving inappropriate relative efficiency scores is due to the lack of suitable DEA-efficient comparator DMIUs. Thus, the proposed approach attempts to estimate suitable input output levels for these missing DEA-efficient comparator DMUs, i.e. Unobserved DMUs. These Unobserved DMUs are based on the manipulation of observed input output levels of specific DEA-efficient DMUs. The aim of the use of these Unobserved DMUs is to improve envelopment, and the specific DEA-efficient DMTJs that are selected as a basis for the Unobserved DMILTs are those that delineate the DEA-efficient frontier from the DEA-inefficient frontier. So, the proposed approach attempts to extend the observed PPS, while assuming that the values of the observed DEA-efficient DMIJs are in line with the perceived views of the DM. The approach was successfully applied to a set of UK bank branches. To illustrate that no approach is all-purpose, and that each has its strengths and weaknesses and, therefore, its own areas of application, a brief comparison is made between the approach of weights restrictions and the approach proposed in this thesis. This thesis is divided into three sections: A - Overview of the research area; B - An alternative perspective for incorporating values in DEA; C - The use of UDMUs to express the DM's values to improve envelopment