Article ID: | iaor1997836 |
Country: | United Kingdom |
Volume: | 24 |
Issue: | 3 |
Start Page Number: | 229 |
End Page Number: | 244 |
Publication Date: | Jun 1996 |
Journal: | OMEGA |
Authors: | Boussofiane A., Dyson R.G., Thanassoulis E. |
Keywords: | statistics: data envelopment analysis |
This paper compares data envelopment analysis (DEA) and ratio analysis as alternative tools for assessing the performance of organisational units such as bank branches and schools. Such units typically use one or more resources to secure one or more outputs, the inputs and/or outputs being possibly incommensurate. The assessment of District Health Authorities in England on the provision of perinatal care is used as a vehicle for comparing the two methods. The comparison focuses on how well the two methods agree on the performance of a unit relative to that of other units, and on the estimates of targets each method provides for improving the performance of units. It is found that provided the performance indicators capture all variables used in the DEA assessment the two methods agree reasonably closely on the performance of the units as a whole, though this depends on the way the performance indicators are combined into a summary figure of performance. The two methods can disagree substantially on the relative performance of individual units. Ratio analysis, unlike DEA, is not found to be suitable for setting targets so that units can become more efficient. This is mainly due to the fact that DEA takes simultaneous account of all resources and outputs in assessing performance while ratio analysis relates only one resource to one output at a time. However, the two mehods can support each other if used jointly. Ratios do provide useful information on the performance of a unit on specific aspects and they can support the communication of DEA results to non-specialists when the two methods agree on performance.