Incorporating demand constraints into piecewise frontier models of public service provision
针对公共服务提供者需在需求未知前确定投入、且无法直接影响需求的情况，提出一种更合理的收入效率衡量方法，并用DEA估计，应用于澳大利亚昆士兰医院与健康服务网络。
Evaluating the performance of public service providers is often complicated by the fact that they must choose input levels before demands for their services are known. We consider an even more complicated situation in which service providers have no opportunity to directly influence demands. This means that their predetermined inputs may be more than what is required to meet realised demands. In such cases, conventional measures of revenue efficiency will generally mis-classify rational and efficient managers as inefficient. We develop a more appropriate measure of revenue efficiency that accounts for exogenously-determined demands. We explain how data envelopment analysis (DEA) methods can be used to estimate our measure. The methodology is applied to hospital and health service networks in Queensland (Australia). We find that most of these networks were able to maximise the revenue they could obtain from their predetermined inputs.