The inefficiency of health care provision presents a major health policy concern in Germany. In order to address the issue of efficiency comprehensively - i.e. at the level of the entire system of health care provision rather than individual service providers - empirical analyses are often based on data at the regional level. However, regional efficiencies might be subject to spatial dependence, rendering any analysis biased that aims at identifying the determinants of efficiency differentials. We address this issue by specifying a spatial auto- regressive model to explain efficiency scores for German districts which we derive through data envelopment analysis. Regression results suggest that spatial dependence is not a dominant feature in the data. Hence, ignoring spatial interdependence is unlikely to severely bias results of efficiency analyses based on regional data. This holds, in particular, for the role of the states in the efficiency of health production. Significant heterogeneity among states is found in the data regardless of whether or not spatial dependence is accounted for.
Health production data envelopment analysis spatial auto- regressive model