In recent years interest in benchmarking and comparisons of performance measures has increased mong Swedish municipalities. Comparisons of performance measures are used for different urposes. They are used in benchmarking processes, with the aim not only of improving perations but also of providing standards in the context of accountability. Regardless of the urpose of the comparisons of performance measures, it is important that the compared measures re defined and applied homogeneously and consistently, i.e. they must have information orrespondence in order to be comparable; otherwise, erroneous decisions can be made. owever, commensurability is not enough. If the compared measures are to have any information alue, the measures also need to be chosen carefully, so that, in a relevant and reliable way, they eflect the measured objects’ most important qualities. This paper analyses performance easures used in three benchmarking projects in the Swedish water and sewage sector. The easures are analysed in the light of how well they reflect the measured objects’ most important ualities, their relevance, commensurability, and reliability. The findings are distressing; the ppropriateness of the performance measures in these three projects must be strongly questioned. n particular, the measures were too aggregated as to capture, reproduce, and describe the most mportant qualities of the measured objects in a relevant way. As for commensurability, it was so oor that one is entitled to sk hether the comparisons are not doing more harm than good.