Benchmarking road safety performance by grouping local territories: A study in the Netherlands

•We studied possible starting points for benchmarking sub-national regions.•Municipalities and their structural factors relevant for road safety were studied.•Three grouping methods were compared: traditional (provincial), simple and advanced.•Level of urbanisation showed to be the most important factor for simple grouping.•Advanced grouping has advantages when benchmarking rural areas or special issues. The method of benchmarking provides an opportunity to learn from better performing territories to improve the effectiveness and efficiency of activities in a particular field of interest. Such a field of interest could be road safety. Road safety benchmarking can include several indicators, ranging from performance indicated by crash statistics, to indicators that also account for consequences in costs or the underlying state of the road safety system and relevant organisation and processes at actor level. The structure and culture of a territory is identified as a basic context of road safety performance. This is regarded as important information to use in grouping of territories to get more homogenous or equal and comparable conditions to learn from ‘the best in class’. The main aim of this study is to assess the usability of different groupings using the physical structure for benchmarking road safety performance at local territorial level. A traditional grouping of municipalities in the Netherlands was compared with a simple grouping of these municipalities based on their level of urbanisation and an advanced grouping in which more indicators such as differences in demography, growth and road structure were taken into account. As in other studies, urbanisation showed to be the most predominant structural factor for grouping local territories and related to differences in road safety performance. However, if information would be needed for specific target groups, other factors like age and gender distribution or the distribution of the road network can provide valuable additional insight and better homogenous starting points for benchmarking. Especially benchmarking of rural territories may profit from such extra distinctive characteristics.