Coastal hazard management involves the assessment of vulnerability in natural and human environments. Indices incorporating a diversity of indicators have therefore been used extensively to provide spatial analyses of the degree of vulnerability. Such indices are typically applied at global and national scales, and they involve varying degrees of simplification and aggregation of information. The degree of simplification that is desirable depends on the management scale, and higher resolution is required at the local compared to the global scale. To investigate the implications of spatial scale in depicting coastal hazard risk, coastal vulnerability indices were developed at national (Northern Ireland), local authority and site levels. Variables were separated into three sub-indices: a coastal characteristics sub-index concerned with the resilience and susceptibility of the coast to erosion, a coastal forcing sub-index to characterize the forcing variables contributing to wave-induced erosion and a socio-economic sub-index to assess the infrastructure potentially at risk. The three sub-indices were merged to calculate the overall index, which is portrayed in the form of colour-coded vulnerability maps. While a common tripartite index could be employed at national, regional and local scales, the nature of the data used to quantify many of the variables varies according to the scale of management. Some important local variations in vulnerability are masked by simplifications at the national scale. For some variables more detailed information is available as the spatial resolution of the study increases, while others become obsolete as data are of insufficient resolution to differentiate real variability at more detailed scales. The results highlight the importance of spatial scale in developing indices of vulnerability: while a common index architecture can be applied, the selection of variables must take account of the scale at which the hazard is to be assessed. It is likely that limits on index development will also be imposed by data availability at various scales.