UTAS IRP - Predicted Proportion of Underinsurance (SA1) 2016

This dataset presents the footprint of the proportion of underinsurance across Australia. The data is aggregated to Statistical Area Level 1 (SA1) geographic areas from the 2016 Australian Statistical Geography Standard (ASGS). House and contents underinsurance is understood as homeowners having no house insurance and renters having no contents insurance to cover adverse events. To create this dataset, researchers developed a method to extrapolate the patterns of underinsurance evident in the 2015 Australian Survey of Social Attitudes (AuSSA), an omnibus postal survey of Australian adults (Blunsdon, 2016). To do this, they combined the results of the full model of underinsurance with the 2016 Socio-Economic Indexes for Areas (SEIFA) (Australian Bureau of Statistics, 2019). For this spatial mapping, regression coefficients were converted to probabilities by taking the exponent of each coefficient to generate the odds ratio and then using the formula: probability = odds/(1+odds). For each SA1 unit (containing approximately 150 households), the proportion of residents or households was determined for each predictor variable from raw census data. The level of underinsurance (proportion of people predicted not to have insurance) was then predicted separately for renters and owner-occupiers for every SA1 and a single map generated by weighting the predictions by the proportion of renters and owner-occupiers per SA1. For further information about this dataset and its creation, please refer to the publication: Booth, K., & Kendal, D. (2019). Underinsurance as adaptation: Household agency in places of marketisation and financialisation. Environment and Planning A: Economy and Space. Please note:

The researchers acknowledge some limitations with the data, including the lack of data on rental properties. They do not know whether these properties are insured by landlord-investors and how this may be associated with sociodemographic variables and contribute to the mapping.

This research was in part supported by the Australian Government through the Australian Research Council Discovery Program (DP170100096).