Prepayment meters are normally installed in the UK to address the risk of non-payment from overindebted households and the literature shows a discrepancy of higher energy prices in prepayment meters. This research seeks to understand the spatial aspect of this sorting process, where prepayment meters and higher energy prices are concentrated in the areas of higher fuel poverty. A corollary research question is whether this sorting affects aspects of the consumption of housing services with respect to structural and neighbourhood characteristic. State-of-the-art latent class discrete choice models (LCM) are employed on the choice of prepayment to standard payment meter. LCM approach identifies unobservable subgroups within the population and the housing stock, allowing better understanding the impact of exposure to patterns of multiple risks, as well as the antecedents and consequences of complex behaviours. Therefore, interventions can be tailored to target the subgroups that are affected most; in this case, households vulnerable to fuel poverty affected by market failures that lead to adverse self-selection.