Predicting repeat consumer bankruptcy: A survival analysis of business-related repeat filings in Australia 2007–2021

Abstract

Ongoing legislative responses to the impacts of the pandemic have prompted many countries to evaluate whether their bankruptcy systems remain fit for purpose. Moreover, the current climate highlights the importance of data-driven policy, which the literature identifies as a deficiency of bankruptcy regimes. In Australia, the 2015 reform proposals to reduce the default discharge period from 3 years to 1 year are currently being revised amidst stakeholder concern about potential abuse and repeat bankrupts. Although an extensive body of literature exists on ‘repeat filers’ in the USA, there has been no equivalent study in Australia. Using our data of 153,526 bankruptcies between 2007 and 2021, we conducted a novel application of survival analysis to predict the probability of a repeat bankruptcy comparing business and non-business groups. The results show that this probability peaked in both male and females with non-business-related administrations irrespective of client's age, employment and relationship status. These findings are important as they identify the prospects that certain bankrupt groups have higher rates of repeat bankruptcy, which can inform strategies to improve their survival rate. A significance of our study is the development of a high-quality longitudinal dataset that facilitates the extension of the data models and allows easy updates about targeted questions involving bankruptcy-related policy shifts and impacts on sub-populations. This methodological approach will enable regulators and insolvency experts to address concerns of repeat bankruptcy to guide policy, evaluate reform and extend the evidence base in other jurisdictions.