An efficient approach to structural breaks and the case of automobile gasoline consumption in Australia

In this paper, we propose an iterative approach within a threshold quantile regression framework to identify sequential structural breaks. We apply it to a time series analysis with a focus on establishing a connection between structural breaks in the gasoline consumption function and extreme events in the economy. The estimated structural breaks and piecewise relationship between Australia's automobile gasoline consumption and household disposable income are consistent with economic intuition. Moreover, the heterogeneity in their relation across various quantiles and over different time periods divided by the identified structural breaks is revealed in a systematic manner. The comparison with a state-of-practice approach demonstrates the validity and efficiency of our approach.