Income polarisation, expenditure and the Australian urban middle class

Recent years have seen growing concern about the 'hollowing out' of the middle class, due to processes of polarisation. In this paper, we examine different conceptualisations of polarisation, and introduce the concept of expenditure-adjusted polarisation that considers not only income, but also various key categories of expenditure at a household level: housing, groceries and meals, transport and energy. Analysing longitudinal data from the Household, Income and Labour Dynamics in Australia Survey, we show that the Australian society is significantly more polarised, with fewer middle-income households, when the relative size of income groups in a given year is based on expenditure-adjusted income rather than pre-expenditure income. Such polarisation is particularly prominent when housing expenditure is considered and has distinctive spatial patterns. In contrast, our analysis finds no evidence of a temporal pattern of polarisation in Australia between 2005 and 2019, with no substantial change in the size of income groups over time, regardless of which income measures are used. We argue that a more nuanced conceptualisation of polarisation, and its relation to processes of 'hollowing out' and rising inequality, is needed to inform urban scholarship and policy.

Automated summary

In recent years, there has been increasing concern about the hollowing out of the middle class due to polarization processes. This paper examines different conceptualizations of polarization and introduces the concept of expenditure-adjusted polarization, which considers not only income but also various categories of expenditure at a household level. By analyzing longitudinal data from the Household, Income and Labour Dynamics in Australia Survey, the study shows that Australian society is significantly more polarized with fewer middle-income households when income groups are based on expenditure-adjusted income. Housing expenditure plays a prominent role in this polarization with distinct spatial patterns. However, there is no evidence of temporal polarization in Australia between 2005 and 2019, indicating no substantial change in the size of income groups over time regardless of the income measures used. The study argues for a more nuanced conceptualization of polarization to inform urban scholarship and policy.