RDP 2004-09: Co-Movement of Australian State Business Cycles 1. Introduction

Are Australian state business cycles synchronised? While there is a wealth of literature addressing the issue of cyclical co-movement among countries, very little focus has been given to the issue at a state level, despite its importance for policy-makers. The extent of co-movement of state activity has implications for our understanding of the economy as a whole, and for the role of national policy instruments in smoothing business cycles. Furthermore, an understanding of the sources of cyclical co-movement could help policy-makers improve their response to economic shocks, given that purely regional shocks may involve a different adjustment of capital and labour than shocks affecting all states.

An examination of the growth of economic activity across Australian states, using state final demand as the measure of activity, suggests that economic cycles are quite closely synchronised (Figure 1).[1] This is particularly true of the three largest states, whose business cycles appear to have been very similar since the mid 1980s. The cycles of the three smaller states appear more disparate, but even these states share some cyclical similarities.

Figure 1: Cycles in State Economic Activity
Year-ended growth in state final demand
Figure 1: Cycles in State Economic Activity

The aim of this paper is to identify the extent and sources of business cycle co-movement among Australian states. While there are a number of studies focusing on regional co-movement in other countries, there has been little research on this issue for Australia. A notable exception is Dixon and Shepherd's (2001) study of unemployment in Australian states, which found a high degree of correlation between state unemployment rates, along with evidence that cyclical activity in the five largest states is driven by a common cycle.[2] Section 3 of our paper expands this research by focusing on more direct measures of the business cycle, and by looking at a range of alternative measures of co-movement. Consistent with Dixon and Shepherd, we also find that state cyclical activity is quite closely linked. Sections 4 and 5 then look at what may account for this co-movement, a question which, to the best of our knowledge, has not previously been addressed in Australia. In support of the findings in Section 3, our results indicate that Australian state activity is dominated by a common cycle, with region-specific fluctuations and spillovers of fluctuations between states making less substantial contributions.


For reasons discussed shortly, we focus on state final demand as a measure of activity in this paper. We confine our analysis to the six states of Australia. [1]

There are other studies that look at synchronisation of state labour markets over a longer time frame than a cycle (for example, Debelle and Vickery 1999). [2]