RDP 9510: Modelling Inflation in Australia[*] 1. Introduction

In the 1980s, the Australian inflation rate averaged around 8 per cent a year. At the beginning of the 1990s, the inflation rate fell substantially and now averages a much more moderate 2 per cent a year. This path of inflation has occurred alongside major changes in the Australian economy, including substantial reductions in tariffs, a shift to a more flexible and productivity-based system for setting wages, and a greater focus on international competitiveness. Moreover, the Reserve Bank of Australia has made a strong commitment to the preservation of low inflation, seeking to maintain an underlying inflation rate of around 2 to 3 per cent; see Reserve Bank of Australia (1994a, p. 3).

To understand better the behaviour of inflation and the role that a central bank may play in its determination, this paper develops an empirical model of the Australian consumer price index (CPI). The underlying economic theory is a mark-up model for prices, but the resulting empirical model also has elements relating to purchasing power parity and the Phillips curve. The empirical model clarifies the relative importance of factors determining consumer price inflation. Further, the structure of the inflationary process in Australia does not appear to have changed over the 1980s and 1990s. Rather, the recent fall in inflation is explained in terms of changes in the determinants of inflation itself.

Sections 2 and 3 briefly describe the economic theory and the data. Using quarterly series over 1977–1993, Section 4 analyses the CPI and its long-run determinants as a system, testing for and finding cointegration between them. Weak exogeneity also appears valid, so Sections 57 model the CPI as a single-equation conditional error correction model, obtained from an autoregressive distributed lag for the CPI. The error correction model is highly parsimonious and empirically constant, with an equation standard error of 0.25%; and its economic interpretation is straightforward. As an error correction model, this model of Australian CPI captures long-run effects that were ignored in some previous models, which were in first differences only. Including the error correction term in the empirical model of Australian CPI ties the model more closely to its theoretical underpinnings and improves the goodness-of-fit. Section 8 concludes. Appendix 1 describes the construction of the data; Appendix 2 documents the design of the empirical error correction model; and Appendix 3 evaluates an alternative, slightly more complicated model.


The authors are staff economists in the Economics Group, Reserve Bank of Australia, Sydney, Australia and the Division of International Finance, Federal Reserve Board, Washington, D.C., United States respectively. The views expressed in this paper are solely the responsibility of the authors and should not be interpreted as reflecting those of the Reserve Bank of Australia, the Board of Governors of the Federal Reserve System, or other members of their staffs. The second author gratefully acknowledges the generous hospitality of the staff at the Reserve Bank of Australia, where he was visiting when much of this research was undertaken. We wish to thank Darren Flood and John Irons for valuable research assistance; Palle Anderson, Carol Bertaut, David Bowman, Julia Campos, Tony Hall, Dale Henderson, David Hendry, Katarina Juselius, Deb Lindner, Jaime Marquez, Doug McTaggart, and Adrian Pagan for helpful comments and discussions; Tony Hall for providing the data in McTaggart and Hall (1993); and Jurgen Doornik and David Hendry for providing us with a beta-test version of PcGive 9.00. All numerical results were obtained using PcGive Professional Versions 8.10 and 9.00 β01; cf. Doornik and Hendry (1994). This paper is being simultaneously circulated as Research Discussion Paper No. 9510 by the Reserve Bank of Australia and International Finance Discussion Paper No. 530 by the Board of Governors of the Federal Reserve System. [*]