Victoria’s Economic Bulletin: An application of the Flexible Fourier Form to the Australian Housing Market

Victoria’s Economic Bulletin, Volume 6, Number 1

Published by:
Department of Treasury and Finance
Date:
1 June 2022

Jack McCorkell1,3 and Jonathan Dark2

1 Revenue Forecasting Team, Economic Division, DTF.
2 Department of Finance, University of Melbourne; Revenue Forecasting Team, Economic Division, DTF.
3 For contact author’s details and/or questions on this article, please email: veb@dtf.vic.gov.au

Disclaimer: The views expressed in this paper are those of the authors and do not necessarily reflect the views of DTF.
Suggested Citation: McCorkell J and Dark J (2022) An Application of the Flexible Fourier Form to Australian Housing Market. Victoria’s Economic Bulletin, June,vol 6, no 1. DTF.

Abstract

We apply the Flexible Fourier Form or FFF (a semi-nonparametric non-linear function) to Australian house prices from 1980 to 2021. First, we decompose real house prices into a non-linear long-term (LT) component and a short-term (ST) component. Real house prices and their LT component have a stable long-run relationship (or cointegrate) with income (real disposable income and real GDP), but not when income is on a per capita basis.

This is consistent with real house price growth outpacing growth in income per person, as well as a non-trivial part of overall income growth being caused by population growth. Results also show that the FFF decomposition reveals other cointegrating relations with rent and affordability measures that are undetected when using raw house prices.

Finally, the short‑term component of house prices is shown to be cyclical and strongly related to sales volumes (but not income). Second, we use the FFF to control for structural breaks within Vector Autoregression (VAR) and Vector Error-Correction Models (VECM). Unlike the conventional VAR/VECM, where variables revert to their (constant) unconditional mean, inclusion of the FFF means that each variable reverts to a deterministic value that evolves over time in accordance with smooth structural change. We show that when allowing for structural change, the VAR and VECM yield greater variable interaction and additional insights into the dynamics of Australian house prices.

Updated