Assessing Early Policy Impacts of the National Housing Accord Using ARIMAX Time Series Forecasting
DOI:
https://doi.org/10.61173/0f922c53Keywords:
ARIMAX Forecasting, Housing Policy Evaluation, Housing Supply, National Housing AccordAbstract
This paper aims to address the research gap on evaluating the National Housing Accord’s effects post-implementation. It applies ARIMAX time series forecasting to the Monthly Building Approvals from the Australian Bureau of Statistics prior to the commencement of the National Housing Accord. The forecasted values are then compared to the actual values to assess the short-term impacts of the Accord to date. It was found that the actual values had an increasing trend that deviated from the flat forecasted trend, indicating a positive effect of the National Housing Accord on building approvals. However, it remained within the 80% confidence interval of the forecasting model, not allowing a definite conclusion to be drawn. The main limitations of the paper discussed included the short time frame and the selection of exogenous variables for the model. It suggested that future research should carry out a similar methodology, but at a later stage of the Accord and after revisiting the selection of exogenous variables. Overall, this paper represents an initial step toward quantitatively evaluating the effects of the NHA on housing supply.
References
correlated with the current error term. This means that our the housing system 2024. 2024. forecasts may have been negatively impacted by such bi- [4] Moallemi M, Melser D. The impact of immigration on ases, making it another limitation of this model. housing prices in Australia. Papers in Regional Science. 2020;99(3):773–786. 5. Conclusion [5] Productivity Commission. In need of repair: The National Housing and Homelessness Agreement. Canberra: Productivity This paper assessed the short-term impacts of the National Commission; 2022. Housing Accord using ARIMAX time-series forecasting. [6] Australian Government Department of the Treasury. The results suggest that building approvals have shown Delivering the National Housing Accord. Canberra: Treasury; an upward trend since the Accord’s implementation, compared with the flat trend forecasted by the ARIMAX 2023. model. However, this positive effect remains within the [7] Hyndman R. J., Athanasopoulos G. Forecasting: Principles model’s 80% confidence interval, meaning it is not yet and Practice (3rd ed.). OTexts, 2021 statistically significant, and no definitive conclusion can [8] Lennert F. Time series modeling: Why ARIMA models beat be drawn. A key limitation is the short timeframe, as the linear regression for temporal data. Statistical Horizons; 2025. Accord has only been in effect for 10 months. This means [9] Hyndman, R. J., & Athanasopoulos, G. Forecasting: that all the effects of this policy are not exactly clear yet. Principles and Practice (3rd ed.). OTexts, 2021. In addition, it also restricts the analysis to building ap- [10] Saunders C., Tulip P. A model of the Australian housing provals, which, while a useful leading indicator, do not market. RBA Research Discussion Paper, 2019. directly measure increases in housing supply. Further lim- [11] Australian Bureau of Statistics. Producer Price Indexes, itations stem from the ARIMAX model itself and from the Australia. Canberra: ABS; 2025. choice and treatment of exogenous variables, which may [12] Mansur S. A., Abdul Hamid A. R., Yusof N. A. Rising trend have reduced forecast accuracy. It is recommended that in construction cost and housing price. Journal of Advanced this methodology be reapplied once the Accord has been Research in Business and Management Studies. 2020 Oct in place longer, allowing clearer effects to emerge and 23;3(1):94–104. enabling analysis of dependent variables such as hous- [13] Housing prices and rents in Australia 1980–2023: Facts, ing completions and commencements. Future research explanations, and implications. Applied Economics Working should also revisit the choice of exogenous variables, Paper. 2023 testing alternative specifications to improve model accu- [14] Ong V., Leishman C. The economics of housing supply: racy. Overall, this paper provides an early assessment of Key concepts and issues. NSW research report, 2024. the National Housing Accord’s short-term impacts using [15] NSW Government. HomeBuilder. Revenue NSW, 2025. ARIMAX time-series forecasting. It represents a first step [16] Devine G. Why rising apartment approvals are not in addressing the research gap on evaluating the Accord’s Australia’s housing crisis saviour. API Magazine, 2025. effects post-implementation. While no statistically sig- [17] Majka M. ARIMAX: Time series forecasting with external nificant impacts were observed in the first 10 months, the observed differences in trends suggest that the NHA may variables. LinkedIn. 2024. positively influence future housing supply, with effects [18] Mikusheva A., Sølvsten M. Linear regression with weak exogeneity. Quantitative Economics. 2025;16(2):367–403.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 by the authors.
This work is licensed under a Creative Commons Attribution 4.0 International License.
