Key findings
- The GDP growth ‘nowcast’ published by the Reserve Bank of New Zealand – Te Pūtea Matua in the Monetary Policy Statement refers to the official projection for the first quarter following the last available datapoint. This Note explains the dynamic factor model which joins the suite of models that the Reserve Bank utilises to nowcast GDP growth in New Zealand.
- The accuracy of the GDP growth nowcasts from the dynamic factor model is comparable to that of the Reserve Bank’s official projections. At times during the assessment period, the Reserve Bank's projection outperforms the dynamic factor model nowcast, while at other times, the dynamic factor model performs better.
- The Reserve Bank’s nowcasts of GDP growth perform substantially better during the COVID-19 era, highlighting the crucial role of expert judgement.
- This Note serves as the technical background material for the Kiwi-GDP dashboard on the Reserve Bank’s website.
Why we did this research
Macroeconomic management involves decision-making that relies on economic datasets which are often incomplete. The frequency and timing of data releases differ, with some data subject to significant delays, substantial revisions, or both. The Reserve Bank of New Zealand – Te Pūtea Matua faces similar challenges, especially with the National Accounts data for gross domestic product (GDP) and its components. Stats NZ releases the GDP figures two and a half months after the reference quarter.
To address the delay in the availability of GDP data, the Reserve Bank relies on a broad range of high frequency indicators as well as qualitative survey data and has developed various tools and models to nowcast and forecast economic activity. Economists call this process of assessing today’s economic situation ‘nowcasting’.
This Note presents the performance of a dynamic factor model (DFM) – part of the suite of models that the Reserve Bank uses to address challenges related to nowcasting quarterly GDP growth.
The analysis presented in this Note is in line with the monetary policy research agenda of the Reserve Bank (theme: forecasting and modelling).
What data have we used?
The DFM is estimated using monthly and quarterly data over the period January 2000 to December 2024. The table below summarises the data series.