Stats NZ is investigating the development of a monthly indicator of economic activity using administrative data (admin data) from Inland Revenue GST filings.
As an admin-data-first organisation, existing sub-annual economic indicators are based on GST filings supplemented with survey data for the largest and most complex businesses. Extending this hybrid approach to a monthly indicator would create substantial additional processing and survey burden. By adopting an admin-only approach, Stats NZ can produce more frequent and timely financial statistics with broad industry coverage.
This paper explores the key characteristics of GST data, and the limitations for producing timely, fit-for-purpose monthly statistics. Although GST is filed on a range of different frequencies, using just monthly and two-monthly (every two months) filers results in a useful indicator with strong coverage.
We constructed an experimental monthly series to determine both the inherent suitability of GST data on its own and to highlight any conceptual differences compared with other relevant Stats NZ outputs. We used a simple disaggregation approach to transform two-monthly GST filings into monthly data based on the behaviour of monthly filers in the same industry.
Download the document below, or read the summary of key points online.
Summary of key points
A summary of the key points of this paper follows:
- GST data has broad coverage across economically significant market activity and is already used in existing economic statistics. Monthly and two-monthly filers account for most reported GST value, and could provide a useful monthly signal with broad industry coverage.
- GST data records taxable transaction value rather than production or value added. A GST-based measure should therefore be interpreted as an indicator of activity, not as a direct measure of economic output.
- Stats NZ receives sufficient GST data on the ninth day of the second month following the end of a period, in most cases. This is typically 39 to 40 days after the end of each period.
- To capture most of the transaction value of GST, GST filed by businesses every two months also needs to be included in the data. This requires a method of apportionment.
- The experimental GST-based series showed that a coherent monthly signal can be derived from GST data, particularly at higher levels of aggregation and in industries where GST turnover is closely aligned with the output of businesses.
- Further work is needed before a production design can be recommended and implemented. This work includes decisions about the methodologies, treatments and trade-offs between quality assurance, treatment of capital transactions, and outliers, revisions, and seasonal adjustment.
ISBN 978-1-991431-39-4
