Modelled Territorial Authority GDP methodology

Source data

Our experimental Modelled Territorial Authority Gross Domestic Product (MTAGDP) series is based on the earnings (wages and salaries) of people in employment by industry and territorial authority area.

The data is attained from a custom Annual LEED (Linked Employer-Employee Data) dataset provided by Statistics New Zealand (Stats NZ).

The methodology

The MTAGDP series uses Statistics New Zealand's (Stats NZ’s) published and custom Regional GDP series for 15 regional council regions as a benchmark to estimate GDP for New Zealand’s 66 territorial authority (TA) areas and the industries in them.

These estimates have been derived by a mathematical method called iterative bi-proportional matrix balancing (also known as RAS) that allocate the earnings data to align with the official Regional GDP totals.

As LEED earnings data is not available for the final year in the MTAGDP series, a hierarchical forecasting procedure is employed to generate the TA estimates for that year. Previously, the industry breakdown was not available for this final year (as it is not provided in the official Regional GDP series), but for the first time we make it available in this release.

Note that the MTAGDP estimates for the final year are provisional figures rather than final estimates and may be subject to revision in future releases.

2024 methodology revision

The experimental MTAGDP series was first developed in 2015. In this release we have made an update to the methodology to improve the accuracy and consistency of the estimates. This revision involved the following changes to the method of producing these estimates:

• further exploration into forecasting methodologies
• additional clarity around the components of the proportional fitting process

Consequently, historical MTAGDP estimates for some territorial authorities (TAs) and some industries have been revised in the post-revision releases. These revisions are due to the replacement of the two-stage raking methodology used previously, with the bi-proportional balancing (RAS) method following a standard matrix balancing methodology.

Remaining differences are due to standard annual revisions in the Regional GDP series.

More details of the MTAGDP 2024 methodology revision can be seen in the latest MTAGDP methodology document:

MTAGDP Methodology 2024 [PDF, 456 KB]

The previous MTAGDP methodologies can be seen in the documents below:

MTAGDP Methodology 2018 [PDF, 306 KB]

MTAGDP Methodology — prior to 2018 [PDF, 3.1 MB]

Why estimate MTAGDP

Although Regional GDP figures have been available from Stats NZ since 2013, there are a number of research questions and policy needs that require finer-scale information (ie, both in terms of the industry classifications and the spatial granularity).

This project was initiated in order to partially fulfil this need, providing an internally consistent, robust estimate of GDP at the Territorial Authority (TA) level.

How estimates compare to Regional GDP

Estimates of MTAGDP were calculated to align with official published statistics for Stats NZ's Regional GDP.

There are small differences between the population estimates in the MTAGDP and the Regional GDP. MTAGDP uses year-ended June population estimates while Regional GDP estimates year-ended March figures by taking the mean of two year-ended June figures, and benchmarking that to the New Zealand total for the year-ended March. These differences are small and have a negligible effect on GDP per capita trends and comparisons across industries or regions.

Quality control applied to estimates

Since the original proof of concept in late 2014 for MTAGDP estimates, we have made several presentations and held question and answer discussion sessions with the New Zealand Institute of Economic Research (NZIER) and the Stats NZ's National Accounts team. Our Business and Economic Development team has also validated and error-checked results.

Assumptions when using estimates

The base data for generating MTAGDP is the earnings (wages and salaries) of people employed by TA and industry, sourced from custom Annual LEED data provided by Stats NZ. This data is used on the assumption that the distribution of earnings is well correlated to the distribution of GDP by geographical area and industry.

How often MTAGDP is updated

We update the experimental MTAGDP series annually as soon as practical after Statistics New Zealand’s Regional GDP series is released at the end of March, typically June.

MTAGDP includes estimates for 16 regional council areas and 66 territorial authority areas for every year of the time series beginning 2000. The breakdown by the 18 industry categories (plus a further 4 sub-industry categories) are available for every year of the time series bar the latest year, and align with the industry groupings used in Stats NZ’s Regional GDP series.