infonews.co.nz
INDEX
TRADE

Investigating how overseas merchandise trade data can change after publication

Thursday 1 December 2011, 4:33PM

By Statistics New Zealand

299 views

All of our published overseas merchandise trade (OMT) statistics are provisional for the first three months after publication. This allows us to include late data and make changes as new information becomes available; the data is final when published for the fourth time. We recently reviewed the changes to OMT data during the provisional period, to help users understand how the published versions of data may change. This page outlines how the four versions of OMT data can vary.

Key findings

We found evidence to confirm our expectations that the first release of total exports data is robust. Total imports data remains stable after the second release of data (after one revision). This is because in the first release of imports data, crude oil values are estimated. In the second month after release (version 2) we include actual crude oil values in the published imports data (see the 'Imports' section below). Changes in subsequent versions of data can be considered as refinements rather than significant adjustments.

About overseas merchandise trade data

OMT data provide information on the import and export of merchandise goods between New Zealand and other countries. OMT statistics are published monthly and made available to users as soon as possible, based on a pre-determined monthly release schedule. Statistics are generally published 18 working days after the reference month.

OMT statistics published by Statistics New Zealand are provisional for the first three months after publication and are subject to change during that period. See Why overseas merchandise trade data can change for more information about how and why our published trade data may change.

Overseas trade indexes published by Statistics New Zealand use OMT data and overseas merchandise trade data is used to calculate parts of New Zealand's balance of payments and gross domestic product. These statistics are published based on version 2 of the OMT data.

Investigating the level of change between versions of trade data

In 2011, Statistics NZ investigated how much change there is between the four published versions of OMT data. We investigated:

  • where most of the change occurs
  • whether the level of change is acceptable
  • whether we can shorten the provisional period.

Our analysis focused on data about the value of exported and imported goods. We did not investigate how quantities series or any other data dimensions may change between versions.

The analysis was based on 12 months of published exports and imports data at the chapter level of the New Zealand Harmonised System Classification (we used data from June 2009 to July 2010). We analysed all four versions of data for each month.

Exports and imports show different patterns

Our analysis revealed differences in the patterns of change between data versions for exports and imports.

Exports

The value of total exports was consistent across the 12-month period. The greatest change between versions was 0.5 percent. The smallest average change in value of total exports occurred between version 1 and version 4 and the largest between version 1 and version 2, showing that a slight increase was offset by a slight decrease in later versions.

Our analysis of exports showed high levels of change between versions for some chapters of the New Zealand Harmonised System Classification. Generally, these were chapters in which New Zealand did not export much and accounted for small proportions of total exports.

Conversely, there were low levels of change between versions in the chapters that include high-value exports, such as dairy and meat. Between version 2 and version 3, 97.7 percent of all changes were no greater than 1.0 percent, indicating that exports data at the chapter level does not change significantly from version 2.

Imports

We found that total imports data changed more than total exports data, in particular between first release and version 2. The differences between version 2 and version 3, version 3 and version 4, and version 2 and version 4 were all 0.5 percent or less. When we compared version 1 with version 2, and version 1 with version 4, we found differences ranging from 0.5 percent to 10 percent. There were no differences greater than 10 percent.

The high level of change in total imports data between version 1 and version 2 was due to changes in the data in chapter 27 of the NZHSC, Mineral fuels, mineral oils and products of their distillation; bituminous substances; mineral waxes, of which a key component is crude oil. When overseas merchandise trade statistics are first released, we estimate the value of crude oil imports. To do this, we use actual quantities of crude oil imports and country-of-origin data (provided by the New Zealand Customs Service, based on information from the refinery at Marsden Point), together with estimated prices. We update the statistics the next month when we receive import documents for actual shipments. This process causes the data in this chapter to vary more than other chapters between version 1 and version 2.

The analysis of imports by chapter showed high levels of change between versions in some cases. Other than chapter 27, which showed significant change between version 1 and version 2, most of the chapters with high levels of change were chapters in which New Zealand imports low levels of goods. These chapters accounted for a small proportion of total imports. The lowest level of change in imports by chapter occurred between version 2 and version 3. The percentage change between these versions was no greater than 1.0 percent in 98.5 percent of cases.

Changes in data after first release are not significant

Overall, we consider the level of change to be acceptable. The first release data is robust (with the exception of crude oil imports, where initial values are estimated and expected to change after the first release of data). Changes in subsequent releases can be considered as refinements to the data, rather than significant adjustments.

No plans to shorten the provisional period

Currently data is not final until four months after the first release of data. This investigation has confirmed that the majority of change to exports and imports data occurs between the first and second versions. Data does not change significantly after the second release. This suggests that we could shorten the provisional period. However, Statistics NZ has no plans to make such a change at this time.

More information

For more information please contact overseastrade@stats.govt.nz.

Overseas merchandise trade statistics provide statistical information on the importing and exporting of merchandise goods between New Zealand and other countries.