September 01, 2018

BERL Data Newsletter - September 2018

The BERL data newsletter has been designed to provide commentary on data issues and news, including methodological changes, new measures, reviews, delays, and other technical issues and news.

This newsletter covers:

  • New migration measure
  • Changes to the international visitor survey

New migration measure

On 26 August 2018 Immigration Minister Iain-Lees Galloway and Customs Minister Meka Whaitiri, announced that come November 2018 filling out a departure card would become a thing of the past. The cards are being scrapped because they are no longer needed for their original purpose, which was to account for all passengers crossing the New Zealand border. In fact their only current purpose was to capture information for Statistics New Zealand to produce migration and tourism statistics.

In May 2017 Statistics New Zealand introduced a new outcome based measure of migration based on the actual travel histories of people crossing the border into and out of New Zealand. With the removal of the departure cards, this measure will become the new measure of permanent migration into and out of New Zealand.

So what does this mean for users?

  • A need to recognise that the new timely data series is an estimate and that this data will be revised several times prior to the data being finalised 17 months later.
  • The new data series may not pick up changes in trends as quickly as the previous intentions measure.
  • Statistics NZ stated that it will take five months to classify at least 80 percent of crossing (departure and arrivals) with certainty.

The new outcomes measure uses New Zealand immigration and customs data to pinpoint the actual date of arrival into New Zealand for inward migration, or departure data out of New Zealand for outward migration. For inward migration Statistics New Zealand can then determine if the migrant stayed in New Zealand for at least 12 months within a 16 month period after they first arrived in the country. If this condition is meet the migrant is classified as a permanent migrant arrival otherwise they are classified as a visitor.

For outward migration Statistics New Zealand can determine if someone leaving New Zealand remains out of the country for at least 12 months over a 16 month period. Again if they meet this condition they are classified as a permanent migrant departure, otherwise the person is classified as an overseas traveller.

Because it takes 16 months from a person’s arrival or departure date to determine the classification of each visitor and permanent migrant, the new outcome measure while more accurate is not very timely with data only able to be released 17 months in arrears. To overcome this issue of timeliness Statistics New Zealand have developed a probabilistic predictive model that will estimate the number of permanent migrants and visitors based on the available characteristic variables of travellers. This estimate will then be revised once the actual number of migrants and visitors is known 17 months later.

Using the new outcomes measure Statistics New Zealand was able to determine that the previous measure which used peoples stated intention on the length of stay was undercounting the number of permanent arrivals and permanent departures. According to Statistics New Zealand a comparison of the two measures between October 2014 and March 2015 showed that under the new outcome measure permanent arrivals were 2.0 percent higher, permanent departures were 2.7 percent higher, and net migration was 1.3 percent higher than what had been stated under the previous intentions measure.

The full paper on the new method can be accessed here.

Changes to the International Visitors Survey

Statistics New Zealand at the end of July 2018 completed their review of the International Visitors Survey (IVS). Statistics New Zealand’s review assessed the sample design of the survey; the questionnaire design of the survey; survey operations; governance of the survey; and the dissemination of the survey.

The outcome of the reviews has been 10 recommendations for the Ministry of Business, Innovation and Employment (MBIE) to implement to improve the accuracy and robustness of the IVS data, and to improve the dissemination and explanation of the results of the survey.

The 10 recommendations for MBIE are:

  1. Improve the survey design documentation and metadata.
  2. Adjust the survey design to real-world changes.
  3. Improve the sample allocation and respondent recruitment processes.
  4. Improve the online questionnaire.
  5. Improve editing and imputation systems and monitoring.
  6. MBIE, Statistics NZ and Kantar TNS work better together.
  7. Improve governance of the survey.
  8. Improve engagement with stakeholders and customers.
  9. Improve explanation of the survey findings.
  10. Assess the combined effect of the recommended improvements to the survey processing system before deciding to revise historical series.

So what does this mean for users?

  • That in the future the accuracy of the IVS data will be improved as the training of survey interviewers is improved, and the sampling allocation for the survey is implemented correctly. It turns out that far too many Australians and far too few Chinese visitors were being surveyed, with interviewers not being set any targets or quotas around main markets; that survey interviewers were only being trained on the job by their more experienced colleagues; and that the high usage of chat-mail over email was leading to low response rates from Chinese and other Asian visitors.
  • Improved access and usability of the IVS microdata from MBIE. The review has recommended that MBIE improve the metadata provided alongside the IVS microdata.
  • Changes in the questions in the IVS. The review has recommended that MBIE review the survey questions every two years to ensure the questions take into account real world market changes.

The full paper can be accessed here.