Experimental monthly filled jobs

Project Description

 Give feedback Download information

Latest update

26 October 2016
Updated to include:

  • estimated totals for the latest reference month with a 10-week lag (July 2016)
  • 'actual totals' for September 2015 because it's the latest reference month associated with a lag of 12 or more months.

See all updates

Description

This experimental series is a measure of monthly labour market statistics. It uses simple linear regression to rate-up the incomplete monthly job totals, based on the historical relationship between complete and incomplete data.

The series will appeal to customers who are interested in changes and levels of monthly filled jobs. It provides an early indication of changes in the labour market that current measures do not; and is designed to complement existing labour market measures, not replace anything.

What you’ll see

You’ll see a nine-year back series of estimated monthly filled jobs by ANZSIC06 industry division and earnings for all industries combined. The time series is available in CSV files and Excel spreadsheets that can be downloaded.

You can download the following:

  • CSV files containing the time series for monthly filled jobs and earnings.
  • Excel spreadsheets containing graphs of the time series.
  • A technical paper that gives background information about the series. It includes details about the coverage, and makes comparisons with other Statistics NZ’s labour market outputs, methodology, assumptions, and limitations.

We’d like your feedback about

  • Whether the data meets your needs
  • Any concerns that would stop you using the data
  • Any other information you would like to have in the data.

In response to feedback, we aim to introduce improvements to the series in subsequent releases. To comment on the series, click on the ‘Give feedback’ button above. Ongoing publication of the series depends on interest from our customers.

Data sources

  • Job records are from the Employer’s Monthly Schedule (EMS) tax form.
  • Attributes of employers (eg industry coding) are from Statistics NZ’s Business Register.

Data quality

We are confident of the quality of the series because:

  • At least 90 percent of the total number of job records from the EMS is available at the time of the estimation process.
  • We’ve applied standard methods in the transformation of the data.
  • The regression model is the best-fit model, based on current data.

Monthly movements in the experimental series and a series of actual totals are generally similar. Almost 100 percent of the job records in the EMS are available within 12 months of the end of the reference month. We include a series based on this data in the CSV files and spreadsheets – for comparison with the experimental series.

Known issues

Note: We’ll add to this list as we hear of other issues.

  • A number of new employers in recent reference months haven’t been assigned an industry code, because of delays associated with updates to the Business Register. They are placed in the ‘not elsewhere included’ industry division. The number of jobs belonging to these employers is not large enough to bias estimates at the aggregate level.
  •  Large employers with partial EMS data can cause artificial monthly movements at the aggregate level. We haven’t implemented a method to detect and deal with them. An example of the effect is the dip in monthly jobs belonging to the information media and telecommunications industry in April 2012. 
  • Delays in processing the EMS data by Inland Revenue can significantly affect the receipt of data. This can result in underestimates of monthly totals. We haven’t implemented a process for monitoring the data supply and dealing with significant decreases in the receipt rate. Two examples are the periods: July 2007 to January 2009; and April 2012 to November 2012. The totals for months in these two periods were underestimated in the experimental series. The first period may have been affected by a temporary slowdown in EMS data processing, as a result of KiwiSaver being introduced in July 2007. Changes in tax codes for student loan borrowers, introduced in April 2012, may have affected processing in the second period.

26 September 2016
Updated to include:
 

  • estimated totals for the latest reference month with a 10-week lag (June 2016)
  • 'actual totals' for August 2015 because it's the latest reference month associated with a lag of 12 or more months.

25 August 2016
Updated to include:

  • estimated totals for the latest reference month (May 2016)
  • 'actual totals' for June 2015 because it's the latest reference month associated with a lag of 12 or more months.

25 July 2016
Updated to include:

  • estimated totals for the latest reference month (April 2016)
  • 'actual totals' for June 2015 because it's the latest reference month associated with a lag of 12 or more months.

27 June 2016
Updated to include:

  • estimated totals for the latest reference month (March 2016)
  • ‘actual totals’ for May 2015 because it’s the latest reference month associated with a lag of 12 or more months.

23 May 2016
Updated to include:

  • estimated totals for the latest reference month (February 2016)
  • ‘actual totals’ for April 2015 because it’s the latest reference month associated with a lag of 12 or more months.
  • upwards revisions to monthly filled jobs and earnings belonging to December 2015 and January 2016.

1. Previously published estimates for December 2015 and January 2016 were underestimated because of the exclusion of jobs belonging to a number of new employees. 

2. Attributes belonging to persons in the tax system had previously been updated quarterly, resulting in a delay in deriving the ages of new employees.

3. We are now updating these attributes every month, resulting in more timely inclusion of new employees aged 15 years and over.

4. This update has resulted in:

  • Upward revision of 1.2 percent and 1.8 percent respectively for filled jobs in December 2015 and January 2016
  • Upward revision of 0.4 percent and 0.7 percent respectively for earnings in December 2015 and January 2016

Revisions to monthly filled jobs by ANZSIC divisions are in the spreadsheet: Revisions to estimated monthly filled jobs - May 2016 release.xlsx

22 April 2016
Updated to include:

  • estimated totals for the latest reference month (January 2016)
  • ‘actual totals’ for March 2015 because it’s the latest reference month associated with a lag of 12 or more months.


Related initiatives

Experimental monthly series of unadjusted job totals

This experimental series is made up of counts of job records from the Employer Monthly Schedule (EMS) tax forms that are available six weeks after the end of the reference month. It doesn’t include adjustments that account for the contribution from records that are available if the lag is greater than six weeks.

 

Page updated 26 October 2016

Project Details