Stats NZ

Employment indicators: Paid jobs week ended 17 May 2020

Employment indicators provide an early indication of changes in the labour market.

This new experimental weekly series aims to provide an indicator of employment and labour market changes in a more a timely manner than the monthly series.

The weekly employment indicators use the timelier and more detailed payday filing that has been available from Inland Revenue since April 2019. Our experimental series includes three measures that trade-off between quality and timeliness.

Due to the nature of the administrative data that these indicators draw from, the accuracy of the data improves the further out from the reference week it relates to. Stats NZ advise strong caution in making decisions based on this data.

See About the data for a data quality statement and information about the series.

Key facts

The most accurate measure, ‘Number of paid jobs – 27 days’, shows that in the week before the alert level 4 lockdown began and non-essential businesses closed there were over 2.216 million paid jobs. This fell to 2.108 million in the last week of alert level 4, before rising back to almost 2.199 million jobs in the week ending 17 May 2020.

The more timely, but less accurate, measures indicate that number of paid jobs continue to rise throughout May and dip again into the first week of June.

Created with Highcharts 5.0.14NumberNumber of paid jobs, by timeliness measure (6, 13, and 27 days), week ended 5 May 2019–7 June 20206 days delay from reference period13 days delay from reference period27 days delay from reference period2019-05-052019-05-122019-05-192019-05-262019-06-022019-06-092019-06-162019-06-232019-06-302019-07-072019-07-142019-07-212019-07-282019-08-042019-08-112019-08-182019-08-252019-09-012019-09-082019-09-152019-09-222019-09-292019-10-062019-10-132019-10-202019-10-272019-11-032019-11-102019-11-172019-11-242019-12-012019-12-082019-12-152019-12-222019-12-292020-01-052020-01-122020-01-192020-01-262020-02-022020-02-092020-02-162020-02-232020-03-012020-03-082020-03-152020-03-222020-03-292020-04-052020-04-122020-04-192020-04-262020-05-032020-05-102020-05-172020-05-242020-05-312020-06-071 250k1 500k1 750k2 000k2 250kStats NZ

Number of paid jobs, by timeliness measure (6, 13, and 27 days), week ended 5 May 2019–7 June 2020

Created with Highcharts 5.0.14NumberNumber of paid jobs, by timeliness measure (6, 13, and 27 days), week ended 5 May 2019–7 June 20206 days delay from reference period13 days delay from reference period27 days delay from reference period2019-05-052019-05-122019-05-192019-05-262019-06-022019-06-092019-06-162019-06-232019-06-302019-07-072019-07-142019-07-212019-07-282019-08-042019-08-112019-08-182019-08-252019-09-012019-09-082019-09-152019-09-222019-09-292019-10-062019-10-132019-10-202019-10-272019-11-032019-11-102019-11-172019-11-242019-12-012019-12-082019-12-152019-12-222019-12-292020-01-052020-01-122020-01-192020-01-262020-02-022020-02-092020-02-162020-02-232020-03-012020-03-082020-03-152020-03-222020-03-292020-04-052020-04-122020-04-192020-04-262020-05-032020-05-102020-05-172020-05-242020-05-312020-06-071 250k1 500k1 750k2 000k2 250kStats NZ
Week ended6 days delay from reference period13 days delay from reference period27 days delay from reference period
2019-05-051,438,1601,982,7702,077,690
2019-05-121,542,7501,916,3902,057,780
2019-05-191,550,1801,966,6502,070,140
2019-05-261,631,2101,987,8002,105,540
2019-06-021,615,7702,019,6502,116,310
2019-06-091,482,6301,969,9602,086,090
2019-06-161,626,4002,005,6802,093,440
2019-06-231,601,9602,008,2902,103,890
2019-06-301,640,8302,006,1602,101,300
2019-07-071,474,1901,949,2402,076,690
2019-07-141,584,6901,962,2802,085,500
2019-07-211,608,9502,022,5302,109,350
2019-07-281,663,2602,044,8702,127,260
2019-08-041,662,9402,060,9902,127,860
2019-08-111,589,6301,988,3102,111,260
2019-08-181,627,3902,028,9902,128,220
2019-08-251,688,2102,071,0902,132,180
2019-09-011,690,2202,056,4102,132,270
2019-09-081,541,5901,979,9402,086,820
2019-09-151,623,7801,978,6502,104,200
2019-09-221,625,8902,042,5702,104,400
2019-09-291,663,8302,019,0602,117,740
2019-10-061,672,8702,038,6102,136,830
2019-10-131,607,5502,043,7702,144,200
2019-10-201,686,7002,038,2802,150,750
2019-10-271,661,3202,074,8102,157,800
2019-11-031,704,9502,100,5502,185,350
2019-11-101,594,6102,077,3402,180,650
2019-11-171,608,6502,087,4102,202,640
2019-11-241,669,1302,135,2402,234,590
2019-12-011,728,5302,155,6502,239,060
2019-12-081,629,0402,150,1402,212,250
2019-12-151,731,2202,026,5402,233,900
2019-12-221,703,6401,982,3102,212,580
2019-12-291,516,4202,001,2302,124,990
2020-01-051,624,3902,039,6302,120,980
2020-01-121,616,5402,011,8402,119,150
2020-01-191,688,3102,080,1202,162,320
2020-01-261,735,7902,103,3802,184,850
2020-02-021,722,4102,138,5402,197,140
2020-02-091,550,4801,993,1402,186,250
2020-02-161,612,7702,056,0402,137,410
2020-02-231,607,7002,006,8802,076,410
2020-03-011,737,3102,139,5002,219,710
2020-03-081,626,7302,090,0802,202,400
2020-03-151,679,9502,094,5602,195,650
2020-03-221,651,0102,156,6802,215,910
2020-03-291,737,6702,076,4302,207,880
2020-04-051,528,6401,991,6502,179,730
2020-04-121,320,6301,925,4302,127,110
2020-04-191,553,5801,940,5102,093,520
2020-04-261,570,9001,935,3402,107,540
2020-05-031,583,6501,983,4702,192,520
2020-05-101,513,0301,970,3302,174,890
2020-05-171,605,9602,101,9702,189,500
2020-05-241,740,6902,152,470
2020-05-311,791,0402,144,070
2020-06-071,604,130

About the data

Data quality statement

Data in this release are produced with minimal processing and are raw job counts for the respective weeks.

These are not official statistics, but an effort to provide the latest available employment data at a time of heightened interest. These indicators use administrative data, so while they may provide early indication of changes in the labour market, they are also significantly influenced by Inland Revenue’s and employers’ operational practices. Data has not been seasonally adjusted, and there has been no imputation or methodological adjustments to account for quality issues relating to operational practices.

See Payday filing  for more information.

Definitions and metadata

About the three measures in the weekly series of paid jobs:

  • Number of paid jobs – 6 days
    This measure counts number of paid jobs using payday filing from the first six days after the reference week ends. This measure has been shown to capture on average 75 percent of the total paid jobs for the period, although this measure does exhibit large fluctuations varying between 65–80 percent. This is the most timely but poorest quality indicator.

  • Number of paid jobs – 13 days
    This measure counts number of paid jobs using payday filing from using payday filing from the first 13 days after the reference week ends. This measure has been shown to capture on average 93 percent of the total paid jobs for the period, although this measure does exhibit large fluctuations varying between 87–96 percent.

  • Number of paid jobs – 27 days
    This measure counts number of paid jobs using payday filing from the first 27 days after the reference week ends. This measure has been shown to capture on average 98 percent of the total paid jobs for the period but can vary between 95–99 percent. This is the least timely but highest quality indicator. 

About industry data

The CSV download includes a breakdown at a broad industry level for each of the timeliness measures. The industry classifications are:

  • Primary industries – ANZSIC06 divisions A (Agriculture, Forestry and Fishing) and B (Mining)
  • Goods-producing industries – ANZSIC06 divisions C (Manufacturing) to E (Construction)
  • Services industries – ANZSIC06 divisions F (Wholesale Trade) to S (Other Services)
  • Total – All ANZSIC06 divisions, including those that could not be classified.

Paid jobs are classified at the enterprise level and thus reflect the industry of the employer, not the nature of the employee’s work. Additionally, not all enterprises can be matched to an industry, so the total paid jobs is greater than the sum of paid jobs of the three industry breakdowns.

FAQs for weekly employment indicators

What is the reference period for paid jobs in the weekly employment indicators?
The reference week is from Monday to Sunday – the dates on the graph refer to the end of the week (Sunday).

Who is covered in the weekly employment indicators?
Workers with wages or salaries taxed at source, employees on paid leave, non-New Zealand residents, armed forces, and self-employed with wages or salaries. Workers who were on paid leave, or receiving the wage subsidy, during the COVID-19 lockdown would be included in the count, however workers who were on unpaid leave would not. Contractors are excluded.

This number represents the number of unique employer-employee relationships and does not directly show the number of people in paid employment, since one person can work for more than one employer.

What is the data source?
Payday filing data is from employers via Inland Revenue. Since April 2019, employers have reported PAYE tax data to Inland Revenue each payday, rather than monthly.

Payday filing was not adopted by all employers until August 2019, when use of the old Employers Monthly Schedule finally ceased. This means there is a slight undercount in the number of paid jobs between April and August 2019, although no measures have been taken to account for this.

How often will this data be updated?
We will update each series, each week on a Thursday, until the end of October 2020. At which point we will review whether the series will continue. This series is designed as a temporary product in response to requests for more timely information to support the COVID-19 response.

Why are there three different values for each reference week?
We have created a snapshot of what the data would have looked like each week, based on the data that would have been available 6 days, 13 days, and 27 days after that week ended. This allows for quality comparisons between the different measures.

What changes are not reflected in the data?
This series only counts the number of unique employer-employee relationships, and does not measure the volume of work, hours worked, or earnings associated with each job. For example, if the number of employees stay the same, but the number of hours worked or wages received change, this won’t be reflected in the data.

Why did you include the more timely, lower quality series if Stats NZ advises against their use?  
We have included all three measures in response to stakeholder engagement. Different customers have different wants and needs for the data. Some customers would prefer more timely data and are willing to accept the lower quality that comes with the timeliness.

Why does the quality reduce as timeliness increases?
Employers have up to 10 working days after the end of the pay period to submit the PAYE data for that period. While most employers pay their staff weekly or fortnightly, some pay their staff on a monthly basis. This means that in some cases Inland Revenue won’t receive information about a pay period for several weeks or even months after the pay period ends. The longer we wait to produce the indicator, the more information we receive from employers and therefore the more reliable the estimates.

Analysis of the historical data from the two more timely measures shows significant volatility, that is. week to week changes predominantly reflect differences in operational practices rather than actual changes in the labour market.

Why don’t these numbers match the monthly employment indicators (MEI) numbers?
Our monthly employment indicators are also produced from PAYE data, but they use more robust statistical processes (for example, estimates for business non-response) to improve the quality of the data before it is released. The MEI indicators are higher quality but are less frequent than the weekly series. The MEI also provides more detailed information across industries and regions and has a longer back series dating back to April 1999.

Will future updates include more detailed analysis?
We have prioritised getting the data out as quickly as possible. This comes at a trade-off with the level of analysis and detail that we are able to provide. We expect our customers will do their own more detailed analysis on the data.

We intend to explore the possibility of providing additional cuts of the data, for example a regional or gender breakdown, and will explore the possibility of gross earnings indicator in future releases. We will prioritise the order of development based on customer feedback.

What’s next?
We are looking forward to feedback from our customers on this dataset, seeing their needs and finding out how we can meet them. As this is an experimental series, we are happy to answer questions about the data and would appreciate feedback to help us make improvements.

ISBN 978-1-98-858391-4

Next releases

Employment indicators: Paid jobs week ended 24 May 2020 will be released on 25 June 2020.

Employment indicators: May 2020 – Infoshare tables will be released on 29 June 2020.

Related topics
Related releases