Weekly Payroll Jobs and Wages in Australia methodology

These weekly estimates are part of a suite of new products the ABS are releasing to provide up-to-date information on the impact of the COVID-19 coronavirus on people and businesses in Australia. This release provides indicative information on the economic impact of the COVID-19 coronavirus on employees, including changes in payroll jobs, changes in total wages paid, and changes in average weekly wages per job.

These weekly estimates are derived from Single Touch Payroll (STP) data, which is provided to the Australian Taxation Office (ATO) by businesses with STP-enabled payroll or accounting software each time the business runs its payroll. STP data includes both business and job level tax information and superannuation information. The data are combined with other administrative data from the Australian taxation system to determine additional classification attributes, such as the age and sex of employees.

The ABS would like to acknowledge the critical support from the Australian Taxation Office (ATO) in enabling the ABS to produce these statistics.

Payroll jobs reported to the ATO through STP are in scope of these estimates.

A payroll job is a relationship between an employee and their employing enterprise, where the employee is paid in the reference week through STP-enabled payroll or accounting software and reported to the ATO. Where an employee is paid other than weekly, the established payment pattern is used to identify jobs in weeks outside the payment week. For further information, please see 'Imputation' below.

Owner managers of unincorporated enterprises are not included in these estimates as they are not in scope of STP-enabled software reporting to the ATO.

Approximately 99% of employers the ATO classifies as ‘substantial employers’ (those with 20 or more employees) are reporting through STP. Small employers (those with 19 or less employees) began transitioning to STP on 1 July 2019 and over 80% are reporting through STP. As a result, not all jobs in the Australian labour market are captured within these estimates.

The estimates are considered experimental and use an index to provide an indication of movements in total jobs and total wages (rather than level estimates), for the purpose of assessing the economic impact of the COVID-19 coronavirus on employees.

For context, high level analysis suggests that there were approximately 600 000 fewer payroll jobs in STP-enabled businesses on 27 June 2020 than on 14 March 2020.

These estimates will be published fortnightly on Tuesdays during the COVID-19 coronavirus period. The timing has been optimised around when the largest proportion of both weekly and fortnightly payments are made (thereby minimising the contribution from imputation at the end point of the series).

These estimates have a 17 day time lag after the reference week. While estimates are published fortnightly, weekly data are available within the release.

These estimates are available at the state and territory and national level. All geographic variables are based on the individual's residential address as stated on their income tax return. Statistical area 4 (SA4) estimates are updated in alternate releases for payroll jobs only. Where an address is not available from the income tax return the residential address as reported on the STP file is used.

Estimates are also provided for the 19 Australia and New Zealand Standard Industry Classification (ANZSIC) divisions, with selected sub-division estimates updated on alternate releases for payroll jobs only. All industry variables are based on the ABS Statistical Business Register.

STP data is reported on a year-to-date cash accounting basis at the time the payment was provided to an employee rather than at the time the payment was earned. To produce indicative real time estimates of wages and jobs, the calendarisation method is used to convert STP data from a cash basis to an accrual basis.

The calendarisation method uses the start and end date of the payment period reported through STP to calculate the payment frequency, known as ‘periodicity’. A daily pay rate is then calculated by dividing the total payment by the total number of days in the payment periodicity. For example, a payment made weekly will be divided by seven to calculate a daily pay rate. Where the commencement or termination date for a job occurs within a payment period, the periodicity for that job is adjusted to exclude the days before commencement or after termination.

As new businesses commence reporting through STP, the ABS determines when the business appears in the dataset for the first time and distributes the year to date wages across all past payment periods for the financial year. This ensures that weekly estimates are not an artefact of increased up-take of businesses onboarding to STP, however can cause revisions to earlier parts of the time series.

Payment data extracted from the STP system each week is always partially incomplete, given businesses have different payment frequencies. For example, a business with a fortnightly payroll will only provide STP data every second week. To produce reliable weekly statistics, an imputation method is used to account for differing reporting patterns.

If an employee has not yet had payment data reported and they have not been flagged for termination, it is assumed that their payment status is consistent with their previous reporting record. The previous calculated daily rate will be imputed for the current period.

The data indicates that 38% of employees are paid weekly; 47% are paid fortnightly; 11% paid monthly; 4% are paid quarterly or paid infrequently.

No imputation is required for employees paid weekly. Business reported data is included for all employees paid weekly due to the 17 day lag between the reference week and the release of estimates. Actual payments for most employees paid fortnightly and some employees paid monthly are also available.

Imputation is not applied for the small proportion of employees who are paid quarterly or infrequently as there is not an established pattern of payment to extrapolate forward.

No imputation is applied for new employees without historical payment information, until a pattern can be determined.

Once STP data is converted to an accrual basis using the calendarisation method and imputation is applied, the data is aggregated to produce the following weekly estimates:

• Total Payroll Jobs: The average of the 7 days of employee counts.
• Total Wages Paid: The sum of all daily rates for all jobs for the week. This is the total value of wages for the week.
• Average Weekly Wage by job: This is calculated by dividing the total wages value by the total payroll jobs number.
   

The estimates are presented as an original series only. Seasonally adjusted and trend estimates are not yet available. A number of years of data will be required before seasonal patterns can be observed and adjusted for.

The calendarisation and imputation methodologies applied to the estimates account for calendar related variations, such as the number of days in a month, and different payment frequencies.

Note: All estimates presented in this publication are presented for weeks ending on a Saturday.

Estimates are produced for weekly change in the number of payroll jobs, change in wages paid and change in average weekly wage by job. These estimates are available at the national, state and territory and ANZSIC division level by selected personal attributes, including sex and 10 year age group. These estimates are presented as percentage change movements and indexes.

SA4 and ANZSIC sub-division level estimates are updated on an ad hoc basis for changes in payroll jobs only.

The estimates include indexes to present changes in the labour market during the COVID-19 coronavirus period.

In order to compare changes over time, the week Australia recorded its 100th confirmed coronavirus case (i.e. week ending 14th March 2020) is used as the reference period for constructing the indexes and given an index value of 100.0.

Index values relative to the reference period are updated in each release to reflect revisions to underlying data. Index values commencing week ending 4 January 2020 are available via Table 4 in the Data downloads section to support users to undertake additional longer time series analysis. It is important to note the high degree of seasonality in the start of the series, which reflects the seasonal trend of a proportion of people joining or returning to employment following the Christmas and New Year holiday season. ABS analysis focuses on the COVID-19 period, which is not affected by a high degree of seasonality.

The indexes allow comparison between two points in time. The points in time can be adjacent (this week and the previous week) or many weeks apart. Movements in the index from one period to another can be expressed as either points or percentage change and these are rounded to one decimal place.

The following example illustrates the method of calculating changes in index points and percentage changes between any two periods:

These indexes differ from the ABS' suite of price indexes, including the Wage Price Index, which measure changes in price over time unaffected by quality or quantity and should not be directly compared.

A payroll job is a relationship between an employee and their employing enterprise, where the employee is paid in the reference week through STP-enabled payroll or accounting software and reported to the ATO. Where an employee is paid other than weekly, the established payment pattern is used to identify jobs in weeks outside the payment week.

Weekly estimates are derived from STP, which is provided to the ATO by businesses with STP-enabled payroll or accounting software. While STP effectively supports employer reporting obligations and ATO operational requirements, it is not designed primarily to support statistical purposes. As a result, it is important to note some of the inherent limitations of the data and how the ABS uses statistical methods to best address these limitations in producing information on jobs and wages over time.

Given the high frequency and volume of STP reporting, all data underlying job and wages estimates are revised for each release, and reflected in updated percentage change movements and indexes. Revising underlying data across the time series allows the estimates to incorporate newly available business reported data and replace instances of temporarily imputed data with actual data.

Also, as new businesses commence reporting through STP, the ABS determines when the business appears in the dataset for the first time and distributes the year to date wages across all past payment periods for the financial year. This ensures that weekly estimates best reflect change in the labour market over time, rather than changes in the coverage of businesses. This approach also necessitates revisions across the time series.

Wages are subject to a higher degree of reporting variability and revisions than jobs, reflecting a greater degree of seasonal variation in payments, and changes in working hours and overtime. While the ABS accounts for employees being paid with different frequencies, there are points in the year when additional reporting activity is more likely to occur, which may flow through to published estimates.

Firstly, care should be exercised when focusing on the most recent weekly movements in wages, as these will reflect a higher degree of reporting variability and contribution from imputation.

Secondly, reporting variability is higher at the end of the financial year, especially in the final weeks of June, as some accrued earnings may be paid (or reported) as a lumped amount. This makes it more challenging to apportion some payments to the weeks they were actually earned, in a timely manner.

In preparing for this release, the ABS observed higher than usual week-to-week changes in wages indexes in June, particularly for the Health care and social assistance industry. The ABS applied a smoothing treatment to changes in wages in this industry through June, to minimise the influence of lumped payments on the latest estimates. The lumping of payments in the unadjusted data may relate to fringe benefits tax and other end of financial year payments that are more common in this industry. Receipt of more complete data for the end of the financial year, combined with further analysis of end of year payments, may result in larger than usual revision in subsequent releases. Further information on this will be provided in the next release.

Reduced wages through May 2020, which are more pronounced in industries that were particularly impacted by COVID-19, may reflect a combination of:

• Reduced wages being paid by COVID-19 affected employers, who may also have reduced hours over the period - particularly for jobs that weren't eligible for JobKeeper support;
• Seasonal changes in wages; and
• Potential payroll reporting changes for some employers during May which are not yet able to be identified in the data, pending further payments. This could include changes to payment frequency or payment categorisation, both of which could affect the imputation methodology and how payments are effectively apportioned onto an accruals basis.
   

The first 'JobKeeper' payments were received by employers in the first week of May 2020. When employers use these payments to pay employees, it is included in wages indexes. The initial payroll deadline for the JobKeeper program was 8 May. Backpay to employees reported by businesses in early May has been converted, as best as possible onto an accruals basis and reflected in revised wages estimates through April. Further information on the treatment of JobKeeper in ABS economic statistics can be found in Economic measurement during COVID-19: Selected issues in the Economic Accounts, May 2020 (cat. no. 5261.0).

The gender information in the underlying data includes jobholders who were identified as a gender other than ‘male’ and ‘female’, or where no gender was specified. These data are not included in the calculation of male and female estimates but are included in all other calculations. It is not possible to produce earnings measures for non-binary populations, at this stage.

STP data does not include information on hours worked or hours paid for. Analysis of monthly hours worked can be found in Labour Force, Australia (cat. no. 6202.0) and the detailed releases (cat. no. 6291.0.55.001 and 6291.0.55.003). STP data are also unable to account for job attachment where a payment has not been made, where a jobholder was temporarily stood down without pay.

The estimates do not include information on the employment status of employees, i.e. full time or part time. This will be explored as part of ongoing work between the ABS and the ATO. In the interim, this information is also available within the suite of Labour Force releases.

STP data is supplied by the ATO to the ABS under the Taxation Administration Act 1953, which requires that such data is only used for the purposes of administering the Census and Statistics Act 1905. Any discussion of data limitations or weaknesses is made within the context of using the data for statistical purposes, and is not related to the ability of the data to support the ATO's core operational requirements.

Legislative requirements to ensure privacy and secrecy of this data have been adhered to. In accordance with the Census and Statistics Act 1905, results have been confidentialised to ensure that they are not likely to enable identification of a particular person or organisation.

The ABS would like to acknowledge the critical support from the Australian Taxation Office (ATO) in enabling the ABS to produce these statistics.

Weekly Payroll Wages and Jobs in Australia provide weekly changes in the total number of payroll jobs and total wages at the national, state and territory and ANZSIC industry level by selected personal attributes of jobholders, including sex and age group. Statistical Area 4 (SA4) and industry sub-division estimates are also available on alternate releases for changes in payroll jobs only.

The estimates are compiled using Single Touch Payroll (STP) data provided to the Australian Taxation Office (ATO) by businesses with STP-enabled payroll or accounting software each time the business runs its payroll. Approximately 99% of substantial employers (20 or more employees) and over 80% of small employers (those with 19 or less employees) are reporting through STP.

The STP data are reported on a cash basis (the time when the payment was made) rather than an accrual basis (the time when the payment was earned). Therefore to produce real time estimates of weekly jobs and earnings, the STP data need to be converted from a cash basis to as close to an accrual basis as possible, using a calendarisation method.

In addition to the cash reporting, the STP payment data that are extracted for a specific week cycle will be incomplete due to different payment and reporting frequencies. For example, a business that normally pays its employees on a fortnightly basis will not have any payment data available in every weekly STP dataset. To produce reliable statistics on a timely basis, an extrapolation imputation method is applied to account for these different payment and reporting habits.

The method includes the following steps:

• The periodicity, or payment frequency, is calculated using the start and end date of the payment period;
• A daily pay rate is then calculated by dividing the total payments by the payment frequency. For example, weekly pay is divided by seven;
• Where the start or termination date for a job occurs within the payment period, an adjustment is made to the periodicity for that job to exclude the days before commencement (or after termination).
   

This method allows the data to be aggregated and analysed for different reporting periodicities as all the records are broken down to a common frequency (daily).

The STP payment data that are extracted for a specific week cycle are always partially incomplete due to varying payment and reporting frequencies. For example, a business with a monthly payroll will only provide STP data every fourth week.

The data indicates that 38% of employees are paid weekly; 47% are paid fortnightly; 11% are paid monthly; 4% are paid quarterly or infrequently.

To produce reliable statistics on a weekly basis, an imputation method is used to account for differing reporting patterns, with revisions made when additional data become available.

If an employee has not yet had payment data reported and they have not been flagged for termination, it is assumed that their payment status is consistent with their previous reporting record. The previous calculated daily rate will be imputed for the current period.

No imputation is required for employees paid weekly. Business reported data is included for all employees paid weekly due to the 17-day lag between the reference week and the release of estimates. Actual payments for most employees paid fortnightly and some employees paid monthly are also available.

Imputation is not applied for the small proportion of employees who are paid quarterly or infrequently as there is not an established pattern of payment to extrapolate forward.

No imputation is applied for new employees without historical payment information, until a pattern can be determined. This means that there is an inherent and unavoidable lag before new payroll jobs will appear in the data, after the initial pay period, particularly for new jobs with employers who have less frequent payment and reporting periods.

Once STP data is converted to an accrual basis using the calendarisation method and imputation is applied, the data is aggregated to produce the following weekly estimates:

• Total Payroll Jobs: The average of the 7 days of payroll job counts. A payroll job is a relationship between an employee and their employing enterprise, where the employee is paid in the reference week through STP-enabled payroll or accounting software and reported to the ATO. Where an employee is paid other than weekly, the established payment pattern is used to identify jobs in weeks outside the payment week.
• Total Wages: The sum of all daily rates for all jobs for the week. This is the total value of wages for the week. For employees who work for the whole week in a specific job, the weekly wages are the sum of 7 days of the daily rate. For the employees who work partially for the whole week in a specific job (i.e. who just started the work or terminated the work during this period), the weekly wages are the sum of the actual periodicity for that job excluding the days before commencement (or after termination).
• Average weekly wage by job: This is calculated by dividing the total wages value by the total payroll jobs number.
• These estimates are aggregated to the national, state and territory and ANZSIC division level by selected personal attributes, including sex and 10-year age groups.
   

The data underlying these estimates are revised for each release, and reflected in percentage change movements and indexes.

Revising underlying data across the time series allows the estimates to incorporate newly available business reported data and replace previously imputed data with actual data.

As new businesses commence reporting through STP, the ABS determines when the business appears in the dataset for the first time and distributes the year to date wages across all past payment periods for the financial year. This ensures that weekly estimates best reflect change in the labour market over time, rather than changes in the coverage of businesses. This approach will result in revisions across the time series.

For more information on this methodology please email labour.statistics@abs.gov.au

The ABS has recently released experimental estimates of changes in payroll jobs and changes in wages in Weekly Payroll Jobs and Wages in Australia. The estimates are compiled in near real time and published fortnightly, using Single Touch Payroll (STP) data provided to the Australian Taxation Office (ATO) by businesses with STP-enabled payroll or accounting software each time the business runs its payroll.

This information provides complementary insights to the Labour Force statistics on employment. While changes in payroll jobs can provide an indication of changes in how employed the population is, it is important to remember that the measures are different.

The table below highlights key differences that should be considered when comparing changes in payroll jobs with changes in employment.

For further information, please email labour.statistics@abs.gov.au

Weekly Payroll Jobs and Wages in Australia estimates are presented as ‘original’ data series. The release does not include the other standard time series types of seasonally adjusted or trend data that are found in other labour statistics releases (eg. Labour Force).

Generally, three to five years of data are required before good seasonally adjusted data can be produced. Since Single Touch Payroll (STP) is a relatively new program (with most large employers moving to the platform during 2018-19 and smaller employers through the second half of 2019), it is not yet possible to produce seasonality adjusted series (with seasonal elements removed) or trend series (with both the seasonal elements and irregular fluctuations removed).

This means that some of the change in a Weekly Payroll Jobs and Wages data will not necessarily reflect changes in the economy due to COVID-19. Some changes may also reflect seasonal changes in the labour market, particularly in wages. For example, during April 2020 there was a number of public holidays and school holidays that may have impacted on the level of wages during this period.

A particularly high degree of seasonal change can be seen in the payroll jobs and wages data for January and February. Summer is a period of pronounced seasonality in Australia, with considerable labour market activity before Christmas, and a combination of public holidays, school holidays and lower business activity in the period after Christmas.

Similar patterns of seasonal change can be seen in other labour statistics, such as monthly Labour Force statistics, which can provide important context for interpreting changes in payroll data, until seasonally adjusted data become available.

An increase in both payroll jobs and wages was observed between January and March 2020. This pattern was broadly consistent with seasonal increases in both the number of total employees and weekly hours worked that are observed each year in the original Labour Force series.

The following two graphs compare changes in payroll jobs and wages series with original data from Table 8 and Table 9 of Labour Force, Australia, Detailed - Electronic Delivery, April 2020 (cat. no. 6291.0.55.001). They show a similar seasonal changes in the early period of each year in monthly Labour Force data (with 2019 and 2020 presented) and related Weekly Payroll Jobs and Wages data (for which data are only available for 2020).

It is also important to note that the increase in wages and jobs between January and March is not a result of more employers reporting through STP. The methodology used in Weekly Payroll Jobs and Wages in Australia controls for changes in employer coverage to ensure that the index series are as comparable as possible over time.