Weekly payroll jobs and earnings methodology
The ABS is committed to delivering timely and accurate information about the impacts of the coronavirus (COVID-19) on Australian society and the economy, through producing a range of existing and new statistical products.
The new Weekly Payroll Jobs and Wages in Australia publication is one of these and aims to provide weekly changes in the number of jobs, changes in wages and average weekly wages by jobs at the national, state and territory and industry level by selected personal attributes, including sex and age group. 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. Over 99% of large employers (20 or more employees) and 85% of small employers (those with 19 or less employees) are reporting through STP.
The ABS needed to develop methods for imputation, estimation and quality assurance and were assisted by staff from the ATO and the Treasury in compiling the statistics and producing the publication.
The method for deriving weekly estimates from the STP data is based on the method used by the UK Office for National Statistics for producing statistics from similar payroll data. The STP data are reported on a cash basis, i.e., the time when payment was made rather than earned. To produce timely estimates of weekly jobs and earnings, the STP data is converted from a cash basis to an accrual basis. The conversion involves calculating the payment frequency and a daily pay rate for each job. Weekly aggregates are then derived from the daily data.
The STP payment data that are extracted for a specific week cycle are always incomplete due to varying payment and reporting frequencies. For example, a business with a monthly payroll will only provide STP data every fourth week. We developed a method to impute for the missing payment data ensuring that reliable statistics could be produced on a weekly basis. The method assumes that if an employee has not yet had payment data reported and they have not been flagged for termination, their payment status is consistent with their previous reporting record and their previous calculated daily rate is used for the current period. We also developed a strategy for revising estimates when the latest STP data becomes available.
The methods for imputation and treatment of unusual cases was initially kept simple due to the limited development time for the first publication. Several enhancements have since been identified to further improve the measurement quality of labour market changes and reduce the magnitude of the applied revisions. These include adjusting for Fringe Benefit contributions on an accrual basis, adjusting for short-term growth or decline in the pay rate for imputed jobs, incorporating treatments for new jobs, and accounting for irregular payments. These proposed method enhancements are currently being assessed using historical data and will be implemented once proven to be effective and robust.
For more information, please contact Summer Wang at firstname.lastname@example.org.