Australia (2)

Title of the survey

Survey of Average Weekly Earnings (AWE)

Organization responsible

Australian Bureau of Statistics (ABS)

Periodicity of the survey

Quarterly.

Objectives of the survey

The primary aim of this survey is to provide reliable quarterly estimates of levels and movements in average weekly earnings for Australia, states and territories. To achieve this aim, employers are asked to provide the total regular weekly earnings (including overtime) paid to employees and the number of employees involved. The major users of the results of this survey are Commonwealth and state government agencies, employers' and workers organizations. The AWE is essential for monitoring the Australian economy and formulating economic policy. It is also used for legislative or statutory purposes.

Main labour topics covered by the survey

Employment and earnings.

Reference period

Any part of the last pay period ending on or before the third Friday of the middle month of each quarter (i.e. February, May, August and November).

Coverage of the survey

Geographical

The whole country.

Industrial

All branches of economic activity, except: agriculture, hunting, forestry and fishing; the Australian permanent defence forces; private households employing staff; and overseas embassies, consulates, etc.

Establishments

All types and sizes of establishments which have paid employees.

Persons

All employees in the branches of economic activity covered. Employees based outside Australia are excluded. Excluded are the following persons who are not considered as employees: directors who are not paid a salary; proprietors and partners of unincorporated businesses; self-employed workers such as subcontractors, owners-drivers, consultants and persons solely paid by commission without a retainer; and unpaid family workers.

Occupations

All occupations are covered, but data are not collected by occupation.

Concepts and definitions

Employment

Employees are all wage and salary earners who received pay for any part of the reference period. All permanent, temporary, casual, managerial and executive employees are included. Part-time and casual employees who may have worked only a few hours during the reference period are included. Employees on paid leave (e.g. annual leave, long-service leave); employees who commenced or terminated during the pay period and employees on workers' compensation who continue to be paid through the employer's payroll are also included. Casual employees who work on an irregular basis and who were not paid during the relevant pay period, employees on leave without pay, on strike or stood down without pay for the whole of the pay period, and employees on workers' compensation who were not paid through the payroll, are excluded. Also excluded are home workers and workers sub-contracted from other companies. The following categories of workers are separately identified:

Earnings

Data are collected separately on total gross weekly earnings (including overtime earnings) and weekly overtime earnings. Data on total gross weekly earnings are collected separately for full-time adults, by sex, and all other employees, by sex. Data on weekly overtime earnings are collected for full-time adult employees only. Total gross weekly earnings include ordinary and overtime earnings, penalty payments, shift and other allowances; payments for leave (i.e. one week's proportion of pay in advance for leave taken during the reference period, excluding leave loadings); commission and retainer; payments under profit sharing schemes usually paid each pay period; incentive and piece-work payments; all workers' compensation payments paid through the payroll; and bonuses that are paid regularly. Excluded are: back pay and pay in advance; severance, termination and redundancy payments; leave loading; bonuses and other payments that are not paid regularly each pay period (e.g. Christmas bonuses); allowances which are reimbursements to employees for travel, entertainment, meals and other expenditure incurred in conducting the business of their employer; commissions where a retainer is not paid; imputed value of fringe benefits; fringe benefits tax; workers' compensation payments not made through the payroll. Total gross weekly earnings refer to weekly earnings paid during the reference period to all permanent, temporary, casual, managerial and executive employees, prior to taxation and any other deductions (e.g. superannuation, board and lodging). If the pay period is fortnightly or monthly, the responding establishment is required to include only one week's proportion of earnings. Weekly overtime earnings refer to that part of total gross weekly earnings paid for hours worked in excess of award, ordinary or agreed hours of work for full-time adult employees (separately for males and females).

Wage/salary rates

Not relevant.

Hours of work

Not relevant.

International recommendations

The ABS definition of earnings complies with the international recommendations, with the following exceptions: the value of payments in kind (such as food and drink, fuel, clothing, etc.) and the value of fringe benefits such as cost-of-living or dearness allowance, house rent allowance, transport allowance or family allowance paid directly by the employer are excluded from the data collected.

Classifications

Industrial

Data on employment and earnings were classified according to the Australian Standard Industrial Classification (ASIC-1983). ASIC is convertible to the International Standard Industrial Classification of All Economic Activities (ISIC), Rev. 2, 1968 at the four-digit level. From September quarter 1994, the survey data have been classified according to the Australian and New Zealand Standard Industrial Classification (ANZSIC), 1993. This classification is convertible to the ISIC, Rev. 3, 1990 at the four-digit level.

Occupational

Not relevant.

Others

The survey data are classified by state and territory, sector (public or private), size of establishments, and according to the following characteristics of workers: full-time adults (by sex) and all other employees (by sex).

Sample size and design

Statistical unit

The statistical unit used to compile AWE statistics across all industries is termed the state-management unit. It comprises all activities of a business within a particular state or territory. Each unit is classified to an industry which reflects the predominant activity of the business in the state or territory. Generally the sampling unit and the reporting unit are the same. However, in some instances, the reporting unit can differ from the sampling unit: (i) sampling units can be split into two or more reporting units based on predominant industry classification (these are the industry dissected units); (ii) sampling units can be split into reporting units based on a location of payroll data.

Survey universe / sample frame

The survey sample is selected from a common statistical frame, the ABS Business Register, which covers all sectors of the economy and most forms of entities engaged in economic activity. It includes companies, sole proprietorships, self-employed professionals, all government agencies, community service organizations, religious organizations, etc. However, businesses are usually only included in the register if they have paid employees, with one major exception: agriculture, where establishments are included if they have an estimated value of agricultural operations of at least 5,000 dollars, even if they have no paid employees. The register is regularly updated to take account of new businesses, takeovers and mergers, changes in industry classification, changes in employment and businesses which have ceased operations. Major sources for updating the information on the register include Australian Taxation Office, group employer registrations, unit surveys carried out by the ABS Business Register, updating of large enterprise groups by the ABS Business Register and feedback from the ABS collection areas. The unit recorded on the ABS Business Register and used in creating the population framework for the AWE, is called the management unit. This unit is the highest level accounting unit within a business, having regard for industry homogeneity, for which accounts are maintained. In nearly all cases, it coincides with the legal entity owning the business (company, partnership, trust, sole proprietorship, etc.). The operations of the management unit are dissected by state, using location and establishment ASIC and employment information stored on the ABS Business Register, to form state-management units (SMUs). The SMUs are then used to compile AWE statistics. The framework population consists of all management units on the Register at a point of time, except for those sectors of the economy which are excluded from the scope of the survey, as noted under Coverage of the survey. The management units are then replaced with management-state-industry units. The framework population is approximately 600,000 units. This includes units which are in scope and found to be defunct or have permanently no paid employees.

Sample design

The AWE survey is a single stage, single phase, stratified sample, the sampling unit being the SMU. Statistical units are stratified by state, public and private sector, industry and employment size. An equal probability sample is selected from each stratum, using a selection technique called synchronized sampling. This technique has the following attributes: The sample consists of approximately 5,000 selected units which include units within the scope of the sample framework which are found to be defunct or permanently have no employees. The survey is conducted quarterly and the sample of business units is updated each quarter to reflect changes in the ABS Business Register. In order to reduce respondent burden, especially on the small businesses, it is desirable to have planned sample rotation. In the AWE, there is presently a planned rotation of 5 per cent sampled units per quarter. Sampled units are generally no longer than five years in the survey. In future, it is planned to increase the rate of sample rotation from five years to three years.

Field work

Data collection

The survey is conducted by mail by the ABS, using permanent staff in all aspects of survey collection, compilation and publication. The questionnaires are mailed to the respondents about 14 days before the survey reference date and are due to be returned to the ABS within two to three weeks after the reference date.

Survey questionnaire

This consists of three main sets of questions on: An additional question enables the respondent to include, where relevant, comments on the reasons why the figures for a given quarter are significantly different from those for the previous quarter. Instructions are provided along with the questionnaire, which cover: In November of each year, more detailed breakdowns of data on earnings and additional information on hours of work are collected through this survey.

Substitution of sampling units

There is no substitution of sampling units in the case of total non-response. A sampled unit which is a non-respondent unit is considered to be not in the sample for the current survey, though still in the population.

Data processing and editing

All incoming questionnaires are initially scrutinized for completeness. The data are then edited, using automatic input-output editing processing for consistency and accuracy. The editing system involves the following edits: logical, physical, historical and ratio. Whenever a large change is detected or data are deemed inconsistent, the system identifies the record for further editing and verification by the clerical system. Follow-up letters and telephone calls are used for non-respondents. Each unit receives only one mail reminder, which may be followed up by one or more telephone reminders. If the reference period used for data collection presents abnormal circumstances (e.g. the operations of the sampled unit were temporarily disrupted by an industrial dispute, breakdown, fire), the respondent is required to provide details from the previous normal pay period. However, if the disruption is expected to be long-term, the respondent is required to provide data for the reference pay period and to comment on any unusual factors affecting the information supplied. No other adjustment is made by the ABS.

Types of estimates

Estimates of average weekly earnings are derived by dividing estimates of weekly total earnings by estimates of employment. Statistics are compiled by sector, state, industry level, for various categories of employees.

Construction of indices

None.

Weighting of sample results

Estimates of levels of earnings are derived using the ratio estimation method, within each stratum, with ABS Business Register employment data as the benchmark, for all strata where the benchmark is known and non-zero. In the remaining strata, the number raised estimation method is used. Ordinary-time level estimates are the difference between gross estimates and overtime estimates. Estimates of averages are calculated as the quotient of the estimate of totals at the appropriate level, and employment aggregated to the appropriate level.

Adjustments

Non-response

In the sample sector, variances on estimates for strata with non-responding units are calculated by the appropriate variance formulae using the number of respondents as the sample size. In the complete enumeration (CE) sector, benchmark variables are the total number of employees and total weekly earnings and the ratio estimation method is used in the presence of non-response.

Other bias

Every effort is made by ABS to ensure that any bias (due to changes in industrial classification, representativeness of the sample, etc.) is reduced to a minimum. No significant biases have been identified for specific correction action.

Use of benchmark data

See Weighting of sample results. Final estimates for the AWE are derived using the ratio estimation method with ABS Business Register employment data as the benchmark.

Seasonal variations

Major AWE published series are seasonally adjusted. The method used is the X-11 ARIMA package from Statistics Canada, which is a variant of the US Bureau of Census Method II seasonal adjustment programme. The seasonal factors are reviewed annually to take account of each additional year's original data. Trend estimates: Trend estimates are published with the seasonally adjusted series. They are calculated using a centred seven term Henderson moving average of the seasonally adjusted series. Estimates for the three most recent quarters cannot be calculated using the centre average method; instead, an asymmetric average is used.

Indicators of reliability of the estimates

Coverage of the sampling frame

Total coverage for all industries within the scope of the survey is aimed for.

Sampling error / sampling variance

Standard errors are calculated separately for levels and quarterly changes in earnings.

Non-response rate

The non-response rate is approximately one per cent.

Non-sampling errors

Inaccuracies may occur because of incomplete coverage of the underlying population, imperfections in reporting by respondents, errors made in recording, entry or processing of data, etc. Efforts have been made to reduce non-sampling error by careful design of questionnaires, detailed checking of returns and quality control of processing.

Conformity with other sources

The survey results are compared to the results of other surveys, taking into consideration the differences in methodology.

Available series

Published tables include statistics of average ordinary time earnings and total earnings by sex and for all employees, along with estimates of quarterly changes, also by sex and for all employees, by sector, states and territories, in original, seasonally adjusted and trend terms.

History of the survey

The AWE survey was introduced in September 1981 and replaced the average weekly earnings series based on payroll tax data which dated back to the early 1940's. Since then, it has been conducted on a regular quarterly basis. Since 1981, additional data items have been collected in the November survey, each year. This additional data collection was discontinued after the November 1991 survey.

Documentation

Australian Bureau of Statistics: Average Weekly Earnings, Australia, Preliminary (catalogue No. 6301.0) (quarterly; Canberra); published approximately ten weeks after the survey reference period. idem: Average Weekly Earnings, States and Australia (catalogue No. 6302.0) (quarterly; ibid.); published approximately 13 weeks after the survey reference period. This publication also contains computations of standard errors. For more detailed data on earnings and hours of work collected in May of each year, see: idem: Distribution and Composition of Employee Earnings and Hours, Australia (catalogue No. 6306.0) (annual; ibid.). All publications contain some methodological details. For methodological information, see also: ABS Information paper: Average weekly earnings: new series to replace former payroll tax based series (catalogue No. 6336.0, March 1982); idem: Information Paper; New Statistical Series: Employment, Average Weekly Earnings, Job Vacancies and Overtime (catalogue No. 6256.0, June 1984); contains background information on the developments of this sample survey. idem: Statistics on Wages, Earnings, Income and Labour Costs - A guide to their concepts, measurements and usage; Occasional paper No. 1986/1 (ibid.). In addition to the data that appear in national publications, the ABS can also make available information which is not published. For example, statistics of average weekly earnings classified by industry are available on a special data service. This information may be made available in one or more of the following forms: microfiche, photocopy, data tape, floppy disk, computer printout or clerically extracted tabulation. Generally a charge is made for providing unpublished information.

Confidentiality / Reliability criteria

The following rules are applied to test whether data in a confidential cell (considered to be important to explaining the overall statistical result) should be published or otherwise released: (i) the Australian Statistician can exercise the right to release confidential data relating to General Government (e.g. an official body other than a Government business undertaking), except where the Minister responsible for the body has shown such disclosure would be likely to enable the identification of that particular organization; (ii) confidential data relating to other enterprises (including Government business undertakings) are subject to different rules. The respondent(s) contributing to the confidential cell are approached, through a formal legal process, to obtain permission to publish. If permission is given, the authority to publish lasts three years, at the end of which the authority must be renewed. Under no circumstances does the ABS publish confidential data in this category without the permission of the respondent.

Other information

Data supplied to the ILO for publication

Statistics of average hours paid for (including overtime) and average hourly earnings, derived from the November survey of the AWE, up to 1989 and 1990 respectively, are published in Tables 11 to 15 and 16 to 20 of the Yearbook of Labour Statistics.