• Summary
• Downloads
• Explanatory Notes
• Related Information
• Past Releases

Creating a Prototype Linked Employer-Employee Dataset, With Example Productivity Analysis

The Advanced Data Analytics section has recently completed work on a prototype linked employer-employee dataset (LEED), and performed preliminary productivity analysis using it. The aim of this project was to assess the feasibility and value to the ABS of potentially producing a full LEED in the future.

The prototype LEED was assembled by linking a number of tax datasets – Business Activity Statements, Business Income Tax, Pay As You Go tax records and Personal Income Tax – with the ABS’ Business Longitudinal Database (BLD) and Business Characteristics Survey records, which together contain detailed information on business characteristics. The Pay As You Go records in particular provided the link between employers and employees to connect their records deterministically. The prototype focussed on small and medium size BLD firms with an employee count between 1 to 199 employees for the 2010-11 financial year.

By linking survey and administrative data, a rich prototype dataset emerged, including firm level information on turnover, capital, and non-capital expenses, and employee level information on age, gender, income and occupation. Such linked employer-employee datasets can be used for a number of analytical purposes – for example, to produce labour market statistics on job flows, multiple job holders and employment tenure; or for firm level productivity analysis to incorporate employee characteristics.

As an example of one of these analytical uses, the prototype dataset was used to describe the characteristics of productive firms, in terms of firm and employee level attributes. Multilevel modelling was used to capture the contextual effects in the data, of employees nested within firms, and firms within industries. In particular, both a two-level model (firms in industries) and a three-level model (employees in firms in industries) were investigated to check whether results were consistent across them. The focus was on descriptive analysis rather than causal analysis, and a number of employee and firm characteristics were included. Results were largely in line with expectations, with characteristics like higher capital expenditure, higher operational expenditure, more permanent employees and more high-wage employees associated with more productive BLD firms.

The results of this project were presented at the Asia Pacific Productivity Conference in July, and the paper will be made available on the ABS web site. This project documented the methodological challenges to create a full LEED, provided clear evidence that it would have a number of important analytic uses in answering policy questions, and proposed areas for future research.

Further Information
For more information, please contact Andreas Mayer (02 6252 7140, andreas.mayer@abs.gov.au) or Joseph Chien (02 6252 5917, joseph.chien@abs.gov.au)

The ABS Privacy Policy outlines how the ABS will handle any personal information that you provide to us.