Acting Second Commissioner Law Design and Practice, Kirsten Fish
Transfer Policy Institute's A-Life conference – Opening remarks
18 March 2021
Thank you Bob and welcome everyone to the inaugural A-Life conference. What a fantastic line up of speakers you have in store for you including Professor Claus Kreiner from the University of Copenhagen joining to deliver the conference keynote! I’m delighted to be here and to see so much interest in ATO data and its research potential.
Over recent decades, there has been growing recognition of the immense value in government data and increasing calls for Governments to be doing more with that administrative data for public policy research.
Back in 2010, The Henry review on Australia’s Future Tax system recommended that: “Governments should make available data and publish analyses of taxes and transfers to inform and encourage community debate about the performance of the tax and transfer system.” (page xxiv A Future Tax System – Part 1)
Then in 2015, the then Prime Minister released the Australian Government Public Data Policy Statement. That statement acknowledged that “[the] data held by the Australian Government is a strategic national resource that holds considerable value for growing the economy, improving service delivery and transforming policy outcomes for the Nation.” The government recognised “the importance of effectively managing this national resource for the benefit of the Australian people” and committed to optimising the use and reuse of public data and collaborating with the private and research sectors to extend the value of public data for the benefit of the Australian public. Through that statement, Government agencies committed to making high value data available for use by the public and academia, and to publish appropriately anonymised government data by default.
Data tells us how people are behaving. It’s policy in action and it is something that as administrators of the tax system we see everyday. Of course in the ATO we are very proud of our long history of bringing these insights to the attention of policy makers to help inform robust and sound decision making. For over 100 years the ATO has published Taxation Statistics and in more recent times has made routine releases of other non-sensitive data on data.gov.au.
A-Life is our commitment to build on these efforts.
The story of A-Life
The creation of A-Life was motivated by the establishment of the Tax and Transfer Policy Institute in 2013. As most of you may know, the Institute was established to enable more empirical analysis and evaluation of Australia’s tax and transfer system. It aims to foster a richness and diversity in tax and transfer policy research, exploring issues and practical solutions to the critical tax and transfer policy challenges facing governments.
So, a new ATO team, pioneered by Thomas Abhayaratna, was established to explore the possibility of making more data available for this purpose. Specifically, the intention was to provide a high-quality longitudinal sample file for policy research. This marked the beginning of A-Life.
From the outset, the project was based on four principles:
- To make a high-quality panel dataset available for research
- To ensure the data was representative of the tax paying population
- To make the data as easy as possible to use; and above all
- To ensure that taxpayer information remains private and confidential.
As we developed the initial version of the file, it was really important to us to ensure it was fit for purpose for the end users that would be using the data for research. In this, the engagement and involvement of academics was invaluable.
The team worked closely with Department of Treasury Sir Roland Wilson scholars. Shane Johnson developed a 'Tax Simulator' which played an important role in helping us detect and fix problems in early versions of the data. And Nathan Deutscher pioneered what we intend to be a future A-Life product – the ‘A-Life: Family Links’ file which will enable intergenerational analysis. Working with these and other scholars from the outset has added an invaluable academic lens to the development of the file and has helped us to achieve a high quality dataset that is research ready.
Other academics have assisted our project in other ways, such as Professor Michael Martin from the ANU. As a professor in statistics and the Chair of ANU’s Human Research Ethics Committee, Michael provided important feedback and input into the A-Life sampling methodology, as well as advice that helped to guide decisions on the A-Life governance arrangement – such as the need for researchers to seek ethics approval from their university as part of the project application process.
Like any project, this one has faced its share of challenges. To provide some insight, A-Life provides hundreds of variables for analysis. Looking all the way back to 1990, the vast number of policy changes overlayed with internal database upgrades that an organisation such as ours undertakes over such an extended period meant that simply tracking and mapping these variables over time was an initial challenge.
Another challenge was writing code, or data extraction rules, that provide an accurate account of all tax records. The team tackled this, applying a methodical iterative process and a suite of integrity checks to identify and correct errors and assure the accuracy of the data.
Finally, maintaining privacy and confidentiality is top of mind for everything that the team (and the rest of the ATO) does, and the team has implemented numerous assurance mechanisms. This included obtaining an expert opinion and Risk of Reidentification Assessment by the Australian Institute of Health and Welfare – experts in individual level data for research - to ensure the A-Life files strikes the optimal balance between our obligations under the Tax Administration Act and making data available for public policy research.
Outcome
The result of these efforts and the ATO’s investment in A-Life is not only a dataset that is accurate and ready to use, but one that is much better suited to academic research. This marks a quantum leap forward for public policy research in Australia.
The recent December 2020 A-Life update makes data available to 2018. This means that we now have almost 30 years of high-quality longitudinal data available. A-Life offers a longer period relative to other commonly used Australian datasets we see used in policy research. It also offers a higher degree of statistical precision to robustly analyse sub-groups of interest. In the 2018 year alone, A-Life has almost 1.5 million individual records. And with over 300 variables generated from the income tax return and superannuation systems, A-Life provides detailed information on income tests and other information required for policy analysis and evaluation that is not available in this form elsewhere.
This is quite a remarkable achievement for Australia.
Future/next steps
With A-Life, we have a great foundation that we will continue to build upon.
I would like to acknowledge and thank the group of researchers who have been enthusiastic ‘early adopters’ of A-Life. We recognise that your time is valuable, and that a bit of a ‘leap of faith’ is required as an early adopter for new sources of evidence like A-Life.
To date, 22 Australian universities have signed up to the A-Life Heads of Agreement. This has led to approvals for almost 80 researchers who are examining a wide variety of policy related questions.
And we see that early A-Life research has already produced valuable insight. The projects employing this data range from evaluation of specific policy, examining income distribution and dynamics, labour market analysis, intergenerational mobility, taxpayer bunching, and retirement income analysis.
This includes the real, important and practical contribution this research has made informing the Government’s recently published Retirement Income Review (published on 20 November 2020). A-Life provided longitudinal analysis that revealed how different income cohorts contribute and draw down superannuation savings over time, the longitudinal analysis enabled by A-Life provided new perspectives on this issue. Aided by A-Life analysis, the detailed Review will influence the tone of retirement income policy debate for decades to come.
Over the next two days we are excited to hear from the 14 presenters to learn more about their projects.
For those who are yet to use A-Life, we really encourage you to do so and hope that the presentations over the next two days help seed some ideas. We hope that the momentum generated to date will continue as more researchers use and trust A-Life as a high-quality source of evidence.
To this end, the ATO will continue to invest in A-Life and will do what we can to make more data available where possible. And we are committed to continuing to ensure that the data is accurate and fit for purpose.
As the A-Life project matures, I am hopeful that the frequency of published papers on Australia’s tax and superannuation systems will continue to increase, and that it will reduce our reliance on research that is conducted in other countries – and, importantly, on conclusions which may not necessarily hold in the Australian context.
Finally, I’d like to thank Professor Robert Breunig and his team from the ANU for their work in making this conference happen. The conference is not only a great milestone for A-Life, but a great milestone for public policy research in Australia.
I hope you enjoy participating in the conference. Thank you.
Speech by Acting Second Commissioner Law Design and Practice Kirsten Fish to the Tax and Transfer Policy Institute's A-Life conference, 18 March 2021.