Data can fuel innovation, improve service delivery and underpin genuine evidence-based policy — but public servants aren’t remembering to make it available. A new 18-month roadmap and training from the Department of the Prime Minister and Cabinet aims to shift public service culture on data sharing.
A recent review of data management found “competing priorities and the lack of an overarching strategy” holding back the Commonwealth’s ability to turn data into public value. A more coordinated approach is needed for the innovation push to succeed, according to a cross-agency team that recently reported their findings.
Prime Minister Malcolm Turnbull’s Innovation Statement gets the ball rolling today with a new Public Data Policy Statement.
Public servants will also benefit from a new data skills training program run by PM&C.
The major change is all agencies will now have to make “non-sensitive high-value data” open by default. Most open data will be free of charge to access but fees will apply for “specialised data services”. According to the review, some agencies have been “unnecessarily” charging for data, to the frustration of researchers and other agencies trying to get at it:
“Many expressed a view that data collected by governments has already been paid for by taxpayers, and should be available free of charge wherever possible, particularly where the marginal cost of provision is low.”
The data management report criticised the lack of a strong mandate for the government to use and release public sector data. That should no longer be a problem but its authors also argue there is no “strong culture of publishing data to foster economic opportunities” in the Commonwealth.
The Australian Public Service does not have the necessary incentives, skills or organisational arrangements in place for this to change, they write. There is also no way to get an accurate picture of what data the government has as a whole, but it is clear there is a lot of it.
Data sharing between federal bodies and levels of government is held back by barriers, both real and perceived, according to the report, and it is the intersections between different sets of data where the biggest value is derived. Department of Health secretary Martin Bowles reflected on this exact issue in a recent speech:
“When I was first appointed I naively asked whether I could provide the Pharmaceutical Benefits Schedule and Medical Benefits Schedule data to our states and territories, I was given 100 pages of legal advice around privacy explaining why this was impossible.
Get the Juice - the Mandarin's free daily newsletter delivered to your inbox.You’ll also receive special offers from our partners. You can opt-out at any time.
“When I then said — I want to do this. How could we make it happen? I received just four pages of advice. They now have the data. You need to ask the right question.”
The review project was commissioned in April by the head of PM&C, Michael Thawley, who kept an eye on it together with Finance mandarin Jane Halton and then Communications boss Drew Clarke, who now runs the Prime Minister’s Office. One of Thawley’s deputy secretaries, Heather Smith, led a team of staff from the three departments as well as the Australian Bureau of Statistics and Australian Taxation Office.
With the government’s renewed focus on innovation and stimulating the science, technology, engineering and mathematics fields, they concluded “now is the right time to invest in data capability” to inform policy, improve service delivery and provide the valuable modern resource to Australian businesses.
“Sustained action and commitment was key” for other nations whose governments are creating more public value out of their data than Australia. New Zealand focused on using data to get more bang for the taxpayer’s buck, while the United Kingdom and United States looked first to what the private sector could do with open public sector data.
Smith’s team set out clear steps for ministers and public service leaders and said that in the meantime:
“A lot can be done now within existing policy settings, so it is therefore proposed that reform starts with several projects that demonstrate value, uncover barriers, and lead to better designed policy and services that improve people’s lives.”
Data management has to improve. In many cases, valuable data is hard to find, and even harder to share and leverage in new ways. There is also a strong public perception of risk around digital information, with good reason, so any changes have to align with community expectations as well as privacy legislation. Trust-building must be part of the push for change.
The report’s 15 recommendations are presented chronologically, sketching out an 18-month roadmap to unlock the value of public sector data which probably isn’t being followed to the letter, but doesn’t have to be. The first six months advocates going for quick wins through high-value projects, building external partnerships, boosting capability and publishing whatever data is easiest first. A whole-of-government data policy and governance framework, the centrepiece of the recommendations, would follow.
As well as making a lot of data open by default, the new public data policy provides the kind of clear direction that is needed in spades.
Notably, the report argued for “a requirement for evidence-based policy development to motivate data use in policy”.
Other elements of the mooted plan are a “trusted access model” for data sharing inside the APS, a whole-of-government data catalogue and a priority list to rank datasets according to public value. The final steps include published data management standards, a consistent regime for cost-recovery and efforts to optimise the legislative environment, possibly with help from the Productivity Commission.
There is praise in the report for APS data pioneers like the ABS, ATO and Digital Transformation Office, but points out that while data.gov.au has been established, it could have a lot more datasets. It only lists 6700 datasets currently — compared with 25,461 and 132,865 for its respective UK and US counterparts — over 75% of which are from just four agencies, less than a quarter can be accessed through APIs, and there are “big gaps” in areas like health, employment and socioeconomic data.
The review team found most agencies do not routinely release data to data.gov.au — “In consultations, many said it simply does not enter their mind to do so” — and a view that the open data repository is under-funded.
Turnbull’s Innovation Statement refers to ensuring data.gov.au is the “central place” and the “one-stop shop for government data” but does not mention any increase in resourcing.
Alongside the new data policy, the Innovation Statement announced $75 million in new funding for Data61, the new CSIRO research group that was established when National ICT Australia (NICTA) lost its federal funding and agreed to merge with the larger agency.
Data61 is charged with turning its data analytics expertise to helping the APS meet Turnbull’s expectations, by joining up the many disparate government datasets and publicly releasing them on open data platforms.
It is also expected to “improve industry cyber security and develop new cyber security architectures”, form a network of data researchers to link them with the business sector, and deliver data analytics training to the private sector. Its new funding takes effect on July 1.