Covid has accelerated data-driven policymaking
Why is it so hard to integrate data insights into policymaking - and what can be done to achieve greater leverage from existing data assets, while maintaining and building citizen trust?
This was the focus of the opening international plenary panel discussion of the Government Data Show today, an online conference bringing together over 500 UK public sector data professionals. The event runs until 23rd September.
"Policymakers just want to know that the data will deliver value," noted Deborah Anton, Interim National Data Commissioner with Australia's Department of the Prime Minister and Cabinet. "Covid has made government rely upon data for decision-making, and our experience is that it has reinforced why it's good."
Deborah was appointed as the Interim National Data Commissioner in 2018, where she has led the development of reforms to improve the use of Government data to deliver public benefit. This has culminated in the Data Availability and Transparency Bill, which aims to modernise government data sharing to improve government service delivery, inform government policies and programmes and enable greater research and development.
One example she shared in the 45-minute discussion was the Australian Tax Office's 'Single Touch Payroll' system, which was able to give policymakers real-time insight into what was going on in the economy: "When ministers see this, then they want more - and then we in the public sector need to lift and do more."
One key impediment to 'do more' with data is the tricky issue of interoperability, where Australia's federal structure meant that State-developed bushfire warning apps used different data points to determine whether residents should stay or evacuate, and didn't consistently measure the severity and extent of the fires.
"Standards failed us with bushfires with each state government having their own app. This comes back to whether policymakers can agree to move their positions - common standards matter if you want good policy outcomes. To have a national picture on family violence the data has to be consistent across jurisdictions. So you have to have leadership and people's willingness to move to somewhere different from where they are now."
Progress is being made, Deborah argues, pointing to regular data and digital minister meetings across Australia, with attention now on the value of commercial off the shelf solutions in driving the interoperability agenda, rather than bespoke platforms that create new obstacles to pooling data.
"There's a growing belief that when Government uses data, it is bad. This is obviously extremely problematic and we need to urgently try and change this mindset," said Ott Velsberg, Chief Data Officer of Estonia.
Ott argues that the governance structures are in place - with a head of profession for government data users, and a graduate data network to support professional development. What needed to change to build trust with policymakers is how data is used across the public sector.
"We have changed our strategy of how we use data," explains Ott. "Previously it was technical, now it's about asking the right questions and getting agreement from all participants. Our role now, as data professionals, it to provide central support at every level possible, while helping the business side to talk through the questions it needs answers for, and identify the problem they're trying to solve."
"To earn trust from policymakers we have to earn it from citizens, and that comes from empowering people so that they understand when their data is being processed by Government," he continued. "GOV.UK makes it very clear how your Government processes your data. Here in Estonia we are shortly launching our consent management service to share government held data with third parties."
"If something goes awry with our cybersecurity and digital ID, we are transparent about it and the risks that arise. We need to change the mindset that systems should be working 100% of the time. Ultimately it's about putting people in charge of their own data, and try to be as transparent as possible."