People power: why digital tech demands very human attributes

Rosemary Huxtable: “Part of changing the trust equation is having responsive services where people feel that they’re at the centre of decision-making; that they’re involved in the design of policy."

To improve trust in government, civil servants must deploy digital tech to meet citizens’ needs. And as officials heard at Global Government Summit 2019, that requires an understanding of human psychology, shrewd decision-making, and great leadership

“We have a trust deficit, and we see that playing out in the public debate,” said Rosemary Huxtable, Secretary of Australia’s Department of Finance. And, addressing top officials from nine countries gathered in Singapore for the Global Government Summit, she argued that having an effective, high-quality public service is key to winning people’s trust.

“Every time a citizen engages with a government service, their expectations are either met or not met. At that moment, they’re coming to a view about their government; and that’s the politicians as much as the public servants,” she said. “So part of changing the trust equation is having responsive services where people feel that they’re at the centre of decision-making; that they’re involved in the design of policy. And ministers are just as interested in that as we are, if not more so.”

Speaking in the session on the cross-government management of operations and service delivery – ‘Unifying functionality: sharing data and establishing platform services’ – Huxtable explained that her government has identified a number of “signature initiatives where there’s a very strong connection with citizen trust and engagement,” and worked to build a sense of “shared responsibility” across government for successful delivery. These include reforms to welfare, veterans’ and business services, she added.

Change direction, not pace

Public servants, Huxtable continued, often respond to budgetary pressures by “doing everything in pretty much the way we always had, but running faster and working harder.” But this is unsustainable: “Now we’re beginning to look more fundamentally at our operating models, and how we work together and collaborate, and how we use technology to drive efficiency and effectiveness.”

This thinking produced Australia’s Digital Transformation Strategy, Vision 2025, and the establishment of an A$500m (US$345m) Modernisation Fund; all the government’s projects, she noted, are to some extent underpinned by digital transformation. Working collaboratively across government, Huxtable added, departmental secretaries are leading initiatives to understand and improve citizen engagement, foster innovation and modernise the public sector.

The need to transform services, Huxtable argued, has hastened an existing shift back towards a more centralised model of public sector management – enabling central leaders to drive progress on issues such as data sharing and shared services. “We’re coming back into a centralisation curve, and I think the digital environment makes that inevitable,” she commented. “So it’s about how we do that effectively, and avoid some of the risks we’ve had in the past.”

Zooming in on performance

Better use of data is key to Australia’s plans – and Huxtable pointed out that, by pooling data centrally and applying analytics techniques, civil service leaders can improve the quality of management information. Taking the results of Australia’s 2018 public sector-wide staff survey, she explained, officials have created a data analysis tool that provides departmental leaders with information down to the team level – helping them to understand what makes teams effective and productive.

“At the departmental level, there’s about a five percentage point difference between staff engagement levels,” she said. “But if you go down to the team level, you can see a very big distribution within every entity.” And by cross-referencing engagement levels with employees’ responses to individual questions, analysts can clearly identify some of the determining factors – such as the extent and quality of managers’ communications. “Showing people these correlations can drive real change at the team level,” she added.

Huxtable’s focus on using data to improve decision-making struck a chord with Steve Bennett – an American scientist and former civil servant now working as Director of Global Government Practice at event knowledge partners SAS. “There’s an international definition of analytics by Informs, the international association for analytics professionals: ‘Analytics is the scientific process of transforming data into insights for decision-making’,” he commented. “I like this definition, because it focuses on the point that the value of analytics lies in helping leaders make decisions better and faster.”

Analysing analytics

In particular, Bennett argued, analytics can help overcome the problem that “humans are terrible at making decisions”. Given a set of data, he explained, people often make poor decisions due to three weaknesses in human psychology. “One is biases in judgement; the second is difficulty in dealing with quantitative information; and the third is irrationality under uncertainty.”

On the first, he explained, “our brains are built to be biased in how we receive and process information.” For example, “availability bias” means that “things that are vivid or recent in our imagination drive our estimation of reality”. And this affects us all, he said: in his work with government experts, Bennett found that their judgements of the threats posed by chemical and biological weapons were skewed by recent news coverage.

Another example is “anchoring bias,” he added. “The first information you receive is often given more weight than information that comes later – even if the later information is of higher quality.”

On assessing quantitative information, he continued, “we’re really bad at dealing with numbers that are very large and very small. We underestimate things that are very large, and overestimate things that are very small. So we can do maths on a spreadsheet and get the numbers right, but intuitively we don’t weigh the numbers in the way we should.” As a result, people are often far more motivated and engaged by the story of, for example, a single child’s death than by a report on the killing of thousands.

Take the right risks

Third, said Bennett, “we’re terrible at making decisions when there’s uncertainty – and all decisions in government involve uncertainty!” For example, when asked whether they’d rather be given $7000 or an 80% chance of winning $10,000, most people opt for the safe money – but they’d be better off taking the risk, which produces an “expected value” of $8000. “We tend to undervalue probability when we’re looking at a gain,” he explained.

On the other hand, given the choice between being required to pay out $7000 and having an 80% chance of coughing up $10,000, most people take the chance – though this is again irrational, leading to an average loss of $8000. “People become risk-seeking when they’re looking at a loss, as opposed to risk-avoiding when they’re considering a gain,” said Bennett. “And this shows up in geo-politics. It explains the Cuban missile crisis and many other government decisions. This really drives the way a lot of decisions are made under uncertainty.”

Analytics can compensate for human biases, he argued: “It provides an evidence backstop that helps improve the quality, consistency and effectiveness of decisions.” And he gave a series of examples of the technique in action. In New Zealand, he said, officials began using public data to identify single parents, then reached out to them to offer support: the result was a 10% fall in unemployment among a pilot group.

Artificially intelligent, but organically biased

Bennett also highlighted the potency of Artificial Intelligence, noting that “it has some advanced capabilities and benefits, but lots of challenges as well.” And, Nancy Chahwan, Chief Human Resources Officer at the Treasury Board of Canada Secretariat, argued that these techniques must be handled carefully. For AI and analytics systems can end up developing their own biases – either ‘learning’ them from skewed decisions in ‘training data’, or picking them up from programmers. These systems are created by people, she noted, “and that’s where we see the biases come up – bringing huge ethical risks.”

Mikhail Pryadilnikov, Deputy Director of the Analytic Centre for the Russian Government, pointed out two more issues around the use of analytics. Better evidence only has value if elected leaders are willing to use it, he said: “A good example is the wall that Trump is trying to build. All the civil servants are saying that this is not where the risks lie: that terrorists are not coming from Mexico.” Yet Trump is fixed on building his border wall – and it remains popular with his supporters.

Second, Pryadilnikov noted that China tops the league for public trust in government, and “they have no reservations about putting sensors everywhere”; but western populations both have far greater control over their data, and are much more reluctant to share it with public bodies. “Does that mean China is going to be ahead of us, because they’re going to know everything about [citizens] and they’ll be using this data to provide better public services?” he asked. “Maybe there’s a trade-off.”

Fair point, responded Bennett: “There’s a ‘could we?’ versus ‘should we?’ dimension to this, and the boundary is being pushed by nations which have less of a ‘should we?’ restriction. They’re willing to go further, and they’re limited only by what they can do. But in some cases, our western democratic values will cause us to self-limit what we can do with technology.”

Give people control, & they’ll share it

But Taimar Peterkop, the State Secretary of Estonia – widely recognised as one of the world’s most advanced digital governments – had a solution to the problem. In Estonia, he said, citizens “really have control over their own data, and that creates transparency and trust.” Under the system he called “Big Brother reversed”, Estonians can both choose how to share information with government bodies, and see exactly which public servants are accessing and using their data. Having provided citizens with both that sense of control and really accessible, high-quality digital services, he explained, the government has found that people are willing to allow data-sharing to improve the services they receive.

For example, Estonia has created data links allowing citizens to share data between their banks and the tax authorities – so 90% of the population can submit their annual tax return in less than five minutes. That’s a powerful incentive to allow data sharing; and take-up is very high.

Estonia’s tax return system provided powerful evidence for another of Bennett’s key messages: that the true power of digital is realised not by putting a new front-end on legacy systems, but by rebuilding business processes around entirely new models – providing more convenient, more efficient and more effective services. “Back office integration is about more than putting up a web page, or making an app that accesses the legacy process,” he said. “It’s about end-to-end digital transformation.”

And what capabilities do governments need to generate that kind of change? Ng Chee Khern, Permanent Secretary of Singapore’s Smart Nation and Digital Government Group (SNDGG), had some answers.

The ability to transfer data is an obvious one – but, he noted, exchanging data is only helpful if departments use the same data standards and definitions. Powerful cyber-security and an effective digital identity verification system are also crucial. And changes are required to systems of financial management and project approval: using traditional approaches, he argued, “you can’t do ‘Agile’ development, because from the word go they want you to tell them the exact plan of what you’re going to do – and it’s not possible.” With Singapore’s Ministry of Finance, said Ng Chee Khern, SNDGG is working to build resourcing and financial systems that can support digital projects.

Civil services must also find ways of building effective multidisciplinary teams, he said: “If you want to do digitalisation well, you have to find a way to bring five different communities together: the business owner; the IT guy; the data guy; the cyber-security guy; and the competency development and HR training guy. And these are people with very different cultures, used to working with very different bosses.” So recruitment and staff management practices must change – both to bring in digital professionals, and to equip them for working in the unique environment of public services.

Across the boundaries

All this, of course, demands effective management. As Huxtable pointed out, “the leadership that we show – not just at the senior level, but throughout our organisations – is absolutely critical to building a modern, adaptable public service that embraces digital opportunities.” And Peterkop stressed that realising those opportunities will demand substantial changes to civil service leadership cultures – moving away from traditional hierarchies and risk aversion, and empowering private suppliers and delivery professionals.

“What we’ve done in Estonia, we’ve done together with the private sector,” he explained. “The private sector is always willing to go further, to explore new things, and we’ve created an atmosphere where we allow them to do that. Digital Estonia hasn’t happened because of top-down leadership, the policy managers and digital planning: it’s happened because politicians trusted their digital advisors, and allowed them to experiment. You need this agility when you digitalise.”

“In the digital age, we’re in the first kilometre of a marathon; it’s a long road ahead,” commented Leo Yip, Head of the Singapore Civil Service. “And it’s not possible for us to envisage what’s ahead, in terms of technologies and opportunities. But there’s no alternative: it’s what citizens expect, and the onus is on us to marshal the systems, processes and capabilities to deliver.”

So there’s a chicken and egg problem here. Better public services will improve citizens’ trust in governments; but when public trust is low, people are reluctant to allow governments to use their data – slowing progress on improving services. “We need a better narrative when we talk to citizens,” argued Gabriel Makhlouf, then New Zealand’s Secretary to the Treasury (he has since become Governor of the Central Bank of Ireland). “At the moment, digital is either presented as a way of saving money, or as a way of giving government lots of information: it’s very producer-focused.”

“I don’t think there’s a powerful enough story as to the difference digital will make,” he concluded. “Citizens are increasingly in need of narratives – on services and technologies, as well as the issues around globalisation. We need to show how digital services will impact them, and help to improve their lives.”

Related Posts

• 

  Ministers need to better understand the tension between urgency and realism in government if they want to deliver
  Ministers need to better understand the tension between urgency and realism in government if they want to deliver
• 

  Australian Public Service to tap talent outside Canberra
  Australian Public Service to tap talent outside Canberra
• 

  G7 leaders ‘take stock’ of AI amid calls for shared governance standards
  G7 leaders ‘take stock’ of AI amid calls for shared governance standards
• 

  UK civil servants seek legal routes to oust bullies; US proposes hiring feds responsible for emerging tech: management & workforce news in brief
  UK civil servants seek legal routes to oust bullies; US proposes hiring feds responsible for emerging tech: management & workforce news in brief