Our new policy project explores Mission 3 of the NDS: ‘transforming government’s use of data to drive efficiency and improve public services’. We believe that once this mission is achieved, it will support Mission 1 – ‘unlocking the value of data across the economy’. In turn, this supports the overall ambition of the NDS: ‘to help organisations of every kind succeed – across the public, private and third sectors’. Through transforming its own use of data, the UK government can set an example for others and show how its own data practices can enable data flows and data use across sectors.
As part of our project, we’re mapping the different teams responsible for data across the UK government and public sector landscape, and the data-related initiatives they’re undertaking. We’ll be examining them against the six components of a trusted and trustworthy data ecosystem that we outline in our manifesto: infrastructure, capability, innovation, equity, ethics, and engagement.
Data equity, fairness and rights
This week’s blogpost focuses on data equity:
Everyone must benefit fairly from data. Access to data and information promotes fair competition and informed markets, and empowers people as consumers, creators and citizens.
One aspect of equity is around diversity and inclusion. The importance of data to decision making, data-driven innovation through technologies like artificial intelligence, and data’s role as important national infrastructure means it is vital that data is as representative as possible. But certain groups can be under-represented in data and others over-represented: for example, the ODI has looked at the importance of collecting information on diversity and protected characteristics to highlight who is and isn’t using digital public services and therefore who might be discriminated against.
The Public Sector Equality Duty, which came into force in 2011, is designed to eliminate discrimination and advance equality of opportunity between those with protected charcteristics and those without, but it may not always occur to us that data gaps and the way we collect data can lead to bias. This can reinforce existing inequalities: the 2020 A Level algorithm fiasco was a prominent example of bias in the use of data and algorithms, relying on previous results from schools and colleges and therefore discriminating against individuals from schools that didn’t have a tradition of strong results.
But another aspect is around fair competition and informed markets, and empowering consumers. Regulation and legislation like the General Data Protection Regulation (GDPR) guarantees people’s rights as citizens and consumers, protects their privacy and gives them more control over rights about data about themselves. In the NDS, the government pledges to protect such rights but balance them against a ‘pro-growth’ regime which benefits businesses and isn’t too ‘burdensome’. The right to data portability in GDPR – which allows individuals to obtain and reuse personal information – enables consumers to move between different services, which may be good for competition and allow the consumer to get a better deal.
Data equity in government
There are fewer initiatives and organisations explicitly focused on data equity compared to other data-related subjects like infrastructure, capability or ethics. But the initiatives that do exist can also support data equity across the wider economy and society.
Some of this is through the government collecting and publishing data that shines a light on inequalities within and beyond its own operations. For example, the Equality Hub, which works with the Cabinet Office, Government Equalities Office, Social Mobility Commission, Race Disparity Unit and Disability Unit, is using data to understand inequity and inequalities across society. An equalities data audit, building on the Race Disparity Audit, includes data on geographical inequalities – part of the government’s ‘levelling up’ agenda – and the intersection between different characteristics, such as ethnicity, gender, disability status and socio-economic background. The UK Statistics Authority’s Inclusive Data Taskforce is seeking to improve the data held on a number of these characteristics, while the (much-criticised) Report of the Sewell Commission on Race and Ethnic Disparities includes a recommendation to ‘use data in a responsible and informed way’.
A review of how algorithmic bias might affect significant decisions about individuals has also been conducted by the Centre for Data Ethics and Innovation. It looked at financial services and recruitment as well as the public sector, and its recommendations and resources will be of use beyond government.
The government’s wider role in data equity
Other government actions should encourage greater equity across the economy. For example, the NDS notes government funding for degree conversion courses to data science and AI, ‘including £10m for up to 1,000 scholarships for people from diverse backgrounds’; these efforts to diversify data science degree students should lead to greater diversity in the collection, use and application of data in the future. The NDS Forum – designed to help ‘operationalise’ the strategy and keep stakeholders involved – says it will ‘prioritise diversity of representation, bringing together multidisciplinary expertise and fresh perspectives from up and down the country’.
When it comes to the private sector and markets, the work done by the Department for Business, Energy and Industrial Strategy on Smart Data – ‘putting consumers in control of their data and enabling innovation’ – stands out. It was referenced multiple times in the NDS, with developments including the creation of a cross-government Smart Data working group to support initiatives and benefit consumers across different sectors. The new Digital Markets Unit is tasked with approaching competition and harms in digital markets. The Information Commissioner’s Office obviously has an important role, given its responsibility for enforcing legislation like GDPR. And the Equalities Hub is also responsible for the Gender Pay Gap Service: organisations with more than 250 employees are required to publish gender pay gap data, to help understand and illustrate the problem.
Examples from elsewhere
Other jurisdictions are also working on tackling data inequity. The United States is establishing an ‘equitable data working group’, noting that a lack of disaggregated data on protected characteristics ‘has cascading effects and impedes efforts to measure and advance equity’. The European Commission has an equality data handbook and other initiatives underway.
There is also increasing action around the inequities created by tech monopolies on both sides of the Atlantic, with Europe’s Digital Markets Act and the US government’s appointment of Lina Khan as Federal Trade Commission chair prompting complaints from Facebook and Amazon.
We’d love you to help us map relevant UK public sector teams and initiatives, as well as relevant examples from other sectors or other countries. Our mapping document is open to comments and contributions so that we can crowdsource to help us fill out the gaps, and to get your perspectives on government data initiatives. It is open for comment until Friday 17 September 2021 (extended from 10 September). You can also email email@example.com or submit via our anonymous Google Form.
Alongside our broader engagement with the UK National Data Strategy, some of the ODI’s wider portfolio of work in this area includes:
- The UK National Data Strategy 2020: equity, fairness and rights
- One year of Diversity, Equity and Inclusion at the ODI
- Inclusive data: perspectives from a roundtable discussion
- The ODI responds to the UK Statistics Authority Inclusive Data Consultation
- Objective data? Reflections on the Commission for Race and Ethnic Disparities report
- The weird and the wonderful: reflections on the Commission for Race and Ethnic Disparities report
- The dividing line: how we represent race in data
- Data enabled cities
- Monitoring Equality in Digital Public Services (report)
- About Data About Us