There are many unanswered questions about data: who controls or accesses it, how to govern it, how much it is worth and who has the rights to that value? Much of the debate to date has focused in general terms on either the privacy costs of the growing use of online data or the broad economic potential of its use, but the creation of value from data of different kinds, and its capture by different entities, need to be better understood for effective policy.
There is little distinction in the public debate about different types and uses of data. For all the excitement about ‘big data’, in effect, the answers to such questions are either unclear or contested. So we aim to develop a taxonomy of ‘data’ for the purpose of understanding its value creation and capture, and to help build a shared understanding among researchers, policymakers and industry stakeholders about the categories.
Data consists of many different forms of information and meaning so we will be asking questions about how to value data of different types, based on its distinctive economic characteristics. This will affect the policy implications for data governance and regulation.
First, government has to make policy decisions that rest on the value of data to the economy as a whole. These include decisions to invest in maintaining datasets the government makes available as open data, and decisions to regulate concerning data sharing or openness. The Treasury recently published (2018) a discussion paper pointing to the economic potential of data, but also the challenges around unlocking that potential. The European Commission’s JRC noted the large array of policy questions but ended: “We conclude that there are no easy answers for regulators how to overcome market failures in data and information markets. This paper is a call for more research on these questions.” (Duch-Brown et al, 2017). A greater understanding of the value of data would help identify where the benefits of greater investment in and sharing of data are worth the costs. Given the public good characteristics of data, it seems likely that there is considerable untapped potential value in enabling greater sharing and joining up of data sets.
Second, even where it is recognised that greater sharing of data brings benefits, such as in transport, or for individuals, those making decisions about sharing data they hold need to understand what value they are giving away and what benefits they will receive. The question of how to distribute the benefits arising from data needs to answer the question of what value the provided data has, and how the benefits of data from multiple organisations and individuals can be fairly distributed. These are live policy questions. Implementing data trusts effectively will require understanding how to distribute value from users to contributors. The question of market power based on data aggregation is one of the considerations debated by the Furman Review. There is also considerable debate about mechanisms for paying people for personal data.
While many studies have investigated the value of data, and particularly public sector data, to the wider economy (McKinsey, 2013), in individual sectors (Transport Systems Catapult, 2017) or from particular organisations (Deloitte, 2017), most treat data in the abstract. However, the value of different types of data can be very different: it may be reference data, streaming data, historical data, statistical data or personal data; it may have different levels of completeness, accuracy or representativeness; it may depreciate more or less rapidly; it may be unique or commonplace; its marginal value may differ depending on context. This project will identify these facets in order to develop a more nuanced understanding of how to value data. It will also explore some unresolved questions such as complementarities between different types of data and non-linearities in the valuations.
Data is also not like standard economic goods, despite the cliché about it being ‘the new oil’. Our aim is to explore the distinctive characteristics of data with a view to contributing to a more systematic understanding of how to value and govern this key resource in modern societies. We want to develop a framework to recognise these distinctive features. For example, data is ‘non-rival’: it can be used without being used up, so that many people can use the same data. Data is shared, not exchanged. Market mechanisms are therefore unlikely to deliver socially optimal outcomes, in either price or quantity produced.
Data also often involves externalities: much of it gains its value mainly from being combined with other data. However, although aggregated value is therefore often greater than the sum of individual values, where the tipping point occurs is unclear. There is also value (private and social) to be created from combining separate data sources; there are often significant complementarities if individual sources of data can be combined with other data.
We would also aim to use this framework to contribute to the live policy debate about data regulation and governance. A key issue for us is to ensure that the benefits of the data revolution are captured as fully as possible but also that they bring wide social benefits, and do not become or remain a vector for the unequal distribution of wealth and power. Effective policy in this area will require understanding the distinctive characteristics of data.
Project information on Valuing data: foundations for data policy is also available online on the Nuffield Foundation website.
The project has been funded by the Nuffield Foundation, but the views expressed are those of the authors and not necessarily the Foundation.
About the author
Professor Diane Coyle, Inaugural Bennett Professor of Public Policy
Professor Coyle co-directs the Institute with Professor Kenny. She is heading research under the progress and productivity themes. Learn more