By Thilo Klein (OECD- Paris21) and Stefaan Verhulst (The GovLab)
Executive Summary: “New data sources from the private sector have enormous potential to complement and enhance official statistics by, for example, providing more timely data and developing new areas for analysis. The increasing number of research projects using this data underlines its potential. However, progress to operationalize the use of such data sources in the world of official statistics is proving to be slow due, in part, to barriers in gaining access to data from private corporations. Such barriers take the form of concerns on the part of private companies about losing their competitive advantage; legal constraints concerning privacy and confidentiality of client information; and the costs of setting up the necessary infrastructure and training staff for a non-core business related activity.
Nonetheless, several business models that enable data exchange between private corporations and official statistics are emerging, including:
- in-house production of statistics
- transfer of data sets to end users
- remote access to data
- trusted 3rd parties (T3Ps)
- moving algorithms rather than data.
Each of these business models, along with their associated risks and technical and governance requirements, are examined in this paper.
Incentives for private companies to share their data include the mutual benefits accrued from working with National Statistical Offices (NSOs), the potential to develop new analytical skills, improve their reputations, generate revenue, meet regulatory compliance and demonstrate corporate responsibility. There is growing recognition among many companies – albeit slowly – of these many incentives for making data available for public good.
After examining various data sharing models and incentives, the paper makes a number of recommendations for policy actions. These recommendations suggest the need for NSOs to enter into partnerships with private providers:
- Corporations should consider electing “data stewards” to act as focal points for data access.
- A network to share experiences and know-how around data sharing and evidence of impact should be created.
- A repository of case-studies that highlight innovation in sharing practices should be developed.
- A decision tree to help assess benefits and risks of sharing corporate data should be defined.
- A standardized safe environment for sharing data without risk of compromising customer privacy should be developed.