Published on Data Blog

Principled data access: building public-private data partnerships for better official statistics

This page in:

Illustration of cloud storage
Image credit: geralt /

Building an encompassing statistical system—one that straddles the three pillars of economy, society, and environment—is the key objective for national, and international statistical organizations.  Statistics stand at the root of policymaking and execution, as in Drucker’s famous quote: “what gets measured gets managed”. The ability of statistical systems to hold governments, businesses, and peoples to account bears a profound impact on ‘people, planet, and prosperity’ (the three interconnected pillars of action named by the 2021 G20 Italian Presidency,).

Official statistics are a key compass for policymakers . Yet, they fail to serve their role unless they meet a stringent set of requirements. Official statistics need to be readily accessible, impartial, reliable, timely and accurate. They must also be flexible enough to capture emerging phenomena, and to remain salient even in the presence of major disruptions.  The COVID-19 crisis revealed both weaknesses and gaps in our statistical systems. Digitalization, accelerated by the pandemic containment measures, caused acute measurement challenges as new activities, skills and behaviors emerge, hardly captured by conventional statistical sources. The challenge is reinforced by the increasing recognition of diversity, which entails even more disaggregated and granular information.

As recently pointed out by the World Bank, digitalization provides also a unique opportunity. The abundance of the data it generates could yield societal and economic benefits if put to use to fill such information gaps. Today, this happens to a limited extent. Official statisticians have little access to new data sources, which remain largely confined within the private sector, fragmented across collectors and bound by commercial and regulatory ties. It is hard to underestimate the welfare-enhancing value of this data. Were privately held data available for official statistics and for the public interest, they could vastly improve public decision-making, enhance our appraisal of incipient risks to the global economy, and increase our ability to face the next global disruption.

word cloud: statistics

Statistical authorities can best face this challenge by leveraging one specific trait of the digital society—the abundance of data on a very large and diverse set of domains of human life, including some that have never been quantified before. The amount of raw information that is produced daily by all sorts of devices, from smartphones to industrial machinery, is unprecedented and increasing exponentially. This data, together with new analytical techniques, can greatly advance the cause of evidence-based policymaking. Official statistical agencies, which represent social rather than private interests, need to play a primary role in this process. Depending on circumstances, they need to be given access to certain privately collected information that is relevant to specific policy objectives or, at the very least, a right to use it at various degrees of aggregation. When policymakers choose to resort to indicators sourced by the private sector, official statisticians must be able to audit their methodological quality. 

In a new article, published in the Occasional Papers series of the Bank of Italy, we argue that, when it comes to business-to-government data access, the challenges of trust, usability, and sustainability should be met through the development of private-public data partnerships hinged on globally shared principles.  The word “partnership” here is meant to be legally neutral, encompassing frameworks of cooperation that may be widely different in how they address the economics of data access and use. Choices on which arrangements are legally acceptable and/or desirable should fall within the remit of individual governments, without compromising on the independence of official statistics and compliance with international standards.

Retro styled golden compass (sundial) and old white nautical chart close-up.
Image credit: Alex Stemmers / Shutterstock

Therefore, we believe that a set of principles guiding the adoption of private-public data partnerships aimed at ensuring the official statistics are complete, timely, and informative. The proposal builds on three prongs: The first group set the stage by stating general principles for data partnerships, applicable to all parties involved. The second group introduced principles for official statistical agencies. Finally, the third group outlined principles for private-sector participants. Such principles are inspired by a logic of burden sharing in order to avoid that either the public sector or the private one are faced with excessive and unreasonable tasks (for a precursor of our initiative see also the Recommendations for Access to proprietary data proposed by the United Nations Committee of Experts on Big Data and Data Science for Official Statistics in 2017).

In sum, we provide a foundation for broader, structured access to new data sources on the part of official statisticians, for the purpose of evidence-based policymaking and research. Whenever the data refers to individuals, such access has to be accompanied by a careful assessment of privacy and confidentiality implications for individuals, and the implementation of safeguards to guarantee full anonymity and preserve trust. These considerations also extend to other entities such as businesses that have sensitive and economically valuable data under their custody about their operations and their relations, including but not limited to intellectual property. Finally, it is important to note that such proposal does not aim to undercut existing initiatives for public-private cooperation in the data sphere, especially those stewarded by international organizations for statistical capacity building in low-income countries, but rather to enhance multilateral consensus and coordination in this field (the Development Data Partnership provides an excellent benchmark). Our work benefited from extensive discussions and suggestions from colleagues in many G20 central banks, finance ministries and statistical offices.


Claudia Biancotti

Guest blogger / Director, Banca d’Italia

Oscar Borgogno

Guest blogger / Researcher, DG for Economics, Statistics and Research, Banca d’Italia

Giovanni Veronese

Guest blogger / Economist, Banca d'Italia

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000