The IFC together with the International Statistical Institute (ISI) organised a High-Level Meeting on Data Governance in Tunisia on 22 November 2019. The event was a key occasion to discuss the issues faced by public institutions in general, and national statistical authorities in particular, when dealing with, and using, official data. It also proved to be a useful opportunity to show how statistics can play a decisive role in measuring, monitoring and evaluating the implementation of major international initiatives supporting development. This is particularly the case in respect of United Nations (UN) 2030 Agenda for Sustainable Development Goals (SDGs), which requires countries to produce a standardised list of indicators, and the Agenda 2063 for the development of Africa.
A first main message from the meeting is the need to have an all encompassing approach to data governance when collecting, managing, disseminating and making use of official statistics. This should cover all the related principles, policies and procedures, structures, roles and responsibilities. Another key lesson is the importance of proper data governance frameworks for those organisations composing national statistical systems (NSSs), especially National Statistical Offices (NSOs) – or National Statistical Institutes (NSIs) – and central banks’ statistical departments,3 to reap the full benefits of the ongoing “data revolution”. Such governance frameworks should cover the entire organisations and be an integral part of their strategic plans.
As regards first data collection, traditional statistical surveys and censuses can be usefully complemented with new types of information, eg alternative data sources including administrative records and “the internet of things” (big data) – described by some as the new oil of the 21st century (The Economist (2017)). This can be a great opportunity for those less developed statistical systems, not least considering the high costs associated with setting up and maintaining standard exercises. One risk from this perspective is the hoarding of the vast data resources collected outside the official perimeter of statistics if one fails to incorporate them properly to support the measurement of economic indicators.
Second, there are clear challenges related to NSSs’ management of the evolving data ecosystem. In particular, what is unclear is how to deal with “organic” information sources, whether private commercial data sets or public registers that were not initially set up for a statistical purpose, and which may not pass the test of time. Sticking to long-established and internationally agreed practices and standards, preserving sufficient “traditional” statistical capacity in the NSS and favouring a complementary use of both traditional and alternative data sources are central to maintaining well-founded trust in official statistics.
Third, turning to data dissemination, digitalisation techniques allow for easier, almost cost-free access to information for the public. However, the increasing complexity of economic and financial activities in a data-rich world puts a premium on statistical education and financial literacy. In addition, official statistics are essential to provide reference, objective information and in turn support economic development and well-being.
Fourth, there has been a growing interest globally for the better use of data for policy purposes, especially when designing, calibrating, assessing and modifying policy actions. But the development of such indicator-based frameworks is facing important obstacles, reflecting existing limitations to effective and seamless data access and the sharing of official statistics (IAG (2017)) – for instance, when trying to make use of information collected from supervisory reports. One way to go is to promote the exchange of experience among institutions and countries in addressing these challenges in an effective and practical way.
Looking forward, well-defined data governance frameworks can be instrumental in supporting official statisticians’ task to collect and analyse data of the highest quality possible. However, an institution-level approach to data governance should be complemented by a broader focus covering the entire production and use of statistics, including alternative sources. For instance, ensuring the following of adequate Codes of Principles by private data providers, clarifying the responsibilities in the national governance landscape, and establishing proper international guidelines and cooperation mechanisms.
Perhaps more importantly, while NSS organisations are facing a decline in their traditional function of “data collectors”, they have a key role to play as reference custodians of the quality of the data used by society. Needless to say, establishing sound data governance frameworks can be a central element in supporting this “data curator approach”.