In 2019, the Irving Fisher Committee on Central Bank Statistics (IFC) of the Bank for International Settlements (BIS), together with the European Central Bank (ECB) and the Central Bank of Malaysia, decided to analyse the issues posed by the integration of micro data sets into the framework of official statistics, for the purposes of policymaking.
This was an opportunity to review the expansion of micro (granular) data sets made available since the Great Financial Crisis (GFC) of 2007–09 and their possible contribution to macro compilation exercises. The review underlined the importance of broadening official statistical frameworks to benefit from the growing availability of granular data to assist the assembling of macroeconomic aggregates and/or facilitate the linkage between micro- and macro-level statistics.
This stocktaking exercise also documented that the “micro data revolution” can bring important analytical benefits and effectively support central bank policies, with a greater ability to “zoom in” on particular areas of interest and to assess the distribution of macroeconomic aggregates within reporting populations. One significant example is that granular and micro data already assist in assessing how finance can contribute to the greening of economies. The initiative also proved a welcome opportunity to highlight the potential of new types of granular information in “unusual” circumstances. Indeed, and as observed subsequently with the sudden outbreak of the Covid-19 pandemic in early 2020, micro data sources can be a valuable complement to the “traditional” official macro statistics available in times of stress, especially to alleviate compilation disruptions, assess pressure points, and facilitate the implementation of targeted policy measures.
However, there are also important challenges associated with dealing with these data sets, for instance as regards their quality, confidentiality and manner of accessing them. Indeed, the task of integrating (granular and) micro financial information in macro frameworks has proved more complex than initially thought. To ensure concrete progress in the future, attention should focus on:
- building effective micro data collection frameworks based on a comprehensive data strategy helping to contain reporting burden;
- accessing and making use of more granular sources of information, with the need to overcome the challenges related to their size and complexity with a view to transforming simple data points into knowledge;
- promoting the exchange of experience, eg as regards access to micro data sets and external research projects, the development of diversified staff skills, and the combination of different types of data sets;
- developing new and adequate analytical tools, for instance to enhance data quality assurance processes, extract summary indicators from a wealth of data points, and develop machine learning (ML) / text mining / network analysis approaches to maximise the potential of granular (including micro-level) information; and
- bridging the gap between micro- and macro-level statistical exercises, which can be instrumental in enhancing the understanding of how the financial system functions and interacts with the economy, assessing distributional issues and facilitating sectoral analyses.
In supporting the above tasks, there is a growing interest in making use of granular information from private sources that are not part of the official statistical offering. What is unclear, though, is how data producers located outside the national statistical system should feature vis-à-vis the fundamental principles that govern the production of appropriate and reliable official statistics and adhere to certain professional and scientific standards. Helpfully, a number of private firms have already adopted dedicated and transparent mission statements and principles to address these concerns, and such initiatives may deserve to be strongly encouraged by the official (international) statistical community.
Lastly, there are specific communication challenges, as well as potential legal difficulties, for policymaking institutions like central banks when (confidential) granular analytical insights are used as the foundation for their decisions.