Alberto Naudon: Statistics and beyond - new data for decision making in central banks

Welcoming remarks by Mr Alberto Naudon, Board Member of the Central Bank of Chile and Chair of the Irving Fisher Committee on Central Bank Statistics, at the 12th biennial IFC conference "Statistics and beyond: new data for decision making in central banks", Basel, 22 August 2024.

The views expressed in this speech are those of the speaker and not the view of the BIS.

It is my pleasure to welcome you all to the twelfth Biennial Conference of the Irving Fisher Committee on Central Bank Statistics (IFC) of the Bank for International Settlements (BIS). As you are well aware, this committee is a global network whose fundamental raison d'etre is to promote discussions around statistical issues that are of interest to central banks.

IFC membership has significantly expanded in recent years and has become quite universal. This is evident from the rich representation of organisations gathered here today. As I look around this room, I see diversity and I have been informed that we have participants from almost 70 diverse nations, with about 140 participants representing about 80 institutions. Let me thank in particular the high-level representatives of the IMF, the ISI, the OECD and the World Bank for their presence with us today. We are also pleased to welcome many participants attending virtually from all over the world. Our gathering clearly reflects the great collaboration achieved between central banks as well as within the global statistical community.

The presence of each one of you here today is essential. It will enrich our discussions, broaden our perspectives, and strengthen the IFC network. And this is particularly welcome because the theme of this conference is of utmost importance for central banking. In today's rapidly changing world, good statistics can play a pivotal role in central banking.

Firstly, central banks are important statistical producers serving the public good. We are well aware of the need to ensure the quality of this information and to develop more comprehensive and flexible datasets. This calls for a good dialogue with other stakeholders, including other authorities, national statistical offices, market participants, and academia. It also requires continuous support to the main international statistical initiatives, such as the revisions of the national accounts or the Data Gaps Initiative sponsored by the G20. Just as importantly, we require a good dissemination policy to bring the statistical products to different users. Achieving good communication with citizens is essential so that they can clearly understand what we are doing.

Secondly, central banks are also heavy data users. They need accurate and reliable data as well as sophisticated techniques to make informed decisions and evaluate their effectiveness, in turn supporting the promotion of monetary and financial stability. But, as producers as well as users of data, central banks have to cope with the advent of new technologies and the proliferation of data sources. On the one hand, the data revolution, the development of artificial intelligence and IT innovation are opening up exciting new possibilities, for instance of making use of non-conventional and/or unstructured information and dealing effectively with "big data". On the other hand, these rapid changes are bringing new ethical challenges to ensure a trustworthy use of AI, such as responsibility, explainability, accountability, bias, privacy, and data protection. Another key challenge is to preserve the relevance of good statistics in today's societies in front of alternative facts and fake news.

The IFC has been already working actively to shed light on these issues and to facilitate central banks' cooperation in addressing them. Our most recent contributions are dealing with the topics of micro data, data science, statistical communication, data governance and climate change statistics. The event today will allow us to learn more of so many experts from around the globe and make further progress.

Before turning to the presentations of this conference, let me share with you some thoughts on important aspects that will be covered more extensively during the various sessions during the following days.

• First, we are living in an era of unprecedented changes. New forces are reshaping our world in ways we could not have imagined a few decades or even years ago. Therefore, we are facing new challenges at a greater speed. For example, digitalisation is transforming every aspect of our lives, in turn questioning the way payments are organised and recorded. The development of sustainable finance is attracting a lot of interest in order to help track and mitigate the impact of climate change; but the availability of relevant, good-quality environmental data has become a key concern for central banks and financial supervisors.
• Second, and fortunately, data science can help us tackle the various issues we are facing. In particular, new tools can support the development of and dealing with large, complex and unstructured information data sets. Moreover, untapped, "alternative" information sources can provide more real-time, frequent, timely and detailed insights. No wonder that central banks have been actively exploring the adoption of data science tools in recent years, especially in the context of the IFC. As we will hear today, there are already promising results – for instance to analyse labour market conditions, gauge economic sentiment, predict consumer behaviour, improve the offering of official statistics or detect anomalies and improve data quality. In particular, we are clearly at a turning point with the emergence of large language models (LLMs) and generative AI.
• These new techniques can clearly revolutionise the way we work with data, by expanding beyond traditional statistical series, developing new types of indicators and extracting fresh insights. But, clearly, we need to understand better how the new models being developed can complement, or perhaps replace, our existing toolkits. We are very lucky to have present with us today, Fabrizio Ruggeri, President-Elect of the International Statistical Institute, who has accepted to present his views on these issues.
• Third, the amount and type of data mentioned above, encourage us to establish strong data governance frameworks. These are needed to manage the growing and increasingly complex pool of information available. Central banks are actively engaged with other public and private entities to make progress on this front. The objectives are to be able to meet growing information demands while ensuring strong data quality. This calls in particular for developing robust data standards that can serve as a common language for exchanging information, ensuring that data are only collected once, and providing transparent information about the data themselves – that is, data about the data, or metadata.
• Fourth, the combination of stronger governance frameworks and sophisticated techniques can help us to benefit from the wealth of granular data available in today's society. These data offer multitude of opportunities, providing central banks with insights that were previously inaccessible through aggregate indicators. However, working with micro-level information presents important challenges especially to protect privacy and confidentiality – which are fundamental concerns for official statisticians and public institutions. I am happy to note that a specific session devoted to these issues has been organised during this conference that also includes a presentation on Statistical Disclosure Control by INEXDA, the International Network for Exchanging Experience on Statistical Handling of Granular Data.
• Indeed, and this is my fifth point, these initiatives should not be taken in isolation as they require close international collaboration. We are very fortunate to have at our disposal international statistical manuals like the System of National Accounts (SNA) or the Balance of Payments Manual (BPM). They provide a common language to discuss and analyse economic and social phenomena. But they need to be constantly refined to adapt to our rapidly changing societies. The central banking community therefore actively supports the ongoing revision of these standards, not least to develop adequate methodologies for measuring new phenomena that constantly emerge.
• Lastly, it is also important to reflect on our past while we grapple with current and future problems. An important development in this context was the recent IFC decision to host the historical monetary and financial statistics (HMFS) central bank network. This network brings together statisticians and academic experts to exchange approaches to compile historic data and to facilitate learning from each other. Noteworthy advancements have already been made in the dissemination of long-term statistics concerning credit, interest rates, housing prices and, more recently, central banks' balance sheets.

I am confident that the insights shared and the discussions held during this conference will facilitate our understanding of how data governance, IT advancements, and practical applications of data science are shaping modern statistical analyses and in turn supporting central banks' public mandates. It is through such gatherings that we can learn from one another, exchange ideas, and build bridges that transcend borders. I am confident that, by working together, we can better harness the power of data to strengthen the resilience and effectiveness of policy actions.

I wish you all fruitful and engaging discussions.