© Frank Rumpenhorst

"Without data, policies cannot move forward," said Min Zhu, Deputy Managing Director of the International Monetary Fund (IMF), at the opening of the 3rd IMF Statistics Forum in Frankfurt am Main. At this two-day conference, which was hosted by the Bundesbank and held outside Washington, DC, for the first time, around 150 data producers, policy‑makers, academics and other users discussed how to further improve the provision of statistics. Data are also of vital importance for central banks, as Bundesbank President Jens Weidmann stressed in his opening address. "Monetary policy ultimately depends on the availability of good statistics," he said. Macro and microeconomic analyses are not possible without them. This, he said, was also a reason why last year the Bundesbank established a Research Data and Service Centre where internal analysts and external researchers can access micro data, ie individual-level data, for meaningful analysis while complying with data protection requirements.

Financial crisis the key turning point

Great progress made in the field of statistics has often been driven by new demands, Weidmann said. And, in the recent past, the financial crisis has led users to demand more of statistics, he added. "One of the lessons of the financial crisis is that stable consumer prices do not automatically guarantee financial stability," said Weidmann. "We have also learned that just because individual banks are sound, one cannot conclude that the entire banking system is stable." Risks to financial stability, can arise, for instance, if individual banks are too big to fail, too interconnected to fail or if many small institutions are exposed to the same risks, he added.

Former ECB chief economist and Executive Board member Otmar Issing noted that a globalised world required not only micro data (at the individual level) and macro data (containing aggregated data from groups) but also statistics which show how individuals, institutions and sectors are networked. In light of these new requirements he saw the financial crisis as representing the greatest turning point in the history of statistics. In the years following the crisis, central banks and supervisory authorities were given new mandates and instruments which require a broader set of data, noted Issing. Demand for granular data, ie special micro data, emerged virtually overnight, he said. In order to describe and bring to light financial interlinkages, both sides of a financial relationship need to be accounted for, necessitating "from-whom-to-whom" data. Mr Issing stressed that statistics needed to reflect a fast-changing world.

Official statistics indispensable

In various panel sessions, academics, statisticians and policy-makers discussed which data sources could be used in future to satisfy the rising demand. Using the city of New York as an example, Julia Lane, a professor at New York University, showed how big data applications can be harnessed for research purposes. Big data are large data sets that are very often generated in near-real time as a by-product of transactions or web-based communication. Members of the audience debated with panellists on the extent to which these data sources could complement official statistical agencies, as data producers, in the future. Big data will also be increasingly valuable for statistical purposes, said Stefan Bender, Head of the Deutsche Bundesbank's Research Data and Service Centre. However, these data are often selective, incomplete and erroneous, he warned. While they could be used as an additional source, he noted, they could not replace official statistics.

In order to satisfy increased user requirements, in future it will be increasingly important to share data, according to Patrick McGuire from the Bank for International Settlements (BIS). He said it was important, for example, that microprudential supervisors pass on information to central banks, which monitor risks to the entire financial system. Furthermore, he also called for international organisations, for instance, to be given limited access to data, although he conceded that this was often legally difficult.

Using data from authorities

According to Walter Radermacher, Director General of Eurostat, the times in which data were captured for one purpose and used solely for that purpose are over. Luigi Guiso, professor at the Einaudi Institute for Economics and Finance in Italy, even heralded the arrival of a new age for research in the fields of social sciences and economics – the age of administrative records. He highlighted successful examples of combined data sharing, such as the use of official data by statistical agencies and researchers in Scandinavia. This, he said, is rather difficult in other countries, such as Germany, owing to data protection laws.

The panellists agreed that more common standards were needed for working with and developing statistics. Patrick McGuire called on the international community to work together to provide standard guidance in collecting micro data. Bundesbank President Jens Weidmann pointed to the German Panel of Household Finance (PHF) survey as a model for a new breed of micro data sets that could be used for research purposes. The PHF is a survey of households' assets that is seamlessly embedded in the Eurosystem's European Household Finance and Consumption Survey. The latter has revealed a large disparity in the distribution of wealth across the euro area. The detailed information on assets and liabilities identifies risks for financial stability and possible liquidity constraints, enables a better understanding of the channels for monetary transmission and enables the study of the distribution effects of monetary policy. Radermacher underscored the need to make statistics internationally comparable, calling for every single statistical product to have an international barcode, like a product in a supermarket. He said that this has now been decided for European data for the first time, but that this was a long time coming.

Access for external researchers

According to Bundesbank Deputy President Claudia Buch, from a cost-benefit perspective it was important not only to share data, but also to analyse it jointly. This could also prevent additional burdens being placed on survey respondents. The limited capacities of official statistics offices also made this necessary, she added. Professor Buch encouraged academics and external analysts to work more closely with official statistics producers. This, she said, was the very reason behind the Bundesbank's Research Data and Service Centre, which allows external researchers access to comprehensive micro data, as Stefan Bender pointed out.

Professor Buch called on participants to open up and explain what they do. She noted that the public often greeted the collection of microdata with scepticism. The important thing, she said, is to explain to the public what one intends to do with these micro data. It is not about obtaining information on specific individuals, she stressed. Rather, central banks and policy-makers need anonymised micro data in order to condense these and be able to draw meaningful conclusions about certain groups. This, said Ms Buch, helps improve the identification of risks to financial stability overall.