Welcome remarks at the 3rd IMF Statistical Forum
- 1 Introduction
- 2 The 3rd IMF Statistical Forum: Programme
- 3 Responding to new demands: The Bundesbank’s Research Data and Service Center
- 4 The Panel on Household Finances as an example for the use of micro data
- 5 Conclusion
Dear Mr. Zhu,
Dear Otmar Issing,
Ladies and gentlemen,
Good morning and welcome to the 3rd IMF Statistical Forum. It is probably safe to say that the Bundesbank has seldom had the opportunity to welcome such a large body of international statistics experts.
The objective over these two days is to bring together data producers, policy makers, academics and other users to discuss how to further improve the provision of statistics for micro and macroeconomic analyses.
For central banks this is of vital importance. Monetary policy ultimately depends on the availability of good statistics. The monetary and economic analyses that precede any deliberation on the appropriate monetary policy stance are built on a large amount of data on monetary aggregates, financial market data, inflation expectations or labour market indicators.
I am therefore particularly glad that the Bundesbank is hosting the first IMF Statistical Forum to be held outside Washington.
And, when I tell you that the Bundesbank is attaching great importance to statistics, I am not merely paying lip service to you.
The importance the Bundesbank attaches to statistics and statistical methods is shown, for instance, by the fact that Carl Friedrich Gauss, one of the most influential statisticians of all time, was portrayed on the 10 Deutsche Mark banknote for a long time.
Building on the work of Carl Friedrich Gauss, statistical methods have been perfected and the process of data collection has been refined again and again.
Often, the great progress made in the field of statistics has been driven by new demands. If you will, the supply of data and new methods has always responded to changes in demand.
Governments, for example, have always had a demand for statistics on the wealth of their citizens, as this helped them to collect taxes. In the 17th century, this led Sir William Petty, an Englishman, to count chimneys in order to draw inferences about the wealth of an economy. His works can be seen as the foundation of modern social and economic statistics.
But to find examples of how the supply of statistics has responded to demand one does not have to go back as far as the 17th century. In the 1930s, the Great Depression provided another big push for the development of statistics. At that time, the economic crisis had left its scar on workers, consumers and companies in the United States and Europe. Politicians were desperate for data allowing them to describe and analyse the economic situation. In 1930, the United States Congress therefore declared the estimation of national income to be a governmental task.
And, in the recent past, the financial and economic crisis again defined new requirements for statistics.
One of the lessons of the financial crisis is that stable consumer prices do not automatically guarantee financial stability. And we have also learned that just because individual banks are sound, one cannot conclude that the entire banking system is stable.
Risks for the stability of banks 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.
This insight has led governments and central banks around the world to assume macroprudential mandates. The objective is to ensure that the financial system as a whole can fulfil its central macroeconomic function, even in times of financial market stress.
But the new task of macroprudential supervision and regulation brings with it new demands for statistics: We need, for instance, to identify suitable indicators that signal when “financial stability” is in danger. This task is not to be underestimated as financial stability is a more complex concept than price stability.
And today we still know little about the effectiveness of macro prudential instruments. In evaluating the effect of these new instruments, we are often only able to conduct cross-country analyses that stretch over comparatively short periods of time.
To improve our understanding of macro prudential measures, we need to go beyond the use of aggregated statistics: The US experience, for instance, shows that a loosening of credit restrictions for subprime borrowers can only be identified if micro data are available.
We also need data allowing us to better understand the financial links between different banks.
And to ensure that the right conclusions are drawn from the data, we should make more anonymised micro data available to the public - at least as far as it is possible with regard to our commitment to protect the confidentiality of data pertaining to individual banks, firms or citizens.
2 The 3rd IMF Statistical Forum: Programme
The demands for new data that come from regulators, central banks and supervisory bodies are also reflected in the programme for this conference: In the first session, you have the opportunity, for instance, to discuss how micro data can be used to evaluate the effects of regulatory changes.
The second session addresses the questions of how and what micro data collected by public authorities can be made available to the public via data hubs.
And the third session looks at how data can be used to understand financial interconnectedness, for example by linking a buyer and a seller of a security or by identifying who is lending money to whom.
But this conference is not only about financial stability: The fourth session examines the statistical challenges associated with the measurement of the macroeconomic effects resulting from changes in the price of natural resources. The recent decline in the oil price and its consequences for inflation and financial markets may serve as an example for the importance of this topic.
Last but not least, the fifth session is concerned with the measurement of the material conditions of households.
By including this session in the conference, the programme committee has proven to have the right instinct concerning topical issues: This October, the Royal Swedish Academy of Science decided to award the Nobel Prize in Economics to Angus Deaton. Angus Deaton was awarded the prize not least for his extensive research into consumption and poverty in developing countries.
3 Responding to new demands: The Bundesbank’s Research Data and Service Center
Ladies and Gentleman, the demand for new and better statistics is not only reflected in the programme for this conference. It has also already left its mark on the organisation of the Bundesbank.
The Bundesbank is not only a user but also a large producer of high quality statistics. Aggregate data that are relevant for macroeconomic analysis are regularly published in our Monthly Report and our time series database.
But we also collect a large amount of micro-level data. These sometimes highly sensitive data range from banks’ monthly balance sheet statistics to micro data on foreign direct investment. To meet data protection rules, we can only make these data available to the public under certain restrictions.
To facilitate access to these data, last year we established our Research Data and Service Center. It is located in the 20th floor of the Trianon tower here in Frankfurt and provides workspaces for 12 guest researchers. A staff of 12 persons advises these guest researchers on data selection and data access.
The Center does not lift restrictions on data access in place to meet data protection rules, but it makes it easier for non-commercial researchers to use the data to the largest possible extent.
4 The Panel on Household Finances as an example for the use of micro data
One example of the many sets of micro data available in the Research Data and Service Center is our Panel on Household Finances. It provides survey data on households’ balance sheets, pension incomes, employment and demographic characteristics of German households. The second wave of the survey was conducted in 2014 and the first results of this second wave will be published at the beginning of next year. Similar panels are now surveyed in all member states of the euro area under the guidance of the ECB, allowing empirical analyses to exploit variations not only over time but also across countries.
Monetary policy makers can gain important insights from these data. They allow us, for instance, to draw conclusions about the number of liquidity constrained households. Such information cannot be obtained from aggregate data on household debt alone, because highly indebted households may still have access to bank credit if they possess sufficiently valuable assets. In the end, central banks can therefore better analyse the effectiveness of the lending channel.
By being able to obtain information on the asset and liability side of households’ balance sheets, we are also better able to identify financial stability risks.
But the Panel on Household Finances does not only yield important insights for central bankers. It is also a great treasure trove for academic researchers. To date, 60 researchers in Germany and more than 150 users abroad are using the anonymised data from the Panel on Household Finances. Their projects cover a broad range of topics. There are, for instance, projects on the impact of monetary policy on income distribution, on the importance of residential property for retirement saving or on measuring poverty and consumption.
So it is probably fair to say that the Panel on Household Finance is one example where the demand for deeper analysis is already producing new statistics and important insights.
The advance of modern information technology in all areas of society means that we are able to analyse ever larger data sets. But does this justify collecting every possible statistic?
There is no doubt that more data and new data processing technologies can facilitate the work of researchers, supervisors and statisticians. But it increasingly also raises serious questions of data protection. And it also requires taking into account that the burden of collecting data and preparing it for processing has ultimately to be borne by businesses, banks and households. The famous saying, "there is no free lunch", holds for data collection too.
But maybe, this aspect will already be addressed during this conference.
Ladies and gentlemen, you have two very interesting days ahead of you and I do not already want to strain your attention too much. I wish you all a very successful conference and hope that the discussions will be intensive and informative.