Central bank statistics: What did the financial crisis change?
Speech by Jürgen Stark, Member of the Executive Board of the ECB, delivered at the fifth ECB Conference on Statistics, Frankfurt am Main, 20 October 2010
Ladies and gentlemen,
The focus of this conference has been on the changes that have been brought about by the financial crisis in terms of central bank statistics. It has been pointed out, correctly, that our analysis of the data available prior to the emergence of the crisis highlighted vulnerabilities in the financial system. However, those data were not sufficient to foresee the scale and severity of the crisis.
In the meantime, some progress has been made, particularly in terms of closing the most urgent gaps in the statistics (for example, on loan securitisation by banks and the activities of the securitisation vehicles), but more information will be needed, especially on the extent of the interlinkages within the financial system. Thus, central bank statistics will need to cover a wider range of financial sub-sectors, instruments, markets and geographical areas. However, it is not just changes to central bank statistics that are required. The most recent phase of the financial crisis has also exposed gaps in government finance statistics. To support the Stability and Growth Pact and maintain credible budgetary surveillance, we need more reliable and timely data.
Let me elaborate on these points.
The European System of Central Banks (ESCB) was greatly aware of the growing financial imbalances in the years preceding the crisis.  A number of risks and vulnerabilities had been highlighted in various official reports. For example, the ECB’s Financial Stability Review drew particular attention to:
the situation of the sub-prime segment of the US mortgage market (and the risk of contagion to other markets);
the key role of hedge funds in fuelling liquidity growth in financial markets (and, again, the risk of contagion to other markets); and
the difficulties in identifying risk concentrations, owing to the proliferation of structured finance products and other credit risk transfer instruments.
Nevertheless, both the scale and severity of the financial crisis came as a surprise to most, not least because macro-prudential risk was not sufficiently taken into account. So, the question arises as to which tools and information are needed for macro-prudential oversight?
The effective conduct of macro-prudential oversight requires a large set of tools that are supported by a comprehensive information base. These tools need to be revised and updated regularly. Consequently, the information base needs to be sufficiently flexible to adjust swiftly to changing needs, as well as to better capture innovation in financial markets.
Macro-prudential indicators, as well as early warning indicators and models, are important risk surveillance tools. They should indicate when a financial system, market or intermediary enters a “danger zone”, which usually suggests a risk of a crisis. In the past, most central banks have used a number of forward-looking risk indicators, but these have not necessarily been comprehensive enough to cover the entire financial system.
The establishment of the European Systemic Risk Board (ESRB) and the ECB’s involvement therein will require a wide range of initiatives and responses in the field of statistics. The data available prior to the crisis were not sufficient to grasp the full extent of the vulnerabilities in the financial system and to better comprehend at least two phenomena:
First, the wide usage of risk-based instruments: the years leading up to the crisis saw considerable growth in the securitisation and credit risk transfer market. Through these instruments, risks were being passed on from the originators to various investors, not only outside the banking sector, but also within it. The structuring of collateralised debt obligations was also a very profitable business for banks in the years before the crisis, with many accumulating sizeable amounts of risk-based instruments that later on became impaired assets.
Second, the large degree of interconnectedness and interdependence among financial intermediaries (in particular those operating outside the regulated financial markets). The ECB had looked into the interdependencies in the financial sector for the purposes of its monetary analysis. For example, it had analysed the growing number of interlinkages between the euro area banking sector and the rest of the world. However, it did not have the full set of information that would have enabled it to identify how much of this activity was being conducted by special purpose vehicles and offshore auxiliaries.
This conference has highlighted a number of gaps in statistical reporting, such as the two I have just mentioned. It has also discussed some work streams that had been initiated by the ESCB before the outbreak of the financial crisis. At least three new avenues were pursued and these have proved to be very effective in addressing needs for both monetary policy and financial stability purposes.
First, there is the greater integration of various sets of information. One example of this is that, since 2007, the ECB and Eurostat have been publishing integrated quarterly euro area financial and non-financial accounts by institutional sector. Over the coming years, priority will continue to be given to improving the completeness, geographical coverage, consistency and timeliness of the quarterly sector accounts. It will also be important to complement these sector accounts with distributional information on indebtedness, wealth and consumption in the household sector. The forthcoming Eurosystem Household Finance and Consumption Survey will certainly help to fill this gap.
Second, there are the efforts that have been made to achieve greater granularity and flexibility in datasets based on micro-databases and registers, so that data need only be reported once and the reporting burden is kept to a minimum. This is important because the same source data can be used more flexibly to derive various (micro and macro) statistics in order to meet new data needs, for example for financial stability analysis. I attach particular importance to the ongoing work to gather statistics on the holdings of securities.
Third, there is the wide range of new ESCB statistics that have been published recently (or are soon to be published) in order to close the most urgent gaps. Here, I refer to statistics on investment funds, insurance corporations and pension funds, loan securitisation by banks and the activities of securitisation vehicles.
These initiatives notwithstanding, further progress is needed not only in the field of central bank statistics, but also in other fields.
Particular attention needs to be given to the preparation of a comprehensive statistical basis for macro-prudential oversight. This will necessitate the collection of additional data in terms of central bank statistics.
First, there will be a need for more timely, harmonised and frequent – at least quarterly – data for the EU consolidated banking sector. The same requirements will apply to individual EU countries and to certain major non-EU economies. The data should include firm-level information on a set of large financial institutions. In this respect, I should underline that the objective is no longer to cover only the euro area, but the EU as a whole. For some types of analysis, for example those to gain a picture of global developments, we may even need to cover countries outside the EU. The statistics need to focus much more strongly on new risk-based instruments and exposures to certain asset classes and geographical areas.
Second, there will be a need for more detailed statistics on financial institutions outside the banking sector. Here, I refer mainly to insurance corporations and pensions funds, but also to institutions, such as hedge funds and securitisation vehicles. Some of these statistics either already exist or are soon to be released. As far as possible, these statistics should follow the same format as the information collected from the banking sector.
Third, the major gaps in statistics on the interlinkages within the financial system (including exposures between institutions in the banking sector and institutions in non-regulated markets) need to be filled. The collapse of Lehman Brothers two years ago illustrated the relevance of such interlinkages. In particular, it highlighted that such interlinkages may be more relevant than the absolute size of the balance sheet of the institutions.
It will take some time for these enhancements to materialise. Hence, we can expect to receive ad hoc requests from the ESRB relatively frequently in its first few years. Such ad hoc requests will often translate into ad hoc surveys because they are able to capture financial innovation and changing exposures, and accommodate more swiftly new data requests from the ESRB. Later on, we may have to improve the quality of the data and I would expect these data requirements to be met with regular supervisory and statistical data that are comparable across countries and across institutions.
Further improvements are also needed in other statistical fields.
As the sovereign debt crisis in May this year has further underlined, reliable and timely government finance statistics are key to the implementation of the Stability and Growth Pact and the credibility of its budgetary surveillance. Budgetary forecasts also need to improve in parallel. To this end, the principle of the professional independence of statistical authorities needs to be fully adhered to by all Member States, in order to avoid any potential political interference in the compilation of the data for an excessive deficit procedure.
I would like to conclude my remarks today by assuring you that the ESCB will further develop its statistics so as to better align them with the data needed for ECB monetary policy and financial stability purposes. In particular, they need to focus much more strongly on new risk-based instruments and the exposure of banks and other financial institutions outside the banking sector to certain asset classes and geographical areas. Moreover, we need to address the gaps in the information on interlinkages within the financial system. However, these changes to central bank statistics will not be enough. In particular, the importance of timely and reliable government finance statistics cannot be overestimated. If the credibility of one specific set of European statistics is damaged, there may be spillover effects on other sets of data in Europe. Most importantly, damaged credibility may expose unexpected vulnerabilities, which could lead to a general loss of confidence in the markets. As today is World Statistics Day, it would also be a very fitting occasion to call for joint efforts to further build trust in statistics.
At the same time, we also have to accept that we will not be in a position to measure everything we would like to measure. This relates in particular to those entities that are located outside the EU, but may nevertheless have a significant impact on EU financial markets and institutions. Hedge funds are a case in point. International initiatives, for example under the auspices of the G20, are very welcome but are likely to find solutions only in the longer term.
Ladies and gentlemen, let me finish by congratulating the chairs, the speakers, the discussants and the many distinguished participants in this conference. You have provided excellent contributions and reflections on strategic issues for the future development of statistics. This conference has no doubt been a very good opportunity to enhance our cooperation, to identify user requirements and to remove obstacles to the supply of financial stability statistics. It has also provided useful insights into the ESRB. I hope that our fruitful cooperation will continue in the future and thank you very much for your attention.
See, for example, Financial Stability Review, ECB, June 2007.