The development of euro area statistics from an ECB perspective
Speech by Jean-Claude Trichet, President of the ECBdelivered at the third ECB Conference on Statistics,Frankfurt am Main, 4-5 May 2006
Colleagues, fellow central bankers, ladies and gentlemen,
It is a pleasure for me to address you on the occasion of the closure of the Third ECB Conference on Statistics. With this conference, we have continued our biennial tradition of holding a dialogue between the compilers of statistics and policymakers, academics and other users on the development of statistical requirements. This tradition was started four years ago by Eugenio Domingo Solans, the first ECB Executive Board member responsible for statistics, who left the ECB in May 2004 and sadly passed away soon afterwards.
The subject of this conference has been financial statistics for a global economy, and I would like to use this occasion to share with you some reflections upon the four themes discussed during this conference, namely:
Are financial statistics good enough to capture globalisation?
What are the challenges for national data collection in a global world?
Which are the statistical implications of financial stability and financial integration? and finally
What can we say on global statistical governance?
I will obviously focus on the necessary development of euro area statistics and I would like to begin with a brief review of the recent developments in these statistics.
Recent developments in euro area statistics
An evaluation of current euro area statistics should start with three observations. First, official statistics provide the foundation for both the economic and the monetary analysis of the ECB and therefore for its monetary policy. In addition, they are indispensable for the other functions of the ECB, for other economic policy-making in Europe and for a smooth functioning of the financial system. This also means that high demands are placed on these statistics.
Second, it is still less than a decade ago that the statistical requirements for the European Economic and Monetary Union were agreed and that the first euro area statistics became available. It can now be concluded that the compilers of statistics in Europe have gone a long way towards harmonising their national statistics and meeting the requirements in a relatively short period of time, and I would like to compliment them on this achievement.
Third, this progress has been possible thanks to excellent cooperation within Europe’s statistical community, which consists of national central banks (NCBs), national statistical institutes (NSIs), Eurostat and the ECB. In particular, I would like to pay tribute to the Statistics Committee of the European System of Central Banks (ESCB) and the European Committee for Monetary, Financial and Balance of Payments statistics (CMFB) for the very fruitful role they play.
In fact, since I last addressed this audience, in April 2004, significant progress has been made, although also still much remains to be done. Concerning the euro area statistics compiled by the ECB, in close cooperation with the NCBs, I would like to recall some recent enhancements:
Since January 2005, the euro area balance of payments has contained a quarterly geographical breakdown of the counterparties involved in transactions with residents.
Furthermore, the international investment position (i.i.p.) statistics, which show the balance sheet of the euro area, are now available at a quarterly frequency.
In a different field, the ECB now regularly compiles a set of statistical indicators that shed light on the state of euro area financial integration in many different financial markets.
Another improvement concerns the semi-annual release by the ECB of a residential property price indicator for the euro area.
The ECB’s publication of the quarterly government revenue, expenditure and deficit for the euro area, based on contributions from NSIs, was recently complemented by fairly detailed data on government debt and the deficit-debt adjustment.
The ECB’s financial analysis benefits from separate statistics that are now available for securities issued at fixed and floating rates, and zero-coupon bonds.
Finally, the accessibility of ECB statistics and the national breakdowns has been enhanced as the same set of tables is now available on the websites of the ECB and of most of the Eurosystem NCBs, in the national languages of the central banks concerned.
In addition to the financial and monetary statistics compiled by the Eurosystem, the ECB makes intensive use of general economic statistics compiled by Eurostat and the NSIs. First and foremost, this concerns the Harmonised Index of Consumer Prices (HICP) for the euro area, which has continued to be both timely and reliable. In this context, the gap between inflation perceptions and actual inflation is now decreasing in many euro area countries. As you know, this gap originated when the euro was introduced and it became complex for consumers to process price information. The ECB also supports the further development work in this field, including the priority attached to compiling a reliable indicator for the price development of owner-occupied housing services and a further enhancement and harmonisation of adjustments for changes in the quality of goods and services. The latter work should also benefit the estimation of GDP volume growth, which is another key indicator for the ECB.
GDP and the national accounts statistics on which it is based have recently been revised in many countries. Whereas methodological improvements are welcome, it is important that the timing of such revisions, and of the release of the Principal European Economic Indicators more generally, be coordinated among countries. Otherwise, European users could be confronted with constantly changing euro area aggregates. In this context, recent ECB analysis found that, since the start of the euro area, the revisions of the principal quarterly euro area indicators have been fairly modest and significantly smaller than the revisions of the same indicators for the largest euro area countries and for the United States. This is explained by the fact that revisions of national statistics largely cancel out at the euro area level. At the same time, this result reinforces the importance of aggregate euro area statistics for the analysis of the euro area economy.
In view of the still growing importance of euro area statistics, I am particularly looking forward to a forthcoming new milestone, i.e. the first release of integrated financial and non-financial euro area accounts for the institutional sectors – namely households, non-financial corporations, financial corporations and the government – and for the transactions with the rest of the world. These euro area accounts will provide a wide-ranging overview of the euro area economy and enable a much more comprehensive analysis to be made of the interactions among sectors and between non-financial and financial developments. In turn, this will permit an even more comprehensive cross-checking of the economic and monetary pillars of the ECB’s monetary policy analysis. Furthermore, these accounts will provide users for the first time with regular key economic indicators for the euro area, such as household savings and corporate after-tax profits. They will be compiled in close cooperation between the ECB, the NCBs, Eurostat and the NSIs, and will be developed in two steps. The first annual accounts will be published soon and the inaugural release of the quarterly euro area accounts is expected in the spring of next year.
These and other recent enhancements in euro area statistics are not only of great importance for policy-making in Europe but also contribute to a better analysis of the global economy. However, continuing globalisation poses some particular challenges to statistics, as has been clearly demonstrated by the excellent contributions to this conference. Drawing from these contributions, I will now turn to the themes you have discussed.
Are financial statistics good enough to capture globalisation?
Globalisation stems from economic interlinkages, which are statistically reflected in the balance of payments and international investment position statistics and in the “rest-of-the-world” accounts of the national accounts framework.
In a world of increasing financial integration, statistics on cross-border transactions and positions also gain in importance. However, at present the analysis of international economic inter-linkages is hampered by the fact that, for all pairs of countries, bilateral transactions statistics do not mirror each other; this is the asymmetry problem. Related to this, the sum of the financial transactions in the balance of payments typically does not match the sum of the non-financial transactions, the difference being called “errors and omissions”. To eliminate both types of discrepancy, it seems best to employ a multi-pronged strategy. At least, it would be expedient if the legal framework protecting statistical confidentiality were to allow the exchange of such confidential information among the compilers concerned, for the strict purpose of statistical quality improvement.
Globalisation can only be analysed in a meaningful way if the concepts, definitions and classifications underlying the statistics are standardised. Fortunately, this is largely the case, as these standards are laid down in, for example, the 1993 System of National Accounts and the IMF Fifth Balance of Payments Manual. These global guidelines are being updated and the ECB attaches particular importance to the reinforcement of their mutual consistency and to the outcome of this process in general. For instance, it is of key importance that these statistical standards prescribe a faithful representation of economic reality as reflected in monetary transactions and do not impute an artificial price to non-monetary transactions or recognise assets and liabilities that are not viewed as such by the economic actors. It should then also be acknowledged that welfare cannot be adequately captured by any single statistic, including GDP.
The revised statistical standards should also reflect the fact that modern economies are knowledge economies, in which the quality of the labour force is the key success factor. This requires the regular compilation of disaggregated labour accounts as an integral part of the national accounts. This, in turn, would substantially enhance the analysis of productivity growth and shed a better light on the evolving global distribution of labour, in which skill levels play a crucial role.
In general, the statistics not only attempt to capture the ever increasing globalisation, but they are also affected by it. This is a subject to which I now turn.
Challenges for national data collection in a global world
Globalisation has certainly benefited from a progressive deregulation of cross-border transactions. At the same time, though, this deregulation has adversely affected the quality of national statistics. For instance, equal prices for cross-border and national payments in euro are difficult to match with widely divergent reporting obligations related to these two types of payment. Traditionally, however, many balance of payment reporting systems were based on cash settlements, and a transition to a more direct reporting system requires a substantial investment for the statistics compiler concerned, which is typically the NCB.
More generally, globalisation has intensified competition among countries and this has also led to an increasing emphasis on the administrative burden put on business. Although the statistical response burden accounts for at most a few percent of the total administrative burden, it has also come under close scrutiny. Indeed, statistics compilers should constantly attempt to minimise the costs for respondents and to make use of already existing information to the maximum extent. This requires, however, that statisticians have unrestricted legal access to all relevant administrative records, while of course maintaining their very strict confidentiality standards. This applies, for instance, to all tax records, including the cross-border data that become available through the EU Savings Directive. This being said, I am convinced that it is not possible to obtain good quality, timely statistics on the basis of administrative records alone. In other words, the high-quality statistics that are indispensable for the ECB’s monetary policy unavoidably place some burden on business, including small and medium enterprises, even if this burden has to be minimised. It is therefore important that this data collection modality is adequately anchored in a legal framework, in order to guarantee representative survey results. At the same time, as regards the collection of data from the financial sector, supervisors and statisticians should intensify their cooperation in order to align, and wherever possible integrate, their data collection efforts to a maximum extent.
Whereas globalisation complicates the compilation of national statistics and also makes them somewhat less meaningful, particularly in those fields where the decisions are now made at the euro area level, the collection of euro area statistics should continue to make the best possible use of the enormous statistical expertise at the NCBs and the NSIs. However, this does not imply that for many years to come each and every statistic should be compiled by the relevant authorities in each and every country. Instead, it is perhaps time to draw conclusions from the fact that the collection of statistics, particularly the increasingly automated retrieval of statistical data from the ledgers of larger corporations, is typically characterised by significant economies of scale. Similarly, significant efficiency gains would be reaped if sample sizes for statistics, particularly the very timely, high frequency statistics, are attuned to European needs. It goes without saying that countries that still want to spend additional resources to maintain nationally representative samples will be free to do so.
Statistical implications of financial stability and financial integration
Although this conference has devoted a single session to the statistical requirements for financial stability and financial integration analysis, these requirements are not easily aligned at first sight. In fact, financial stability analysis puts new demands on data availability, if only because it requires consolidated data and strengthens the case for making available a whole range of micro-data. At the same time, however, macro-statistics are typically much more reliable than micro-data, particularly those in the tail of the distribution, if only because significant misreporting by an “average” firm almost invariably means that it ends up in the tail. In any case, this analysis would be served by the setting-up of reliable and flexible pan-European micro databases. In fact, the ECB, in close cooperation with the NCBs, embarked on such an ambitious project several years ago, when it designed the Centralised Securities Database (CSDB) to contain an accurate representation of individual securities issued, or potentially held, by euro area residents.
Another requirement for financial stability analysis is related to the increasing tendency to transfer risks among institutions and among countries. This also points to important remaining gaps in our current euro area statistics, particularly concerning insurance corporations and pension funds, the collateralisation of lending business, investment funds (which include hedge funds), and special purpose entities (which include securitisation vehicles).
Concerning statistical indicators of financial integration, I have already alluded to the ECB’s contribution and I look forward to the release of an enlarged set of indicators this autumn. In addition, the contribution of the ECB to the short-term paper (STEP) market in Europe could be mentioned, in particular through the regular publication of statistics on volumes and prices.
Global statistical governance
I have given some examples of trends that all point to a more global, and particularly more European, coordination of statistics. It is conceivable that over time the European statistical systems will develop in the direction of genuine systems with regional competence centres, coordinated by the European institutions. The mutual interdependence of statistics compilers would then increase, which in turn calls for a further strengthening of international statistical governance. In this regard, it may be mentioned that the ECB has subscribed to a set of universal core principles for international statistics. Similarly, it may be worthwhile exploring an elaboration of the Eurosystem mission statement in the field of statistics.
Good governance is particularly important for statistics that have a direct administrative and policy use in Europe, such as the indicators for assessing the convergence of EU Member States that do not yet belong to the euro area and the statistics underlying the excessive deficit procedure. While Eurostat is responsible for these statistics, it can benefit, when deciding upon conceptual issues, from the considerable statistical expertise available at the ECB, the NCBs and the NSIs, as brought together in the CMFB. In any case, the compilation of government finance statistics – and all other statistics for that matter – should not be subject to political considerations. The compilers of these statistics must act in full scientific independence, and the government must ensure that sufficient resources are made available for their task. At the same time, it is important to regularly review the cost/benefit ratio of existing statistics that were agreed upon many years ago.
Ladies and gentleman, credible euro area statistics lie at the heart of the ECB’s monetary policy-making. At the same time, these statistics are embedded in and affected by a globalising world. The key to meeting the challenges this poses lies with further enhanced international cooperation among statistics compilers. Another key success factor is continuous cross-fertilisation between users and producers of statistics. An example of such cooperation in the near future will be the first release of integrated financial and non-financial euro area accounts for the institutional sectors, which will permit an even more comprehensive cross-checking of the economic and monetary pillars of the ECB’s monetary policy analysis. In addition, users and policy-makers should keep in mind that indispensable high-quality statistics unavoidably place some burden on business, including small and medium enterprises, although this burden can be minimised if statisticians obtain appropriate access to administrative data collected for other purposes.
This conference has indeed provided a significant contribution to the cross-fertilisation between users and producers of statistics and I would like to thank you all very much for your participation and for your attention.