The financial crisis exposed a number of gaps in our ability to monitor systemic risks. Addressing even a few of these may help to reduce the risk of future crises. The crisis has revealed at least five elements which matter in the monitoring of systemically important financial institutions:
Efficiency demands that we draw on the existing reporting frameworks for our global statistics, but with better data design, and with more coordination across organisations. For its part, the BIS stands ready to help wherever possible. The BIS already collects statistics on the international activities of large banks - the so-called BIS banking and derivatives statistics - and will continue to push for improvements to these statistics.
Thank you all for participating in today's conference, and welcome to Basel and to the BIS. The immediate task that faces this group over the next two days is to draw up a concrete plan to implement the 20 recommendations on data gaps spelled out by the joint IMF-FSB report to the G20 last October. In my view, the broader task this group has set for itself - a rethinking and a redesign of the global data collection framework - is one of the most important things the international community can do to prevent the next crisis.
The financial crisis exposed many holes in our reporting system. Repairing even a few of these may help to significantly reduce the risk of future crises. And that, in the end, is all that we can really hope to accomplish.
To my mind, the crisis taught us at least five things which matter in the monitoring of systemically important financial institutions (SIFIs).
First, consolidation matters. We live in a world where the balance sheets of financial firms span the globe. Yet the aggregate statistics collected around the world have paid little attention to the actual geography of these firms' operations. One must keep in mind that balance sheet stresses build up across the consolidated balance sheet, in the form of poor asset quality as well as currency, maturity or interest rate mismatches between assets and liabilities. Inability to "see" the consolidated balance sheet, either at the individual bank level or at the headquarter country level, means that the build-up of stresses at the systemic level cannot be monitored.
Thus, a key lesson of the crisis is that the appropriate unit of analysis is banks' worldwide consolidated positions. At the level of individual institutions, some information on banks' consolidated global balance sheets is available in their regular financial reporting, and this information is collected in BankScope and Bloomberg. However, these disclosures lack the essential breakdowns with which to measure balance sheet stresses: that is, information on the currency, maturity and counterparty type, for both asset and liability positions, along with off-balance sheet exposures. In short, the publicly available information on individual banks' international positions falls far short of what is needed for sound monitoring.
At the national level, statistics compiled by national authorities, the IMF, the OECD and the BIS do not provide a complete picture either. For example, the flow of funds statistics, the balance of payments statistics, the IMF's Coordinated Portfolio Investment Survey and the BIS locational banking statistics all rely on residency-based data. Such data are insufficient for identifying vulnerabilities in any particular consolidated national banking system. Currently, the BIS banking statistics come closest to providing the needed consolidated picture of banks' international positions (for 20 or so national banking systems), albeit not at the level of individual banks.
Second, liabilities matter. It was uncertainty about the scale of losses on banks' assets that was the proximate cause of the crisis. However, it was the dislocations in banks' funding markets that made this a global crisis. Funding in the interbank market, and from non-bank money market funds, became impossible for all but the shortest terms; funding in the repo market became available only against high-quality collateral; and funding in the swap market became much more expensive. In short, major dislocations occurred in every important short-term funding market. Yet we cannot really "see" any of these markets in our aggregate data. And when we cannot see them, we cannot assess the degree of maturity mismatch embedded in the system.
Third, currency matters. The funding crisis was really a crisis of dollar funding. The flaw in the funding models of so many banks was that they failed to appreciate the hidden vulnerabilities in the excessive maturity mismatch in their funding of US dollar assets. The off-balance sheet borrowing of dollars through FX swaps that they relied on covered the exchange rate risk. But the short-term tenor of these instruments left them vulnerable to rollover risk, and liquidity in this market was simply taken for granted. No measure of effective maturity mismatch is possible if cross-currency funding positions are not taken into account.
Fourth, interconnectedness matters. The degree of interconnectedness of a particular financial institution - and not only its size - is a key indicator for measuring systemic importance. To measure this, we need information on bilateral interbank liabilities, and on the system-wide aggregate exposures to particular counterparties. But such data on bilateral exposures are reported only to supervisors because of confidentiality and legal restrictions.
Fifth, non-banks matter as well. Off-balance sheet entities such as structured investment vehicles (SIVs) obscured the build-up of stresses in the financial system, and they exacerbated the problems when they had to be moved back onto banks' balance sheets. Moreover, other non-banks - in particular, pension funds, insurance companies and large corporates - should not be excluded from systemic monitoring exercises.
Without the ability to see, at various levels of aggregation, the on- and off-balance sheet consolidated positions of large bank and non-bank financial firms, and at least some measures of the maturity mismatch on their balance sheets, we cannot hope to really understand the risks at a systemic level.
Of course, new and better data come with new and possibly higher costs, many of which will be borne by taxpayers and reporting firms around the world. The concrete agenda which ultimately emerges here must bear this fact in mind.
Each of the organisations represented in this room already collects and disseminates data covering some important piece of the international financial system. Much of these data ultimately come from reporting agencies in member countries. Efficiency demands that we draw on the existing reporting frameworks for our global statistics, but with better data design, and with more coordination and sharing of information across organisations.
For its part, the BIS stands ready to help wherever possible. The BIS already collects and disseminates statistics on the international activities of large banks - the so-called BIS banking and derivatives statistics. These statistics cover the international activities of more than 7,000 reporting entities - or roughly 95% of all international bank assets - and are based on underlying data reported by more than 40 central banks. This BIS / central bank reporting framework has functioned smoothly for more than 30 years, giving the BIS considerable experience in the handling of confidential international banking data. This framework has also proved to be flexible, leading to improvements to the reported data, including derivative exposures. The BIS, along with the CGFS, will continue to push for improvements to these banking and derivatives statistics.
To conclude, we look forward to two days of productive discussions, and I wish you all the best for a successful event. Thank you for your attention.