1
Introduction and Major Outcomes of the Workshop
Goals of the Workshop
The financial reform plans currently under discussion in the United States recognize the need for monitoring and regulating systemic risk in the financial sector. To inform those discussions, the National Research Council held a workshop on November 3, 2009, to identify the major technical challenges to building such a capability. The Workshop on Technical Capabilities Necessary for Regulation of Systemic Financial Risk was organized in response to the letter of August 27, 2009, from Senator Jack Reed of the Senate Banking, Housing, and Urban Affairs Committee to the National Academies specifically requesting an appraisal of the data and analytical tools available for systemic risk regulation (Appendix A). Senator Reed identified the following key issues that need to be examined “in considering reforms to financial oversight,” specifically with respect to systemic risk:
-
What data and analytical tools are currently available to regulators to address this challenge?
-
What further data-collection and data-analysis capabilities are needed?
-
What specific resource needs are required to accomplish the task?
-
What are the major technical challenges associated with systemic risk regulation?
-
What are various options for building these capabilities?
Because every systemic event is unique with respect to its specific pathology—the various triggers and the propagation of effects—the workshop focused on the issues listed above for systemic risk in general rather than for any specific scenario. Thus, by design, the workshop explicitly addressed neither the causes of the current crisis nor policy options for reducing risk, and it attempted to steer clear of some policy issues altogether (such as how to allocate new supervisory responsibilities). More than 40 experts representing diverse perspectives participated in the workshop (Appendix B).
Some Underlying Observations
A basic observation shared by several workshop discussants is that recent decades have seen rapid change in the financial system—driven by innovation and deregulation—that has altered the mechanisms and pace of financial intermediation to such an extent that regulatory tools, processes, and data have fallen behind. The far more numerous and increasingly complex linkages among financial institutions of all types, with essential linkages extending beyond the banking sector and beyond domestic U.S. institutions, suggested to many discussants that the monitoring of systemic risk has become a more urgent and far more complex problem than in the past. George Sugihara of the Scripps Institution of Oceanography emphasized the importance of understanding systemic risk holistically as a “dynamic,” nonlinear problem, as opposed to atomistically as a decomposable, static problem that can be addressed by simply aggregating risks across independent firms.
It was widely acknowledged at the workshop that the United States currently lacks the technical tools to monitor and manage systemic financial risk with sufficient comprehensiveness and precision. While some of the building blocks are available, many workshop participants pointed to major gaps that remain. Andrew Lo of the Massachusetts Institute of Technology presented a simple mnemonic for capturing the range of information that a systemic risk regulator will need to monitor: namely, the “four L’s” of leverage, linkages, liquidity, and losses across the financial system. Assembling a holistic perspective will require significant additional data as well as new models and research. Myron Scholes of Stanford University pointed out that even with information on leverage and linkages, liquidity and losses can only be simulated with interacting models. Other elements identified in the workshop were capital, maturity mismatch, and risk concentrations.
Christine Cumming of the Federal Reserve Bank of New York added that risk at the firm level cannot be truly assessed unless the much broader context of overall risk positions and risk dynamics in the financial system is understood. Decision makers at financial institutions, given access to more reliable knowledge about their total risk exposure with respect to proposed actions, should be better able to manage those risks. This more complete understanding could provide help to the following:
-
Institutions, in recognizing how they share in creating and being affected by systemic risk;
-
Markets, in setting values; and
-
Regulators charged with moderating markets and firms.
Market efficiency will be enhanced by improved intelligence about what is going on in the system as a whole. Yaacov Mutnikas of Algorithmics observed that risk analysis has developed almost exclusively to manage firm-specific risks, and that the aggregate of firm risk is not necessarily equal to systemic risk. Firm-based analysis ideally takes into account the market responses and stresses that information about losses in other financial firms produces, so it provides partial analysis of feedback effects. Full analysis of system risks, however, must incorporate more complex interactions, which, as recent experience has shown, can be especially dangerous. Furthermore, individual firms have their own scenarios of concern, which are not necessarily those of greatest significance to the overall system. Thus, to manage systemic risk, new analysis capabilities and appropriate data will be needed.
Major Points for Policy Makers
Although this one-day workshop was not aimed at developing consensus conclusions, there were some recurring themes that are relevant for policy makers:
-
Today’s tools of financial risk analysis will need to be augmented to provide information needed for the regulation of systemic financial risk. As implied above, existing capabilities to value individual instruments and manage firm-specific risks and capture system-wide exposures are not a sufficient foundation for systemic risk management.
-
The new understanding that is necessary for systemic risk management calls for new, or extended, mathematical models. These models would be designed to capture better the extensive linkages among firms and markets, the dynamic interactions among the firms
-
and markets, and the potential for any of these to change according to the state of the system (e.g., in shifting from normal times to distressed times).
-
Similarly, the new understanding of systemic risk management will require the use of more and better data. The creation and validation of those models will rely on some data that are not currently used and perhaps not currently available. However, there were a variety of views expressed at the workshop as to what data should be collected. One view argued that complete transaction records in cash and derivatives markets would be the appropriate level of data collection. A more widely held view was that no one knows enough to say what data are needed.
-
A new understanding of systemic risk management is just beginning to develop. Inevitably, systemic risk management capabilities will be built up iteratively, starting with the imperfect data and models that are currently available and refining both as research improves our understanding.
Several workshop participants commented that the questions in Senator Reed’s letter are valid and generally approachable and that they could be answered through a careful study, which would constitute a solid step toward setting up a systemic regulatory capability. Nevertheless, it will take a long-term, multidisciplinary effort to build up this capability fully, and the structure is not in place yet for that effort. A good deal of research is needed to guide the development of effective systemic financial risk regulation.