Professor of Finance
Leonard Stern School of Business, Kaufman Management Center
New York University
Co-Rapporteur, Workshop on the Technical Capabilities Necessary for Systemic Risk Regulation
Board on Mathematical Sciences and Their Applications
National Research Council
The National Academies
Subcommittee on Security and International Trade and Finance
Committee on Banking, Housing and Urban Affairs
February 12, 2010
Good morning, Mr. Chairman and members of the Subcommittee. My name is Robert Engle. I am a Professor at the Leonard Stern School of Business at New York University. I recently co-authored a report of the National Research Council that summarized a workshop on Technical Capabilities Necessary for Systemic Risk Regulation. The Research Council is the operating arm of the National Academy of Sciences, National Academy of Engineering, and the Institute of Medicine, chartered by Congress in 1863 to advise the government on matters of science and technology. This workshop and its report were sponsored by the Alfred P. Sloan Foundation. After I summarize our report, I would like to give you my personal response to the questions posed in your letter as well as the pending legislation on the National Institute of Finance.
Key findings of the National Research Council’s report are as follows:
1. With better data and better analytical tools, the participants were confident that we could measure, monitor and ultimately lessen systemic risk.
2. Research to accomplish this goal is already underway in academic and regulatory institutions and is being presented in scholarly and practitioner meetings. This research is now primarily based on publicly available market, size and leverage data. The research however has far to go.
3. Additional data across asset classes and with counterparty, position, collateral and valuation information would be very helpful. But many participants at our meeting said it was not yet possible to determine which specific data would be needed.
4. Compilation and analysis of such data would require standardization and classification that does not yet exist and which would be valuable for the industry as well as for the regulators.
5. Data alone will not be sufficient to understand risk, illiquidity, bank runs, bubbles and other features central to assessing systemic risk. These can only be examined within models, and research is needed to develop some of those models and improve others. As Andrew Lo of MIT summarized, systemic regulators need to know the four Ls: leverage, linkages, liquidity and losses.
I request that the full report be attached to my testimony.
In response to the specific questions and the pending legislation on the National Institute of Finance, I would like to make several points. These comments are my own, not the National Research Council’s. I have endorsed the concept of the National Institute of Finance in a letter from Harry Markowitz which is also signed by William Sharpe, Robert Merton, Myron Scholes, and Vernon Smith, all Nobel Laureates. These comments suggest the aspects I find particularly strong or weak about the proposal.
First, let me address the status of data available to regulators.
6. Data collection by supervisory agencies takes various forms. Some obtain detailed information in response to specific requests of entities that they supervise, but it is not automatically generated and reported. Because of privacy issues, these data cannot be shared even within a regulatory body or with other regulatory entities. Thus only partial views of all information are available.
7. Useful data sets exist within clearing houses, self-regulatory organizations such as DTCC or FINRA and surveillance operations in exchanges or regulators. These data will have positions, and counterparties at a fine level of disaggregation. Upon specific request, regulators have been able to access these data but generally cannot share them.
8. Risk reports that summarize major firm risks at a point in time are automatically submitted to supervisory agencies. However, these generally do not have counterparty information, and their information might not be freely available throughout the agency.
9. In summary, regulators have access to more data than the public, but the collection is fragmented, discontinuous and possibly quite delayed in time. Improved data availability would enhance all regulatory processes.
Further comments on the NIF.
10. The data collection and aggregation functions of the NIF proposal are staggering. Some aspects are likely to be more important than others. In particular, the OTC contracts are the most complex and have the greatest chance of being systemic. Many of these have substantial counterparty exposure so they are inherently more systemically risky. Thus, collection of OTC position and transaction data could be the most important starting point. Establishment of some standardization and common identifiers for contracts, as noted above, would be an important step for making such data more useful.
11. Systemic risk assessment requires selecting entities to monitor and then obtaining data to monitor them. Consequently, a great deal of data on entities that are not risky can be ignored, once the selection is accomplished. This selection process is not completely obvious, because small entities can sometimes pose systemic risks, so it needs to be done through careful analysis and revisited from time to time.
12. The independence of the NIF is important in order to insulate it from corporate or government pressures. However, this could be satisfied by an NIF that was embedded within a regulatory agency that was itself independent and had systemic regulatory oversight responsibilities. There could be important cost savings from such embedding as well.
13. International collaboration is of great importance. Systemic crises are global and the markets are global, hence regulators that only see domestic positions will miss important features. The NIF should immediately coordinate with relevant international agencies to collaborate on data acquisition and analysis. This coordination is important within the US and it is even more important for the smaller countries and regulators around the globe. Sovereign crises such as those experienced in the past in Ireland, Iceland, Latvia and Hungary and in Greece, Portugal and Spain today emphasize the interlinking of our financial system.
14. The security of these data must be ensured. Otherwise, compliance may be difficult to achieve.
15. Congress and the NIF should have as a goal making partially aggregated or delayed versions of the OTC data public so that counterparties can better price and manage the risks that they face. Better individual firm risk management contributes to better systemic risk management. Transparency in financial markets is a powerful supplement to regulation.
An archived webcast of the hearing can be found on the Senate Banking, Housing, and Urban Affairs Committee's Web site.