Notes from the Vault
Larry D. Wall
April 2014
Policymakers need access to good information if they have any hope of reducing the probability and cost of financial instability. The importance of information was highlighted during the recent financial crisis, as both government policymakers and private markets were frequently surprised (to the downside) by new information about risks and exposures. This year's Atlanta Fed's Financial Markets Conference, cosponsored by the Center for Financial Innovation and Stability, sought to draw attention to the changing informational environment in financial markets with a panel session titled "Seeing Clearly: Will We View the Next Curve in the Road with More Information?"1 This session looked at the extent to which we will have a clearer picture of the threats to financial stability. The panel discussion highlighted some significant improvements in the information environment since 2008 but also some significant, and as yet unaddressed, needs. The panel concluded with some important warnings about unavoidable limits on the quality and usefulness of the information.2
Importance of information before and during crises
Supervisors and market participants do not necessarily agree on the value of highly detailed information prior to a crisis. "Ignorance is bliss" is the phrase that Til Schuermann, a partner at Oliver Wyman, borrowed from MIT Professor Bengt Holmstrom to describe the private-sector benefits of shared ignorance in good times. All else being equal, the volume of trade and liquidity is much higher in securities where market participants do not need to worry that the other side of the deal has private information about the security's value. In good times, this can be done by common agreement that certain debt securities are so low risk that there is no value in gathering additional information.3
In good times as well, prudential supervisors are supposed to be looking for future problems. Here a lack of information can be an acute problem. In his presentation, Andreas Lehnert, a Federal Reserve Board deputy director, pointed to a lack of such basic information as what percent of residential real estate borrowers were "borrowing" their down payment using second mortgages (sometimes called "piggyback mortgages") before the financial crisis. Clearly, this type of information is important for valuing the mortgages themselves but also for an assessment of the aggregate risks building in the financial system.
After a crisis has started, market participants and supervisors become keenly aware of what they don't know and eagerly seek additional information to guide their responses. Supervisors cannot take effective action until they understand what the problems are and how these problems can propagate through the financial system. Schuermann, Lehnert, and Patricia Mosser (a deputy director at the U.S. Treasury's Office of Financial Research) were all part of the Federal Reserve team trying to mitigate the damage in 2007–09, and each of their talks referenced their anxious search for key information on financial sector developments during the crisis.
Similarly, ignorance is no longer bliss for private investors during a crisis, as they realize that some people have very valuable private information about asset values. Market liquidity dries up as private investors become more and more uncertain about asset values and no longer trust the supervisors. Information release is critical at this stage, according to Schuermann, to "Transform uncertainty into risk" because investors say, "That (risk) I can manage!" Additionally, markets must regain trust that the supervisors understand the condition of key intermediaries and markets.
Improved information
The gaps in supervisory information have been substantially reduced, according to Lehnert and Patricia Mosser. Lehnert pointed to the enormous amount of data being gathered on large banking organizations by supervisors through the capital stress testing process and through increased monitoring of liquidity. Mosser noted these efforts and also referenced several other new sources such as transactions data from swap execution facilities and securitization securities disclosure.
Additionally, the accounting authorities have changed accounting rules in an attempt to provide investors with a more accurate picture of the firm. One example of a set of rules that allowed misleading accounting treatment are those associated with loan securitization. Lehnert noted that the supervisors received an unpleasant surprise at the amount of loans supposedly sold through securitization that returned to some banks' balance sheets during the crisis. Accounting authorities have since changed the rules on accounting for securitization, according to University of North Carolina Professor Robert Bushman, to prevent this misleading treatment.
But fewer data gaps does not mean no data gaps. As one example, Lehnert pointed to inadequate information on some markets, such as that for repos (repurchase agreements). Mosser observed that merely having a lot of data may not be very helpful if the data are not based on standard definitions or lack common identifiers that allow it to be aggregated in meaningful ways. She discussed efforts by the Office of Financial Research to improve both.
Governance of financial data
Banks' financial statements are critical to both supervisors' and private investors' understanding of conditions. Yet Bushman points out these statements are not the underlying economic reality, but rather a depiction of that reality. Some items can be defined and measured in relatively uncontroversial ways, such as the cash in a bank's vault. For some other items, the definition can be controversial, such as the precrisis securitization rules that allowed banks to create the appearance but not the reality of having sold the credit risk of a pool of loans. Further, even if an item is well defined, its measurement can be difficult and depends in part on how firms exercise the discretion provided by the accounting standards.
Bushman discussed two theories of accounting discretion that have conflicting implications. Senior management under one hypothesis uses its discretion to signal its private information to investors and help investors form better estimates of the firm's true value. The other hypothesis holds that management uses its accounting discretion to further the personal interests of senior management. Given the many opportunities to exercise discretion, each of the hypotheses could be accurate for some group of firms in some circumstances. To get a better sense of which hypothesis seemed more accurate for banks' loan loss accounting, Bushman reviewed the accounting literature on banks that delayed expected (loan) loss recognition. That literature finds that banks delaying loss recognition subsequently had a higher risk of a severe drop in equity and greater financing frictions during downturns. These findings suggest that bank managers were using their discretion to further their own interests rather than helping investors better evaluate their bank's likely losses.
Thus, Bushman's discussion highlighted the dangers of relying solely on financial statements' depiction of a firm's financial condition as an accurate measure of its underlying economic condition. Bank examinations by prudential supervisors and stress tests conducted by supervisors can mitigate the limitations inherent in relying on financial statements, but they cannot eliminate the problem.
Identifying stability risks in good times
While adequate, accurate, and timely data are necessary for evaluating risks prior to a crisis, these conditions are not sufficient, as highlighted in Lehnert's presentation. He pointed to the need also to have sufficient imagination to contemplate potential high-loss scenarios. As an example, he asked the provocative question of whether investors thought, in 2005–07, that house prices could decline 30 percent. His research with Kristopher Gerardi, Shane M. Sherlund, and Paul Willen came back with a resounding no. A Lehman Brothers analyst report in 2005 had as its worst case that housing prices would decline 5 percent per year for three years followed by 5 percent annual appreciation thereafter. This scenario was called "Meltdown." Indeed, even as house prices declined from October 2006 to July 2007, Lehnert quotes analysts who were talking about signs of price "stabilization" and prices that were "near bottom." Not until September 2007 did one analyst title the report "UGLY! Double digit declines in August and September."
Lehnert's recommendation was to make information collection more flexible and to look for signs of problems, such as industries where firms have been earning "supernormal profits" over a substantial period of time. His view is similar to my recommendation that financial markets should be subject to periodic reviews with more frequent and intensive reviews targeted at markets growing rapidly, experiencing rapid volume growth in new financial instruments, and/or experiencing rapid price appreciation in the collateral underlying the market. Including markets where participants are earning supernormal profits would be an excellent addition to the set of triggers for more intensive reviews.
Conclusion
With these changes and new awareness, will supervisors and investors be "Seeing Clearly: Will We View the Next Curve in the Road with More Information?" There is reason to believe that we will all be more prepared than before, but likely never as prepared as we would like to be.
The session clearly highlighted improvements in the information environment that will allow supervisors and investors to have a better understanding of the financial system going forward. However, the session also pointed out several limitations of the data. Investors do not necessarily want, and will not generate, an abundance of information in good times if that reduces liquidity. Financial statements continue to be the representations of senior managers who have an incentive to be opportunistic in using their accounting discretion. Finally, even accurate and complete data are not enough absent an ability and willingness to imagine crises scenarios.
Given these limitations on our ability to use data to anticipate future problems, the goal of preventing all future financial crises seems unobtainable. However, the combination of more data, better data, and better strategies for imagining future problems offers the potential for reducing the likelihood and cost of future crises.
Larry D. Wall is the executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Paula Tkac for helpful comments on the paper. The views expressed here are the author's and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please e-mail atl.nftv.mailbox@atl.frb.org.
_______________________________________
1 Papers and slide presentations from the Financial Markets Conference are available online.
2 Edward DeMarco, a senior deputy director at the Federal Housing Finance Authority, also talked about the importance of data standards in a different context. In his keynote speech, DeMarco highlighted how different data mortgage application standards between Fannie Mae and Freddie Mac both increased the risk of data errors and reduced competition.
3 For an earlier Financial Markets discussion of the role of "information-insensitive" debt, see Gary Gorton's 2009 paper.