About


The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.

Authors for macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.


« Improvement in the Outlook? The BIE Panel Thinks So | Main | Building a Better Jobs Calculator: Choose Your Own Payroll/Household Employment Ratio »

April 22, 2013


Too Big to Fail: Not Easily Resolved

As Fed Chairman Ben Bernanke has indicated, too-big-to-fail (TBTF) remains a major issue that is not solved, but “there’s a lot of work in train.” In particular, he pointed to efforts to institute Basel III capital standards and the orderly liquidation authority in Dodd-Frank. The capital standards seek to lower the probability of insolvency in times of financial stress, while the liquidation authority attempts to create a credible mechanism to wind down large institutions if necessary. The Atlanta Fed’s flagship Financial Markets Conference (FMC) recently addressed various issues related to both of these regulatory efforts.

The Basel capital standards are a series of international agreements on capital requirements reached by the Basel Committee on Banking Supervision. These requirements are referred to as “risk-weighted” because they tie the required amount of bank capital to an estimate of the overall riskiness of each bank’s portfolio. Put simply, riskier banks need to hold more capital under this system.

The first iteration of the Basel requirements, known as Basel I, required only 30 pages of regulation. But over time, banks adjusted their portfolios in response to the relatively simple risk measures in Basel I, and these measures became insufficient to characterize bank risk. The Basel Committee then shifted to a more complex system called Basel II, which allows the most sophisticated banks to estimate their own internal risk models subject to supervisory approval and use these models to calculate their required capital. After the financial crisis, supervisors concluded that Basel II did not require enough capital for certain types of transactions and agreed that a revised version called Basel III should be implemented.

At the FMC, Andrew Haldane from the Bank of England gave a fascinating recap of the Basel capital standards as a part of a broader discussion on the merits of complex regulation. His calculations show that the Basel accords have become vastly more complex, with the number of risk weights applied to bank positions increasing from only five in the Basel I standards to more than 200,000 in the current Basel III standards.

Haldane argued that this increase in complexity and reliance on banks’ internal risk models has unfortunately not resulted in a fair or credible system of capital regulation. He pointed to supervisory studies revealing wide disparities across banks in their estimated capital requirements for a hypothetical common portfolio. Further, Haldane pointed to a survey of investors by Barclays Capital in 2012 showing, not surprisingly, that investors do not put a great deal of trust in the Basel weightings.

So is the problem merely that the Basel accords have taken the wrong technical approach to risk measurement? The conclusion of an FMC panel on risk measurement is: not necessarily. The real problem is that estimating a bank’s losses in unlikely but not implausible circumstances is at least as much an art as it is a science.

Til Schuermann of Oliver Wyman gave several answers to the question “Why is risk management so hard?” including the fact that we (fortunately) don’t observe enough bad events to be able to make good estimates of how big the losses could become. As a result, he said, much of what we think we know from observations in good times is wrong when big problems hit: we estimate the wrong model parameters, use the wrong statistical distributions, and don’t take account of deteriorating relationships and negative feedback loops.

David Rowe of David M. Rowe Risk Advisory gave an example of why crisis times are different. He argued that the large financial firms can absorb some of the volatility in asset prices and trading volumes in normal times, making the financial system appear more stable. However, during crises, the large movements in asset prices can swamp even these large players. Without their shock absorption, all of the volatility passes through to the rest of the financial system.

The problems with risk measurement and management, however, go beyond the technical and statistical problems. The continued existence of TBTF means that the people and institutions that are best placed to measure risk—banks and their investors—have far less incentive to get it right than they should. Indeed, with TBTF, risk-based capital requirements can be little more than costly constraints to be avoided to the maximum extent possible, such as by “optimizing” model estimates and portfolios to reduce measured risk under Basel II and III. However, if a credible resolution mechanism existed and failure was a realistic threat, then following the intent of bank regulations would become more consistent with the banks’ self-interest, less costly, and sometimes even nonbinding.

Progress on creating such a mechanism under Dodd-Frank has been steady, if slow. Arthur Murton of the Federal Deposit Insurance Corporation (FDIC) presented, as a part of a TBTF panel, a comprehensive update on the FDIC’s planning process for making the agency’s new Orderly Liquidation Authority functional. The FDIC’s plans for resolving systemically important nonbank financial firms (including the parent holding company of large banks) is to write off the parent company’s equity holders and then use its senior and subordinated debt to absorb any remaining losses and recapitalize the parent. The solvent operating subsidiaries of the failed firm would continue in normal operation.

Importantly, though, the FDIC may exercise its new power only if both the Treasury and Federal Reserve agree that putting a firm that is in default or in danger of default into judicial bankruptcy would have seriously adverse effects on U.S. financial stability. And this raises a key question: why isn’t bankruptcy a reasonable option for these firms?

Keynote speaker John Taylor and TBTF session panelist Kenneth Scott—both Stanford professors—argued that in fact bankruptcy is a reasonable option, or could be, with some changes. They maintain that creditors could better predict the outcome of judicial bankruptcy than FDIC-administered resolution. And predictability of outcomes is key for any mechanism that seeks to resolve financial firms with as little damage as possible to the broader financial system.

Unfortunately, some of the discussion during the TBTF panel also made it apparent that Chairman Bernanke is right: TBTF has not been solved. The TBTF panel discussed several major unresolved obstacles, including the complications of resolving globally active financial firms with substantial operations outside the United States (and hence outside both the FDIC and the U.S. bankruptcy court’s control) and the problem of dealing with many failing systemically important financial institutions at the same time, as is likely to occur in a crisis period. (A further commentary on these two obstacles is available in an earlier edition of the Atlanta Fed’s Notes from the Vault.)

Thus, the Atlanta Fed’s recent FMC highlighted both the importance of ending TBTF and the difficulty of doing so. The Federal Reserve continues to work with the FDIC to address the remaining problems. But until TBTF is a “solved” problem, what to do about these financial firms should and will remain a front-burner issue in policy circles.

Photo of Paula Tkac By Paula Tkac, vice president and senior economist, and

Photo of Larry Wall Larry Wall, director of the Center for Financial Innovation and Stability, both in the Atlanta Fed’s research department

April 22, 2013 in Capital and Investment , Capital Markets , Financial System | Permalink

TrackBack

TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00d8341c834f53ef017d43058473970c

Listed below are links to blogs that reference Too Big to Fail: Not Easily Resolved :

Comments

Yes, solving the Too Big to Fail problem is really, really difficult if the people making policy don't want to solve it, as they evidently don't. Clearly, the political influence of such organizations is at work impeding the process. Sadly, the regulators/enablers (both international and domestic) are terrified of imposing the kind of "dumb but simple" workable solution that would drastically impact the business models of these very politically powerful organizations. No, let's keep putting increasingly sophisticated band aids on tumors--that's the spirit, fellows!

Posted by: William Meyer | April 23, 2013 at 09:37 AM

Nice summary of the problems here. In addition to the problems you mention, I worry about a couple of others. One was noted recently by Sarah Gordon in the FT. She writes:


There is a compelling body of evidence suggesting that the people most likely to go into the riskier areas of financial services are precisely those least suited to judging risk. Susan Cain’s recently published book Quiet cites a series of studies that suggest that extroverts tend to be attracted to the high-reward environments of investment banking, deals and trading. And, troublingly, these outgoing people also tend to be less effective at balancing opportunity and risk than some of their more introverted peers. (http://www.ft.com/intl/cms/s/0/a917de18-abef-11e2-9e7f-00144feabdc0.html#axzz2Rxxxb9nc)

A second thing that makes me nervous is the fact, as you point out, that the FDIC an exercise its new powers to resolve a systemically important institution only if the Fed and the Treasury agree to let it do so. Would the Treasury really be able to resist political pressures and make such a decision objectively? Even the Fed, independent though it is, might find it hard to do so.

Posted by: Ed Dolan | April 30, 2013 at 12:56 PM

There is waiting in the Platonic cave a critically important paper that derives Campbell's Law and Goodhart's Law from Goedel's incompleteness theorem. As long as we allow a financial system to exist in which ever more complex debt instruments are allowed to be created, then for every control regime devised by some Basel XXXIV or whatever, there will be a new class of derivatives that escapes those regulations and destabilizes the system.

There are three possible ways to escape this endless escalation of risk-creating games: (1) impose a formal language for creating financial instruments that has been proven to be sub-Turing in its expressive power, (2) implement institution-scale TBTF, preventing any single institution or implicit consortium of institutions from acquiring enough significance to endanger the system as a whole, or (3) implement sovereign TBTF, preventing the failure of any nation-state from endangering the financial systems of others. It's not too late to start considering mechanisms to allow the Eurozone to spin off disruptive members in the same way that a lizard sheds its tail when in mortal danger. Cyprus is almost there already.

Unfortunately the US cannot spin off insolvent states, much as some of them might wish to leave. If only we could let North Carolina go, and take the Bank of America with it!

Good luck!

Posted by: George McKee | April 30, 2013 at 10:04 PM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

Google Search



Recent Posts


Archives


Categories


Powered by TypePad