Living Wills or Phoenix Plans: Making sure banks can rise from their ashes

Wills are for when you die. Living wills guide your affairs when you lose the capacity to act.

We’re all mortal and fragile – not just people, but firms and banks, too. The Dodd-Frank Act of 2010 requires systemic intermediaries (and many others) to create “living wills” to guide their orderly resolution in bad times.

In August, these Dodd-Frank living wills made front-page business news when the FDIC and the Fed rejected those submitted by the biggest banks as inadequate. That should come as no surprise. In their current form, we doubt that living wills would do much to secure financial stability.

In the absence of special arrangements, the legal system creates a will for every company. When a firm dies, established bankruptcy procedures prescribe a court-tested priority of claims on which to base the distribution of its assets, allowing either liquidation or, where the firm still has value as a going concern, restructuring.

For a financial intermediary, the going-concern value collapses faster in death than that of a nonbank – with potential damage to the financial system as a whole. As a result, traditional bankruptcy procedures are widely viewed as ineffective for financial firms. Instead, beginning in the 1930s, U.S. laws established customized administrative techniques (like the now-established practices of the FDIC) to resolve small, simple, domestic banks. But even these special mechanisms don’t work for large, complex, international ones.

All of this has led to an enormous amount of thinking and discussion about how to resolve the largest intermediaries using living wills, contingent convertible capital, and bail-in of debt holders, among others things. This week, leading regulators from the United Kingdom and the United States will gather for a “war game” to test some of these evolving mechanisms for containing the failure of a big internationally active bank. Their efforts to make resolution credible under multiple (potentially conflicting) bankruptcy regimes are necessary and admirable. Yet, the problem will not go away quickly.

Put simply, some institutions likely remain too big to fail. A few numbers make this clear. If we adjust for accounting, at the end of 2013, there were 10 banks in the world with assets estimated at over $2 trillion, three of which are U.S.-based. (You can find a list of the top 100 banks in the world as of the end of 2013 here, and the accounting adjustments here.)

Size alone is not decisive. What makes an institution too big to fail is a combination of size, complexity and interconnectedness. As a proxy for complexity, consider the number of legal subsidiaries in the largest banking groups. According to Avraham, Selvaggi and Vickery, there are seven U.S. financial holding companies with more than 1,000 subsidiaries, many of which are outside the United States (see their chart that we reproduce below). JPMorgan Chase has over 3,000, as does Goldman Sachs. And both the Bank of America and Morgan Stanley have more than 2,000 legal subsidiaries each.

Source: Avraham et al. “A Structural View of U.S. Bank Holding Companies,” FRBNY Economic Policy Review, 2012.

Source: Avraham et al. “A Structural View of U.S. Bank Holding Companies,” FRBNY Economic Policy Review, 2012.

Dodd-Frank Title I sees living wills as the device to overcome the size-and-complexity challenge. The goal is to pre-plan the restructuring or liquidation of an inadequately capitalized financial institution without putting other intermediaries, or the financial system as a whole, under stress.

You can find public versions of these resolution plans on the Federal Reserve and the FDIC websites here and here. They do not make for reassuring reading. The posted materials typically do not specify the precise steps that a resolution manager – in this case, the FDIC – would have to take to restructure or liquidate the institution. Even if they did, these living wills are designed with the presumption that the rest of the financial system is reasonably healthy so that the resolution of the bank in question is an idiosyncratic affair. The plans offer no advice on how to limit systemic risk if a resolution must occur when the entire financial system is under stress, as was the case in the recent crisis.

We see three possible responses to this problem. Regulators can break up the large, complex banks now, so they are small and simple enough to resolve without putting the system at risk. They can require that banks issue so much capital that they will never be at risk of insolvency. Or, we can raise capital requirements to some extent, while also implementing a mechanism for virtually automatic recapitalization if systemic intermediaries nevertheless become insolvent.

If there are large economies of scale and scope in banking, as one recent analysis argues, then breaking up the biggest institutions would be socially costly. Not only that, their political power makes it exceedingly difficult for regulators to act – especially in good times when the systemic threat seems distant.

Like many academicians, we support a further substantial increase in capital requirements (see this earlier post), together with stringent stress tests, to contain systemic risks. However, increasing capital levels sufficiently to forestall any chance of a banking disruption would drive us toward a system of narrow banks with 100% equity capital. In our view, a narrow-banking regime would push risk-taking to shadow banks without making the financial system safer. Eventually, in a crisis, financial intermediation would shrivel, damaging the real economy.

This brings us to the third solution: a sharp increase of capital requirements combined with virtually automatic (or rules-driven administrator-assisted) recapitalization in bad times. This is a phoenix plan, rather than a dissolution plan. [Acharya, et al, in a chapter on “Resolution Authority” call this approach the “academic concept of a corporate living will.”]  If we can’t break up banks in good times or resolve them without intensifying systemic risks in a crisis, a phoenix plan would make them capable of rising from the ashes.

Like the bankruptcy code, such a rules-based plan would protect property rights and reduce uncertainty. By constraining the discretion of regulators, it also would make the promise to “bail in” the creditors of big institutions more credible, reducing the financial market distortions associated with “too-big-to-fail.” An economics textbook might say that a phoenix plan is more time consistent than today’s living wills and would highlight this example as one of many where rules beat discretion.

Imagine the following simple approach (like that of Acharya et al). Let the capital structure of a bank’s long-term liabilities be clearly stated and then honored if and when necessary. That is, think of the bank as having a hierarchy of long-term debt ranging from the most senior (call it tranche A) to the most subordinated (tranche Z for zombie!). Whenever a bank’s capital position is deficient – say, because the market value of its equity sinks below a threshold ratio to its book assets – the resolution authority automatically makes some of the debt into new equity, starting with the Z tranche and then climbing up the alphabet until there is sufficient capital to return the bank above the regulatory minimum. Provided that there is sufficient long-term debt to absorb the losses, the concern remains a going one. [The resolution authority could still replace management and shut down risky activities in an effort to prevent a serial failure.]

Transparent specification of the subordination structure of a bank’s long-term liabilities – the order in which debt will be “converted” by the resolution authority – means that a living will can be a reasonably short document. [Note that the debt as defined here is not “convertible” in the legal sense. Rather, as in a standard bankruptcy procedure, a portion of it becomes equity as part of the resolution process.]

But let’s not overstate the attractiveness or simplicity of the phoenix plan. No scheme can eliminate policy discretion, as crises often lead governments to change the rules on the fly (think of the 2008 TARP legislation that followed the failure of Lehman).

More important, safeguarding the financial system will still require other complex rules and enforcement. Above all, in addition to more equity, each systemic intermediary must issue an adequate supply of long-term debt in good times to absorb its potential losses in the worst times. Were it instead to issue short-term debt, the intermediary would be vulnerable to a run. In addition, (as one of us has argued here) rules must prevent other leveraged intermediaries from owning this risky long-term debt, because it is in fact a close substitute for equity capital. Otherwise, losses incurred by the holders of this debt would transmit one intermediary’s insolvency to the broader financial system.

And there is no free lunch, at least not for the debt issuers. It remains to be seen whether unleveraged buyers of phoenix debt would price it in a way that is economically viable for systemic intermediaries. After all, in the worst state of the world, it is designed to be no different from equity. Making it attractive to buyers might require banks to issue more equity, to streamline and simplify their operations, and to become more transparent. But that kind of market discipline would go a long way to making the financial system safer, reducing our reliance on the omniscience of regulators.