PrefBlog

Mark Flannery has published a paper with the captioned title dated 2009-10-6, which elaborates and defends his original idea for Contingent Capital that has been previously discussed on PrefBlog (he used to call them Reverse Convertible Debentures):

The financial crisis has clearly indicated that government regulators are reluctant to permit a large financial institution to fail. In order to minimize the transfer of future losses to taxpayers or to solvent banks, we need a system for assuring that large institutions always maintain sufficient capital. For a variety of reasons, supervisors find it difficult to require institutions to sell new shares after they have suffered losses. This paper describes and evaluates a new security, which converts from debt to equity automatically when the issuer’s equity ratio falls too low. “Contingent capital certificates” can greatly reduce the probability that a large financial firm will suffer losses in excess of its common equity, and will provide market discipline by forcing shareholders to internalize more of their assets’ poor outcomes.

Specifically:

• a. A large financial firm must maintain enough common equity that its default is very unlikely. This common equity can satisfy either of two requirements:
  • o Common equity with a market value exceeding 6% of some asset or risk aggregate. For simplicity, I’ll discuss the aggregate as the book value of on-book assets.
  • o Common equity with a market value exceeding 4% of total assets, provided it also has outstanding subordinated (CCC) debt that converts into shares if the firm’s equity market value falls below 4% of total assets. The subordinated debt must be at least 4% of total assets.
• b. The CCC will convert on the day after the issuer’s common shares’ market value falls below 4% of total assets.
• c. Enough CCC will convert to return the issuers’ common equity market value to 5% of its on-book total assets.
• d. The face value of converted debt will purchase a number of common shares implied by the market price of common equity on the day of the conversion.
• e. Converted CCC must be replaced in the capital structure promptly.

There are two problems with this proposal. First, there is a continued dependence upon official balance sheets which, however reflective they are of the truth, are subject to possible manipulation and will not be trusted in times of crisis. Second, the conversion of CCC into equity at contemporaneous market values has a destabilizing effect upon capital markets, brings with it the (admittedly slight) potential for death-spirals and (most importantly for some, anyway) leaves CCC holders immune to the potential for gaps in the market.

Dr. Flannery cedes the first point regarding balance sheets:

Market values are forward-looking and quickly reflect changes in a firm’s condition, including off-book items, which GAAP equity measures might omit. In contrast, GAAP accounting emphasizes historical costs and provides managers with many options about when and how to recognize value changes. These options are manipulated most aggressively when the firm has problems – exactly when rapid re-capitalization is required to ameliorate those problems. A trigger based on GAAP equity value thus guarantees that the trigger will be tripped long after a financial firm enters distress, and perhaps long after it has become insolvent. (Recall how many troubled banks and holding companies during 2008 were “well capitalized” or “adequately capitalized” according to Basel’s book-valued calculations.) A trigger based on GAAP equity value thus guarantees that the trigger will be tripped long after a financial firm enters distress, and perhaps long after it has become insolvent. (Note how many troubled banks and holding companies during 2008 were “well capitalized” or “adequately capitalized” according to Basel’s book-valued calculations.)

I suggest that if the trigger is set according to a pre-determined equity price, then the potential for jiggery-pokery is reduced substantially as, for instance, bank management will have no incentive to manipulate the balance sheet, or to benefit from prior efforts at manipulation. It will be recalled that Citigroup’s problems first made the news due to its off balance sheet SIVs. It will also be recalled that banks have substantial nod-and-wink exposure to defaults experienced in their Money Market Funds that are not recognized until well after the fact.

Additionally, basing the trigger solely on the market price of the common has the great advantage of being separated from accounting and regulatory considerations – the redundancy is important! I suggest that such redundancy with respect to the leverage ratio vs. the BIS ratios is, essentially, what saved the North American banking system.

It is not clear why Dr. Flannery, having thrown out book value for the equity (numerator) part of the trigger, continues to believe in its adequacy for use in the assets (denominator) component.

With respect to the conversion price, Dr. Flannery states:

My proposal in Section 3 would convert CCC face value into shares at a rate implied by the contemporaneous share price. With a contemporaneous-market conversion price, CCC bonds have very low default risk. With a sufficiently high trigger value, the CCC investors will almost surely be fully repaid either in cash or in an equivalent value of shares. Relatively safe CCC bonds whose payoffs are divorced from share price fluctuations should trade at low coupon rates in liquid markets.

With all respect, I consider this to be a bug, not a feature. CCC bonds should have a higher default risk than other bonds – defining default to be a recovery of less than expected value – otherwise there is little incentive for buyers of such bonds to enforce market discipline. I suggest that CCC bondholders should be exposed to gap risk – if the equity trades on day 0 fractionally above the trigger price (however defined) and management makes announcements that evening that cause the stock to gap downwards overnight, I suggest that it is entirely appropriate for unhedged CCC bondholders to take a loss. Exposing equity but not CCC to gap risk will make it harder to recapitalize the bank through new equity issuance.

When discussing the Squam Lake commentary on this issue, Dr. Flannery asserts:

CCC bonds with a market-valued trigger and a fixed conversion price could effectively recapitalize over-leveraged firms. However, the fixed conversion price adds an element of equity risk and uncertainty to the CCC returns. A high conversion price might give shareholders an incentive to induce conversion as a means of selling equity cheaply. A low conversion price would make bondholders eager to bid down share prices (if possible) to trigger conversion. Such strategic considerations are unrelated to the firm’s credit condition and add nothing to the regulatory goal of stabilizing under-capitalized financial firms. Occam’s razor thus supports the separation of equity risk from credit risk in CCC, further abetting transparent solutions for troubled banks.

I do not find this argument convincing. If, as I have suggested, the trigger price is also the conversion price, then the statements A high conversion price might give shareholders an incentive to induce conversion as a means of selling equity cheaply. A low conversion price would make bondholders eager to bid down share prices (if possible) to trigger conversion. becomes not applicable.

Her Majesty’s Treasury has released a discussion document on the role of banks titled Risk, reward and responsibility: the financial sector and society, which discusses contingent capital among other things:

In the recent crisis existing subordinated debt and hybrid capital largely failed in its original objective of bearing losses. Going-concern capital instruments often failed to bear losses because banks felt unable to cancel coupon payments or not call at call-dates (even though it was more expensive to refinance), in part for fear of a negative investor reaction as well as due to the legal complexity of the instruments. Gone-concern capital such as Lower Tier 2 has often failed to bear losses in systemic banks as governments have been forced to step in to prevent insolvency in part to prevent further systemic impacts on debt-holders such as insurance companies. CRD 2, the first in a series of forthcoming packages amending the Capital Requirements Directive, sets out criteria for the eligibility of hybrid capital instruments as original own funds of credit institutions. It also provides a limit structure for the inclusion of hybrid capital instruments in own funds.

Box 3.D reviews academic proposals for Contingent Capital:

Debt-for-Equity Swap – Raviv (2004) [1]
The proposal is for debt that pays its holder a fixed income unless the value of the bank’s capital ratio falls below a predetermined threshold (based on a regulatory measurement). In this event, the debt is automatically converted to the bank’s common equity according to a predetermined conversion ratio (the principal amount may change upon conversion).

Contingent capital certificates – Flannery (2009) [2]

Similar to the above, contingent capital certificates are debt that pays a fixed payment to its holders but converts into common stock when triggered by some measure of crisis. In contrast to the above this would be a market-based measure, with conversion occurring if the issuer’s equity price fell below some pre-specified value. The converted debt would buy shares at the market price of common equity on the day of the conversion rather than at a predetermined price.

Capital Insurance – Kashyap, Rajan and Stein (2008) [3]

Under this proposal, the insurer would receive a premium for agreeing to provide an amount of capital to the bank in case of systemic crisis. The insurer would be required to hold the full insured amount, to be released back to the insurer once the policy matures. The policy would pay out upon the occurrence of a ‘banking systemic event’, for which the trigger would be some measure of aggregate write-offs of major financial institutions over a year-long period. Long-term policies would be hard to price and therefore a number of overlapping short-term policies maturing at different dates are proposed.

Tradable Insurance Credits – Caballero, Kurlat (2009) [4]

The central bank would issue tradable insurance credits, which would allow holders to attach a central bank guarantee to assets on their balance sheet during a systemic crisis. A threshold level or trigger for systemic panic would be determined by the central bank. An attached tradable insurance credit is simply a central bank backed Credit Default Swap (CDS).

1 Bank Stability and Market Discipline: Debt-for-Equity Swap versus Subordinated Notes. Raviv, Alon. 2004.

2 Contingent Tools Can Fill Capital Gaps, Mark Flannery, American Banker; 2009, Vol. 174 Issue 117.

3 Rethinking Capital Regulation Kashyap, Rajan, Stein, paper prepared for Federal Reserve Bank of Kansas City symposium on “Maintaining Stability in a Changing Financial System”, Jackson Hole, Wyoming, August, 2008

4 The “Surprising” Origin and Nature of Financial Crises: A Macroeconomic Policy Proposal, Ricardo J. Caballero and Pablo Kurlat, August 2009

As has been discussed on PrefBlog (as recently as last week), Flannery’s proposal makes most sense to me. The Capital Insurance proposal has been used in Canada, with the RBC CLOCS, but I am not convinced that such elements are reliable in terms of a crisis – to a large degree, this will simply shift the uncertainty and fear of a crisis onto the insurance providers.

Update, 2010-6-13: The Kashyap paper is available on-line.