Julie Dickson, Superintendent of the Office of the Superintendent of Financial Institutions, has delivered a speech on contingent capital titled Too-big-to-fail and Embedded Contingent Capital. This speech is long overdue; most regulators would have delivered the speech prior to writing an op-ed for foreigners, but OSFI, as we all know, has its own special way of doing things.
In accordance with OSFI’s standards, there are not only no footnotes in the published speech, but there is next to no acknowledgment of the international debate concerning contingent capital and there are some breathtaking examples of intellectual dishonesty.
As a result of the crisis, there is now a widely held presumption that governments will support institutions that are perceived to be too-big-to-fail. In rating bank debt, rating agencies now explicitly acknowledge that some banks are likely to be supported by government because they are deemed systemically important.
“Now” is pitching it a little strong. Moody’s made its assumptions about government support explicit prior to the crisis. I will also note that the Bank of International Settlements (and OSFI itself) allows bank paper to be risk weighted according to the credit rating of the sovereign, which implicitly assumes sovereign support.
She overstates the benefits of contingent capital:
Another advantage of embedded contingent capital is that it avoids any need to create a systemic risk fund, which could lead to concerns about what to do with such a fund over time. Instead, investors with a financial interest would decide what each bank should pay when contingent capital was issued – with riskier banks penalized by the market. Thus, regulators would not have to develop a specific charge on systemically-important institutions, which is extremely difficult to do.
That depends, doesn’t it? If there’s any leverage at all in the bank, then its losses can exceed the capital; it’s only a matter of degree.
Additionally, her statement that the system results “riskier banks penalized by the market” is somewhat – not completely – at variance with her desire to have contingent capital priced like debt. As has been discussed on PrefBlog, the Fed has found that sub-debt pricing is not well correlated with risk and there is not much theoretical difference between the risk of sub-debt as it is and her vision of sub-debt that is contingent capital.
The conversion trigger would be activated relatively late in the deterioration of a bank’s health, when the supervisor has determined that the bank is no longer viable as currently structured. This should result in the contingent instrument being priced as debt. Being priced as debt is critical, as it makes it far more affordable for banks, and therefore has the benefit of minimizing the impact on the costs of consumer and business loans.
This objective dooms the plan to failure. You cannot get something for nothing. If you encourage your average bozo bond investor to buy sub-debt ‘because it’s just like debt’, he’s going to be awfully surprised and hurt when he finds out that it isn’t. As evidence, I can cite what happened when Deutsche Bank refused to honour its sub-debt pretend-maturity. Conversion – or the prospect of conversion – will in such a case exacerbate the panic.
An identifiable conversion trigger event could be when the regulator is ready to seize control of the institution because problems are so deep that no private buyer would be willing to acquire shares in the bank, or when a government injects capital into (or otherwise provides guarantees to) a bank. Upon occurrence of a trigger event, each contingent security would convert into common equity.
This one-trigger-fits-all approach will make it virtually impossible for a bank to issue contingent capital when it’s starting to get into trouble.
A range of conversion methods is being analyzed. For example, each contingent security could convert into any number of common shares determined by dividing the par value of the contingent security by the average mid-day market value of common shares during the last several (to be defined) trading days.
Conversion at market price, no matter what that market price is, will lead to death-spirals. Later on, she pretends concern regarding ‘trading against the trigger’. Trading against the trigger is what death spirals are all about.
An additional concern – to me – is that a straight market-value conversion means that converted contingent capital holders will have taken no loss at all due to the deterioration in the bank’s health. There are a number of contradictory elements to Ms. Dickson’s plan:
- She wants it to be priced like debt
- but absorb losses prior to government intervention
- and provide market penalties for riskier banks at time of issue
- but not make the buyers take any losses at all on conversion
All methods seem to generally convert par-to-market value of shares, and need to achieve the outcome that the more senior the capital security, the more consideration provided.
This is a complete fabrication. The Newcastle Building Society issue, for instance, essentially converts at book value. The Rabobank issue works like straight insurance, with no actual conversion. The UK FSA proposes issues where the conversion price is pre-set. The Lloyds issue converts at the market price at time of issue.
There is no consensus on conversion prices,
To her credit, she does address the question of earlier triggers, although she insists on framing the discussion with a trigger based on reported capital levels:
Question #1: Why not require conversion earlier – for example conversion when a bank is still healthy but trips a tier 1 target yet to be defined?
Answer: An earlier trigger does have some appeal. It would create an incentive for management to issue equity long before a forced conversion takes place, and thus deals with any reluctance management might have to take early action. If conversions are early they might also be more common place and thus help demonstrate that market discipline is real – after all, actions speak louder than words.
The problem is that early and frequent conversions would mean that contingent capital would be priced more like equity, which greatly increases its cost. The higher the cost of equity, the higher the resulting cost of credit to consumers and business.
A trigger for conversion well before non-viability (at relatively high levels of capitalization for example) could be destabilizing. As well, it is much more likely to be associated with forbearance, and creative interpretation of the data, than a trigger at non-viability (a specific regulatory capital or financial target trigger would be subject to potential manipulation or arbitrage).
It is also impossible to know in advance when conversion is desirable and equitable based on a pre-set capital or financial target trigger. A trigger at non-viability would mean that a solution is necessary and pressing. This also means that procrastination is less likely.
The likelihood is that banks, rating agencies, and investors have incentives to seek triggers which are far too late in the process to provide capital to a failing bank and achieve the intended benefits. Triggers that are set far beyond when a supervisor would actually act to close an institution would minimize the cost of contingent capital by greatly reducing the chances of conversion, and thereby simplify the ratings and sale process. But, to the extent that conversion will not occur when required, it is unlikely to offer the intended benefits (reduction of moral hazard, providing an expedited resolution mechanism, ensuring that all capital bears losses when governments invest capital).
The objection she cites in the second paragraph isn’t a bug, it’s a feature. It should be apparent that you can’t get something for nothing; that contingent capital should be priced differently from senior debt; and that proximity to the trigger will increase the accuracy of the pricing of the risk – i.e., it’s very hard to calculate whether RBC will go bust in ten years time. It’s much easier to calculate whether it will lose money next year.
I must say, it’s rather odd for OSFI to start obsessing about the banks’ cost of capital after two solid years of bragging about how much capital they required them to hold!
As for her third paragraph – well, the potential for such manipulation by management is a major reason why I prefer market based triggers.
Her last paragraph is somewhat incoherent with respect to ‘rating agency motivations’. DBRS has published its classification of triggers; Moody’s has announced it will not rate contingent capital with a trigger based on regulatory discretion; quite reasonably, they imply that all else being equal, greater certainty will imply a higher rating on the instruments. S&P’s comments imply that earlier triggers will enhance the rating of senior instruments.
Update: Copy-paste journalism from the Globe & Mail.