The question should move from “how does the number of covered bonds outstanding influence the return for other creditors?” to “how does the number of covered bonds outstanding influence the probability that bail-in will be an ineffective resolution tool?”
A few years ago it was very fashionable at covered bond conference to have arguments about asset encumbrance and the damage that excess covered bonds could do to other creditors. It became fashionable for two reasons, a very real probability of limits on covered bonds being imposed in the European Union and the introduction of new jurisdictions that had historically other forms of encumbrance on balance sheet and therefore a greater focus on the problem. The new jurisdictions were led in this by Australia which has always had concerns about the subordination of retail depositors. Canadian and Singaporean regulators went along with it.
Note that the debate was not instigated by rating agencies downgrading unsecured bonds or by investors in them demanding more of a risk premium. It was instigated by people worrying on their behalf.
Fortunately the debate has largely gone away now – although its residue in terms of fixed limits on issuance in some jurisdictions stubbornly persists. Whether it went away because everything that could be said was or because those of us who downplayed the risk won the argument is unclear. To recap, the best argument was a mathematical one – the impact on the loss given default for an unsecured creditor is relatively small for even a very large number of covered bonds. From a rating perspective the impact on default probability of having access to such a robust funding source is much more significant, and positive.
Are we going to see a return of this debate in a new form?
The old debates were predicated on a bank failure and a division of assets amongst creditors. Unsecured creditors always got less and covered bond creditors got full recovery. But this scenario is, post bank recovery and resolution directive, highly unlikely to ever actually happen. We need to recalibrate accordingly.
Because covered bonds are exempt from bail-in under the directive, the question should move from “how does the number of covered bonds outstanding influence the return for other creditors?” to “how does the number of covered bonds outstanding influence the probability that bail-in will be an ineffective resolution tool?”.
Requirements for minimum amounts of bail-in-able debt are supposed to eliminate this risk. Similarly supervisory rights to early intervention – asset quality reviews and decisions on point of non-viability, for example – also reduce it. But neither eliminates the risk, they just make it more remote.
How then to recalibrate the ‘how many is too many?’ for the resolution directive age? The rating agencies are starting to provide quantifiable differences between default and resolution probabilities for individual banks but so far these seem insufficiently sophisticated to form the basis for a meaningful model of the cost of less bail-in.
I fear this is a discussion we may have to return to.
Event, Articles and Videos that might interest you
No related events !