APRA recently announced that it had concluded that AT1 instruments had not been shown to act effectively in a going concern scenario and were worse than Tier 2 capital in a resolution scenario leaving it with three choices:
- Maintain the status quo while monitoring to see whether a solution emerged from international developments (either Basel redesigning the instruments or some kind of market solution)
- Unilaterally redesign AT1 to make it more effective in Australia via some combination of a higher trigger for conversion, a new discretionary trigger and/or restrictions on the investor base
- Simplify the capital adequacy framework by replacing AT1 with Common Equity Tier 1 (CET1) and/or Tier 2
APRA summarised their conclusion in favour of the “Simplify” option as follows:
“Replacing AT1 with more reliable forms of capital will enable banks to more quickly and confidently use their capital buffers in a crisis scenario and is expected to reduce compliance costs for banks. It will also strengthen the proportionality of the prudential framework by embedding a simpler approach to capital requirements for small and mid-size banks compared to the new requirements for large banks,” Mr Lonsdale said.
Under APRA’s proposed approach:
- Large, internationally active banks would be able to replace 1.5 per cent AT1 with 1.25 per cent Tier 2 and 0.25 per cent Common Equity Tier 1 (CET1) capital.1
- Smaller banks would be able to fully replace AT1 with Tier 2, with a corresponding reduction in Tier 1 requirements.
One thing we can all agree on is that phasing out AT1 does make the framework simpler, I have questions though:
- How does removing AT1 makes it easier to use capital buffers in a crisis?
- Does the option to convert to CET1 have no value in a going concern scenario?
- Just how risky are these AT1 instruments from an investor perspective?
- Will the debt market happily accept the smaller banks switching from AT1 to Tier 2 (i.e. if the debt market and rating agencies do see some value in having AT1 then will the smaller banks be required to hold CET1 instead)?
Does removing AT1 make it easier to use capital buffers in a crisis?
APRA questions whether AT1 discretionary distributions will in fact be cancelled when a bank is under stress and losses are eating into the capital buffers because “… the market signal effects from canceling distributions are considered more detrimental than the minor benefit of additional financial support“,
I don’t disagree that adverse signalling is a concern with any capital management action but I am not sure that this particular example is as big an issue as it is presented to be. This is partly because the dollar value of the AT1 distributions is so small that they don’t really help very much. By all means suspend them at the time that ordinary dividends have been suspended completely but otherwise respect the seniority of these instruments in the capital stack and loss hierarchy.
This brings us to adverse signalling concern that seems to be a significant factor in APRA’s analysis. Clearly you can cut ordinary dividends without the adverse signalling constraint so why is this a problem for AT1 distributions?
- I suspect the problem is that the seniority of AT1 to CET1 creates the reasonable expectation on the part of AT1 investors that ordinary dividends should be required to do the initial heavy lifting on loss absorption and capital conservation and that AT1 distributions should come into play only when the ordinary dividend is cut to zero
- APRA itself also acknowledges that the actual financial support created by cancelling the AT1 distributions is “minor” which begs the question why should the A1 distributions be cancelled when the loss absorption and capital conservation requirements can be easily achieved by increasing the reduction in ordinary dividends.
- Once you get to the point where ordinary dividends are completely cancelled, there is room for AT1 distributions to be cancelled as well without the adverse signalling concern, especially if the bank is reporting losses.
Sometimes this is not the way it has played out. Part of the reason that this logical, predictable hierarchy of loss absorption can be disrupted lies I think with the overly complex way that the Capital Conservation Ratio (CCR) is applied. An explanation of why the CCR is overly complex is a whole topic in itself that I will leave for another day. The people who best understand this point are those working in stress testing who have had to calculate the CCR in a stress scenario (you really have to get into the detail to understand it).
The CCR issue could be addressed by simplifying the way it is calculated and I imagine there would be broad agreement that simplicity is always a desirable feature of any calculation that has to be employed under conditions of stress and uncertainty. The main point I want to make is that “simplifying” the application of the way the CCR is used to conserve capital might be a useful place to start before you decide to get rid of AT1 completely.
The bigger question I think is whether the option to convert AT1 into common equity has no value, either in a going concern scenario or in resolution.
APRA argues that
“Rather than acting to stabilise a bank as a going concern in stress, international experience has shown that AT1 absorbs losses only at a very late stage of a bank failure. … If AT1 was used in Australia APRA considers it would not fulfil its role in stabilising the bank before non-viability was reached.”
In one sense, late conversion is a feature as opposed to a bug in the design of AT1 instruments. They are designed to convert as a last resort when the 5.125% CET1 Point of Non-Viability is crossed so it should not be a surprise that they convert relatively late in a stressed scenario. Perhaps more importantly the conversion option is calibrated to a point where a bank has run through the standard set of options for conserving and rebuilding capital buffers.
APRA acknowledges that the trigger could be set higher so that conversion happens earlier but asserts that “... this could undermine recovery plan actions and the usability of capital buffers“. So far as I could determine, the only reason cited to support this conclusion is that “A higher trigger, if exercised, may operate as a negative signal to the market at the very time when regulators are aiming to restore confidence“.
The statement regarding market signals is interesting. The conversion trigger, wherever it is set, is something that the “market” is well aware of, so the potential for adverse signalling seems to imply that the situation in the bank is much worse than was commonly understood by the market. If that was not the case then conversion is simply an outcome that the market should have been anticipating. A related concern for APRA is that international experience has shown that banks have failed with CET1 ratios much higher than 5.125 per cent, further reducing the potential value of the PONV trigger in their eyes.
It is obviously true that banking has an unfortunate history of seemingly solvent banks very quickly being found out not to be solvent (or maybe just perceived to be insolvent – much the same thing in practice), let’s call this Scenario 1. This still leaves a lot of other scenarios in which a bank has taken a large hit and just needs an infusion of new common equity to rebuild and survive.
Remember that APRA is proposing to replace 1.5% of AT1 with .25% of CET1 (and 1.25% of T2) for the big banks, while the smaller banks simply swap AT1 for Tier 2. For the bigger banks, the extra 0.25% CET1 arguably makes no real difference to the outcome for a Type 1 scenario, but a bigger block of convertible capital could be useful in the other adverse scenarios. For the smaller banks, they just have less options. So the capital framework proposed by APRA is simpler for sure but it is not obvious that it is better.
A related question is just how risky are these instruments from the investor perspective
The theory is that AT1 should be less risky than common equity because they pay investors a predetermined income distribution and have a more senior position in the loss hierarchy. Clearly we are all comfortable with retail investors holding common equity so in theory it should be ok for them to invest in something less risky.
This nice theoretical framework can get a bit messy in the real world for a couple of reasons:
- Firstly, the CCR as currently implemented can result in AT1 distributions having to be cut in conjunction with the ordinary dividend thereby disrupting the simple hierarchy in which AT1 only comes into the frame once cutting the ordinary dividend has totally exhausted its capacity to satisfy the CCR
- Secondly, there have been instances in which loss absorption has been achieved by the write down of the AT1 instrument rather than by conversion to common equity (thereby absorbing loss in a way that benefits common equity at the expense of AT1 investors and violating the loss hierarchy benefits that the AT1 investor though they were paying a premium for.
- Even if conversion does play out as expected, there remains a risk that the value of the ordinary shares received on conversion continues to fall before the AT1 investors can convert them into cash. Indeed, the simple fact that a block of shares is being issued to investors who are not long term holders can itself contribute to a decline in the share price.
Most of these problems (not all) seem to be fixable:
- Simplifying the way in which the CCR must be met would help reduce the potential for AT1 distributions to be required to absorb loss sooner than would be expected under the principle that AT1 ranks senior to ordinary equity
- Similarly, loss absorption at the PONV can also be implemented in ways that respects the hierarchy of loss that AT1 investors believed they had contracted to. I am open to someone making the case that the option to write off AT1 has some value but mostly it just seems to introduce complexity and raise doubts about the seniority of the AT1 instruments.
The potential for AT investors to receive less than par value via the conversion is I think a risk that can’t be eliminated but I am not sure that it need to be eliminated. In the event that the shares received on conversion are sold for less than the conversion price, AT1 investors are still likely to have realised lower losses than those experienced by ordinary shares under the same scenario. Investing at this end of the capital stack is about relative risk and reward, the real question is whether the risk is correctly valued.
It is possible that APRA still concludes that retail investors are unable to accurately assess and price the risks. In that case, you retain the option of restricting the sale of the instruments to sophisticated investors and institutional money. That seems to me like a perfectly reasonable compromise that retains the useful features of these AT1 instruments. I just don’t see that the case has been made to completely eliminate the options that AT1 offers.
Last, but not least, we come to the question of what the debt market thinks about the proposal for smaller banks to completely replace AT1 with T2.and perhaps more importantly what do the rating agencies think?
I have to confess that I don’t know. It does seem to me however that something is being lost. The consultation on the proposal is currently open and more informed minds than my own will I hope be addressing that question in the feedback they offer.
I am speculating here but the only way that I can rationalise the policy option proposed by APRA is that they have concluded that the CET1 levels they have established in response to Unquestionably Strong target set by the 2014 Financial System Inquiry have rendered other forms of “going concern” capital redundant. The rating agencies will clearly have an opinion on this question.
I have been a long way from the front line on these issue for some time but that is my 2 cents worth – as always please let me know what I am missing.
Tony – From the Outside
From the outside 
