Category: Regulatory Capital

DBRS has released a commentary titled European Banks’ AT1 Instruments at a Crossroads: Regulatory Reassessment and Market Implications and I have tucked away a copy HERE:

This commentary assesses the ECB’s December 2025 recommendations within the broader policy debate on the effectiveness of Additional Tier 1 (AT1) instruments as going-concern capital. While the ECB’s recommendation provides limited details, its stance points to two potential pathways: a structural redesign of AT1 instruments or complete removal from the going-concern capital stack.
…
AT1 as a Going Concern Capital Instrument: Operational Challenges
AT1 instruments, while designed to be going-concern capital capable of absorbing losses through conversion or write-down, have shown certain functional and operational limitations. The 2023 Credit Suisse AG case—occurring outside the EU—represents the only major instance where AT1 instruments absorbed losses ahead of CET1 and outside a formal resolution process. This case created uncertainties, triggered legal challenges, and underscored doubts about the instrument’s effectiveness as a going-concern buffer. By contrast, past cases such as Banco Popular Español S.A. in 2017 show that AT1 triggers typically activate only at the point of non-viability, effectively functioning as gone-concern tools rather than early-stage stabilisers.

Structural features, including quantitative contractual triggers, further constrain the usability of AT1 instruments in going-concern situations. Most EU instruments feature a 5.125% CET1 trigger (with higher thresholds in the UK and Switzerland and none in Canada). Yet experience shows that authorities tend to intervene before quantitative triggers are breached, particularly when crises are liquidity-confidence driven despite capital ratios remaining above thresholds.

Market behaviour has also raised issues related to AT1s: coupons, though discretionary and noncumulative, are rarely cancelled; issuers routinely call instruments at the first call date (typically after five to seven years), subject to regulatory approval. This is because the coupon reset mechanism often makes the post-reset cost unattractive, and failing to call could be perceived by investors as a sign of weakness. If the issuer does not call, the bond rolls into a reset period with a new coupon based on a five-year swap rate plus the original spread (typically in the range of 350 basis points (bps) to 450 bps).

Potential Policy Scenarios
In our view, regulators will likely review some of the following options: (1) raise quantitative triggers (potentially credit negative for AT1s); (2) remove contractual triggers and treat AT1 more like equity (as in Canada); (3) tighten call and coupon conditions to reduce adverse signalling and improve going-concern credibility; or, in the most radical scenario (4) eliminate AT1 instruments, replacing them with CET1 and/or Tier 2 capital. A CET1-only replacement would increase capital costs and effectively raise capital requirements, while removal without a substitute would diminish systemwide loss-absorbing capacity, underscoring the policy trade-off between simplicity, cost, and resilience.

Australia Backs Out
In December 2024, the Australian Prudential Authority (APRA) announced its decision to eliminate AT1 instruments from the prudential capital stack, reflecting the regulator’s assessment that AT1 has not operated as a dependable going-concern buffer and instead introduces contagion and legal-execution risk. Moreover, the Australian AT1 market differs from other jurisdictions because a significant shareof banks’ AT1 instruments is held by retail investors—primarily high-net-worth individuals—adding further complexity and making loss absorption potentially more challenging than if these instruments were predominantly held by institutional investors.

Under the revised framework, which is effective from 1 January 2027, Australian banks will replace AT1 instruments with CET1 and Tier 2 capital. All outstanding AT1 instruments will be fully phased out by 2032 by gradually removing regulatory recognition and making no changes to the existing legal terms, including subordination, of these outstanding instruments. APRA also recalibrated the minimum leverage ratio to 3.25% of CET1, from 3.50%, preventing unintended tightening linked to the withdrawal of AT1 capacity.

Those who have been reading my commentary on this stuff for the past fifteen years will know that of the four alternative policy responses outlined:

• (1) raise quantitative triggers (potentially credit negative for AT1s);
• (2) remove contractual triggers and treat AT1 more like equity (as in Canada);
• (3) tighten call and coupon conditions to reduce adverse signalling and improve going-concern credibility; or, in the most radical scenario
• (4) eliminate AT1 instruments, replacing them with CET1 and/or Tier 2 capital.

I prefer #1: going to a higher trigger for loss absorption. Canada’s low trigger regime is just plain stupid, having very little effect on market discipline as OSFI has vapours at the idea of banks having to confess that they haven’t been 100% up to scratch.

Update, 2026-03-12: The ECB’s recommendations were published in a press release dated 2025-12-11 and titled Simplification of the European prudential regulatory, supervisory and reporting framework:

Recommendation #2 suggests that the going-concern loss-absorbing capacity of the capital stack could be improved by adjusting the design or the role of AT1 (and Tier 2) instruments. Two alternatives can be considered. First, the features of AT1 instruments could be enhanced to further ensure their loss-absorption capacity in going concern and provide additional clarity to banks and investors on the going-concern loss-absorption properties of AT1 instruments. This option would be Basel-compliant. It would not modify the role of AT1 (and Tier 2) instruments, and would therefore not reduce the overlap with gone-concern requirements. Alternatively, non-CET1 instruments could be completely removed from the going-concern capital stack. This could be achieved either by fully or partially replacing them with CET1 instruments or by eliminating them without any replacement in the going-concern framework. This alternative would (i) decrease the complexity in the going-concern framework, as only one type of instrument would have to be considered in going concern; and (ii) reduce the interplay between the different requirements. However, unless non-CET1 instruments were fully replaced with CET1 (implying a tightening of CET1 requirements), this alternative would raise difficult questions with regard to maintaining resilience and Basel compliance, potentially conflicting with the principles guiding the formulation of the high-level recommendations. In addition, irrespective of the calibration, it would lead to changes in the regulatory CET1 demand, raising questions of capital neutrality.

This recommendation aims to strengthen the quality of capital required under the EU regulatory framework, align the functioning of the going-concern capital stack with its intended purpose and thereby increase transparency for banks’ creditors.

Readers will remember that I am very interested in the IAIS deliberations regarding the definition of Insurance company Tier 1 Limited Capital (which includes preferred shares); I take the view that rules comparable, if not identical to the bank NVCC rules will be imposed by OSFI at some point in the future.

I do not expect OSFI to act until a global standard is agreed upon.

Those who have followed my arguments to support my position may well be getting impatient, although not as impatient as I am. So, I’ll just pass along the news that the comment period for the IAIS Public Consultation: Risk-based Global Insurance Capital Standard (ICS) Version 2.0 has expired:

The purpose of this consultation document (CD) is to solicit feedback from stakeholders on the ICS ahead of the completion of ICS Version 2.0, scheduled for late-2019, before the monitoring period begins on 1 January 2020. This CD covers both issues related to the ICS Version 2.0 monitoring period and the technical aspects of the design and calibration of ICS Version 2.0.

This CD is the third IAIS consultation in a multi-year process to develop the ICS. The IAIS issued its first and second ICS consultation documents in December 2014 and July 2016, respectively. In addition, the IAIS has conducted three quantitative Field Testing exercises in the development of the ICS – in 2015, 2016 and 2017. Currently, the IAIS is conducting its fourth quantitative ICS Field Testing exercise, with data to be submitted in August 2018.

At the same time as this consultation on ICS Version 2.0, the IAIS is also consulting on the Common Framework for the Supervision of IAIGs (ComFrame). While ICS is part of the ComFrame, it was agreed by the Executive Committee of the IAIS in June 2017 that ICS Version 2.0 would be adopted as a stand-alone document in 2019. As such, there are two separate consultation documents.

The consultation document, downloadable from the above page, contains the critical (for our purposes) question:

173. The IAIS is considering whether to set an additional criterion requiring Tier 1 Limited instruments to have a principal loss absorbency mechanism (PLAM). Such mechanisms would provide a means for financial instruments to absorb losses on a going-concern basis through reductions in the principal amount and cancellation of distributions. Without such mechanisms these instruments might only provide going concern loss absorbency through cancellation of distributions.

Click for Big

I will also note that:

7. Comments must be sent electronically via the IAIS Consultations webpage.1 All comments will be published on the IAIS website unless a specific request is made for comments to remain confidential.

I will be keeping a sharp eye out for publication of comments received, I assure you, and will pass them on.

OSFI has disgraced itself yet again with its response to the International Association of Insurance Supervisors’ 2016 Insurance Capital Standard Consultation.

The question is:

Q70 Section 5.3.1: Should Tier 1 Limited financial instruments be required to have a principal loss absorbency mechanism?

OSFI’s answer, found in the document “Section 5 Capital resources (Public)” that is linked on the above page, is “No”.

The follow-up question is:

Q70.1 Section 5.3.1 If “no” to Q70, should the principal be considered to provide loss absorbency on a going concern basis? Please explain how the instrument demonstrates loss absorbency on a going concern basis.

OSFI answers “Yes”, with the explanation:

Tier 1 Limited and Unlimited instruments provide loss absorbency on a going concern basis through the discretion the issuer has to not pay or cancel coupons on the instrument and the non-cumulative nature of such payments. The principal amount of such claims is only extinguished in resolution (regardless of accounting).

OSFI does not support principal loss absorbency mechanisms whereby instruments can be written down or converted into equity under going concern/early triggers (and that are not at the discretion of the supervisory authority) due to concerns that such triggers can lead to financial instability and adverse signalling regarding the issuer’s financial condition (as observed with CoCos issued by European banks earlier this year, for example). OSFI would only support such mechanisms where they result in a full and permanent write-off of the instrument at the point of non-viability where the IAIG has entered into resolution.

Note: What is an “IAIG”?:

An IAIG is a term under ComFrame for insurance groups or financial conglomerates that exceed thresholds on international activity and size. The IAIS defines an IAIG as a large, internationally active group that includes at least one sizeable insurance entity. There are two criteria for an insurance group to be identified as an IAIG: 1) International Activity — premiums are written in not fewer than three jurisdictions, and percentage of gross premiums written outside the home jurisdiction is not less than 10% of the group’s total gross written premium; and 2) Size —based on a rolling three-year average, total assets of not less than USD 50 Billion, or gross written premiums of not less than USD 10 Billion.

However, it is heartening to observe that the other four IAIS full members who provided public answers (European Insurance and Occupational Pensions Authority (Europe; the developers of the “Solvency 2” regime), BaFin (Germany), Financial Supervisory Service (Korea) and the National Association of Insurance Commissioners (USA)) all answered question 70 with “Yes”.

So I continue to believe that “Deemed Retractions” will eventually apply to Insurers and Insurance Holding Companies; I believe that while OSFI may well continue its ridiculous insistence on “low-trigger” conversions, it will adopt a global standard once the rest of the world agrees on conversion.

I will also note that in Canada, forcible conversion of Tier 1 capital for banks is also low-trigger, but this did not stop OSFI from demanding NVCC compliance for bank preferred share issues, which in turn led to “Deemed Retraction” for bank issues.

Now, is all that clear as mud? Sorry, but I’ve got PrefLetter to get out and don’t have much time for linking to previous material on this issue.

Update, 2017-4-19: As noted above, OSFI’s response included:

such triggers can lead to financial instability and adverse signalling regarding the issuer’s financial condition (as observed with CoCos issued by European banks earlier this year, for example).

For a review of the performance – and reasons behind this performance of European CoCos, see Europe’s CoCos Provide a Lesson on Uncertainty:

Contingent convertible bonds (CoCos) issued by European global systemically important banks (G-SIBs) as part of their total loss-absorbing capacity (TLAC) are meant to enhance financial stability by forcing investors to absorb losses when a bank is under stress. Coupon payments are made at issuers’ discretion while loss absorption can be triggered at regulators’ discretion. This study investigates price effects of four press releases by Deutsche Bank AG in February 2016 related to the bank’s willingness and ability to make its upcoming CoCo coupon payments. Expected cash flow models capture changes in CoCo default risk, while event dates capture uncertainty effects. The price of a European G-SIB peer group portfolio declined a statistically significant 2.0-2.5 percent over two days in response to Deutsche Bank’s first press release. Deutsche Bank’s efforts to allay its own CoCo investors’ concerns appeared to increase concerns among CoCo investors generally. The results show potential negative effects of regulatory discretion.

Click for Big