AT1s: Deutsche Bank loses by split decision

In Additional Tier 1s (AT1s), investors spend a lot of time thinking about calls, since whether banks decide to call their AT1 bonds or not can have a big impact on price action in the market.

As a reminder, the vast majority of AT1 bonds are called as expected at the first opportunity. However, every so often the asset class faces the prospect of an issuer not calling, and particularly in the case of a high profile issuer, this can stir a fair bit of emotion and debate among AT1 market participants.

Early on Friday, Deutsche Bank (DB) announced a call of its $1.25bn 7.5% AT1 and a simultaneous non-call of its $1.25bn 4.789% AT1. The former will be redeemed on April 30, while the latter will remain outstanding with a reset coupon of around 8.46% and a new call date of April 30, 2030. The prices of both bonds were broadly unchanged post the announcement, though this means that the non-called 4.789% bond for now continues to trade marginally below par, an undesirable outcome from the perspective of bondholders.

There are a few factors worth mentioning here.

First, it is worth reiterating that for a number of reasons, the vast majority of AT1s are called at their first call dates. Call decisions are always driven by economics, but the bank’s assessment incorporates related factors such as bondholders’ expectations, its reputation with the AT1 investor base, and the potential impact of a non-call on the rest of its debt stack. In most cases, a non-call normally happens because the issuer is unable to refinance the existing deal in the market at a cheaper level than the reset coupon, and as a result the prices of the bonds tend to drop when the non-call is announced to reflect this discrepancy (obviously an undesirable outcome for bondholders).

Second, DB’s history of calling capital instruments is somewhat patchy relative to its peers. Back in 2008, for example, to the disappointment of investors, the issuer became the first major European bank not to call its Lower Tier 2 instruments. More recently, in March 2020 the bank chose not to call its $1.25bn AT1 (this is the same 4.789% bond that has not been called also today). Notwithstanding these instances, the issuer has done considerable work in turning around its perception among creditors in recent years which helped offset the negativity associated with strictly economic calls and little weight given to bondholders’ expectations.

Third, and moving to the decision today, the bank elected to call one of its AT1s while leaving the other one outstanding. We note that calling and replacing both deals would have seen DB incur an FX loss of around €400m at today’s FX rates (around €170m for the $7.5% bond and around €230m for the $4.789% bond) thanks to the change in the rates since these securities were issued in USD back in 2014. This is unhelpful for a bank still working to meet a return on tangible equity (RoTE) target of 10% for full-year 2025, when European banks are seeing on average around 13% RoTE (DB itself reported an RoTE of just 4.7% for 2024). A €400m FX loss is material when set against DB’s projected net income of €6.32bn for 2025.

In the context of the above, DB indicated as recently as February that it intended to take a “deal-specific approach” to AT1 calls in 2025, while also emphasising its focus on FX losses in the decision process. This made complete sense to us considering the importance of its RoTE target.

While acknowledging its clearly stated intention, we believe this split decision undoes some of the progress DB has made in bringing creditors back on side, at a time when its returns remain sub-par versus peers in what is a very supportive operating environment. Indeed, the cleanest outcome would have been for both bonds to be called.

But failing that, we think the issuer should have skipped the call on both issues rather than picking and choosing between individual bonds, since the economics of a non-call were not dramatically different between the two deals (the $7.5% deal would have traded up slightly over par). Skipping the call on the 7.5% AT1 would have come at a cost for DB; before the decision announcement we thought a replacement deal could be priced at around 8.375%, far lower than the reset coupon. However, economically speaking, the effect of not calling the 7.5% bond would have been the same as calling and replacing a hypothetical AT1 that would have reset at a lower coupon than the market level for a new deal (i.e. accepting a higher cost of capital in order to meet bondholders’ expectations). Clearly, the bank was not willing to skip the call on both deals even when this route would have meant avoiding the FX losses; the message to investors is that in future decisions, if FX losses aren’t an issue, there is even less chance of DB putting its reputation with bondholders over the economic considerations.

As things stand, this move suggests to us that DB will continue to trade with a spread premium to its peers in the credit markets, not least due to its lagging returns but also due to its purely economic treatment of AT1 bondholders – in other words, investors will perceive higher extension risk and will likely demand higher reset spreads in order to participate in DB’s new issues, at least for as long as the bank continues to deliver relatively weaker returns. DB is still set to bring €1.5-3bn of capital instruments to the markets this year, and we believe that following this non-call decision, the bank will have to pay up for these bonds.