Large banks are not “the next SVB”

A lot of people are digesting the news of the UBS acquisition of Credit Suisse (CS). For me, it’s a sad outcome because I know a lot of great CS people that have been working tirelessly.

Lots of people are wondering whether another large bank will be next. CS faced a loss of confidence and deposit outflows, similar to Silicon Valley Bank (SVB). However, the issues at CS that led to the loss of confidence were specific to the bank, and then became exacerbated by broader concerns about the industry. Other large banks haven’t faced the same challenges as CS in recent years. Therefore, I think it’s unlikely that, by the end of 2023, we will see a large bank in trouble caused by reasons similar to CS.

However, it is possible that we will see a bank in a similar situation to SVB, caused by similar risks. That is where I have focused my attention and this post. That said, I believe it will not be a large bank.

I am in the fortunate position to have worked in bank treasury, so I understand what went wrong in SVB relatively well and I understand liquidity disclosures well. Therefore, when this LinkedIn post started being circulated last week, it was relatively easy for me to look into:

I haven’t had time to research each of the banks, so I took the only large bank near the top of the list, Bank of America (BoA). I thought BoA presented the largest systemic risk, so I thought I should research it first. My conclusion is that BoA doesn’t have the same risk as Silicon Valley Bank (SVB) and I explain why below.

I should highlight that BoA is a potential client of Finteum , so I am somewhat biased in their favour. I have not had any confidential discussions with anyone in BoA, so I don’t have any insider knowledge.

Parallels between Bank of America and SVB

Most people agree that the biggest problem at SVB was the unhedged interest rate risk in the held to maturity (HTM) securities book, which created a large unrealised loss. The LinkedIn post above looked at other banks with large HTM securities portfolios and large unrealised losses (relative to the capital base). It would be easy to conclude that those banks might get into the same trouble as SVB. However, like many things in life, the situation is not that simple.

It’s true that BoA had a large $109bn unrealised loss on its books from its HTM securities as of 31 Dec. To an outsider, it might look like BoA is positioned similarly to SVB, having bought large amounts of fixed rate securities with long duration. However, I believe this is not the case and will not create the same disastrous consequences as for SVB, for 4 main reasons.

Contrasts between Bank of America and SVB

1.??????Bank of America has access to a wider range of contingent funding sources

SVB should have had a better contingency funding plan in place. BoA has many sources of contingent funding. As a large bank, BoA regularly accesses the repo markets for funding (both bilateral and triparty repo). BoA is obliged to regularly test its access to Federal Reserve facilities. BoA can also access the Fed’s new Bank Term Funding Program (BTFP). BoA has many contingent funding sources that SVB didn’t have, which means BoA is less likely to be forced into selling securities and realising a loss, like SVB was.

2.??????Bank of America is less likely to realise a loss in a liquidity stress scenario

BoA is much less likely to sell its HTM securities in a liquidity stress scenario. This is important because, in contrast with SVB, in a stress scenario BoA wouldn't need to realise the (currently unrealised) losses on the HTM securities, wouldn't be forced to raise additional capital promptly, wouldn't face negative publicity and wouldn't incur additional losses from fire-selling assets.

Unlike SVB, BoA is far more likely to survive a severe stress scenario without needing to monetise its mortgage-backed securities (MBS). BoA’s LCR disclosure shows that in a severe stress scenario it could face outflows of $505bn of liquidity over a 30-day period. To fund most of that, BoA could use its $174bn of cash and monetise $252bn of “Level 1” securities, which are mostly US Treasuries. That only leaves $12bn of outflows to cover, which means BoA could survive this severe event by repoing only $12bn (3%) of its $427bn Agency MBS, or by accessing $12bn (3%) of its $348bn capacity at the FHLB and the Fed (see 10-K pg. 55).

3.??????Bank of America has a more diverse asset and liability base?

In SVB, the HTM portfolio represented 43% of the total assets of the bank and the AFS portfolio was another 12%, so 55% in total. Also, SVB’s deposit liabilities were highly concentrated – a relatively small number of organisations deposited funds there.

BoA’s assets are $3.05tn, so the AFS + HTM portfolios are 25%. It’s not small, and the HTM loss, at 48% of the capital base, looks large. But it’s much less significant than at SVB. In addition, BoA’s liability base is much less concentrated than SVB. If there was a withdrawal of funding from BoA, it wouldn’t be as rapid. This means BoA would have more time to react and monetise its securities in different ways.

4.??????Bank of America's deposits are stickier

For an interest rate risk manager, it's normal to buy long-dated fixed-rate securities and hold them in HTM, as an interest rate hedge for liabilities. Banks create behavioural assumptions about how long they will keep deposits for and whether the interest paid on those deposits will change when interest rates change. BCBS368 created some standards and caps for this, but as far as I understand it hasn't been implemented in the US, so BoA and SVB would have created their own assumptions.

BoA's processes for creating assumptions and monitoring the risks would be more robust than SVB's processes were, and more heavily scrutinised by regulators. Using their assumptions, BoA and SVB might have both judged that they were fully hedged for interest rate risk, when they bought HTM bonds with long duration as a hedge for deposits that would be sticky and not sensitive to interest rates. In BoA's case, this judgement is far more likely to be accurate, since BoA's customer deposits have more FDIC insurance and are less concentrated, so would be less sensitive if BoA didn't change deposit rates along with interest rates. Since their assumptions are probably better, BoA is less likely than SVB to be badly hedged for interest rate risk and is not "the next SVB".

Conclusion

In conclusion, I don’t think any bank is “the next CS” because the challenges there were specific to CS. I think it’s far less likely that BoA will need to liquidate the HTM book, so it’s unlikely BoA will be “the next SVB”. Since all large banks are held to similarly high standards, I believe that if I looked into other large banks on the list, I would draw a similar conclusion.

I wrote a more detailed analysis for Finteum 's clients of what went wrong at SVB from a treasury perspective - please send me a direct message if you would like to receive it.

* note, having spoken with more treasury professionals on this, I edited the article to add point 4. "Bank of America's deposits are stickier" - the conclusion is unchanged. I haven't worked on interest rate risk in the banking book (IRRBB), so thank you to those who helped.