5 min read 26 Jun 23
We begin by looking at Silicon Valley Banks (SVB). The chart below compares a typical US bank balance sheet with that of SVB.
Before 2019, SVB’s balance sheet would have looked a lot closer to that of a typical bank. During the pandemic, however, SVB grew rapidly as investments in private equity and tech start-ups – its primary customer base – spiked. While system-wide US bank deposits grew by just 35% from 2019-2022, SVB’s deposits tripled1.
As deposits increased, SVB decided to invest in government bonds and mortgage-backed securities rather than cash equivalents at near-zero interest rates. To maximise returns, the bank invested in longer-dated bonds. Putting over half of the bank’s assets in long-dated securities might be permissible for large, well-diversified banks, but for SVB it turned out to be poor risk management because their deposit base was neither well-diversified nor ‘sticky’.
SVB’s depositors were clearly sophisticated: they were mostly private equity investors or the funds and start-ups in which they invested. Hence they were well aware of developments in financial markets and ready to move their money if they felt it was threatened. Another unique feature of SVB’s deposit base was that 90% of it was above the $250,000 FDIC insurance limit in the US – making the depositors much more likely to leave if panic set in.
As interest rates rose rapidly in 2022, the value of SVB’s long-dated bonds fell. A 2018 softening in US banking regulations meant that unrealised losses on these bonds did not impact regulatory capital for “small” and “mid-size” banks, including SVB. But in March 2023, SVB’s management decided to sell some of the bonds, crystalising a loss, likely in order to avoid further losses should rates rise. To cover this loss, they announced a plan to raise capital. This brought the problem in the SVB’s balance sheet to the market’s attention, surprising investors who had viewed SVB as a great success story and had thought its capital position to be sound. Investors fretted about SVB’s capital adequacy, and its share price plunged.
Customers started to withdraw their money, knowing there was little downside in doing so, but potentially catastrophic risk from not doing so. And because these were sophisticated customers, they moved quickly. Panic spread over social media, and more depositors rushed to pull their money out. Before SVB was taken over by the regulators, it had lost $40 billion – 25% of its deposit base – in just five hours. The pace of these moves, accelerated by digital banking and social media, clearly caught many regulators and private investors by surprise. In the end, it was a combination of higher rates, poor risk management decisions and a unique business model that took down SVB.
Credit Suisse, meanwhile, had several features that distinguished it from other European banks. First, given its focus on wealth management, a portion of Credit Suisse’s deposit base, was relatively concentrated, sophisticated and uninsured, though to a lesser degree than SVB’s. Second, it had suffered a long string of high-profile losses in recent years – amounting to billions, raising some questions about the sustainability of its business model. Third, it had lost both its chairman and CEO within two years, leaving a strategic void for a period.
In October 2022, speculation on social media led to a meaningful decline in deposits – effectively, the first run on the bank. Here, the regulation – particularly on liquidity – actually worked in preventing a collapse at the time. Despite a sharp decline in deposits, the bank embarked on a restructuring plan to raise capital and rebuild liquidity. However, the combination of SVB’s collapse and further negative headlines created panic among depositors in March, and effectively a second bank run in six months. This led to Credit Suisse’s ultimate failure and the Swiss authorities decision to step in and force a sale to UBS.
Perhaps the most interesting aspect of Credit Suisse’s failure was how different parts of its capital stack (i.e., its bonds and equity) were treated. Their senior debt was left untouched, and now are liabilities of UBS. But this was not true of the contingent convertible bonds (“CoCo”), which were written down to zero. CoCo are actually hybrid securities – a bit like bonds (regular payments) and a bit like equity (perpetual) - meant to take losses only in the event that a bank’s equity is wiped out. The chart below shows where they sit in the capital structure. Despite sitting above equity, Credit Suisse’s CoCo were completely written down to zero while equity holders received some payment in UBS shares. This was, in our view, a fairly blatant, surprising and unfortunate contradiction of the creditor hierarchy.
Since the fall of SVB, the markets have calmed down considerably, though neither bank bond spreads nor bank stock prices have fully recovered to the levels seen in February. The key question now is whether the failures of SVB and Credit Suisse were symptoms of systemic problems in the sector or were unique cases that won’t be repeated going forward. As is often the case, the answer lies between these two extremes, though as far as large, diversified banks go, we do not believe a systemic banking crisis is likely in the near-term for two main reasons.
First, large banks are fundamentally much stronger today than they have been in decades. Capital and asset-quality metrics have improved massively since the Global Financial Crisis. Profitability has improved too; among the G-SIBs (Global Systemically Important Banks) and other national champions, Credit Suisse’s string of large losses was an outlier. So whatever challenges lie ahead – and there are several – most large US and European banks are in reasonably good shape to meet them.
Second, regulation of interest-rate risk is much stronger for European banks and systemically important US banks than it was for smaller US banks – so this crucial element in SVB’s demise is unlikely to be a material concern for either European banks or the largest US banks. It has become a concern for US regional banks, but should be a manageable one for the larger, more diversified among these.
These factors have kept and should keep contagion risk low. But there are other potential challenges facing banks, and we can’t rule out other failures completely given the importance of confidence in banking systems. As we’ve seen with SVB, Credit Suisse and also First Republic Bank, it’s hard to restore confidence in a bank once it’s lost. This is especially true in the age of social media.
From a system-wide perspective, they include shadow banking risks, further fallout from higher rates, political risk, and of course a “hard landing". While we are optimistic the large diversified banks will weather whatever economic storm comes in the next year or so, we suspect the challenges may be most acute for smaller US banks and specialty lenders most heavily exposed to commercial real estate (CRE). Higher interest rates combined with structural changes in both office and retail segments (working-from-home and online competition trends) have begun to push commercial property prices lower and stress CRE investor cashflows. CRE loans make up about a quarter of all US bank lending, though the figure is 20% excluding multi-family CRE which is likely to be resilient. This 20% however is unevenly distributed: only 10% of loans at the largest 25 US banks are non-multifamily CRE, while the for all other US banks stands at 36%. Broadly speaking, the smaller the cohort of US banks, the higher their exposure to CRE.
Fortunately, most bank CRE lending has been done at reasonably low loan-to-value ratios (65% or lower), so loan losses should be manageable. However, while excessive loan losses can take down a bank, another type of loss – a loss of confidence – can be equally devastating and emerge much more quickly. As seen with SVB and Credit Suisse, plummeting confidence can do the job largely on its own.
The value of investments will fluctuate, which will cause prices to fall as well as rise and investors may not get back the original amount they invested. Past performance is not a guide to future performance. The views expressed in this document should not be taken as a recommendation, advice or forecast.