Observations on the Silicon Valley Bank

The spectacular implosion of Silicon Valley Bank (SVB) has occupied headlines since the weekend. This is a good summary of the sequence of events

In 2021, at the height of an investment boom in private technology companies, SVB received a flood of money. Companies receiving ever larger investments from venture funds ploughed the cash into the bank, which saw its deposits surge from $102bn to $189bn, leaving it awash in “excess liquidity”. Searching for yield in an era of ultra-low interest rates, it ramped up investment in a $120bn portfolio of highly rated government-backed securities, $91bn of these in fixed-rate mortgage bonds carrying an average interest rate of just 1.64 per cent. While slightly higher than the meagre returns it could earn from short-term government debt, the investments locked the cash away for more than a decade and exposed it to losses if interest rates rose quickly. When rates did rise sharply last year, the value of the portfolio fell by $15bn, an amount almost equal to SVB’s total capital. If it were forced to sell any of the bonds, it would risk becoming technically insolvent. The investments represented a huge shift in strategy for SVB, which until 2018 had kept the vast majority of its excess cash in mortgage bonds maturing within one year, according to securities filings...

A sale by SVB of $20bn of securities to mitigate a steep drop in deposits had focused investors’ attention on vulnerabilities in its balance sheet. They dumped its stock, wiping $10bn off its shares and crashing the market value of the bank — worth $44bn just 18 months earlier — to below $7bn... Customers had initiated withdrawals of $42bn in a single day — a quarter of the bank’s total deposits — and it was unable to meet the requests. The Federal Deposit Insurance Corporation — the US bank regulator that guarantees deposits of up to $250,000 — moved into the bank’s Santa Clara, California, headquarters, declared it insolvent and took control.

This and this are very good articles that point to all the important issues. Noah Smith has two very good posts here and here. 

This gives a good sense of SVB's importance to the Silicon Valley economy

SVB provided banking services to half of all venture-backed tech and life sciences companies in the US and played an outsized role in the life of entrepreneurs and their backers, managing personal finances, investing as a limited partner in venture funds and underwriting company listings.

This article by Michael Moritz of Sequoia Capital captures the importance of SVB to the Silicon Valley economy. 

It's reported that around 10,000 startups with deposits more than $250,000 have exposure to SVB, thereby putting their daily operations at risk. At the end of 2022, it held $157 bn of deposits across just 37000 accounts. 

The Fed has quickly stepped in with a new lending facility to provide extra funding to eligible institutions to prevent bank runs.

The so-called Bank Term Funding Program will offer loans of up to one year to lenders that pledge collateral including US Treasuries and other “qualifying assets”, which will be valued at par. The programme will eliminate an institution’s “need to quickly sell those securities in times of stress” and would be enough to cover all uninsured US deposits, the Fed said. The facility is backstopped by the Treasury, which put up $25bn. The discount window, where banks can access funding at a slight penalty, remained “open and available”, the central bank added.

SVBs problems today are a mix of illiquidity and insolvency. It took deposits and invested a lot in bonds and loans, whose prices kept going up when the rates were low. Now that rates are rising, the prices of these fixed income assets have fallen, thereby sharply eroding the asset value of SVB. This exposes the Bank to the classic bank run risk, whereby its assets are neither adequate nor liquid to meet the sudden surge in deposit redemption requests. But it's likely that with greater liquidity support, the Bank can manage the current situation. The question is who should bear the costs?

The Silicon Valley Bank (and others like Signature) were effectively a captive saving and lending institution for Silicon Valley startups. Start-ups used to raise capital from venture capital funds and deposit the proceeds with SVB to draw them down as per requirement. Besides, the presence of the deposits also meant that, unlike other banks which don't lend to startups, SVB offered them loans. SVB also invested as a limited partner a part of its rising deposits into the same VCs which were funding its startup clients. Finally, SVB ended up managing the personal finances of many tech entrepreneurs and VC investors. This was a mutually beneficial partnership for all concerned - startups, VCs, and SVB. 

Needless to say, there are already loud self-serving demands for government bailouts from the same stakeholders who are also among the most vociferous opponents of government regulation. Garry Tan, the President and CEO of Y-Combinator, described this as an "extinction-level event for start-ups and called for government help to make sure "American tech industry is not set back by a decade and we can save thousands of workers jobs". He also asked tech entrepreneurs to write to their local Congressman for immediate government help. Bill Ackman has warned of a run on all but the biggest banks if the government does not guarantee all of SVB's deposits or if the lender is not acquired by JP Morgan, Citigroup of BofA. Michael Moritz makes the veiled threat that if "adequate steps aren't taken to ensure that tens of thousands of entrepreneurs can meet their payrolls and other obligations, the US hold on any number of groundbreaking technologies will be substantially weakened"!

John Thornhill summarised the hypocrisy surrounding the episode and its aftermath,

It may stick in some throats that the US and UK financial authorities have had to engineer an emergency rescue for an institution, and an industry, that is so fond of railing against government intervention and lobbying against stricter regulatory oversight... Janet Yellen, Treasury secretary, had invoked a “systemic risk exception” to justify the support... The SVB fiasco also shines an unforgiving spotlight on the hypocrisy of some of the biggest venture capital players on both sides of the Atlantic, who privately urged their portfolio companies to pull their money from the bank and then later publicly called for government support. SVB collapsed on Friday as a result of a classic bank run after customers withdrew $42bn of deposits. Just like many of the banking titans after the global financial crisis of 2008, tech tycoons appear to favour the privatisation of profits and the socialisation of losses. There are few libertarians in a financial foxhole.

The first-level losers in case depositors are forced to take haircuts on their deposits are the depositors, primarily the lenders. To the extent that the main investors in startups are the venture capital firms, the largest losers from a startup going under are the VCs. It's therefore only natural that VCs bear the costs of any liquidity support for SVB. Government bailouts are effectively a bailout of the VCs and their investors. 

It's therefore only appropriate that the VCs should be encouraged to step in and provide temporary credit backstops to the startups who have made deposits to SVB. Alternatively, the VCs with the largest exposure to SVB should be brought together to provide liquidity support to SVB. Instead of being nationalised with taxpayer bailouts, SVB should be taken over by these exposed VCs. The US government has already announced that no losses from the resolution of SVB or Signature will be borne by taxpayer and the shortfall would be funded by a levy on the rest of the banking system. 

This is also a teachable moment to all those who oppose regulation of private capital markets. When the volumes of money involved become in the order of tens or even a few hundreds of billions, then contagion becomes inescapable. The interconnections in public and private markets start to become evident and pose systemic risks. 

The VC industry has benefited from externalisation of several social costs and appropriation of all private benefits. Fundamentally, as John Thornhill has written, the SVB "provided services for risky, collateral-light tech start-ups that were far from ideal customers for traditional financial institutions". Never mind the risks, which if materialised could be socialised. Another is the absence of a debt counterpart to the VC equity capital, which meant that entities like SVB and Signature stepped in to meet the felt need. This allowed startups to raise leverage and thereby further boost the returns to VC investors. Then there is the non-trivial cost of treasury management of the VC capital flooding into the startups. Instead of structuring and managing the flows into the start-ups, the VC managers are incentivised to push out as much of their dry-powder and leave the treasury management concerns to the startups, who invariably go with the herd. It's only appropriate that these social costs are internalised.  

On a related note, regulators have several lessons from this episode. There are several points of disturbing regulatory failures.  

It's now known that at the end of last year 96% of SVB deposits were not covered by the FDIC's deposit protection coverage. This is in contrast to 30-40% for the major Wall Street banks. Such excessive build up of deposit protection risk was  well known to all concerned, including the regulators. 

A major source of failure of a bank is correlatedness. SVB suffered from multiple layers of correlatedness risks. One, SVB took deposits from and also lent out to startups. This meant that once the startups started to struggle, their debt servicing abilities got squeezed and they has to draw on their cash runway parked as deposits faster than expected. Two, a disproportionate share (atleast in value) of their depositors being startups also meant that SVB was exposed to risk of a reversal of the credit boom squeezing VC credit flows to startups. Another correlatedness risk which was ignored concerned SVB's disproportionate exposure to loans made to startup firms, which meant a high likelihood of concentrated withdrawals if startup funding dried-up. Finally, there is the commonplace herd-like behaviour associated with both startups and VCs.

The run on the bank that laid SVB low has been held up as an example of the herd-like behaviour often displayed by tech investors. A number of venture capital firms urged companies they had invested in to take their cash out of SVB after the bank said it was seeking to raise more capital.

Then there is the colossal failure of the SVB management. Its management did not think anything amiss when it assumed such massive correlatedness risks - taking deposits from and lending to a group of borrowers whose valuations were already heavily stretched, and all this at a time when rates were ultra-low and a reversal on the interest rate cycle was inevitable. Sample this

“It turned out that one of the biggest risks to our business model was catering to a very tightly knit group of investors who exhibit herd-like mentalities,” said a senior executive at the bank. “I mean, doesn’t that sound like a bank run waiting to happen?”

The failure of SVB mimics the several problems with VC industry in general. In particular herd mentality. The startups herded into SVB with their deposits, the VCs encouraged the trend, and SVB management kept buying those bonds and taking deposits.

Update 1 (14.03.2023)

After all the brave talk of no bailouts, the US authorities finally bowed to pressure and fears of systemic spillovers, and have announced that all SVB depositors would be fully repaid through an emergency lending facility. 

The so-called Bank Term Funding Program will offer loans of up to one year to lenders that pledge collateral including US Treasuries and other “qualifying assets”, which will be valued at par. The programme will eliminate an institution’s “need to quickly sell those securities in times of stress” and would be enough to cover all uninsured US deposits. The facility is backstopped by the Treasury, which put up $25bn. The discount window, where banks can access funding at a slight penalty, remained “open and available”, the central bank added. The regulators said all depositors of SVB would have access to their money on Monday, as would those of Signature, which was closed by the New York Department of Financial Services before being placed under FDIC control and marketed for sale. Officials on Sunday said no losses stemming from the resolution of either SVB or Signature’s deposits would be borne by the taxpayer. Any shortfall would be funded by a levy on the rest of the banking system. They added that shareholders and certain unsecured debtholders would not be protected... The senior Treasury official denied that the move represented a bailout because shareholders and bondholders of the two banks had been “wiped out”. 

The decision to cover the entire deposit is a crossing the rubicon moment of moral hazard for deposit insurance. Roger Lowenstein writes in NYT

Federal deposit insurance was introduced 90 years ago during the heart of the Great Depression. Ever since then, small depositors within the F.D.I.C. limit of coverage have slept soundly. Now, in light of the bank failures of the last few days and the F.D.I.C.’s extension of coverage, why will any depositor worry about risk? Having bailed out depositors of two banks in full, how will the government refuse others?... Until the 1970s, the F.D.I.C. limit on deposit coverage increased only slowly. But in 1980, as banks came under pressure from soaring inflation, Congress raised the cap to $100,000, over the objections of the F.D.I.C. itself. In the 2008 crisis, the limit was raised to $250,000. And after the failure of IndyMac in 2008, the F.D.I.C., when possible, quietly protected uninsured depositors. In the rescue of S.V.B. on Friday and of Signature Bank in New York two days later, the F.D.I.C. overtly ignored the cap and rescued all depositors, irrespective of size. This is a breathtaking leap.

This from former FDIC Chair Sheila Bair

Preventing “systemic risk” was repeatedly used as a rationale for bailing out Wall Street during the 2008 financial crisis. The 2010 Dodd-Frank Act was supposed to have fixed all of that by strengthening regulation and banning government bailouts. Yet, banking regulators have now decided that the failure of two midsized banks, Silicon Valley Bank and Signature, pose systemic risk, requiring the Federal Deposit Insurance Corporation to pay off their uninsured depositors. At combined assets of $300bn, these two banks represent a minuscule part of the US’s $23tn banking system. Is that system really so fragile that it can’t absorb some small haircut on these banks’ uninsured deposits? If it is as safe and resilient as we’ve been constantly assured by the government, then the regulators’ move sets dangerous expectations for future bailouts. The uninsured depositors of SVB are not a needy group. They are a “who’s who” of leading venture capitalists and their portfolio companies. Financially sophisticated, they apparently missed those prominent disclosures on the bank’s websites and teller windows that FDIC insurance is capped at $250,000.

From Ken Griffin of Citadel, the hedge fund

The US is supposed to be a capitalist economy, and that’s breaking down before our eyes... There’s been a loss of financial discipline with the government bailing out depositors in full. It would have been a great lesson in moral hazard [to leave the uninsured depositors exposed to risk. Losses to depositors would have been immaterial, and it would have driven home the point that risk management is essential.

Meanwhile in the UK, the government convinced HSBC to take over SVB UK with 3300 startup clients for £1.

Update 1 (19.03.2023)

This is a very good article on the hypocrisy of the VC and start-up industry in loathing government when times are good but calling for government help when things go wrong. 

SVB is the latest example that highlights the reality that being "cool" in banking is being "risky". This about some of the risks,

Its client base — which was concentrated among companies or entrepreneurs typically with large deposits — was always higher risk in a bank run than a diversified bunch of retail customers. Meanwhile the lending side of SVB’s strategy — the money making part of it — had some related problems. The bank did not have a lot of leverage to impose higher rates on its borrowers, partly because it relied on them so much for their deposits. Neither was it well-positioned to lend to other industries, given its focus on one single ecosystem...

Another problem was that the bank overlooked the important but boring job of risk management. At a time of low rates, its decision to invest in long-dated held-to maturity securities to boost profit was not irrational. But a more anxious and less cool bank might have paid more attention to the mismatch risk on the balance sheet. They could have used swaps to hedge the portfolio against higher interest rates, or invested in more suitable securities, a former European bank executive remarked. Another red flag was SVB’s spectacular growth, the European bank veteran added: “Banks rarely grow much faster than their peers. When one does, it means they are taking more risks.”

Scott Galloway has this to say about the moral hazard with deposit insurance,

I am a founder, director, or investor at four firms that have approximately $20 million in deposits at SVB. We did not pull out a dollar. We didn’t keep our deposits at SVB because we’re moral or felt an obligation to save the bank, but because I went to the FDIC site and found that 73 banks have failed in the last 10 years, and all had their deposits backstopped. I didn’t lose a minute of sleep.

Martin Wolf has two graphics which capture the scale of challenges faced by banks. The first is about the the massive spike in unrealised losses in the banking system due to the rising interest rates.

And the second is about SVB's erosion of its Common Tier equity due to the unrealised losses in its assets.