At the beginning of last week, Silicon Valley Bank (SVB) was the 16th largest bank in the country. By the end of the week, it had collapsed into receivership and its deposits transferred to the newly-created, federally-owned Deposit Insurance National Bank of Santa Clara. 

Here is how it happened.

Deposits flow in, deposits flow out

Banks generate profits by borrowing at low interest rates and lending at higher interest rates. The borrowings come from three sources. Investors lend banks money for a period of time at a predetermined interest rate. Shareholders give banks money in exchange for a share of future profits. Depositors park their funds with banks in return for interest, the ability to write checks and make payments from their accounts, and for safe-keeping. These three sources of funding constitute the liabilities side of a bank’s balance sheet—the record of its financial position. 

The balance sheet also tracks the destination of those borrowed funds—these are the bank’s assets. Some of the funds end up as loans to companies and individuals. Loans are usually long term, whereas deposits are short term (most depositors can ask for their money back at any time). This presents a risk for banks—what if many depositors ask for their money back at the same time? This is called a bank run.

To bridge this risk, banks invest a large portion of their cash in stocks and bonds that can be liquidated quickly to cover deposit outflows. By far the biggest portion of these investments is government bonds, and other forms of government-backed debt, like federally-insured mortgage bonds. These types of bonds pay lower interest rates than do most companies on their loans. But banks hold government bonds because they are considered “risk-free”, and because they are easy to sell. 

A bank’s assets must always equal its liabilities (this is an accounting truism).

SVB experienced a surge of deposits during the pandemic, as its clientele—mostly tech companies—were flush with cash. Between the end of 2019 and the first quarter of 2022, the bank’s deposit balances more than tripled. SVB couldn’t rapidly find good companies to lend these funds to, and instead invested most of the cash in government bonds. In fact, the surge of deposits led SVB to be disproportionately invested in bonds rather than loans.

In 2022, as the Federal Reserve raised interest rates, venture capital funding dried up. Tech companies started to burn through cash. SVB’s deposits slowed and then reversed. 

Compared to other banks, SVB was particularly exposed to deposit flight. Its depositors were mostly companies, who tend to move cash quickly, compared to slower-moving individual customers.

They were also concentrated in a single industry, and one known to be prone to group-think. Some believe the earlier failure of Silvergate, another small tech-focussed bank, sent a wave of panic through the tech community. Investor Peter Thiel allegedly whipped that panic into a spiral of withdrawals from SVB after advising his companies to diversify their deposit accounts.

In order to avoid these kinds of spirals, the federal government insures deposits of up to $250,000. But SVB’s corporate depositors tend to have larger deposits than individual customers. If SVB failed, these companies stood to lose the value of their deposits over the federally-insured cap. At the end of 2022, SVB had $173bn in deposits. All but $8bn or so of these were above the cap.

What started as a trickle of deposit outflows quickly became a deluge. 

Marking to market

SVB sold some of its government bonds to raise cash and pay depositors back.

But the bank had bought those bonds in prior years, when the Fed was setting interest rates close to 0%. Today, in the face of high inflation and seeking to slow the economy, the Fed has raised its key rate to almost 5%. Older bonds that pay a lower interest rate were still liquid, but only at a lower price. It turned out that the risk lurking underneath SVB’s liabilities—rising interest rates—also haunted its assets.

SVB had a ticking time bomb on its balance sheet. The bank had taken advantage of a regulatory loophole that, under some circumstances, allows banks to ignore the deterioration in the market value of their bonds. In its accounting, SVB assumed those bonds were worth their purchase price. 

As a result, for every dollar of deposits SVB redeemed, it had to sell much more than a dollar’s worth of bonds on its balance sheet. It was rapidly running out of bonds to sell. Moreover, with every sale SVB recognized a loss that it had until then ignored.

By the end of last year, SVB’s equity—the institution’s true value—was a veneer stretched thin over a bloated balance sheet. The gap between where SVB’s bonds were marked on its balance sheet and their fair value in the marketplace had widened to over $15bn—more than its total shareholder equity. If the bank did have to sell the bonds to satisfy withdrawals, it would be bankrupt.

SVB’s solution was to ask investors for more money to keep the bank afloat and pay depositors. But by then it was too late. Either because they were waiting for a better deal, or not wanting to catch a falling knife, investors balked. The bank collapsed on Friday with insufficient cash to pay depositors.

Bad Luck, Stupidity, or Greed?

From one angle, this is a story about regulatory failure. SVB should have been compelled to hold more high quality bonds, and to recognize the deterioration in their value in real time—long before they were sold at beaten down valuations. In such a world, SVB would have had to raise capital early and often, and would likely not have faced a bank run. 

But regulations can only go so far. SVB’s management could have taken a number of actions to protect itself. It could have insulated its book of government bonds from deterioration using interest-rate swaps. It could have invested in shorter-dated bonds, which are less sensitive to rising interest rates. As deposits gushed in, it could have put restrictions on withdrawals. These measures would have steadied its balance sheet, albeit at a potentially severe cost to profitability.

A concern for profitability over stability may in fact be the best explanation for the failure of executives at SVB to prepare for risks that were lurking in plain sight. Stability is expensive. SVB preferred taxpayers pay for it.

In writing this article, I relied on the excellent work of Marc Rubinstein at Net Interest, Noah Smith at Noahpinion, Brad DeLong at Grasping Reality, Robert Armstrong at The Financial Times, Matt Levine at Bloomberg News, Adam Tooze at Chartbook, Joseph Politano at Apricitas Economics, and Matt Klein at The Overshoot. I heartily recommend all of these publications.