A long queue outside a popular bakery once led to a bank run in Hong Kong. The 1985 story goes that the bakery queue extended out of its door across the front of a nearby bank, which caused passersby to mistakenly think depositors were lining up to take their money out. As the rumour spread the bank was insolvent, panicked customers rushed to the bank to withdraw their deposits, until their fear created a self-fulfilling prophecy.

Though plausible, this story is probably apocryphal, but it does illustrate how fear feeds on itself to trigger bank runs.

The queues outside of bank doors were noticeably short during the run on the Silicon Valley Bank earlier this month, because most of the action happened online. Rumours that the bank was in trouble immediately went viral through social media networks. At the touching of screens, customers withdrew their funds at a rate of more than a million dollars per second. Within ten hours the outflow had reached $42 billion, a quarter of the bank’s deposit base.

Bank deposits are insured in the U.S. up to $250,000, but close to 94% of Silicon Valley Bank accounts exceeded this deposit-insurance cap. No uninsured depositor wants to be last in line during a bank run.

The Silicon Valley Bank went from being the U.S. 16th largest bank to its 2nd biggest bank failure within the space of 36 hours. The speed with which the madness of crowds can be harnessed through online networks, combined with the ease of online banking, has exposed the increasing fragility of the banking system.

‘Cash in the bank’ is supposed to be near as risk-free. This is why banks get away with paying so little in comparison to stocks and bonds. As at the end of last year, only half of Silicon Valley Bank’s U.S. deposits even paid any interest. The Silicon Valley Bank held $175 billion in customer deposits, with which it largely bought U.S. treasuries. Since the principal and payments are guaranteed by the U.S. government, customers should have been able to sleep easy, because the bank was being cautious with the money in its care.

The arcane part, albeit crucial, bonds prices are not guaranteed if you sell them before maturity. The bank had bought mostly long-dated bonds, with which it locked-in an average yield of 1.6%. This strategy looked clever while interest rates were close to zero. But these past twelve months interest rates have risen rapidly, so treasuries could now be bought that yield 4%. Now its bond portfolio looked deeply unattractive. This forced the bank to discount the prices of its bonds sold to cover customer withdrawals. When it announced $2 billion in losses from bond sales, the bank’s uninsured depositors were spooked and so begun a classic bank run.

Sean Croucher, portfolio manager of Asia Pacific Investment Advisors Limited in his January 2021 report warned “Long-term U.S. treasuries are priced to decline -19% in the event interest rates rise by just 1% - this will come as a shock to many who view government bonds as risk- free investments” The Silicon Valley Bank bond portfolio reflected a lack of risk management, which would lead to its bankruptcy. It is worth reflecting on why nobody expected this - until 8 March, just before news broke of its bond losses, its shares had been trading up 15% since the start of the year.

Silicon Valley Bank specialised in serving venture capitalists and technology startups. These clients are natural risk takers, willing to bear high risks in the pursuit of high rewards. They probably thought their bankers were the opposite. But bank executives’ compensation is tied to bank earnings and share price, so they are motivated to take on risk. Silicon Valley Bank didn’t need to buy long-term bonds, but in doing so to boost short-term profits, they gambled with their customers’ money and lost. Silicon Valley Bank was founded over a poker game 40 years ago.