A modern version of an old story.

— Mar 22, 2023

Bank runs are almost as old as banking. But familiarity does not make them any more comfortable when they happen. The failure of Silicon Valley Bank, the distressed sale of Credit Suisse and turbulence in the wider financial system had left markets on edge and rekindled memories of 2007 to 2009.

To understand the bigger picture, it is worth stepping back and looking at both the immediate triggers of the current problems and their longer term drivers. Whilst a full-blown banking crisis is still far from the base case, the events of the past couple of weeks may well be enough to slow the pace of central bank tightening ahead.

The basic business model of banking is what economists tend to dub ‘maturity transformation’. They transform short term liabilities into long term assets by accepting deposits withdrawable on demand and making longer term investments of loans. That is a crucial part of the functioning of the financial system and one for which they are, in theory, rewarded with a stream of profits. But those profits are far from risk-free, the model itself contains an always present danger in the mismatch between short term liabilities and longer-term assets. As anyone who has watched either It’s a Wonderful Life or Mary Poppins knows, if enough depositors get twitchy at the same time and move to withdraw their deposits at the same time then the bank will not have ready cash to meet the flow.

The dynamics of a bank run can quickly become self-reinforcing. It is an unusual case where acting seemingly irrationally can be the entirely rational thing to do. If rumours spread that a bank is in trouble  and it appears that, say, 50% of deposits are likely to be withdrawn in a short period of time, then it makes perfect sense to want to be one of the 50% of deposit holders who gets their cash out even if one suspects that the rumoured problems are overblown.

Silicon Valley Bank, which had a balance sheet of around $200bn, seemed – just 18 or so months ago – to have a solid business model. It was not just a bank in Silicon Valley but was almost the bank for Silicon Valley. It had become the go-to provider of banking services for venture capital backed, California based tech start-ups with a solid business in tech companies outside of the Sunshine State too.  It provided a high level of client care, hands-on relationship management and helped connect firms with the potential to work together. It was very much a part of the tech community. It was that community which helped bring it down so quickly. Whilst the bank almost certainly thought of itself as having thousands of individual depositors, many were actually funded by the same few dozen venture capital firms. The real economic decision makers when it came to financial matters, such as where firms held their cash, numbered in the hundreds rather than the thousands. When a few large VC funds told their investees to move their cash many others quickly joined them.

The more interesting question is not how the run happened so quickly but why it took place at all. That involves stepping back to look at the longer term factors behind SVB’s recent troubles, which go much deeper than the tech sector.

The really interesting difference between SVB’s failure and the banking crisis of 2007 is that this one does not involve the same kind of reckless lending decisions and loans gone bad. The subprime crisis was at least easy to grasp – banks had extended loans to people who could not afford them. This time around SVB’s assets consisted  almost entirely of supposedly ultra-safe long dated US government bonds and government-guaranteed mortgage backed securities. This is not really a story about credit risk.

Once upon a time US commercial bankers used to joke that they lived by a 3-6-3 rule: pay depositors 3%, charge borrowers 6% and always be on the golf course by 3pm. But in a world of ultra-low interest rates, the 3-6-3 rule no longer applied.

SVB did not really operate as the textbook model of a bank suggests. It paid depositors very little – although it did offer a high level of service – and it did not really make many loans at all. Instead it bought long dated safe bonds. The rule was less 3-6-3 and more – before 2021 anyway – 0-1.5-0 – pay depositors nothing and earn 1.5% from the yield on US government debt, with little hope of making it to the golf course in the working day.

That model seemed to work. Customers were happy enough with the bank’s services that they did not demand a high return on their cash balances and the bank itself seemed to be earning – just about – enough from its ultra-safe long term assets to generate a profit. But the pandemic and its aftermath put paid to that. The initial impact of the pandemic seemed like a blessing for SVB. Money poured into VC funds and via their investee firms into SVB leading to rapid expansion in its balance sheet. But after that came the rise in inflation and related rapid tightening in central bank policy.