Why banks fail: the six classic killers
Why banks fail: thin capital in a recession, property booms, concentration, funding runs, interest-rate risk and rogue trading, from Barings to SVB.
Take office →Bank failures look dramatic and unique from the outside. From the inside they rhyme. Almost every failure in modern history is a version of one of six stories, often two at once.
1. Thin capital meets a recession
The most common story. A bank grows its loan book faster than its capital, runs close to the regulatory minimum, and then a recession arrives. Losses that a well-capitalised bank would absorb push this one below its requirement. Lesson: the buffer is the strategy. See what the CET1 ratio is.
2. The property boom
Real-estate lending pays well and feels safe because it is secured, until property prices fall and every loan sours at once. The Nordic banking crisis of the early 1990s, Ireland's banks in 2008 and a string of Danish banks in 2008–2012, such as Roskilde Bank, all began with rapid property lending in a boom. Lesson: cap concentration in commercial real estate, especially late in the cycle.
3. Concentration
One borrower, one sector, one counterparty that is too big. When the investment firm Archegos collapsed in 2021, banks with large exposures lost billions. Credit Suisse alone lost about $5.5 billion. Lesson: single-name and sector limits exist for the good days, when cutting exposure is cheap.
4. Funding runs
A bank can be solvent and still die if its funding walks away (see how bank runs happen). Northern Rock (2007) relied on wholesale markets that froze; Silicon Valley Bank (2023) relied on uninsured tech deposits that ran in a day. Lesson: fund with sticky, insured, primary customers and hold liquidity before you need it.
5. Interest-rate risk
A bank that funds long-term fixed-rate assets with short-term deposits can be broken by rising rates without a single bad loan. The US savings-and-loan crisis of the 1980s and SVB's collapse both had this at their core. See interest-rate risk and duration.
6. Culture: rogue trading, fraud and bad deals
Barings, Britain's oldest merchant bank, collapsed in 1995 after a single trader in Singapore hid losses of about £827 million. Overpriced acquisitions do slower damage: paying a huge premium at the top of the market (as in the takeover of ABN AMRO in 2007 by a consortium led by RBS, which needed a government rescue in 2008) leaves goodwill that cannot absorb losses when the cycle turns. Lesson: controls must grow before the business does, and acquisitions are best made in recessions, at a discount.
The common thread
Every killer works fastest on a bank that is thinly capitalised, illiquid and growing fast at the top of the cycle. Boring banks (fat buffers, conservative underwriting late in the cycle, sticky funding) survive, and buy the others at the bottom.
Frequently asked questions
What is the most common reason banks fail?
Credit losses in a downturn hitting a bank with too little capital, often after a period of rapid lending growth, especially in property.
Why did Silicon Valley Bank fail?
SVB combined interest-rate risk (long-term bonds that lost value as rates rose) with a concentrated, largely uninsured deposit base that withdrew very quickly once confidence broke.
Can well-run banks fail?
It is rare. Banks with strong capital, ample liquidity, diversified funding and disciplined lending usually survive crises, and often grow by buying weaker rivals.
Keep learning
Run a bank. Stay one.
A bank in a real jurisdiction in a real year, run from a terminal: the funding stack, the capital stack, the credit desk, the agencies, the supervisor and the board. Free, in the browser.