‘Gradually, then Suddenly: How Banks Go Bust’

“How did you go bankrupt?” Bill asked.
“Two ways”, Mike said “Gradually, then suddenly”.

Ernest Hemmingway “The Sun Also Rises” (1926)

Part 2: Gradually

Solvency Risk 

It is often claimed that banks benefit from higher interest rates, but this is true only if they can increase the yield on their assets faster than the cost of their funding (predominantly deposits) rises, without the increase in the cost of money causing credit losses in their loan books or losses in their securities portfolios. When interest rates rise banks are notoriously slow in raising their deposit rates precisely because they rely on the inertia of savers to fatten their net interest margin.

Banks in the United States currently have $19.2 trillion of deposits (See Fig. 1) on which in Q4 2022 they were paying an average cost of 1.30% ($250bn per annum) which are funding assets of $23.6 trillion (See Fig. 2) that yield an average income of 4.11% ($970bn per annum). This results in annualised net interest income of $720bn (See Fig 7: US Bank Net Interest Income). As the cost of deposits rises with a lag to the Fed Funds rate, bank profitability will now fall, since it is unlikely that banks will be able to pass on higher interest costs to their borrowers without causing defaults, or invest in assets with higher yields, without more default risk.

 Fig 7: US Bank Net Interest Income⁹

As an illustration, if banks actually paid the current Fed Funds rate of 5% on their deposits it would cost $960bn per annum which would result in only $10bn of annual interest income at current asset yields, which with just $252bn of non-interest income, but $541bn¹⁰ of non-interest expenses, the industry would be loss making to the tune of $279bn at the pre-provision profit level. Although the US banking system is well-capitalised overall with $2.2 trillion of equity, there will be outliers in terms of banks more disadvantaged by the rising costs of deposits, which are less able to weather the net interest income storm. Lack of profitability – rather than lack of ready cash – can still cause more banks to fail. 

The best way for banks to manage rising interest rates is to have a high degree of liquidity (either from assets maturing or new deposit inflows) so that more new money can be lent at - or invested in securities - at higher interest rates. Conversely, banks which have fixed their returns on assets at low interest rates either by making long-term loans or through buying long duration securities, will be particularly vulnerable to having their net interest income squeezed, since all banks will have to continue to fund their assets with a higher cost of deposits. If the yield curve is inverted (with long term interest rates below short-term) this compounds the problem since most deposits, typically instant in duration, will reprice quickly and yield more than assets that are longer-term and wont reprice to the same degree.

Fig 8. US Banking Sector Deposit Beta (since 1984)¹¹

At the end of 2022 there were $19.5 trillon of deposits in the US banking system receiving an average interest rate of 1.3% at an annualised cost of $180bn (See Fig. 1 and Fig 8. US Banking Sector Deposit Beta). The biggest problem US banks currently face is retaining these deposits since depositors have the option of switching to money-market funds which (by investing in short duration government bonds or depositing at the Federal Reserve in the Reverse Repo Facility (See Fig 9. Where are all the deposits going?) are able to nearly match the current Fed Funds rate of 5%. 

Fig 9. Where are all the deposits going?¹²

Irrespective of any perceived credit risk by depositors (without which banks with riskier investment strategies would have the same cost of funding as those with low risk, a compelling argument against full deposit insurance) deposits should continue to migrate into money market funds (See Fig. 10: US Money Market Assets). Since there are so few opportunities to safely reinvest at this point of the credit cycle in assets with yields that will offset the higher cost of funding, banks will not pro-actively pay-up for deposits. Nevertheless, as Fig 8. Illustrates, deposit costs always catch up to the Fed Funds rate eventually.

Fig. 10: US Money Market Assets¹³

The Savings and Loan Crisis comparable

Although only two banks have officially “failed” so far, 2023 is already the second worst year on record for bust banks in terms of assets. But put into its proper perspective, “failed” banks so far in 2023 account for just 1.4% of system assets compared to 7.8% in 2008, meaning the current crisis is so far more comparable to the Savings and Loan Crisis of the late-1980’s, where hundreds of small banks went bust every year (See Fig 11: History of Failed US Banks 1934-2023). 

Fig. 11: History of Failed US Banks 1934-2023¹⁴

From 1982-1991, more than 1,400 banks failed, with the common cause of failure, funding long-term fixed-rate mortgages with short-term deposits in an era where the average Fed Funds rate exceeded the yield on those assets. In other words, the Savings and Loan banks found themselves having to pay more to their depositors than they were making on their mortgages, which as today with Treasuries and MBS, regulators – focusing on credit rather than duration risk - had designated as a low-risk asset. The Savings and Loan crisis played out over a decade since banks like today were not required to mark assets to market, meaning the industry went bust not suddenly but gradually.

US Banks: stuck between a rock and a hard place.

Banks which continue to lose deposits will now attempt to increase their liquidity, which will make less credit available to the real economy. This credit crunch will in turn lead to their customers in the real economy to focus on their own cash-flow, which will constrain economic activity and depress asset prices, meaning it will be more difficult to liquidate assets without realising losses. Gradually this downturn in the credit cycle will lead to more bank failures.

It is likely that this new credit crunch will accelerate the process of disinflation which began in the summer of 2022. Although trending downward, inflation has still been stickier than expected and is unlikely to fall back to the 2% level – that would justify monetary easing - without a significant economic crisis. 

After continuing to hike rates in March, central bankers have also made clear through their actions and rhetoric, that any monetary policy U-turn - involving cutting rather than raising interest rates, which would relieve pressure on US bank funding costs- will only take place after a crisis and not before. Therefore, the US banking system looks stuck between a rock, with rising deposit costs eating into net interest margins, and a hard place, of economic crisis causing elevated credit losses. Neither is attractive from an investment perspective.

Barry Norris
Argonaut Capital
April 2023