The FED has left the US Banking System Destabilized after Unprecedented Interest Rate Cycle
The failure of Silicon Valley Bank is exposing systematic vulnerabilities in the US banking system that will likely force the FED to pivot.
Summary
Over the last 28 years, US banks used to pay slightly less than the one-year treasury rate on their deposits. Currently, US banks pay 4.2%(!) less than 1-y treasuries, which represent a three-standard-deviation event.
The Silicon Valley Bank failure exposed to the broad public that bank deposits may also become unsafe and illiquid. As a result, the US Banking system is now facing increased risks of deposit outflows that may cause a chain reaction in bank failures.
US banks are trapped: After years of zero-interest rate policy, their loan books and securities yield far less than the risk free rate, preventing them from offering depositors competitive interest rates without sacrificing profitability.
US banks are also handicapped in terms of meeting large scale withdrawals with asset sales, as they have accumulated $620b in unrealized losses on securities.
I believe the FED must act quickly. Only a significant reduction in interest rates can resolve the current disequilibrium by disincentivizing deposit withdrawals and strengthening the asset base of banks.
If you, like me, were slightly concerned about the unprecedented magnitude and speed of interest rate increases by the FED to fight inflation, your intuition was just proven correct. This article describes how destabilized the US banking system has become at this point. The bank run on Silicon Valley Bank (SVB) may cause a domino effect at a time when the US banking system is the most vulnerable since 2008. The FED must act fast to stabilize the banking system and avoid serious harm to the economy.
The Fall of Silicon Valley Bank
The collapse of SVB represents the second largest bank failure in the history of the US. Remarkably, a bank run led to the demise of the 40-year old bank within just 48 hours. How could this happen?
I will discuss some unique aspects of the SVB failure. Yet, I believe these factors only led SVB to collapse first and it would be a fatal mistake to conclude the rest of the US banking system is fine. There are clear warning signs that numerous other banks are facing similar difficulties and may already be in the process of experiencing a bank run. Each further bank run would accelerate the demise of other vulnerable banks. I will now discuss why the whole US banking system is in a vulnerable state, even before the SVB collapse.
The Unprecedented Interest Rate Hike Cycle by the FED
The FED has just executed the most extreme interest rate hike cycle in recent history. What makes this event even more remarkable, is that it follows a prolonged period of zero interest rate policy. As we will see, this policy has put the US banking system in a uniquely vulnerable position.
The Current State of the US Banking System
The figure below shows a condensed version of the aggregate balance sheet of the US Banking System, based on the 2022 Q4 release of the FDIC.
For those unfamiliar with the accounting of banks. Banks finance their lending and investing operations by:
Deposits from customers (81% of funding sources). Customers used to get compensated for their savings by a deposit interest rate that was on average 0.5% less than the interest on short term treasuries in the money market based on the last 28 years. Moreover, customers expect their funds to be save. From the perspective of the bank these customer deposits represent liabilities, as depositors may withdraw their funds on a short term basis.
Banks may have further borrowings, e.g. from other banks or the FED (currently 9.3% of funding sources), as well as their own equity, or bank capital (also 9.3% of current funding sources). The less equity, the higher the leverage, and thus the higher the risk of a blow-up of the bank.
These funding sources for banks are normally not “free,” even though during ZIRP, they were effectively close to free. According to the FDIC statistics for Q4 2022, the total interest expense relative to total assets and liabilities was 1.1% annualized, which is very low by historical standards. This is possible only because banks basically still pay no interest on the deposits of their customers, despite short term risk free rates of 4-5%.
To generate a profit and healthy returns on equity, banks invest their funds in:
Loans (51% of assets)
Debt Securities (25% of assets)
Cash and other assets (24% of assets)
In Q4 2022, the annualized interest rate on those assets was only 4.1%. This is lower than the current Federal Funds Rate, yet alone the 1y or 2y treasury rate. This is the case because these debt instruments were issued during times of much lower interest rates and will take many years until they mature. Essentially, banks are currently holding risky loan term loan portfolios that are yielding less than the current risk free rate.
Nevertheless, banks currently still enjoy a relatively healthy net interest margin of 3%, which allows them to operate profitably with solid returns on equity. This is possible only because banks still pay effectively no interest on the deposits to their customers.
Liquidity Management and the Stickiness of Deposits
The banking business model is predicated on maturity transformation. Banks borrow money on a short term basis to lend it out on a long term basis. This principle works well as long as they don`t lose their deposits, in which case they would have to liquidate long term assets to meet the withdrawals. The essence of liquidity management is to get this trade off right.
Banks have some control over how sticky their deposits are. If the bank is financially healthy, customers have little reason to worry about whether or not their deposits are safe. Moreover, customers are normally paid a healthy interest rate on their deposits. This interest rate represents the return on “risk-free” savings from the perspective of bank customers. For the banks, these deposits represent a low cost form of debt.
Over the last 28 years, US banks have paid about 50 basis points less in interest rate on their clients’ deposits compared to the one year treasury rate. During times of financial crises, some banks try to attract more deposits by paying higher interest rates. At the height of the global financial crisis, US banks paid on average 2% more on deposits than the 1-y treasury rate.
The Current Unattractiveness of Bank Deposits Represents a 3-Standard-Deviation Event
How attractive are bank deposits currently? Savings on bank accounts are often thought of as risk free. Another alternative for a risk free short term investment is to invest in one-year US treasuries, or in a money market mutual fund.
In the figure above I calculate the spread between the interest rate on 1y treasuries vs. the average interest rate on bank deposits in the US. The data is from the quarterly FDIC statistics. Since 1995, US banks have on average paid 0.5% less on their deposits compared to the 1-year interest rate. However, this spread is relatively volatile. At the peak of the Global Financial Crisis (GFC), US banks paid on average 2% more than 1 year treasuries. The GFC was also associated with an enormous liquidity crunch, hence it made sense for many banks to attract deposits with high interest rates paid and thereby bolster their liquidity.
The 2010s have largely been characterized by ZIRP. Initially banks were offering customers more than 1-year treasures, but this changed over the years. Already in the mini rate hike cycle of 2018/2019, banks were not passing on higher short term interest rates to their depositors.
Now what happened after covid in 2022 is truly unprecedented. While short term rates are currently at 4.7%, banks still offer their depositors close to nothing on their deposits. In the context of the last 28 years, the current spread between 1y treasuries and the average interest rate on deposits represents a three standard deviation event; the spread we are currently seeing has a probability of only 0.3%. It is an extremely unlikely and also unprecedented constellation. To put it in plain English: The interest rates offered by banks today have never been as uncompetitive on a relative basis.
For the liquidity management of banks, this is a problem. Is your deposit base really sticky if you are paying your customers 4.2% than they could get with 1-year treasuries? As we will see in the next section, banks have unfortunately little other choice.
Banks are Trapped: Paying more than 0% on Deposits Would Destroy Net Interest Margin and Bank Profitability
Why are US banks not improving the attractiveness of their deposits if this is the crucial aspect of liquidity management? Well, banks used to pass on short-term risk free rates to their customers, but now they can’t.
The figure below shows the evolution of five interest rate measures since 1995:
The blue line is a proxy for the interest rate received by banks on their loans and security portfolios.
The orange line is a proxy for the cost of debt (deposit rate + borrowing) .
The grey line is the net interest margin (calculated as the difference between 1.+2.).
The yellow line is the 1-year treasury rate as a proxy for the market-based short term risk free rate.
The red line is the difference average bank deposit rates and 1-year treasuries.
As we can see in the graph above, there is normally a very high correlation between 1-year treasuries (yellow line) and the interest rate on deposits (orange line). Since 2022, this relationship is completely broken. While one year treasures have increased to 4.7%, the “cost of funding earnings assets” is still stuck at very low levels. As a result of this, we now observe the highest spread between 1y treasuries and deposit rates in 28 years.
The figure also answers why banks cannot increase the interest rates on deposits. As a result of ZIRP since 2010, the yield on earnings assets (blue line) has fallen below 4% in 2012 and only briefly returned slightly over that benchmark with the mini rate hike cycle in 2018/2019. After covid, this yield fell again significantly to hit new all time lows.
In 2022, something unprecedented happened: For a first time the earnings yield on bank assets (loans and securities) has fallen below the short term risk free rate. At the end of Q4, the earnings yield on loans and debt was only 3.5% per FDIC, whereas the rate on one year treasuries was 4.7%.
If banks were to increase the interest rate on deposits, they would directly and materially affect their net interest margin. The grey line shows that the net interest margin in the US banking system is historically pretty stable. During better years for banks it was up to 4%, but it has stabilized around 3% in recent years.
So far, banks were able to defend a 3% net interest margin only by not passing on higher risk free rates to their depositors. We can see how more competitive deposit rates would quickly eliminate bank profitability:
If deposit rates increase by 1%, NIM drops from 3% to about 2%.
If deposit rates increase by 2%, NIM drops from 3% to about 1%.
If deposit rates increase by 3%, NIM drops from 3% to about 0%.
Even during the Global Financial Crisis, NIM did not fall substantially below 3%.
Conclusion: Banks are trapped, they cannot offer competitive interest rates on savings accounts without giving up profitable operations. The implications of this situation are enormous. Essentially the FED’s interest rate policy has pushed the US banking system into a very vulnerable situation.
Over time, banks will be able to earn higher yields on their loan and securities portfolios, which will allow them to also partially pass on higher rates to their depositors. But this takes time. Due to the longer maturity of loan portfolios, the blue line (earnings yield on the bank assets) is a strongly lagged function of the yellow line (short term interest rates).
In other words, the FED should have increased interest rates at a much slower pace. The banking system can deal with higher interest rates over time, but it cannot absorb a 5% interest rate shock within just 12 months!
Now, things may go on seemingly fine until something breaks, and this has just happened. With the benefit of hindsight bias and the data analyzed in this section, it may seem obvious. Note however that the data here is only for the aggregate US banking system, or the average bank. Of course some banks will be in better shape with higher equity ratios, shorter term loan portfolios, possibly very sticky customer deposits.
And other banks will be worse than the average with even lower yielding loan portfolios, less equity, and a very nervous customer base. Silicon Valley Bank was exactly that - the weakest link in the chain. Silicon Valley bank loaded up on long term mortgage backed securities yielding less than 2% in 2021. The majority of assets for Silicon Valley Bank consisted not of loans, but of securities. With a long maturity securities portfolio, Silicon Valley Bank was among the banks that were the least able to increase the interest yield on their assets. At the same time, the depositors of Silicon Valley Bank are often cash burning startups that find themselves in a liquidity crunch. So they had to withdraw funds to finance cash burn, and they were among the most incentivized to look for higher yielding investments for their cash holdings to minimize cash burn.
When a Bank Run Meets a Trapped Banking System
Banks section above explained how banks are trapped currently. They cannot offer more than a three-standard-deviation unattractive level of interest rates to their depositors without sacrificing profitability.
At the same time, the bank run on SVB itself is a three-standard-deviation event given how rare bank runs have become nowadays.
What happens when two negative three standard deviation events meet each other? The combination of such two events will probably not wonderfully turn into a beautiful white swan.
Bank customers are notoriously lazy. They are expensive to acquire, but then even the most disadvantageous conditions barely make a customer vote with its feet. Paying zero percent on deposits when short term rates are at 5% perfectly showed how long things can go surprisingly well.
Now the failure of SVB is drastically changing the perception of the US banking system. The bank run suddenly brings to everyone’s attention that bank deposits not only pay little to no interest, but they are not even riskless and may become illiquid. I could not imagine a stronger event to finally cause bank customers to act. In a flight to safety and a flight to fair interest rates, we may see a mass movement in bank deposits at a time when the US banking system desperately needs these deposits to keep operating profitably.
Lets be frank, the situation is extremely delicate. Two days ago, Janet Yellen has admitted more banks are in trouble.
"There are recent developments that concern a few banks that I'm monitoring very carefully, and when banks experience financial losses, it is and it should be a matter of concern." Janet Yellen on March 10.
That was even before the final collapse of SVB. To my surprise, the FED and the US government have failed to issue a strong communication before the weekend. In the age of social media, this pace of action is unforgivably slow given that the US is now obviously facing a major banking crisis.
Over the weekend, pictures of long queues in front of ATMs at US banks are making the round on social media. I have no doubt that countless market participants will take action on Monday morning to de-risk their situation.
It is now a race between the unfolding of actions and a firm and effective response by the FED and the US government to fix the situation.
Unrealized Losses on Securities Have Further Destabilized US Banks
Investment securities represent 25% of assets on the balance sheet of US banks. These securities have suffered enormous mark-to-market losses as a result of the FED’s interest rate hikes. By the end of Q4 2022, US banks have accumulated unrealized losses of $620b according to the FDIC statistics. Were all banks to realize those losses, their asset base would shrink by 2.6% and their equity ratios would shrink from 9.3% to 6.7%.
Accounting rules allow banks to report these securities at cost, not at the current market value. This practice is effectively smoothing reported earnings and the procedure makes sense in normal times. If banks never have to sell those assets, they will still realize the interest income until maturity. As long as the net interest margin stays above 3%, banks can elegantly maneuver through any market volatility.
Well, that’s a big IF for many banks now. In fact, it is the reason why SVB fell first. SVB had among the most outsized losses relative to book equity, whereby the equity was effectively wiped out on a mark-to-market basis, taking the unrealized securities losses into account.
The graphic above shows what would happen to the tier 1 equity of some major banks if they were to realize losses on their securities.
It is evident that the last thing banks would like to do now is to sell their securities at a loss. Yet, some weaker banks in the system may have no other choice to meet deposit withdrawals. In the end securities are still way more liquid and sales of securities have lower transaction costs than potential sales of parts of the loan portfolio that represents more than 50% of assets.
Once the weaker banks in the system start selling their securities, these asset sales would cause the prices of those securities to fall and the interest rates to move up further. As the asset prices fall further, the unrealized losses of all other banks would increase more, bringing further banks into troublesome situations that may trigger further bank runs. At this point, this is a systemic risk that cannot be ignored by the FED and may force them at some point to step in with quantitative easing as the ultimate buyer of those debt instruments.
The Big Picture
As discussed, bank deposits are as unattractive as never before. Accordingly, deposits the least sticky in at least three decades. Yet banks are trapped, they cannot make deposits more attractive without sacrificing net interest margin at a time when appearing financially healthy was rarely more important.
On the asset side, the situation with the unrealized losses on securities is probably by itself a three standard deviation event, too. Banks will want to do everything to avoid realizing losses and central bankers should do everything to prevent banks to be forced into such a situation. It could cause a vicious circle that would further destabilize the system.
Bank capital / equity already looks rather precarious on a mark-to-market basis. But what happens if loans (more than 50% of assets) were also marked-to-market? What about those 30-year mortgage loans that were issued at interest rates below 2% in 2021? What about future losses on the loan books of banks in a recession, which seems to be the goal of the FED at this point to control inflation?
When focusing on net interest margin and return on equity, the banking sector still looks like it is in solid shape. However, this article has demonstrated that this is an incomplete picture. Right now, the banking sector is extremely vulnerable on both the asset and the deposit site. Several three-standard deviation events are aligning, and we have not even seen the interest rate hike cycle by the FED affecting the real economy.
How can the US Banking System be Stabilized?
To recap, the US banking system is not in equilibrium, as:
Interest rates on savings deposits are three standard deviations below short term risk free rates. This situation makes it rational to move deposits into 1-year treasuries or money market funds. Normally, bank customers are very lazy to change things, but with the SVB bank run, these deposits have not only proven to yield to far too little, but also that they can become unsafe and illiquid as experienced right now by SVB clients.
The assets of US banks are under pressure. Unrealized losses on securities become a real problem when those may have to be liquidated. Moreover, we are just at the beginning of the FED rate hike policy impacting the real economy and hence the loan books of commercial banks in the US.
What should regulators do?
The situation is extremely dangerous. First thing, deposits of all banks must be declared safe to prevent any further bank runs caused by a fear of bank runs. It should be the highest priority that this clear message arrives at the markets on Monday morning. I am however not sure if the vulnerability of the overall situation is clear to all decision makers. Any delay in a firm response will most likely have severe adverse consequences.
Declaring deposits save / insured is not enough. Depositors may still fear the illiquidity associated with a bank run. The vulnerability of the deposit base is only healed if banks offer deposit rates that are sufficiently attractive. Remember, in the liquidity crunch of 2008/2009, US banks were paying 2% more than 1-year treasuries. Today this would equate to a deposit rate of 7.7%(!). We are not yet at this point. But to disincentivize depositors from withdrawing, banks are well advised to increase their deposit rates fast.
The problem is, they can’t. Lets assume banks increase the interest rate on deposits to 3%, which is still 1.7% less than what the money market pays. This increase in the cost of debt for banks would wipe out their net interest margin as their loan book and securities yield to little. Jerome Powell and the FED have put US banks into an impossible situation. Hence, Jerome Powell and the FED are the only ones who can fix the situation and they better realize this sooner, rather than later.
The FED must Cut Interest Rates Now
How can the US banking system get back into an equilibrium state? By reversing the excessive interest rate increases of the past 12 months.
If the FED cuts the Federal Funds Rate to 2.0%, bank deposits yielding 0% would immediately become relatively more attractive, without the US banks having to sacrifice their net interest margin. To be frank, 1-year treasuries would still be significantly more attractive. If the level of stress in the banking system increases, banks might be forced to increase rates to 1% or 2% to attract deposits. But in order to be able to do so, they would need a strong enough capital base.
The FED cutting interest rates significantly would have a second important effect. It would strengthen the capital/equity base of US banks. Debt securities would immediately increase in value as interest rates fall, hence reverting the unrealized losses on securities in the balance sheets of banks. Even if some banks were still forced to sell securities to meet deposit withdrawals, they would at least be able to do so at a much smaller loss. US Banks would immediately be intrinsically more healthy, which by itself is the best remedy against bank runs.
In sum, the FED cutting interest significantly would stabilize the US banking system as follows:
Reestablish the competitiveness of bank deposits relative to short term interest rates, thus erasing the primary financial incentive for bank customers to withdraw deposits.
The loan and securities portfolio of banks would again yield more than the risk free rate, resolving the second major disequilibrium
Unrealized losses on securities would diminish as a result of the interest rate decrease, this would strengthen the true equity / capital base of US banks, eliminating another major reason for deposit withdrawals.
In sum, the unprecedented speed and magnitude of interest rate hikes have pushed the US banking system into a disequilibrium. It is exclusively in the hands and in the responsibility of the FED to fix this situation with an emergency interest rate cut.
Should the FED not act fast and decisively enough, I fear substantial domino effects within the US banking system that would have deep impacts on the real economy. I would not be surprised if the magnitude of the crisis would exceed the global financial crisis. I have still some hope that the FED can and will resolve the situation. But the time is running fast and basically all forces are working against the US banking sector and the FED (let alone short sellers having identified their next victims).
A Word on Interest Rates and Inflation
Yes, interest rates are the primary tool to fight inflation, which was way too high in 2022. Yet, inflation targeting as the sole goal of monetary policy can have important disadvantages. A major banking crisis would threaten the whole financial system and the functioning of the economy with unpredictable consequences. The FED has increased interest rates at a speed and magnitude far above than can be stomached by the banking system and economy after more than a decade of ZIRP. The consequences are on the table now and when the worst case events materialize, I am certain there will be nobody left claiming a major banking crisis is preferable to having inflation run slightly above target for a while.
The graphic below shows that the US CPI ex shelter has barely increased since June 2022, when the FED has become serious in its fight against inflation. As many readers know, Shelter is measured in a strongly biased way with lags of more than 12 months compared to real time data. While the January CPI release still showed shelter increasing some 7-8% year over year, real time data for housing prices shows they are falling since June 2022 and rents are falling since September 2022. If shelter was measured correctly, we would not be talking much about inflation at this point.
While we have heard nothing from the FED or the US Treasury on last Friday, president Biden has stated he is “confident” inflation numbers next week will come in fine. This leak has not calmed markets, but I am sure inflation will not be a hot topic of discussion anymore after the release of the February CPI next week.
This will give the FED the perfect excuse to revert its irresponsible interest rate policy of the last 12 months. Interestingly, no other central bank in the world has been this aggressive in fighting inflation - for good reasons as we are learning now.