In March 2023, a series of problems suddenly emerged in a few financial institutions in the United States and Europe, bringing back memories of the Great Financial Crisis. Silvergate, Signature Bank (both exposed to the crypto industry), and Silicon Valley Bank (SVB) were in the spotlight in the U.S., while Credit Suisse, a Global Systemic Important Bank, was involved in Europe. The problems of the institutions were complex and mainly due to individual actions that cannot be generalized among different banks.
SVB, was the first institution that raised concerns about the situation of the entire banking system. Although SVB was not the first bank to encounter problems, one of the major difficulties for the institution was a common issue among other banks in the U.S. The case of SVB highlighted the problems stemming from a portfolio that was highly exposed to interest rate risk. Although this risk exposure may also be common to other entities across the U.S, some specificalities, among them its poor risk management framework, were also important to understand the failure of the institution.
So far, regulatory institutions have been able to contain the negative spillovers coming from SVB’s failure through quick action aimed at solving liquidity pressures. However, the emergence of new problems, beyond the banking sector, cannot be ruled out. Additionally, the situation has complicated the actions of major central banks in the short term, as it has brought about contradictions between their principal mandates of controlling inflation and the risks to financial stability.
What happened at SVB?
The situation faced by Silicon Valley Bank was just a reminder of one of the main risks embedded in the banking business, the maturity mismatches. In a general way, a maturity mismatch arises when the expected cash outflows are greater than the expected cash inflows over a period. This situation is really common for a bank as its financing (i.e. deposits) is considered to be redeemable at a short notice and part of its revenue generating activities (i.e. loans, investments or both) is linked to the long-term.
In the presence of these maturity mismatches, the level of confidence around a bank is utterly important to avoid liquidity crises. If the confidence around a bank evaporates, investors run to withdraw their funds from the institution. In this situation the bank, once its liquidity buffers have been exhausted, is forced to sell its assets at a loss, to honor its commitments with clients. If the liquidity crunch is short-lived, a bank can withstand the pressures arising from the outflow of funds on its own. If the liquidity pressures continue and the losses coming from the fire sales dent its equity base, a solvent institution might become insolvent as a result of the lack of short-term liquidity.
Liquidity issues and bank runs have been a common issue in financial history and consequently, financial authorities have created a sort of public “safety net” to avoid widespread crisis of confidence. In a general way this “safety net” combines different layers of regulation and supervision, capital and liquidity requirements, the lender and market-maker of last resort functions as well as the deposit insurance. From the private side, the public “safety net” can also been complemented by other financial intermediaries stepping as lenders of last resort and the introduction of some “circuit breakers” with dubious results, as evidenced by the so called “corralitos financieros”.
Before the post-mortem of SVB, it is worth having a quick look at the 2022 annual report. As shown by the consolidated balance sheet at the end of 2022, SVB was an entity with 211 billion in assets whose business model was more on the investment side[1]. When looking at the asset side, loans accounted for 35% of total assets and its securities portfolio accounted for 57%. From the liability side, the financing of the bank was mainly reliant on deposits as they represented around 88% of total liabilities. Without differentiating among asset classes, total equity was around 16 billion, representing just an 8% of total assets. Additionally, it is worth mentioning that the bank saw an important increase in deposits since 2019 as a result of the huge forced savings coming from the Covid-19 shock and the bank’s exposure to the 2021 tech boom.
Although the huge increase in deposits could be seen as a proxy for potential future financial problems, it is not possible to accurately assess the financial health of an institution based solely on the numbers mentioned above. A more detailed look at the composition of assets and liabilities, the hedging strategy of the bank, as well as regulatory changes introduced in 2018, might provide a clearer picture.
First, from the asset side, the risks coming from some of the activities were somehow hidden or diluted due to the current accounting standards. As shown in the investment securities activities of the consolidated financial statements, SVB trading book was heavily exposed to interest rate[2] and wrong way[3] risks. This is explained by the huge bet they made on the “lower for longer” scenario through Treasuries and Mortgage-Backed Securities (MBS). If it is true that Treasuries and, to a lesser extent MBS, are considered the risk-free assets par excellence as a result of its low or non-existent credit and liquidity risk, they continue to be exposed to interest rate risk. As a result, in an environment characterized by a fast tightening of financial conditions, the successive hikes of rates by the Fed were depreciating these two assets.
But, if it was that clear that these assets were quickly depreciating in price, why did it take so long for investors to realize SVB’s trading book losses? Well, the truth is that the current accounting standards allow for the “partial” recognition of losses from the securities portfolio. This has to do with the different treatment of the securities portfolio coming from the categorization of assets between Hold-To-Maturity (HTM) and Available-For-Sale (AFS). When it comes to the HTM, the possible losses or profits derived from these assets are not directly categorized as profits or losses as they are not reported in the income statement. The reason for this is that these assets, shown in the balance sheet at amortized cost, are in principle kept until maturity and hence, not sold. In other words, the changes in the value of the asset are not 100% materialized from an accounting perspective. Regarding the AFS assets, the possible losses or gains coming from these exposures are shown to investors through “other comprehensive income”, reported in the income statement. Looking at the numbers, SVB moved securities with a carry value of $8.8 billion from AFS to HTM in 2021, postponing the recognition of losses. In 2022, the bank reported $91 billion and $26 billion of HTM and AFS assets, respectively.
Secondly, regarding liabilities, SVB considered deposits as a “sticky” source of financing. If it is true that deposits are one of the safest sources of financing for banks, there should be a distinction among the type of deposits and, more importantly, the level of deposits covered by the deposit guarantee. In the case of SVB, the concentration of deposits among tech startups and venture capital firms made the U.S. deposit insurance not to be as effective as expected. At the end of 2022, the estimated uninsured deposits were around 95% of the total deposits as a result of the deposits amount, higher than $250.000 million, or their origin outside the U.S.
Thirdly, it is possible to talk about the poor risk function of SVB, mainly explained by the inability of the Board to correctly identify, monitor and mitigate the risk exposures of the bank. As reported in the annual report, the amount SVB had on interest rate derivatives was only 563 million for a trading book of $107 billion, heavily exposed to interest rate risk. The reasons behind this risky behavior had to do with the benign expectations of the Board regarding the evolution of interest rates as well as the desire to boost short term profits[4]. Concerning the levels of equity, it is easy to see that $16 billion was insufficient to withstand in a good manner the possible losses from the trading book. Just taking into account the HTM portfolio, there were $15 billion of unrealized losses.
Moreover, some analysts flagged that SVB had to rely on the liquidity provided by the Federal Home Loan Bank System (FHLBS)[5] at the end of 2022. The necessity to tap FHLBS liquidity is considered problematic as the government-sponsored institution is seen as the “lender of next to last resort”. As showed in 2008, some of the financial institutions that came under financial stress had to rely on the liquidity of FHLBS. Financial institutions with problems have incentives to draw on FHBLS liquidity as they can be provided with cheaper sources of financing, due to the government-sponsored status. Additionally, FHBLS financing might also be attractive when compared to the traditional repurchase agreements since, under certain conditions, the financial institutions are not required to pledge collateral (i.e., margin calls) when market conditions change.
Finally, the regulatory changes introduced in the 2018 Economic Growth, Regulatory Relief, and Consumer Protection Act exempted SVB from some requirements. Since SVB was a firm with $100 billion to $250 billion in total assets, the regional bank was not subject to the standardized Basel III liquidity requirements[6] or the requirement to conduct and publicly disclose the results of company-run capital stress tests.
The problems for SVB financial group began at the start of 2022, as evidenced by its market value, due to the bank’s high exposure to the tech sector and its investment choices. Despite this, the institution was able to enter 2023 with relatively few issues. However, the situation for SVB changed rapidly in March 2023 due to the voluntary liquidation of Silvergate Bank. Although Silvergate made some bad bets on the crypto world, some of the problems stemmed from marking HTM assets as AFS, recognizing the losses in the income statement.
The problems of Silvergate Bank did not take a lot of time to spill on SVB, as the bank was also highly exposed to high rate-sensitive assets and had a thin equity cushion. On the 8th of March, SVB published a statement in which they acknowledged the materialization of 1.8 billion in losses coming from the sale of 21 billion in securities from its AFS portfolio. Additionally, in the same statement they also announced a $2.25 billion capital raise to shore up the bank’s financial position. These announcements lead to a Moody’s downgrade of SVB’s parent company.
Contrary to the bank’s expectation, the announcement of losses started an eye-popping bank run of $42 billion on 9th March. However, the situation did not last too much time as the U.S. authorities decided to close the bank on 10th March, transferring the assets to the Deposit Insurance National Bank of Santa Clara for its liquidation, effectively leading to the second largest banking failure in the U.S.
In addition to the closure of the bank, the authorities decided to invoke the systemic risk exception on 12th March, fully protecting the deposits of SVB and Signature Bank[7] with the end goal of stemming the emergence of a crisis of confidence. It is important to mention that as regulators made clear that even small bank failures can have systemic implications, the willingness to use the systemic risk exception can be consider a de facto guarantee on all deposits despite the uncertainty surrounding the case-by-case judgements.
In addition, on March 12th, the Federal Reserve (Fed) created a new lending facility to backstop the banking system and prevent further spillovers. The new program, called the Bank Term Funding Program (BTFP), allows financial institutions to obtain loans with a maturity of up to one year. To be eligible, institutions must pledge U.S. Treasuries, public agency debt, or MBS at par value with no haircuts applied as collateral. Following Bagehot’s principle, the interest rate applied to the loans is 10 basis points above the one-year overnight index swap rate.
It’s worth mentioning that the creation of the BTFP under section 13(3) of the Federal Reserve Act gives the Fed more flexibility because the facilities created under this section allow the central bank to provide liquidity to other financial intermediaries besides banks. With the creation of this facility, the Fed eliminates the need to sell Treasuries and MBS quickly at a discount or borrow from other private financial intermediaries, effectively creating an insurance against interest rate risk for the entire financial system.
In addition, since March 12th, the Fed has also been providing liquidity through the discount window, which is the primary liquidity facility for the banking sector, on the same terms as the BTFP regarding margins. Furthermore, the Fed has extended liquidity to the bridge banks that the FDIC established for SVB and Signature (Graph 1).
According to the Fed’s balance sheet, as of April 18, the total amount of loans provided under BFP was $73.9 billion. The discount window credit and the bridge bank financing amounted to $69.9 billion and $172.6 billion, respectively. It is important to note that financial institutions made more extensive use of the discount window in March compared to the Great Financial Crisis and the Covid-19 recession. In just the first week after the start of the banking turmoil, the use of the discount window peaked at $152.8 billion, which is higher than the $110.7 billion and $50.7 billion required in 2008 and 2020, respectively. As for the BTFP, it may be used less than the discount window due to the stigma associated with the facility. Additionally, the pre-pledge collateral of the discount window can also help when explaining this. Financial institutions might prefer to hold collateral to meet other needs, instead of pledging it to obtain BTFP financing.
According to the data on banking deposits, depositor sentiment has shifted following the banking turmoil in March. Initially, large banks saw an inflow of deposits as depositors preferred to park their money with these institutions (Graph 2.a). However, this was short-lived as bank accounts offered lower yields compared to other investment opportunities. In contrast, small banks experienced significant deposit outflows through the whole month of March, possibly due to their similarities with SVB and Silvergate (Graph 2.b). This indicates that, even after the different actions of the authorities, small U.S. banks might still face challenges.
Some broader implications.
Following the financial shock initiated by SVB, attention was turned to the HTM portfolios, which were qualified as the “ticking time bomb” in which U.S. banks were sitting. According to the FDIC’s Banking Profile of 4Q 2022, the level of unrealized losses for U.S. banks coming from the AFS and HTM portfolios was around $620 billion in the fourth quarter (Graph 3).
Similar to SVB’s situation, the losses for U.S. banks were mainly coming from the longer-term maturities of assets, although these levels of unrealized losses were not as worrisome as some might think. Drawing conclusions about the state of the US banking sector based solely on these numbers is, at a minimum, somewhat presumptuous. This is because the general picture given by FDIC’s data does not provide any information about some important parameters needed to understand the expected losses of individual banks as well as their ability to withstand those losses. Among others, some of the relevant parameters missing here were the maturity profile and composition of banking assets, the level of hedging against possible losses, the levels of capital and liquidity of banks, the level of covered and uncovered deposits, the liability composition, the possibilities to substitute their short-term liquidity and the existence of central bank liquidity facilities, insuring banks against the depreciation of Treasuries and MBS.
If it is not possible to rule out the emergence of problems in the banking sector related to the recent market developments, it is more likely that the problems come from other parts of the financial system. As shown by the recent mini-budget crisis, the nonbank financial intermediaries (NBFIs)[8] are probably the institutions that are more exposed to the sudden price corrections of financial assets. As highlighted by the International Monetary Fund in the April 2023 Financial Stability Report, the current environment of high inflation and tightening financial conditions, together with a combination of poor liquidity, excessive use of leverage and the interconnections with other financial intermediaries might prompt and exacerbate fire sales and investor runs across different financial institutions. These problems could lead to increased levels of price volatility and the contagion of problems across the financial system, ultimately affecting the real economy.
Another of the debates triggered by the deposit run on SVB has been the design of the deposit insurance system and whether or not it is necessary to increase or eliminate the limit[9]. The deposit insurance is one of the cornerstones of the “public safety net” around the banking system. As previously mentioned, the engagement of banks in liquidity and maturity transformation makes them to be susceptible to suffer bank runs. This individual bank runs, together with the existence of imperfect information, the herd behavior and the incentives associated with the “first-mover advantage” can lead to a crisis of confidence affecting the whole banking system. In this context, the deposit insurance acts like a circuit breaker removing the incentives that depositors have to withdraw their deposits.
However, the deposit insurance comes also with some disadvantages as it reduces the necessity of depositors to monitor the risk-taking behavior of the institutions in which they deposit their funds. This can lead to a state in which banks, in order to attract more depositors, increase their risk-taking as a result of search-for-yield strategies. Regulators, aware of the existence of moral hazard, have tried to reduce these perverse incentives through the supervision and enforcement of risk-based rules, the creation of a bank levy to finance the deposit insurance fund and the imposition of an upper bound in the amounts of deposits guaranteed. In the case of SVB the auto-imposed limitations of the deposit insurance proved insufficient to limit deposit outflows from the bank. As a result of this, some U.S. lawmakers have called for the effective removal of the $250.000 million upper bound. Additionally, it should be noted that, as happened in 2008, the deposit insurance limit has led to outflows of deposits towards Money Market Funds (MMFs), financial intermediaries included under the category of the NBFIs. Given the negative effects these institutions can have on asset prices and the overall liquidity of the system during periods of market turmoil, the possible reforms of the deposit insurance should also take into account this leakage towards less regulated and more volatile institutions.
On a more macroeconomic level, March’s banking stress also raised concerns about the interplay between monetary policy and financial stability. In an environment characterized by stickier-than-expected inflation the facilities aimed at solving the liquidity pressures seem to be at odds with the balance sheet reduction policies (i.e., quantitative tightening) and the rate increases. However, a broader and less monetarist view of the economy can lead to different conclusions about the inflationary outlook. In this sense, even if the central bank is alleviating liquidity constraints though its facilities, the overall effects on credit and, ultimately on aggregate demand, might not be correlated with the liquidity that is being pumped into the system. In here, it is important to understand that the provisions of liquidity or the purchase of bonds under the quantitative easing policies do not have to translate into new credit creation, pushing aggregate demand upwards[10]. In addition, the increase in the levels of uncertainty as a result of the banking stress might also reduce the levels of credit further, as banks and economic agents might increase their risk aversion decreasing respectively lending and borrowing. This reduction of credit might help to cool the labor market and contribute to the reduction of headline inflation in the near term. In this regard, looking at Fed’s meeting in March, the central bank argued that their decision to increase rates at a slower pace was partially explained by the expected reduction in credit creation as a result of the banking turmoil.
When talking about the “separation principle” it is important to highlight that a clear communication by the central bank might alleviate the tensions arising from the monetary policy and financial stability interactions. As shown by the actions of the Bank of England during the mini-budget turmoil, stablishing a clear distinction between the different goals of the central banks as well as a clear strategy to reverse the liquidity-related balance sheet expansions might be able to alleviate liquidity constraints without having a noticeable impact on the inflationary outlook. The situation in the United Kingdom proves that the “separation principle” is still possible.
[1] This is relevant as an environment characterized by increasing rates is positive for commercial banks as its business model benefits from higher rates. Other things equal, the net interest Iincome of commercial banks grows as a result of higher interest rates.
[2] The interest rate risk in the trading book refers to the potential losses in fixed income that may arise from rate increases.
[3] The wrong way risk appears when the likelihood of default is positively correlated with general market risk factors.
[5] The FHLBS includes 11 cooperative owned wholesale banks that act as a general source of liquidity for some member financial institutions. Even if the FHLBS was conceived to address mortgage lender’s liquidity shortfalls, successive periods of financial instability broadened the scope of the system.
[6] Basel III includes two liquidity ratios: the Liquidity Coverage Ratio (LCR) and the Net Stable Funding Ratio (NSFR). The LCR ensures that banks have enough cash to meet outflows during at least 30 days, while the NSFR requires banks to have enough stable funding to support their operations for at least 1 year.
[7] With the systemic risk exception, the Federal Deposit Insurance Corporation (FDIC) can act without any limit in order to avoid adverse effects on economic conditions and financial stability, even if the actuations do not have the least cost for the FDIC. In practical terms, the systemic risk exception allowed the FDIC to guarantee all deposits, including those above standard insurance amount of $250,000 per depositor and insured bank.
[8] NBFIs, sometimes referred as the “shadow banking” sector, are financial institutions whose activities and risks resemble those of the commercial banks, without having the same funding, backstops and regulations than these. The interconnexions of NBFIs with other parts of the financial system, (through short-term borrowing and lending operations as well as the commonality of assets), their size and the impossibility to substitute the core activities in the short-term, make them to be the perfect candidates to create and spread systemic risk. It is also important to note that the existence of NBFIs comes also with some benefits as these institutions increase the sources of funding and, in some cases, act as shock absorbers.
[9] In addition to the deposit guarantee, policymakers have also raised concerns about the necessity to reduce the asset threshold with the aim to enforce toughest regulatory requirements. Moreover, from the supervisory side, some have called for the inclusion of less benign macroeconomic scenarios in the stress-test exercises.
[10] Away from the post-Covid years and as the decades after the Great Financial Crisis have shown, the eye-popping balance sheet expansions of the major central banks have led to similar increases in the amount of excess reserves in the banking system. Since the liquidity was kept within the boundaries of the banking sector, central banks were not able to steer inflation as intended.
