Deposit Insurance

Should we expand coverage?

Jul 30, 2023

FDR signs the Banking Act of 1933, creating the FDIC

Today’s post is a particularly long one, even by my standards, with a lot of background for those new to the topic of deposit insurance. I have organized this post with caption headings so you can select which sections to read and which to skim or skip. Those of you already well versed in this subject might wish to read just the introduction and the final few sections; others may benefit from some of the earlier material before moving on to deposit insurance reform. Whatever you chose to do, I hope you will find this material interesting, informative and perhaps even provocative.

In my last post, I wrote about the risks and regulation of regional banks following the failures of SVB, Signature Bank and First Republic. I discussed some of the reasons these banks failed, highlighting among other things the large concentration of uninsured deposits at all three firms and in the regional banking sector more generally. I noted a number of potential regulatory responses to these events, including enhanced liquidity and capital requirements, more supervisory attention to interest rate risk and possible modifications to regulatory capital accounting. I acknowledged the possibility of increased M&A activity in response to these developments. And I asked whether as part of any future regulatory reforms we should also consider expanding FDIC insurance coverage as means of more directly mitigating runs at banks with large amounts of uninsured deposits.

I am not sure where I stand on the question of expanding FDIC insurance coverage, but ‘skeptical’ is probably a fair characterization. Having said that, I am not entirely satisfied with the current state of affairs either. As things now stand, FDIC deposit insurance is legally capped at $250,000 (per depositor, per account category, per bank), which covers virtually all individual and most small business deposits but only 55% of total deposits, down from 80% in 2010. Unlike other funding sources, bank depositors have not historically been paid to take risk—or more precisely they have sacrificed yield for liquidity—and so deposits (particularly demand deposits) tend to run in times of trouble, as we saw this past March. Unless of course the deposits are insured, in which case they don’t run, even at insolvent banks and in times of financial turbulence.

But do we really want to ensure all bank deposits, just because of our recent experience with SVB et al? Won’t this create even more moral hazard risk than we have now, increasing systemic risk if not in the depository banking sector than in other parts of the financial economy? Who will pay for the increased coverage and how much additional regulatory burden will be imposed on the banking industry? And if we increase deposit insurance for banks, will we need to do something similar for non-bank financial institutions, which also fund their businesses with liquid liabilities and are themselves quite capable of wreaking havoc on our financial system, as we saw during the global financial crisis?

I do not have good answers to most of these questions, but this is an interesting and important topic for banks and bank regulators, which we should all probably understand better than we do now.

Let’s explore.

What are bank ‘deposits’? Deposits are liabilities created by commercial (depository) banks in favor of their customers, which can take various forms. Customer deposits can be held by individuals, trusts, business corporations, governmental entities, financial institutions or others. Customer deposits can be held in transactional or operational accounts (including checking and business payment accounts), the purpose of which is to facilitate customer payments. Or they can be in savings or investment accounts, meant to provide not a medium of exchange but rather a store of value (with some nominal investment return). And some deposits are held in sweep accounts, which automatically transfer funds from an investment account into a transactional account as cash is needed to pay customers’ bills.

Deposits can be core deposits held by regular customers of the bank or brokered deposits provided by funding sources with no other ongoing business relationship with the bank. Some deposits pay interest (savings and investment accounts), while others do not (transactional accounts). Some deposits have fixed maturities (time deposits and CDs) and others do not (including demand deposits, which have no maturity date but which can be converted to cash by the depositor on demand). Some deposits are collateralized, or benefit from preferential payment priorities in the event of bank failure. And some deposits are insured while others are not.

How are bank deposits created? I think it is fair to say that the average man on the street does not fully understand how commercial banks work, let alone how deposits come into existence. The popular understanding among folks of a certain age seems to be that bank deposits are created when bank customers hand the teller a bag full of coins and dollar bills to put into their checking or savings account for safekeeping. The bank then takes this cash and lends most of it out to other customers of the bank, keeping a small amount in the vault to meet daily liquidity needs while hoping and praying that depositors won’t show up en masse to withdraw their cash before the bank’s borrowers have paid back their loans and refilled the bank’s coffers.

This is more or less how George Bailey explained banking in the film It’s a Wonderful Life, as customers of the Bailey Building & Loan lined up at the teller window to withdraw their cash from the bank in a classic 1930’s bank run: “You are thinking about this place [the BB&L] all wrong, as if I had the money in a safe. The money’s not here. Why your money’s in Joe’s house, that’s right next to yours, and in the Kennedy house and in Mrs. Maitland’s and in a hundred others.” This Hollywood explanation is actually pretty good as far as it goes, but it is far from the whole story even if we update it for the advent of digital banking. When the Bailey Building & Loan lent Mrs. Maitland the money to buy (or build) her house, it didn’t hand over to her the cash that Tom and Joe and the others had put on deposit. Instead it simply credited Mrs. Maitland’s bank account with newly created deposits in the amount of the loan, which she then used to pay for her house by writing checks against her account at the bank. The amount of the deposits credited to the accounts of Mrs. Maitland and other borrowers far exceeded the amount of cash deposits that the BB&L had collected from its customers, let alone what it kept in the vault (even before Uncle Billy turned over all the vault cash to the cross-town bank). Which of course is how all depository-lending banks work.

Deposits and fractional reserve banking. We call this economic model fractional reserve banking, which to many people seems like a form of alchemy if not black magic. But however mysterious this model may seem, this is in fact how commercial banking works. Banks lend out and invest multiples of the amount of money they hold as reserves, and the banks’ reserve ratio (reserves/loans) determines the amount of credit banks can create from any given quantity of reserves. Most bank deposits are created not by customers showing up at the teller with bags of cash, but rather as a byproduct of the aggregate lending (and investment) activities of individual banks conducted in the ordinary course of business. When a commercial bank makes a loan, the size of its balance sheet does not stay constant with offsetting movements in asset accounts (eg loans up, cash down). Instead, the size of the bank’s balance sheet increases by the amount of the loan, with a corresponding increase in both assets (loans) and liabilities (deposits). (For the customer, of course, the labels are reversed: The loan is a liability and the deposit is an asset.) As the bank makes more loans, both its loan book and its deposit base increase by corresponding amounts. The more deposits a bank takes in (or creates), the more money it can lend out or invest. Which generally works well, until it doesn’t.

As one might expect, this fractional reserve banking model is inherently unstable insofar as demand deposits are prone to run from time to time, as we saw with the fictional BB&L and in real life during the 1930s, the 1980s, 2008-9 and again last March. And while some people believe we should put an end to bank runs by adopting a full-reserve banking model, most of the experts do not. (Read here and here.)

Deposits are both a funding source and a product of the bank. We often refer to deposits as a ‘funding source’ for bank assets, which is true. Most banks fund their loans primarily with deposits, which may constitute 90% of a bank’s total liabilities and 80% or so of its total assets. In commercial terms, however, what really happens is that in the process of making loans (and investments) banks create deposits, as the BB&L did with Mrs. Maitland. And while we often say that commercial banks ‘take deposits and make loans’, we should perhaps rephrase this and say that banks ‘make loans and create deposits’. Deposits are not simply ‘funding sources’ for the bank; they are also ‘products’ of the bank. And the type of deposits which individual banks create and hold has a lot to do with the type of customers they choose to serve and the particular services they provide.

Where did the deposits go? In the weeks and months following the collapse of SVB and Signature Bank, the banking industry reported a significant migration of deposits out of regional banks and into the large money center banks, in what was described by the press as a depositor flight to safety. And as the Fed has raised interest rates over the past 16 months or so, we have also seen a large movement of funds out of low-rate bank deposits and into higher yielding bank and non-bank products, most notably money market funds. But where did all these funds go?

When customers pull their deposits from one bank and put them on deposit at another bank, the aggregate amount of deposits in the banking system does not change. Deposits go down at Bank A and deposits go up at Bank B, as one would expect. And this is also what happens when bank customers transfer funds from their bank deposit accounts into money market funds held at a brokerage firm, at least initially. (For a good description of how this works, read this article about Charles Schwab which runs both a brokerage firm and an affiliated commercial bank, with large customer fund flows between the two.) In this case, deposits go down at Bank A and deposits go up at Bank C (where the MMF holds its customer deposits pending investment). When the MMF subsequently invests these new customer deposits by purchasing money market securities, deposits go down at Bank C and deposits go up at Banks D, E and F (the banks which hold as deposits the proceeds received from the sale of securities to the MMF). In all these examples, deposit funds simply moved from one bank to another, with no change in aggregate deposits in the banking system.

Yes, there are exceptions to this. There are always exceptions. For example, if the MMF uses the newly received customer funds not to buy bank CDs or corporate commercial paper from the issuer or a secondary market seller (in which case deposits would just move between various bank accounts), but rather to buy T-bills direct from the US Treasury, the funds would flow out of the MMF’s bank account and into the US Treasury’s account at the Fed, reducing (at least initially) the amount of deposits held in the banking system. Or if the MMF lends the funds to the Fed in a reverse repo transaction, deposits would again leave the banking system and end up at the Fed. And something similar happens when the Fed sells securities from its own portfolio to a non-depository bank or broker dealer; in this case the funds move from the broker-dealer’s bank account to the Fed. In all these examples, aggregate deposits in the banking system would decline (and would increase if we reversed the transactions), but this is not usually the case.

As a general matter, deposits regularly move between banks but they do not usually leave the banking system entirely, at least not in large size or for long. [For more on this topic, read Steven Kelly’s explanation in Without Warning.]

Aggregate Deposits. As noted, the movement of deposits between banks is something very different from the creation and destruction of aggregate deposits in the banking system. But there are times when aggregate deposits in the banking system do expand and contract, sometimes quickly and in large amounts. In the aftermath of the global financial crisis, aggregate bank deposits more than doubled, from $6 trillion (2007) to $13 trillion (2019). During Covid, bank deposits again increased by 50% or so, from $13 trillion (2019) to $18 trillion (2022). (See the FRED data here.)

Much of this change in the aggregate amount of bank deposits resulted from the Fed’s expansionary monetary policy in response to the GFC and Covid, when the Fed’s balance sheet increased by $8 trillion and commercial banks eventually used the newly created reserves to increase outstanding credit (loans and investments) and hence deposits. (See FRED data here and here.) During Covid, other factors also contributed to the increase in bank deposits, including large customer draws on bank lines of credit (particularly early on in the pandemic), federal fiscal stimulus (think of all those ‘stimmy’ checks that hit bank accounts) and increased household saving (reduced spending) during lockdown.

If you are wondering why any of this is important—this matter of how bank deposits are created and destroyed and what causes changes in the aggregate amount of bank deposits—it is because in our financial system commercial bank deposits are the largest component of our money supply, and large changes in the money supply can have very significant impacts on real economic activity, which in turn impacts the lives and livelihoods of real human beings, not just bankers and economists. And while the financial health and stability of individual banks is important, what is even more important is the health of the overall banking system and with it the stability of our money supply and conditions in the real economy. Bank runs matter because deposits are a form of money, which drives real economic activity.

Deposits as money. The term ‘money’ means different things to different people. Some people think of money simply as the currency in their wallets. Others would include in the definition of money the funds immediately available to spend from their bank checking accounts, and possibly also their savings and money market account balances, which can be easily converted to spendable funds. And while some might include the current market value of longer-term investment securities in their brokerage accounts, or even in their various retirement accounts, these investment funds have more to have more to do with personal wealth and household savings than with money per se.

To the economist, the term ‘money’ has a more precise definition which highlights the role of various financial assets as a medium of exchange as well as a store of value. Economic money has to be ‘liquid’, and liquidity entails not just marketability (the ability to convert an asset quickly to cash at some price) but also the stability or certainty of the market value (price) of the asset. Currency in hand is entirely liquid as is the cash in a demand account at a solvent bank. The funds held in a savings account or a money market fund may be almost as liquid as demand deposits, depending on the the terms of withdrawal and the nature of any underlying investments. The funds held in all of these deposit or deposit-like accounts are considered by economists to be 'money’ of some sort. But this is not the case with a bond or equity investment fund, which may be marketable but is not liquid, because of its price (value) volatility. And neither of course are crypto assets, including my favorite crypto oxymoron, so-called ‘stablecoins’.

Economists and central bankers have various ways to measure the aggregate money supply (or money stock)—M0, M1, M2, MZM, etc—but for our purposes we can think of the money supply as having several components: currency in circulation, central bank reserves, transactional deposits held at commercial banks, and certain other bank deposits and liquid financial instruments (eg retail MMF accounts). [For the latest Fed money stock data, see here.]

What are ‘insured’ and ‘uninsured’ deposits? Before we delve further into the FDIC deposit insurance program, we should first clear up one occasional misunderstanding regarding the terms ‘insured’ and ‘uninsured’ deposits. Deposit insurance is not like life insurance, where the insured policyholder chooses how large a death benefit (or annuity) to procure. Or like auto insurance, where the insured can decide to buy liability insurance but not collision or comprehensive coverage, or coverage on one car but not another. This is not at all how deposit insurance works.

Individual banks decide whether or not to join the FDIC insurance program, but once they have been accepted by the FDIC as Insured Depository Institutions all of their deposits are automatically insured. But the deposits are insured by the FDIC only up to the legal coverage cap, currently set at $250,000 per depositor, per account category, per bank (IDI). Deposit balances below the cap are considered to be ‘insured’ and deposit balances above the cap are considered to be ‘uninsured’.

As an example, let’s look at Silicon Valley Bank, a large but niche bank which at the time it failed had ca $150 billion of uninsured deposits, representing 90% of its total deposits and 80% of its total liabilities. This exceptionally high concentration of uninsured deposits did not happen because SVB’s management or board of directors chose not to purchase deposit insurance, as explained above. Rather, SVB’s high concentration of uninsured deposits was a direct consequence of the fact that SVB was primarily a bank for business customers, in this case VC firms and technology companies which periodically held large cash balances on deposit at the bank. At the time it failed, SVB’s top ten customers held an average of over $1 billion each on deposit at the bank, and its largest customer, Circle Internet —ironically, the sponsor of the USDC ‘stablecoin’—had over $3 billion on deposit at SVB (and large amounts at other banks as well). These deposits were all uninsured, because of their size. And it appears that a large portion consisted of uncollateralized demand deposits, which were highly liquid and in the end prone to run.

A brief history of US deposit insurance. Before we discuss the future of FDIC deposit insurance, perhaps it would be a good idea to quickly recap some of the relevant history, if only to understand how we got to where we are today. The history of deposit insurance in the United States is popularly thought to begin with FDR’s signature of the Banking Act of 1933, the so-called Glass Steagall Act which separated commercial banking from investment banking and also created the FDIC. (See photo image above.) But this is not quite right. The US had a long and checkered history with state and industry-sponsored deposit insurance programs well before 1933, which goes a long way to explaining how and why the FDIC came into existence.

The first US deposit insurance program was launched by the state of New York in 1829, with five other states implementing similar programs from 1831-1858. These insurance programs covered two different forms of bank money: deposits and bank notes (an historical relic), generally without limitation as to amount. Bank supervision was an essential element of all these programs, but operated somewhat differently in each program. Insurance payouts were funded (in advance or retroactively) through assessments on the capital stock or insured obligations of participating banks, without state guarantees. All of these programs eventually failed, in large part due to the emergence of free banking following the 1836 closing of the Second Bank of the United States.

During the Civil War, Congress created the national banking system with the passage of the National Banking Acts of 1863 and 1864. In the early years of national banking, the primary form of bank credit obligations was bank notes rather than deposits. The holders of bank notes did not require insurance, because by law the notes had to be collateralized by US Treasury obligations and also had a backstop guarantee provided directly by the US government, which was unlimited in amount. Within a few years, however, deposits supplanted bank notes as the primary form of circulating bank money. Deposits were not collateralized or guaranteed by the government, unlike bank notes, creating renewed pressure for federal government insurance. Between 1866 and 1933, Congress considered over 150 different proposals for some form of federal deposit insurance, generally in the wake of some financial crisis, most notably the Panic of 1907. But none of these bills were enacted into law.

In the absence of federal legislation, a number of states adopted their own deposit guarantee programs. From 1908-1917, eight states (primarily western agricultural states) adopted deposit guarantee programs of various sorts, funded by bank assessments. In none of these programs did the state itself guarantee bank obligations. None of these programs survived the economic instability of the 1920s and all were closed by the end of the decade.

During the 1920s, an average of more than 600 banks per year failed, which was 10 times the rate of failure during the preceding decade. These bank failures primarily involved small rural banks, many of which were poorly managed, and were not widely considered to be of major public policy concern. In 1930, however, things took a marked turn for the worse when 1,350 banks failed, including a number of large banks, resulting in big depositor losses and a widespread panic. Over the next three years, over 9,000 additional banks failed, culminating in the famous (some would say infamous) ‘banking holiday’ declared during the initial days of FDR’s first term. (Listen here to FDR’s first fireside chat. FDR may not have been George Bailey, but he was pretty good on the subject of banking.)

The FDIC was created by the Banking Act of 1933 and the first federal deposit insurance program went live in 1934. Coverage was initially capped at $2,500, which was quickly raised to $5,000, equivalent to over $100,000 in today’s dollars. The deposit insurance provisions of this legislation were very controversial at the time, opposed not only by the banking industry but also by FDR and even by the co-sponsor of the legislation, Senator Carter Glass of Virginia. Opponents had a number of objections. Many thought deposit insurance simply would not work, citing past experience with the state programs. Others had concerns about moral hazard risk. Some questioned the cost of the program and who would pay for it. And of course some opponents saw in deposit insurance one more example of ‘creeping socialism’ under FDR. The public strongly supported deposit insurance, however, and in the end this public support carried the day. Congress passed the Banking Act with deposit insurance intact and FDR signed it into law in June of 1933.

During the first full year of deposit insurance, the number of failed banks fell dramatically, with only nine insured banks and fifty two uninsured banks being closed in 1934. The amount of deposits in the banking system increased over 20%, the quality of bank assets increased markedly, and by the end of the year over 98% of all deposits were held in insured institutions. During the balance of the Great Depression, a total of 370 more banks failed, an average of ca 50 per year, most of which were small banks. Bank failures became more rare with the advent of WW2 and the end of the Great Depression. Twenty insured banks failed in 1942 and fewer than ten failed in each of the subsequent 32 years, an unprecedented period of banking system stability widely attributed in large part to the establishment of FDIC deposit insurance.

During the 1980s and early 1990s, over 2,900 US banks and savings institutions failed for reasons I won’t go into but which you can read about here. But with one significant exception, Continental Illinois Bank in 1984, runs on deposits played very little role in this crisis. The FDIC was forced to guarantee the uninsured deposits at Continental, and the FDIC and FSLIC had to be recapitalized as a result of the losses incurred paying off insured deposits at the failed banks, but deposit insurance seems to have done its job mitigating bank runs.

Over the decades, the FDIC coverage cap has been periodically increased, from $5,000 in 1934 (the equivalent of over $100,000 today) to $250,000 in 2008 (made permanent in 2010), which is where it stands today. As noted below, this is a much higher cap than exists in other countries.

What is the purpose of deposit insurance? Deposit insurance is generally considered to have two overarching objectives: to protect consumers (small bank depositors) and to enhance the stability of the depository banking system by mitigating bank runs. Capping insurance coverage at $250,000 makes sense from a consumer protection perspective, and it may even be considered generous (by international comparisons at least), with the vast majority of individual and many small business deposit accounts fully insured at this level. But FDIC insurance currently covers only 55% of the dollar amount of all outstanding deposits, with most business deposits largely uninsured, which may or may not be the optimal amount of coverage for the purpose of mitigating systemic run risk. And so we should expect to see some tension between these two objectives—consumer protection and financial stability— in the public debate over any future proposals to expand FDIC insurance coverage.

Has FDIC insurance been effective? Yes, remarkably so, at least if we define ‘effective’ as reducing the occurrence (frequency) of runs at insured institutions. As discussed above, the number of bank failures dropped significantly following the establishment of deposit insurance in 1934, and more to the point deposit runs on insured institutions ceased almost entirely from that point on. During the US banking (and S&L) crisis of 1980-95, we experienced the failure of over 2,900 banks and savings institutions with over $2.2 trillion of assets, but with one major exception (Continental Illinois in 1984) there were few if any significant runs on the deposits of insured institutions. And while there were deposit runs on some commercial banks during the GFC, most notably those with large sub-prime lending operations, this paled in comparison to the wholesale funding runs on a wide variety of other non-depository financial institutions.

This does not mean, of course, that that the benefits of deposit insurance have been without cost. FDIC deposit insurance is funded by assessments on the banking industry, the cost of which has been borne by bank customers, lenders and/or shareholders. And it may well be that the moral hazard costs associated with deposit insurance have also been large, increasing the amount of embedded risk in our depository and non-depository financial institutions.

Do other countries have deposit insurance? Yes. All major economies have deposit insurance schemes, broadly similar to the FDIC program but with generally lower coverage amounts (eg EUR 100,000 in the EU). Most of these programs are funded by assessments on the banking industry, but not all of them have government guarantees as in the USA. Deposit programs across the world seem to be targeted primarily to protecting retail (individual) deposits rather than business deposits. In the US, FDIC insurance protects all deposits, including business deposits, which is not always the case in other countries. In the event that the US decides to raise the FDIC coverage cap, other countries may feel pressure to do something similar, as we saw in the UK when Ireland (and much of the EU) raised its coverage limits following the global financial crisis. Foreign nationals with funds deposited in US banks are entitled to the same FDIC coverage as US citizens and residents.

The FDIC Deposit Insurance Fund. The FDIC is the insurer, regulator and supervisor of Insured Depository Institutions (IDIs) and is charged with resolving failed IDIs. The FDIC also manages the Deposit Insurance Fund (DIF), the assets of which are used to fund the FDIC’s operations and pay off insured deposits at failed banks. The DIF is funded by assessments (insurance premiums) paid by the IDIs, and its obligations are guaranteed by the full faith and credit of the United States government. In addition to the premiums (assessments) charged to member banks, the FDIC also generates revenue from the interest earned on funds invested in U.S. government obligations. Revenues from assessments and investment income increase the DIF balance, while operating expenses and losses (primarily from bank failures) reduce the balance.

The DIF is funded by assessments on over 4,700 insured depository institutions. The amount of assessment (premium) charged to individual banks is calculated with two components: an Assessment Base and an Assessment Rate. Since the passage of Dodd Frank in 2010, the FDIC’s Assessment Base is determined by banks’ total liabilities (total assets less tangible equity) rather than the amount of total deposits, as had previously been the case. (Perhaps surprisingly, the FDIC’s Assessment Base is not and never has been based on the amount of banks’ insured deposits.) The change in calculating the Assessment Base from total deposits to total liabilities was intended by Congress to shift some of the FDIC funding burden from smaller to larger banks, which historically relied more on non-deposit funding sources, on which it did not p[ay FDIC assessments. This is in fact what has happened, with the Assessment Base now more equally allocated among banks in proportion to their total assets.

The Assessment Rates which the FDIC uses to set premiums for individual banks are risk-based, using the FDIC’s supervisory CAMELS ratings, with somewhat different methodologies applied to small, large and complex banks. This is a pretty esoteric subject, which I don’t really understand, but one has to wonder what if any changes to the FDIC’s risk assessment methodology may be made in the wake of the failures of SVB et al, which in hindsight may be viewed as having taken on a lot more risk than was understood at the time their deposit insurance Assessment Rates were calculated.

At the end of Q1 2023, the DIF balance stood at $116 billion, down $16 billion or so from the prior quarter. This represents a Reserve Ratio of ca 1.1%, which is below the legal minimum Designated Reserve Ratio (DRR) of 1.35% and far below the FDIC’s long-term DRR policy goal of 2%. (The Reserve Ratio equals the value of DIF assets divided by the total amount of insured deposits.) The obligations of the DIF are guaranteed by the full faith and credit of the US government, and I would not be surprised to see some pressure on Congress to improve the funding status of the FDIC, with or without further changes to the amount of insurance coverage.

SVB and Signature. For most of modern US financial history, deposit insurance has been remarkably successful in reducing the risk of bank runs, as noted several times above. And during this time it has been generally understood that when a commercial bank (or savings bank) failed, the FDIC (or FSLIC) would guarantee deposits only up to the coverage cap, leaving the balance of deposits unpaid pending distribution of liquidation proceeds to the bank’s various creditors. But when the $200 billion (asset) Silicon Valley Bank failed in mid-March, and the $100 billion Signature Bank failed a few days later, the FDIC did something quite extraordinary. With Fed and Treasury Department approval, and after consultation with the President, the FDIC exercised its Systemic Risk exception (SRE) authority under the FDIC Improvement Act of 1991 and paid off immediately and in full all of SVB and Signature’s deposits, 90% of which were uninsured, leaving the FDIC with an estimated loss $22+ billion. (Read here, here and here.)

The FDIC’s decision to protect the uninsured depositors of SVB and Signature was unexpected but not unprecedented. During the 2008-9 global financial crisis, the FDIC invoked its SRE authority to guarantee certain obligations of several too big to fail banks (Cit and BofA) and preempted an industry-wide funding run by issuing a limited guarantee of future bank debt and uninsured transactional bank deposits. (To learn more about the FDIC’s Temporary Liquidity Guarantee Program, read here.) And when Continental Illinois Bank failed in 1984, the FDIC guaranteed both insured and uninsured deposits, using other legal authority available at the time. But the FDIC generally took a very different approach during the US banking crisis of 1980-95, when more than 2,900 US banks and thrifts failed, with collective assets of over $2.2 trillion. With the exception of Continental Illinois, the FDIC paid off deposits only up to the coverage cap and issued liquidation certificates to the holders of unpaid deposit balances, forcing uninsured depositors to take their place in the liquidation queue along with the other bank creditors (and the FDIC).

Which of course is exactly what many people think the FDIC should have done with both SVB and Signature.

Did SVB fail because it ‘relied’ on uninsured deposits? There is rarely just one reason that a bank or other financial institution fails, and this is true of SVB as well as Signature Bank and First Republic. In the case of SVB, the proximate cause of the bank’s failure was a massive run by a concentrated group of bank customers, communicating amongst themselves on social media, who in just over 24 hours withdraw (or attempted to withdraw) uninsured deposits representing close to 80% of the bank’s total funding. No bank could survive a run of this magnitude and timing, and certainly not a bank with as many other risk factors as SVB.

But is also true that SVB would likely not have failed if the bank had not made a $120 billion bet on long-duration government bonds, if the Fed had not raised rates as high and as quickly as it did, if SVB’s customer base had been more diversified and less lemming-like in its behavior, or if SVB’s shareholders had not reacted so badly to the bank’s attempted capital increase. Or for that matter if SVB had simply made sure that its emergency lines of credit were working properly at the Fed and the FHLB. But for any one of these factors, including of course its high concentration of uninsured deposits, SVB might still well be with us today, nursing a low share price and finding itself a potential takeover target, but nevertheless solvent and still serving its customers. Which by most accounts it did quite well, until the very end.

There is no doubt that SVB’s large concentration of uninsured deposits was a major contributor to its failure, perhaps the main contributor, but it is perhaps misleading to say that SVB failed because of its ‘reliance’ on uninsured deposits. SVB didn’t so much ‘rely’ on uninsured deposits as it chose to serve a particular customer set with certain banking needs, including the management of periodically large amounts of cash deposits, which because of their size were mostly uninsured by the FDIC. At the time it failed, SVB’s ten largest customers had on deposit at the bank an average of over $1 billion each, and its largest depositor, Circle Internet, had over $3 billion on deposit, all of which was uninsured. These were customers (and deposits) that many other banks would like to have had, but they chose instead to bank with SVB.

This relationship between a bank’s customer base and its funding strategy is sometimes misunderstood or mischaracterized, as demonstrated in this combative exchange between Representative David Scott (D-GA) and former SVB CEO Greg Becker during a recent House Financial Services hearing:

Rep. Scott: “Who made the decision to maintain this reliance on uninsured deposits—given the warnings also by our federal regulators? Who made this decision, Mr. Becker? This foolish, irresponsible, and deceitful decision—who made it?”
SVB’s Becker: “Congressman, as I said, that’s been our business model for as long as I can remember-”
Rep. Scott (interrupting): “Who made the decision, my friend?! Was it you?!”

No doubt Representative Scott was looking for a simple answer to a complex problem, and ideally one that would make headlines or garner him a spot on the evening news. But to answer Representative Scott’s question—”Who made this decision [to rely on uninsured deposits]….this foolish, irresponsible, and deceitful decision?—Mr. Becker might rightly (if provocatively) have said: You did, Representative Scott. You and your colleagues in Congress. The decision not to insure SVB’s deposits was made not by SVB’s management or its board but by the members of Congress, who in 2008 and 2010 passed legislation increasing FDIC insurance coverage to $250,000 (up from $100,000 previously), leaving the vast majority of deposits at SVB and other business banks uninsured.

Who should pay for the FDIC’s bailout of uninsured depositors at SVB and Signature? The FDIC’s decision to pay off the uninsured depositors of SVB and Signature Bank will cost it about $16 billion (revised estimate), which it is required by law to collect through a special assessment on the banking industry. The FDIC announced the terms of its proposed special assessment in May and the public comment period has just ended. The FDIC has proposed to base its special assessment on the amount of uninsured deposits at individual banks as of December 31, 2022. The assessment will apply only to uninsured deposits in excess of $5 billion, which the FDIC believes will exempt all but 113 of the 4,700 IDIs from paying any special assessments. The vast majority of the $16 billion special assessment will be paid by the largest banks, particularly those in the top 10% (total assets) which have the largest amounts of uninsured deposits, with approximately 60% of the total charged to a handful of banks designated as Global Systemically Important Banks: JP Morgan, BofA, Citi, Wells Fargo, BNY Mellon and State Street.

In setting the SRE special assessment, the FDIC is required by law to consider the types of entities that benefitted from its action, the effects on the industry, and such other factors as the FDIC deems appropriate and relevant to the action taken or the assistance provided. The public comment period for the FDIC’s special assessment proposal has now ended and the range of views expressed is quite interesting. I won’t recap all the comments here, but suffice it to say that there is great disagreement within the banking industry over which firms did and did not benefit from the FDIC’s bailout of the uninsured deposits at SVB and Signature and therefore who should pick up the tab. The GSIBs think they did not need or benefit from the FDIC’s intervention and are being overcharged as a result. The large custody banks feel the same way, and observe that not all uninsured deposits are equally risky (including many of their own) and so should not all be assessed equally. And mid-sized banks think that perhaps the uninsured deposit exemption amount should be increased $10 billion, letting them off the hook for making contributions, presumably on the theory that the FDIC would never had done for these banks what it did for the larger SVB and Signature. [For more on the public comments, read here,]

As an interesting post-script, you may have seen recent press reports about the FDICs cautionary letter to banks, reminding them of their obligation to make accurate call report disclosures. The FDIC’s proposed special assessment to recoup the cost of the SVB and Signature bailouts, announced in May, was based on banks’ December 2022 call report disclosures, which contained the banks’ estimates of their own insured and uninsured deposits at that point in time. In the wake of the FDIC’s special assessment, however, it seems that a number of banks have decided that perhaps their December 2022 call reports were not accurate, with overstated estimates of their own uninsured deposits at that time. And as a result these banks believe that their share of the FDIC special assessment should also be reduced. As expected, this is not getting a very warm reception from the FDIC.

What might deposit insurance reform look like? On May 1, the FDIC helpfully released this report to Congress, Options for Deposit Insurance Reform. This is a very interesting document and I encourage you all to read or at least skim it. The report begins by recounting the historical success of deposit insurance in enhancing the stability of our commercial banking system, essentially eliminating bank runs for much of our recent history. The FDIC goes on, however, to note the large increase in the aggregate amount of uninsured deposits in recent years—by my calculations an increase of close to $7 trillion, up 7x from 2010—and to suggest that deposit insurance as currently structured may prove to be a less potent systemic risk mitigator going forward than it has been in the past. And because most of this increase in uninsured deposits has occurred in the largest banks, those in the top 10% and particularly the top 1% of assets, future deposit runs may be much larger and more consequential than in the past.

The FDIC report outlines three broad approaches to potential deposit insurance reform, all of which would increase the current $250,000 coverage cap. Limited Coverage would maintain the current structure of deposit insurance, with a fixed coverage cap that applies across depositors and types of accounts, but would increase the current cap to some larger amount. Unlimited Coverage would provide an unlimited amount of insurance for all deposits at insured depository institutions. Targeted Coverage would allow for different levels of deposit insurance coverage across different types of accounts, with larger (perhaps unlimited) coverage for operational or transactional business accounts.

The FDIC report does not make a clear recommendation on what form deposit insurance reform should take, leaving this for Congress to debate and decide. But among the alternatives, the FDIC does express a pretty clear preference for some form of Targeted Coverage, which it believes will capture many of the financial stability benefits of expanded Limited Coverage while mitigating many of the undesirable consequences of Unlimited Coverage The biggest problem with Targeted Coverage, as identified by the FDIC, is defining which specific accounts would be eligible for expanded coverage (it suggests ‘business payment accounts’) and limiting the ability of banks and depositors to game the system. The FDIC acknowledges that expanding Targeted Coverage would require a significant increase in bank assessments, but not to the same extent as Unlimited Coverage.

In what other ways might we mitigate the risk associated with uninsured deposits? The problem with uninsured deposits is that they are prone to run when banks gets into financial difficulty. And so one approach to mitigating run risk is simply to focus on making banks less likely to get into financial difficulty, for example by tightening up on the regulation and supervision of bank risk activities, capital adequacy and/or liquidity requirements, all of which are currently under consideration by the regulators. The problem with increased regulation, of course, is that it also tends to reduce bank profitability and potentially bank lending (and investment) activity, which may be sub-optimal from a macro perspective. And while runs are most common at insolvent banks with uninsured deposits or other liquid funding sources, a bank does not have to be insolvent in order to experience a debilitating run, as we saw at SVB. Absent the run, SVB would have remained solvent under both GAAP and regulatory capital accounting standards and with plenty of liquidity to boot.

One could also address the run risk associated with uninsured deposits by requiring vulnerable banks to increase their usage of other forms of less liquid funding, like long-term bonds and equity. As we have discussed in other posts, ‘loss absorbing capital’ plays an important role in capital adequacy regulation and can also help reduce liquidity risk. And so we might require banks with large concentrations of uninsured deposits to hold additional amounts of equity and/or long-term debt, providing an additional buffer against loss and thereby reducing run risk. More directly, we could also require banks to collateralize some portion of their uninsured deposits with Treasury securities or other types of high grade collateral, as is done with deposits held by governmental customers (eg municipalities).

An alternate approach would be to expand the lender of last resort facilities provided by the Fed or others (eg the FHLB). In the wake of SVB’s collapse, the Fed created a new Bank Term Funding Program, which enables banks to borrow for up to one year against UST and agency collateral at no discount to face value. This new facility relieved the liquidity pressure on banks which (like SVB) were sitting on large unrealized losses in their investment portfolios. Had the BTFP been put into place earlier, it is quite possible that SVB would not have failed, although the BTFP did not in the end save First Republic, whose unrealized losses were primarily in its loan book rather than its investment portfolio.

Should depositors be responsible for monitoring bank risk? Many people object to deposit insurance coverage on the grounds of moral hazard, believing that insurance coverage relieves depositors of their (moral?) responsibility for monitoring bank risk and thereby encourages banks to take more risk than is optimal. Moral hazard is a very real concern, which we should not dismiss lightly. But one has to ask whether depositors are in fact the parties best suited to monitor bank risk and to discipline those banks which take on too much risk. No one realistically expects individual or small business depositors to play this role, as they are not well suited to the task and given their small deposit amounts have little financial incentive to do so. But many people believe that large depositors are both well suited and sufficiently motivated financially to undertake this responsibility. In the wake of SVB, Signature and First Republic, however, I think we have to question this belief.

At SVB, depositors had $160 billion at risk in the bank, almost all of it uninsured, and the top ten depositors had average deposit balances of over $1 billion each. Many of these folks were quite sophisticated financially (or owned by firms which were). And yet SVB’s depositors appear to have paid essentially no attention to the bank’s growing solvency and liquidity risk until the very end, when they all freaked out simultaneously and tried to withdraw 80% of the bank’s total funding, $140 billion or so, in a matter of just over 24 hours. Which is not what I think most people have in mind when they task depositors with the responsibility for monitoring bank risk.

Why insure only deposits and not other forms of bank and/or non-bank funding? This is a very good question, which is why I asked it. The answer most commonly given is that commercial bank deposits, and particularly demand deposits, play a unique role in our financial system. As discussed above, bank deposits comprise a large share of our money stock, and changes in the money supply impact the amount of activity in the real economy. During periods of financial instability, or in the face of deposit runs, banks tend to pull back on lending and investment activity which reduces the aggregate amount of deposits in the banking system and hence the amount of money circulating in our economy. Which in turn impacts activity in the real economy, meaning jobs and income and the cost of living. And this is why we insure, and regulate, depository banks.

I am not entirely satisfied with this answer, but its the best I’ve got for now. As we all know, non-bank financial institutions also fund themselves with liquid (runnable) funding sources, as well as large amounts of financial leverage, and the failure of these firms can also wreak havoc in our financial system, causing even solvent and liquid commercial banks to cut back on lending and reduce deposits (and money). We saw this in spades during the GFC, which was primarily a global run on the liquid funding sources of non-bank financial institutions. When Lehman Brothers failed, it was leveraged something like 40x (2.5% equity/assets), and its $700 billion balance sheet was funded with $250 billion or so of overnight funding, which dried up completely even though a large portion consisted of collateralized repo. And something very similar happened at Fannie & Freddie, AIG and the money market funds. Yes, there was a traditional deposit run at WaMu, IndyMac and Wachovia, but the biggest problems were generally outside the depository banking system.

No one thinks we should insure the short-term debt obligations of most non-depository financial institutions, which by the way are generally much better capitalized and more liquid today than they were fifteen years ago. But if we increase deposit insurance coverage for commercial banks there will be a certain if unpredictable impact on the amount, cost and stability of funding for the shadow banks as well. Which may or may not be a good thing for the overall stability of our financial system.

For more on this subject, I encourage you to read this Ezra Klein interview with Morgan Ricks, author of the excellent book, The Money Problem, which I also recommend.

Will deposit reform actually happen? I really have no idea, but it does not appear that reforming the federal deposit insurance program has gained much traction in Congress. (Read here.) Which is understandable now that the contagion risk from SVB has dissipated, the federal debt ceiling can has been kicked down the road for another year, the economy seems headed for a soft landing and as we are entering a presidential election year with the public’s attention focused on shall we say ‘other matters’.

The creation of FDIC deposit insurance in 1933 only became a reality because were in the middle of the Great Depression, with thousands of banks having already failed, unemployment at 25% or higher, millions of homes and farms foreclosed and much of the US economy in ruins. And even then, the passage of deposit insurance legislation faced a steep uphill political battle, which in the end was won only because the public insisted on it and the newly elected President got on board, overcoming entrenched political and industry opposition along the way.

Fortunately, this is not the economic situation we are facing today. But the risk of runs on banks with large amounts of uninsured deposits has not gone away, and the FDIC’s emergency action at SVB and Signature has left everyone confused about which bank deposits are and are not effectively government guaranteed. Moral hazard seems to be running amok, the Fed continues to raise rates and regulators are tightening the screws on our biggest and most successful banks, which during the SVB fallout were perceived as safe havens in a brewing storm.

But one thing remains crystal clear. Our elected representatives in Congress, like Representative Scott, are very concerned about bank risk management failures and they are on the case, looking for someone to blame.

Someone else, that is.

Links

FDIC Guide to Deposit Insurance

Deposit Insurance FAQs, FDIC

A Brief History of Deposit Insurance in the United States, FDIC, 1998

FDIC Annual Report, 2022

How Does Deposit Insurance Work, Lee and Wessel, Brookings, March 21, 2023

The Regulation of Private Money, Gorton, NBER, May 2019

Ezra Klein Interviews Morgan Ricks, NY Times, March 24, 2023

Fed to Consider Tougher Rules for Midsize Banks After SVB, Signature Failures, WSJ, March 14. 2023

Congress Cools on Expanding Deposit Insurance, Politico, April 26, 2023

Options for Deposit Insurance Reform, FDIC, May 1, 2023