Silicon Valley Bank and the 2023 Regional Banking Stress

Key Takeaways

• The decisive number was in a public filing two weeks early. SVB Financial Group's 10-K for 2022 reported $15.16 billion of gross unrealized losses on held-to-maturity securities against $16.00 billion of SVBFG stockholders' equity, so a loss it was not required to recognize in earnings or capital was worth about 95 percent of the equity it did report.
• The funding sat almost entirely outside the insurance safety net. The same filing put estimated uninsured deposits in United States offices at $151.5 billion against $173.1 billion of total deposits, plus $13.9 billion of foreign deposits under no United States insurance regime at all.
• The securities book was extraordinarily long. Of $91.33 billion of held-to-maturity securities at carrying value, $86.04 billion matured after ten years, about 94 percent of the book in the bucket least able to become cash without a loss.
• The run was measured in hours. The Federal Reserve records over $40 billion of outflows on 9 March 2023 with $100 billion more expected the following day, against a bank the FDIC sized at roughly $209.0 billion of assets.
• Emergency borrowing went from routine to record in one week. Discount window primary credit rose from $4.58 billion on 8 March 2023 to $152.85 billion on 15 March, and the new Bank Term Funding Program lent $11.94 billion in its first days.
• Money moved within the banking system rather than leaving it. In the week to 15 March, deposits at small domestically chartered commercial banks fell $149.3 billion while deposits at large domestically chartered banks rose $53.5 billion.
• Equity investors barely registered it. The S&P 500 fell about 4.8 percent from 6 to 13 March and finished the month higher than it started, while the two-year Treasury yield fell 57 basis points on 13 March alone, its largest one-day decline since October 1987.

What Happened to Silicon Valley Bank in March 2023?

Silicon Valley Bank was the banking subsidiary of SVB Financial Group, a Delaware holding company listed on Nasdaq as SIVB. It banked the venture capital economy: startups, growth companies and the funds that backed them. That gave it a deposit base which was large, corporate, mostly uninsured and unusually well connected, and every one of those adjectives mattered in the week it failed.

On the evening of Wednesday 8 March 2023 the holding company filed a Form 8-K announcing that it had that day sold approximately $21 billion of available-for-sale securities at an after-tax loss of approximately $1.8 billion, and that it was raising roughly $2.25 billion of new capital. Thursday 9 March was the run. The Federal Reserve's review of the episode records that deposit outflows exceeded $40 billion that day and that management expected $100 billion more on the Friday. There was no Friday. The California Department of Financial Protection and Innovation closed the bank on the morning of 10 March and appointed the FDIC as receiver.

The FDIC's initial resolution was conventional and cold. It created a Deposit Insurance National Bank of Santa Clara, promised insured depositors full access by Monday morning, and offered uninsured depositors an advance dividend plus a receivership certificate for the rest. Whatever a company held above the insurance limit, whether that was a few hundred thousand dollars or a quarter's payroll, it was now certain of $250,000 and holding a claim for the rest. That is the standard outcome of a bank failure, and within forty-eight hours it was judged unacceptable. On Sunday 12 March the Treasury Secretary, the Federal Reserve Chair and the FDIC Chairman jointly invoked the systemic risk exception, guaranteed every deposit at both Silicon Valley Bank and a newly closed Signature Bank of New York, and announced the Bank Term Funding Program the same evening.

Chronology of the acute phase

Dated events with the closing level of the S&P 500 and the constant maturity two-year Treasury yield on the same day.

Read that table for its two shapes rather than its individual rows. The equity column barely moves. The yield column collapses. Whatever was being repriced in March 2023, it was not the earnings of American companies in general, and a case study that opens with an index chart will find almost nothing to show.

Why Was Silicon Valley Bank's Balance Sheet Unusual?

Every explanation that stops at "rising rates hurt bond portfolios" is describing a condition shared by essentially every bank in the country. What was specific to Silicon Valley Bank was the combination, and it is legible from three numbers in the bank's own annual report.

It had roughly tripled in two years. Total assets went from $71.00 billion at the end of 2019 to $115.51 billion at the end of 2020 and $211.31 billion at the end of 2021, which is growth of about 198 percent across the two years. Deposits followed the same curve, from $61.76 billion to $189.20 billion. The Federal Reserve's review treats that stretch as the point supervision fell behind, noting the firm reached over $211 billion in assets without becoming subject to heightened supervisory or regulatory standards. A bank that triples in two years has to put the money somewhere, and in 2020 and 2021 the somewhere available was long-dated securities yielding very little.

It put the money in the longest assets available. At the end of 2022 the held-to-maturity book stood at $91.33 billion of amortized cost and $91.32 billion of net carrying value, against a fair value of $76.17 billion and gross unrealized losses of $15.16 billion. Of that carrying value, $86.04 billion matured after ten years. The accounting is the crucial detail: securities in that bucket are carried at amortized cost, so the loss never touches reported equity or regulatory capital while the bank holds them. Set against $16.00 billion of total stockholders' equity, that footnote described an institution whose economic capital was close to zero if it were ever forced to sell. Why a long bond loses so much value when rates rise is set out in bond duration explained, and the arithmetic can be run in the bond price and yield to maturity calculator.

It funded that with money that had no reason to stay. The 10-K put estimated uninsured deposits in United States offices at $151.5 billion against $173.1 billion of total deposits, plus $13.9 billion of foreign deposits not subject to any United States federal or state insurance regime. The standard limit of $250,000 per depositor, per insured bank, per ownership category is meaningful for a household and irrelevant to a company running payroll. Roughly seven-eighths of the funding base therefore had a rational, unhedged incentive to leave at the first credible sign of trouble. What that line does and does not cover is explained in FDIC deposit insurance.

SVB Financial Group balance sheet at year end, from the group's own filings with the Securities and Exchange Commission. All figures in billions of United States dollars. The equity column excludes noncontrolling interests, which added $291 million at the end of 2022.

The 2022 row is the one worth staring at. Assets stopped growing and deposits fell by about $16 billion over the year, which is what the end of the venture funding boom looks like on a bank's liability side. The share of deposits that paid interest rose from 33 percent to 53 percent over the same twelve months, meaning the cheap money was already being replaced by expensive money. None of that was hidden. It was in a filing anybody could read.

What Did the March 8 Announcement Say, and Why Did It Backfire?

The 8 March disclosure is the most instructive document of the whole episode, because on its own terms it was a competent piece of balance sheet management, and it destroyed the bank inside a day.

The bank sold approximately $21 billion of available-for-sale securities. The investor materials filed with the 8-K describe what was sold: United States Treasuries and agency securities, yielding 1.79 percent, with a duration of 3.6 years. It booked an after-tax loss of approximately $1.8 billion and set out to raise approximately $2.25 billion, made up of a $1.25 billion common stock offering, a $500 million offering of depositary shares representing mandatory convertible preferred, and a $500 million private subscription from General Atlantic. The same materials recorded term borrowings increasing from $15 billion at 31 December 2022 to $30 billion as at the date of the announcement.

Now look at the yield on what was sold against the date it was sold. On 8 March 2023 the two-year Treasury note closed at 5.05 percent, a level it had not reached since 15 June 2007. The bank was liquidating a portfolio earning 1.79 percent to buy assets earning something close to five, and the investor letter projected a payback period of approximately three years. In isolation that is obviously the right trade.

The problem was never the trade. It was what the trade admitted. Until 8 March a depositor could treat the held-to-maturity note in the 10-K as an accounting artefact, because the bank intended to hold those securities to maturity and would therefore never realize the loss. The moment it sold a large securities portfolio at a loss and asked the market for capital in the same sentence, it demonstrated that it might need liquidity badly enough to crystallize losses, converting a disclosed hypothetical into an observed behavior. It did so with the offerings only announced and not yet priced, so the bank spent the whole of the following session publicly short of capital without having raised any. No capital ever arrived, because the bank was in receivership two days after the announcement, and the General Atlantic subscription had been made contingent on the common offering closing first.

The disclosure trap. A bank whose asset losses are unrealized has an incentive never to act on them, because acting is itself the signal. Silicon Valley Bank was punished not for the state of its balance sheet, which had been public since February, but for doing something about it. This is worth holding onto when reading any institution's decision to raise capital in a stress: the announcement is simultaneously the remedy and the confession, and the market frequently prices the second before the first can settle.

How Fast Was the Run, and Why Did Speed Decide the Outcome?

Over $40 billion left in one business day. Against total assets the FDIC put at approximately $209.0 billion, that is close to a fifth of the bank in a single session, with roughly half the bank again expected to leave the next morning. No liquidity plan built on conventional assumptions survives an outflow at that rate, because a bank's assets cannot be monetized that fast without either a buyer standing ready or a central bank lending against the collateral immediately. Three features made the outflow faster than any historical template a stress test would have used.

The depositors were a network, not a population. A retail base of a million households learns about trouble through news, at different times, and reacts unevenly. Silicon Valley Bank's depositors were startups and the venture funds that sat on their boards, communicating in real time through the same group chats and investor mailing lists. A single fund advising its portfolio companies to move cash is a coordinated instruction to hundreds of depositors at once. Deposit concentration in a business sense had become deposit correlation in a statistical sense.

Moving money no longer requires moving. The FDIC recorded seventeen branches for a bank with $175.4 billion of deposits, which tells you the branch network was almost irrelevant to how the money was held or moved. Transfers were initiated from a phone, so the binding constraint on the speed of the run was the payment system's operating hours rather than physical logistics.

For an uninsured corporate depositor, leaving early is close to free. If the bank survives, moving the operating account cost a few days of administrative inconvenience. If the bank failed and you stayed, you might miss payroll for months while a receivership paid out. That payoff structure does not require panic. It makes withdrawing the correct decision at a much lower level of suspicion than most people assume, which is why a deposit base that is overwhelmingly uninsured is structurally unstable regardless of how well the bank is run.

The resulting distinction matters more here than any capital ratio. Solvency asks whether assets exceed liabilities given time; liquidity asks whether cash can be produced today. Silicon Valley Bank's held-to-maturity portfolio was United States government and agency paper with negligible credit risk, and the eventual recovery on those assets was never the real question. The bank failed because it could not convert good assets into cash on Thursday.

Which Warning Signs Were Visible in Advance, and Which Only in Hindsight?

This episode has an unusual property for financial history: the fragility was disclosed with precision, in a filing available to anyone, two weeks before the failure. That makes the hindsight question sharper rather than easier, because the honest finding is that public information supported a clear judgment about fragility and gave almost no purchase on timing.

Signals classified by whether they could be acted on with the information available before 8 March 2023.

The uncomfortable part is the top row. An analyst reading that 10-K in late February 2023 could have said, correctly and with figures, that this bank had no economic capital left if forced to sell its securities and that seven-eighths of its funding could leave at will. That is a complete description of the failure mechanism, published before the failure. What that analyst could not have said is when, or whether at all. Plenty of banks carried large unrealized securities losses in 2023 and did not fail, because their depositors did not leave: the FDIC put unrealized losses on securities across the whole banking system at $515.5 billion in the first quarter of 2023. The condition was general. The failure was not.

Hindsight check. The question to ask of any Silicon Valley Bank warning sign is not "was it visible" but "would acting on it have been distinguishable from acting on the same signal at fifty other banks that survived". For the held-to-maturity loss, it would not have been. What separated SVB was the concentration and insurance status of its deposits, which is why that variable, and not the securities loss, is the part worth generalizing. Why prior knowledge of an outcome makes the causal chain look more obvious than it was is covered in cognitive biases in trading.

What Did the Systemic Risk Exception and the Bank Term Funding Program Do?

The official response ran on two tracks addressing different problems, and confusing them is the most common analytical error about March 2023.

The systemic risk exception dealt with the incentive to run. The joint statement of 12 March declared that all depositors at Silicon Valley Bank and Signature Bank would be made whole, above and beyond the $250,000 insurance limit. It was equally explicit about who was not being rescued: shareholders and certain unsecured debtholders were not protected and senior management was removed. The statement said no losses would be borne by the taxpayer, and that any loss to the Deposit Insurance Fund from protecting uninsured depositors would be recovered by a special assessment on banks as required by law. That promise was honoured traceably. In November 2023 the FDIC finalized a special assessment covering approximately $16.3 billion of cost attributable to protecting uninsured depositors, levied at an annual rate of about 13.4 basis points across eight quarterly periods on 114 banking organizations, with no organization under $5 billion of total assets paying anything.

The Bank Term Funding Program dealt with the mechanics of the run. Announced the same evening, it offered depository institutions loans of up to one year against Treasury securities, agency debt and mortgage-backed securities, with one design choice that made all the difference: collateral was valued at par rather than at market price. The entire mechanism of the failure had been a bank unable to raise cash against securities worth less than face value, and the facility removed exactly that constraint. The Treasury made $25 billion available from the Exchange Stabilization Fund as a backstop. The uptake shows how binding the constraint had been.

Federal Reserve balance sheet lending to depository institutions, Wednesday levels in billions of United States dollars, from the H.4.1 release. The H.4.1 also reports a daily average per week, which is far lower during a spike: for the week to 15 March 2023 that average was $84.96 billion against the $152.85 billion outstanding on the Wednesday.

Discount window borrowing rose more than thirtyfold in a single week and then drained away over about seven weeks as the term facility took its place. That substitution is the whole story of the second half of March 2023 in two columns. Nor was the stress confined to the United States: on 19 March the Federal Reserve, the Bank of Canada, the Bank of England, the Bank of Japan, the European Central Bank and the Swiss National Bank jointly announced that their seven-day dollar swap operations would run daily rather than weekly from the following morning, and central banks do not change the operating frequency of standing facilities for a purely domestic problem.

One further point cuts against the reflex that a banking crisis forces a central bank to stop tightening. The Federal Reserve's record of target changes shows the federal funds range moving to 4.75 to 5.00 percent effective 23 March 2023, eleven days after the systemic risk exception. Monetary policy and financial stability policy were run as separate instruments against separate problems, a change from the pattern described in the 2008 financial crisis and explained further in Federal Reserve policy rates and forward guidance.

Why Did the Bond Market Convulse While Stocks Barely Moved?

March 2023 is the clearest case in this library of an event that was enormous in one market and almost invisible in another, and that divergence is the most useful thing an investor can take from it.

On the equity side, measured close to close, the S&P 500 went from 4,048.42 on 6 March to a low of 3,855.76 on 13 March, a fall of about 4.8 percent, and ended the month at 4,109.31, above where it stood before the announcement. An investor checking a broad index once a fortnight would have seen a mildly negative period and nothing more. On the rates side, the same days were historic. The two-year Treasury yield closed at 5.05 percent on 8 March, a level last seen on 15 June 2007, and at 3.76 percent on 23 March. The session of 13 March alone produced a 57 basis point decline in that yield. Measured against the full daily constant maturity series, which begins on 1 June 1976, that is the largest one-day fall in the two-year yield since 20 October 1987. The only comparable move in the intervening thirty-five years was the 54 basis point drop on 13 September 2001, when the Treasury market reopened after the September 11 attacks, which is the company that week keeps.

Daily constant maturity Treasury yields, in percent, around the failure.

Two things are happening in that column at once and should be kept apart. Part of the move is a flight to the safest instrument available, which is what always happens when the safety of bank deposits is in question. The larger part is a repricing of the expected policy path: a market that had spent the first week of March expecting rates to stay higher for longer spent the second week deciding a banking accident had shortened that path. The two-year note is sensitive to both, which is why it moved further than the ten-year.

The practical consequence is that a portfolio's experience here depended almost entirely on what it owned, in a way that broad market commentary could not capture. A holder of short-dated Treasury securities had an excellent fortnight. A holder of a diversified equity index had an uneventful one. A holder of the equity or unsecured debt of the wrong regional bank lost everything, since the joint statement protected depositors and explicitly did not protect them. The phrase "the market" performs no useful work here. How these relationships shift between calm and stressed periods is treated in how correlations change across regimes and in financial conditions, credit spreads and liquidity.

Where Did the Deposits Actually Go?

A run withdraws money from a bank. It does not withdraw money from the banking system, and the distinction determines who was hurt. The Federal Reserve's weekly H.8 release, which reports assets and liabilities of commercial banks in the United States, records where the money went with unusual clarity.

Deposits at commercial banks in the United States, seasonally adjusted weekly levels in billions of United States dollars, from the Federal Reserve's H.8 release.

In the week that contained the failure, deposits at small domestically chartered banks fell by $149.3 billion while deposits at large domestically chartered banks rose by $53.5 billion. That is a migration, not an evaporation. Corporate treasurers who had spent Friday discovering that their operating cash was an unsecured claim on a mid-sized bank spent the following week moving it to an institution they assumed would be protected in any circumstance, and some of it into government money market funds and Treasury bills instead. The all-bank column fell by about $420 billion across the five readings in that table, and by about $335 billion in the three weeks after 8 March alone, which is money leaving deposits altogether for higher-yielding cash. That flow had been running for a year as short-term rates rose, and the failure accelerated it. For anyone thinking about where operating cash actually sits, the mechanics are in Treasury bills as cash equivalents and bank sweep programs, since a sweep changes which institution holds the money and therefore which insurance applies.

The migration is also why the episode did not end in March. A smaller bank losing deposits to a larger one loses its cheapest funding first and replaces it at market rates, compressing net interest income for quarters afterwards. That slow squeeze, rather than any single day of withdrawals, is what eventually caught First Republic.

When Did the Stress End, and What Ended It?

Three clocks ran at once in this episode and they stopped roughly a year apart, which is why the answer depends entirely on which one is being read.

The depositor clock stopped on 13 March 2023. Once all depositors at the two failed banks had access to all of their money and any bank could borrow against par-valued collateral, the specific incentive to withdraw from a solvent institution was removed. That was three days after the closure, and it is the reason no comparable single-day outflow was reported afterwards.

The market clock stopped within about three weeks. The S&P 500 closed above its 6 March level before the end of the month. The two-year Treasury yield, having fallen from 5.05 percent to 3.76 percent, ended March at 4.06 percent, which is a partial retracement rather than a return to the starting point. That gap is meaningful: the equity market treated the episode as closed while the rates market continued to price a materially different policy path.

The bank balance sheet clock ran for more than a year. First Republic Bank was closed on 1 May 2023, seven weeks after Silicon Valley Bank, with approximately $229.1 billion in total assets and $103.9 billion in total deposits as of 13 April 2023. It was larger than Silicon Valley Bank on the FDIC's own measure, and JPMorgan Chase assumed all of its deposits at an estimated $13 billion cost to the Deposit Insurance Fund. On the funding side, borrowing under the Bank Term Funding Program did not peak until 24 January 2024, at $167.77 billion, and the facility only stopped making new loans on 11 March 2024.

That last clock is the one most commonly lost. An episode whose acute phase lasted seventy-two hours required official support for closer to twelve months. Defining recovery by when the headlines stopped would have produced an answer wrong by a factor of about a hundred, which is a general hazard in measuring the duration of any market event and a reason to define the measure before running it, as in the drawdown and recovery calculator.

Why Is Silicon Valley Bank a Poor Template for the Next Bank Failure?

Four features of this failure were specific enough that expecting them again is likely to be the wrong preparation.

The concentration was extreme even among concentrated banks. Roughly seven-eighths of deposits uninsured, drawn overwhelmingly from one industry whose participants shared boards, investors and communication channels. The Federal Reserve's review names the pairing directly, listing a highly concentrated business model and a reliance on uninsured deposits among the vulnerabilities that left the bank acutely exposed. A screen for banks with large securities losses would have caught hundreds of institutions; a screen for this funding structure caught very few.

The asset problem carried no credit risk. Silicon Valley Bank did not fail because its loans went bad, but because Treasury and agency securities had fallen in price as rates rose. Most bank failures a reader has encountered, including those described in the 2008 financial crisis, ran through credit losses instead. The closest structural precedent is the savings and loan crisis, where rising rates also destroyed the value of long assets funded by short liabilities, and the instructive difference is speed: that damage took years to surface and years more to resolve, while SVB's took two days.

The policy response had a tool built for exactly this problem. A facility lending at par against government securities neutralizes a mark-to-market liquidity shortfall almost perfectly, and would do nothing for a bank whose assets had fallen because borrowers stopped paying, since those assets are not worth par under any accounting. Reassurance that a central bank can stop a run of this shape does not transfer to a run of a different shape.

The macro setting was the mirror image of most crises. This failure was a direct consequence of the fastest tightening cycle in decades, described in the 2022 rate shock, and the Federal Reserve kept raising rates eleven days after invoking the systemic risk exception. Contrast the 2020 COVID crash, where the shock came from outside the financial system and the response was to ease as hard and fast as possible. A crisis caused by high rates and a crisis solved by cutting rates are not the same species.

Common Myths About the Silicon Valley Bank Failure

"It was a bailout." The word obscures who got what. Depositors were made whole. Shareholders in SVB Financial Group and Signature Bank were not protected, certain unsecured debtholders were not protected, and senior management was removed, all stated explicitly in the joint statement. The roughly $16.3 billion cost of protecting uninsured depositors at the two banks was recovered through a special assessment on other banks, not from public funds. Whether guaranteeing uninsured deposits after the fact was wise policy is a real argument. Describing it as a rescue of the bank's owners is factually wrong.

"It was a crypto failure." Silicon Valley Bank's held-to-maturity book was $91.33 billion of government and agency paper funded by venture-backed corporate deposits. Signature Bank's closure two days later generated a lot of narrative crossover, but the mechanism at Silicon Valley Bank was interest rate duration funded by uninsured corporate cash, and no digital asset appears anywhere in the chain.

"Nobody could have seen it coming." Both numbers that describe the failure were in the Form 10-K of 24 February 2023, quoted twice above. The genuine difficulty was not visibility but discrimination, because many banks carried a securities loss of that shape and very few carried a deposit base of that shape.

"The banking system nearly collapsed." Deposits moved from smaller banks to larger ones and into Treasury bills, the S&P 500 finished March 2023 higher than it began, and each of the three failures was resolved with depositors made whole. That was a serious, expensive and genuinely dangerous episode, and it was not 2008. Treating every bank failure as systemic is as unhelpful as treating none of them that way.

"Held-to-maturity accounting hid the loss." It kept the loss out of reported equity and regulatory capital, which is a real and consequential effect, and it hid nothing: the loss was disclosed to the nearest ten million dollars in the notes. That the information did not move the share price or the deposit base until March is a fact about how filings are read, not about what they contained. Reading the notes rather than the headline figures is the subject of fundamental analysis.

What a Reader Can Actually Carry Forward

The temptation is to convert this case into a rule about avoiding regional bank shares. That is the least useful reading available, and it ignores the part of the episode that touches almost everybody.

What generalizes

• Cash in a bank is a claim on that bank, not a holding of money. Above the insurance limit, an operating balance is an unsecured claim whose value depends on the institution's condition, as a great many depositors discovered over one weekend in March 2023. The practical response is to know where the balance sits and whether the arrangement spreads deposits across institutions, which is what FDIC deposit insurance exists to explain.
• An unrealized loss is a real loss that has not been forced yet. Marking an asset at cost changes the accounting, not the economics. The question is whether the holder can afford to wait, and Silicon Valley Bank could not, because its liabilities were callable on demand while most of its securities book matured beyond ten years.
• Concentration on the funding side is a risk in its own right. Portfolio risk is usually discussed in terms of what you own. This episode was decided by who the bank owed money to and how fast those creditors could talk to each other, which is the domain of risk management.
• Test the liability side, not just the asset side. A stress test asking what happens if these securities fall 15 percent would have shown Silicon Valley Bank surviving. One asking what happens if a quarter of deposits leave in two days would have shown it failing. Both were runnable in February 2023, and only the second mattered. Stress testing and scenario analysis covers how to build the second kind.

What does not generalize

• The thirty-six hour timeline. It depended on a specific depositor network. A bank with retail deposits and a normal insurance profile fails, when it fails, over months.
• The complete protection of uninsured depositors. That required a discretionary invocation of the systemic risk exception by three officials over a weekend. It is a decision, not a rule, and it was not the initial response on 10 March.
• The absence of credit losses. A failure driven entirely by interest rate duration on government securities is historically unusual, and it is why a par-value lending facility could resolve it so cleanly.

The one question worth asking now

Not "could my bank fail", which almost nobody can answer, but a narrower and fully answerable one: if every balance you hold above $250,000 at a single institution were frozen for six months, what would stop working? For most individuals the answer is nothing, because the balances sit below the line. For a business with payroll, or anyone holding a large cash position in one place, the answer is specific and unpleasant, and the fix is administrative rather than analytical. That is the part of March 2023 that is genuinely portable, and it requires forecasting nothing at all.

References

Every figure on this page was verified against the following sources, each retrieved on 26 August 2026:

• SVB Financial Group: Annual Report on Form 10-K for the Year Ended December 31, 2022: every SVB balance sheet figure quoted, including the $91.327 billion held-to-maturity book, its $76.169 billion fair value and $15.160 billion of gross unrealized losses, the $86.038 billion maturing after ten years, $151.5 billion of estimated uninsured deposits in United States offices, $13.9 billion of foreign deposits, and interest-bearing deposits rising from 33 to 53 percent.
• SVB Financial Group: Current Report on Form 8-K, March 8, 2023: the completed sale of substantially all of the available-for-sale portfolio, the approximately $1.8 billion after-tax loss, and the $1.25 billion common offering, $500 million of depositary shares and $500 million General Atlantic subscription that make up the $2.25 billion raise.
• SVB Financial Group: Exhibit 99.2 to the Form 8-K of March 8, 2023, Mid-Quarter Update: the $21 billion sale size, United States Treasuries and agency securities as what was sold, the 1.79 percent yield and 3.6 year duration, and term borrowings increasing from $15 billion at 31 December 2022 to $30 billion. These figures are in the exhibit, not in the 8-K body.
• SVB Financial Group: Exhibit 99.3 to the Form 8-K of March 8, 2023, Message to Stakeholders: the approximately three-year payback period and the $1.75 billion sought from the two public offerings before the General Atlantic subscription.
• Federal Reserve Board: Review of the Federal Reserve's Supervision and Regulation of Silicon Valley Bank, Key Takeaways: asset growth from $71 billion to over $211 billion, outflows of over $40 billion on 9 March 2023 with $100 billion more expected, and the 10 March closure.
• Federal Reserve Board: Federal Reserve Board Announces It Will Make Available Additional Funding to Eligible Depository Institutions, March 12, 2023: one-year loans, collateral valued at par, and the $25 billion Exchange Stabilization Fund backstop.
• Federal Reserve Board: Joint Statement by Treasury, Federal Reserve and FDIC, March 12, 2023: full depositor access from 13 March, the non-protection of shareholders and certain unsecured debtholders, and recovery of Deposit Insurance Fund losses by special assessment rather than from taxpayers.
• Federal Reserve Board: Coordinated Central Bank Action to Enhance the Provision of US Dollar Liquidity, March 19, 2023: the six central banks and the move to daily dollar operations.
• Federal Reserve Board: Bank Term Funding Program Will Cease Making New Loans as Scheduled on March 11, January 24, 2024: the 11 March 2024 end of new lending.
• Federal Reserve Board: Open Market Operations: the 4.75 to 5.00 percent target range effective 23 March 2023.
• FDIC: FDIC Creates a Deposit Insurance National Bank of Santa Clara to Protect Insured Depositors of Silicon Valley Bank: $209.0 billion of assets, $175.4 billion of deposits, seventeen branches, and the initial depositor treatment.
• FDIC: FDIC Establishes Signature Bridge Bank, N.A., as Successor to Signature Bank: $110.4 billion of assets, $88.6 billion of deposits and the 12 March 2023 closure.
• FDIC: JPMorgan Chase Bank, National Association Assumes All the Deposits of First Republic Bank: the 1 May 2023 closure, $229.1 billion of assets, $103.9 billion of deposits and the estimated $13 billion fund cost.
• FDIC: Final Rule on Special Assessment Pursuant to Systemic Risk Determination: the $16.3 billion attributable to protecting uninsured depositors, the 13.4 basis point rate and the 114 banking organizations assessed.
• FDIC: FDIC-Insured Institutions Reported Net Income of $79.8 Billion in First Quarter 2023: system-wide unrealized losses on securities of $515.5 billion.
• FDIC: Deposit Insurance: the $250,000 coverage limit.
• Federal Reserve Board: H.4.1 Factors Affecting Reserve Balances, Week Ended March 15, 2023: the $152.85 billion Wednesday level for primary credit and the $11.94 billion for the Bank Term Funding Program, alongside the $84.96 billion and $2.44 billion daily averages for the same week that the table note distinguishes.
• Federal Reserve Bank of St. Louis, for every yield, index close and weekly balance quoted: 2-Year Treasury Constant Maturity Rate, Series DGS2 (including the ranking of the 13 March fall against the full series from 1 June 1976), 10-Year Treasury Constant Maturity Rate, Series DGS10, S&P 500, Series SP500, Loans, Primary Credit, Wednesday Level, Series WLCFLPCL, Loans, Bank Term Funding Program, Net, Wednesday Level, Series H41RESPPALDKNWW, Deposits, Small Domestically Chartered Commercial Banks, Series DPSSCBW027SBOG, Deposits, Large Domestically Chartered Commercial Banks, Series DPSLCBW027SBOG and Deposits, All Commercial Banks, Series DPSACBW027SBOG.

Figures deliberately not stated. This page gives no percentage decline for any bank stock or bank sector index, no figure for Silicon Valley Bank's share price on 9 March 2023, no count of total bank failures in 2023, no terms for the First Citizens acquisition of Silicon Valley Bank's assets, and no numbers relating to the Credit Suisse resolution in Switzerland, because no primary or institutional source verified for this page supplied them. Two peer-comparison findings once attributed here to a Government Accountability Office review were also removed: gao.gov refused every request made while checking this page, and a document that cannot be read cannot be cited. The asset-growth figure it had supported survives, because that is arithmetic on the balance sheet above rather than anyone's finding. Where the mechanism is documented but the magnitude is not, as with the pace of withdrawals on 9 March beyond the Federal Reserve's own $40 billion figure, the mechanism is described and no number is supplied.

Method note: figures described as computed were derived by Swoopr Investment from the daily and weekly series named above. Percentage changes in the S&P 500 are close to close and price only, so they exclude dividends and are not intraday. Deposit levels from the H.8 release are seasonally adjusted weekly averages, which is why they differ from any single institution's balance on a given day. The FDIC reports Silicon Valley Bank at the insured-bank level while SVB Financial Group's Form 10-K reports the consolidated holding company, which is why the two asset and deposit totals for 31 December 2022 differ; both are stated with their source rather than reconciled.

This is educational content about a historical episode. It is not investment advice, it is not a forecast, and nothing here should be read as a claim about the condition of any bank today.

Frequently Asked Questions