2023 SVB and Rate Risk
2023 SVB and Rate Risk
Silicon Valley Bank collapsed in March 2023 when a silent duration mismatch became a funding crisis. Bonds that were supposedly "safe" and "held to maturity" proved illiquid and deeply underwater, forcing a forced sale that revealed the bank's insolvency.
Key takeaways
- SVB accumulated a £91 billion bond portfolio of low-yielding long-dated Treasuries and mortgage-backed securities, locked in at 2022 levels
- When rates rose to 4%+ in 2022–2023, those bonds were worth 20–30% less than their book value, a loss of £15–20 billion
- The held-to-maturity (HTM) accounting treatment hid the losses from investors and regulators for months
- Rapid deposit outflows forced SVB to sell bonds at market prices, crystallizing losses and revealing a negative net worth
- The crisis demonstrated that bond duration risk and funding duration mismatch remain critical vulnerabilities in the financial system
The SVB Setup: Excess Deposits and Low Rates
Silicon Valley Bank was a regional lender to tech startups and venture capital firms. In the 2021–2022 period, as venture capital funding surged and tech companies raised enormous rounds, SVB saw its deposit base explode. Companies and funds were raising billion-dollar rounds and parking cash at SVB, their go-to bank for tech financing.
In 2021, SVB's deposits grew from £80 billion to £120 billion. In 2022, they continued growing. By early 2023, SVB had ballooned to nearly £210 billion in assets, mostly funded by this explosive deposit growth. The deposits were "hot" money—venture funds and companies that would withdraw them as quickly as they arrived.
During the 2020–2021 period of near-zero interest rates and quantitative easing, SVB faced a classic banker's problem: what do you do with £100+ billion in deposits when short-term lending yields almost nothing? Money market funds offered near-0%. Treasury bills offered 0.5–1%. There was no way to earn meaningful returns without going further out on the yield curve.
SVB's response: stuff the balance sheet with long-dated, low-yielding bonds. They bought £91 billion in U.S. Treasuries and mortgage-backed securities (MBS), many with maturities of 10–15 years. The yields looked acceptable in 2021: 1.5–2.5%. But when rates started rising in 2022, those yields became deeply underwater.
This was a classic "duration mismatch" problem: SVB had short-duration liabilities (demand deposits that could leave at any time) and long-duration assets (bonds with 7–10 years of duration, locked into low yields).
The Hidden Losses Grow
As the Fed raised rates from 0% to 4.33% through 2022, the market value of SVB's bond portfolio collapsed. A 10-year Treasury with a 1.5% coupon becomes worth much less when the market yield is 4%. The duration losses were severe: a 10-year bond with 8 years of remaining duration would lose roughly 20% in value for a 250 basis point yield rise.
By the end of 2022, SVB's bond portfolio was worth roughly £75–80 billion in market terms, down from a book value (historical cost) of £91 billion. The bank had accumulated a hidden loss of £11–16 billion.
But here's the accounting trick: SVB classified most of these bonds as "held to maturity" (HTM) under U.S. accounting rules. HTM bonds are not marked to market on the balance sheet. Instead, they're carried at amortized cost. This meant that SVB's balance sheet showed £91 billion in assets, even though the true market value was only £75–80 billion.
The bank's actual capital (assets minus liabilities) was far lower than reported. If the deposit base held steady and the bank could hold all those bonds to maturity (collecting the 1.5–2.0% coupons despite rates being 4%), the strategy would eventually work. But it required perfect stability. One stress—either rising rates OR deposit flight—would expose the losses.
The Deposit Flight
In early March 2023, several factors collided:
First, the collapse of Silvergate Bank (a smaller crypto-focused lender) on March 8 spooked the banking sector. Regulators signaled concerns about regional banks. Investors and depositors began asking: are my regional banks stable?
Second, venture capital was pulling back. The venture bubble of 2021–2022 had burst, and startups were raising less, burning through cash faster, and drawing down deposits at lenders like SVB. The explosive deposit growth of 2021–2022 reversed almost immediately.
Third, and most importantly: a venture capitalist named Peter Thiel, an early SVB depositor, publicly recommended that portfolio companies withdraw deposits from SVB. Thiel's reasoning was simple: if deposit outflows accelerated, SVB would be forced to sell bonds at a massive loss. Once that became public, the bank would become insolvent, and deposits above £250,000 (the FDIC insurance limit) would be at risk.
This was brilliant analysis and fatal advice. By recommending withdrawals, Thiel ensured the precise scenario he was warning about: a deposit run that forced bond sales at massive losses.
On March 8–9, SVB saw roughly £40 billion in deposit withdrawals. Panicked depositors were getting their money out while they still could. The bank needed liquidity immediately.
The Death Spiral
To raise cash quickly, SVB had two options:
-
Borrow from the Federal Reserve's discount window (the lender of last resort). But this required collateral, and SVB's bonds would be valued at a steep discount to market value.
-
Sell bonds into the market at current market prices.
SVB chose a combination. They announced they would raise £2.25 billion in new capital and sell a chunk of bonds to raise additional liquidity.
The announcement did the opposite of reassure. It confirmed that SVB needed to raise capital, which meant the bank was in trouble. More depositors panicked.
On March 10, SVB announced that it had sold £1.4 billion in bonds at a realized loss of £915 million. This crystallized the hidden losses. The market suddenly understood: SVB had £15+ billion in underwater bonds and was rapidly running out of liquidity to fund deposits.
By the afternoon of March 10, the bank was insolvent on a mark-to-market basis. Liabilities (deposits of £210 billion plus borrowings) exceeded the true value of assets. The bank ceased to exist as a going concern, though it took until late evening for regulators to officially close it.
The Cascade of Contagion
SVB's collapse sent shockwaves through the regional banking sector. Investors suddenly realized that many regional banks had similar mismatches: deposits exceeding loans (meaning they held huge bond portfolios), bonds largely classified as HTM (meaning hidden losses were widespread), and deposit bases that could become unstable in a crisis.
Signature Bank, which had similar characteristics (large deposit base, significant HTM bond holdings), saw deposit outflows and failed by March 12. First Republic Bank, another regional bank with similar characteristics, saw stress and ultimately failed in May 2023.
The contagion was not a financial system-wide collapse (unlike 2008), because the Federal Reserve and FDIC moved quickly. The Fed established an emergency facility to lend to banks against their bond collateral at par value (full book value, not discounted market value). This removed the immediate liquidity pressure. The FDIC temporarily insured all deposits at SVB and other troubled banks, removing the incentive to run.
But the underlying lesson was stark: the bond market had dealt a blow to the banking system, and duration risk was still underestimated.
The Real Culprit: Rate Risk and Maturity Mismatch
The SVB story is often told as a case of poor risk management or lax regulation (all true), but the fundamental problem was rate risk on bonds combined with maturity mismatch.
A bank with a stable funding base and long-term liabilities can safely hold long-duration bonds. A bank with short-term deposits and a 50-basis-point spread between deposit costs and bond yields cannot. SVB's model broke because:
- The spread between deposit costs and bond yields was very tight (maybe 75 basis points in 2021–2022)
- The funding base was unstable (hot deposits from venture capital)
- The bond portfolio was massive relative to loan origination (£91 billion bonds vs. ongoing lending)
- The rates scenario was a tail risk: a rise from 0% to 4% is unusual, but it happened
Any competent risk management would have forced SVB to maintain higher capital buffers, shorter-duration bond positions, or a combination of both. SVB's management chose instead to reach for yield by loading up on duration—the exact wrong strategy for a bank with unstable, short-term funding.
The Duration Unwind of 2023
Why This Matters for Bond Investors
The SVB collapse teaches several lessons:
Duration risk is real and persistent. Even professional investors and large institutions can underestimate how much a 250–300 basis point rate rise can hurt a long-duration portfolio. SVB's management seemingly believed that rates would stabilize at 0–1% for years. They were badly wrong.
Maturity mismatch is still dangerous. Banking regulators have stress-tested interest rate risk since 2008, but clearly the stress tests at SVB were inadequate. When you finance long-term assets with short-term liabilities, you need a very large equity cushion.
Accounting can hide reality. The HTM classification allowed SVB to report a healthy balance sheet while being deeply insolvent. This made it harder for depositors, investors, and even regulators to assess true financial health.
Liquidity and solvency are intertwined. SVB was insolvent on a mark-to-market basis (assets worth less than liabilities). But they became insolvent in a technical sense because they couldn't access liquidity to fund deposit withdrawals. A bank can be solvent in the long run but illiquid in the short run—and if it can't bridge that gap, it fails.
For individual investors, the SVB crisis reinforces the importance of diversification and not underestimating interest rate risk, especially when yields are abnormally low.
Related concepts
Next
By mid-2023, the immediate SVB-contagion panic had subsided, helped by emergency Fed lending and FDIC insurance for uninsured deposits. But the episode raised questions about the durability of higher-rate environment. Would the Fed continue tightening, or would banking stress force a reversal? The answer would come within months.