March 2023EconomicHistoric

Silicon Valley Bank (SVB) Collapse 2023

Year

2023

Region

United States

Markets

Banking، Macro، Stocks

Crisis

Bank Run، Liquidity Crisis، Contagion

Severity

Historic

Reading

4 min

Sources

Core lesson

Liquidity and solvency are not the same thing. SVB's assets may have been worth enough over time, but it could not turn them into cash fast enough to meet withdrawals — and that gap is what kills institutions and accounts alike.

◵ Practice the pattern · Practice framework

⏱ 30-second summary

A two-day digital bank run drained $42 billion from Silicon Valley Bank and forced regulators to seize it — the largest US bank failure since 2008 and a lesson in how fast deposits flee in the smartphone era.

Timeline

2020–2021
Deposits flood in, parked in long bonds
Amid the tech boom, SVB's deposits more than tripled. Unable to lend it all, the bank bought tens of billions in long-dated Treasury and mortgage bonds — building hidden duration risk.
2022
Fed rate hikes sink bond values
The Federal Reserve raised rates sharply to fight inflation. The market value of SVB's bonds fell, leaving large unrealised losses on assets it might be forced to sell.
March 8, 2023
$1.8B loss and a capital raise
SVB sold about $21B of bonds at a ~$1.8B loss and announced a capital raise. The disclosure alarmed its tightly connected depositor base instead of reassuring it.
March 9, 2023
A $42B run in one day
Word spread through VC networks and chats; the share price collapsed and depositors tried to withdraw roughly $42 billion in a single day — a digital-age bank run at unprecedented speed.
March 10, 2023
Regulators seize SVB
California regulators closed the bank and named the FDIC receiver — the largest US bank failure since 2008. Fear spread to other regional lenders.
March 12, 2023
Authorities backstop all depositors
Treasury, the Fed, and the FDIC jointly guaranteed access to all deposits — insured and uninsured — to stop a wider panic and contain contagion.

What happened?

Silicon Valley Bank was the banking hub for US tech startups and venture capital. During the cheap-money boom of 2020–2021, deposits flooded in faster than the bank could lend, so it parked tens of billions in long-dated US Treasury and mortgage bonds. Those bonds are very safe if held to maturity — but when the Federal Reserve raised interest rates sharply in 2022, their market value fell. SVB was sitting on huge unrealised losses, a classic case of duration risk: borrowing money that can leave on demand while investing it in assets that only pay off years later.

On March 8, 2023, SVB sold about $21 billion of those bonds at a roughly $1.8 billion loss and announced a plan to raise fresh capital. Instead of reassuring people, the disclosure spooked them. SVB's customer base was unusually concentrated — a tight, well-connected community of founders and VCs, most with balances far above the $250,000 insurance limit. Word spread instantly through group chats and venture firms, and depositors raced to pull their money. On March 9 alone, customers tried to withdraw about $42 billion in a single day.

On Friday, March 10, regulators closed the bank and the FDIC took over as receiver — making it, at the time, the second-largest bank failure in US history. Fear quickly jumped to other regional lenders (Signature Bank failed two days later), the early signs of financial contagion. To stop the panic, on March 12 the Treasury, Federal Reserve, and FDIC jointly guaranteed that all SVB and Signature depositors — even uninsured ones — would have access to their money. The Fed's own April 2023 postmortem concluded the failure was driven by poor risk management at the bank and weak supervision.

Why markets reacted

The trigger looked technical — a bond sale and a routine-sounding capital raise — but markets and depositors read it as proof the bank was short of cash. In a run, perception becomes reality: once people believe others will pull their money, the rational move is to pull yours first, and that collective dash is exactly what makes a bank insolvent.

SVB's depositors were unusually homogeneous and connected, so they panicked together and moved together. With modern apps, billions could leave in hours rather than days, and bank stocks across the sector sold off as traders priced in flight-to-safety and the risk of contagion to other lenders holding similar long-dated bonds.

What traders usually get wrong

Treating a falling regional-bank stock as a cheap bargain on the way down, without grasping that a deposit run can wipe out equity holders in days — fundamentals do not matter if confidence evaporates.

Ignoring funding and liquidity risk because the balance sheet 'looks fine' — the bonds were technically safe, yet the duration mismatch was the whole problem.

Assuming a single name's troubles stay contained, and being caught off guard when fear jumped to other lenders and dragged correlated positions down together.

Adding leverage or oversizing into a fast-moving panic, when gaps and halts make it impossible to exit at your planned price.

The risk lesson for traders

Liquidity and solvency are not the same thing. SVB's assets may have been worth enough over time, but it could not turn them into cash fast enough to meet withdrawals — and that gap is what kills institutions and accounts alike.
Concentration is a hidden risk. A funding base (or a portfolio) made of similar players who all act at once turns a shock into a stampede; diversification of exposure and of counterparties is a defence.
Speed has changed. Information and money now move at the pace of a group chat, so the time you have to react to bad news is far shorter than the old playbooks assume.
When everyone reaches for cash at once, even safe assets get sold and contagion can spread between unrelated names — assume correlations can spike in a panic.

Practise this lesson in Map.Trade

Practice framework

Practice the pattern, not the exact event. Rehearse how a 'technical' headline (a capital raise, a forced bond sale) can trigger a liquidity scramble, and drill the habit of checking funding and crowding risk before assuming a falling name is a bargain.

Replay Lab →Practice scenarios →Testing Workflow →Risk review →