Banks as a System — Regulation and Deregulation

Core idea

Banks are not islands — they are nodes in a network. They lend to each other overnight in the federal funds market to plug short-term reserve gaps. They share customer panic when one of them looks shaky, because depositors at the next bank wonder if they should be next in line. And they live under a regulatory regime that exists precisely because the network amplifies both their usefulness and their failures. The 160-year pendulum of US banking — National Bank Act of 1863, Federal Reserve Act of 1913, Glass-Steagall, FDIC, the S&L crisis, Glass-Steagall's repeal, the 2008 collapse, Dodd-Frank, Dodd-Frank's partial rollback, the 2023 collapses of SVB, Signature, and First Republic — is the story of regulators trying to keep up with a system that keeps finding new ways to fail.

The two structural features that drive everything else are: interconnection (every bank's balance sheet partly depends on other banks) and fractional reserves (no bank can pay all its depositors at once). Together they mean confidence is the load-bearing wall. When confidence cracks, the rest follows fast.

Authors' framing: Banks borrow from each other to manage daily liquidity. They run on confidence. They fail when confidence fails. Regulation exists because the consequences of failure aren't contained to the failing bank — they spread.

Why it matters

It demystifies the federal funds rate

The Fed's most-watched policy lever — the federal funds target rate — is literally the rate banks charge each other for overnight reserve loans. When you understand the interbank lending market, "the Fed raised rates" stops being mysterious. The Fed sets a target for the price of overnight money between banks, and that price is the foundation under every other interest rate in the economy (The Time Value of Money and Interest Rates).

It explains why bank runs are so destructive

A single bank failure used to be contained. In a connected, fractional-reserve system, the failure becomes a signal: if Bank A failed, maybe Bank B is next. Depositors at Bank B race to withdraw. The race makes the failure happen. This is the logic behind every panic from 1907 to 2023 — and the logic behind every modern defense (deposit insurance, the lender-of-last-resort role of the central bank, stress tests).

It puts the deregulation-then-crisis pattern in context

The 1980s S&L crisis, the 2008 financial crisis, and the 2023 regional bank failures all followed the same script: deregulation → riskier behavior → a precipitating shock → systemic stress → public bailout → reregulation. Knowing the pattern lets you read political fights over bank rules with sharper eyes — the lobbying for looser rules tends to peak right before the next crisis.

Key takeaways

Mental model — how one bank borrows from another

Copy sourceZoom

Mental model — anatomy of a bank run

Copy sourceZoom

Mental model — the regulation pendulum

Copy sourceZoom

Practical application

Diagnose a banking-news story

Steps for evaluating a bank account's safety

Example: a stylized version of the 2023 SVB failure

A regional bank serves a concentrated industry (tech startups). It takes in a lot of deposits during a boom — far more than it can lend out locally. So it parks the excess in long-duration Treasury bonds, which is conservative on paper.

Then two things happen at once:

1. Interest rates rise sharply. The market value of the bank's long-duration bonds falls — not because anyone defaulted, but because new bonds yield more. Unrealized losses pile up on the bond portfolio.

2. The startup ecosystem cools. Customers start drawing down deposits to fund operations. The bank has to sell some of those bonds to meet withdrawals — turning unrealized losses into realized ones.

The combination causes the bank to announce a large securities loss. Within hours, on group chats and Twitter, sophisticated depositors with balances far above the $250K FDIC limit do the math: if the loss makes the bank insolvent, only the first $250K of my $20M is safe. So they wire their money out — and tell their networks to do the same. By the next morning, $40 billion has left the building and the bank cannot meet withdrawal demands. Regulators close it within 48 hours.

What broke? Three things stacked: a concentrated depositor base (so contagion within the cluster was fast), a portfolio that lost value when rates rose (a risk the bank had not adequately hedged), and a regulatory exemption (under the 2018 rollback, this bank was not subject to the strictest stress tests). Each could have been survivable alone. Together they produced a textbook 21st-century bank run — one in which the line forming wasn't at the door, but in everyone's group chats.

Caveats

Related lessons