Module 04 showed that conventional banking is structurally fragile. No society tolerates recurring collapse of its payment system, so every developed economy has built a safety net to contain that fragility — and each piece of it targets a specific failure mode we have already met. This module takes the net apart: how deposit insurance is funded and why it is capped, how a lender of last resort breaks the fire-sale loop, the genuinely hard problem of telling an illiquid bank from an insolvent one in real time, and what capital and liquidity regulation actually require. It ends on the sting in the tail — that the same net which tames fragility also builds a wall around the industry.
The safety net is best understood as one side of a bargain first sketched in Module 01. Society wants the benefits of banking — pooled savings funding long-term investment, a smooth payment system, a working money supply — without periodically watching it collapse. So the state agrees to backstop the system. In exchange, it claims the right to decide who may operate a bank and to dictate, in detail, how they must run it. Stability flows one way; control flows the other.
The net itself has three main pillars, and the cleanest way to hold them is to remember that each one is engineered to defeat a specific failure mode from Module 04:
Together they have made outright banking panics far rarer in developed economies than they were a century ago. But each pillar carries a cost — in money, in moral hazard, or in competition — and the costs are as important to understand as the benefits. We take the pillars in turn, then add up what they jointly imply for who gets to be a bank.
The safety net is not one thing but three tools, each aimed at a different way banks fail. Deposit insurance stops runs by removing the incentive to run; the lender of last resort stops fire sales by supplying liquidity; capital and liquidity rules make banks harder to break in the first place. Read every detail that follows as an answer to the question: which failure mode does this address, and at what cost?
Deposit insurance is a government-backed promise that, if your bank fails, you will still get your money back up to a stated limit. Its genius is psychological as much as financial: because you know you are protected, you have no reason to join a run, so the run that would have destroyed the bank never starts. It works by making itself rarely needed. This is the direct answer to the Diamond-Dybvig coordination trap from Module 04 — it eliminates the bad equilibrium by removing the incentive that creates it.
Where does the money come from? Typically not, in the first instance, from taxpayers. Banks pay premiums into an insurance fund, and that fund pays out when a member bank fails. In the United States the fund is run by the Federal Deposit Insurance Corporation (the FDIC) and premiums are risk-based — riskier banks pay more — which also nudges behavior. The government typically stands behind the fund as a final guarantor, which is what makes the promise fully credible, but the routine cost is borne by the industry itself.
When an insured bank does fail, the insurer rarely just mails cheques. More often it arranges a resolution: it finds a healthy bank to take over the failed bank's deposits and good assets (a "purchase and assumption"), often over a single weekend, so that depositors wake up on Monday with their accounts simply transferred and never notice an interruption. The on-demand promise is honored not by the failed bank but by the net standing behind it.
Deposit insurance stops runs before they start. By guaranteeing repayment, it removes the reason any covered depositor would rush to withdraw — so the coordination trap that topples sound banks cannot spring. It is funded mainly by bank-paid premiums into a fund, with the government as ultimate backstop, and failures are usually handled by transferring deposits to a healthy bank rather than by paying everyone out in cash.
Deposit insurance is never unlimited. It covers balances only up to a ceiling — roughly $250,000 per depositor per bank in the United States, €100,000 across the European Union, £85,000 in the United Kingdom. Above the cap, you are back to being the unsecured creditor of Module 03, fully exposed to the bank's failure.
Why cap it at all, rather than guarantee everything? Two reasons pull in the same direction. First, an unlimited guarantee would put the fund — and ultimately taxpayers — on the hook without limit. Second, and more subtly, a total guarantee would mean no one ever monitored their bank's soundness, eliminating whatever market discipline large, sophisticated depositors might otherwise impose. The cap is a compromise: protect ordinary households completely, while leaving big depositors with a reason to care whether their bank is sound.
The compromise has a sharp edge, and Module 04's SVB case sits right on it. Businesses cannot operate within a $250,000 limit — they hold millions in operating cash to make payroll — so the bulk of a business-focused bank's deposits are typically uninsured. Those depositors are exactly the ones with the incentive, the sophistication, and now the technology to run at the first sign of trouble. The cap that preserves discipline in normal times is the same cap that leaves a business bank acutely run-prone in a crisis. This is why, in the worst moments, authorities sometimes abandon the cap entirely and guarantee all deposits — as several governments did in 2008, and as U.S. regulators did for SVB's depositors in 2023 — trading away the discipline to stop the contagion.
Deposit insurance handles the small depositor. But a bank can still be drained — by uninsured depositors, by other banks pulling their funding, by a wholesale-market freeze — and find itself unable to meet demands despite holding sound but illiquid assets. This is the fire-sale doom loop of Module 04: forced to raise cash now, the bank dumps assets at ruinous prices, realizes losses, and turns its liquidity problem into insolvency. The tool built to break that loop is the central bank acting as lender of last resort.
The idea is old and is usually stated as Bagehot's rule, after Walter Bagehot, who set it out in 1873. In a panic, the central bank should:
The mechanism breaks the doom loop precisely. Instead of selling its illiquid assets into a falling market, the troubled bank pledges them to the central bank and borrows against them, getting the cash it needs to meet withdrawals while keeping the assets. No fire sale, no forced losses, no contagious markdown of everyone else's holdings. When the panic passes, the bank repays and reclaims its collateral. A liquidity crisis is bridged rather than allowed to manufacture an insolvency. In modern systems this facility is often called the discount window.
Bagehot's rule contains a crucial qualifier that is easy to state and nearly impossible to apply: the central bank should rescue banks that are illiquid but solvent — sound institutions caught in a panic — and should not prop up banks that are genuinely insolvent, whose assets are truly worth less than their liabilities. Lending to the illiquid is bridging a temporary gap; lending to the insolvent is throwing good money after bad and rewarding failure.
The trouble is that in the heat of a crisis, the two are almost impossible to tell apart. Module 04 showed why: the fire-sale dynamic means a run makes banks look insolvent — asset prices collapse precisely because everyone is selling — so at the moment you must decide, a merely illiquid bank and a deeply insolvent one can look identical on the screen. And the decision must be made in hours, with incomplete information, knowing that guessing wrong in either direction is costly: refuse a solvent bank and you let a fixable panic destroy it and spread; rescue an insolvent one and you have quietly bailed it out with public support.
This judgment problem is why the clean Bagehot rule blurs, in practice, into something messier — and why lender-of-last-resort actions repeatedly shade into outright bailouts. And bailouts breed the safety net's deepest cost: moral hazard. If banks come to expect rescue when things go wrong, they have less reason to hold thick capital or avoid risky bets — they keep the gains in good times and expect the public to absorb the losses in bad ones. The largest institutions, whose failure would be most contagious, enjoy the strongest implicit guarantee of all: the market believes they are too big to fail, which lets them borrow cheaply and grow larger still. The safety net, built to contain fragility, can quietly encourage the very risk-taking that produces it.
You cannot reliably distinguish illiquid from insolvent in the middle of a panic, and you must decide anyway. That single fact pushes lender-of-last-resort support toward bailouts, and bailouts toward moral hazard and "too big to fail." The third pillar — capital and liquidity regulation — exists largely to manage this: if banks are forced to hold more of their own capital at risk and more liquidity in reserve, they fail less often, and the agonizing illiquid-or-insolvent call has to be made less often.
The first two pillars clean up after fragility; the third tries to reduce it at the source. If banks fail less often and less badly, the insurer pays out less and the lender of last resort faces the impossible call less often. The global rulebook for this is the Basel framework, written by a committee of national supervisors that meets in Basel, Switzerland, and implemented (with local variation) by each country. It has evolved through successive versions — Basel I, II, and the post-2008 Basel III — each tightening the requirements after the previous regime proved inadequate.
The rules come in two families, matching the two ways a bank dies from Module 04. Capital requirements target insolvency: they force banks to fund themselves with more loss-absorbing equity, so larger losses can be sustained before the bank goes under. Crucially, capital is measured against risk-weighted assets — riskier assets require more capital behind them — with a simple leverage ratio as a backstop that ignores risk weights so the system cannot be gamed. Liquidity requirements, added after 2008 revealed that well-capitalized banks could still die of thirst, target illiquidity: they force banks to hold enough liquid assets and stable funding to survive a stress period without the central bank.
| Requirement | Targets | What it forces | Rough Basel III minimum |
|---|---|---|---|
| CET1 capital ratio | Insolvency | Highest-quality capital against risk-weighted assets | 4.5% + 2.5% buffer |
| Leverage ratio | Insolvency | Capital against total assets, ignoring risk weights (anti-gaming backstop) | 3% |
| Liquidity Coverage Ratio | Illiquidity | Enough high-quality liquid assets to survive ~30 days of stressed outflows | ≥ 100% |
| Net Stable Funding Ratio | Illiquidity | Enough stable funding to support assets over a one-year horizon | ≥ 100% |
Supervisors reinforce these ratios with stress tests — exercises that model how a bank's capital would hold up under a severe hypothetical recession or market shock, and that can force a bank to raise more capital or curb payouts if it would fall short. The details are technical and they shift over time, but the logic is simple and worth carrying: capital rules guard solvency, liquidity rules guard against runs, and both exist so that the expensive, hazard-prone first two pillars are called upon as rarely as possible.
The safety net is not one institution but a division of labor, and the division differs by country. Typically a central bank acts as lender of last resort and often as a supervisor; a deposit insurer guarantees deposits and resolves failed banks; and one or more supervisors write and enforce the capital and liquidity rules. Sometimes these are separate bodies, sometimes combined. Basel coordinates the rules internationally, but each jurisdiction implements and runs its own net.
The cross-country variation is large and consequential:
The safety net works. Developed economies suffer far fewer ruinous panics than they did before deposit insurance, lenders of last resort, and capital rules existed. But step back and ask what the net requires, and a consequence emerges that the next module is built around.
Every pillar of the net is something the state extends only to institutions it can trust and control. It will guarantee the deposits of — and lend as last resort to, and let hold an account at the central bank — only a bank that is chartered (licensed after intense scrutiny), capitalized to Basel standards, supervised continuously, and examined on its risk-taking. The protections and the obligations are a package: you cannot get the trust-conferring backstop without submitting to the control, and you cannot submit to the control without first being admitted to the club through a charter that is slow, costly, and often simply unavailable.
From the perspective of a newcomer — a fintech, a technology company, an innovator with a better idea — this is the heart of the problem. The very features that make a bank trustworthy enough to hold the public's money (the guarantee, the central-bank account, the regulated balance sheet) are gated behind a barrier that is extraordinarily hard to cross. You cannot easily offer a safe place to keep money, because the things that make it safe — insurance, a central-bank account, the lender of last resort — are reserved for chartered banks, and becoming a chartered bank is a years-long, capital-intensive, often-rejected undertaking. The safety net, in other words, does not just stabilize the incumbents; it protects them from competition, almost as a side effect.
That is the near-monopoly you asked us to keep in view. It is not the product of any single villain; it falls out of the logic of the safety net itself. Module 06 examines it directly — the charter, the central-bank account, the deposit-insurance moat — and looks at it from the fintech's vantage point, where conventional banking can look less like a competitive market and more like a government-protected oligopoly.
Six questions on the three pillars of the safety net, what each one targets, and the costs each carries. The questions test the reasoning rather than memorization of exact ratios.