How Banks Work and How They Create Money

By Pritesh Yadav 11 min read

Most people think a bank is a safe box. You put your money in, it sits there, and the bank lends "some of it" to other people. That picture is comforting, simple, and almost entirely wrong. The truth is stranger and far more powerful: banks create most of the money in the economy, out of almost nothing, every time they make a loan. Understanding how that works is the single biggest leap in understanding modern money. Let's build it up carefully.

First, what a commercial bank actually is

Commercial bank
A business that takes deposits from the public and makes loans. Your high-street bank — Chase, HSBC, SBI — is a commercial bank.
Central bank
The "bank to the banks" — the US Federal Reserve, the Bank of England, the Reserve Bank of India. It issues the most basic form of money and sets interest-rate policy. It does not hold your personal account.
Deposit
Money you keep in a bank account. Crucially, it is the bank's liability — a promise to pay you back on demand. It is not the bank's money; it owes it to you.
Loan
Money a bank lends to a borrower. To the bank this is an asset — a promise from the borrower to pay the bank back.

Hold onto that last pair, because students get it backwards: your deposit is the bank's liability; the loan is the bank's asset. The bank owes you; the borrower owes the bank.

Two kinds of money: base money vs broad money

There are really two layers of money, and confusing them is the root of most confusion about banking.

TypeWhat it isWho creates it
Base money (also "M0", "central bank money", "high-powered money")Physical cash + the reserve balances banks hold at the central bankOnly the central bank
Broad money (M1/M2/M3, UK's M4)Cash in your pocket + all the deposits the public holds at commercial banksMostly commercial banks, through lending

Here is the figure to remember above all others. The Bank of England, in 2014, pointed out that in the UK roughly 97% of the money people actually use is commercial-bank deposit money, and only about 3% is physical cash. Almost all "money" is just numbers in bank computers — created by banks, not printed by the government.

Key takeaway: The notes in your wallet are central-bank money. The far larger balance in your bank app is commercial-bank money. The economy runs mostly on the second kind.

The big idea: lending CREATES money

The textbook story says a bank gathers savers' deposits and then "lends them out" to borrowers — a middleman, or financial intermediary, shovelling existing money from one person to another. That's a useful starting picture, but the Bank of England's landmark 2014 paper, "Money creation in the modern economy," says plainly that it is backwards.

What really happens: "Whenever a bank makes a loan, it simultaneously creates a matching deposit in the borrower's bank account, thereby creating new money." The bank does not dig into a vault. It types a number into your account. The loan (an asset) and the new deposit (a liability) pop into existence on the bank's balance sheet at the very same instant, with a keystroke.

Analogy: A regular shop can only sell apples it already owns. A bank is more like a referee who can create the points: when it grants you a £200,000 mortgage, it doesn't move £200,000 from a saver's account — it writes "£200,000" into your account and "£200,000 loan owed to us" on its own books. New money exists that didn't a second ago.

Follow the ripple. You borrow to buy a house. The bank credits your account — broad money in the economy has just risen. You pay the seller; the deposit moves to the seller's bank. Between the two banks, they settle using reserves at the central bank. But notice: the total amount of deposits in the economy didn't change when you spent the money — it only changed when the loan was created. And it falls again when loans are repaid: paying back a loan destroys that money.

Key takeaway: Loans create deposits, not the other way round. New lending expands the money supply; repaying loans shrinks it.

Fractional reserve banking and the money multiplier

Fractional reserve banking means a bank keeps only a fraction of deposits as ready cash (reserves) and puts the rest to work in loans. The reserve ratio (r) = reserves ÷ deposits.

The classic teaching example: imagine r = 10%. Someone deposits $100. The bank keeps $10 and lends $90. That $90 gets spent, lands in another bank, which keeps $9 and lends $81 — and so on. Add up the infinite chain and you get $1,000 of deposits built on $100 of base money. The money multiplier is 1 ÷ r = 1 ÷ 0.10 = 10.

 THE TEXTBOOK MULTIPLIER (r = 10%)
   $100 deposit ──► keep $10, lend $90
                         │
                         ▼
   $90 redeposited ─► keep $9,  lend $81
                         │
                         ▼
   $81 redeposited ─► keep $8.10, lend $72.90 ...
   ─────────────────────────────────────────────
   Total deposits = 100 + 90 + 81 + ... = $1,000
   Multiplier = 1 / 0.10 = 10
Common mistake: Believing this is how money is really made today. The Bank of England says the multiplier runs the wrong way in practice. Banks are not reserve-constrained moment to moment — they lend first (limited by profitable opportunities, borrower demand, capital rules, and the interest rate the central bank sets) and then get whatever reserves they need afterwards. Reserves don't get "multiplied up." Tellingly, the US cut its reserve requirement to 0% in March 2020 — and banks kept lending. Treat the multiplier as a useful intuition pump, not a literal lever.

So what really limits lending? Three things: capital requirements (regulators force banks to hold a buffer of their own money against losses), loan demand and profitability (banks lend when it's worth it), and the central bank's interest rate (cheaper rates encourage more borrowing). Central banks today steer the economy by setting the price of money, not by rationing a fixed quantity of reserves.

The balance sheet — see it once and it clicks

 BANK BALANCE SHEET (after making a $100 loan)
 ┌─────────────────────────┬─────────────────────────┐
 │        ASSETS           │       LIABILITIES        │
 ├─────────────────────────┼─────────────────────────┤
 │  Loan to customer  $100 │  Customer deposit  $100  │ ◄ both created
 │  Reserves at CB    $ 20 │  (depositors' money)     │   at the same moment
 │  Govt bonds        $ 80 │  Equity (capital)  $100  │ ◄ loss cushion
 └─────────────────────────┴─────────────────────────┘
   Long-term / illiquid       Short-term / on-demand   ◄ MATURITY MISMATCH

Read it left to right. On the left (assets) the bank holds long-lived, hard-to-sell things: loans, government bonds, and a liquidity cushion of reserves. On the right (liabilities) it owes depositors money payable instantly. The equity / capital is the bank's own money — the buffer that absorbs losses before depositors are hurt (set by international "Basel" rules). Reserves are the liquidity buffer; capital is the solvency buffer. Different jobs.

Why banks are fragile: maturity transformation

Maturity transformation is the core trick of banking — and its core danger. Banks fund long-term, illiquid assets (a 25-year mortgage, a long government bond) with short-term, on-demand liabilities (deposits you can withdraw this second). That mismatch is exactly what makes banks economically valuable: they let society fund long projects while savers keep instant access to cash. But it is also why banks can collapse.

Analogy: Imagine borrowing money that's repayable today to buy a house you can only sell in ten years. Perfectly fine — unless every lender shows up demanding their money back at once.

This creates a critical distinction: a bank can be solvent but illiquid. On paper its assets are worth more than it owes — it isn't bankrupt. But it cannot turn those long-term assets into cash fast enough to satisfy a sudden flood of withdrawals. Forced to dump assets at fire-sale prices, a solvent bank can be pushed into genuine insolvency.

Bank runs and the Diamond–Dybvig insight

A bank run is when depositors collectively panic and rush to withdraw. Because withdrawals are first-come-first-served, the rational move — even if you trust the bank — is to run first, before the cash runs out. Everyone reasoning this way causes the very collapse they feared.

Economists Douglas Diamond and Philip Dybvig modelled this in 1983 (work that won them, with Ben Bernanke, the 2022 Nobel Prize in Economics). Their key result: a deposit-taking bank has two possible equilibria — a "good" one where only people who genuinely need cash withdraw, and a "bad" one where everyone panics. The bad one is a self-fulfilling prophecy: a perfectly sound bank can be destroyed purely because depositors fear a run.

Case study — the Great Depression: Between 1930 and 1933, waves of runs swept the United States. Roughly 9,000 banks failed, wiping out around 9 million savings accounts. This was the founding trauma that produced deposit insurance.
Case study — Silicon Valley Bank, March 2023: SVB poured a huge share of its assets into long-dated US Treasuries and mortgage bonds. When the Fed raised interest rates sharply through 2022 to fight inflation, the market value of those bonds fell, creating large unrealized losses. Note: these were safe government bonds — the failure was interest-rate and maturity-mismatch risk, not bad loans. Worse, about 88% of SVB's deposits were uninsured (above the $250,000 cap) and concentrated among tech start-ups who all talk to each other. On March 9, 2023, depositors tried to pull roughly $42 billion in a single day — about a quarter of all deposits — by tapping their phones. Regulators closed it on March 10: the second-largest US bank failure in history, and the first true "smartphone-and-Twitter" run.
Common mistake: Assuming a failed bank must have made reckless loans. SVB shows a bank can die from interest-rate risk on safe bonds plus a liquidity panic — with essentially no loan defaults at all.

The standard fix: deposit insurance

If runs are caused by fear, the cure is to remove the fear. Deposit insurance guarantees small depositors will be paid back even if the bank fails, so they have no reason to run. In Diamond–Dybvig terms, it eliminates the "bad equilibrium" for covered savers.

The US created the FDIC (Federal Deposit Insurance Corporation) in the Banking Act of 1933 (the Glass–Steagall Act), signed by President Roosevelt on June 16, 1933; coverage began January 1, 1934. The same law separated ordinary banking from risky investment banking (a separation repealed in 1999).

Country / bodyInsured limit (per depositor, per bank)
USA — FDIC$250,000 (raised from $100,000 in the 2008 crisis; made permanent in 2010)
UK — FSCS£85,000
EU schemes€100,000
India — DICGC₹5 lakh (raised from ₹1 lakh in 2020)

The funds come first from premiums the banks themselves pay into an insurance pool, not directly from taxpayers. The effect is profound: by promising small savers their money is safe, you stop the panic before it starts.

Common mistake (and a real debate): Insurance isn't free of side effects. If banks know they're backstopped, they may take bigger risks — this is moral hazard. And SVB reopened an old argument: with so many deposits sitting above the $250,000 cap, is the cap too low, leaving big depositors flight-prone? After SVB, US regulators invoked a "systemic-risk exception" to guarantee all its deposits — protecting even the uninsured, which critics say rewards exactly that risk-taking.
Key takeaway: Banking's value and its fragility come from the same source — borrowing short and lending long. Deposit insurance tames the panic, but trades away some discipline in return.

Key Takeaways

  • Your deposit is the bank's liability (it owes you); the loan is the bank's asset (the borrower owes it). Students often get this backwards.
  • Banks create money by lending: a loan and a matching new deposit appear together with a keystroke. Repaying loans destroys money.
  • About 97% of usable money is commercial-bank deposit money; only ~3% is physical cash.
  • The money multiplier (max money = base ÷ reserve ratio) is a teaching heuristic, not the real mechanism — lending is limited by capital, demand, and the central bank's interest rate.
  • Maturity transformation — funding long, illiquid assets with instant-access deposits — makes banks valuable and fragile; a bank can be solvent yet illiquid.
  • Bank runs are self-fulfilling (Diamond–Dybvig, 1983; Nobel 2022); SVB in 2023 showed a digital run can drain a quarter of deposits in a day.
  • Deposit insurance ($250k US, £85k UK, ₹5 lakh India) stops runs by removing fear — at the cost of some moral hazard.

Continue reading