Fractional-Reserve Banking

How fractional-reserve banking works

Walk into a bank and deposit $1,000 in cash. The bank's balance sheet now shows:

• Asset: $1,000 in vault cash (or reserves at the Fed).
• Liability: $1,000 owed to you, the depositor.

The bank does not need all $1,000 sitting around. On any given day, only a small fraction of depositors withdraw, so the bank can keep just enough on hand to satisfy normal redemptions and lend the rest. With a 10% reserve target, the bank keeps $100 in reserves and lends $900 to a borrower.

The borrower spends the $900 on a couch from a furniture store. The store deposits the $900 at its own bank, which keeps $90 in reserves and lends $810 to the next borrower. The cycle repeats. At each round, 90% of the inflow is re-lent, and a new deposit is created.

The total deposits eventually conjured from the original $1,000 are the sum of an infinite geometric series:

$1,000 + $900 + $810 + $729 + $656.10 + ... = $1,000 / 0.10 = $10,000

The original cash is still there, just $1,000 of it. But on the balance sheets of the banking system as a whole, $10,000 of deposits now exist. The extra $9,000 is bank-created money — not coined or printed by the central bank, but generated by the lending process.

Worked example with a 10% reserve ratio

Round	Bank receives	Reserves held (10%)	Loan made (90%)	Cumulative deposits
1	$1,000.00	$100.00	$900.00	$1,000.00
2	$900.00	$90.00	$810.00	$1,900.00
3	$810.00	$81.00	$729.00	$2,710.00
4	$729.00	$72.90	$656.10	$3,439.00
5	$656.10	$65.61	$590.49	$4,095.10
...	...	...	...	...
∞	—	$1,000.00	$9,000.00	$10,000.00

By round 30 the system has converged to within a penny of the $10,000 limit. The reserves total $1,000 — exactly the original cash injection. The loans total $9,000 — the bank-created money. In aggregate the system has not made anything from nothing; it has stretched the same underlying base of central-bank money across a much larger structure of deposit liabilities and loan assets.

Fractional reserve vs full reserve

	Fractional-reserve banking	Full-reserve (narrow) banking
Reserves against demand deposits	Fraction (3–10% historically; 0% today)	100%
Source of loan funding	Demand deposits + savings + equity	Term savings + equity only
Money multiplier	≈ 1/RR (theoretical 10× at 10%)	1× (no deposit-driven money creation)
Bank-run risk	Inherent to demand deposits	Eliminated by construction
Deposit insurance need	Essential (FDIC, FSCS, etc.)	Unnecessary for demand deposits
Cost of credit	Lower; deposits are cheap funding	Higher; loans require explicit savings
Central-bank monetary control	Indirect via reserves and rates	Direct over the entire money stock
Real-world adoption	Universal	None at national scale

Counterarguments and the full-reserve debate

The most enduring critique of fractional-reserve banking is that it allows commercial banks, not the central bank, to control the money supply. Every loan creates new deposits; every loan repaid extinguishes them. The central bank can influence this only indirectly — through rates, reserves, capital ratios. Three serious alternative proposals have been made over the past century:

• The Chicago Plan (1933). Eight Chicago economists, including Henry Simons, Frank Knight, and Paul Douglas, proposed 100% reserves against demand deposits during the Great Depression. Roosevelt opted for FDIC instead. Benes and Kumhof (IMF Working Paper 12/202, 2012) revisited it favorably with a modern macro model.
• Friedman's "A Monetary and Fiscal Framework" (1948). Argued for 100% reserves combined with a fixed money-growth rule, eliminating the discretion-and-instability of fractional-reserve banking.
• Iceland's Sigurjónsson Report (2015). Commissioned by the Prime Minister after Iceland's banking collapse, recommended sovereign money — the central bank as sole creator of new money — as a serious alternative. Not adopted but extensively debated.

Defenders of fractional-reserve banking — most mainstream economists, every functioning treasury — argue that the cost of full-reserve banking is severe. Demand deposits are the cheapest funding banks have. Forcing all loans to be funded by explicit savings would raise borrowing costs across the economy, contract credit, and slow growth. The post-2008 prudential framework (Basel III capital, Liquidity Coverage Ratio, Net Stable Funding Ratio, deposit insurance, central-bank lender-of-last-resort) is, in this view, the right way to manage fractional-reserve risks without giving up the system's productive properties.

The empirical question — does fractional-reserve banking systematically generate financial crises that full-reserve would prevent? — is harder than it looks. Crises happen in non-bank financial systems too (shadow banking, money-market funds, repo markets); the 2008 crisis featured all three. Whether removing fractional banking would shift the same instabilities elsewhere or eliminate them entirely is contested.

Variants and modern adjustments

• Capital-based regulation (Basel III). The binding constraint on lending in modern banks is risk-weighted capital, not the reserve ratio. A bank with a 10% Common Equity Tier 1 ratio can still lend aggressively; a bank with a 4% CET1 ratio cannot, regardless of reserves.
• Interest on reserves (IORB). Since 2008, the Fed pays banks interest on the reserves they hold. This made the reserve ratio operationally irrelevant — banks hold reserves voluntarily for the interest, not for legal reasons.
• Liquidity Coverage Ratio. Forces banks to hold enough High-Quality Liquid Assets to cover 30 days of stressed outflows. Effectively a modern, market-based version of a reserve requirement.
• Central-bank digital currencies (CBDCs). A retail CBDC would let citizens hold central-bank money directly. Full uptake would in effect collapse fractional-reserve banking by drawing deposits out of commercial banks.
• Stablecoins. Fully-reserved tokenized dollars (USDC, USDT) approximate narrow banking outside the regulated system, sidestepping fractional-reserve banking but raising new risks (issuer transparency, redemption guarantees).

Common pitfalls

• Confusing the textbook story with current US reality. The reserve ratio in the US has been 0% since March 2020. The constraint binding bank lending today is capital and liquidity, not legal reserves.
• Believing banks lend out depositors' money. Mechanically, a bank making a new loan creates a fresh deposit; it does not hand over a specific depositor's cash. The Bank of England's "Money creation in the modern economy" (2014) makes this point explicitly and from the horse's mouth.
• Treating the multiplier as automatic. The 1/RR ceiling assumes every loan dollar gets redeposited and re-lent. In reality, leakages — currency held outside banks, excess reserves, demand for loans — keep the realized multiplier well below the ceiling.
• Confusing "bank-created money" with counterfeiting. The new deposit is matched dollar-for-dollar by a loan owed to the bank. When the loan is repaid, the new money is destroyed. The system is reversible; counterfeiting is not.
• Overestimating the multiplier in a trap. When banks are unwilling or unable to lend (post-2008, parts of Japan), reserves pile up and the multiplier collapses toward 1. The textbook number is an upper bound, not a forecast.