Fundamentals

Fractional reserve banking: what it is and why it is debated

By Daniel Sardá · Published on August 31, 2026

9 min read1,902 words

Fractional reserve banking allows demand deposits to coexist with loans and other assets. This structure facilitates credit, but it requires liquidity, capital, and rules capable of containing its risks.

Fractional reserve banking is a banking arrangement in which deposits that can be withdrawn on demand are not backed one-for-one by cash or central bank reserves. A bank holds resources to process payments and withdrawals, but it also holds loans, securities, and other assets.

This creates an apparent paradox: customers see an available balance even though the bank does not keep an equivalent amount of banknotes in a vault. This does not mean that the same money has been lent twice. It means that a deposit is a debt the bank owes its customer and that the institution holds a diverse portfolio of assets against that debt.

Key idea: An account balance is a bank liability owed to the depositor; reserves are assets the bank uses to settle payments and meet obligations. They are not the same thing.

What lies behind an account balance

For an individual, a deposit of 100 monetary units is an asset: the customer can use it to make payments or request a withdrawal under the account's terms. For the bank, that same deposit is a liability because it represents an obligation to the customer.

On the other side of the balance sheet are the institution's assets: loans it expects to collect, securities, central bank reserves, vault cash, and other resources. The bank earns income partly because some of these assets yield interest, while it must provide depositors with access to funds and other services.

This distinction prevents confusion among three elements:

• Deposits: balances that commercial banks owe their customers.
• Reserves: balances that banks hold at the central bank and use, among other purposes, to settle payments between institutions.
• Cash: banknotes and coins held by the public or in bank vaults.

Not every liquid asset is a reserve, either. A bank may own securities that it can sell relatively quickly, but that does not make those securities central bank reserves. As the Basel Committee explains, modern liquidity regulation considers a broader set of high-quality liquid assets and the outflows expected under stress.

A loan of 100, viewed through the balance sheet

Suppose a bank approves a bank loan of 100 monetary units and credits the money to the borrower's account. At the outset, it records two entries:

• +100 in assets: the claim to repayment of the loan.
• +100 in liabilities: the new deposit available to the borrower.

The bank does not need to transfer a particular person's deposit into that account: it creates a loan and a deposit simultaneously. This mechanism is documented by the Bank of England and the Bundesbank.

What happens if the borrower uses the 100 units to pay someone who banks elsewhere? The borrower's balance falls, the recipient's deposit rises, and the originating bank must settle the transaction. It normally does so by transferring reserves to the other institution. If it does not have sufficient reserves at that moment, it must obtain them through incoming payments, market funding, asset sales, or access to central bank facilities, depending on the applicable rules.

Creating a deposit therefore does not eliminate the need for funding and liquidity. It simply clarifies the sequence: the loan entry can originate the deposit, while subsequent payments create settlement needs.

When the borrower repays the principal with deposited money, the reverse occurs: the loan on the asset side and the deposit on the liability side both decrease. The bank money created when the loan originated is destroyed as the principal is repaid.

Useful distinction: A bank can create the deposit when it makes a loan, but it cannot guarantee that the deposit will remain there. When the money is transferred, the institution must be able to settle the payment.

Why a bank cannot lend without limit

This is not an unlimited license to create money. A bank must find creditworthy customers willing to borrow, assess repayment prospects, and cover costs and risks. It also faces constraints involving capital, liquidity, regulation, funding, and monetary policy.

Bank capital requires clarification. It is not a box of cash set aside to pay withdrawals. It is the residual difference that absorbs losses when asset values fall. The more risk a bank takes, the greater the possibility that defaulted loans will erode this cushion.

A reserve requirement may also exist: a legal obligation to hold reserves relative to certain liabilities. But there is no universal positive percentage. The design varies across countries and over time. For example, the U.S. Federal Reserve has kept its reserve requirement ratios at 0% since March 26, 2020, while the Eurosystem calculates minimum reserves on certain liabilities.

A statutory ratio of 0% does not mean that banks operate without reserves, capital, or liquidity rules. It means only that this particular requirement imposes no positive minimum. Institutions still need resources to make payments and may hold reserves for operational or prudential reasons.

The money multiplier: an intuition, not a machine

Many textbooks present a familiar example: if the reserve requirement is 10%, an initial deposit could support a chain of loans and redeposits reaching a multiple of the initial reserve. The formula `1/r`, where `r` is the reserve ratio, is useful for illustrating a relationship under very specific assumptions.

Those assumptions are demanding: every bank must lend as much as permitted, the public must redeposit all the money, the ratio must remain fixed, and no other constraint can be decisive. In practice, banks do not mechanically wait to receive reserves and then multiply them by a predetermined ratio. They assess lending opportunities and then manage capital, funding, liquidity, and settlement within the prevailing institutional framework.

The multiplier is useful as a teaching model, not as a literal description or automatic prediction. Nor does an increase in deposits by itself imply higher inflation: spending, demand, productive capacity, loan repayment, and monetary policy all matter.

Liquidity, solvency, and the risk of a bank run

Fractional reserve banking combines liabilities that may be claimed quickly with assets that mature later or cannot be sold immediately without a loss. This transformation helps finance projects and provide readily accessible accounts, but it creates a vulnerability: many depositors may demand their money before the bank collects its loans.

It is essential to separate two questions:

• Liquidity: Can the bank meet its payments now?
• Solvency: Are its assets worth more than its debts, with enough capital to absorb losses?

A bank can be economically solvent yet temporarily lack the means of payment. It can also have enough cash today but be insolvent because its assets have permanently lost value. In a crisis, the two problems can reinforce each other: doubts about solvency accelerate withdrawals, while urgent asset sales to raise liquidity can crystallize losses.

A bank run occurs when many depositors try to withdraw or transfer their funds within a short period. The classic Diamond–Dybvig model shows why demandable liabilities that fund illiquid assets can produce fragility even when intermediation provides an economic service. But not every run is a baseless panic: some respond to real losses, mismanagement, or credible information about the institution.

Warning: Liquidity is not the same as solvency. Emergency lending can buy time; it cannot make permanent asset losses disappear.

What safeguards surround the system?

Modern institutions combine several defenses because each addresses a different risk.

Liquidity management seeks to ensure that a bank can withstand expected outflows and stress scenarios. The international standard known as the LCR, for example, requires institutions within its scope to hold enough high-quality liquid assets to cover net outflows over 30 days of stress. It is not a promise of immunity but a prudential barrier.

Capital requirements compel owners and investors to provide a cushion against losses. Supervision seeks to detect concentrations, poor valuations, and unsafe practices. Resolution mechanisms determine how to intervene in a troubled institution without treating all its obligations alike.

Deposit insurance can discourage runs by protecting balances up to specified limits. And the central bank can act as lender of last resort, providing liquidity under defined conditions. None of these safeguards is free or infallible.

If depositors, managers, or creditors expect the state to absorb every loss, their incentive to monitor risk may weaken: this is the problem of moral hazard. Coverage limits, risk-adjusted premiums, equity capital, supervision, and credible resolution procedures seek to allocate that responsibility more effectively, although their success depends on design and enforcement.

From a liberal perspective, what matters is knowing what the bank promised, who bears the losses, and whether the rules are general, known, and applied without privilege. A clear contract should distinguish a deposit account—a demandable debt—from a custody service. Hiding that distinction would be misleading; acknowledging it does not eliminate the risks.

What would change under 100% reserve banking?

Proposals for 100% reserve banking require full backing for deposits covered by the regime. Under a strict version, money intended for payments would be separated from credit intermediation. Loans would have to be financed with capital, term debt, investment funds, or other liabilities that do not promise immediate availability on the same terms.

The appeal is easy to understand: if every covered deposit is backed by reserves, the liquidity mismatch within those accounts disappears. The difference between storing money and financing credit also becomes more visible. Advocates argue that this could strengthen contractual discipline and reduce certain types of bank run.

The outcome, however, depends on which deposits are covered, how the transition is financed, and where intermediation moves. Eliminating the mismatch in one category does not prevent credit losses or runs on other short-term debt. Nor does it by itself resolve how credit will be allocated or what public backing the remaining institutions will receive. Studies such as The Chicago Plan Revisited explore scenarios of this kind through models; they are neither definitive proof of real-world effects nor a conclusive institutional position of the IMF.

To assess the debate: The decisive question is not only how many reserves exist, but also what contract is offered, how assets are financed, who absorbs losses, and what incentives public guarantees create.

An architecture of commitments, not a magic percentage

Fractional reserve banking is best understood as a balance-sheet architecture: demandable deposits on the liability side, loans and other assets with different maturities and risks, and liquid resources available to meet payments. Its coordinating function is to connect saving, payment services, and credit. Its fragility comes from promising availability while some assets cannot be converted into money immediately without cost.

Reducing the topic to “holding a fraction” conceals too much. The reserve requirement is only one element, and in some systems it is not even the main operational constraint. Capital, liquidity, credit quality, contracts, competition, supervision, and bankruptcy or resolution mechanisms determine who takes risks and who is responsible when something goes wrong.

The serious debate therefore does not end with a choice between fractional and full reserves. It begins by making obligations and their limits visible. Market discipline requires comprehensible information and credible losses for those who chose to bear the risk; public rules, where they exist, should prevent protection from becoming an invitation to act recklessly. This institutional combination, more than any isolated formula, defines the system's stability and legitimacy.