Lending, borrowing and liquidations

Credit functions without a conventional loan desk

Lending and borrowing systems make assets available for a period in exchange for an obligation to repay or return value. In conventional finance, a lender usually evaluates a borrower and records a contractual claim. In decentralized finance (DeFi), a protocol may instead administer deposits, collateral, interest calculations, and enforcement through smart contracts. The fact that software runs the process does not mean economic risks disappear. It changes which decisions are automated, who has discretion, and what evidence a reviewer can inspect.

In this module, collateral means assets committed to support a debt. Overcollateralization means that the protocol requires collateral with a measured value greater than the outstanding borrowing. The extra margin is intended to absorb market movements, fees, and delays. It does not ensure that a borrower or lender can never incur a loss. Some systems use pooled liquidity, some match counterparties, and others involve specialized credit arrangements. It would be inaccurate to describe all DeFi borrowing with a single formula.

How pooled lending works

Depositors, borrowers, and utilization

In a pool-based lending arrangement, suppliers contribute eligible assets to a common pool. Borrowers receive assets from the pool subject to the protocol's requirements. A smart contract tracks supplied balances, debt balances, and rates based on rules in force. The utilization rate is a measure of how much available liquidity is borrowed. If most liquid assets are lent out, immediate withdrawals may be constrained even when a depositor's accounting claim remains visible.

Some protocols adjust variable interest rates as utilization changes. A high utilization level can trigger higher borrowing rates under a predefined schedule, encouraging more supply or debt repayment. It is not a guarantee that either response will occur. Interest may accrue continuously or at discrete accounting intervals according to the implementation. A quoted annual percentage rate is an annualized description of a particular rate assumption, not necessarily a locked contractual return. Fees, changing rates, token values, and interruptions can make actual outcomes differ.

An accounting token representing a supplied position may reflect a claim on protocol assets, but its transferability and redeemability depend on code and rules. It should not automatically be described as a bank deposit, a legally guaranteed balance, or immediate cash. Legal characterization may require separate review. This course deliberately distinguishes economic description from legal conclusions.

Collateral and the health of a position

Many protocols use a collateral ratio, loan-to-value ratio (LTV), or health factor to monitor safety. LTV generally compares debt value with eligible collateral value, after specified valuation conventions. A health factor may incorporate protocol-specific liquidation thresholds; its exact formula is not universal. A collateral asset can fall in market value while the amount owed remains unchanged, increasing the ratio of debt to collateral and narrowing the margin of safety.

For example, assume a position carries 150 units of collateral measured at a total market value of $150 and 75 units of debt valued at $75. Its simple LTV is 50%. If the collateral's measured value falls to $100 while debt remains $75, LTV rises to 75%. Whether that position may be liquidated depends on the applicable threshold, asset-specific parameters, fees, price sources, and state at the moment of evaluation. This example is arithmetic for conceptual learning, not an instruction to open a position or a forecast.

Liquidation is an automated response to a threshold

Trigger, executor, and outcome

Liquidation is the process in which some or all of a borrower's collateral may be sold, transferred, or claimed under protocol rules when a condition is met. It is not necessarily a court-supervised insolvency proceeding. In some designs, third-party liquidators monitor positions and submit transactions to repay debt in exchange for collateral plus an incentive. In others, auctions or different mechanisms are used. The key distinction is between an automated contract condition and a legal determination about a person's broader finances.

A liquidation threshold expresses when a position becomes eligible for an action; it does not guarantee that action occurs instantly. Someone or something may need to submit an eligible transaction, the network must process it, and available markets must absorb the collateral. During rapid market moves, block congestion, or disruptions to price feeds, execution can be delayed or produce unexpected outcomes. The remaining debt or collateral depends on protocol rules and liquidity conditions.

A liquidation incentive is designed to compensate parties for executing an eligible action and taking execution risk. If collateral values fall too quickly or markets lack depth, liquidation proceeds can fail to cover the debt. A shortfall may be described as bad debt and can affect suppliers or protocol backstops. Reserve funds, insurance-style mechanisms, and governance actions might absorb losses in some designs, but their capacity and legal meaning are not uniform.

The price feed is part of credit risk

Collateral systems must decide what assets are worth. An oracle is a mechanism that supplies data from outside a smart contract's own state, such as an asset price. A price feed can lag the market, fail temporarily, use thinly traded references, or be manipulated. If an oracle reports an incorrect price, a position could appear healthier or riskier than economic conditions justify. The exposure is not merely a software defect; valuation choice and market quality are substantive risk decisions.

Multiple data sources, bounded update intervals, independent controls, and emergency procedures can reduce some failures. They do not remove the possibility that several sources share the same weakness. A professional review should identify the oracle, accepted collateral, reference assets, liquidation assumptions, fallback behavior, and governance powers to change parameters. The next lesson examines oracles as an infrastructure dependency more broadly.

Interconnected positions and systemic effects

Some participants borrow against one asset, acquire another, and place it into a separate protocol. Those positions can connect pools, trading venues, bridges, and price feeds. A decline in one asset may trigger collateral sales; those sales may depress prices and cause additional positions to breach thresholds. This reinforcing pattern is sometimes called a liquidation cascade. It is a possible stress mechanism, not an assertion that all market declines cause one.

A lending pool's apparent total value locked (TVL), meaning assets accounted for in deployed contracts, can be misleading as a measure of independent economic capital. Assets may be borrowed, moved, and represented more than once across protocols. Treating aggregated balances as entirely distinct could overstate the economic resources available to absorb losses. Similarly, the presence of collateral does not by itself prove that it can be sold quickly at the observed reference price.

Legal and governance boundaries

The legal treatment of lending depends on the activity, asset, counterparties, jurisdiction, and contractual framework. Some crypto-asset regimes do not list lending as a standalone licensed service; that narrow point should not be generalized into an exemption from all financial, consumer, insolvency, or other laws. A service may perform several activities, and other applicable regimes can matter.

FATF Recommendation 15 calls for risk assessment and appropriate controls for new technologies and relevant virtual-asset service activities. FATF's 2021 guidance assesses certain DeFi arrangements by actual involvement or control rather than relying on a marketing label. IOSCO's Policy Recommendations for Decentralized Finance (DeFi) (December 2023) address responsible persons, conflicts, disclosure, and risk management. These documents are relevant reference points; they are not identical to enforceable laws in every country.

For public-sector and legal professionals, the questions are practical: Which contracts establish obligations? Which asset secures the debt? What triggers seizure or sale? Who sets parameters? Are there administrators who can suspend actions? What records show the actual events? A transaction log can reveal a liquidation, but may not explain every legal entitlement or the complete financial position of an affected party.

Worked example: interpreting a liquidation record

An observer sees a debt repayment event, a transfer of collateral to another address, and a fee paid to execute a transaction. Those facts support an initial hypothesis that a liquidation occurred. Confirmation requires checking the called contract, event definitions, the affected debt position, collateral valuations, and the applicable protocol version. It would be unsound to infer that the borrower voluntarily sold the collateral or that the receiving address was the original lender. A liquidation mechanism can distribute assets according to code without a conventional buyer-seller agreement.

Exam focus: Define collateral, LTV, utilization, liquidation threshold, health factor, and bad debt. Explain why automated enforcement depends on price data, liquidity, and network execution. Recognize that collateralization reduces some exposures but does not eliminate smart-contract, liquidity, or legal risk.