DeFi Borrowing and Liquidations Explained: Health Factor and Collateral Risk
Decentralised lending protocols commonly allow eligible users to post digital assets as collateral and borrow other assets. That convenience does not make the borrowing safe: collateral values can fall, debt values can rise and automated liquidation conditions can be reached faster than a borrower can react.
This guide explains the health-factor arithmetic used by lending systems such as Aave, distinguishes loan-to-value from liquidation thresholds, and illustrates why leverage and oracle values can compound crypto-market risk. It is a technical and educational analysis, not instructions to borrow.
Key takeaways
- A lending position is defined by the value of eligible collateral, outstanding debt and protocol-specific risk parameters.
- Health factor falls when risk-adjusted collateral loses value relative to the debt.
- Health factor below one makes an Aave position eligible for liquidation under its current rules.
- Automated liquidation does not guarantee a fixed loss or orderly execution, especially when markets gap.
Essential concepts
Collateral
Assets supplied to support borrowing and accepted for this purpose by a protocol under defined eligibility rules. Some supplied assets may be disabled as collateral or subject to different risk weights.
Loan-to-value (LTV)
A ratio linking a permitted amount of debt to eligible collateral value under the market’s borrowing configuration. Initial borrowing capacity is not the same as the liquidation trigger.
Liquidation threshold
A protocol parameter establishing the risk-weighted collateral coverage below which liquidation can be permitted. Thresholds vary by collateral and can be updated through governance.
Health factor
A numerical ratio of liquidation-weighted collateral value to total debt value under the protocol’s valuation conventions. A ratio close to one indicates limited buffer; it is not a guarantee of a predictable liquidation price.
Liquidation bonus
An incentive paid in eligible collateral or value to a liquidator who repays a borrower’s qualifying debt. It can increase economic loss to the borrower beyond the change in collateral price.
Oracle valuation
A price reference provided through a particular data feed and aggregation arrangement. Oracle price and executable market price may differ under rapid or illiquid conditions.
Liquidator
An eligible participant or system performing a permitted repayment-and-collateral-transfer transaction. Operational competitiveness and available network block space affect whether liquidation transactions execute.
Isolation mode
A protocol feature restricting certain collateral or debt combinations to manage risk concentration. Exact restrictions depend on the version and chain, and can change.
Debt accrual
Growth of outstanding loan obligations through interest or other applicable charges. Debt can change even when the posted collateral price stays constant.
Collateral correlation
The extent to which collateral and borrowed assets tend to move together in price. Correlation can change under stress and is not equivalent to a contractual hedge.
Bad debt
An amount that cannot be covered fully through the protocol’s recovery or liquidation mechanisms. Insurance funds and safety modules, where present, have their own limits and conditions.
How DeFi health factor works in practice
How overcollateralised crypto lending works
A borrower supplies an eligible token to a lending protocol and receives another token as a debt obligation according to the pool’s rules. The initial borrowing capacity is constrained by the risk parameters and market valuation of the collateral. The borrower can experience changes in both sides of the balance sheet, especially when the debt is denominated in a volatile asset.
This arrangement is economically different from a mortgage with a human credit committee or a retail overdraft. Smart contracts may evaluate the position using updated oracle data and allow third parties to liquidate once conditions are satisfied. The specific privileges, market pause settings and insolvency treatment depend on the protocol.
The health factor formula and its limitations
Aave describes health factor as liquidation-weighted collateral value divided by outstanding debt value. With $10,000 collateral, an 80% liquidation threshold and $6,000 debt, the simplified health factor is 1.333. A value below one signifies liquidation eligibility, but the point of actual transaction execution may be affected by oracle updates and other transactions in the same block.
For multiple collateral assets the threshold is weighted, and debt may include interest and distinct denominated assets. A protocol can also have eMode, isolation or other risk configurations that alter parameters. A spreadsheet must not treat every collateral token as though it always had the same threshold.
LTV is not a promise about liquidation levels
Borrowing at 50% LTV does not mean assets can fall 50% before liquidation. A market may permit borrowing up to one parameter while liquidation uses a separate, typically higher, threshold. The remaining buffer depends on both collateral and debt prices, the relevant feed and accumulated debt.
Suppose a token used as collateral is denominated in ETH while the debt is a dollar stablecoin. A rapid ETH decline lowers the dollar value of collateral, reducing health factor. The same borrower taking debt in another crypto asset may experience a different risk because both the numerator and denominator change.
How third-party liquidations work
Once an eligible position becomes liquidatable under the contract, a third party may repay some debt and claim collateral under protocol rules. The liquidation incentive compensates for gas, execution and adverse price risk. The precise amount that can be repaid depends on version, market, debt size and protocol parameters and can be less or more than a simple fixed percentage.
A borrower can lose part of the collateral, while the outstanding debt may also be reduced. This is not identical to a market order closing an exchange margin trade. The procedure occurs through smart contracts, with oracle values and blockchain inclusion affecting the exact sequence.
Risk from oracle prices and congestion
The contract does not directly observe a universal executable market price. It uses data feeds designed to reflect specified market information. If an oracle is delayed, malfunctioning or materially different from available trading quotes, liquidation triggers can behave unexpectedly. Some protocols include safeguards, but these require verification rather than assumption.
Network congestion can increase gas costs and delay repayment or collateral-top-up transactions. A borrower may see a deteriorating health factor and still be unable to transact in time. Automated monitoring can help operationally but cannot remove gap risk or guarantee the availability of liquidity.
Why leverage can magnify crypto drawdowns
Borrowed proceeds may be used to obtain further exposure, increasing the ratio of gross assets to equity. A 20% decline in the collateral token can therefore reduce the borrower’s net equity by materially more than 20%, depending on debt and other positions. Repeat borrowing across protocols can create linked liquidation conditions that intensify selling pressure.
On-chain liquidations can interact with decentralised exchange liquidity. Large liquidation swaps may move prices, reducing collateral values elsewhere. Describing a position as ‘overcollateralised’ at initiation does not mean it will remain solvent through a stressed market.
A responsible lending-risk worksheet
List collateral quantity, token identifiers, oracle source and timestamp, applicable threshold, debt quantity, debt rate, current health factor and costs of repayment or withdrawal. Model a fall in collateral price, a rise in debt price and interest accrual over time. Confirm that the borrowing market permits each token as collateral under the exact network and version.
The research note should also check governance changes, protocol reserves, liquidity available for repayment, transaction fees and smart-contract upgrade controls. Never assume that increasing collateral will always be possible, that a price feed will update on schedule or that a stablecoin remains at parity during stress.
Worked numerical example
A fictional position has $12,000 collateral, an 80% liquidation threshold, and $7,000 debt. Its simplified health factor is 12,000 × 0.80 ÷ 7,000 = 1.3714. If the collateral’s dollar market value falls 20% to $9,600 and debt remains constant, health factor becomes 9,600 × 0.80 ÷ 7,000 ≈ 1.0971. If collateral falls further to $8,500, health factor becomes 0.9714, below one in this teaching model. Aave’s actual liquidation quantities and bonus depend on the specific market rules; the example does not forecast the collateral price or exact transaction cost.
Branded comparison table
| Measure | Illustrative value | Why it matters |
|---|---|---|
| Collateral value | $12,000 | Assets exposed to price changes |
| Liquidation threshold | 80% | Risk-weighted collateral |
| Borrow value | $7,000 | Debt denominator |
| Initial health factor | 1.3714 | Buffer above one |
| After collateral falls to $8,500 | 0.9714 | Eligibility threshold crossed |
Research checklist and limitations
Record the exact chain, contract address, asset denomination, timestamp, data source and protocol version. Distinguish prices observable in a market from oracle marks, model output and interface estimates. Verify the transaction or contract terms rather than inferring behaviour from a headline. If a protocol is upgradeable, its previous settings may not describe its current or future operation.
Use scenario analysis that allows market prices, liquidity and network performance to deteriorate together. Include network charges, slippage, debt accrual, settlement timing, counterparty dependencies and any liquidation penalties or fee incentives. A single successful historical transaction does not demonstrate that the same process works under stress. Explain what can be learned from public on-chain records and what cannot be inferred without independent evidence.
Finally, separate technological correctness from commercial outcomes. A smart contract can execute precisely according to its source code yet produce an unfavourable result because of oracle updates, collateral terms, transaction ordering or a user’s assumptions. This article aims to improve informed risk analysis, not to direct readers to leverage, exploit, or trade a cryptocurrency.
Frequently asked questions
Does a health factor above one make a DeFi loan safe?
No. It describes a modelled buffer at the time of observation. Volatility, oracle updates, debt growth and execution delays can reduce it rapidly.
Is LTV the same as the liquidation threshold?
No. The initial borrowing cap and the point at which liquidation can occur are different protocol settings.
Can borrowing against a stablecoin remove liquidation risk?
No. Depegging, debt denomination, interest, market liquidity and protocol risks remain relevant.
Why can a small further price fall cause a large loss?
Leverage magnifies changes in net equity, and liquidation penalties or selling pressure may add to the effect.
Related Traders MBA guides
DeFi Lending and Pools; Crypto Liquidation Cascades; Stablecoin Peg and Reserve Risk. Explore the Cryptocurrency learning category.
Sources and further reading
Important notice: General technology and financial-market education only, not personalised investment advice. Numerical examples are hypothetical. Crypto-assets and derivative or lending protocols may result in severe losses, liquidation, loss of access or contract failures.