How an overcollateralized crypto loan works
The principle is related to the guide explaining how crypto debt can arise. The borrower deposits an asset, such as ETH, and receives a smaller amount in stablecoins. The protocol maintains a buffer so it can sell the collateral if its value falls.
Loan-to-value, or LTV, compares the debt with the value of the collateral. With $10,000 in ETH and $5,000 borrowed, the LTV is 50%. The protocol also sets a liquidation threshold, generally above the maximum opening LTV. The gap between the two provides a safety margin.
Interest accrues over time. As a result, the debt can increase even when the collateral’s price remains unchanged. The rate often varies according to pool utilization: as demand for loans approaches available liquidity, borrowing costs can rise.
Posting collateral does not guarantee that you will retain the asset. Once the position crosses the threshold, a liquidator repays part of the debt and receives collateral along with a bonus. This automated execution protects the protocol, not the borrower’s purchase price.
LTV, liquidation threshold and health factor
Maximum LTV indicates the initial borrowing capacity. The liquidation threshold determines when a position becomes eligible for a sale. Confusing these two parameters can lead borrowers to take on debt too close to the limit.
Aave notably uses a health factor calculated from the collateral value, the weighted threshold and the debt. According to the official Aave liquidation documentation, a factor below 1 makes the position liquidatable. The platform points out that there is no universally safe level, because asset volatility and correlation change the risk.
A position with a factor of 1.05 has only a narrow buffer. A sharp decline, an increase in debt or a change in the oracle price can push it below 1 before the user has time to react. Adding collateral after an alert also requires time, gas and a functioning network.
The dashboard should display the price threshold that would trigger liquidation. If the interface does not provide it, calculate it using the protocol’s parameters and verify the result with a second source. Do not base a decision on a dollar-denominated rounded figure.
The total cost goes beyond the advertised rate
A variable annual rate is not a fixed promise. It can change block by block. Borrowers must include origination fees, deposit, borrowing, repayment and withdrawal transactions, as well as the spread from any conversion.
Yield paid on the collateral can offset part of the cost, but it may expose users to another token. Incentive programs add rewards whose price can fall. Compare interest in the debt currency, not only by the number of tokens received.
A liquidation adds a penalty, or a bonus paid to the liquidator. The borrower then loses more collateral than the debt amount repaid. Potential tax costs also depend on the country: a forced disposal of an asset may create a taxable event.
For comparison purposes, build three scenarios: planned repayment, a higher interest rate and a sharp decline in the collateral. Add Ethereum gas costs if the protocol operates on Ethereum. A small position can become unprofitable when transactions are expensive.
Why liquidations happen so quickly
Liquidators use bots to monitor positions. As soon as an oracle pushes the factor below the threshold, they submit a transaction with competitive fees. Users are not necessarily given a grace period.
A brief price wick can be enough. The protocol follows its oracle, not the average price later seen on a mainstream app. If the collateral market lacks depth, a forced sale can intensify the decline.
Correlation does not eliminate risk. Borrowing a stablecoin against another can sometimes reduce apparent volatility, but a depeg can widen the gap. Liquid-staking collateral can diverge from the underlying asset at the worst possible time.
Parameters can change through governance. A more conservative threshold, a cap or an asset freeze can alter how the position must be managed. Check announcements and do not assume the figures will remain constant throughout the loan.
The critical role of oracles
A protocol does not automatically know a token’s dollar value. It relies on an oracle. Delayed, manipulated or unavailable data can trigger an incorrect liquidation or prevent a necessary action.
Check the source, update frequency, deviation threshold and fallback mechanism. A feed aggregated across several markets is more resilient to a local anomaly, although it does not eliminate every risk.
Some protocols use a time-weighted average price. This approach limits an instantaneous manipulation but reacts more slowly during a crash. Others combine several oracles and halt certain functions when discrepancies emerge.
Users cannot correct the oracle at a critical moment. They must therefore assess this component before depositing funds, review past incidents and avoid a narrow buffer on an illiquid asset.
Lending crypto assets is not the same as making a bank deposit
Liquidity providers receive a yield paid by borrowers and sometimes incentives. Their capital depends on the smart contract, the quality of the collateral and the proper functioning of liquidations. An unrecoverable debt can create a shortfall.
On a centralized platform, the company may reuse assets, lend to counterparties or impose withdrawal restrictions. The terms determine whether the customer holds a claim and what priority that claim would have in an insolvency.
A check of an exchange’s reserves provides some information, but does not always reveal liabilities, collateral pledged elsewhere or distressed loans. A high rate may compensate for significant credit risk.
In DeFi, code makes certain rules visible, but an administrator may hold emergency keys. Audits reduce risk without guaranteeing the absence of vulnerabilities. Review permissions, operating history and shutdown mechanisms.
Centralized borrowing or a DeFi protocol
A centralized platform often simplifies the interface, conversions and support. It requires identification and controls custody. Users depend on its solvency, jurisdiction and withdrawal policy.
A DeFi protocol offers on-chain execution and makes its parameters verifiable. Users manage their keys, gas and interactions. An address error or misunderstood signature does not always come with a remedy.
The choice is not simply a matter of “trust” versus “trustlessness.” DeFi relies on code, the oracle, governance and the assets involved. The centralized model relies on contracts, controls and a company. Each chain of dependencies must be identified.
For a company, accounting, tax and segregation-of-duties requirements may make one solution more suitable than the other. A committee should set permitted assets, minimum buffers and caps rather than leaving the decision to a single operator.
Reducing risk before borrowing
Borrow well below the maximum. A generous buffer absorbs some volatility and reaction time. The appropriate level depends on the asset, duration, network and ability to add collateral.
Keep a reserve of the borrowed asset so you can repay without selling under pressure. That reserve should not depend on the same protocol or bridge. Test a small repayment transaction before increasing the position.
Set up several independent alerts for the health factor and collateral price. An alert is no substitute for a buffer: the network may become congested, the application may go down or the owner may be asleep during a sharp move.
Avoid stacking yields. Depositing a borrowed asset into another protocol and then using the receipt as new collateral creates a liquidation chain. Each layer adds an oracle, a smart contract and liquidity risk.
Comparison checklist
Record the borrowed asset, collateral, opening LTV, liquidation threshold and health factor. Note the current rate, how it is calculated, the caps and the liquidation bonus. Check available liquidity and the fees for every step.
Identify the oracle, administrator keys, audits and incidents. For a platform, read the custody terms, restrictions and jurisdiction. For DeFi, confirm contract addresses using the official documentation.
Model a 20%, 40% and 60% decline in the collateral. Calculate the time and funds required to restore the buffer. If the result requires an impossible reaction, the position appears too large.
A crypto loan can provide liquidity without immediately selling an asset, but that flexibility comes at a price. The borrower accepts liquidation risk, a variable rate and technical dependencies. The lender accepts protocol or counterparty risk. A sound decision starts with the loss threshold, not the advertised yield.
