DeFi liquidation risk in 2027 will not depend solely on a drop in the deposited token. Rising debt, an appreciating borrowed asset or changes to protocol parameters can also weaken a position. Effective monitoring must track both sides of the balance sheet and the market’s exact rules.
This guide explains how to build a stress test based on official documentation consulted on November 9, 2026. The figures used are fictional; they do not describe any recommended threshold or guaranteed condition for 2027. Understanding the mechanism is not protection against liquidation.
In brief
Our guide to crypto lending and collateral covers the basic concepts. Here, the goal is to measure the distance between a current position and its threshold, then check whether corrective action would remain genuinely possible under adverse conditions.
- The liquidation threshold is not a prudent borrowing level to target.
- The protocol’s price reference may differ from a chart viewed elsewhere.
- An alert does not repay the debt and does not guarantee that a transaction will arrive in time.
Distinguishing debt initiation from liquidation triggers
A borrowing limit authorizes the opening or increase of a debt position. The liquidation threshold determines when collateral no longer provides sufficient coverage for that debt under the market’s rules. These parameters may be separate. Treating them as two names for the same margin can lead to poor decisions.
In Aave’s explanation of the health factor, the value of collateral weighted by its thresholds is measured against the value of the debt. A ratio above one does not guarantee absolute safety. It reflects a state calculated using prices and parameters that may later change.
A position record should identify the relevant version and assets before any calculation. Rules for partial and full liquidation, as well as liquidator incentives, should not be copied from another market. It is also important to distinguish a supplier with no outstanding debt from a borrower exposed to the seizure of their own collateral.
A supplier without borrowing exposure still faces other risks, including those related to the contract, the asset and potential market losses. But that deposit alone does not create a personal debt that must be kept below a threshold. Confusing these roles makes it impossible to determine who an alert applies to and what action would be relevant.
Calculating the potential decline before the threshold without confusing percentages
Consider a fictional position with 8,000 units of collateral, debt of 4,000 and a threshold of 80%. Its simplified ratio is 8,000 multiplied by 0.8 and divided by 4,000, or 1.6. These values are used solely to illustrate the mechanism; they do not recommend any borrowing level.
If the debt and threshold remain unchanged, the collateral value corresponding to a ratio of one is 5,000. The decline from 8,000 to 5,000 is 37.5%. This distance is not equal to 60% because the ratio is 1.6, nor is it simply the difference between two debt and collateral percentages.
Now assume that the debt value rises by 10%, to 4,400. The collateral value corresponding to the same threshold then reaches 5,500. The distance from the initial 8,000 falls to 31.25%. Debt growth therefore reduces the margin even without a withdrawal or new borrowing.
This calculation assumes a single collateral asset, a single borrowed asset, a fixed threshold and consistent pricing. A multi-asset position requires appropriate weighting and scenarios. These assumptions should be recorded rather than presenting the calculated distance as an exact, immutable liquidation price supplied by the protocol. Adding collateral, making a repayment or changing the debt requires the calculation to be redone; keeping the old threshold in an alert can then provide information that has become false, even if the original formula was correct.
Assessing debt as closely as collateral
Debt is initially denominated in units of the borrowed asset. Borrowing a volatile token does not create a fixed obligation in the currency used for spending. If its price rises, the value to be repaid may increase. Collateral denominated in another unit does not automatically offset that change.
Interest is a second driver. A moderate rate over a long period can increase the debt if no repayment is made. A position close to its threshold has little room for this accumulation. Monitoring must therefore keep the amount owed up to date, not just the quantity received on day one.
A strategy that borrows to reinvest may appear to generate a net return as long as both rates remain favorable. But a change in funding costs, a decline in income or a discount on an asset can disrupt the balance. Expected income is not a repayment reserve that is already available.
The guide to a €500 crypto portfolio helps distinguish gross assets from net wealth. Borrowed funds must be recorded as debt even when they have been redeposited into another service. Adding all balances without deducting liabilities obscures the actual exposure.
Checking the oracle and shared dependencies
The price used to assess coverage comes from a mechanism defined by the market. The official page on Morpho oracles explains, among other things, the distinction between this reference and an indicative price shown in the application. An alert based on another price may therefore miss the relevant signal.
The oracle address, its sources and its conversion assumptions must be identified. A token assumed to represent another asset may be valued under a rule that does not exactly track the secondary market. The analysis should ask what happens if that relationship changes, rather than assuming it will always remain equal to one.
Two assets that are usually correlated can diverge during a period of stress. Collateral and debt with similar economic exposure do not make every scenario impossible. Discounts, liquidity and pricing sources should be examined separately, especially when the assets depend on the same issuer, bridge or redemption mechanism.
The risks documented by Morpho also point to the possibility of bad debt if collateral no longer covers obligations before liquidation. The mechanism is designed to manage credit risk; it does not promise that every price move, technical incident or liquidity shortage will be absorbed without loss.
Understanding what a liquidator can do
In the description of Morpho liquidations, a third party may repay an eligible debt in exchange for collateral under the market’s rules and incentives. This is not the same as a voluntary sell order that the user could freely cancel after the threshold had been crossed.
The amount that can actually be liquidated and the benefit granted to the liquidator depend on the protocol and its version. An old seizure percentage should not be generalized to every situation. The position record should refer to the applicable document, with its date, rather than retaining an isolated figure found in a tutorial.
Liquidation does not guarantee an automatic profit for the party carrying it out either. Network costs, competition and the conversion of collateral all matter. For borrowers, these details mainly show that an automated, competitive mechanism can act faster than occasional manual monitoring.
Falling below the threshold should therefore be treated as a situation to avoid, not as a planned exit method. Waiting until the last moment to choose between repayment and adding collateral reduces the available options. The strategy should define in advance what it is prepared to do, with which funds and on which network.
Preparing an executable response rather than a simple notification
A stress test can combine a decline in collateral, an increase in debt and higher transaction costs. The objective is not to predict the next move, but to check what response would be possible. A scenario that requires an immediate bank transfer over a weekend relies on availability that may not exist.
Emergency funds should be distinguished from capital that is already committed. If they are held in the same protocol or in an asset that has become illiquid, their availability cannot be assumed. The asset needed to pay transaction fees on the relevant network must also be available, without assuming that an interface will cover every cost.
The guide to crypto fees can help quantify this process. Several successive operations may be required; a notification does not execute them. Congestion, a pending transaction or a problem accessing the wallet can make operational margin smaller than calculated margin.
Signature security is a separate issue. Hardware wallets protect keys without guaranteeing execution within a given timeframe. No support service should receive the recovery phrase to “prevent a liquidation.” A financial emergency can also be used as a pretext for an attempted theft.
For 2027, effective monitoring therefore requires a dated calculation, adverse scenarios and a practicable exit. Relevant margin is not determined by price alone: it depends on obligations, oracles, available funds and the time required to act. If this chain cannot be explained, a reassuring ratio is not enough to validate the position.
