Loans Without Liquidations¶
The DeFi Lending Landscape¶
DeFi Lending Today¶
DeFi lending today relies on liquidation.
When a borrower takes out a loan, they must deposit collateral worth significantly more than the amount borrowed. If the value of that collateral falls too far, the protocol liquidates the position and sells the collateral to repay lenders.
Liquidations are the mechanism that protects lenders from collateral price risk.
A New Type of Lending¶
Liquidation is not the only way to manage the risk posed in the drop in collateral prices.
If the risk of a collateral price crash can be hedged at the moment a loan is issued, liquidation is no longer necessary. Instead of managing collateral risk through forced sales, the risk can be priced and transferred to a market at the moment the loan is created.
This makes it possible to construct a different type of lending product entirely: fixed-term loans with no liquidation risk.
Hedging the Loan at Issuance¶
Atomic Hedging¶
The mechanism that makes this possible is atomic hedging.
When a loan is issued, the protocol simultaneously purchases prediction market positions covering the collateral price outcomes at the loan’s maturity that would normally result in insolvency.
Because the hedge and the loan are created in the same transaction, the cost of the hedge can be included directly in the loan pricing.
Loan Pricing¶
The borrower therefore pays interest on two components:
- the amount borrowed
- the cost of the hedge that protects lenders from collateral price crashes
From the borrower’s perspective the loan behaves like a traditional fixed-term credit instrument.
There is no liquidation threshold to monitor and no risk of forced collateral sales while the loan is active.
Market Structure¶
Terminal Price Distribution¶
The hedge is drawn from a prediction market describing the distribution of the collateral asset’s price at the loan’s maturity.
For example, a market may represent the terminal price of ETH on 31 May, divided into a series of price buckets such as:
- ETH < $1600
- $1600 – $1800
- $1800 – $2000
- $2000 – $2200
- ETH > $2200
Each bucket represents a possible range for the asset’s price when the market resolves.
Hedging the Insolvency Region¶
The lending protocol purchases positions in the buckets corresponding to price regions where the loan would become undercollateralised.
These positions do not need to cover the full value of the loan. Instead they are sized to cover the potential shortfall between the outstanding loan value and the liquidation value of the collateral in each price bucket.
At maturity the collateral can be liquidated at the prevailing market price. The prediction market payout then covers any remaining deficit, ensuring that lenders are repaid.
Loan-to-Value and Hedge Size¶
A borrower taking on a higher loan-to-value ratio must hedge a larger portion of the downside price distribution, and therefore purchase more protection from the prediction market.
Shared Expiry Markets¶
In practice these markets can be standardized around common expiry dates, allowing many loans to share the same prediction markets while concentrating liquidity into a single terminal price distribution.
What Happens if Prices Collapse?¶
Early Repayment¶
The borrower must repay the loan before the expiry of the loan term in order to recover the collateral.
If the borrower repays the loan before expiry:
- the collateral is returned
- the loan position is closed
The prediction market hedge remains held by the lending side of the protocol and can either be sold back into the market or held until the market resolves.
Automatic Settlement¶
If the loan is not repaid before expiry, the protocol settles the position automatically.
At settlement the collateral is liquidated at the prevailing market price.
If Collateral Covers the Loan¶
If the collateral value is sufficient to cover the loan:
- the lenders are repaid
- any remaining collateral value is returned to the borrower
If Collateral Falls Short¶
If the collateral value is insufficient to repay the loan:
- the liquidation of the collateral covers part of the debt
- the prediction market position pays out an amount designed to cover the remaining shortfall
Together, the collateral liquidation and the prediction market payout ensure that lenders are repaid.
Insurance Held by Lenders¶
Any surplus generated by the prediction market hedge remains with the lending side of the protocol.
In effect, lenders hold the insurance that protects the loan.
Liquidation during the life of the loan is therefore replaced by deterministic settlement at maturity combined with market-priced insurance against collateral shortfall.
The risk of a collateral price crash is transferred to participants in the prediction market.
A Shift in Where Risk Lives¶
Risk in Traditional DeFi Lending¶
In traditional DeFi lending systems the protocol itself must manage collateral risk.
This is why lending platforms rely on a complex set of parameters and safeguards, including:
- liquidation thresholds
- collateral whitelists
- oracle price feeds
- conservative risk configurations
These mechanisms are necessary because the protocol itself is bearing the downside risk of collateral volatility.
Risk in the Hedged Model¶
In the hedged model this responsibility moves elsewhere.
Instead of the lending system managing collateral risk internally, the downside exposure is sold to counterparties in prediction markets at the moment the loan is created.
As long as a prediction market exists for the asset’s price distribution over the loan horizon, the protocol can hedge the relevant outcomes immediately.
This means the set of assets that can theoretically be used as collateral expands significantly.
Any asset with a liquid prediction market for its price distribution could potentially support lending.
Oracle and Market Microstructure Risk¶
Risk in Current Lending Systems¶
In existing lending protocols the borrower absorbs many types of market-structure risk.
Events such as:
- delayed oracle updates
- flash crashes
- temporary market dislocations
can trigger liquidations even if the underlying asset price later recovers.
Risk in the Hedged Model¶
The hedged lending structure introduces a third possibility.
Instead of placing this risk on either the borrower or the lending protocol, the exposure is sold to prediction market participants at the moment the loan is issued.
If a flash crash or oracle anomaly pushes the collateral price through the liquidation threshold, the prediction market payout covers the lender’s exposure.
The lending system itself remains neutral.
Why This Doesn't Exist Yet¶
The Architectural Constraint¶
For atomic hedging to work, the prediction market trade must execute in the same transaction that creates the loan.
The system must know the exact cost of the hedge before issuing the loan so that it can price the credit instrument correctly.
Centralised Execution¶
Today’s prediction markets cannot provide this guarantee.
Although settlement may occur on-chain, trade execution typically happens inside centralised matching engines or platform-controlled order books.
Because of this, other protocols cannot deterministically execute prediction trades as part of their own transactions.
Without atomic execution, the hedge cannot be reliably incorporated into the loan issuance process.
This architectural constraint prevents prediction markets from functioning as composable financial infrastructure.
Prediction Markets as Financial Infrastructure¶
Beyond Forecasting¶
Prediction markets are usually viewed as standalone applications used to forecast events.
But they can also serve a broader role: pricing and transferring risk.
In this sense prediction markets resemble a programmable form of insurance market.
Rather than simply producing probabilities about future events, they allow financial systems to hedge exposure to those events directly.
PredictionSwap¶
If prediction markets could execute trades atomically on-chain, they could be embedded directly into financial products like the lending structure described above.
PredictionSwap is designed to remove the structural inefficiencies that currently force prediction markets to rely on centralised execution.
By allowing prediction market trades to occur fully on-chain and atomically, prediction markets can become composable infrastructure rather than isolated platforms.
Fixed-term loans without liquidation risk are one example of what becomes possible when prediction markets can be integrated directly into financial systems.
Further Reading¶
For more detail on the structural issues that prevent existing prediction markets from operating this way see: