DAI collateral ratio: why volatility shifts market stability

It did not remove the mechanism that makes DAI vulnerable to crypto-market volatility: collateralized borrowing through Vaults.

But one distinction matters from the first sentence. Most DAI created through crypto-backed Vaults is secured by collateral whose dollar value must exceed the outstanding debt by a defined buffer. DAI can also enter circulation through the Peg Stability Module, or PSM, in a 1:1 USDC–DAI swap that does not use the same over-collateralized Vault structure. The collateral ratio is therefore a core risk metric for Vault-generated DAI, not a description of every unit of DAI in circulation.

The $1.00 price of DAI is not set by a single market maker, an issuer, or a reserve manager. It emerges from several mechanisms working together: Vault collateral, liquidation rules, arbitrage, stablecoin conversion facilities, oracle prices, and the willingness of independent Keepers to act when positions become unsafe.

That distinction becomes critical when the market moves quickly. A high collateral ratio can absorb a moderate drawdown. A falling ratio can push a Vault toward liquidation. A congested blockchain or thin auction market can then turn a manageable price move into a system-level problem.

The Mechanics of Over-Collateralization in Sky Protocol

When a user deposits ETH into a Maker or Sky Vault, the user is opening a loan denominated in DAI against that collateral. The protocol does not assess the borrower’s income, credit history, or intention to repay. It assesses the value of the collateral against the amount of DAI created as debt.

For an ETH-A Vault configured with a 150% liquidation ratio, $150 of locked ETH secures $100 of DAI debt. The position has a 50% nominal buffer before the collateral reaches the liquidation threshold. That buffer is not free capital for the borrower. It is the system’s protection against the fact that ETH can lose value before a liquidation transaction is executed.

The basic calculation is straightforward:

Collateralization ratio = market value of collateral ÷ outstanding debt × 100

If the value of the collateral falls while the DAI debt remains unchanged, the ratio falls with it. Debt accrues through the applicable stability fee, so the ratio can also deteriorate gradually even when the collateral price is flat. Conversely, a rising ETH price increases the ratio without the borrower repaying any DAI.

The risk is not only the size of an individual position. It is the distribution of positions across the system. A protocol with many lightly buffered Vaults is more exposed to a rapid ETH sell-off than one in which borrowers maintain substantial excess collateral. This is why dai stablecoin collateral ratio fluctuations are best understood as a system of moving exposures rather than as a single number printed on a dashboard.

Over-collateralization exists because crypto collateral is volatile and continuously repriced. A position that appears safe at one moment can approach liquidation after a sharp market move. If the system accepted $100 of ETH for $100 of DAI debt, even a small fall in ETH would leave insufficient collateral to repay the debt. The buffer gives oracles, Keepers, and auctions time to respond.

The architecture has evolved since single-collateral DAI launched in December 2017. The November 2019 introduction of Multi-Collateral DAI opened the system to several asset types, each with its own risk parameters. The August 2024 transition to Sky Protocol preserved DAI while introducing USDS as the forward-issuance asset and the Sky Savings Rate as a yield mechanism for holders. The names and interfaces changed, but the central question remained the same: how much collateral stands behind each unit of debt, and how quickly can that collateral be sold if the position fails?

The collateral ratio is a buffer against time: it gives the protocol room to sell collateral after prices have already started moving.

Vault collateral is not the same as PSM liquidity

It is tempting to describe all DAI as crypto-backed and leave the explanation there. That is too broad. The PSM was designed to provide a different path for DAI creation and conversion.

Through the PSM, a participant can deposit an approved stablecoin such as USDC and receive DAI at a one-to-one rate, subject to the module’s limits and fees. That transaction does not create a conventional ETH-backed Vault with a 150% or 130% liquidation threshold. The USDC is treated as the asset supporting that conversion, and the module’s risk is tied to the stablecoin, its reserves, its issuer, and the operation of the conversion facility rather than to the liquidation of a volatile crypto position.

The distinction affects how DAI supply should be read:

• DAI created against ETH or another volatile asset is linked to a Vault-level collateral ratio and liquidation threshold.
• DAI obtained through the PSM is linked to the approved stablecoin held by the module and the module’s available capacity.
• DAI acquired on the open market may have originated through either route; the token itself does not reveal its issuance path.
• Aggregate stability therefore depends on the mixture of Vault collateral, PSM assets, market liquidity, and outstanding debt.

The PSM reduces one type of volatility while adding another type of exposure. It can make the peg easier to defend around $1 because users have a direct conversion route. It does not eliminate risk; it changes the source of that risk.

Vault Parameters: Comparing ETH-A and ETH-B Liquidation Thresholds

Different Vault types carry different liquidation ratios, stability fees, debt ceilings, and auction parameters. ETH-A and ETH-B illustrate the basic trade-off between capital efficiency and protection.

These figures describe the configurations used in the draft’s comparison and should be read as Vault-specific parameters, not as a permanent rule for every ETH position. Governance can change risk settings, and the relevant parameters may differ across versions, collateral types, and implementation periods.

An ETH-A position reaches liquidation territory when the collateral value falls to roughly 1.5 times the DAI debt. With ETH-B, the lower 130% threshold allows the borrower to extract more DAI from the same amount of ETH. That improves capital efficiency in a rising or stable market. It also leaves less room for a fast decline before the protocol must intervene.

The higher ETH-B stability fee is part of that pricing structure. A borrower receives more leverage but pays more for the position. The fee does not make the collateral safer; it compensates the system for accepting a tighter buffer and can make the position less attractive to maintain over time.

The trade-off can be summarized without pretending that one Vault type is universally superior:

• A wider liquidation buffer gives Keepers more room to execute before the collateral no longer covers the debt.
• A tighter buffer allows greater borrowing against the same collateral but increases sensitivity to intraday price movements.
• A higher stability fee raises the cost of maintaining debt and may discourage prolonged leverage.
• Lower fees can support borrowing demand, but they do not offset the risk of a position that is too close to liquidation.

This is the protocol’s primary risk-management surface. Liquidation ratios are generally set for a Vault type rather than recalculated individually for every borrower. Governance may propose changes through executive votes, but a position remains exposed to the parameters that apply to it until those rules change or the borrower adjusts the debt.

The same logic applies beyond ETH. Stablecoins, real-world assets, and tokenized treasuries do not carry the same price-volatility profile as ETH, so their parameters can be designed differently. Less volatile collateral may support a smaller buffer, while assets with weaker liquidity or more complex settlement arrangements may require more conservative treatment even if their quoted price appears stable.

That is why the phrase crypto-collateralized stablecoin stability can be misleading when used as a blanket label. Stability is not a property of the word “collateralized” alone. It depends on what the collateral is, how liquid it remains during stress, how the oracle values it, and whether buyers are available when the system needs to sell it.

The Role of Keepers and Auction Dynamics During Market Crashes

When a Vault’s collateralization ratio falls below its liquidation threshold, automated liquidators known as Keepers can trigger the liquidation process. Keepers are independent operators running software that monitors Vaults, oracle prices, gas conditions, and auction opportunities. They are not employees of Sky and do not receive an instruction from a central dealing desk.

The liquidation system depends on their incentives. Once a position is taken over, its collateral is offered through an on-chain auction or another protocol-defined sale mechanism. Buyers compete for the collateral, and the resulting proceeds are used to cover the Vault’s DAI debt, accrued fees, and any applicable penalties.

A discount to the oracle price can compensate Keepers and auction participants for execution risk, inventory risk, and transaction costs. The exact incentive depends on the relevant system parameters. It should not be treated as a guaranteed fixed discount across every market condition or every generation of the protocol.

Under ordinary conditions, competition can push the auction price toward the market reference. The system has time to process transactions, participants can estimate the collateral’s value, and available liquidity is sufficient to absorb the sale. During a crash, all three assumptions can fail at once.

Several problems can appear:

1. Many Vaults can become unsafe together. A sharp fall in ETH does not target one borrower. It can move a large group of positions toward liquidation at the same time.

2. Auction participants may reduce risk. Keepers that normally bid aggressively may preserve capital, widen their discounts, or stop participating if gas costs and price uncertainty become too high.

3. Market depth can disappear. The oracle may show a price at which the collateral is theoretically worth a certain amount, while executable bids are materially lower.

4. Transactions compete for block space. A liquidation that is profitable in one block may become uneconomic or insufficient in the next.

5. The collateral itself can keep falling during the auction. A sale that would have covered the debt at the start of the process may no longer do so by the time it settles.

This is where the DAI peg becomes dependent on external market infrastructure as well as smart-contract rules. The protocol can specify a liquidation threshold, but it cannot force an independent Keeper to provide unlimited capital at any price. It can design incentives and auctions; it cannot guarantee orderly liquidity in a market-wide panic.

The impact of ETH price on DAI is therefore indirect but significant. ETH does not need to be used to buy DAI for its price to matter. If ETH is the collateral behind a large amount of DAI debt, a fall in ETH weakens the positions that support that debt. If liquidations clear smoothly, the system can remove unsafe exposure. If they do not, the protocol may inherit a shortfall.

The Peg Stability Module and Stablecoin Collateral

MakerDAO introduced the Peg Stability Module after the March 2020 market dislocation. Its purpose was to make it easier to exchange approved stablecoins and DAI around the target price and to reduce reliance on volatile collateral for every unit of new DAI.

The PSM permits a one-to-one USDC–DAI conversion under its configured conditions. A participant can deposit USDC and receive DAI without opening an over-collateralized ETH Vault, or move in the opposite direction by returning DAI and receiving USDC. The mechanics are fundamentally different from borrowing DAI against ETH.

This is the qualification the collateral discussion requires: DAI supplied through a PSM conversion is not backed by a Vault with a collateral ratio above 100%. It is supported by the stablecoin held in the module on a 1:1 basis, subject to the module’s parameters and the solvency and accessibility of the underlying stablecoin. Saying that every DAI unit must be backed by collateral worth more than the debt would therefore misdescribe the PSM layer.

The PSM offers three practical benefits:

• It gives traders and arbitrageurs a direct route for moving between DAI and an approved dollar stablecoin.
• It reduces the amount of DAI issuance that must rely on ETH or another volatile asset.
• It can help bring the market price back toward the target when DAI trades away from $1.

The trade-off is equally clear. The PSM reduces exposure to ETH liquidation cascades, but it increases exposure to the stablecoin held inside it. USDC brings issuer, reserve, custody, legal, and regulatory risks that do not appear in the same form in an ETH Vault. If access to the reserve is restricted or the stablecoin itself trades below its intended value, a nominal 1:1 conversion route may not provide the same economic protection as it does in normal conditions.

The PSM also changes the composition of DAI supply. Two days with the same total DAI supply can have very different risk profiles if one has more exposure to volatile crypto Vaults and the other has more exposure to stablecoin conversion facilities. Dai supply volatility factors include borrower demand, changes in stability fees, ETH price movements, PSM usage, governance decisions, and the willingness of users to repay or refinance debt.

When ETH rises and borrowing conditions are attractive, crypto-backed Vault activity can expand. When ETH falls or market participants prefer a more stable instrument, demand for PSM conversions may increase. These flows are not a perfect automatic hedge. Users do not always act immediately, and PSM capacity, fees, and market liquidity can constrain the response. Still, the ability to shift between collateral channels is a major part of the system’s defense.

The PSM does not make DAI risk-free; it exchanges part of the liquidation risk of volatile collateral for the reserve and issuer risk of the stablecoin held inside the module.

Network Congestion and the Risk of Bad Debt Accumulation

The Keeper system is most important when conditions are least comfortable for the people operating it. During a sudden volatility spike, two risks become closely connected: delayed price information and delayed execution.

Oracles provide the reference prices used to calculate Vault collateralization ratios. If the reference price lags the spot market during a crash, a position may appear safe to the protocol even though its collateral is already worth materially less in an executable market. When the oracle updates, several Vaults may cross their liquidation thresholds at once.

The delay does not have to be large to matter. In a market moving quickly, the difference between the price at which a Vault is flagged and the price at which its collateral can actually be sold can determine whether the auction repays the debt in full.

Network congestion compounds that timing problem. Liquidation transactions compete for block space with trades, transfers, oracle updates, and other activity. Gas costs may rise. A Keeper may submit a transaction that fails to confirm before the next price move. Another operator may see the same opportunity but decide that the required capital and execution risk no longer justify participation.

The protocol’s nominal threshold is therefore not an instant guarantee of solvency. It is the point at which the system attempts to begin a process. The process still requires a functioning chain, updated prices, available Keepers, adequate auction liquidity, and collateral that can be sold at a price sufficient to cover the debt.

When the sale proceeds are not enough, the result is bad debt: DAI liabilities that cannot be fully recovered from the collateral associated with the failed Vault. Bad debt is not the same as a temporary DAI price deviation. A market price can return to $1 while the protocol still carries a deficit from an earlier liquidation failure.

The March 2020 event, an early major stress test for Multi-Collateral DAI, demonstrated how quickly these weaknesses could compound. The episode helped motivate a greater role for stablecoin collateral and the PSM, but it did not make liquidation risk disappear. It showed that a decentralized credit system can fail not because its arithmetic is wrong, but because the surrounding market cannot execute the arithmetic fast enough.

Mitigating that risk requires more than raising every liquidation ratio. A higher ratio gives a position a larger buffer, but it can also reduce capital efficiency and discourage borrowing. The system must balance several variables:

• the expected volatility of the collateral;
• the liquidity available in normal and stressed markets;
• the reliability and update behavior of the oracle;
• the capital requirements for Keepers;
• the design of the auction mechanism;
• the debt ceiling assigned to the collateral;
• the exposure introduced by stablecoin-based modules such as the PSM.

No individual setting solves all of these problems. A conservative liquidation ratio can still be overwhelmed by a chain halt or a market with no bidders. A deep Keeper network can still struggle if the oracle reference is delayed. A large PSM can support the peg while increasing concentration in one external stablecoin.

Operational Risk in TradFi Terms

For institutional users evaluating DAI as a settlement asset, the distinction from USDT or USDC is operational rather than philosophical.

USDT and USDC are exposed primarily to issuer and reserve risk. The relevant questions include whether reserves are sufficient, whether redemption remains available, how assets are custodied, and whether legal or regulatory action restricts access. DAI is exposed to a different set of failure modes: collateral prices can fall, oracle updates can lag, liquidation auctions can clear below expectations, Keepers can become inactive, and bad debt can accumulate.

That does not make one category automatically safer. It makes the risk map different.

For DAI generated through crypto-backed Vaults, the collateral ratio is the first line of defense. It determines how much price movement a position can absorb before liquidation becomes necessary. For DAI associated with the PSM, the key question is not whether the Vault is 150% or 130% collateralized. The question is whether the stablecoin held by the module remains redeemable and useful at the intended one-to-one relationship.

This distinction matters for settlement desks and treasury operators. A payment team may see DAI trading close to $1 and conclude that the underlying system is stable. That conclusion is incomplete. The market price reflects current liquidity and arbitrage conditions; it does not reveal how much of the supply is supported by ETH Vaults, how concentrated the collateral is, how close positions are to liquidation, or how much conversion capacity remains in the PSM.

For cross-border settlement rails, treasury operations, and merchant acquisition pipelines that touch decentralized stablecoins, the practical question is not simply whether DAI is more decentralized than USDT. It is whether the system’s collateral and execution architecture can absorb the volatility of the assets it accepts.

That assessment should track at least four moving parts:

1. Collateral composition. ETH-backed debt carries a different risk from USDC-backed conversion liquidity or tokenized real-world assets.

2. Position health. A high aggregate collateral ratio can conceal a concentration of lightly buffered Vaults.

3. Execution conditions. Keepers need functioning infrastructure, reasonable gas costs, and buyers willing to take the other side of auctions.

4. Exit liquidity. The ability to convert DAI through markets or the PSM matters most when participants are trying to do the same thing at once.

The system’s resilience is therefore dynamic. It can look strong while ETH is liquid, auctions are competitive, and PSM conversions are available. The same architecture can become strained when ETH falls rapidly, gas markets tighten, and stablecoin liquidity becomes one-sided.

DAI’s collateral ratio is not a promise that losses are impossible. It is a design choice about where the losses should be absorbed and how much time the system has to respond. Over-collateralized Vaults place a buffer between volatile collateral and DAI debt. The PSM creates a separate one-to-one route that avoids Vault liquidation but introduces stablecoin counterparty exposure. Keepers and auctions connect the two layers to real market liquidity, where the clean logic of a smart contract meets the less orderly reality of a crash.

That is the operational reality behind DAI. Its stability depends neither on over-collateralization alone nor on the $1 market price alone. It depends on whether each layer — Vaults, oracles, Keepers, auctions, the PSM, and the underlying collateral markets — continues to function when volatility makes every delay more expensive.