AaveDAO debates Dai collateral limit after $600M mint with eUSD backing

AaveDAO risk management contractor Chaos Labs proposed a 12% reduction in the Dai loan-to-value ratio after the stablecoin became partially backed by eUSD.

The Aave decentralized autonomous organization (AaveDAO) continued to debate limits to Dai (DAI) collateral on April 5, as risk management consultants Chaos Labs offered a new proposal to reduce Dai loan-to-value ratios (LTV) by 12%. Previously, Aave Chan initiative founder Marc Zeller had argued for a 75% reduction.

Aave is a crypto lending platform that runs on multiple blockchain networks. It allows borrowers to take out loans in one cryptocurrency while depositing another as collateral. It is governed by holders of the Aave token, who collectively form AaveDAO. Dai is an algorithmic stablecoin backed by multiple forms of crypto collateral, including USDC (USDC), Ethereum (ETH), and others. Dai is issued by the Maker protocol, which is governed by MakerDAO.

MakerDAO was criticized on the AaveDAO forums on April 2 after it minted 600 million DAI and deposited it in a vault with decentralized lending protocol Morpho. An April 1 proposal on the MakerDAO forums attempted to increase the vault's minting limit to 1 billion DAI, potentially leading to an even greater supply of the stablecoin.

Morpho vault with 100 million Dai. Source: Morpho.

MakerDAO claims that the newly minted Dai will be lent out to end-users who deposit stablecoin eUSD to the Morpho protocol. Thus, they claim that the new coins will be properly backed by stable collateral. In response, critics claim that eUSD is a risky asset and that MakerDAO is being overly aggressive in using it for collateral purposes.

On April 2, Zeller proposed that the Aave LTV for Dai be set to zero instead of its current 75%, which would essentially ban Dai from being used as collateral for any new Aave loans. Zeller claimed this change is necessary because of “concerns regarding the inherent risk nature of DAI as collateral” after its alleged “aggressive actions.” The proposal is still in the discussion stage and has not been put to a formal vote.

On April 5, AaveDAOs risk management consultant, Chaos Labs, analyzed the risks posed by the new vault and suggested a 12% reduction in LTV instead. This would still allow borrowers to use Dai as collateral in new loans but would require them to hold a greater collateral-to-loan ratio to compensate for the perceived increased risk. Currently, Dai depositors to Aave can only borrow 75% against the value of their Dai. The proposal from Chaos Labs would lower this amount to 63%.

Related: Aave deploys DeFi protocol on BNB Chain

One of the key issues in the debate is the nature of stablecoin eUSD, which is issued by the Ethena protocol. According to Ethena documents, eUSD is backed by two components: an amount of Lido Staked Ether (stETH) that market makers deposit into the protocol and a corresponding futures short position that hedges the underlying ETH represented by this deposit.

Ethena protocol official website. Source: Ethena.

This backing should be “delta neutral” or incapable of declining or increasing in value, according to the documents. Thats because, if the price of ETH increases, the value of the stETH deposit should increase while the value of the short futures position should decrease, causing gains and losses to net at zero. On the other hand, a decline in the price of ETH should have the opposite effect, causing a loss in the value of the deposit that exactly equals a gain in the value of the short position. Either way, the value of the eUSD token should remain stable regardless of which direction ETH price goes.

In addition, the documents claim that holders can earn a yield from staking eUSD, which purportedly comes from two sources: First, the stETH deposit earns staking rewards from the Ethereum network, which can be captured by eUSD stakers. Second, the futures price for ETH is currently much higher than its spot price, creating a “basis” or spread between them, which can also be captured by stakers. According to the protocols official website, eUSD staking pays a yield of approximately 37% as of April 5.

Despite this characterization of eUSD as delta-neutral, critics have claimed that it can become under-collateralized in two different circumstances. First, in a crypto bear market, the price of ETH futures could become lower than the price of spot ETH. If this occurred, the yield on the position would turn negative, implying that stakers would need to pay to hold eUSD instead of receiving payments for holding it. Critics say this would lead to a wave of redemptions, causing bad debt in the system and leading the eUSD value to go below $1.

In an April 2 thread on X, Yearn.finance founder Andre Cronje argued that eUSD must eventually become “unbacked” due to “negative funding rates,” stating:

“while things are going great now (because market is positive and shorting funding rates are positive [because everyone is happy being long]), eventually that turns, funding becomes negative, margin/collateral gets liquidated, and you have an unbacked asset.”

Second, a problem with Lido‘s staking network could lead stETH’s value to decline relative to ETH. This could cause the collateral backing each eUSD to be worth less than $1. Decentralized finance educator BowTiedIguana expressed this concern in a reply to Cronjes post:

“One issue that I've never seen acknowledged or discussed is that ETH is not a perfect hedge for stETH. If something bad happens on the tech side (low probability but very high impact) the unstaking queue becomes very long and the prices of the two assets diverge significantly.”

On April 3, Ethena protocol developers hosted a public discussion on X with MakerDAO founder Rune Christenson and Morpho development team CEO Paul Frambot, along with Zeller from Aave Chan.

In the discussion, Zeller argued that MakerDAO and Morpho did not have adequate guardrails to protect Dai from its exposure. The rest of the participants claimed that eUSD poses some risk to the Maker protocol but that the development teams from all three projects (Maker, Ethene, and Morpho) were managing this risk properly.