Tornado Cash sanctions raise concerns among lending platforms, NFTs lending under liquidation risk, Fei Labs dissolving Tribe DAO, an interview with Liquity's co-founder, and more...
Issue #6 of The State of DeFi Lending newsletter
The Merge is coming… time to be cautious.
Welcome to issue #6 of The State of DeFi Lending, a bi-weekly newsletter covering the highlights of lending markets in DeFi.
In this issue we cover:
Some Short Updates about Fei dissolving, Compound III going live, Acala hack, Alchemix and more…
Lending Platforms getting prepared for The Merge (Ethereum PoS fork).
NFTs lending exposed to liquidation.
OFAC sanctions of Tornado Cash ripple effects on stables and lending platforms.
An interview with Rick Pardoe, Lead Engineer at Liquity.
We know there are many DeFi newsletters and substacks out there.
But few if any are focused only on DeFi lending.
That’s the gap we are here to bridge.
Short Updates
Fei Labs propose to dissolve Tribe DAO (and got some ex-partners angry).
Compound deployed its first USDC market on Compound III, pending a gov vote to go live.
Acala is out on a tracing mission to recover over $3b aUSD minted incorrectly.
Alchemix launching its new Vault Migration tool.
Jet Protocol, a lending protocol on Solana, launches v2 in beta mode.
MIM proposes to lock 1/3 of its treasury CRV for stronger voting power on the Curve wars.
News Recap
Leading DeFi lending platforms getting prepared for The Merge
The upcoming Ethereum PoS fork (aka The Merge) might create some short-term effects on lenders and borrowers due to speculations on PoW ETH to maintain some non-zero value (at least for a short while). Compound, Aave, Euler, and other protocols are preparing for the merge by DAOs voting to clearly signal which fork their platform will support, and by setting new parameters to allow a smoother and safer experience for their users during this period of uncertainty.
Compound plans to significantly grow the max borrow rate for cETH market and to add a max borrow cap in order to ensure continuous withdrawals during the merge. No governance proposal on this was published yet.
Aave proposed a DAO vote for selecting PoS Ethereum mainnet as the only chain to operate its markets and governance, shutting down any Aave deployments on other Ethereum forks post the merge.
Liquity announced that they will maintain operations on ETH PoS chain without any expected issues for users
But though Liquity main oracle is Chainlink, which will only support the PoS fork, it also has a fallback oracle, Tellor, which is controlled by TLR holders that might keep operating on PoW ETH fork at least for a while. In this case, some scenarios are possible for LUSD throve owners to extract some value out of their troves -
Euler signaled they will also raise the borrowing interest rates for ETH to prevent full utilization of ETH market as 100% utilization will make it impossible for users to withdraw their ETH.
ICYMI - Euler also published its full a-z ETH PoW game plan towards the merge:
NFT lending at risk of mass liquidations
According to CirrussNFT there is $57m worth of ETH in NFTs collateralized on BendDAO and other NFT lending platforms such as NFTfi. Some of them are at immediate risk of getting liquidated, which at current market conditions could potentially trigger cascading liquidations scenario as the floor price of BAYC and MAYC will be impacted.
Read the details in his thread -
It seems some of these liquidations have already started -
US Gov sanctions on Tornado Cash raise questions regarding the true decentralization nature of lending platforms and CDPs.
Recent sanctions by US Gov. OFAC on Tornado Cash made Circle, the company behind USDC, blacklisting addresses that were interacting with TC.
The DeFi community didn’t appreciate this move…
The blacklisting and sanctions were followed by banning accounts on some of the leading lending protocols’ frontends, raising discussions on the decentralization essence of DeFi.
Though the protocols themselves might be fully decentralized and accessible to anyone at the smart contract level, the user interfaces, which are the main path for DeFi users to interact with these protocols, are maintained by centralized companies which are open to regulatory scrutiny in case they do not comply with the US sanctions.
MakerDAO’s frontend Oasis went out with a similar message -
But the frontend issues for MakerDAO were the less concerning part of USDC blacklisting accounts. As DAI is currently being backed mainly by USDC, the potential of it censoring accounts that are related to MakerDAO started turmoil discussions, including an idea to market-sell the $3.5b USDC from its PSM to ETH.
But some didn’t like the concept -
Meanwhile, the flames around the blacklisting are getting cooler, but the need for a decentralized stablecoin still remains.
While MakerDAO’s DAI is being criticized for its centralized USDC dependency, Liquity’s LUSD getting strong support from decentralization advocates
Seems like the OFAC sanctions and the censorship made by USDC and its potential effect on DAI, played into the hands of Liquity’s LUSD stablecoin. LUSD can be minted only against ETH collateral, has immutable contracts, and has no governance, making it more attractive for decentralization advocates.
5Qs
For our 6th issue we are happy to have Rick Pardoe, Co-Founder and Lead Engineer at Liquity, answering 5 questions about the lending markets in DeFi.
Stable coins that are backed by debt provide a partial Compound-like lending functionality to users, however typically with much lower interest rates (and with LUSD it is even interest free). What is the black magic that enables it? What are the down sides in a stable coin based lending platform? What are your thoughts on the advantages and disadvantages with the future Aave hybrid approach, who plan to build a stable coin on top of their traditional lending market?
The core of this model is collateralization: the stablecoin supply must at all times remain backed by collateral worth at least as much as the value of the outstanding stablecoins.
Liquity uses pure ETH as collateral - it’s the most robust and decentralized collateral, but it’s also volatile, so the system needs to be overcollateralized. This ensures that LUSD is fully backed even if the value of the underlying ETH drops suddenly.
The more responsive the system is - i.e. the more rapidly you can clear out bad debts - the more capital efficiency you can get, and the better loan-to-value ratios you can offer. Liquity has novel rapid and automatic liquidations. The system can almost instantly batch-liquidate hundreds of loans in a single transaction, and the LUSD required to pay down the bad debt tends to be readily available in the system’s Stability Pool. That’s why Liquity can offer an unprecedented high leverage with a minimum 110% collateral ratio.
Liquity also doesn’t use interest rates. The LUSD price floor of $1 is enabled by a hard redemption mechanism that guarantees 1 LUSD can always redeem for $1 worth of ETH (fees aside). This acts as a natural counterbalance to excess leverage. Conversely, LUSD has a hard price ceiling at $1.10 where it becomes profitable to borrow, mint and sell LUSD for arbitrage profits - even if the Trove used to mint LUSD is liquidated.
A downside of the stablecoin based lending model in general is that it couples borrowing demand to stablecoin demand. The total LUSD supply is constrained by the market appetite for borrowing and leverage. This makes it challenging to scale LUSD, though it does mean that it grows organically and sustainably.
As for Aave’s GHO - we’re certainly curious to see how that pans out! I have to say we’re not fans of DAO-governed interest rates though. If you must use interest rates to control supply, a well-tuned algorithmic approach makes more sense to us.
How will the GHO price floor be implemented? If it redeems for loan collateral, that changes the Aave system drastically. Otherwise, they’ll need something like DAI’s PSM - and if ultimately GHO is backed by another stablecoin, that kind of defeats its purpose. Overall though we’re big fans of Aave’s strong engineering and security, so we’re excited to see GHO hit the market.The LUSD stable coin is fully immutable, without any governing entity. This helps in positioning it as resilient to censorship, and gives certainty on its interest rates. But after more than a year since going live, what do you see as the downsides for having no governance?
No governance is great in a lot of ways: it gives users confidence that the rules of the system will never change, it frees up dev capacity for new projects, and ultimately it’s true to the spirit of blockchain. Users can trust the code - they don’t need to trust the future whims of human admins.
However, immutable systems pose unique challenges. As an engineering team you have to get everything right first time: since we can’t make any changes to Liquity’s system code now that it’s live, we had to be absolutely sure that the system was as secure as possible before launch. We placed a heavy focus on testing and adversarial thinking, we did several security audits with top firms, and we were generally very diligent about finding and mitigating failure modes and worst case scenarios.
Robust economic modelling was also key: we had to know that the system would survive heavy ETH price crashes and extreme network congestion.To that end we commissioned a thorough economic stress-test by Gauntlet Network, as well as external academic macroeconomic modelling. The Liquity system was really put to the test just a few weeks after launch - May 19 brought a widespread crypto collapse which liquidated billions in value broadly, and $100m worth of debt in Liquity. Thankfully, our system behaved exactly as intended and passed the test with flying colors.Does Liquity have a multichain strategy? E.g., Compound was extensively forked by other projects, and recently introduced a Business Source Licence to mitigate forking. Does Liquity work on something similar?
Liquity has been forked many times also. We’re not in the game of restrictive licenses. Imitation is the best form of flattery! Most forks actually fizzle out - they tend to be get-rich-quick cash grabs, and they often mess with the immutabilty of the system.
Some Liquity forks have even evaporated as the value of the collateral collapsed. Launching a system like Liquity needs more than just a copy/paste of our code - it needs a robust collateral choice, and users must have high confidence in the team and security practices in order for it to attract sticky capital.
We’re certainly exploring cross-chain and L2 opportunities. We wouldn’t deploy multiple Liquity instances, since we only want to grow one stablecoin - rather, the focus is on bridging LUSD to different venues. Even though the Liquity Protocol is still only available on mainnet, LUSD has been bridged to the Optimism rollup. LUSD’s liquidity fragmentation is a serious concern, so we’re prudent when it comes to bridging LUSD to new chains, since there is a strong need for LUSD liquidity on mainnet for liquidations when ETH crashes.
I have to say that at heart we’re quite the ETH maximalists - it’s the only chain that’s really proven itself over the years while maintaining decent decentralization. We’re excited about some of the zk-rollup based L2s and the potential for batched Liquity actions which amortize gas costs and settle to Liquity on L1, without need for a system redeployment.Do you think DeFi lending will eventually become a winner takes all market with one big platform having 99% of the market?
The DeFi lending market does have strong network effects and amplifying feedback loops. When real value is at stake, users gravitate to the battle-tested and most secure protocols. The longer a lending protocol survives without a major exploit, the more trust and confidence it instills, and the more attractive it becomes to users with large borrowing or lending appetites.
However, there are different lending models: double-sided lending markets, collateralized stablecoins, even undercollateralized borrowing. We’ll see different leaders in different categories.
Also, dApps make different trade-offs. With Liquity we went full-tilt with decentralization and immutability: “automated, algorithmic, set-in-stone” are core Liquity value propositions. Other protocols prefer to sacrifice these in exchange for flexibility and manually adding features over time.
Another trade-off is capital efficiency. Liquity loans have an unprecedented minimum collateralizaton ratio of 110%, however, as a borrower in Liquity you need to be more attentive to your position than you might in other protocols which are less capital efficient but more hands-off.
I think we’ll continue to see protocols making different trade-offs to serve different user wants and needs.MakerDAO is making big efforts to maintain an exact 1:1 peg, which comes with a centralization cost (relying on USDC). Rai on the other hand aims for a floating peg, where the coin is not pegged to USD, but rather just supposed to have relatively low volatility. What are your thoughts on RAI-like use cases? Could a Liquity-like system support such use case as well?
RAI is an interesting niche financial construct. Their intent was to create something akin to a “dampened ETH”, and they sort of admitted it didn’t turn out that way.
There are probably other ways to do a dampened ETH collateral - you could potentially construct it via options, kind of an “inverse Squeeth”. Though if you wanted something more algorithmic, RAI’s PID controller approach is smart.
Could something like RAI be used in a Liquity-like system? Since RAI is unpegged and has algorithmic interest rates, it’s very different in kind from LUSD. RAI’s decentralized and algorithmic nature could make it a nice collateral type, though it’s a subtle instrument: the bull case for RAI (and thus demand for leverage against it) is not as clear as for pure ETH.
The broader question is: how big is the demand for a low-volatility asset unpegged to fiat currency? As long as people pay for their goods, services and bills in fiat, then pegging to fiat is probably the way to meet the demand where it’s at.
That said I don’t think we’ve seen the end-game for collateralized stablecoin design. There’s still plenty of room for immutable designs with improved capital efficiency and a tighter peg. We have a few ideas up our sleeves here!
Till next time…