What’s Next for Lending Protocols?
By Daniel Dal Bello, Director.
May 25, 2021–11 min read.
Satoshi once said that “the core design is set in stone” which might have inadvertently moulded views of Bitcoin as a “pet rock”. In my opinion, that’s fine. However, DeFi which is still extremely nascent, finds benefits from its rapid experimentation and implementations to find product-market fit.
In this post we discuss the growing lending environment within DeFi, several of the staple protocols, and new emerging players.
We see that even in something as fundamental as borrowing and lending with innovations and brand new primitives popping up beyond the “Big 3”: Aave, Compound and MakerDAO.
These three are the undisputed leaders within DeFi, especially within the realm of borrowing and lending. Collectively, they have over $18bn in loans outstanding that grew exponentially since late 2020.
These are some of the core considerations to a lending protocol:
- Interest rate determination
- Variable vs. stable interest rates
- Collateralisation Rate/Liquidation Ratio
- Liquidation engine/mechanism
- Method of value accrual
- Types of assets collateralised
New protocols or projects typically seek to put a spin on any or all of the considerations. Alchemix, which will be discussed later, is unique because its spin doesn’t quite fit into the categories and many praise it for its real 0-to-1 innovation or the creation of a new primitive.
Over-collateralisation and Forced-liquidations
Classic lending protocols all follow an over-collateralised model: you have to post up more collateral in dollar terms than you borrow in dollar terms. For example, borrowing Dai (minting Dai in reality) from MakerDAO via the ETH-A vault has a 150% liquidation ratio (“LR”). This means that if you deposit $15,000 in ETH, you are allowed to mint up to $10,000 in Dai that can then be used for any other purpose. You can think of it as a debt-equity ratio of 0.667. To redeem the underlying ETH up to a value of $15,000, you need to repay your debt of $10,000 in Dai.
In this example, 150% LR = $15,000 of ETH collateral / $10,000 in Dai x 100%.
You might do this to leverage your ETH, retaining full price exposure to your underlying position while giving you more firepower to speculate elsewhere (e.g. yield farms). In reality, you’d borrow much less since any fall in the price of ETH will cause the forced liquidation in enough underlying collateral to repay 100% of the debt. This safety net ensures the solvency of MakerDAO and the ecosystem built around it (which is pretty much the entirety of DeFi).
In contrast, buying a house is a form of under-collateralisation. You might make a 20% downpayment while borrowing the remaining 80% of the house value. In this example the term Loan-to-Value Ratio is typically discussed, in our context, we could say the debt-equity ratio is 4.0. The difference is that this leverage is supported by your credit score, ability to continually produce income, and the underlying asset— on-chain credit scoring in DeFi is something not yet solved.
Both AAVE and Compound have similar over-collateralised models with varying liquidation ratios. In adverse market conditions with high volatility or oracle failures, forced liquidations might occur. Another similarity is that debt can be kept perpetually and has no maturity. A key difference is that these two protocols only liquidate enough collateral to cover 50% of the debt.
Similar to a typical loan, DeFi loans accrue interest and have to be paid back upon redeeming collateral. Between the Big 3, their interest rate models differ.
Interest Rate Models
MakerDAO charges what’s known as a stability fee, which is essentially the interest rate, to Dai minters/borrowers. In this case, there’s no spread between the borrow rate and lending rate — there’s only the stability fee, which in the case of the ETH-A vault, is currently 5.5%. This rate is different for each of the many vaults (holding different assets) within the MakerDAO ecosystem, and is also determined by governance proposals and votes. Income is accrued indirectly to holders of the MKR token via buy-back-and-burns. On the off-chance that the protocol suffers a loss, MKR is minted and sold, diluting MKR holders.
Both Aave and Compound have different interest rates for both lending/supplying and borrowing. The supply rate is typically lower than the borrow rate, creating a spread between them. At its core, both use variable interest rate models that have double-slope interest rates that spike up when the asset’s utilization rate (total borrows/total supply) hits a level set by governance. Within Compound, all assets may have a different interest rate curve, sometimes without the “kink”.
Aave introduces stable interest rates as an addition to its variable interest rates. Loans with stable rates can be taken at a premium in exchange for predictability.
A variety of competitors with similar characteristics do exist but will not be explored in this article (Cream, Venus, Liquity…). Instead, we’ll only discuss a few lending protocols that are innovating in a big way on a variety of the Big 3’s key characteristics.
Non-liquidatable loans: Ruler Protocol
Ruler protocol is a relatively new lending protocol largely known for its introduction of non-liquidatable loans. Its other key features are interest rates that are purely determined by demand and supply, rather than altered by an interest rate curve, and fungibility in loans. Ruler also provides both fixed- and variable-rate loans. As of writing, TVL has grown to $86mn with $12m in loans.
Their non-liquidatable loans are in stark contrast to the Big 3 that use varying degrees of forced liquidations. The catch is that these loans also have expiry or maturity dates requiring punctual repayment, failing which will forfeit the borrower’s entire collateral (split pro-rata amongst “creditors”/rcToken holders). Thus, the underlying collateral could fall significantly below the value borrowed (e.g in a market-wide liquidation cascade) before rebounding positively without any liquidation risk. Scroll to the bottom of this link to get a better explanation of the platform’s non-liquidatable mechanism.
By being non-liquidatable, borrowers also do not have to rush to post more collateral in adverse market conditions. This process is typically a nightmare given massive spikes in gas costs sometimes up to 1,000 gwei momentarily. It’s possible that your collateral in traditional loans get liquidated while your transaction to post more collateral is pending in the mempool.
Overall, borrowing on Ruler might be an exceptional way to get leverage without risks of liquidations in a volatile market. However, the catch is that borrowers need to be highly cognizant of their loan expiry dates or risk losing it all.
Self-paying Loans: Alchemix
Alchemix is one of the more interesting projects that has launched this year, bringing with it “self-paying loans”. It’s described as “borrowing on your future yield”, where users can borrow up to 50% of the value of their deposited Dai (and potentially other assets in the future) for a loan that is also non-liquidatable. As of writing, Alchemix has attracted $200m in Dai deposits and over $1.2bn in TVL farming the ALCX governance token.
Somehow, it managed to catch the attention of non-crypto-natives, causing quite a stir due to its counter-intuitive description.
This is a good way to see Alchemix (adapted from 0xdef1’s explanation): (1) You deposit money into a bank account that pays you a certain interest rate. (2) You have an associated credit card that you can draw down 50% of your bank deposit, but with no interest rate and no monthly payments. (3) The interest you earn on your deposit ultimately pays off your credit card loan.
If you deposited Dai and minted (borrowed) alUSD, you can then swap the alUSD into Dai on Curve, effectively borrowing Dai. Unlike depositing a volatile asset to mint Dai in MakerDAO, you have what can be seen as a match between assets and liabilities — stablecoin to stablecoin. Market volatility is not an issue here and the only thing that is required for the redemption of collateral is time.
In reality, it’s not so simple. As a borrower, you’re essentially betting that the underlying yields (via Yearn and its underlying farms) remain positive for the entire loan tenure. If a bear market were to hit in 1 year from now, there’s a good chance yields compress significantly, even turning negative. Furthermore, you’re expecting the peg between Dai and alUSD to maintain on Curve (via Crv.to).
Alchemix v2 is quickly approaching, bringing with it direct value accrual to ALCX governance token holders and more.
Uncollateralised loans: TrueFi
A vast departure to typical loans within DeFi are uncollateralised loans via TrueFi. That’s right, borrowers do not have to post any collateral to borrow from the protocol. This makes it extremely capital efficient for borrowers who do not have to post up any collateral, especially on an over-collateralised basis. Alternatively, lenders may be compensated with higher yields as compared to lending to other money markets. An interesting thing about TrueFi is the use of the TrueUSD (“TUSD”) stablecoin, created by the same TrueFi team.
Instead of collateral, a credit prediction market (“CPM”) is used instead, where holders of the platform’s TRU token vote to “Approve” or “Decline” a loan request. Users can lend TUSD to the platform in a permissionless manner, earning the native TRU token as well as any interest accrued on loans created. Users are encouraged to act as a backstop in the event of a loan default, in return earning a share of loan origination fees.
As you might have gathered, TrueFi is also different from most projects because it is not permissionless. Rather borrowers have to be whitelisted via a KYC process currently open to institutional borrowers only. Institutional borrowers using the platform have borrowed anywhere from $1m to $10m each. Alameda, Multicoin, Poloniex and Wintermute are some of the platform’s users.
In the real world, uncollateralised loans manifest in the form of unsecured loans. These come in the form of personal loans, student loans or credit card loans, issued out on the basis of the borrower’s creditworthiness. In our industry, credit scoring has yet to be solved since borrowers are effectively anonymous. As such, the whitelisting of borrowers via KYC is a necessary process to bridge this gap for unsecured loans. Should a default occur, real legal action can actually be pursued.
The CPM is the other key to the puzzle that enables TrueFi’s uncollateralised loans. TRU stakers are given the stkTRU token as receipt. This in turn is used for “voting” on loan applications — a “yes” is a bet that the loan will not default, while a “no” is the opposite bet. Should an approved loan be paid back in full, “yes” voters are rewarded with more TRU; alternatively, defaults punish “yes” voters by burning a portion of their TRU. Over time, you can expect that certain borrowers will develop a stronger credit score, promoting greater loan utilization and more fees accrued by lenders.
The true test of TrueFi’s platform will come once the initial defaults appear, whether that be from the current whitelisted institutional borrowers, or a new wave of whitelisted borrowers.
Fixed-Rate Loans: Notional Finance
Notional Finance introduces fixed-rate loans which may not appear to be too special given that Aave already provides such a service. However, a key difference is that both lend and borrow rates are at fixed-rates as compared to Aave’s variable lending and fixed/variable borrows. Furthermore, Notional Finance loans have fixed maturity dates. These fixed parameters are extremely useful in bringing about greater financing predictability.
Despite borrowing and lending only in USDC, ETH can currently be used as collateral with a minimum LR of 140%.
Since launching in early 2021, it has attracted over $10m in TVL with more than $6m in fixed-rate loans issued.
However, fixed-rate loans are generally not as well utilized as variable-rate loans. In fact, as you might have been able to tell, most competing lending platforms only offer variable-rate loans (Cream, Venus, Compound…).
One reason might be due to capital efficiency. Picture the scenario below:
- Protocol’s USDC market: close to full utilization with fixed-rate loans
- Average fixed-rate: 10%
- Average market rate for USDC borrowing: 50%
Borrowers have no incentive to “refinance” at higher fixed rates such that the protocol will have difficulty attracting new liquidity. That’s why Aave has a function that allows a “rebalancing” procedure — fixed-rate loans take rates far below market rates and may be moved to higher rates to better incentivise more depositors to enter. Such is the nature of asset-liability mismatches. Since financial markets generally operate on fixed rates, Notional Finance is tackling a huge total addressable market with a good opportunity to onboard more capital into DeFi.
Key to Notional Finance is its fCash token mechanism that represents claims on cash flows at a specific point in the future. This is a necessary complexity since both lenders and borrowers can have a variety of maturity dates on their loans. Borrowers mint fCash and sell it for Dai, similar to how Ruler Protocol operates. To redeem their collateral at maturity, the opposite needs to occur.
In recent months, we’ve seen how quickly cross-chain exploration has grown. Growth on a new chain (e.g BSC, AVAX, Solana, Fantom) typically begins with the development of DEXs (e.g BSC’s Pancakeswap), followed by lending protocols (e.g BSC’s Venus). On Ethereum, the core infrastructure is relatively mature such that we’re seeing massive innovation on existing models.
Something else that should be keenly looked out for are lending protocols that deal with under-collateralisation (not un-collateralised). Any protocol that is able to successfully deliver this should enjoy reflexive growth from capital efficiency.
The next stage is bringing in institutional money that is increasingly eager to capitalize on the lucrative yields found within DeFi. Aave’s Stani Kulechov has suggested that institutional players are familiarizing themselves with the basics of DeFi such as Aave via their private pools. It's only a matter of time before they branch out and begin to explore other platforms with real innovations being developed.
Author’s note: there are more unique lending protocols currently in the market that might be covered in a future article.
The contents of this article are solely the opinion of the author in which he has exposure to a number of the projects mentioned. Nothing here should be construed as financial advice.