The Web3 Lending Landscape
Last week, we gave an overview of the different use-cases for lending protocols. Today, we’ll be diving into the differences (and limitations) between some of the leading blockchain lending platforms.
To better approach this study, we decided to divide the different lending protocols into distinct generations. Here’s an overview:
- Gen 0: CeFi
- Gen 1: ERC20 lending protocols
- Gen 2: Peer-to-peer NFT lending protocols
- Gen 3: Peer-to-pool NFT lending protocols
- Gen 4: Arcadia Finance lending protocol
Without further ado, let’s dive in.
Gen 0: CeFi
Lending is not a concept that was invented on-chain. It has existed for at least 3000 years, depending on who you ask. Loans are simply contracts where banks act as an intermediary between the borrowers and the liquidity providers. If you get a mortgage for a house, interests and principal are paid-off monthly. If you can’t pay back the loan, the house gets sold and the proceeds are used to repay credit providers. The mortgage in this case is a collateralized loan.
Another example of a collateralized loan comes in the form of margin trading on centralized (crypto) exchanges. In this type of collateralized loan, you’re borrowing an asset from the exchange with the promise you’ll return the asset on a later date. To ensure you’ll keep your promise, you agree with the exchange to leave some collateral as guarantee.
While these types of collateralized loans are value-adding for both consumers and businesses, the current system is not without shortcomings:
- Users have to trust the financial institutions to fully act in their best interest. There’s no way for users to verify what’s happening behind the scenes.
- Innovators can only build on top of the end-points of this black-box. End-points that are controlled by incumbents (but initiatives as PSD2 are already a big step forward!).
That there is a problem with managing collateralized loans within a black box is best illustrated with the Great Financial Crisis. The 2008–2009 crisis highlighted what happens when the expected collateralized value of loan isn’t on par. The initial housing mortgages were repackaged with many layers of securitization (Mortgage → MBS → CDO → CDO²…).
Since the traditional financial system is very opaque and fragmented, it was impossible for the final investors to properly assess the risk of the underlying collateral. Because these risks were not properly assess, even individuals with very bad credit score received loans with low interest rates, which (among other factors) led to a housing bubble and the final implosion of the system.
The biggest losers were the investors that bought the securitized mortgages backed by worthless houses. Please note that the banks and pension funds that bought these securities were NOT the final investors. These institutions are nothing more than trusted intermediaries that invest funds on behalf of the real final investors: employees and people with saving accounts.
A Blockchain-based financial system promises to solve a lot of what’s wrong with CeFi, but in the process creates a new set of problems.
Gen 1 Lending Protocols: Single ERC20s
Contrary to their centralized counterparts, blockchain-based lending protocols work based on a pre-defined set of rules set forth in a smart contract that allows each participant in the system to know exactly what they are signing up for. All trust assumptions are explicitly codified, and therefore can be verify in real-time. This was a big unlock and gave rise to a number of first movers: MakerDAO/Oasis, AAVE, Compound and CREAM. At their core, these protocols enable users to deposit a curated list of ERC20 tokens as collateral and borrow against them. For the first time, a transparent and permissionless lending system emerged.
In MakerDAO, each user creates their own vault to store collateral against which DAI is minted. AAVE, Compound and CREAM pool assets of lenders together, forming lending pools used by borrowers.
A common theme with these Gen 1 lending protocols is that users take out a single loan against a single asset. Users with multiple assets are forced to open and manage multiple positions separately. This is expensive and cumbersome for users. Furthermore, a significant decrease of value in one asset always requires users to either top-up the position, repay the loan or get shreked. An increase in value of asset A doesn’t compensate for another position holding asset B which decreased in value.
Another design aspect of these protocols is their inability to quickly add new types of assets to their protocol. That is, their protocol architecture was built around a single asset type. Therefore, increasing the range of allowable asset(s) (types) would go together with a complete protocol version upgrade.
Gen 2 Lending Protocols: Hello NFTs
Blockchains continued to gain traction and with it, new asset types were introduced. Case in point were NFTs. In a relatively short period of time, the capital locked in these NFT assets grew to billions of dollars. As capital in the space continued to grow, so did the interest from users to leverage said capital. Many protocols emerged to solve this problem for users, most popularly, NFTfi, Arcade and Defrag (to name a few). They enabled users to get liquidity with an NFT as collateral for the first time.
As such, a new type of lending protocol emerged to provide peer-to-peer loans against NFTs. But under-the-hood, these protocols are very inefficient. First, because they work in a peer-to-peer manner, liquidity wasn’t instant. That is, users have to first find a counter-party willing to underwrite the loan. Second, taking out a loan meant temporarily losing ownership of the underlying NFT. Users wouldn’t be able to use their NFT to attend a gated event for example.
The big upside of these peer-to-peer NFT lending platforms is that one-offs and rare items within NFT collections can be appraised and thus leveraged. But Gen 2 lending protocols cannot offer the instant liquidity that made the 1st generation lending protocols popular. As a result, it remained difficult to make NFTs composable with other protocols given that the outcome couldn’t be atomically predicted.
Another frequently forgotten issue is the loss of ownership. Although there is much more to talk about this topic, a big part of the intrinsic value of NFTs comes in the form of utility, whether that means using your NFT to attend an exclusive event, verify your identity to participate in an airdrop or something else entirely. The reality is that any NFT provided as collateral in a Gen 2 protocol cannot be used for anything else.
Gen 3 Lending Protocols: Instant NFT Liquidity
Improving on their peer-to-peer predecessors, 3rd generation lending protocols introduced the concept of instant liquidity to NFT markets. As of now, it’s mainly JPEG’d that’s already live.
Users can deposit a single NFT and receive immediate liquidity without having to go through an appraisal process or underwriters first. In the case of JPEG’d, users can deposit blue-chip NFTs, like CryptoPunks, as collateral and get instant liquidity. This is a nice step forward in the right direction. However, it’s limited in its current form.
First, similar to their gen 2 counterparts, users losing ownership of their NFTs is a big drawback. Second, users can only use a single NFT as collateral per position. In other words, users that wish to open a loan against a collection of NFTs need to open individual positions for each asset. If you happen to have 10 NFTs, you’re looking at 10 individual loan positions. Third, users are at the mercy of the collection’s floor price volatility. That is, whereas ERC20 lending protocols allow users to top up their open positions to avoid liquidation, NFT lending protocols do not. Because of this, NFT lending protocols work with limited LTVs (collateral ratios of 300%) to the detriment of end-users.
Gen 4 Lending Protocols: Arcadia Finance
Both Gen 2 and Gen 3 lending protocols aim to bring new asset types (NFTs) to the world of DeFi. However, both Gen 2 & 3 lending protocols remain stuck in the “single asset, single loan” way of working. Sure, being able to lend against your CryptoPunk or Bored Ape is nice. But what about lending against your CryptoPunk and your Bored Ape at once? What about being able to top up your position, just like you would do with an ERC20 loan, to avoid a liquidation? What about combining your NFTs with your LP tokens or yield bearing tokens? At Arcadia Finance, we asked ourselves why there was no way for users to get a credit line against their entire portfolios, specially when that translates to a better user experience and lower fees.
You can feel where we’re going here. We wouldn’t be making this article if we weren’t going to end the comparison with Arcadia Finance. We have built Arcadia Finance to allow users to open a vault, deposit a combination of different types of assets, and get instant liquidity. The advantages of this approach cannot be overstated. By allowing users to pool their assets, users don’t have to open and manage multiple individual positions when trying to use more than one asset as collateral for a loan. This comes with many benefits, including lower collateral thresholds, better UX for users, lower fees, and complete composability of any asset, including NFTs, with the wider DeFi ecosystem. What’s more, users can still actively manage their assets while they’re being used as collateral. Users are able to sign transactions, swap, stake, and even change ranges in your Uniswap v3 position. In short, users can borrow against any asset without losing ownership and utility in the process.
For the interested readers, Matt Aaron holds a nice list of all NFT lending platforms currently live or in development: https://app.charmverse.io/share/e412fd3e-1551-4ec2-9482-438aa3a18f7e.