Author: Nicole Cheng (Investment Manager of OFR)
Advisor: JX (Partner of OFR)
In the past year, we witnessed tremendous growth in the NFT space, but yet we have realised one unignorable fact: the more non-fungible tokens we collect, the less liquid our portfolio becomes. As the proliferation of NFT technology brings more novel applications, there is also a growing need for NFT financialization to unlock optimal capital efficiency.
In fact, NFT is a low-velocity asset very similar to real estate, while in the traditional world, properties are often pledged as security for loans. Think of an NFT-backed loan as a home mortgage, where users are able to lend or borrow money by using these low-velocity assets as the collateral of the loans made. The intermedium that facilitates this process is called the NFT collateralised lending and borrowing protocol. In this report, we will focus on research around such protocols, including the pricing mechanisms and different types of approaches based on the counterparties.
To be considered as collateral, an NFT needs to obtain enough consensus around its value, to an extent where the mainstream holds a belief that its value is not going to fade anytime soon. This requires high trading volume and good reputations of the creators, both in a consistent way. Some of the most-recognised collaterals include CryptoPunk, BAYC, MAYC, Azuki, and Doodles, which are also the so-called ‘Blue Chip’ NFT collections. If we want to fit these ‘Blue Chip’ NFT back into the home loan example, they are undoubtedly the primate cities, while ‘Blue Chip' that have the rarest traits is the luxury residential areas within the primate cities.
However, NFTs are highly volatile assets, even the value of a ‘Blue Chip’ collection may be subject to significant fluctuation. The floor price of BAYC has hit a record high in ETH-demonated valuation before the Otherdeed mint, followed by a more than 50% drawdown afterward. A long-standing challenge that every collateralised lending protocol faces in its design is: How to unbiasedly determine the value of the underlying NFT collaterals? Existing players have adopted a few different solutions:
Time Weighted Average Prices (TWAPs)
Oracles such as Chainlink would fetch and publish the time-weighted average price of both sales and floor prices to create a blended price to value the NFT. Such a model can mitigate the outlier events by taking the average of multiple prices over a predefined period of time, thereby increasing the difficulty of potential malicious manipulation events.
However, using TWAPs in the valuation of NFTs has some major drawbacks - TWAPs can only be applied to collections with an active market and large transaction volumes, of which are less prone to oracle attacks. The capital efficiency in the TWAPs method is also lower as protocols tend to set a smaller Loan-to-Value ratio to avoid the impact of extreme market conditions.
Example: BendDAO, JPEG’d, Drops DAO, Pine Protocol, DeFrag
In the peer appraisal approach, NFTs are appraised by the users to give a prediction of the value of the NFTs. Appraisals that occurred in a Peer-to-Peer fashion can be applied to a broader range of NFT collections, as it does not impose a strong restriction on the quality of collections like the TWAPs method does. By leveraging individuals or curator committees with reward incentives, it enables a price discovery of NFTs which can then give a rather objective value to the NFT collaterals. However, the valuation costs are significantly higher than other approaches through incentive provision, with a less efficient process and potential inaccurate outcomes.
Example: Taker Protocol, Upshot V1
Liquidity Pool-based Pricing
One of the issues associated with the peer appraisal approach exists in its inability to provide a real-time price for NFTs. This is no longer the case in the liquidity pool-based pricing approach, where every NFT put onto the protocol is actively traded by the effective lenders in a pool, thereby generating constant spot pricing on the NFT equals to the total ETH in the pool. Once the NFT is locked in a pool by the borrower, traders can start depositing ETH into the pool to their perceived NFT value. If the NFT is overvalued, traders may lose its ETH in the case of a public auction; In the opposite case of NFT being undervalued, traders will fill up the pool until they believe it reaches the true market value of this NFT in an effort to take advantage of profits in the case of sale. Through encouraging the traders to speculate on the NFT pool, the value of NFT will be less-biasedly determined in a dynamic way.
Although some of the examples listed above do not fall into the range of an NFT lending protocol, these pricing mechanisms play a vital role in determining the maximum loan amount as well as whether to execute a liquidation event. Once the value of an NFT is determined, depending on the type of counterparties, these protocols can be divided into two models:
This approach is theoretically applicable to all NFTs and it is easier to reach a consensus on the underlying value of an NFT. Think of it as an open market where lending protocols act as a facilitator to provide the marketplace. On one side, NFT holders can create a loan with their desired terms and on the other side, a fund provider can browse the platform to decide to who they want to lend their money. Once they accept the loan offer, a contract will be created and the NFT used for collateral will be sent to an escrow account guarded by the protocol. Simultaneously, the loan will be transferred to the borrower together with a promissory note NFT.
As the lender and borrower come to an agreement for loan terms such as duration, LTV ratio, and APR, the systematic risk could be mitigated as defaults only happened locally. However, its customisable capability comes along with lower time-to-liquidity and scalability as lenders and borrowers will need to wait until it reaches a mutual agreement.
Example: NFTFi, Arcade, MetaStreet
Rather than a ‘bid-and-ask’ loan that may never be accepted, this is more of a ‘let the market decide’ approach, where liquidity provided by lenders will be pooled together to share the interest paid by the borrowers depending on the supply and demand of each side. If the borrowers are unable to pay back the loan or the NFT collateral faces a liquidation situation due to the price drop, the protocol will auction the NFT and the proceeds will be returned to the borrowers.
The amount of funds available to the borrowers can be significantly increased with the Peer-to-Pool approach, while the borrowers can instantly get access to liquidity as they do not need to wait for lenders to confirm the terms. However, this also means that they will need to rely on a reliable price feed from the oracles to generate the loan terms automatically, while the approach can only be applicable to mainstream NFTs as long-tail NFT assets are more sensitive to price manipulation.
Example: JPEG’d, DeFrag, BendDao, MetaLend, Pine, Drops DAO
For comparison purposes, I’ve put up a table consisting of some important metrics when evaluating an NFT lending protocol. Some protocols decided to set a cap on the Loan-To-Value (LTV) ratio to limit the potential of defaults, while the ratio is normally higher for NFTs with greater liquidity and demand. Supported collections varied a lot with peer-to-peer protocol outperforming most of the peer-to-pool protocols in terms of the coverage range. Note that most of the protocols are consistently increasing their supported collections while making adjustments to the pricing mechanisms and LTV ratio.
Despite all the controversy surrounding the NFT collateralised lending protocol, we have yet to expect more NFT lendings and financialization primitives entering the space, providing the NFT collectors a gateway to unlocking greater value from the digital collectibles. Taking a step forward, if one day there is a sustainable amount of NFTs locked in the lending protocol, these protocols may turn out to hold some degrees of pricing power for the NFTs. There is plenty of untapped potential awaiting us and I have no doubt financialization is going to be one of NFT’s strongest narratives this year.