The article is penned by Alice Henshaw. She is a smart contract engineer at Fluidity. Fluidity are a New York-based company working on DeFi, and are best known for creating decentralized exchange Airswap. Previously Alice worked at ConsenSys where she designed and implemented smart contract systems responsible for over $100M USD in transaction volume. She is a graduate of Oxford University with a degree in Computer Science.
Decentralized Finance (or DeFi) represents one of the most promising near-term use cases for public blockchain technology. By replacing traditional financial intermediaries with immutable open-source software, DeFi has the potential to reduce costs and expand access to traditional financial products. And because many of the underlying protocols are open-source and interoperable, participants can freely mix and match various financial primitives like lego blocks to create entirely new products and forms of financial engagement between users.
While DeFi can theoretically encompass a wide array of decentralized financial products, the application that has gained the most traction to date is lending. Ethereum-based DeFi protocols like MakerDAO and Dharma, for example, have originated roughly $40 million in loans in the last 30 days (and have $100 million in total outstanding loans), up from virtually zero a few months prior. Though these figures are impressive (particularly given the nascent state of the industry), they of course pale in comparison to virtually any particular sub-segment of legacy credit markets. The American mortgage market alone, for example, currently stands at $15 trillion in outstanding debt.
In order for DeFi credit markets to continue to scale and develop into a legitimate alternative to bank financing, they must build out the capacity to accept real world assets (like real estate) as on-chain collateral for loans. Doing so will allow current collateralization ratios and interest rates to go down and also expand access to credit by widening the pool of assets that can be borrowed against. This article explores some of the key technical and legal obstacles that must be overcome for this key milestone to be achieved.
At its core, DeFi is premised on the notion that smart contracts executed by decentralized networks can replace existing financial intermediaries and, in doing so, enable a more open, accessible and efficient financial system. As noted, one area where this vision is already gaining traction is decentralized lending. While many companies are seeking to serve this nascent market, two that stand out are MakerDAO and Dharma. While the mechanics of the two protocols differ, both generally rely on smart contracts that loan a digital asset in exchange for locking up another digital asset as collateral. When the principal and interest on the loan are repaid into the contracts, the collateral is then unlocked and returned. In the event that the borrower does not repay, or if the collateral falls in value below a certain threshold, the collateral is liquidated in order to recover the value of the loan.
This model has the potential to create an alternative source of credit, particularly for underserved borrowers who either have no access to loans whatsoever, or whose only resort is to turn to payday lenders or check cashing services to obtain credit, typically at highly onerous rates. In addition to expanding access, DeFi lending can also unlock efficiencies by eliminating intermediaries, speeding up loan approval processes and potentially enabling lower fees and interest rates on loans.
Notwithstanding these benefits, widespread adoption of DeFi lending applications will depend on their ability to accept real world assets as collateral for loans. Though both MakerDAO and Dharma are planning to expand into other forms of collateral, they remain restricted currently to only ETH (in the case of MakerDAO) and ETH, DAI and USDC (in the case of Dharma). Not only does this add friction for many borrowers (who must first purchase one of these currencies before taking out a loan), it also makes the terms of these loans economically unattractive in many cases.
For example, due to the volatility of assets like ETH, both MakerDAO and Dharma require any loan to be collateralized by an amount of ETH that equals at least 150% of the amount borrowed, far above the collateralization level required on a traditional secured loan. In addition, the current stability fee (which functions similar to an interest rate) for MakerDAO loans is 17.5% annually, which is on par with unsecured credit card debt. In order to make these terms more economical and expand access, DeFi protocols must develop the capacity to accept less volatile, real world assets as collateral.
Establishing legal ownership of real world assets on-chain
Legal ownership of assets like real estate is determined today by the collection of paper records kept in county and state offices. For example, any ownership change of real estate must be recorded with a “deed” in the recorder’s office where the property is located. Similarly, lenders secured by a piece of property that seek to establish priority over competing lenders with respect to that property must generally file a “UCC-1” with the relevant secretary of state’s office. In either case, any subsequent purchasers or lenders must first check the chain of public records kept at these various offices to ensure that competing claims to the property do not exist.
To date, no jurisdiction in the U.S. has allowed real estate ownership to be represented exclusively on-chain. While it is possible to create a token that represents a deed, any disputes as to ownership of that property will be decided by the actual paper deed residing in the applicable county recorder’s office, not the specifications within the token or the wallet address that holds it. In other words, if a deed token were transferred to a new purchaser, but the paper deed was not also updated to reflect the transfer, the purchaser would likely have difficulty establishing ownership if the seller re-sold the property to a third-party. The same problem would exist for property that is pledged as collateral for a loan: if a UCC-1 is not filed with the state, the lender will not have priority over competing lenders, regardless of what the token says.
While legal ownership of certain assets like real estate must currently be registered through recording offices, other assets (including securities) can be legally transferred peer-to-peer more easily. For example, if a security is tokenized and the token is designed as a “bearer instrument”, then the holder of the token is, by definition, the owner of the security. Utilizing this model we can represent legal ownership of property on-chain in the short term, without having to change the underlying recording system.
For example, let’s say Bob wants to tokenize his house to use as collateral for a DeFi mortgage. Bob can set up an LLC to purchase the property, and record this LLC as the official owner on the paper deed. Equity in the LLC can then be tokenized as a bearer instrument on-chain, and these tokens used as collateral for a loan. If Bob then defaults on his loan the collateral equity tokens can be sold, with whoever buys them becoming the owner of Bob’s House LLC and, by proxy the owner of Bob’s house.
While this solution can enable DeFi loans secured by real world assets, it is a rather inelegant fix, as it requires setting up an LLC for each asset that will collateralize a loan. Given the time and costs required, this approach will limit the scalability of using off-chain collateral, especially for smaller loans.
Of course, an ideal solution would allow the actual property itself to be represented on-chain using a token. In the real estate example, the property could be represented using a Non-Fungible Token (or NFT), which effectively just means it would be unique. Each NFT would store the details of the property and the deed, and when the time came to sell the property any descriptive and legal changes would be updated for the new ownership transferal. In addition, security interests (including priority of competing lenders) could also be encoded into the token. This system would of course require the consent of local recording offices, as well as courts with legal jurisdiction over any property claims. Given the highly localized nature of the legacy system, this approach is not likely to gain widespread adoption in the near term.
Ultimately, the barriers to a seamless on-chain collateral system are more legal than technical in nature. While a fully optimized system may be far off, near-term fixes (including tokenizing the LLC that owns the collateral) can offer significant advantages to DeFi lending applications by greatly expanding the pool of assets that can be borrowed against. The continued growth of DeFi as a true alternative credit source depends on the ability of these protocols to accept off-chain collateral without disrupting the underlying model. Time will tell whether this is possible.