A business proposal that I did together with a group of 2 other MBA students and 2 UChicago Law students.
Decentralized finance in its current state tends to be circular: investors use crypto assets to take out loans in crypto or earn yield on their crypto in the form of crypto. While this has been a big step forward, the industry needs to be able to interact with assets in the real world before it can produce meaningful economic activities. Bringing real world assets on-chain offers a unique opportunity to access diverse off-chain debt markets, while also allowing TradFi institutions to tokenize and issue debt/assets, agnostic of market geography. In fact, DeFi could solve many TradFi problems and unlock new markets and capital access. A blockchain property recording system can establish immutable records of ownership, reduce reliance on trusted third parties and the associated overhead costs, leverage alternative data for better credit decisions and unlock new lending opportunities. Borrowers who donât have access to traditional lines of credit can find financing through on-chain credit. At the same time, investors benefit from increased transparency of the risks they undertake. Such a system should lead capital to flow efficiently if built with sufficient safeguards to withstand potential misrepresented or fraudulent data on-chain, privacy concerns, and lack of standardization across the different jurisdictions.
I. Problems to solve
We are trying to solve 4 problems that currently exist within asset-backed financing â specifically, loans against property â in the traditional financing world:
One of the crucial pain points associated with asset-backed financing is determining ownership. When a trad-fi lender, such as a bank, wants to lend against a property, how does one determine who owns the property? Every city, state and country has different ownership rules, documentation, and criteria. The paperwork varies across jurisdiction and in many cases may be absent altogether.
Because of differences in regulations and the lack of a unified platform to maintain data on ownership, banks and lenders are often not able to lend against these properties. Because of this inherent structural issue, banks tend to focus on the âcleanâ cases, i.e. the high-income customers who have detailed ownership records, while ignoring complex cases where ownership exists but cannot be proven. The laws are so specific that the process is not a standard one-size-fits all, even within the same state.
Because of the high amount of time spent on paperwork and processing of these loans, asset-backed financing becomes a fairly expensive endeavor. In developed markets like the U.S., banks charge an approximate 1% processing fee on these loans. This means that if one were to take a loan for $50,000 against a house that they owned, they would have to pay the bank an additional $500 just to be able to get the loan. While this number might seem small, it is actually very meaningful given that interest rates have traditionally been in the 1-4% range. This means that processing charges alone could often be 25% of the interest burden for the borrower.
From our conversations with lenders, we understand that 2-3 personnel hours on average are spent in trying to determine ownership of the property and whether all the relevant paperwork exists. The reason this process is expensive is because there is a lot of manual work involved on behalf of the lender. Back-office teams process applications, determine credit worthiness of the borrower and also check for title ownership of the property. The paperwork in Illinois looks different from the paperwork in New York, for example. As a result, lenders need to train their staff on different regulations. Further, a lot of time is spent trying to determine nuances of local regulations in addition to state regulations.
Research has found that more than half of consumers seeking asset-backed loans struggle with the process, whether thatâs getting connected to a lender, being verified, or navigating complex steps that delay the approval process. Consumers have identified the steps that would enable easier processes. Over 62.2% of those sampled said that quicker, more simplified processes were top of mind. Another 29.7% reported that in-real-time access to updates on the status of their loan would simplify the process, and 24.3% said that access to pre-qualification tools would improve their experience In a separate survey, lenders were asked to identify issues and challenges with their current processes in their organization. Over 35.6% of lenders reported operational issues were a struggle, while 28.1% reported issues in lack of efficiency. The findings show a clear discrepancy between lender processing systems and customer needs. Consumers today desire quick outcomes, yet also expect personalized customer experience. Companies are searching for AI tools that allow them to increase customer satisfaction through digital retailing.
Even if one were to complete the steps above and determine ownership, trad-fi is too siloed and fails to leverage the powerful data it acquires in approving loans. Lenders and holders of this information are in competition with each other, deterring an interest in sharing hard-earned data. And governments may not identify incentives to proactively distribute this data across government or private networks, even if the information-share could facilitate lending processes. For example, if 100 Booth MBA students took loans against property over time and if 0 of them defaulted, then lenders should realize that Booth MBA students are a highly creditworthy set of borrowers. They should be willing to give smaller ticket loans to these students at more reasonable rates, because the expected cost of distress is almost zero. However, if institutions are unaware of this trend, the industry and the individual borrowers suffer, slowing the economic benefit for all. Much of the existing data sits on individual servers, limiting the power of data and the ability to meaningfully understand customer bases.
For a long time, trad-fi has focused on serving the top of the pyramid customers and continue to focus on increasing share of wallet with them. However, we believe that there exists a large customer base (~40% of individuals belong to the low-income category) which is overlooked and sits at the bottom of the pyramid. These customers are often exploited through predatory lending practices (high interest rates, onerous terms, etc) and as a result, continue to be part of a vicious cycle.
These are customers who often donât have their property ownership fully recorded, or have incomplete documentation. Because their data does not get recorded on the blockchain, it remains in local and physical form which means it can often get lost over time. As a result, this customer base often finds itself locked out from access to capital. The US residential market is ~$15 trillion ($13tn in single-family and $2tn in multi-family). We believe that solving some of the problems listed above could ensure that this market grows by 30-40% (~15% credit growth + 25-35% growth in penetration of low income households).
II. Mechanics of our solution
By its nature, high capital real estate properties are beyond the reach of the ordinary investor. Tokenizing property makes for much better market access for both investors and borrowers and simplifies the securitization process. It will enable the use of real estate as collateral for on-chain loans and the transferring property ownership based on predefined conditions.
The process will look as follows:
A property is recorded on-chain as a non-fungible token. It is associated with a certain transaction output, which is called a 'genesis transaction output'. That transaction output belongs to the initial owner recorded by the system. For example, a genesis transaction output is established by the developer of the house, also the initial owner.
When the property is sold or transferred, a transaction output which belongs to the previous owner is spent, and a transaction output which belongs to a new owner is created in the same transaction, which needs to be created according to certain rules.
When somebody needs to identify an owner, he or she will go through the transaction history starting from the genesis transaction up to an unspent transaction output. The owner of the unspent transaction output is the current owner of the property.
In this case, property ownership is associated with a certain private key rather than with a certain person. If we assume that only one person is in possession of that private key, the effect is the same. However, a private key can be lost or stolen. Also, in some cases a legal system (courts) might override ownership which was recorded in the blockchain. We must take this into account if we want to build a robust system. For this, our company will act as a trusted registry.
For ownership transfer, the token can be sent to a multiÂsig address, which requires signatures both from us, the registry and from the owner to unlock. In this case, the owner cannot transfer their property without interaction with us, the registry. Neither can we do transactions without the owner's consent. We can perform additional authentication steps to make sure that transfer is correctly authorized, such as by verifying the ownerâs ID and intention to transfer ownership.
How do we keep information about the asset up-to-date on-chain? In order to fully enable property as a financial vehicle to access on-chain lending, we need accurate, current and trustworthy property valuations. Blockchains do not contain any data about the real estate industry. To connect smart contracts to the traditional data and systems outside the blockchain, we will build out an oracle network to transmit data between on-chain and off-chain environments.
Initially, our oracle network will include property appraisal oracles. We can start by having multiple independent traditional real estate appraisers for each property, which is the current state of the real world. The independent appraiser as an oracle is an expensive and accurate system, but fortunately there is no need for daily or weekly appraisals of the properties as real estate markets arenât that volatile.
As we scale, we will start using algorithmic solutions to value properties. Services like Clear Capital or Estated offer algorithmically-generated valuation information for real estate properties. Each on-chain tokenized asset will be connected through a network of oracles with APIs integrated to multiple private and public sources on up-to-date market valuations, rental rates, regional trends, market condition, regulations, etc. These oracles will aggregate data and build out predictive algorithmic pricing models that get sent to smart contracts. More importantly, the same framework can be applied together with the use of big data to expand beyond asset valuation and into user's alternative credit data, meaningfully improving credit decisions and identifying more applicants which meet credit criteria.
For privacy purposes, the data stored on the networking layer is arranged via a hash mechanism. A third party cannot derive any meaningful information from the data without a specific key that enables them to decrypt the data. However, each territory may have different applicable laws. This will create an issue with the use of public blockchains given each node could be in a different geography subject to different laws and enforcement. For the purpose of complying with regulations, the geographical positioning of the nodes will need to be accounted for.
III. Customer acquisition and scaling
Our business will operate as a two-party lender/borrower platform, and will rely on economies of scale as a primary value driver. Furthermore, network externalities must be present to both the lender and the borrower for our offering to be successful. For this reason, the growth strategy would involve focusing on perfecting the process within specific zipcodes during the early stages of development, before expanding to other regions. This concentrated approach to expansion is a framework that could be replicated across different nations. The same test, perfect, and expand approach would be adopted to work out the legal, regulatory, and different lending practices associated with each target nation.
Before the product is rolled out at the nationwide level, it will be imperative that the pilot creates value on both sides of the platform. On the lender side, it will be important to ensure trust in the accuracy of the assessed home values. To do so, it will be important to incentivize the oracle nodes to validate that the home data is correct and up to date. Satisfaction on the borrower side of the platform will be driven whether or not the service improves loan access, terms, and execution timelines.
Selecting operating regions during the pilot stage will depend heavily on which cities/counties/states have a legal framework that is open to a blockchain based solution, especially one related to lending practices. Because ownership rules and documentation practices can vary, selecting jurisdictions with favorable policies will be helpful in avoiding regulatory headwinds.
To assess whether or not a state has favorable home ownership policies, one could develop a state ranking system based on metrics commonly used to assess the health of the residential market. If average annual income for a state is healthy relative to national averages, however foreclosure rates and home value to income ratios are low, this could be an indicator that lending practices within the state are prohibitive to borrowers. For example, the state of Michigan has an Average Annual Income of $94.5k which is nearly the national average. Despite falling within the middle 20 percentile of states on average annual income, Michigan has a foreclosure rate per 10,000 homes of 0.25 and a Home Value to Income Ratio of 2.87, metrics which are both within the bottom 25 percentile at the national level. In other words, despite Michiganâs healthy economy and its occupantsâ low likelihood of default, the state has a low home value to income ratio, indicating that there may be room for improvement in lending practices.
Once an operating region is selected, it will be imperative to ensure that the contracts on the decentralized oracles network are both robust and accurate. Contracts will be validated by oracle nodes, who will receive rewards for entering correct data into the system. During the early stages, larger rewards may be necessary to incentivize interaction with the network, however this can later be scaled down as economies of scale begin to take effect as a value driver.
IV. Legal risks
The integrity of the data is integral to the establishment of an effective oracle. Not only does the data need to be accurate for lending purposes, but it also must have necessary safeguards to prevent risk of fraud. The business proposal facilitates small business loans by lowering the transaction costs of acquiring mortgage-based loans. Yet, because typical lending requires an individualized process to determine that a potential borrower is the rightful holder of a title, the established Oracle must translate this same degree of care into creating an information base for titles. This legal risk section discusses three areas of risk which the business would need to devote particular attention to: fraud, privacy, and navigating property law and conflicts across a variety of jurisdictions.
To understand the threat of fraud, a bit of context is helpful. In the U.S., we rely on a title assurance system to assure land purchasers have title and this information is held in public records offices distributed throughout the country. By having the title assurance system, one can better access credit, spend less time enforcing their personally-owned mortgage documents and facilitate robust market development, lessening the transaction costs associated with verifying ownership of property.
In the U.S., prior to centralized title assurance systems, proof of ownership through official mortgage/title paperwork was not the only functioning system for property rights. Other characteristics of property were also rewarded, such as being the first to live on a plot of land or being the person to put work into the land. Eventually, ârecording statutesâ were implemented, and with it, a requirement to register title or proof of property ownership with your local government. When recording statutes were implemented, states still leveraged core property principles, such as prioritizing ownership for those who were first to report their title to the state. At this time, the majority of titles have been recorded and are more diligently updated as formal transfers of land occur.
Although some information for the Oracle will come from documents from official government records, the system would be structured so that individuals can submit their own documents and proof of title. Thus, the Oracle will need to have a verification process for ensuring that title submission is accurate and a contingency plan is available for addressing fraud found in the blockchain system. Further, the ability to identify the fraud would be primarily manual and centralized up until market participation is high enough. The âfirst to claimâ treatment may also incentivize faster uptake so that people can ensure their information is accurately recorded before fraud can occur â essentially âclaimingâ your NFT token as the rightful owner.
In exploring issues of fraud, it is important that the company considers the best way to address on-chain/off-chain enforcement. If the fraud exists on the blockchain, the company must know what legal enforcements may be available to address this type of fraud with any sort of meaningful recourse. Further, if the system is decentralized, it is important that the company understands and establishes its authority to make these corrections as they occur.
Data on blockchain can be quite sensitive and subject to privacy and data protection regulations. For example, postal address is regarded as âpersonal informationâ under California Consumer Privacy Act of 2018. Under this statute, customers have the right to request that any collected information about them be deleted. Not only in the US, a right to be forgotten has become recognized as a statutory right for many parts of the world. In the EU, customers have the right to obtain the erasure of their personal data without undue delay if certain conditions are met: personal information is no longer necessary, the subject person withdrew its consent, among other things. Yet, it is nearly impossible to change information on blockchain. So our business has to establish a system where these customersâ requests are completed in due course so as to comply with this requirement. Further, we have to determine who is legally responsible for customersâ requests. Given that our system is decentralized where unassociated individuals perform as nodes, it can be difficult to attribute this legal responsibility to a specific entity.
Each state faces a variety of licensing requirements for those involved in borrowing and lending processes. These licensing requirements often come with background verifications and training to ensure that those handling valuable private information can be trusted with it and to reduce the risk of costly mistakes. For example, in Michigan you must be registered with the state, complete necessary verifications, and renew licenses annually. Further, states may vary in their processes and, to ensure best compliance, the company will need to follow the most stringent standards and ensure that they maintain abreast of any changes to licensing requirements down the lines.
V. Weaknesses
Jurisdictions may navigate property laws and norms differently. Customers will submit different legal documents to prove their title depending on their jurisdiction in which their real property is located. This inconsistency might make it challenging for our business to establish a global and centralizing system on which lenders and customers are willing to rely. Further, in order to establish a global system, serious analysis must consider which laws and norms to prioritize and how to best communicate these to consumers and to lenders.
Lack of standardization can be a major weakness. In particular, because of different real property systems in each region, a data collection process may result in different information collected in different regions. So a process of cleaning and organizing the data to ensure that our business has systematic inputs would be critical to the blockchainâs success. Further, not only will a system to collect data need to remain uniform, but collected data and information need to be absolutely accurate.
Groups with property that is not easily valued will be incentivized in using our system. On the other hand, groups that can get a loan out of a bank in the traditional way may not have a strong appetite in our new system, unless our system provides more benefits such as lower interest rates. Since information on blockchain - zip code, address, record of title, housing valuation among other things can be sensitive, these customers who rely on the traditional systems might be reluctant to share this information with other people. Although data on blockchain is immutable and anonymous, these customers who do not necessarily have intimate knowledge about blockchain technology might be skeptical about the security of blockchain.
VI. Valuation
The total value of the United States housing market is $15 trillion. In the bear and bull case, we expect to create between $120bn and $600bn respectively. If risk is adequately limited, we can assume an aggregate 25% market share. We expect to be able to charge a 1% processing fee on the total value of each home transaction, which yields $600 billion in value created. However, unforeseen circumstances such as significant regulatory headwinds could result in a material adverse effect on the business. In this scenario we maintain the same 1% processing fee, but assign a 5% market share to reflect a more cautious outlook, which yields $120 billion in value created.
If the project is successful, the benefits of our solution include:
The securitization process of residential real estate properties becomes much more efficient with a digital auditable trail of ownership. Assets can become more globally accessible, opening up new markets for investors and creating additional liquidity for borrowers.
The immutable audit trail of ownership for the asset will show greater asset and loan transparency. Full transaction history and metadata on improvements made, amount of capital, past loan criteria and loan status are available on the blockchain for consumers to refer to, eliminating the information asymmetry in the traditional lending process.
Automated real estate transactions can be fully integrated with existing systems to reduce cost of capital and loan processing time by bypassing multiple layers of middlemen
Up-to-date information about the property can be used for better assessing the risks and lending decisions..
Real estate assets can be easily used within other blockchain-based financial products without the cumbersome manual overhead.
By aggregating more and more property data in a particular area, we will be able to build better predictive pricing models for the local housing market.