Web3 credit scoring

Website for the book "Data Science for Web3". With monkeys and Van Gogh style. Add an index section, a call to action to purchase on Amazon and the author social networks.

Home Page

Explore

Modern Analytical Techniques in DeFi

To increase the risk assumption by lending protocols, the creation of a scoring system based on the history of an address could be explored. Here's a brief description of the traditional banking scoring systems: • FICO Score: Developed by Fair Isaac Corporation, the FICO score is the most widely used credit scoring system, ranging from 300 to 850. It is calculated based on a person's payment history, amounts owed, length of credit history, new credit, and types of credit used. Multiple versions exist for different lending purposes such as credit cards, mortgages, and personal loans. • VantageScore: This scoring system also ranges from 300 to 850 but employs a slightly different methodology, focusing on recent credit behaviors and accommodating consumers affected by natural disasters. Both systems have access to sensitive personal and transactional information. They considered for the scorings they produce medical collections accounts up to 2022. The blockchain ethos and the opportunity for undercollateralized lending give us the chance to reevaluate these traditional scoring criteria, adapting them to the limited information available to a lending smart contract, brought by the "0x" address of the potential lender. Potential data points could include: • Age of the address: Indicates how long the participant has been active in the ecosystem, reflecting their experience. • Total assets: Measures the value of the participant’s assets within reportable markets, providing a snapshot of their financial capacity. • Quality of held assets: Differentiates assets by risk levels; owning unknown LP tokens versus BTC, for example. • Liquidation events: Frequent and substantial liquidations may indicate poor financial health and a higher risk of default. • Quality of Interacted Protocols: Given nearly a decade since Ethereum’s inception and the advent of DEFI protocols, certain protocols and practices have earned a reputation. • Account type: Reflects the consistency and source of income, including the regular receipt of salaries and remittances through blockchain systems, which can demonstrate financial stability and trustworthiness.

Rootstock

The development of an undercollateralized loan market in DeFi may necessitate a shift in existing protocol mechanisms that use collateralization ratios to determine APY levels. This adjustment, coupled with the implementation of a reputable address scoring system, could potentially unlock billions in TVL currently held idle in lending protocols. Moreover, by democratizing access to credit data, the dominance of a few companies over sensitive financial information could be challenged. This would pave the way for a system where borrowers, lenders, and regulators can transparently verify transactions and credit histories. Additionally, it could overcome the geographical limitations inherent in traditional scoring systems, which often restrict the flow of funds to areas with fewer financial opportunities but great capital returns. If the internet knows no geographic boundaries, why should money for productive investments be constrained? Implementing such a system within DeFi smart contracts would enable real-time, automated credit scoring for each borrower (address), enhancing transparency and trust across the financial ecosystem and aligning with the core principles of blockchain technology

Blog Posts

Twitter

CR Academia

Blog

Current Situation: Overcollateralized Loans

Packt

References

RootstockLabs

Home

Blockchain Data Science

Website for the book "Data Science for Web3". With monkeys and Van Gogh style. Add an index section, a call to action to purchase on Amazon and the author social networks. https://www.packtpub.com/product/data-science-for-web3/9781837637546

Buy Book

The advent of blockchain analytics platforms like Dune Analytics, Flipside Crypto, Footprint Analytics, OrtegueAI and many others makes it possible to perform such analyses on a multichain basis. Further enhancing the integrity of these scoring systems, advanced machine learning techniques, such as those employed by Coinbase using graph theory, help prevent tampering and provide a robust framework for credit evaluation. Numerous tools, including Nansen and Arkham, are creating robust databases that enable almost real-time updates to models. Excellent educational resources on these analytical techniques include YouTube channels of the analytics platforms mentioned, Twitter accounts of experts like @0xKofi, @0xBoxer, and @andrewhong5297. Additionally, notable literature includes "Investigating Cryptocurrencies" (Understanding, Extracting, and Analyzing Blockchain Evidence) by Nick Furneaux, published by Wiley, and the chapter on "Data Science Techniques for Cryptocurrency Blockchains" in a book by Innar Liiv, published by Springer. Furthermore, my book, "Data Science for Web 3," offers a pedagogical approach to the intersection of data science and blockchain technology.

Home

Website for the book "Data Science for Web3". With monkeys and Van Gogh style. Add an index section, a call to action to purchase on Amazon and the author social networks. Website for the book "Data Science for Web3". With monkeys and Van Gogh style. Add an index section, a call to action to purchase on Amazon and the author social networks.

• European Central Bank. (n.d.). Europe Loan to Deposit Ratio. Retrieved from https://data.ecb.europa.eu/main-figures/banks-balance-sheet/deposits • TAB Insights. (n.d.). South American and European Banks Post Highest Loan-to-Deposit Ratios. Retrieved from https://tabinsights.com/article/south-american-and-european-banks-post-highest-loan-to-deposit-ratios • Federal Reserve Board. (2023). Supervision and Regulation Report: Banking System Conditions. Retrieved from https://www.federalreserve.gov/publications/2023-may-supervision-and-regulation-report-banking-system-conditions.htm • FICO. (n.d.). Explanation of factors influencing FICO scores. Retrieved from: https://www.ficoscore.com/ficoscore/pdf/Understanding-FICO-Scores.pdf • VantageScore Solutions. (n.d.). https://www.vantagescore.com/ • NerdWallet. (n.d.). Medical Bills on Credit Report. Retrieved from https://www.nerdwallet.com/article/finance/medical-bills-on-credit-report • Macrotrends. (n.d.). Fed Funds Rate Historical Chart. Retrieved from https://www.macrotrends.net/2015/fed-funds-rate-historical-chart

An avenue to explore:

Posts and articles related with Data Science and Web 3

Future Directions and Market Impact

Web3 for Data Science Blog

The case for Web3 credit scoring

Well-established on-chain lending protocols predominantly adhere to an overcollateralized framework, which is sensible given by the anonymity inherent in Web3. Overcollateralized lending protocols allow users to borrow assets by providing more collateral than the value of the borrowed assets. In contrast, undercollateralized loans—issued with less collateral than the loan amount, or none at all—are common in traditional finance due to stringent KYC measures. Some experts suggest that the overcollateralized methodology, coupled with the inherent volatility of crypto prices, does not effectively facilitate long-term productive investments. In fact, data shows that the loans are limited to use cases related with trading and exposure to other tokenized assets through leverage, while long-term productive investments are neglected. Another interesting aspect refers to the utilization of the capital locked in these overcollateralized protocols. To evaluate that we can refer to the Utilization Ratio (UR) that is calculated by dividing the borrowing volume by the total value locked (TVL): Where: Borrowing Volume: The amount of funds that have been borrowed in a certain timeframe. TVL: Total Value Locked of such lending protocol in a certain timeframe. If we carry out the calculation above, we can conclude that the utilization of funds locked in DeFi protocols remains notably low. Some notable findings include: • As of April 2024, the average UR on platforms like Compound and AAVE, as reported on DefiLlama, stands at 44%. Considering data for the closed month of April on Token Terminal, the ratio between deposits and active loans was 3% for Compound and 34% for AAVE. • JustLend, the leading protocol on the low-fee Tron chain, reports only 2% utilization. • In Bitcoin sidechains, only two chains are actively reporting these metrics. Utilization is more effective in Tropykus, which is active in Rootstock with 30%, while Merlin’s Avalon Finance reports only 3%. Stacks’ Zest protocol is still in closed beta. We can compare the UR with the loan-to-deposit ratio from traditional banking that is a measure of a bank’s risk management and profitability. If do so, these figures show that DeFi is not leveraging its capital as efficiently as traditional banks, which often maintain a loan-to-deposit ratio significantly higher—ranging from 72% in the US, 94% in Europe, and even higher in South America. This represents a significant opportunity cost for the protocols, as their deposits sit idle in vaults, generating no revenue in a context of high interest rates.