We have been working hard on choosing the most objective and fair criteria. We would be excited if you read our article where we laid out what we’ve achieved so far.
It is true that the lack of knowledge and information for the borrowers in the Defi segment leads to the necessity of over-collateralisation, however the proposed solution to the problem in the article is far away from actual implementation. Although on-chain credit scores are a possibility in the far future, today they are but just a dream.Moreover, there will be no real proof that credit scores computed off-chain will be the result of robustcredit risk modelling and sound quantitative risk management practices.
The use of over-collateralization in the beginning of any project is the most basic form of risk management when there is no source of a trustable measure of creditworthiness. Therefore, if we consider a P2P lending project, the main issue becomes how to create a credit risk assessment framework (and as a consequence credit models and scores) that will serve two main purposes:
- first to provide the lender with quantification of credit risk that can be trusted and
- second to show the borrower that he has been assessed objectively by unbiased statistical methods.
These two elements will provide the necessary anchoring to the negotiation between the lender and the borrower. They could obviously reach an agreement for the loan at a different price level. They could bargain and speculate, but at the end of the day, theme asure of a fair credit risk score (and corresponding risk premium) would be the assessment of the majority of market participants. As long as the offer meets the demand, it signals good efficiency of a risk-based pricing agreed by the market.
The introduction of viable credit risk scores would thus make it possible for real economy businesses with potential to be financed at fair prices in the DeFi space and not to be ripped off by crypto market enthusiasts seeking to ride the next bull market and get an abnormal profit.
In summary, risk scores need to be trustworthy, robust and based on proven risk management practices. At Credefi we have designed a bridge between the current Defi marketplace and the future by merging the traditional risk management tools with the decentralised decision making. Our framework relies on sound technical infrastructure able to handle big data computations and real time risk scoring, and proven risk management practices Both will be continuously applied over the growing lending activity on the platform to arrive at a risk-based pricing that will be representative for all markets we operate in.
The credit score system implemented on the platform applies both quantitative and qualitative criteria.
Qualitative criteria:
- Know your customer (KYC)
- Professional experience
- Age of the borrower
- Reputation and management’s expertise
- Credit history and liabilities in public registries
- Proof of ownership and right to representation
Quantitative criteria:
Credit Policy Rules. Hard credit criteria that set exposure thresholds and minimum levels for specific financial ratios based on the credit product type and borrower’s segment
Application Scorecards. Statistically developed based on our historical credit portfolio
The combination of the criteria above allows for the application of a model that provides to the lender a statistically accurate probability of successful repayment of the loan they choose to finance. Through the collection of historical data on the credit deals done on the platform the models are adapted,statistically validated over time and refined with the increasing amount of data in order to produce risk-based pricing that will be accepted by the market as fair and objective. Only with this approach we can expect to apply the automated truly decentralized lending in the future. The collection and management of data is the only means by which the model can evolve and enable true decentralized decision making in the future.
Prev. article 