With scrutiny of crypto’s stability and security intensifying, lenders in the sector are increasingly under pressure to provide measures that provide investors with the best protection possible.
Michele Tucci, Chief Strategy Officer and Managing Director, Americas, at Credolab, explored how crypto lenders can learn from fiat currency firms when it comes to embracing risk mitigation strategies such as KYC, AML checks and more.
Fiat lenders have come a long way when it comes to the use of data analytics to reduce risk. They are transforming the banking and credit card sectors. The source of the data to analyse is not what you might imagine – standard credit files available from the bureaus.
A few years ago, academic economist Daniel Björkegren realised it is possible to predict if someone will pay back a loan based on their smartphone data. He found that a bank could have reduced defaults by 43% by selecting customers with specific metadata characteristics.
Since then, several global fintechs – we are working with a number ourselves – have developed credit assessments based on this so-called “alternative data”. And it’s not just smartphone metadata: some also make use of psychometric characteristics or even Twitter.
Crypto’s missing link
Strangely, the same things are not happening in crypto lending. The advances in risk mitigation being achieved using analytics by conventional banks and credit card issuers has not fully filtered through here.
It’s a puzzle, as crypto lenders have an even greater incentive to address risk, as blockchain-based currencies are well-known for their volatile nature.
Things are starting to change. We are working with crypto lenders who are keen to learn lessons about risk mitigation using behavioural analytics. But there is still a wide gulf.
Crypto lending in crisis
The crypto lending market expanded rapidly over the two years to 2021, dominated by major players such as Celsius Network, BlockFi, Binance, Bitfinex, Blockchain.com, Genesis Capital, and Nexo. These platforms were offering various services and products, such as interest-bearing accounts, crypto-backed loans, margin lending, and institutional lending.
Crypto lending operates on decentralised platforms, cutting out middlemen and enabling peer-to-peer transactions. This distinction brings both benefits and risks. On one hand, it offers lower barriers to entry, faster transaction times, and global accessibility. On the other, the platforms have to mitigate counterparty risk, and address the potential for fraud and hacking incidents.
This in turn has seen regulatory pressures intensify. Regulatory bodies worldwide are grappling with establishing clear guidelines for crypto lending, aiming to strike a delicate balance between consumer protection and fostering innovation.
In the US, proposed measures aim to regulate unhosted wallets by requiring customer identity verification and reporting for transactions over $10,000. India has introduced a bill to regulate crypto platforms, create a framework for official digital currency, and prohibit the use of private cryptocurrencies for payments.
In the EU, MiCA requires crypto-asset service providers to obtain authorization, comply with AML/CTF rules, and disclose customer identities. The UK government plans to regulate crypto platforms, including trading and lending, to protect consumers and ensure compliance with financial sanctions and reporting suspicious activity.
Failing to learn from the past
What’s odd here is that the challenges faced by crypto lenders are exactly the same as traditional lenders are addressing: cost of capital, risk management, and fraud. But some obvious lessons have not been carried across from one to the other.
As Ernest Lima, partner at XReg Consulting, told Cointelegraph in February: “The market collapse in the last year was spurred by poor practices in this space like weak or non-existing risk management and reliance on worthless collateral.”
Crypto lenders should surely be focusing on tight risk management controls instead of, as Celsius did, hiring a 24-year-old former pornstar as their head of institutional lending.
Crypto secrecy and fraud
The elephant in this particular room is that fraud may be especially worrisome on a crypto platform because the identities of the counterparties are unknown – given the pseudonymous nature of blockchain transactions. Since lending money is always easier than collecting money, how can a crypto lender collect from a user that has not been identified (or identified correctly)?
While it’s true that blockchain technology prioritises privacy and anonymity, there are several ways that crypto lenders can minimise the risk of fraud by borrowers. These include:
- KYC procedures (that are gradually being imposed anyway by regulators);
- Smart contract-based collateralization to ensure that borrowers provide sufficient digital assets as collateral for the loan;
- Peer-to-peer monitoring and transparency that leverages the decentralised nature of blockchain to create transparent and auditable lending transactions;
- And insurance from specialist providers for protection against potential fraud and default risks.
Credit scoring 2.0
What’s still missing, however, is a strong understanding of a borrower’s likelihood of repaying. Better, surely, to avoid default in the first place than have to deal with the repercussions afterwards?
Conventional lenders are now reducing first party fraud and defaults by leveraging behavioural data collected during the loan application process, onboarding, or any other interaction with a user.
By analysing a user’s device fingerprint, app installation patterns, IP address, mouse movements, typing patterns, session analysis, and other behavioural cues, a wealth of valuable insights can be gained to implement preventive measures that strengthen fraud detection and lower the cost of risk.
For example, by using behavioural analysis we developed a model for a Spanish lender aiming to approve more loans but with a lower bad rate. Such a model resulted in an increased approval rate (up 46.7%) and a simultaneous reduction in the bad rate (down 18.8%) due to a higher (+0.12) GINI coefficient (a measurement of the predictive power of a model).
How cryptolenders can deploy behavioural analytics
Finally, can all these lessons really be applied to the cryptosphere?
Yes. While the pseudonymous nature of blockchain transactions poses inherent challenges, crypto lenders can proactively implement these strategies to minimise the risk of fraud by borrowers.
As long as the lender has a mobile app requiring users to opt-in and consent their access to data, and they deploy a suitable SDK into their front-end (mobile app or website), it is possible to analyse this kind of data and provide marketing insights, anti-fraud checks and risk scores.
Keeping in mind that false positives may occur, as legitimate users could share some attributes with known risky and fraud patterns, there is no reason why behavioural data can’t fortify crypto loan application security – as is happening now for fiat lenders – ultimately paving the way for enhanced risk and fraud mitigation strategies.
