The Question of Underwriting in DeFi And What's Developing On That Front

So far 2020, has been a truly magnificent year for Decentralized Finance (DeFi). Since the beginning of the year, the total value locked in DeFi smart contracts has risen fivefold, with the vast majority of this growth coming from the last two months alone.

This was also marked by the emergence and explosive growth of new hopeful DeFi platforms, such as Compound and yearn.finance.

However, there is still a long road ahead for DeFi to become a veritable alternative to traditional finance and one of the most pressing pain points right now is underwriting.

In classical finance, taking out a loan means that you go to a bank, the bank assesses the probability of you being able to pay back the loan, and if the bank deems you creditworthy, you get your loan instantly, often without any collateral.

With DeFi, it is not the same. When you want to take out a loan, you have to provide a much larger amount of money as collateral than the amount you are getting.

Now, this does make sense if you take out a loan on an illiquid asset, for example, if you take out a mortgage on your house. Suppose you need some money to bridge over a minor financial emergency. 

Sure, you could sell your house and use the revenue to pay your bills, but most of us would prefer to keep living in our own house. Instead, you can easily use your house as collateral to get a credit line from your bank.

In contrast, cryptocurrencies are highly liquid assets, so there is no real need to take out a mortgage on your crypto holdings. 

Furthermore, there are instances where you need cash for spending, but you don’t want to liquidate your crypto because you expect the price of your assets to rise.

However, you are not really taking out a loan in the classical sense with DeFi, since you have to put down more capital as collateral than you are taking out, thereby increasing your risk exposure. 

This is more akin to taking out a leveraged position than a loan. In fact, you are likely better off simply selling a share of your holdings and then taking out a leveraged position through either margin trading or perpetual contracts, since this is usually cheaper and liquidation risks can be estimated more easily.

In order to move forward and become attractive to a wider audience, DeFi must take this important step and make real underwriting possible. Here are three promising developments on the road to underwriting in DeFi.

Identity

Naturally, when you lend money and take on the risk that your counterparty defaults on their debts, you want to know the identity of your counterparty. With DeFi, borrowers could be literally anyone. This can only work since as a lender, that counterparty risk is mitigated by the fact that the loan is overcollateralized. 

In order to enable underwriting, however, it is absolutely necessary to determine the borrower’s identity. Otherwise, the borrower could easily refuse to pay back their debt and get away with it.

When you know the identity of your counterparty, you can take legal action against them, which could result in the worst-case of personal insolvency for the borrower. 

While this may sound cruel and coldhearted, these steps are required to ensure credit discipline. With DeFi as it currently exists, there is no possible way to engage in litigation against a tardy borrower. Also, there is no such thing as personal insolvency or a bad credit rating in blockchain-based systems, as people can always generate a new private key and start over with a blank slate.

The bare minimum that is needed in order to make underwriting in DeFi possible is to have some kind of reputation system that is linked to each user’s real-world identity. Luckily, there are projects such as Shyft that work on establishing an identity layer for blockchain users. 

This has multiple use cases, such as user identification and authentication system for online services and a hard identification system for governmental interactions, or services that require KYC. With blockchain-based identity management solutions, users get full control over whom their personal data is shared with, in contrast to centralized authentication systems such as Google’s.

Shyft also establishes a reputation system based around their Relational Merit Token, which users earn for various activities, such as associating their blockchain address with KYC data and positive interactions with other attested Shyft users and other network nodes. Just as easily, Shyft can be incorporated into DeFi platforms to identify its users and assess their creditworthiness, by looking at their past transaction and credit history across the various DeFi products.

Risk Scoring

The next thing we need for underwriting is a scoring system that assesses a potential borrower’s creditworthiness and, by proxy, their default risk. This is very similar to the credit score you know from classical lending, but mapping this to the blockchain environment has proven to be difficult. 

Likely the easiest way to add your credit score to your blockchain identity is to simply link it with your traditional credit score through a credit bureau. However, the real innovation would be to infer a credit rating from your DeFi history or some other data. For example, Colendi is a microcredit platform that uses machine learning to establish a risk score from over 1000 personalized data points, gathered from a user’s behavior on the platform, the user’s smartphone, social media, purchase history, and other factors.

Thanks to this automated credit scoring without the requirement of prior credit history, it will become possible for the over 1.7 billion unbanked people around the world to apply for a loan and establish a credit history in the same breath. If systems like Colendi prove to be successful, they might become a basis for DeFi credit scoring as well.

On a similar note, the credit rating agency Morningstar plans to move into the blockchain sector by rating debt tokens, which are often issued by companies through security token offerings.

While this only provides credit ratings for companies at the moment, this approach could become another part of the puzzle for DeFi, especially when paired with personal identities and their prior credit history. 

Decentralized Insurance

Right now, the closest things we have in the DeFi sector to underwriting are the relatively few projects that work on decentralized insurance. This already involves a certain element of underwriting but in a strictly controlled environment.

For example, Etherisc is a project that is working on several DeFi insurance products, such as insurance against hurricanes, droughts, or floods. 

Their first product that is already live is a decentralized flight delay insurance. This works because it is relatively easy to estimate the risks that a flight will be delayed and data on flight delays are readily available.

In order to insure yourself, you pay a premium in Ether and when your flight is indeed delayed, you get an instant payout in Ether as well.

Another insurance project that aims to tackle a much harder problem is Nexus Mutual with its smart contract coverage. Users of the platform can insure themselves against losses incurred as a result of smart contract hacks.

The pricing of the policy is largely determined by risk assessors who stake their NXM tokens, thereby vouching for the security of the covered smart contract. 

Finally, Augur should be mentioned in this context. This platform allows anyone to create prediction markets on just about any subject. In theory, it is entirely possible to ask the market how likely a certain debtor is to default, thus getting both a market estimate and an opportunity to hedge against this case. 

Currently, this does not look like a viable option for retail DeFi borrowers, since these markets likely wouldn’t be very attractive. For large debtors, such as debt token issuers, it might however be possible to ask how high is the risk that the debtor will default. 

Conclusion

Right now, decentralized loans are the killer application for DeFi, as this is the use case for the two most high-ranking DeFi platforms, Maker and Compound. Together, they make up for almost 50% of all the value locked in DeFi alone. With that said, lending will not be very useful for most people without proper underwriting. Making this a reality is quite possibly the holy grail of DeFi and any platform that finds a satisfactory solution to the underwriting problem will surely become the highest-ranked DeFi platform by value locked in the platform.

There is however a long way to go until we get there and it will most certainly be a puzzle with many pieces. The first thing that is needed is to establish an identity layer on the blockchain. This is necessary in order to ensure borrowers follow up on their debt, as well as rejecting unreliable borrowers while making discounts for disciplined borrowers possible. 

Secondly, these identities must be linked with some sort of credit score in order to assess the risks involved with lending to a specific person. This score can be derived from traditional credit histories, histories about behavior on DeFi platforms, or other factors. The third large piece of the puzzle is risk transformation, which can be achieved through insurance or prediction markets. 

While we are still far away from achieving this goal, the first steps in the right direction are already underway.