There ain’t no under-collateralized loans
A common criticism of the “successful” decentralized finance applications on ethereum (which is dominated by lending platforms) is that the loans are all over-collateralized. In Maker a DAI loan must be kept over 150% collateralized by ETH, with similar ratios required at compound bZx, etc. This is required so that, if the collateral value starts to drop to a point where there would not be an incentive to pay back the loan (below the loan amount + accrued interest) the collateral can automatically be sold to make the borrower whole and provide the expected return. This is all done automagically on the ethereum blockchain using smart contracts and the fact that it is only using assets native to the blockchain allows for this to work quite seamlessly. I would argue that this is still reasonably capital efficient because borrowers that have confidence in the value of their collateral can get some liquidity to run a . business or make purchases without selling an asset and triggering a taxable event (at least in the US).
But when people think of loans they think of mortgages and no one maintains 150% of the value of their outstanding mortgage in some account so that the bank can liquidate it if necessary. These loans are commonly referred to as under-collateralized and, now that we have seen sufficient traction with over-collateralized loans, the concept of under-collateralized loans has gained some attention. I would like to provide some context on this topic to hopefully guide some of the energy around this into productive pathways.
Mortgages are over-collateralized
When being reviewed for a mortgage the issuing bank reviews your pay history (much of this can be done automatically by finding patterns in your bank transactions), make sure you are gainfully employed (W2s), and review any other outstanding debts. They are trying to determine your ABILITY to continue making payments on the loan. Once you pass that test then the only other test is your WILLINGNESS to pay off the loan. Banks have the government and regulators on their side on this one and they know that if you dont pay they can go after other assets you may have or force you into bankruptcy. These things have a very tangible negative affect on your potential future financial position (lowering your credit score, essentially booting you down the financial hierarchy, social stigma, etc.). While there isn't an explicit formula for this, it goes into the considerations with the bank. On top of all that, the house itself, barring a dramatic price drop, is held as collateral and can always be taken back. Mortgages are heavily over-collateralized.
Loan = over-collateralized, everything else is a gift or grant
Family members and friends loan each other money all the time but it is still either a loan or a gift. The lender may not know what they are actually issuing upfront and it doesn't have to be all one or the other.
- Loan — If not repaid negative consequences (ostracized from a social group) occur
- Gift — If not repaid no negative consequences occur
Risk and Scale matter
Like in all financial transactions what matters is risk adjusted return. All of the inputs are taken and the lender creates probabilities for all of the possible outcomes. If they think the reward is worth the risks (or are ok with it being a gift in the long run) then they may choose to move forward. The bigger the loan (from perspective of lender) likely the less risk they are willing to take (for example that social pressures will not be enough to collateralize the full value of the loan). I have a hard time seeing truly under-collateralized loans (no economic, only social incentive to pay back) gaining much traction for the same reason that peer-to-peer loans (as introduced in dharma v1) didn't catch on — it is near impossible to scale pairing two sides that have very heterogeneous evaluation of risks and needs. Bringing everyone together under a protocol (like compound) or building around a specific use case (like MakerDAO) seem like the obvious options. The talk of a “digital credit score” has been a topic for about as long as the crypto industry has existed but to me this has always sounded like the concept for a black mirror episode (or basically the social credit score in china). This requires still bringing social or legal pressures of some sort back into the real world. And if you think it can be done anonymously or pseudo-anonymously I would say that you could be right up to a point. Every digital identity has a price so there will always be some point where someone will burn their identity if it makes economic sense.
Where to build
It is not all bleak though, I think there is great opportunity for innovation around new digital collateral types to use for collateral. What this new construct for tokenization and open digital markets has done is allow for the monetization of assets that had no model for valuation before. Acouple examples are
- $MAGIC, $ALEX, etc. — tokens representing someones time, can be used to book a meeting or some specific task to be done in the future by the token issuer at the request of the token holder
- $PEW — Can be exchanged by the token holder for a retweet by the token issuer
- DAO memberships — If a loan is not repaid, kick someone out of a DAO (rageKick in Moloch DAO)
- DAO Shares — Meta Cartel Ventures is a for profit DAO, its members will have shares representing a claim on assets held by the fund.
But I dont want any PEWs
The concern around using any of these assets for collateral would come down to valuation. While there are generally markets open for for those that are tokens on uniswap the liquidity is very thin which limits the scale at which they could be deployed. There are some possible solutions though
- Borrow from within trusted enclaves — If everyone in X group values your time at Y per unit and have confidence you will follow through (or have a legal contract supporting the asset) then you can borrow from that group directly. If joining daos becomes common practice, members borrowing from static assets from the dao and using personal assets + risk of being kicked out as collateral could work because all actions remain at the smaller scale.
- Buyers of last resort — A counterparty that values these tokens can post a public claim that they would buy X amount of the asset for Y value. That attestation (which would need to be actionable in an exchange somewhere) could then be used as collateral.
The goal with any of these solutions is to keep the evaluation of the assets at a local level but extend that evaluation to a global level to be tapped into by a protocol with liquidity. These new asset types and innovative funding mechanisms have enormous potential, for unlocking value and liquidity completely inaccessible before. If you would like any more information on anything discussed above (This was rant-typed after having versions of this convo with a number of different people) don’t hesitate to reach out.
twitter: @tompkins_jon
telegram: @jontom
